Browse the corpus

Shared decision-making (SDM) is an increasingly referenced and lauded approach to medical decision-making in Western countries, and its use is also spreading to other contexts including China1 and Malaysia.2 SDM is a partnership between patients (and families, where appropriate) and clinicians that considers patients’ values and preferences alongside medical evidence to make the best decisions for a given patient in a specific scenario. In Western countries, arguments for SDM often focus on principles of autonomy and self-determination, particularly in the setting of clinical uncertainty. For many decisions, there is not a ‘right’ answer; SDM enables patients and families to choose the best option for them based on individual values, goals and considerations such as mechanism of administration, cost and side effect profile.

omy and self-determination, particularly in the setting of clinical uncertainty. For many decisions, there is not a ‘right’ answer; SDM enables patients and families to choose the best option for them based on individual values, goals and considerations such as mechanism of administration, cost and side effect profile. Arguments supporting the use of SDM go beyond these principles, however. Research suggests that SDM results in improved knowledge/understanding, satisfaction and trust,3 which are hoped to also lead to better health outcomes. Decision aids (DAs)—tools that guide patients, families and clinicians through the SDM process—increase knowledge, lower patients’ decisional conflict, reduce patient passivity in decision-making and the number of patients unable to decide, and result in more decisions for less-aggressive care.4 Research to date has focused more on these short-term outcomes of SDM rather than its long-term impact on health outcomes.3 5 6 A conceptual model of SDM, however, suggests that SDM can result in short-term, mid-term and long-range benefits for patients, clinicians/other healthcare professionals, organisations and healthcare systems, including improved decision-making, satisfaction, patient experiences, trust, health outcomes, cost-effectiveness and resource utilisation, along with decreases in litigation and professional burnout.6

and long-range benefits for patients, clinicians/other healthcare professionals, organisations and healthcare systems, including improved decision-making, satisfaction, patient experiences, trust, health outcomes, cost-effectiveness and resource utilisation, along with decreases in litigation and professional burnout.6 Approaches to SDM SDM likely best occurs in the setting where a physician and a patient have an established relationship such that the physician knows the patient’s values and goals, informing how options are described and weighed during SDM.7 Even in the acute setting, such as the emergency department or during hospitalisation, understanding a patient’s background (eg, employment) and values prior to formal decision-making can provide important context for decisions. There are multiple models for SDM with different numbers of outlined steps.8–10 When a clinical decision is needed, SDM starts by engaging patients and key supports in the process (table 1). This step requires the clinician to understand who the patient desires to participate in decision-making, such as a family member. In situations where a patient is incapacitated and unable to participate, SDM occurs with the surrogate decision maker. Even if a patient ultimately desires to defer the decision to a physician or family member, it is important to actively engage him/her in the SDM process. Table 1 Steps to shared decision-making

There are multiple models for SDM with different numbers of outlined steps.8–10 When a clinical decision is needed, SDM starts by engaging patients and key supports in the process (table 1). This step requires the clinician to understand who the patient desires to participate in decision-making, such as a family member. In situations where a patient is incapacitated and unable to participate, SDM occurs with the surrogate decision maker. Even if a patient ultimately desires to defer the decision to a physician or family member, it is important to actively engage him/her in the SDM process. Table 1 Steps to shared decision-making Step 1 Engage patients (and other decision makers, if appropriate) in the decision-making process Step 2 Describe the decision and the options available, including each option’s potential benefits and risks, how the options are different and what is unknown about the options (the uncertainty) Step 3 Further assess the patient’s values and goals, specifically as they relate to the available options Step 4 Make the decision together Once patients and families are engaged, the second step (table 1) is specifically describing the decision and outlining the different options. In describing the different options, clinicians should use the available medical evidence to inform patients about the potential benefits and risks. It is also important to highlight when there is something unknown about the options (uncertainty) and to describe how the options are distinct. Differences between options include potential benefits and harms and considerations such as cost and invasiveness. This discussion should be individualised—for example, the balance of benefits and harms of clopidogrel for secondary stroke prevention will be different between a person with a prior history of bleeding gastric ulcers and a person with no such medical history.

and harms and considerations such as cost and invasiveness. This discussion should be individualised—for example, the balance of benefits and harms of clopidogrel for secondary stroke prevention will be different between a person with a prior history of bleeding gastric ulcers and a person with no such medical history. Once the evidence is presented, the intersection between the options and the patient’s values and goals is explored (step 3, table 1). If a patient’s circumstances, values and goals were known prior to the initiation of SDM, the presentation of the options in step 2 should occur in that context. For example, if a patient lives alone and prioritises continued independence, the potential benefits and risks of each intervention are specifically presented with a reference to the likelihood of maintaining independence (eg, from successful treatment) or putting it at risk (eg, due to side effects). Regardless of prior knowledge of patient values, in step 3 a patient’s values as they specifically relate to the decision are explored. What is most important to the patient in this situation—expected functional recovery? Amount of risk? Cost?

from successful treatment) or putting it at risk (eg, due to side effects). Regardless of prior knowledge of patient values, in step 3 a patient’s values as they specifically relate to the decision are explored. What is most important to the patient in this situation—expected functional recovery? Amount of risk? Cost? Finally, a decision is made. Ideally, the patient makes the decision with the help of those friends or family members whom he or she has chosen for involvement. Sometimes patients prefer clinicians to make final decisions. In these circumstances, patients often still express a desire for participating in SDM,11 but request that the clinician select the best strategy after discussion. When the patient defers to clinicians, the burden is on the physician to target the decision to the patient’s stated values and goals, thus still using SDM to make the best individualised decision for that patient in that circumstance.

in SDM,11 but request that the clinician select the best strategy after discussion. When the patient defers to clinicians, the burden is on the physician to target the decision to the patient’s stated values and goals, thus still using SDM to make the best individualised decision for that patient in that circumstance. In many circumstances, particularly those encountered in the outpatient setting, re-evaluation is an important component of SDM. Anticipated ongoing benefits and risks may change based on the development of comorbidities; patients’ values and priorities may change based on their experiences with a medication or shifting life circumstances. In the outpatient clinic, for example, decisions regarding anticoagulation for secondary stroke prevention in the setting of atrial fibrillation should be reassessed over time. This is in contrast to certain acute stroke decisions, such as those regarding tissue plasminogen activator (t-PA), where the window for SDM is small, with little opportunity for re-evaluation.

egarding anticoagulation for secondary stroke prevention in the setting of atrial fibrillation should be reassessed over time. This is in contrast to certain acute stroke decisions, such as those regarding tissue plasminogen activator (t-PA), where the window for SDM is small, with little opportunity for re-evaluation. Values and goals Within SDM, values tied clearly to diagnostic or therapeutic options such as efficacy, toxicity, quality of life, convenience and cost are often emphasised.12 Other values and goals may also inform patient decisions, however, and these can be critical to SDM.13 Global values reflect life priorities or beliefs, which may be religious or cultural in origin; these values impact all decisions. Global values can also represent value traits, such as risk aversion or a desire to try the ‘new’ thing, which also influences approaches to decision-making.13 External values reflect a patient’s choice to consider others’ values and preferences when making a decision.13 This occurs in Western cultures but may be more important in other cultures, such as in mainland China where family involvement in decision-making can reflect mutual benevolence and the Confucian ideal of family harmony,14 or places like Pakistan where the norm is family–doctor–patient triadic decision-making.15 Finally, situational values reflect context-specific factors that influence a decision differently now than in the past or future, such as an upcoming event (eg, a wedding) that may impact how long a patient is willing to remain in the hospital or rehabilitation.13

rm is family–doctor–patient triadic decision-making.15 Finally, situational values reflect context-specific factors that influence a decision differently now than in the past or future, such as an upcoming event (eg, a wedding) that may impact how long a patient is willing to remain in the hospital or rehabilitation.13 The role of evidence-based medicine Although SDM is often emphasised in discussions of personalised and patient-centred care, it is critical to note that this process relies on evidence-based medicine. Evidence-based medicine is foundational to step 2 of SDM (table 1), where patients, families and clinicians discuss the evidence (or uncertainty/lack of evidence) of benefits and harms for each potential option. It is only by knowing the available evidence that patients and families can make informed decisions. To present this evidence, clinicians can reference original research or use tools such as DAs, systematic reviews or evidence-based guidelines, each of which summarises known evidence in response to a specific question or choice. SDM and stroke Most recent publications on SDM in stroke care focus on oral anticoagulation for stroke prevention in atrial fibrillation.16–20 This is a decision where SDM clearly plays an important role given differences in individual risks based on comorbidities, multiple options with different potential benefits, risks, costs and time requirements (eg, for international normalized ratio [INR] monitoring), and obvious value assessments relating to potential outcomes such as stroke and bleeding.

lays an important role given differences in individual risks based on comorbidities, multiple options with different potential benefits, risks, costs and time requirements (eg, for international normalized ratio [INR] monitoring), and obvious value assessments relating to potential outcomes such as stroke and bleeding. Less research exists for other decisions relating to stroke, and currently available DAs may not meet decision aid standards.21 A 2013 review of patient tools designed for decision-making regarding thrombolytic treatment identified that available tools lacked key development stages, presented outcome probabilities poorly and failed to completely describe potential benefits and risks.22 Subsequently, the COMPuterized decsion Aid for Stroke thrombolysiS (COMPASS) tool, a computerised DA for thrombolysis in acute stroke, was developed with clinicians, patients, families and modelling techniques. Using the tool took a median time of only 2.8 min in early pilot testing, but additional study is required.23 24 DAs have particular potential for improving care in this emergent setting, where SDM is challenged by the time limitations for effective thrombolysis, the need to engage patients and families and convey knowledge in the context of the shock and effects of an acute stroke, and the need to incorporate personal values into a decision that relies heavily on physician expertise.25

mergent setting, where SDM is challenged by the time limitations for effective thrombolysis, the need to engage patients and families and convey knowledge in the context of the shock and effects of an acute stroke, and the need to incorporate personal values into a decision that relies heavily on physician expertise.25 Barriers to SDM Research on barriers to SDM is largely conducted in Western contexts. Identified barriers to SDM include physician and patient attitudes towards SDM,26 27 lack of familiarity and experience with SDM,26 lack of continuity of care,27 physician knowledge regarding evidence,26 27 the physician–patient relationship,27 insufficient explanations,27 use of medical terminology,27 the ability of patients and families to understand and use health-related information (health literacy),27 28 lack of resources26 27 and time.10 26 27

ck of continuity of care,27 physician knowledge regarding evidence,26 27 the physician–patient relationship,27 insufficient explanations,27 use of medical terminology,27 the ability of patients and families to understand and use health-related information (health literacy),27 28 lack of resources26 27 and time.10 26 27 Research regarding SDM in China is extremely limited, but identified barriers overlap with those described elsewhere and include lack of resources, time, physician communication skills, patient–physician relationships, the health literacy of patients and families, and unrealistic patient and family expectations.1 Despite these barriers, a recent study found it feasible to implement the use of a statin DA for cardiovascular risk reduction in two teaching hospitals in Northern China.29 Additional barriers identified in this study included lack of privacy for uninterrupted discussions, family dominance within some encounters, lack of applicability of data within Western DAs to Chinese contexts, and low health literacy requiring additional cardiovascular education in order for patients to effectively use the tool.29

riers identified in this study included lack of privacy for uninterrupted discussions, family dominance within some encounters, lack of applicability of data within Western DAs to Chinese contexts, and low health literacy requiring additional cardiovascular education in order for patients to effectively use the tool.29 Facilitators of SDM The most commonly described facilitators of SDM are clinician-related: clinician motivation and the perception that SDM improves the clinical process and patient outcomes.26 Patient-identified facilitators include continuity of care, good relationships between patients and clinicians, trust, adequate time, engagement of various members of the healthcare team (eg, nurses, in addition to doctors), a sense of partnership, encouragement of patients to participate and ask questions, the provision of sufficient information, use of plain language, and patient engagement and ownership in the process.27 DAs are practical facilitators of SDM, although they are insufficient on their own and have some limitations.27 DAs are useful for addressing barriers to SDM, such as lack of familiarity with SDM, physician knowledge regarding evidence, and provision of sufficient and understandable information, as DAs walk clinicians, patients and families through the SDM process and describe the medical evidence in plain language, often using visual aids.

ul for addressing barriers to SDM, such as lack of familiarity with SDM, physician knowledge regarding evidence, and provision of sufficient and understandable information, as DAs walk clinicians, patients and families through the SDM process and describe the medical evidence in plain language, often using visual aids. Although there are few published stroke-related DAs, approaches exist for helping clinicians develop tools for commonly encountered decisions, such as Option Grids.30 In the absence of formal tools, other patient education materials can be helpful in promoting step 2 of SDM, such as those available through neurology and stroke organisations. Ideal tools will be culturally and context-sensitive, something of particular importance as SDM spreads to non-Western contexts.1 29

n Grids.30 In the absence of formal tools, other patient education materials can be helpful in promoting step 2 of SDM, such as those available through neurology and stroke organisations. Ideal tools will be culturally and context-sensitive, something of particular importance as SDM spreads to non-Western contexts.1 29 Conclusions SDM is an increasingly promoted approach for patients, families and clinicians to partner to make the best medical decisions for each individual in a particular moment by using the best medical evidence. Although long-term benefits for patients, families, clinicians, hospitals and health systems have yet to be explored, SDM has known benefits on decision-making and satisfaction and has the potential for improving other outcomes as well. Every decision within stroke care has potential for improvement with SDM, whether relating to thrombolysis, goals of care, diagnostic assessments, rehabilitation strategies or secondary stroke prevention. Although SDM is necessarily context-specific, development of DAs for commonly faced decisions within vascular neurology may improve stroke care. Future research is needed regarding the cultural elements of SDM in general and also within the field of stroke. Contributors: The review was requested by Dr David Zheng Wang. The review was written by MJA. Competing interests: None declared. Provenance and peer review: Commissioned; internally peer reviewed.

Introduction Blood pressure (BP) is elevated in 75% of patients presenting with acute ischaemic stroke1 and is associated independently with poor clinical outcomes.2 3 Lowering elevated BP appears safe in acute ischaemic stroke, but has failed to show clinical benefit.4 There is a specific concern regarding BP lowering in the 15% of patients with significant carotid stenosis in whom cerebral perfusion may be compromised and where reducing BP might extend the ischaemic core and potentially worsen outcome.5 Data on BP reduction in severe ipsilateral or bilateral carotid stenosis are limited, although a meta-analysis found that lower BP was associated with an increased rate of stroke recurrence in bilateral carotid stenosis.6 In the ultra-acute prehospital and acute hospital situation, information on carotid stenosis is often not available and it is unclear whether BP lowering is safe in this group of patients with stroke. The Efficacy of Nitric Oxide in Stroke (ENOS) Trial assessed the safety and efficacy of transdermal glyceryl trinitrate (GTN) and of continuing prestroke antihypertensives in 4011 patients with acute stroke.7 Although GTN lowered BP by 7/3.5 mm Hg at day 1, GTN did not influence functional outcome at 90 days.7 However, when administered within 6 hours of stroke onset GTN improved several clinical outcomes.8 The aim of the current preplanned substudy9 was to assess the safety and efficacy of BP lowering on clinical outcomes in patients with acute ischaemic stroke and carotid stenosis.

influence functional outcome at 90 days.7 However, when administered within 6 hours of stroke onset GTN improved several clinical outcomes.8 The aim of the current preplanned substudy9 was to assess the safety and efficacy of BP lowering on clinical outcomes in patients with acute ischaemic stroke and carotid stenosis. Methods Details pertaining to the ENOS trial protocol, statistical analysis plan, baseline characteristics and main trial results have been published.8 10–12 In summary, the ENOS Trial recruited 4011 patients with acute stroke within 48 hours of onset with high systolic BP (140–220 mm Hg) and randomised them to GTN 5 mg patch or no patch for 7 days. Those participants taking antihypertensive medication prior to their index event were also randomised to continue or stop these drugs for 7 days. Known carotid stenosis was not an exclusion criterion. Written consent to participate was given by patients or relatives/carers in those who lacked capacity. The ENOS Trial was registered (ISRCTN99414122). Patient and public involvement Patients and public were not involved in the development of this preplanned secondary analysis of the ENOS trial.

Methods Details pertaining to the ENOS trial protocol, statistical analysis plan, baseline characteristics and main trial results have been published.8 10–12 In summary, the ENOS Trial recruited 4011 patients with acute stroke within 48 hours of onset with high systolic BP (140–220 mm Hg) and randomised them to GTN 5 mg patch or no patch for 7 days. Those participants taking antihypertensive medication prior to their index event were also randomised to continue or stop these drugs for 7 days. Known carotid stenosis was not an exclusion criterion. Written consent to participate was given by patients or relatives/carers in those who lacked capacity. The ENOS Trial was registered (ISRCTN99414122). Patient and public involvement Patients and public were not involved in the development of this preplanned secondary analysis of the ENOS trial. Carotid stenosis Clinical information on carotid stenosis was collected by investigators during the participant’s index event admission. Clinical imaging using either carotid Doppler, MR angiography or CT angiography was performed as per local protocol. Investigators entered the % of stenosis of both left and right internal carotid arteries using North American Symptomatic Carotid Endarterectomy Trial (NASCET) criteria where available (online supplementary material).13 Data were checked and validated but no central adjudication of carotid imaging was performed. 10.1136/svn-2019-000232.supp1Supplementary data

Carotid stenosis Clinical information on carotid stenosis was collected by investigators during the participant’s index event admission. Clinical imaging using either carotid Doppler, MR angiography or CT angiography was performed as per local protocol. Investigators entered the % of stenosis of both left and right internal carotid arteries using North American Symptomatic Carotid Endarterectomy Trial (NASCET) criteria where available (online supplementary material).13 Data were checked and validated but no central adjudication of carotid imaging was performed. 10.1136/svn-2019-000232.supp1Supplementary data Participants who had a final diagnosis of ischaemic stroke and who had carotid data available were included in this substudy. Grades of carotid stenosis were defined as follows:Unilateral carotid stenosis ipsilateral to the symptomatic hemisphere:13 <30%, 30–<50%, 50–<70%, ≥70%. Bilateral carotid stenosis (% for both carotid arteries): <30%, 30–<50%, ≥50%. Haemodynamic measures BP and heart rate were measured peripherally at baseline (three measurements) and on days 1–7 (two measurements/day), using validated automated equipment (Omron 705 CP).14

Participants who had a final diagnosis of ischaemic stroke and who had carotid data available were included in this substudy. Grades of carotid stenosis were defined as follows:Unilateral carotid stenosis ipsilateral to the symptomatic hemisphere:13 <30%, 30–<50%, 50–<70%, ≥70%. Bilateral carotid stenosis (% for both carotid arteries): <30%, 30–<50%, ≥50%. Haemodynamic measures BP and heart rate were measured peripherally at baseline (three measurements) and on days 1–7 (two measurements/day), using validated automated equipment (Omron 705 CP).14 Clinical outcomes The primary outcome in ENOS was functional outcome measured using the modified Rankin Scale15 (mRS, a seven-level categorical scale where 0=independent and 6=dead) at 90 days. Day 90 secondary outcomes included disability (Barthel Index),16 mood (Zung Depression Scale),17 quality of life (health utility status calculated using the European Quality of Life 5-dimensions three levels version, and Visual Analogue Scale)18 and cognition (telephone mini-mental state examination,19 modified Telephone Interview for Cognition Scale [TICS-M]20 and verbal fluency). Patients who had died by day 90 were assigned a worst score for the outcomes. Safety data were collected on all-cause mortality at day 90, early neurological deterioration (a minimum 5-point reduction overall or >2 points reduction in the consciousness domain from baseline to day 7 on the Scandinavian Stroke Scale [SSS]), symptomatic hypotension, hypertension or headache by day 7. The National Institute of Health Stroke Scale (NIHSS) was calculated from SSS.21 Day 90 outcomes were recorded by trained blinded assessors via telephone at national coordinating centres.

s domain from baseline to day 7 on the Scandinavian Stroke Scale [SSS]), symptomatic hypotension, hypertension or headache by day 7. The National Institute of Health Stroke Scale (NIHSS) was calculated from SSS.21 Day 90 outcomes were recorded by trained blinded assessors via telephone at national coordinating centres. Statistical analysis In line with the ENOS Trial statistical analysis plan and statistical analyses performed in the primary publication, intention-to-treat analysis of data was carried out.11 Data are number (%), median [IQR] or mean (SD). Baseline characteristics between grades of carotid stenoses were assessed using the χ2test for categorical variables and one-way analysis of variance for continuous variables.

ses performed in the primary publication, intention-to-treat analysis of data was carried out.11 Data are number (%), median [IQR] or mean (SD). Baseline characteristics between grades of carotid stenoses were assessed using the χ2test for categorical variables and one-way analysis of variance for continuous variables. Associations between carotid stenosis grades and outcomes were assessed using multiple linear regression, ordinal logistic regression or binary logistic regression with adjustment for baseline prognostic covariates including age, sex, baseline mRS score, history of previous stroke, history of diabetes mellitus, total anterior circulation stroke, nitrate use, baseline SSS, thrombolysis, feeding status, time to randomisation and baseline systolic BP. Analyses involving the whole population were also adjusted for treatment allocation. Associations between BP change from baseline to day 1 and outcome across degrees of carotid stenosis were assessed per 10 mm Hg reduction in BP. Interaction p values were obtained by adding an interaction term to statistical models. Data are mean difference or OR and associated 95% CIs, with significance defined as p≤0.05. Analyses were performed using SPSS V.24 (Chicago, Illinois, USA).

ss degrees of carotid stenosis were assessed per 10 mm Hg reduction in BP. Interaction p values were obtained by adding an interaction term to statistical models. Data are mean difference or OR and associated 95% CIs, with significance defined as p≤0.05. Analyses were performed using SPSS V.24 (Chicago, Illinois, USA). Results Of 4011 participants, 2023 (50.4%) had a final diagnosis of ischaemic stroke and carotid imaging data (GTN 1002 vs no GTN 1021, table 1). One thousand three hundred and nineteen (32.9%) patients with ischaemic stroke did not have carotid imaging, typically those with more severe stroke (carotid imaging: SSS 36.6 [12.4], no imaging: SSS 30.6 [13.7], p<0.001); there was no relationship between country of enrolment and whether carotid imaging was performed (data not shown). Of 2023 participants with carotid data, 148 (7.3%) had 50–<70% ipsilateral stenosis, 213 (10.5%) had ≥70% ipsilateral stenosis and 97 (4.8%) had ≥50% bilateral stenosis. Age and rates of treated hypertension, diabetes mellitus, atrial fibrillation, history of transient ischaemic attack, ischaemic heart disease and peripheral arterial disease differed across grades of ipsilateral carotid stenosis. Those with higher degrees of ipsilateral carotid stenosis were more likely to be male, current smokers, have more severe strokes with higher NIHSS and lower Glasgow Coma Scale Scores, fewer cardioembolic and small vessel disease-related strokes, and more received thrombolysis treatment (table 1). As compared with patients with carotid imaging data, those without had a worse functional outcome at day 90: mRS 4 [3] versus 3 [3], (OR 1.76, 95% CI 1.54 to 2.01, p<0.001).

lower Glasgow Coma Scale Scores, fewer cardioembolic and small vessel disease-related strokes, and more received thrombolysis treatment (table 1). As compared with patients with carotid imaging data, those without had a worse functional outcome at day 90: mRS 4 [3] versus 3 [3], (OR 1.76, 95% CI 1.54 to 2.01, p<0.001). Table 1 Baseline characteristics of all patients with ischaemic stroke with carotid data and by ipsilateral carotid stenosis

lower Glasgow Coma Scale Scores, fewer cardioembolic and small vessel disease-related strokes, and more received thrombolysis treatment (table 1). As compared with patients with carotid imaging data, those without had a worse functional outcome at day 90: mRS 4 [3] versus 3 [3], (OR 1.76, 95% CI 1.54 to 2.01, p<0.001). Table 1 Baseline characteristics of all patients with ischaemic stroke with carotid data and by ipsilateral carotid stenosis All IS GTN No GTN Continue Stop Stenosis <30% Stenosis 30–<50% Stenosis 50–<70% Stenosis ≥70% P value Number of patients 2023 1002 1021 534 525 1431 224 148 213 Age (years) 69.1 (11.4) 68.8 (11.3) 69.4 (11.5) 71.6 (10.5) 70.9 (10.5) 68.3 (11.6) 71.2 (10.7) 73.3 (9.9) 68.9 (10.8) <0.001 Sex, male (%) 1193 (59.0) 599 (59.8) 594 (58.2) 286 (53.6) 283 (53.9) 817 (57.1) 141 (62.9) 91 (61.5) 141 (66.2) 0.036 Premorbid mRS>1 (%) 209 (10.3) 95 (9.5) 114 (11.2) 71 (13.3) 64 (12.2) 135 (9.4) 27 (12.1) 20 (13.5) 24 (11.3) 0.29 Medical history (%) Hypertension 1307 (64.6) 624 (47.7) 683 (66.9) 512 (95.9) 503 (95.8) 903 (63.1) 146 (65.2) 108 (73.0) 144 (67.6) 0.078 Treated hypertension 1072 (53.0) 516 (51.5) 556 (54.5) 533 (99.8) 522 (99.4) 720 (50.3) 134 (59.8) 98 (66.2) 115 (54.0) <0.001 Diabetes mellitus 353 (17.4) 164 (16.4) 189 (18.5) 125 (23.4) 121 (23.0) 245 (17.1) 38 (17.0) 38 (25.7) 28 (13.1) 0.020 Atrial fibrillation 333 (16.5) 169 (16.9) 164 (16.1) 135 (25.3) 116 (22.1) 224 (15.7) 43 (19.2) 36 (24.3) 30 (14.1) 0.024 Stroke 295 (14.6) 150 (15.0) 145 (14.2) 113 (21.2) 97 (18.5) 207 (14.5) 37 (16.5) 25 (16.9) 22 (10.3) 0.22 TIA 286 (14.1) 147 (14.7) 139 (13.6) 91 (17.0) 96 (18.3) 179 (12.5) 43 (19.2) 24 (16.2) 39 (18.3) 0.010 IHD 380 (18.8) 191 (19.1) 189 (18.5) 136 (25.5) 153 (29.1) 248 (17.3) 59 (26.3) 38 (25.7) 34 (16.0) 0.001 PAD 65 (3.2) 29 (2.9) 36 (3.5) 23 (4.3) 22 (4.2) 35 (2.4) 10 (4.5) 8 (5.4) 12 (5.6) 0.018 Hyperlipidaemia 587 (29.0) 293 (29.2) 294 (28.8) 204 (38.2) 216 (41.1) 412 (28.8) 60 (26.8) 52 (35.1) 57 (26.8) 0.29 Smoking, current 573 (28.3) 278 (27.7) 295 (28.9) 111 (20.8) 109 (20.8) 380 (26.6) 71 (31.7) 37 (25.0) 83 (39.0) 0.010 Alcohol >21 units per week 176 (8.7) 92 (9.2) 84 (8.2) 38 (7.1) 29 (5.5) 116 (8.1) 19 (10.9) 11 (7.4) 28 (13.1) 0.10 Side of lesion, right (%) 1047 (51.8) 509 (50.8) 538 (52.7) 278 (52.4) 264 (50.4) 722 (50.5) 109 (48.7) 89 (60.1) 127 (59.6) 0.010 NIHSS (/42), calculated 9.9 (5.3) 9.8 (5.3) 10.0 (5.4) 10.3 (5.5) 10.2 (5.4) 9.7 (5.2) 10.1 (5.4) 9.6 (5.0) 11.6 (5.6) <0.001 GCS <15 (%) 460 (22.7) 222 (22.2) 238 (23.3) 134 (25.1) 145 (27.6) 300 (21.0) 61 (27.2) 32 (21.6) 65 (30.5) 0.006 TOAST classification*

64 (50.4) 722 (50.5) 109 (48.7) 89 (60.1) 127 (59.6) 0.010 NIHSS (/42), calculated 9.9 (5.3) 9.8 (5.3) 10.0 (5.4) 10.3 (5.5) 10.2 (5.4) 9.7 (5.2) 10.1 (5.4) 9.6 (5.0) 11.6 (5.6) <0.001 GCS <15 (%) 460 (22.7) 222 (22.2) 238 (23.3) 134 (25.1) 145 (27.6) 300 (21.0) 61 (27.2) 32 (21.6) 65 (30.5) 0.006 TOAST classification* Cardioembolic 358 (17.7) 181 (18.1) 177 (17.3) 133 (24.9) 117 (22.3) 271 (18.9) 41 (18.3) 24 (16.2) 22 (10.3) 0.021 Large vessel 527 (26.1) 254 (25.3) 273 (26.7) 143 (49.5) 146 (27.8) 200 (14.0) 53 (23.7) 90 (60.8) 180 (84.5) <0.001 Small Vessel 808 (39.9) 402 (40.1) 406 (39.8) 188 (35.2) 199 (37.9) 649 (45.4) 105 (46.9) 34 (23.0) 16 (7.5) <0.001 Other 394 (19.5) 202 (20.2) 192 (18.8) 93 (17.4) 87 (16.6) 333 (23.3) 32 (14.3) 17 (11.5) 12 (5.6) <0.001 Haemodynamics BP, systolic (mm Hg) 166.6 (18.5) 167.1 (18.3) 166.1 (18.7) 165.4 (18.9) 167.7 (17.8) 166.6 (18.6) 166.6 (18.7) 165.1 (17.2) 167.5 (18.3) 0.70 BP, diastolic (mm Hg) 89.2 (13.0) 89.9 (13.1) 88.5 (12.8) 87.5 (13.3) 88.2 (12.6) 90.1 (13.0) 88.2 (13.3) 87.7 (12.9) 86.7 (11.8) <0.001 Heart rate (bpm) 76.8 (14.4) 77.1 (14.5) 76.5 (14.2) 75.8 (14.5) 76.1 (14.5) 76.5 (14.6) 78.4 (13.1) 78.7 (15.4) 75.3 (13.1) 0.039 Time to randomisation [hours] 25.6 [21.2] 24.9 [21.5] 26.0 [21.1] 25.2 [18.8] 23.9 [22.1] 26.0 [21.6] 24.0 [20.1] 23.9 [17.8] 24.8 [19.0] 0.08 Thrombolysis (%) 239 (11.8) 107 (10.7) 132 (12.9) 71 (13.3) 63 (12.0) 156 (10.9) 24 (10.7) 20 (13.5) 39 (18.3) 0.015 *Total may exceed 100% due to mixed causality. χ2 test for categorical variables or one-way analysis of variance for continuous variables across grades of carotid stenosis.

] 23.9 [17.8] 24.8 [19.0] 0.08 Thrombolysis (%) 239 (11.8) 107 (10.7) 132 (12.9) 71 (13.3) 63 (12.0) 156 (10.9) 24 (10.7) 20 (13.5) 39 (18.3) 0.015 *Total may exceed 100% due to mixed causality. χ2 test for categorical variables or one-way analysis of variance for continuous variables across grades of carotid stenosis. BP, blood pressure; GCS, Glasgow Coma Scale; GTN, glyceryl trinitrate; IHD, ischaemic heart disease; mRS, modified Rankin Scale; NIHSS, National Institute of Health Stroke Scale; PAD, peripheral arterial disease; TIA, transient ischaemic attack. Relationship between carotid stenosis and outcome Across all patients and as compared with participants with <30% ipsilateral stenosis, those with ≥70% stenosis had an unfavourable shift in mRS (worse outcome) at day 90 (OR 1.88, 95% CI 1.44 to 2.44, p<0.001, table 2, figure 1); significant associations with worse disability and quality of life, more depression and poorer cognitive scores were also seen. In addition, those with ≥70% stenosis had an increased rate of recurrent ischaemic stroke, clinical deterioration, neurological deterioration, and higher NIHSS Scores at day 7 (online supplementary table 1). Table 2 Functional outcome and death at day 90 by degree of ipsilateral carotid stenosis

Relationship between carotid stenosis and outcome Across all patients and as compared with participants with <30% ipsilateral stenosis, those with ≥70% stenosis had an unfavourable shift in mRS (worse outcome) at day 90 (OR 1.88, 95% CI 1.44 to 2.44, p<0.001, table 2, figure 1); significant associations with worse disability and quality of life, more depression and poorer cognitive scores were also seen. In addition, those with ≥70% stenosis had an increased rate of recurrent ischaemic stroke, clinical deterioration, neurological deterioration, and higher NIHSS Scores at day 7 (online supplementary table 1). Table 2 Functional outcome and death at day 90 by degree of ipsilateral carotid stenosis Stenosis <30% Stenosis 30–<50% Stenosis 50–<70% Stenosis ≥70% n (%) /median [IQR] OR (95% CI) P value n (%) /median [IQR] OR (95% CI) P value n (%) /median [IQR] OR (95% CI) P value Number of participants 1431 224 – – 148 – – 213 – – mRS (/6)* 2 [3] 2 [2] 1.03 (0.80 to 1.33) 0.83 3 [2] 1.21 (0.89 to 1.64) 0.23 3 [2] 1.88 (1.44 to 2.44) <0.001 Death (%) 67 (4.7) 21 (9.4) 1.85 (1.06 to 3.22) 0.030 11 (7.4) 1.43 (0.71 to 2.90) 0.32 25 (11.8) 2.52 (1.48 to 4.27) 0.001 Data are n (%), median (IQR) or OR with 95% CIs. Comparison using logistic or ordinal regression with <30% stenosis as reference group. Adjusted for age, sex, baseline mRS, history of previous stroke, history of diabetes mellitus, TACS, nitrate use, baseline SSS, thrombolysis, feeding status, time to randomisation, baseline SBP, GTN/no GTN and continue/stop. *Ordinal logistic regression.

Stenosis <30% Stenosis 30–<50% Stenosis 50–<70% Stenosis ≥70% n (%) /median [IQR] OR (95% CI) P value n (%) /median [IQR] OR (95% CI) P value n (%) /median [IQR] OR (95% CI) P value Number of participants 1431 224 – – 148 – – 213 – – mRS (/6)* 2 [3] 2 [2] 1.03 (0.80 to 1.33) 0.83 3 [2] 1.21 (0.89 to 1.64) 0.23 3 [2] 1.88 (1.44 to 2.44) <0.001 Death (%) 67 (4.7) 21 (9.4) 1.85 (1.06 to 3.22) 0.030 11 (7.4) 1.43 (0.71 to 2.90) 0.32 25 (11.8) 2.52 (1.48 to 4.27) 0.001 Data are n (%), median (IQR) or OR with 95% CIs. Comparison using logistic or ordinal regression with <30% stenosis as reference group. Adjusted for age, sex, baseline mRS, history of previous stroke, history of diabetes mellitus, TACS, nitrate use, baseline SSS, thrombolysis, feeding status, time to randomisation, baseline SBP, GTN/no GTN and continue/stop. *Ordinal logistic regression. GTN, glyceryl trinitrate; mRS, modified Rankin Scale; SSS, Scandinavian Stroke Scale. Figure 1 mRS at day 90 <30% vs ≥70% ipsilateral stenosis. mRS, modified Rankin Scale.

Stenosis <30% Stenosis 30–<50% Stenosis 50–<70% Stenosis ≥70% n (%) /median [IQR] OR (95% CI) P value n (%) /median [IQR] OR (95% CI) P value n (%) /median [IQR] OR (95% CI) P value Number of participants 1431 224 – – 148 – – 213 – – mRS (/6)* 2 [3] 2 [2] 1.03 (0.80 to 1.33) 0.83 3 [2] 1.21 (0.89 to 1.64) 0.23 3 [2] 1.88 (1.44 to 2.44) <0.001 Death (%) 67 (4.7) 21 (9.4) 1.85 (1.06 to 3.22) 0.030 11 (7.4) 1.43 (0.71 to 2.90) 0.32 25 (11.8) 2.52 (1.48 to 4.27) 0.001 Data are n (%), median (IQR) or OR with 95% CIs. Comparison using logistic or ordinal regression with <30% stenosis as reference group. Adjusted for age, sex, baseline mRS, history of previous stroke, history of diabetes mellitus, TACS, nitrate use, baseline SSS, thrombolysis, feeding status, time to randomisation, baseline SBP, GTN/no GTN and continue/stop. *Ordinal logistic regression. GTN, glyceryl trinitrate; mRS, modified Rankin Scale; SSS, Scandinavian Stroke Scale. Figure 1 mRS at day 90 <30% vs ≥70% ipsilateral stenosis. mRS, modified Rankin Scale. Effects of GTN versus no GTN Those with ≥70% ipsilateral stenosis who were randomised to GTN had a significant shift in mRS to less death or dependency at 90 days (OR 0.56, 95% CI 0.34 to 0.93, p=0.024) (table 3, figure 2). However, GTN did not influence mRS across the other carotid stenosis groups, although there were higher cognitive scores at day 90 in those with 50–<70% stenosis but not in other stenosis groups. Headache, a recognised side effect of GTN, was more common in those with ≥70% stenosis who were randomised to GTN; non-significant increases in headache with GTN were also reported in the other carotid stenosis groups (online supplementary table 2).

ay 90 in those with 50–<70% stenosis but not in other stenosis groups. Headache, a recognised side effect of GTN, was more common in those with ≥70% stenosis who were randomised to GTN; non-significant increases in headache with GTN were also reported in the other carotid stenosis groups (online supplementary table 2). Table 3 Functional outcome and death at day 90 by randomised treatment by degree of ipsilateral carotid stenosis Stenosis 30–<50% Stenosis 50–<70% Stenosis ≥70% GTN No GTN OR (95% CI) P value GTN No GTN OR (95% CI) P value GTN No GTN OR (95% CI) P value Number of participants 102 122 – – 77 71 – – 94 119 – – mRS (/6)* 2 [3] 3 [2] 0.77 (0.47 to 1.27) 0.31 3 [2] 3 [2] 0.71 (0.39 to 1.31) 0.28 3 [2] 4 [2] 0.56 (0.34 to 0.93) 0.024 Death (%) 7 (6.9) 14 (11.5) 0.43 (0.14 to 1.32) 0.14 3 (3.9) 8 (11.3) 0.18 (0.03 to 1.03) 0.054 9 (9.7) 16 (13.4) 0.63 (0.23 to 1.75) 0.37 Continue Stop Continue Stop Continue Stop Number of participants 65 68 – – 52 47 – – 57 57 – – mRS (/6)* 2 [3] 2.5 [3] 0.93 (0.48 to 1.80) 0.82 3 [2] 3 [2] 1.92 (0.86 to 4.27) 0.11 4 [3] 4 [3] 1.46 (0.73 to 2.93) 0.29 Death (%) 9 (13.8) 7 (10.3) 1.59 (0.42 to 6.01) 0.49 6 (11.5) 1 (2.1) 81.00 (1.08 to 6093.98) 0.046 12 (21.4) 6 (10.5) 3.11 (0.85 to 11.44) 0.09 Data are n (%), median (IQR) or OR with 95% CIs. Comparison using logistic or ordinal regression. Adjusted for age, sex, baseline mRS, history of previous stroke, history of diabetes mellitus, TACS, nitrate use, baseline SSS, thrombolysis, feeding status, time to randomisation, baseline SBP and continue/stop or GTN/no GTN, respectively.

dian (IQR) or OR with 95% CIs. Comparison using logistic or ordinal regression. Adjusted for age, sex, baseline mRS, history of previous stroke, history of diabetes mellitus, TACS, nitrate use, baseline SSS, thrombolysis, feeding status, time to randomisation, baseline SBP and continue/stop or GTN/no GTN, respectively. *Ordinal logistic regression. GTN, glyceryl trinitrate; mRS, modified Rankin Scale; SBP, systolic blood pressure; SSS, Scandinavian Stroke Scale; TACS, total anterior circulation syndrome. Figure 2 mRS at day 90 in those with ≥70% ipsilateral stenosis GTN versus no GTN. GTN, glyceryl trinitrate; mRS, modified Rankin Scale. Effects of continuing versus stopping prestroke antihypertensives In those with 50–<70% ipsilateral stenosis, continuing prestroke antihypertensives was associated with more depression, worse disability and quality of life, and poorer cognitive score (TICS-M) at day 90 compared with stopping prestroke antihypertensives, independent of GTN allocation. No effect on mRS was seen nor were these effects replicated in those with more severe (≥70%) ipsilateral stenosis (table 3). Bilateral carotid stenosis Only 97/2023 (4.8%) had ≥50% bilateral carotid stenosis. There were no significant associations across degrees of bilateral carotid stenosis with clinical outcome measures at either 7 or 90 days (online supplementary table 4). Neither GTN nor continuing prestroke antihypertensives influenced outcome in participants with bilateral carotid stenosis (data not shown).

stenosis. There were no significant associations across degrees of bilateral carotid stenosis with clinical outcome measures at either 7 or 90 days (online supplementary table 4). Neither GTN nor continuing prestroke antihypertensives influenced outcome in participants with bilateral carotid stenosis (data not shown). BP change and carotid stenosis The largest fall in BP was seen from baseline to day 1 in those randomised to GTN versus no GTN (7/3 mm Hg overall) and did not significantly differ across degrees of ipsilateral (online supplementary table 2, interaction p=0.22) or bilateral carotid stenosis (interaction p=0.19). When assessed across degrees of ipsilateral carotid stenosis, change in BP from baseline to day 1 did not influence mRS at day 90 (online supplementary table 5). There were no significant interactions between treatment with GTN or continuing prestroke antihypertensives in relation to BP change and outcome. Similarly, across degrees of bilateral carotid stenosis, change in BP from baseline to day 1 did not influence mRS at day 90 (online supplementary table 6).

mentary table 5). There were no significant interactions between treatment with GTN or continuing prestroke antihypertensives in relation to BP change and outcome. Similarly, across degrees of bilateral carotid stenosis, change in BP from baseline to day 1 did not influence mRS at day 90 (online supplementary table 6). Discussion In this preplanned ENOS substudy,9 taking all patients irrespective of treatment allocation, the presence of severe (≥70%) ipsilateral symptomatic carotid stenosis was associated with unfavourable clinical outcomes both early and late after acute stroke. However, treatment with GTN versus no GTN was associated with a significant shift to less death or dependency at 90 days in those participants with ≥70% ipsilateral stenosis. Continuing versus stopping prestroke antihypertensives was associated with worse secondary outcomes in those with 50–<70% ipsilateral stenosis. Modest BP lowering was safe in patients with acute stroke in the context of unilateral and bilateral carotid stenosis.

ays in those participants with ≥70% ipsilateral stenosis. Continuing versus stopping prestroke antihypertensives was associated with worse secondary outcomes in those with 50–<70% ipsilateral stenosis. Modest BP lowering was safe in patients with acute stroke in the context of unilateral and bilateral carotid stenosis. It has long been established that higher degrees of symptomatic carotid stenosis are associated with early stroke recurrence and subsequent dependency after minor stroke and transient ischaemic attack; carotid endarterectomy is therefore recommended.22 This dataset suggests that patients with more severe strokes with large artery disease who may not have been eligible for carotid revascularisation have both poor early and late clinical outcomes as expected. These findings are likely to be even stronger in clinical practice as the 32.9% patients with ischaemic stroke who did not have carotid imaging had more severe strokes and therefore worse clinical outcome than those with imaging.

for carotid revascularisation have both poor early and late clinical outcomes as expected. These findings are likely to be even stronger in clinical practice as the 32.9% patients with ischaemic stroke who did not have carotid imaging had more severe strokes and therefore worse clinical outcome than those with imaging. Although pathophysiological data have demonstrated dysfunctional cerebral autoregulation in severe carotid stenosis,23–25 and cerebral blood flow can become dependent on systemic BP,5 there are limited prospective data assessing BP lowering in acute stroke with carotid stenosis. Previous blinded analysis of ENOS performed during recruitment revealed that BP lowering in the context of carotid stenosis was safe.9 We have reinforced that finding here with no evidence to suggest modest BP lowering is associated with stroke recurrence or other poor outcomes. A post hoc analysis of the Scandinavian Candesartan Acute Stroke Trial (SCAST) found no clear evidence to suggest that BP lowering was detrimental in patients with carotid stenosis, but there were non-significant tendencies towards increased stroke progression and poor functional outcome with candesartan.26 As in the present substudy, the observed BP lowering effect in SCAST was modest (5/2 mm Hg) and therefore differing drug class effects may explain the differences observed between the present analysis and SCAST. Further, continuing prestroke antihypertensives in those with modest ipsilateral carotid stenosis (50–<70%) was associated with poorer outcomes across several secondary domains, but not in those with ≥70% ipsilateral stenosis. This may be due to those with modest stenosis having greater baseline comorbidity (higher mRS, more hypertension, diabetes, previous stroke, ischaemic heart disease and hyperlipidaemia) than those with ≥70% stenosis. This imbalance may represent selection bias, whereby patients with more severe stroke and severe stenosis were not imaged as doing so would not change management. It is unclear whether there are drug class-specific mechanisms that may be harmful in the context of carotid stenosis early after stroke, but we add to evidence from the main ENOS Trial that routinely continuing prestroke antihypertensives should perhaps be avoided until the patient is neurologically stable.7 Whether larger precipitous drops in BP are safe is beyond the scope of this substudy, but given the uncertainty this practice should perhaps be avoided.

e add to evidence from the main ENOS Trial that routinely continuing prestroke antihypertensives should perhaps be avoided until the patient is neurologically stable.7 Whether larger precipitous drops in BP are safe is beyond the scope of this substudy, but given the uncertainty this practice should perhaps be avoided. The shift in mRS to less death or dependency with GTN seen in those with ≥70% ipsilateral carotid stenosis may be related to effects other than BP lowering. As a vasodilator, GTN did not reduce cerebral blood flow in patients with acute stroke despite lowering peripheral and central BP.27 28 Although it is unclear whether GTN improves cerebral blood flow,28 it can be hypothesised that GTN may improve collateral blood supply via surface pial collaterals29 and thereby maintain blood flow to the ischaemic penumbra in the context of carotid stenosis when given early. A similar potential beneficial effect of transdermal GTN administered within 4 hours of stroke onset in patients with ≥70% ipsilateral carotid stenosis was seen in a preliminary analysis of 314 patients from the Rapid Intervention with Glyceryl trinitrate in Hypertensive stroke Trial-2 (RIGHT-2).30 Bilateral carotid stenosis is uncommon (4.8% in this analysis) and therefore data regarding the safety of BP lowering in patients with acute stroke in this context are scanty.6 Impaired cerebral perfusion is more common in this population than in unilateral carotid stenosis and although we found no evidence that modest BP lowering was unsafe, further data are required to address this question.

data regarding the safety of BP lowering in patients with acute stroke in this context are scanty.6 Impaired cerebral perfusion is more common in this population than in unilateral carotid stenosis and although we found no evidence that modest BP lowering was unsafe, further data are required to address this question. This study represents the largest trial-based analysis of BP lowering in patients with acute ischaemic stroke with carotid stenosis to date. However, there are limitations. First, not all patients with ischaemic stroke had carotid imaging studies performed, a deficiency most prominent in patients with severe stroke who would not be appropriate candidates for revascularisation therapy. Second, the analyses across degrees of carotid stenosis are subgroup analyses and, so the results may represent chance. Third, the median time to randomisation in ENOS was 26 hours and so the effect of BP lowering in the context of carotid stenosis within the first few hours of ischaemic stroke remains unclear. Fourth, no adjustment was made for multiplicity of testing due to the exploratory nature of the study. Fifth, imaging information on carotid stenosis was provided by investigators at the site with unknown reporting criteria (NASCET,13 European Carotid Surgery Trial [ECST]),31 and Carotid and Vertebral Artery Transluminal Angioplasty Study (CAVATAS)32) and were not centrally adjudicated. However, data were validated and checked for accuracy. Sixth, it was not possible to adjust cognitive outcome data for cognition at baseline and many of the differences seen at 3 months in association with degrees of stenosis (apart from allocated trial treatment differences) likely reflect baseline status. Seventh, ENOS assessed mild to moderate BP lowering and the results presented here do not provide information on the effects of intensive BP lowering. Last, data on carotid endarterectomy were not captured prospectively in ENOS and were therefore unavailable for the population studied, although in a population of patients with mostly moderate-to-severe stroke it is unlikely that many patients underwent endarterectomy.

ormation on the effects of intensive BP lowering. Last, data on carotid endarterectomy were not captured prospectively in ENOS and were therefore unavailable for the population studied, although in a population of patients with mostly moderate-to-severe stroke it is unlikely that many patients underwent endarterectomy. In summary, this ENOS substudy has demonstrated that severe ipsilateral carotid stenosis is associated with worse clinical outcomes at 7 days and 90 days after acute stroke irrespective of treatment allocation. Transdermal GTN improved functional outcome at 90 days in those with ≥70% ipsilateral stenosis and was safe across all degrees of carotid stenosis whether unilateral or bilateral. Further, modest BP lowering with GTN was safe in the context of carotid stenosis although continuing prestroke antihypertensives was associated with poorer secondary outcomes in those with 50–<70% ipsilateral carotid stenosis. Future studies should establish whether GTN and other BP lowering therapies have specific mechanistic properties that may be of benefit in acute stroke in the context of carotid stenosis. The authors thank the participants and investigators who took part in the ENOS trial.

In summary, this ENOS substudy has demonstrated that severe ipsilateral carotid stenosis is associated with worse clinical outcomes at 7 days and 90 days after acute stroke irrespective of treatment allocation. Transdermal GTN improved functional outcome at 90 days in those with ≥70% ipsilateral stenosis and was safe across all degrees of carotid stenosis whether unilateral or bilateral. Further, modest BP lowering with GTN was safe in the context of carotid stenosis although continuing prestroke antihypertensives was associated with poorer secondary outcomes in those with 50–<70% ipsilateral carotid stenosis. Future studies should establish whether GTN and other BP lowering therapies have specific mechanistic properties that may be of benefit in acute stroke in the context of carotid stenosis. The authors thank the participants and investigators who took part in the ENOS trial. Contributors: JPA performed the analyses and wrote the first draft. PMB conceived the study and is corresponding author and project guarantor. LJW is the trial statistician. DB, EB, VC, HMC, HKC, RC, JG, ACL, GN, SO, GMS, NS, SS, YW, JW are members of the trial steering/international advisory committee. LJW, AB, DB, EB, VC, HMC, HKC, RC, JG, ACL, GN, SO, GMS, SS, YW, JW, NS commented upon and approved the manuscript for publication. Funding: ENOS was funded by Bupa Foundation and the Medical Research Council (G0501797). JPA is funded by NIHR TARDIS (10/104/24) and BHF RIGHT-2 (CS/14/4/30972).

Contributors: JPA performed the analyses and wrote the first draft. PMB conceived the study and is corresponding author and project guarantor. LJW is the trial statistician. DB, EB, VC, HMC, HKC, RC, JG, ACL, GN, SO, GMS, NS, SS, YW, JW are members of the trial steering/international advisory committee. LJW, AB, DB, EB, VC, HMC, HKC, RC, JG, ACL, GN, SO, GMS, SS, YW, JW, NS commented upon and approved the manuscript for publication. Funding: ENOS was funded by Bupa Foundation and the Medical Research Council (G0501797). JPA is funded by NIHR TARDIS (10/104/24) and BHF RIGHT-2 (CS/14/4/30972). Competing interests: PMB is Stroke Association Professor of Stroke Medicine, Chief Investigator of MRC ENOS, and a NIHR Senior Investigator. PMB reports grants from British Heart Foundation and the NIHR Health Technology Assessment Programme, during the conduct of the study; other from Platelet Solutions Ltd, personal fees from Diamedica, personal fees from Nestle, personal fees from Phagenesis Ltd, personal fees from ReNeuron, personal fees from Athersys, personal fees from Covidien, outside the submitted work. Ethics approval: The ENOS Trial was approved by national and local ethics committees and competent authorities in participating countries. Provenance and peer review: Not commissioned; internally peer reviewed. Data sharing statement: Supplementary material are available online. Collaborators: ENOS investigators' details are specified in the online supplementary file. Patient consent for publication: Not required.

Introduction Intraventricular haemorrhage (IVH) is an outcome of haemorrhagic stroke characterised by eruption of blood in the cerebroventricular system caused by rupture of a cerebrospinal artery spontaneously or secondarily. IVH can serve as an independent predictor of worse outcome and increasing morbidity and mortality after intracerebral haemorrhage (ICH), and can lead to severe complications, such as hydrocephalus. Hydrocephalus itself may contribute to increased mortality following IVH,1–3 and can be used as an independent predictor of higher mortality4 5 and worse prognosis.1 Hydrocephalus may be communicating or non-communicating. Communicating hydrocephalus occurs when the flow of cerebrospinal fluid (CSF) is blocked after it exits the ventricles. Non-communicating (obstructive) hydrocephalus occurs when the flow of CSF is blocked along one or more of the narrow passages connecting the ventricles. Current data indicate that hydrocephalus, communicating or non-communicating, develops in up to 67% of patients with intraventricular extension of ICH.3 4

ventricles. Non-communicating (obstructive) hydrocephalus occurs when the flow of CSF is blocked along one or more of the narrow passages connecting the ventricles. Current data indicate that hydrocephalus, communicating or non-communicating, develops in up to 67% of patients with intraventricular extension of ICH.3 4 Both immediate and delayed hydrocephalus are possible following IVH,6 yet the mechanisms by which they occur are unclear. Despite improvement in treatment of IVH and subsequent hydrocephalus in recent years, these conditions are still life-threatening and often lead to death. Mechanistic understanding of the events that lead to IVH and subsequent hydrocephalus will guide both clinical research and basic-science experiments geared towards treating these conditions. Creating a mechanistic framework is important for clarifying the underlying pathophysiology, identifying targets for intervention and improving the monitoring of progression of IVH. With improved monitoring and means of intervention, deterioration of patients’ conditions can be avoided and rate of recovery can be increased. We focus on the mechanism of hydrocephalus after non-traumatic IVH in adults. Although germinal matrix-intraventricular haemorrhage is a frequent phenomenon in premature and very-low-birth-weight (<1500 g) infants, it is beyond the scope of this mini-review, as the anatomy and physiology of IVH in infants and adults are distinct.

the mechanism of hydrocephalus after non-traumatic IVH in adults. Although germinal matrix-intraventricular haemorrhage is a frequent phenomenon in premature and very-low-birth-weight (<1500 g) infants, it is beyond the scope of this mini-review, as the anatomy and physiology of IVH in infants and adults are distinct. Blood clot blockage Currently, the prevailing theory explaining acute IVH-induced hydrocephalus is blood-clot blockage in the CSF drainage pathway. Blood-clot blockage frequently takes place in cerebral aqueducts or in the outlets of the fourth ventricle, whereas tetraventricular hydrocephalus typically results from blockage at the level of the cortical subarachnoid space and, less commonly, in the fourth ventricle outlets.6 After IVH, obstructive hydrocephalus can occur immediately. In these cases, multiple small blood clots form throughout the ventricular system, and obstruct the pathway through the arachnoid villi into the venous sinuses and small blood vessels leading to and from the ependymal cells.7 Based on this classical theory, current therapies predominantly aim at removing the blood and blood clots, slowing down haematoma growth, and reducing haematoma size. Examples of such strategies include administration of recombinant tissue-type plasminogen activator (rt-PA), minimally invasive stereotactic surgery and endoscopic surgery.8

cal theory, current therapies predominantly aim at removing the blood and blood clots, slowing down haematoma growth, and reducing haematoma size. Examples of such strategies include administration of recombinant tissue-type plasminogen activator (rt-PA), minimally invasive stereotactic surgery and endoscopic surgery.8 Advances in our understanding of the mechanism of hydrocephalus following IVH have been aided by observations in experimental animals. Mayfrank et al9 10 used a swine model of IVH to demonstrate that occlusion of CSF outflow by blood clots in the ventricular system and/or distention of the ventricular walls are prominent mechanisms of hydrocephalus development, which also lead to ventricular dilation and elevation of intracranial pressure (ICP). Furthermore, the impact of the hydrocephalus on adjacent cerebral structures and local blood circulation was shown to persist in the animal model even after the acute obstructive phase.9 Similar mechanisms likely control IVH-induced hydrocephalus in humans. The total blood volume and duration of blood and blood-clot presence in the ventricular system are the main factors contributing to the intensity of communicating hydrocephalus, which occurs mostly in a delayed phase when an inflammatory reaction is generated as a result of the debris accumulating as the blood products are broken down.11

ume and duration of blood and blood-clot presence in the ventricular system are the main factors contributing to the intensity of communicating hydrocephalus, which occurs mostly in a delayed phase when an inflammatory reaction is generated as a result of the debris accumulating as the blood products are broken down.11 Barrier impairment Hydrocephalus occurs because of increased production, inappropriate flow or decreased reabsorption of CSF. Barrier-impairment mechanisms leading to hydrocephalus, including dysfunctional ependymal cells, blood–brain barrier (BBB) and the relevant molecular structures are summarised below.

ume and duration of blood and blood-clot presence in the ventricular system are the main factors contributing to the intensity of communicating hydrocephalus, which occurs mostly in a delayed phase when an inflammatory reaction is generated as a result of the debris accumulating as the blood products are broken down.11 Barrier impairment Hydrocephalus occurs because of increased production, inappropriate flow or decreased reabsorption of CSF. Barrier-impairment mechanisms leading to hydrocephalus, including dysfunctional ependymal cells, blood–brain barrier (BBB) and the relevant molecular structures are summarised below. The ependyma is the thin epithelium-like lining of the ventricular system. Ependymal cell integrity is important for CSF production and flow. Postnatal ependymal cells are non-renewable, multicilial, cuboidal epithelial cells lining the entire ventricular system. Their function includes CSF production, which is regulated by the CSF components and by neurotransmitters such as serotonin.6 12 13 The ependymal cells vary in morphology and metabolism during development, and become relatively uniform in adults. They are vulnerable to injury,12 which may occur as a result of increased ICP14 and/or inflammation.15 The structural characteristics and enzymatic function in mature mammalian cerebral ependymal cells ensure integrity of the barrier at the brain–CSF interface, which is responsible for scavenging and detoxifying various substances in the CSF, such as oxidants and pathogens.16 The cilia of ependymal cells have a dual function—as sensory compartments, and as mediators of fluid and cell movement.17

ebral ependymal cells ensure integrity of the barrier at the brain–CSF interface, which is responsible for scavenging and detoxifying various substances in the CSF, such as oxidants and pathogens.16 The cilia of ependymal cells have a dual function—as sensory compartments, and as mediators of fluid and cell movement.17 Inevitably, the ependymal surface is damaged after IVH.6 Mayfrank et al10 observed marked loss of the ependymal cells covering the ventricular walls in their swine IVH model. Two complementary mechanisms underlying ependymal cell damage and leading to hydrocephalus following IVH have been proposed: (1) disruption of the ependymal surface (cells and their cilia) results in collapse of the cerebral aqueduct walls or in aqueductal stenosis and CSF flow occlusion6 and (2) IVH leads to failure regulating the transfer of fluid, ions and small molecules between the cerebral parenchyma and the ventricular fluid, due to injury-caused dysfunction of the ependymal cells.12 The relative contribution of each mechanism awaits further study.

r in aqueductal stenosis and CSF flow occlusion6 and (2) IVH leads to failure regulating the transfer of fluid, ions and small molecules between the cerebral parenchyma and the ventricular fluid, due to injury-caused dysfunction of the ependymal cells.12 The relative contribution of each mechanism awaits further study. IVH causes not only ependymal cell damage, but also disruption of the BBB. The BBB is composed of cerebral endothelial cells and their linking tight junctions.6 In addition to protecting the brain from pathogens and unwanted molecules, the BBB also is important for maintaining the CSF protein content and the osmotic pressure in the brain.6 Krishnamurthy et al18 injected hyperosmolar dextran and fibroblast growth factor (FGF) 2 into the ventricles to mimic the condition of increased protein content and osmotic overload resulting from breakdown of BBB after IVH. These solutions that increased the osmotic load in the ventricles and water influx (through the choroid plexus CSF secretion and/or through the brain) into the ventricles to normalise this osmotic gradient successfully induced secondary hydrocephalus in a model of adult rats. However, whether increasing osmolarity or protein content in the ventricles would be able to maintain hydrocephalus is unclear.

through the choroid plexus CSF secretion and/or through the brain) into the ventricles to normalise this osmotic gradient successfully induced secondary hydrocephalus in a model of adult rats. However, whether increasing osmolarity or protein content in the ventricles would be able to maintain hydrocephalus is unclear. Aquaporins (AQPs) are water channels facilitating water movement across cell membranes. In the human brain, the main water channels are AQP1 and AQP4. AQP1 is a predominant component of the apical membrane of the choroid plexus, whereas AQP4 is the main channel in the basolateral surface of ventricular ependymal cells.19 Choroidal AQP1 appears shortly after the choroid plexus in embryonic development and decreases with ageing. Its substantial contribution to CSF production has been demonstrated by experiments in adult mice and rats.19 In severe chronic hydrocephalic rats,20 upregulation of brain AQP421 and its relocation from astrocyte end-feet to the entire plasma membrane of hypertrophic astrocytes might be protective response mechanisms developed to maintain water homoeostasis, possibly by allowing absorption of transependymal CSF into the brain capillaries.19

nic hydrocephalic rats,20 upregulation of brain AQP421 and its relocation from astrocyte end-feet to the entire plasma membrane of hypertrophic astrocytes might be protective response mechanisms developed to maintain water homoeostasis, possibly by allowing absorption of transependymal CSF into the brain capillaries.19 Additional rodent experiments have shown that AQP4 is important in maintaining the ependymal layer integrity in mice, whereas its role in maintaining the integrity of the BBB and CSF dynamics has been a subject of debate.19 Li et al13 demonstrated that the majority of AQP4-null mice presented smaller ventricle size, decreased CSF production, higher brain water content and decreased expression of the gap-junction protein connexin 43 in the ependymal cells compared to wild-type mice. In addition, another AQP4-null mouse model was reported to develop hyperpermeability of the BBB, emphasising the role of AQP4 in maintaining BBB integrity.22 Feng et al23 found hydrocephalus develops in 10% of AQP4-null mice, which display an incomplete ependymal structure and consequent obstruction of small CSF apertures, such as the cerebral aqueduct and fourth ventricular outlets, leading to deficient water transport. Qing et al24 reported iron overload and increased levels of AQP4 located in the perihaematomal area in an adult rat model of ICH and suggested that both contributed to the development of brain oedema after ICH. They hypothesised that AQP4 was upregulated in response to iron accumulation in the periventricular area to mediate hydrocephalus after IVH because AQP4 expression was shown to correlate with iron concentration in that model, and AQP4 upregulation was inhibited by the iron chelator, deferoxamine.25 However, to the best of our knowledge, there have currently been no reports on studies directly examining the role of AQP1 or AQP4 in the mechanism of hydrocephalus after IVH, and there is a dearth of human data regarding aquaporin function in hydrocephalus.19

ulation was inhibited by the iron chelator, deferoxamine.25 However, to the best of our knowledge, there have currently been no reports on studies directly examining the role of AQP1 or AQP4 in the mechanism of hydrocephalus after IVH, and there is a dearth of human data regarding aquaporin function in hydrocephalus.19 Inflammation and fibrosis The inflammation-related theory of hydrocephalus is well established in neonatal hydrocephalus after IVH, for example, dysfunction of arachnoid granulations due to obliterative arachnoiditis26 or CSF flow obstruction due to fibrotic blockage.27 In contrast, the theory and evidence to support or disprove it are limited in adults. After acute obstructive hydrocephalus, inflammation and subsequent scarring of the arachnoid granulations are major contributors to the secondary reaction, in which the flow of CSF through the cerebral aqueduct, fourth ventricular outlets, basal cisterns and/or arachnoid granulations, is prevented, resulting in communicating hydrocephalus,6 28 a frequent sequel of IVH, in which dilation of all the ventricles occurs, in contrast to obstructive hydrocephalus.15

ry reaction, in which the flow of CSF through the cerebral aqueduct, fourth ventricular outlets, basal cisterns and/or arachnoid granulations, is prevented, resulting in communicating hydrocephalus,6 28 a frequent sequel of IVH, in which dilation of all the ventricles occurs, in contrast to obstructive hydrocephalus.15 Inflammatory reaction in response to blood-breaking products reaching the arachnoid granulations with subsequent development of communicating hydrocephalus have been shown in animal experiments, including a swine, rat and mouse.29 However, blockage is a relative phenomenon. For example, subarachnoid space occlusion may be present in communicating hydrocephalus.6 Complement activation may also play a role in hydrocephalus. However, only when the BBB is disrupted or blood extension into the ventricular system occurs, are components of the complement system (beneficial or detrimental) allowed to pass into the ventricular system and possibly induce immune reaction in the brain parenchyma, including cell lysis and inflammation, leading to hydrocephalus.30 31 Inflammation following IVH is mediated also by the transforming growth factor (TGF) family members, TGFβ1 and TGFβ2, which are among the most abundant and functionally versatile cytokines in the mammalian central nervous system (CNS).32 Normally, TGFβ1 is restricted to the choroid plexus and meningeal cells, two sites that are key to the development of hydrocephalus.33 TGFβ2 is located mainly in neurons and astroglial cells.32

TGFβ2, which are among the most abundant and functionally versatile cytokines in the mammalian central nervous system (CNS).32 Normally, TGFβ1 is restricted to the choroid plexus and meningeal cells, two sites that are key to the development of hydrocephalus.33 TGFβ2 is located mainly in neurons and astroglial cells.32 TGFβ1 is a highly ubiquitous cytokine that can be synthesised by virtually all cells, and almost all cells have been shown to have receptors for TGFβ1.33 Thus, the presence of TGFβ1 in the CSF could originate from the haemorrhage itself or from the choroid plexus.34 TGFβ1 is stored in platelet granules and, therefore, substantial amounts of TGFβ1 gain access to the CSF after IVH as platelets accumulate within intraventricular blood clots.33 TGFβ1 induces upregulation of the cognate genes encoding extracellular matrix proteins, such as fibronectin and laminin, which are important mediators of wound healing and scar formation.34 The major roles of TGFβ2 are scarring and fibrosis.35 36 When these events occur in the ventricular system and TGFβ2 acts in the CSF, the result can be hydrocephalus.

gnate genes encoding extracellular matrix proteins, such as fibronectin and laminin, which are important mediators of wound healing and scar formation.34 The major roles of TGFβ2 are scarring and fibrosis.35 36 When these events occur in the ventricular system and TGFβ2 acts in the CSF, the result can be hydrocephalus. However, when Kaestner and Dimitriou tried to detect the distinct behaviour of TGFβ1 and TGFβ2 following subarachnoid haemorrhage (SAH) and IVH in adults,32 they found that TGFβ2 concentration in the plasma did not change over time, and displayed a parabolic concentration change in the CSF with a peak at day 6 post ictus. In contrast, plasma levels of TGFβ1 increased markedly over time in the early phase after the haemorrhage, whereas the CSF levels constantly decreased. These findings suggest that the mechanisms of development of posthaemorrhagic hydrocephalus in adults is unlikely to be mediated by TGFβ2, while whether it involves the crucial role of TGFβ1 is lacking convincing proof, as well.32 33 37 Blood components Iron Heme is degraded in the brain by hemeoxygenase (HO) into iron, carbon monoxide and biliverdin, the latter of which is subsequently converted to bilirubin by biliverdin reductase.31 In response to haemorrhage, HO plays a dual role. It stimulates protective activity by virtue of the anti-inflammatory, antiapoptotic and antiproliferative actions of one or more of the three heme breakdown products,38 but at the same time promotes harmful effects by causing brain iron overload.39

liverdin reductase.31 In response to haemorrhage, HO plays a dual role. It stimulates protective activity by virtue of the anti-inflammatory, antiapoptotic and antiproliferative actions of one or more of the three heme breakdown products,38 but at the same time promotes harmful effects by causing brain iron overload.39 Normal ependymal cells take up iron from the CSF and prevent iron diffusion to the rest of the brain.6 Thus, destruction of ependymal cells following IVH may be one of the causes for increased non-protein-bound iron—which is cytotoxic—in the CSF, and in turn may increase ependymal cell damage and exacerbate patients’ conditions.6 Gao et al40 found that intraventricular injection of lysed red blood cells (RBCs), but not packed RBCs, resulted in ventricular enlargement in rats 24 h postinjection. Similarly, intraventricular injection of iron also resulted in ventricular enlargement and ventricular wall damage, whereas co-injection of deferoxamine with lysed RBCs was protective against ventricular enlargement. These results suggest that iron, a degradation product of haemoglobin, has an important role in development of hydrocephalus after IVH. Systemic deferoxamine treatment has been found to partially reverse brain iron accumulation, hydrocephalus, bilateral enlargement of the lateral ventricles and hippocampal tissue loss, in an adult rat model of IVH.41 In that model, iron accumulation was found associated with upregulation of HO-1 and ferritin (a key iron storage protein) in the hippocampus and periventricular areas.41

e brain iron accumulation, hydrocephalus, bilateral enlargement of the lateral ventricles and hippocampal tissue loss, in an adult rat model of IVH.41 In that model, iron accumulation was found associated with upregulation of HO-1 and ferritin (a key iron storage protein) in the hippocampus and periventricular areas.41 The role of free iron in IVH-induced hydrocephalus may be tightly linked to the inflammatory response. Complement-mediated erythrocyte lysis may expose the CSF and brain to the damaging effects of free iron ions.6 42 43 Intraventricular macrophages express transferrin receptors and may be involved in iron regulation,44 which is also of possible importance after IVH as haemosiderin, an iron-storage complex containing ferritin, is found within macrophages after haemorrhage.6 In animal models of hydrocephalus, the level of complement receptor type 3 in intraventricular macrophages is elevated, suggesting a possible role of complement activation in hydrocephalus development.45 The precise relationships among complement activation, macrophage action and hydrocephalus after IVH need to be delineated further. Thrombin Thrombin is serine protease that acts as part of the fibrinolytic system to induce blood coagulation after injury when bleeding is involved. Highly regulated pro-coagulant and anticoagulant zymogens and cofactors control the blood coagulation cascade, which comprises a series of proteolytic reactions.

rea of blockage, it cannot be deployed to catch the thrombus and be safely withdrawn, and the sheer force of the device may cause endothelial damage or rupture of the artery because of plaque hardness. In these patients, perhaps IA thrombolysis would be more appropriate. The published RCTs did not address these issues. The fourth question is, does the presence of a conflict between the time window of treatment and imaging analysis on the selection of patients with AIS for intervention exist? The current guideline has stated that a patient with AIS would be a good candidate to intervene if their area of infarction is <70 mL, or ASPECT score is >6 (better collaterals). The question is, should patients with AIS who present to the emergency room beyond 12 h but have imaging findings suitable for treatment be treated? Therefore it remains to be determined whether a patient with AIS should be treated based on the time window or on imaging findings. The fifth question is, should newer thrombolytic agents continue to be studied for AIS? Many new agents such as TNK-TPA3 4 or desmotaplase5 have not performed well in clinical trials, or are still being studied for intravenous thrombolysis. Would these agents then be more effective in patients given IA?

The role of free iron in IVH-induced hydrocephalus may be tightly linked to the inflammatory response. Complement-mediated erythrocyte lysis may expose the CSF and brain to the damaging effects of free iron ions.6 42 43 Intraventricular macrophages express transferrin receptors and may be involved in iron regulation,44 which is also of possible importance after IVH as haemosiderin, an iron-storage complex containing ferritin, is found within macrophages after haemorrhage.6 In animal models of hydrocephalus, the level of complement receptor type 3 in intraventricular macrophages is elevated, suggesting a possible role of complement activation in hydrocephalus development.45 The precise relationships among complement activation, macrophage action and hydrocephalus after IVH need to be delineated further. Thrombin Thrombin is serine protease that acts as part of the fibrinolytic system to induce blood coagulation after injury when bleeding is involved. Highly regulated pro-coagulant and anticoagulant zymogens and cofactors control the blood coagulation cascade, which comprises a series of proteolytic reactions. Liu et al46 reported that intraventricular injection of thrombin (20 U from bovine plasma) in adult rats caused BBB breakdown with reduction of brain microvascular endothelial cell and perivascular astrocyte immunoreactivity. Our group established an IVH model in adult rats to investigate what role thrombin plays in the mechanism of IVH-induced hydrocephalus.47 We found that injection of heparinised blood, in contrast to non-heparinised blood, resulted in decreased hydrocephalus when the rats were examined by MRI between 1 and 28 days postinjury.47 Intraventricular injection of thrombin alone caused significant hydrocephalus, ventricular wall damage and periventricular BBB disruption. Thrombin-induced hydrocephalus was reduced by co-injection of the protease-activated receptor 1 (PAR-1) antagonist SCH79797. Based on these results, we concluded that mediation of thrombin's effect through the PAR-1 pathway is an important contributor to hydrocephalus development after IVH.

iventricular BBB disruption. Thrombin-induced hydrocephalus was reduced by co-injection of the protease-activated receptor 1 (PAR-1) antagonist SCH79797. Based on these results, we concluded that mediation of thrombin's effect through the PAR-1 pathway is an important contributor to hydrocephalus development after IVH. Conclusion Hydrocephalus, a severe complication after IVH, can independently increase the risk of mortality. The mechanisms of hydrocephalus following IVH in adults are distinct from those in infants. Our mini-review focused on research findings in adults and animal models, including blood-clot blockage, barrier impairment, inflammation and blood components, including iron and thrombin. Contributors: YB searched most of the relevant published articles through the PubMed website and wrote this article. MC organised the weekly meetings of all the authors, searched some articles and provided helpful input on the topic. TG, XW and XL also found some useful papers. They attended the weekly discussions and gave the first author valuable suggestions for writing the paper. FG supervised and offered guidance to all the authors about properly choosing a theme, and amended and polished the manuscript before it was finalised. Funding: This study was supported and funded by a grant from the National Science Foundation of China (NSFC) (number 81471168). Competing interests: None declared. Provenance and peer review: Not commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

The fourth question is, does the presence of a conflict between the time window of treatment and imaging analysis on the selection of patients with AIS for intervention exist? The current guideline has stated that a patient with AIS would be a good candidate to intervene if their area of infarction is <70 mL, or ASPECT score is >6 (better collaterals). The question is, should patients with AIS who present to the emergency room beyond 12 h but have imaging findings suitable for treatment be treated? Therefore it remains to be determined whether a patient with AIS should be treated based on the time window or on imaging findings. The fifth question is, should newer thrombolytic agents continue to be studied for AIS? Many new agents such as TNK-TPA3 4 or desmotaplase5 have not performed well in clinical trials, or are still being studied for intravenous thrombolysis. Would these agents then be more effective in patients given IA? The last question is, should intravenous thrombolytics be avoided when IA therapies are being considered? All large published RCTs have pointed out that the combination of intravenous TPA plus IA thrombectomy is superior to intravenous TPA alone. What role did intravenous TPA really play in the combination therapy then? Would new trials be needed to prove that simple IA thrombectomy is as good without intravenous TPA being given first?

lished RCTs have pointed out that the combination of intravenous TPA plus IA thrombectomy is superior to intravenous TPA alone. What role did intravenous TPA really play in the combination therapy then? Would new trials be needed to prove that simple IA thrombectomy is as good without intravenous TPA being given first? As has been reviewed by Campbell et al, we have entered into an exciting era of providing effective and safe treatment to patients with AIS. On the contrary, these proven treatment modalities have raised more questions on patient selection, the use of a time window or imaging studies, the role of IA thrombolysis alone or thrombectomy alone and the use of newer thrombolytics for IA treatment. However, it is clear that we have effective and safe options to treat patients with AIS who present early, and imaging studies show viable penumbra and good collaterals. We have more work to carry out in order to elucidate other questions, as mentioned earlier, and perhaps provide more personalised therapy to our patients in the near future. Competing interests: None declared. Provenance and peer review: Commissioned; internally peer reviewed.

The year 2015 was revolutionary for the treatment of acute ischaemic stroke (AIS) because of publication of multiple large randomised controlled clinical trials (RCTs) on the success of mechanical thrombectomy. Campbell et al provided a very comprehensive review of this topic in this issue of Stroke and Vascular Neurology. Consequently, the European and Chinese AIS guidelines have both given thrombectomy the highest level of recommendation (I,A). While we celebrate this most significant change in 20 years of stroke therapy, we also begin to contemplate what these new treatment modalities will bring.

ssue of Stroke and Vascular Neurology. Consequently, the European and Chinese AIS guidelines have both given thrombectomy the highest level of recommendation (I,A). While we celebrate this most significant change in 20 years of stroke therapy, we also begin to contemplate what these new treatment modalities will bring. The first question is, should the current care protocols be changed? Intravenous thrombolysis has been widely practised in China. The number of patients treated has improved significantly in recent times. The door to needle time has been shortened to a level comparable to the international standard at many large stroke centres. However, the rate of giving intravenous tissue-type plasminogen activator (TPA) is still <10% in China and <5% of patients with stroke would arrive at the hospital within 3 h. Therefore, if we do not make significant changes to the current protocol on acute phase triaging, not many patients with AIS may benefit from these state-of-the-art thrombectomies. We especially need to improve two aspects of care processes. One is to improve dissemination of stroke education to the public, hospital administration and governmental agencies so that they can better understand the benefit of thrombectomy. The second is to standardise the training of the providers. In China, providers who can offer thrombectomy include neurologists, neurosurgeons and interventional radiologists. Their training and ways of providing interventional treatment are quite diverse. Many pursue technical and imaging goals while ignoring the importance of preoperative evaluation of patients. The published trials have taught us that preoperative evaluation of patients with stroke for their stroke subtypes, the location of the blockage, the presence of a penumbra with mismatch and the degree of collaterals is crucial and will impact their outcome. If the training of those performing thrombectomy in China is not standardised, it is perceivable that such treatment modality could be abused or misused and, therefore, lose its effectiveness.

the blockage, the presence of a penumbra with mismatch and the degree of collaterals is crucial and will impact their outcome. If the training of those performing thrombectomy in China is not standardised, it is perceivable that such treatment modality could be abused or misused and, therefore, lose its effectiveness. The second question is whether the research on intra-arterial (IA) thrombolysis alone should be continued? Before the arrival of mechanical thrombectomy, IA thrombolysis was one of the treatment modalities for patients with AIS. From PROACT II (Prolyse in Acute Cerebral Thromboembolism II)1 and MELT (Middle Cerebral Artery Embolism Local Fibrinolytic Intervention Trial Japan),2 we have learned that IA thrombolysis alone could improve the patient's outcome. In China, a retrospective review also showed that IA urokinase performed better recanalisation of the arteries and improved outcome. However, these trials did not have rigorous preoperative imaging analysis and screening. Therefore, if the same preoperative screening of patients and their vasculature were implemented, such as in the ESCPE and SWIFT PRIME trials for IA thrombolysis alone, would the outcome be as good clinically? Compared to IA thrombectomy, simple IA thrombolysis is technically easier to perform and financially more economical. If urokinase is used, then the fee for the entire procedure is around 10 000 RMB. On the contrary, the use of a Solitaire FR thrombectomy device would incur a fee of around 70 000 RMB.

good clinically? Compared to IA thrombectomy, simple IA thrombolysis is technically easier to perform and financially more economical. If urokinase is used, then the fee for the entire procedure is around 10 000 RMB. On the contrary, the use of a Solitaire FR thrombectomy device would incur a fee of around 70 000 RMB. The third question is, can the results of these well-conducted RCTs be applied to routine daily clinical practice? It is well know that every patient with stroke is different. Many patients with stroke may require multiple interventional therapies to clear the clot (Penumbra, Solitaire or thrombolysis). Individualising the treatment plan for each patient with AIS is the likely procedure in the future. For example, many patients with AIS may have in situ thrombus because of high-grade arterial stenosis prior to its occlusion, making it technically difficult to perform thrombectomy. In these patients, the catheter cannot be advanced through the point of stenosis—the device cannot reach the area of blockage, it cannot be deployed to catch the thrombus and be safely withdrawn, and the sheer force of the device may cause endothelial damage or rupture of the artery because of plaque hardness. In these patients, perhaps IA thrombolysis would be more appropriate. The published RCTs did not address these issues.

Introduction Stroke is the leading cause of death in China, responsible for 22.45% of all deaths. Three times as many patients die from stroke as from coronary artery disease. Approximately 65% of Chinese strokes are of ischaemic type.1 Proximal large vessel atherosclerotic stenosis or occlusion accounts for 35–40% of all acute ischaemic strokes (AIS). There is a striking ethnic difference in the distribution of atherosclerosis in the cerebral vasculature of patients with ischaemic stroke. In contrast to the high incidence of extracranial large artery atherosclerosis in the Caucasian population, intracranial atherosclerosis (affecting the middle cerebral artery (MCA), intracranial portion of the internal carotid artery and vertebrobasilar artery) is the predominant cause of ischaemic stroke in the Asian population. In fact, intracranial atherosclerosis is estimated to account for 33–50% of all AIS in the Chinese population.2

herosclerosis (affecting the middle cerebral artery (MCA), intracranial portion of the internal carotid artery and vertebrobasilar artery) is the predominant cause of ischaemic stroke in the Asian population. In fact, intracranial atherosclerosis is estimated to account for 33–50% of all AIS in the Chinese population.2 Early recanalisation of the occluded cerebral vessel is the key to successful treatment of AIS.3 Before December 2014, the only proven and effective treatment for AIS was recombinant tissue-type plasminogen activator (alteplase).4 Intravenous recombinant tissue plasminogen activator (intravenous alteplase) given up to 3 hours of symptom onset is the only effective therapy for patients with AIS, this is supported by a National Institute of Neurological Disorders and Stroke (NINDS) study.5 Although the European Cooperative Acute Stroke Study III (ECASS III) study demonstrated benefit of alteplase in selected patients up to 4.5 hours after symptom onset, the use of intravenous alteplase is only approved for use within 3 hours, by the Chinese Food and Drug Administration (FDA), and the extended time window is seldom used in China. Owing to lack of public awareness of early stroke symptoms, less than one quarter of patients with stroke arrive in hospital within 3 hours.1 6 The delay in patient reporting and limited therapeutic time window, in one study, restricted the use of intravenous alteplase to only 1.6% of patients with acute stroke in China.1 In addition, intravenous alteplase has limited efficacy in successful recanalisation, particularly in patients with large vessel occlusion, where the recanalisation rate is <30%.7 Intra-arterial thrombolysis showed potentially higher recanalisation rate (66.0%) and longer time window (within 6 hour) than intravenous thrombolysis.8 Nonetheless, the relatively high rate of symptomatic intracranial haemorrhage (sICH) (10% vs 2% of placebo) and the long procedural time required for adequate recanalisation are of concern.8

ysis showed potentially higher recanalisation rate (66.0%) and longer time window (within 6 hour) than intravenous thrombolysis.8 Nonetheless, the relatively high rate of symptomatic intracranial haemorrhage (sICH) (10% vs 2% of placebo) and the long procedural time required for adequate recanalisation are of concern.8 The results of five recently published trials for AIS herald the dawn of a new era in the treatment of patients with large vessel occlusion amenable to endovascular intervention.9–14 Treatment strategies for patients with AIS with large arterial occlusion have now changed, with the publication of the Multicenter Randomized CLinical trial of Endovascular treatment for Acute ischemic stroke in the Netherlands (MR CLEAN),9 Endovascular treatment for Small Core and Anterior circulation Proximal occlusion with Emphasis on minimizing CT to recanalization times (ESCAPE),10 Extending the time for Thrombolysis in Emergency Neurological Deficits—Intra-Arterial (EXTEND IA),11 Solitaire With the Intention For Thrombectomy as PRIMary Endovascular treatment trial (SWIFT PRIME)12 and Randomized Trial of Revascularization with the Solitaire FR Device versus Best Medical Therapy in the Treatment of Acute Stroke due to Anterior Circulation Large Vessel Occlusion Presenting within 8 hours of Symptom Onset (REVASCAT)13 studies. The aforementioned studies have shown consistent and persuasive benefits for intra artery treatment (IAT) using advanced technology in patients with stroke because of intracranial large artery occlusion. Based on the above trials, the American Heart Association/American Stroke Association (AHA/ASA) renewed the evidence-based guidelines for the selection of patients with AIS for endovascular treatment, the endovascular procedure and for systems of care to facilitate endovascular treatment.15 Certain endovascular procedures have been demonstrated to provide clinical benefit in selected patients with AIS. Systems of care should be organised to facilitate the delivery of this care.

s with AIS for endovascular treatment, the endovascular procedure and for systems of care to facilitate endovascular treatment.15 Certain endovascular procedures have been demonstrated to provide clinical benefit in selected patients with AIS. Systems of care should be organised to facilitate the delivery of this care. However, these trials were carried out in North America and Europe, where cardiac embolism is the most common cause of large vessel strokes. It is unclear if the results of these studies can be reproduced in the Chinese population, which has a high prevalence of intracranial atherosclerosis. No prospective study investigating the efficacy and safety of Solitaire in the Chinese population is available. The Endovascular therapy for Acute ischemic Stroke Trial (EAST) is designed to evaluate the use of Solitaire in patients with moderate-to-severe stroke, within 12 hours of symptom onset, in the Chinese population. Method/design This multicentre prospective trial will involve 17 stroke centres in China, and plans to consecutively recruit 150 patients in the intervention group and 150 patients in the medical group (control). For patients with AIS who are eligible for enrolment, the patient or patient's legally authorised representative will be informed about the written informed consent requirement of the EAST study before enrolment.

nsecutively recruit 150 patients in the intervention group and 150 patients in the medical group (control). For patients with AIS who are eligible for enrolment, the patient or patient's legally authorised representative will be informed about the written informed consent requirement of the EAST study before enrolment. After agreement of participation in the EAST study, the patient or patient's legally authorised representative will be further informed of all the benefits and risks of intervention and conservative treatment, by a standardised trained doctor. If the patient or patient's legally authorised representative decides that the patient should receive the intervention operation and signs the informed consent of intervention operation, the patient will be assigned to the intervention group. If the patient participates in the EAST study but refuses intervention after notified of all the benefits and risk of intervention and conservative treatment, the patient will be assigned to the control group. The intervention patients will be treated for mechanical recanalisation using Solitaire, within 12 hours after stroke onset, plus standard medical therapy. The control arm will be treated with standard medical therapy alone (figure 1). Figure 1 Endovascular therapy for Acute ischemic Stroke Trial (EAST) study enrolment clinical procedure.

The intervention patients will be treated for mechanical recanalisation using Solitaire, within 12 hours after stroke onset, plus standard medical therapy. The control arm will be treated with standard medical therapy alone (figure 1). Figure 1 Endovascular therapy for Acute ischemic Stroke Trial (EAST) study enrolment clinical procedure. During the trial, multiple indicators will be assessed in all patients at baseline, 24 hour, 7 days (or at discharge, whichever occurs first) and 90 days (see online supplementary I). The Solitaire thrombectomy stent system is provided by Covidien (ev3, Irvine, California, USA). The company will not participate in data collection, analysis, editing, nor in publication strategies. This study is sponsored and conducted by the Cerebrovascular Disease Center of Tiantan Hospital (China), which has the additional responsibility of data analysis. An independent Data and Safety Monitoring Board (DSMB) oversees the conduction, safety and efficacy of the study. The study has been registered—registration number NCT02350283. 10.1136/svn-2016-000022.supp1Supplementary data Study status The study protocol has been approved by the Ethics Committee. Patient recruitment started in January 2015 and will end by June 2016. Patient population Intervention group: Any patient with AIS due to large vessel occlusion, indicated for treatment with Solitaire stent retriever within 12 hours of symptoms onset, and meeting the inclusion and exclusion criteria, is eligible (boxes 1 and 2). Box 1 Inclusion criteria Age >18 years

Study status The study protocol has been approved by the Ethics Committee. Patient recruitment started in January 2015 and will end by June 2016. Patient population Intervention group: Any patient with AIS due to large vessel occlusion, indicated for treatment with Solitaire stent retriever within 12 hours of symptoms onset, and meeting the inclusion and exclusion criteria, is eligible (boxes 1 and 2). Box 1 Inclusion criteria Age >18 years Clinical diagnosis of ischaemic stroke, with stroke symptoms present for at least 30 min and no significant improvement before treatment. No prestroke functional dependence (prestroke modified Rankin Scale score ≤1). National Institute of Health Stroke Scale (NIHSS) ≥8 and <30 at the time of enrolment. Patient can be treated within 12 hours of stroke symptoms onset with minimum of one deployment of the Solitaire device (onset time is defined as the last time the patient was witnessed to be at baseline). Patient is confirmed to have symptomatic intracranial occlusion, based on single phase, multiphase or dynamic CT angiography/MR angiography or digital subtraction angiography (DSA), at one or more of the following locations: carotid T/L, M1 segment of the middle cerebral artery (MCA), or M2 segment of the MCA equivalent affecting at least 50% of MCA territory. Patient or patient's legally authorised representative received information about data collection or, if mandatory, has signed and dated an Informed Consent Form.

Patient is confirmed to have symptomatic intracranial occlusion, based on single phase, multiphase or dynamic CT angiography/MR angiography or digital subtraction angiography (DSA), at one or more of the following locations: carotid T/L, M1 segment of the middle cerebral artery (MCA), or M2 segment of the MCA equivalent affecting at least 50% of MCA territory. Patient or patient's legally authorised representative received information about data collection or, if mandatory, has signed and dated an Informed Consent Form. Box 2 Exclusion criteria Baseline non-contrast CT or diffusion weighted imaging (DWI) reveals a moderate/large core defined as extensive early ischaemic changes of Alberta Stroke Program Early CT Score (ASPECTS) 0–6 in the territory of symptomatic intracranial occlusion or DWI lesion volume >50 mL. Other confirmation of a moderate-to-large core defined in one of three ways: On a single phase, multiphase or dynamic CT angiography (CTA): no or minimal collaterals in a region greater than 50% of the middle cerebral artery (MCA) territory when compared with pial filling on the contralateral side (multiphase/dynamic CTA preferred); On CT perfusion (>8 cm coverage): a low cerebral blood volume (CBV) and very low cerebral blood flow (CBF) ASPECTS<6 in the symptomatic MCA territory; On CT perfusion (<8 cm coverage): a region of low CBV and very low CBF>1/3 of the computed tomography perfusion (CTP)-imaged symptomatic MCA territory. Groin puncture is not possible within 70 min of the end of CTA/MR angiography acquisition. Seizure at onset of stroke. Prior stroke within the past 3 months.

On CT perfusion (>8 cm coverage): a low cerebral blood volume (CBV) and very low cerebral blood flow (CBF) ASPECTS<6 in the symptomatic MCA territory; On CT perfusion (<8 cm coverage): a region of low CBV and very low CBF>1/3 of the computed tomography perfusion (CTP)-imaged symptomatic MCA territory. Groin puncture is not possible within 70 min of the end of CTA/MR angiography acquisition. Seizure at onset of stroke. Prior stroke within the past 3 months. Patient with a pre-existing neurological or psychiatric disease that would confound the neurological and functional evaluations. Presumed septic embolus or suspicion of bacterial endocarditis. Life expectancy of <90 days. Known history of intracranial hemorrhage (ICH), subarachnoid haemorrhage, arteriovenous malformation or tumour. Known disease with increased bleeding risk during the past 3 months, for example, severe liver disease, ulcerative gastrointestinal disease, oesophageal varices, hepatic failure. Major surgery, significant trauma or haemorrhagic disease in past 10 days. Uncompensated hypertension defined as systolic blood pressure >185 mm Hg or diastolic blood pressure ≥110 mm Hg on three repeated measurements at least 10 min apart. Renal failure as defined by a serum creatinine level >2.0 mg/dl or glomerular filtration rate <30 ml/min*1.73 m2. Platelet count below 100 000/mm3. Blood glucose <2.8 or >22.2 mmol/L. Patients receiving oral anticoagulants, for example, warfarin sodium, and coagulant response time (international normalised ratio) >1.5.

Uncompensated hypertension defined as systolic blood pressure >185 mm Hg or diastolic blood pressure ≥110 mm Hg on three repeated measurements at least 10 min apart. Renal failure as defined by a serum creatinine level >2.0 mg/dl or glomerular filtration rate <30 ml/min*1.73 m2. Platelet count below 100 000/mm3. Blood glucose <2.8 or >22.2 mmol/L. Patients receiving oral anticoagulants, for example, warfarin sodium, and coagulant response time (international normalised ratio) >1.5. Administration of heparin within the previous 48 hours and activated partial thromboplastin time (APTT) time exceeding the upper limit of normal for laboratory. Suspected intracranial dissection as a cause of stroke. Clinical history, past imaging or clinical judgement suggests that the intracranial occlusion is chronic. No femoral pulses. Contraindications of DSA examination, severe contrast allergy or absolute contraindication to iodinated contrast. Pregnancy; if a woman of childbearing potential has a positive urine or serum β-human chorionic gonadotropin test. Control group: Includes patients who meet enrolment criteria but refuse intervention.

No femoral pulses. Contraindications of DSA examination, severe contrast allergy or absolute contraindication to iodinated contrast. Pregnancy; if a woman of childbearing potential has a positive urine or serum β-human chorionic gonadotropin test. Control group: Includes patients who meet enrolment criteria but refuse intervention. Treatment plan Interventional treatment with Solitaire All procedures will be performed via a femoral artery approach, under local anaesthesia. Conscious sedation will be administered as needed.16 The patients will be intubated only if there is a significant risk of airway compromise. General anaesthesia will be induced if the patient is combative and the procedure cannot proceed even with conscious sedation, or conscious sedation is considered too risky given the patient's medical conditions and airway compromise. An 8 Fr sheath, 90 cm (Cook) will be introduced into the femoral access and placed into the affected common carotid artery. When an 8 Fr balloon guiding catheter is available, the balloon catheter will be placed into the cervical internal carotid artery. Without balloon guiding catheter, the long sheath may be placed into the cervical internal carotid artery if feasible, or a 6–8 Fr guiding catheter may be inserted into the internal carotid artery safely, as high as possible. A 0.21-inch internal diameter microcatheter (such as Prowler Select Plus, Cordis or Vasco 21, Balt) will be navigated distal to the point of occlusion over a 0.014-inch steerable microwire. Gentle contrast injection is required to confirm microcatheter position and exclude perforation. After gently flushing out the contrast with saline, a Solitaire device will be deployed through the microcatheter. Control angiography will be performed to evaluate the correct placement and expansion of the device. The device will be left in place for 5–10 min, allowing full expansion of the nitinol stent through the thrombus. The fully deployed Solitaire device together with the delivery microcatheter will be gently pulled back as a single unit and recovered. Before this retrieval, the balloon guide catheter will be inflated and manual aspiration performed with a 50 mL syringe through the haemostatic valve, to reverse the flow in the target artery and therefore reduce the risk of thromboembolism. Without balloon-guiding catheter, aspiration will be performed through the long sheath or the 8 Fr guiding catheter.

de catheter will be inflated and manual aspiration performed with a 50 mL syringe through the haemostatic valve, to reverse the flow in the target artery and therefore reduce the risk of thromboembolism. Without balloon-guiding catheter, aspiration will be performed through the long sheath or the 8 Fr guiding catheter. Successful recanalisation will be defined as modified Thrombolysis in Cerebral Infarction (mTICI) ≥2b in all treatable vessels. If the treated vessel is not opened to at least mTICI 2b with a maximum of five passes of the thrombectomy device, the treatment will be considered a failure. No intra-arterial fibrinolytics will be administered at any point during this study, even if the recanalisation attempt is unsuccessful (figure 2). Figure 2 Solitaire revascularisation device clinical procedure. BP, blood pressure; CTA, CT angiography; IFU, instruction for use; MRA, MR angiography; mTICI, modified Thrombolysis in Cerebral Infarction; NIHSS, National Institute of Health Stroke Scale.

Successful recanalisation will be defined as modified Thrombolysis in Cerebral Infarction (mTICI) ≥2b in all treatable vessels. If the treated vessel is not opened to at least mTICI 2b with a maximum of five passes of the thrombectomy device, the treatment will be considered a failure. No intra-arterial fibrinolytics will be administered at any point during this study, even if the recanalisation attempt is unsuccessful (figure 2). Figure 2 Solitaire revascularisation device clinical procedure. BP, blood pressure; CTA, CT angiography; IFU, instruction for use; MRA, MR angiography; mTICI, modified Thrombolysis in Cerebral Infarction; NIHSS, National Institute of Health Stroke Scale. The following steps will be taken if there is a suspicion of underlying atherosclerosis occluding the vessel. If there is difficulty in passing the 0.21-inch microcatheter through the vessel occlusion despite adequate proximal support, angioplasty with a 2 mm balloon will be allowed to facilitate the passage of the microcatheter. If underlying stenosis is discovered after removal of the Solitaire device, a repeat angiogram will be performed with different views to confirm that the stenosis is not due to vasospasm or dissection. Next, the patient will be prepared for intracranial angioplasty and/or stenting for underlying intracranial atherosclerosis, using 300 mg of rectal aspirin. A cone beam CT will be obtained in the angiography suite, to exclude intracranial haemorrhage. If not feasible, the patient will be subjected to a regular head CT. After excluding intracranial haemorrhage, primary angioplasty will be performed to improve flow to the distal vascular territory and reduce the risks of immediate reocclusion. A residual stenosis of 40–50% is considered acceptable. The Solitaire stent will be detached for stenting of intracranial stenosis only if there is haemodynamically significant recoil or dissection. If the patient received primary angioplasty, 300 mg of loading dose of clopidogrel will be given to the patient orally or via nasogastric tube immediately after the procedure. Intravenous IIb3a inhibitor will be administered immediately before stenting. The stenting patients will be loaded with clopidogrel as well immediately after the procedure and the IIb3a inhibitor infusion will be stopped 2 hours after the loading dose of clopidogrel. All patients indicated for intracranial angioplasty and/or stenting will be treated as intracranial atherosclerosis disease cases and will be included in the subgroup analysis. An MR angiography or CT angiography (CTA) will be obtained after the procedure, to assess the patency of the target vessel (figure 2).

idogrel. All patients indicated for intracranial angioplasty and/or stenting will be treated as intracranial atherosclerosis disease cases and will be included in the subgroup analysis. An MR angiography or CT angiography (CTA) will be obtained after the procedure, to assess the patency of the target vessel (figure 2). Groin punctures will be routinely closed with an Angio-Seal (St Jude Medical, St Paul, Minnesota, USA). National Institute of Health Stroke Scale (NIHSS) will be noted and blood pressure (BP) assessment carried out immediately after the procedure. Postprocedure, the patients will be admitted to an intensive care unit. Standard medical therapy will be provided to these patients. BP parameters will be set by the operators and are expected to generally be below 180/100 mm Hg.

Groin punctures will be routinely closed with an Angio-Seal (St Jude Medical, St Paul, Minnesota, USA). National Institute of Health Stroke Scale (NIHSS) will be noted and blood pressure (BP) assessment carried out immediately after the procedure. Postprocedure, the patients will be admitted to an intensive care unit. Standard medical therapy will be provided to these patients. BP parameters will be set by the operators and are expected to generally be below 180/100 mm Hg. Standard medical management Patients assigned to the medical arm will receive the following medical treatment according to the AHA Guidelines for the Early Management of Patients with Acute Ischemic Stroke.17 18 All patients will be admitted to a stroke unit with cardiac monitoring. Patients who received intravenous alteplase will be admitted to an intensive care unit. Airway support, ventilation assistance and supplemental oxygen will be provided to maintain oxygen saturation >94%. Hyperthermia (temperature >38°C) will be treated with antipyretics or physical cooling. The source of hyperthermia will be identified and treated accordingly. Hyperglycaemia and hypoglycaemia will be corrected. Prophylaxis for deep venous thrombosis will be implemented. Swallowing evaluation will be performed within 48 hours before administering patients any oral medication or other intake. BP will be maintained below 180/100 mm Hg. Aspirin (300 mg initial dose) will be started within 24 hours in patients not treated with intravenous alteplase or within 24–48 hours in patients who received intravenous alteplase and were found to have no intracranial haemorrhage on follow-up CT. Patients with low density lipoprotein (LDL)>100 mmol/L will be treated with statin. A check list of the above measures will be provided to the individual centre and treatment team and will be monitored closely to ensure standard medical management and that patients enrolled in the medical arm receive a level of care similar to that received by patients with stroke admitted to primary stroke centres in the USA.

ist of the above measures will be provided to the individual centre and treatment team and will be monitored closely to ensure standard medical management and that patients enrolled in the medical arm receive a level of care similar to that received by patients with stroke admitted to primary stroke centres in the USA. Primary endpoint The primary efficacy endpoint is functional independence as defined by a modified Rankin Scale (mRS) score ≤2 at 90 days or by functional improvement as defined by mRS using shift analysis. The procedural efficacy endpoint is arterial recanalisation of the occluded target vessel, measured by mTICI score equal or superior to 2b right following the use of the study device. The primary safety endpoint is sICH with 24±3 hours postprocedure. sICH is defined as parenchymal, subarachnoid or intraventricular haemorrhage detected by CT or MRI, associated with seizures or new neurological symptoms or signs (headache, change in level of consciousness, or focal neurological deficits), lasting at least 24 hours, and is only regarded as a primary endpoint if it occurs within 30 days after revascularisation.

d or intraventricular haemorrhage detected by CT or MRI, associated with seizures or new neurological symptoms or signs (headache, change in level of consciousness, or focal neurological deficits), lasting at least 24 hours, and is only regarded as a primary endpoint if it occurs within 30 days after revascularisation. Secondary endpoints The secondary endpoints are related to the use of the study device and procedure: serious adverse events (SAEs) at 14 days or discharge; volume of cerebral infarction as measured by a CT scan at 24±3 hours postprocedure; arterial reperfusion measured by reperfusion ratio on CT scan 24±3 hours postprocedure; infarction in participants who achieved mTICI 2b-3 reperfusion without intracranial haemorrhage; death due to any cause at 14 days or discharge and at 90 days; change in NIHSS at 24±3 hours postprocedure; change in NIHSS at 14 days or discharge postprocedure; change in NIHSS at 90±7 days; qualify of life at 90±7 days (EuroQol-5 Dimensions, BI=Barthel Index (EQ-5D), BI); the proportion of patients with a safety outcome; proportion of patients with the composite of: (1) sICH, (2) major bleeding due to femoral artery access complications including groin haematoma and retroperitoneal haematoma and (3) contrast nephropathy; economic (cost-effectiveness) analysis; evaluation of waiver/deferral of consent process; total radiation dose (CT, CTA, angiography) reported as a continuous measure; proportion of patients with malignant MCA infarction; proportion of patients undergoing hemicraniectomy; functional independence as defined by mRS score ≤2 at 90 days in patients with intracranial atherosclerosis who received intervention; functional independence as defined by mRS score ≤2 at 90 days in intervention patients excluding intracranial atherosclerosis; arterial recanalisation of the occluded target vessel measured by mTICI score equal or superior to 2b immediately after the use of the study device in patients with intracranial atherosclerosis; mTICI score equal to or better than 2b at 3 hours in patients with intracranial atherosclerosis; mTICI score equal to or better than 2b in intervention patients without intracranial atherosclerosis; sICH within 24±3 hours postprocedure in patients with intracranial atherosclerosis; and sICH within 24±3 hours postprocedure in patients without intracranial atherosclerosis.

patients with intracranial atherosclerosis; mTICI score equal to or better than 2b in intervention patients without intracranial atherosclerosis; sICH within 24±3 hours postprocedure in patients with intracranial atherosclerosis; and sICH within 24±3 hours postprocedure in patients without intracranial atherosclerosis. Exploratory purpose The influence of different classifications of the Trial of Org 10 172 in Acute Stroke Treatment (TOAST) on 90 days functional independence, arterial recanalisation, sICH and procedure time of thrombectomy. The influence of different sex on 90 days functional independence, arterial recanalisation, sICH and SAEs. The influence of different anaesthesia type on 90 days functional independence, arterial recanalisation, sICH and procedure time of thrombectomy. The influence of antiplatelet drugs and platelet membrane glycoprotein II b/III a receptor antagonists on 90 days functional independence, arterial recanalisation and sICH. The influence of stenosis of occluded artery on 90 days functional independence, arterial recanalisation, sICH and procedure time of thrombectomy. The influence of baseline BP and glucose on primary results.

ycoprotein II b/III a receptor antagonists on 90 days functional independence, arterial recanalisation and sICH. The influence of stenosis of occluded artery on 90 days functional independence, arterial recanalisation, sICH and procedure time of thrombectomy. The influence of baseline BP and glucose on primary results. Statistical analysis The primary outcome will be evaluated by a logistic regression model comparing the intervention and medical groups. The following confounding factors will be considered during the analysis: age, sex, stroke severity, stroke subtype, particularly intracranial atherosclerosis, artery of occlusion, infarct volume, blood glucose, BP, medical history, time of start treatment, time of recanalisation, use of antiplatelet therapy, destruction of blood–brain barrier, and biomarkers. Multivariable regressions will be performed adjusting for potential covariates and adjusting for the propensity score.

rtery of occlusion, infarct volume, blood glucose, BP, medical history, time of start treatment, time of recanalisation, use of antiplatelet therapy, destruction of blood–brain barrier, and biomarkers. Multivariable regressions will be performed adjusting for potential covariates and adjusting for the propensity score. For secondary outcomes, a logistic regression model will be used as well. Evaluation of the safety assessments and additional measurements during the trial (including treatment interruption/discontinuation, lifestyle evaluation and use of concomitant medications) will be based on appropriate summary statistics. Full details of analyses to be performed will be itemised in a separate Statistical Analysis Plan (SAP). The propensity score model will be used to improve the sensitivity of analysis. A non-parsimonious multivariable logistic regression model will be used to determine the propensity for thrombectomy regardless of the outcome. All the baseline characters will be included to calculate the propensity score. A propensity score, indicating the predicted probability of receiving thrombectomy, will then be calculated from the logistic equation for each patient. The propensity score will then be included along with the comparison variable (thrombectomy or control) in multivariable analyses of outcome producing an adjusted OR with 95% CI.

y score, indicating the predicted probability of receiving thrombectomy, will then be calculated from the logistic equation for each patient. The propensity score will then be included along with the comparison variable (thrombectomy or control) in multivariable analyses of outcome producing an adjusted OR with 95% CI. Strengths and limitations of this study This multicentre prospective control study will elucidate the safety and efficacy of thrombectomy with Solitaire stent retriever for patients with moderate-to-severe stroke in the Chinese population due to anterior large-vessel occlusion. The thrombectomy will be performed within 12 hours of symptom onset in patients meeting the inclusion and exclusion criteria. However, this study has limitations. It focuses on the Chinese population and requires caution when applying the findings to patients of other ethnicity. The study has inadequate statistical power to assess the impact of the endovascular treatment on major clinical outcomes. Data and Safety Monitoring Board DSMB members are independent of the researchers and the steering committee. DSMB is responsible for assuring that study participants are not exposed to unnecessary risks, and that the study is conducted according to high scientific and ethical standards. The DSMB is responsible for advising early termination of the study in the event of unexpected safety concerns or if treatment differences were apparent at the prespecified interim analyses.

cipants are not exposed to unnecessary risks, and that the study is conducted according to high scientific and ethical standards. The DSMB is responsible for advising early termination of the study in the event of unexpected safety concerns or if treatment differences were apparent at the prespecified interim analyses. The authors extend their special thanks to the physicians as well as all other personnel who dedicated themselves to the EAST project. Twitter: Follow Yilong Wang at @yilong

cipants are not exposed to unnecessary risks, and that the study is conducted according to high scientific and ethical standards. The DSMB is responsible for advising early termination of the study in the event of unexpected safety concerns or if treatment differences were apparent at the prespecified interim analyses. The authors extend their special thanks to the physicians as well as all other personnel who dedicated themselves to the EAST project. Twitter: Follow Yilong Wang at @yilong Collaborators: The EAST investigators include: Steering Committee: Zhongrong Miao (principal investigator); Yongjun Wang (co-principle investigator); Jeffrey L Saver (USA); Joseph P Broderick (USA); Lei Feng (USA); Xingquan Zhao; Liping Liu; Yilong Wang. Executive Committee: Zhongrong Miao; Yongjun Wang; Yilong Wang; Xingquan Zhao; Liping Liu; Xiaoling Liao; Chunjuan Wang. Clinical sites (principal investigator) in order of participants enrolled: Zhongrong Miao (principle investigator, E-mail: zhongrongm@163.com), Beijing Tiantan Hospital; Yongjun Wang (co-principle investigator, email: yongjunwang1962@gmail.com), Beijing Tiantan Hospital; Xiaochuan Huo, Beijing Tiantan Hospital; Xiaoling Liao, Beijing Tiantan Hospital; Xingquan Zhao, Beijing Tiantan Hospital; Chunjuan Wang, Beijing Tiantan Hospital; Liping Liu, Beijing Tiantan Hospital; Yuesong Pan, Beijing Tiantan Hospital; Hao Li, Beijing Tiantan Hospital; Yilong Wang, Beijing Tiantan Hospital; Feng Gao, Beijing Tiantan Hospital; Ning Ma, Beijing Tiantan Hospital; Dapeng Mo, Beijing Tiantan Hospital; Ya Peng, Changzhou No 1 People's Hospital; Hang Lin, Fuzhou PLA General Hospital; Meng Zhang, Daping Hospital; Wei Wu, QiLu Hospital of ShanDong University; Zaiyu Guo, Tianjin Teda Hospital; Li Liu, Chifeng Municipal Hospital; Changchun Jiang, Baotou Central Hospital; Hua Yang, Affiliated Hospital of Guiyang Medical College; Qiyi Zhu, People's Hospital of Linyi City; Xiaoxiang Peng, Hubei Zhongshan Hospital; Yibin Cao, Tangshan Gongren Hospital; Anding Xu, The First Affiliated Hospital of Jinan University; Shengli Chen, Chongqing Sanxia Central Hospital; Cunfeng Song, Liaocheng Third People's Hospital; Liping Wei, Luoyang Central Hospital affiliated to Zhengzhou University; Hui Liang, Yantai Hill Hospital. Data and Safety Monitoring Board: Hao Li, Haipeng Shen, Yong Jiang, Yuesong Pan. Statistical and Data Management Centre: Yilong Wang; Anxin Wang; Gaifen Liu; Xianwei Wang. Events Adjudication Committee: Liebeskind DS (USA); Joseph P Broderick (USA); Lei Feng (USA); Anding Xu; Peiyi Gao; Wengui Yu.

Yantai Hill Hospital. Data and Safety Monitoring Board: Hao Li, Haipeng Shen, Yong Jiang, Yuesong Pan. Statistical and Data Management Centre: Yilong Wang; Anxin Wang; Gaifen Liu; Xianwei Wang. Events Adjudication Committee: Liebeskind DS (USA); Joseph P Broderick (USA); Lei Feng (USA); Anding Xu; Peiyi Gao; Wengui Yu. Funding: This study is sponsored and conducted by the Cerebrovascular Disease Center of Tiantan Hospital, which is also responsible for data analysis. This study is funded by the National Science and Technology Major Project of China (2011BAI08B02, 2015BAI12B04 and 2015BAI12B02). Competing interests: None declared. Provenance and peer review: Not commissioned; externally peer reviewed.

Stroke is a major problem that threatens the health of mankind. In countries with low-level or mid-level income, stroke is an even more serious problem.1 In some countries, such as in China, stroke has become the leading cause of death. Indeed, stroke continues to disable people and destroy lives. However, the battle of fighting strokes has never stopped. Many have fought hard in the past.2 In 1543, Italian scholar Andreas Vaslius described the anatomy of cerebral circulation for the first time. That opened a window for people to explore the mystery of cerebral vasculature. In 1681, English scholar Thomas Willis described the circulatory connections at the skull base that was named after him: the circle of Willis.3 People have been able to gain more in-depth knowledge of brain anatomy since. In 1658, German scholar Johann Jacobus Wepfer published the first book on stroke, the Apoplexia, which was the earliest systematic and theoretical description of cerebrovascular disease. With a series of landmark discoveries thereafter, the Europeans led the world in this field for over 400 years. It was not until after World War II that the prominence of academic work in cerebrovascular diseases crossed the Atlantic and stayed in the USA ever since.

theoretical description of cerebrovascular disease. With a series of landmark discoveries thereafter, the Europeans led the world in this field for over 400 years. It was not until after World War II that the prominence of academic work in cerebrovascular diseases crossed the Atlantic and stayed in the USA ever since. On the basis of the contents and field of exploration, the history of cerebrovascular disease can be divided into three important historical phases: from the 16th to 18th century, the descriptive phase of the cerebrovascular anatomy, physiology and pathology; from the 19th century to the 1960s of the 20th century, the discovery phase of new technology in the diagnosis and treatment of cerebrovascular diseases; after the 1970s of the 20th century, the dawn of multicentre clinical trials, marking the beginning of evidence-based medicine (EBM).

atomy, physiology and pathology; from the 19th century to the 1960s of the 20th century, the discovery phase of new technology in the diagnosis and treatment of cerebrovascular diseases; after the 1970s of the 20th century, the dawn of multicentre clinical trials, marking the beginning of evidence-based medicine (EBM). EBM has brought unprecedented hope to the war against stroke. Results from many large, multi-international centre, prospective and randomised comparative trials (RCTs) have contributed to the building of a solid foundation for the current clinical practice of stroke. Some of the examples of these RCTs include carotid artery endarterectomy, carotid artery angioplasty and stenting, mono and dual antiplatelet therapies, intravenous thrombolysis and intra-arterial thrombectomy, stroke unit, etc. These multicentre research projects have given new meaning to stroke prevention and treatment. Therefore, to a great extent, EBM symbolised the arrival of international collaboration. Multi-international clinical trials have become a sign of the times showing that the world could work together to protect our population from the injury of stroke. In 2015, medical practice has entered into a new phase of precision medicine. Studies of genomics have been implemented into day-to-day stroke therapies. Pharmacogenomics and protein genomics have opened up new opportunities for personalised stroke therapies. The time of precision stroke therapy has arrived in an unannounced fashion.

l practice has entered into a new phase of precision medicine. Studies of genomics have been implemented into day-to-day stroke therapies. Pharmacogenomics and protein genomics have opened up new opportunities for personalised stroke therapies. The time of precision stroke therapy has arrived in an unannounced fashion. While EMB seeks a common strategy to prevent or treat a condition in different populations, precision medicine would treat each patient by relying on the biological difference in each individual. That is why research on therapies based on ethnic differences is now a hot topic.4 Stroke is different between the East and the West. Compared to the Caucasians, Asians have more intracerebral haemorrhages. On the causes of ischaemic stroke, Asians have more intracranial arteriopathy while Caucasians have more carotid diseases. There are also significant pharmacogenomic differences between the East and the West. For example, on the issue of clopidogrel metabolism, Asians or Chinese carry two loss-of-function variant alleles and are classified as ‘poor metabolisers’ due to having little or no CYP2C19 enzyme activity.5 We have certainly learnt that patients would be placed at risk if given a treatment that has only been studied well in a different ethnic population. Therefore, it is about time for us to put our resources together, regardless of our ethnic background or cultural difference, to continue to explore genomic or pharmacogenomic differences in stroke therapy.

ts would be placed at risk if given a treatment that has only been studied well in a different ethnic population. Therefore, it is about time for us to put our resources together, regardless of our ethnic background or cultural difference, to continue to explore genomic or pharmacogenomic differences in stroke therapy. The Chinese Stroke Association (CSA) was officially established on 19 January 2015. The formation of this national disease-specific association is very significant for the country since China has the most number of patients with stroke. In an effort to fight stroke together and unite the talent from China, North America, Australia, Europe and the world, CSA's official English journal Stroke and Vascular Neurology (SVN) has been born. We hope that SVN can record our effort in conquering stroke and offering guidance in stroke clinical practice. Developing international partnership and hand in hand with the world to conquer stroke; this is what SVN is all about! Competing interests: None declared. Provenance and peer review: Commissioned; internally peer reviewed.

Introduction The socioeconomic prospects of Asia have recently taken a dramatic change including changes in lifestyle, diet and urban localisation. This has brought changes in risk of cardiovascular diseases especially stroke.1 2 Prevalence of diabetes and prediabetes has increased in Asia including Pakistan in the past decade, nearly doubling during this era.3–5 Diabetes and prediabetes both show increased risk for ischaemic stroke and are associated with poor patient outcomes.6 7 With the increased risk of diabetes and ischaemic stroke in Asians, the importance of predictive value of glycated haemoglobin (HbA1c) has been an area of active study in recent years.8 9 The aim of this study was to analyse the role of HbA1c in the risk prediction of ischaemic stroke in Pakistani population without diabetes. We further studied the difference between HbA1c values of individuals with and without diabetes with stroke. Subjects and methods This single centre, descriptive, case–control study was carried out on indoor and outdoor patients. Patients were selected from the stroke registry database of the department of neurology while controls from a Health Check-up database, from January 2014 to June 2015. All the patients of both genders ≥12 years of age were included in the study.

iptive, case–control study was carried out on indoor and outdoor patients. Patients were selected from the stroke registry database of the department of neurology while controls from a Health Check-up database, from January 2014 to June 2015. All the patients of both genders ≥12 years of age were included in the study. The study was approved by the hospital ethical committee and carried out according to international ethical standards of the responsible committee on human experimentation and with the latest version of Helsinki Declaration of 1975. All patients underwent a detailed history taking and physical examination and all relevant investigations were performed.

ital ethical committee and carried out according to international ethical standards of the responsible committee on human experimentation and with the latest version of Helsinki Declaration of 1975. All patients underwent a detailed history taking and physical examination and all relevant investigations were performed. The study was divided into two phases for the purpose of statistical analysis; phase I and phase II (figure 1). In phase I, two groups were defined, from the hospital registry, namely patients with stroke, and controls without stroke. Both these groups consisted of individuals without diabetes. A total of 233 patients without diabetes with ischaemic stroke and 245 age and gender frequency matched individuals without stroke selected from a Health Check-up database as controls were enrolled in phase I. HbA1c levels, fasting glucose levels (BSF), random blood glucose (BSR), lipid profiles and blood pressure (BP) were recorded for all individuals. HbA1c levels were recorded at the time of stroke when patients were initially admitted. BSF, BSR, BP and lipid profile values were taken from the database of the registry of these patients which was already available with the hospital and was not more than 4 weeks old (averaged over previous visits). Height and weight were measured and body mass index (BMI) was calculated by dividing the weight (kg) by the height (m2). Smoking history was noted. In phase II, a subset of patients was selected randomly from the stroke registry, half of them being patients with diabetes and the other half being without diabetes (40 each) and their HbA1c, BSF, BSR, lipid profiles and BP were recorded.

lated by dividing the weight (kg) by the height (m2). Smoking history was noted. In phase II, a subset of patients was selected randomly from the stroke registry, half of them being patients with diabetes and the other half being without diabetes (40 each) and their HbA1c, BSF, BSR, lipid profiles and BP were recorded. Figure 1 Summary of selection of study population and phase division.

lated by dividing the weight (kg) by the height (m2). Smoking history was noted. In phase II, a subset of patients was selected randomly from the stroke registry, half of them being patients with diabetes and the other half being without diabetes (40 each) and their HbA1c, BSF, BSR, lipid profiles and BP were recorded. Figure 1 Summary of selection of study population and phase division. All patients taken in either phase with stroke had suffered from focal symptoms of stroke and were observed within 1 week of symptom onset. Relevant lesions on CT or MR diffusion-weighted imaging and apparent diffusion coefficient (MR-DWI and ADC respectively) brain were identified. Individuals without HbA1c and BSF data were excluded from the study. Patients with a history of diabetes, BSF ≥126 mg/dL or fulfilling ADA criteria for diabetes were also excluded from the phase I analysis.4 The diagnosis of stroke was made by the consultant neurologist on the basis of history revealing focal symptoms, clinical examination supporting it and by neuroimaging ie, CT scan and/or MRI brain. Furthermore, patients from the stroke registry with otherwise unclear or suspicious diagnosis of stroke were also not included for analysis. The following laboratory investigations were performed as a part of routine work up of stroke registry: complete blood counts, liver function tests, renal function tests, electrolytes, ECG, Chest X-ray, erythrocyte sedimentation rate, blood sugar fasting (BSF), blood sugar random (BSR), hepatitis serology, urine routine examination, echocardiography, coagulation profile, HbA1c and lipid profile. Lipid profile panel consisted of total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C) and triglyceride cholesterol (TGL-C) levels. For phase II, diagnosis of diabetes was made based on ADA criteria for diabetes. Patients fulfilling the inclusion criteria were enrolled after taking informed written consent from the patients or relatives. The data collected was entered in a specifically designed proforma prospectively using the non-probability, consecutive sampling technique.

diabetes was made based on ADA criteria for diabetes. Patients fulfilling the inclusion criteria were enrolled after taking informed written consent from the patients or relatives. The data collected was entered in a specifically designed proforma prospectively using the non-probability, consecutive sampling technique. All individuals were examined after overnight fasting. The hexokinase method (Autoanalyzer) was used to measure fasting blood glucose levels. An enzymatic colorimetric test was used to measure TC and TG concentrations. The selective inhibition method was used to measure the level of HDL-C and a homogeneous enzymatic calorimetric test was used to measure the level of LDL-C. HbA1c was measured using an immunoturbidimetric assay with an automatic analyser with a reference value of 4.4–6.4%. The methodology was aligned with the Diabetes Control and Complications Trial and National Glycosylated haemoglobin Standardization Program standards. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) was recorded by the standard clinical method at the brachial artery by the auscultatory method using an appropriate size cuff and after at least 5 min of resting in sitting position using a calibrated mercury sphygmomanometer. Individuals were divided into four groups according to the quartiles of HbA1c level for the assessment of risk for ischaemic stroke as follows: 1st quartile=HbA1c≤5.3%; 2nd quartile=5.3<HbA1c≤5.6%; 3rd quartile=5.6<HbA1c≤6.0%; and 4th quartile=HbA1c>6.0%.

All individuals were examined after overnight fasting. The hexokinase method (Autoanalyzer) was used to measure fasting blood glucose levels. An enzymatic colorimetric test was used to measure TC and TG concentrations. The selective inhibition method was used to measure the level of HDL-C and a homogeneous enzymatic calorimetric test was used to measure the level of LDL-C. HbA1c was measured using an immunoturbidimetric assay with an automatic analyser with a reference value of 4.4–6.4%. The methodology was aligned with the Diabetes Control and Complications Trial and National Glycosylated haemoglobin Standardization Program standards. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) was recorded by the standard clinical method at the brachial artery by the auscultatory method using an appropriate size cuff and after at least 5 min of resting in sitting position using a calibrated mercury sphygmomanometer. Individuals were divided into four groups according to the quartiles of HbA1c level for the assessment of risk for ischaemic stroke as follows: 1st quartile=HbA1c≤5.3%; 2nd quartile=5.3<HbA1c≤5.6%; 3rd quartile=5.6<HbA1c≤6.0%; and 4th quartile=HbA1c>6.0%. The data was analysed using SPSS V.17.0. Descriptive analysis was carried out and reported as mean±SD for continuous variables whereas frequencies and percentages were calculated for categorical variables. Associations of the variables and the presence or absence of ischaemic stroke were analysed in phase I. Case–controlled analyses was performed in the two groups to assess the risk of ischaemic stroke according to HbA1c. Comparison of the mean variables like age, BMI, BSF, HbA1c, HDL-C, LDL-C, TGL-C, TC, SBP and DBP, between patients with ischaemic stroke and control individuals was performed using the Student's t-tests. Logistic regression analysis with ischaemic stroke as the dependent variable was performed with age, gender, BMI, BSF, HDL-C, LDL-C, TGL-C, TC, SBP and DBP, smoking and quartile of HbA1c included in the model in phase I.

n patients with ischaemic stroke and control individuals was performed using the Student's t-tests. Logistic regression analysis with ischaemic stroke as the dependent variable was performed with age, gender, BMI, BSF, HDL-C, LDL-C, TGL-C, TC, SBP and DBP, smoking and quartile of HbA1c included in the model in phase I. In phase II, a comparison was made between age, gender, BMI, smoking, BSF, HbA1c, HDL-C, LDL-C, TGL-C, TC, SBP and DBP values on a subset of patients in two groups; with diabetes versus without diabetes with ischaemic stroke (figure 1). p Values <0.05 were considered statistically significant. Results Phase I Phase I comprised 233 patients and 245 controls. The mean age was 54.6±1.6 (range: 15–105) years in the case group and 50.3±1.4 (range: 17–97) in the control group. Among the patients, 133 (57.0%) were males while 100 (42.9%) were females. Among the controls, 130 (53.0%) were males while 115 (46.9%) were females. Mean age and gender were not statistically significantly different between the two groups (p>0.05; table 1). Table 1 Comparison between patients (with stroke) and controls (without stroke) for patients without diabetes: phase I

Results Phase I Phase I comprised 233 patients and 245 controls. The mean age was 54.6±1.6 (range: 15–105) years in the case group and 50.3±1.4 (range: 17–97) in the control group. Among the patients, 133 (57.0%) were males while 100 (42.9%) were females. Among the controls, 130 (53.0%) were males while 115 (46.9%) were females. Mean age and gender were not statistically significantly different between the two groups (p>0.05; table 1). Table 1 Comparison between patients (with stroke) and controls (without stroke) for patients without diabetes: phase I Variables Patients (with stroke) (n=233) Controls (healthy matches) (n=245) p Value Age (years) 54.69±1.60 50.33±1.44 0.96 Gender/sex (males) n (%) 133 (57.08) 130 (53.06) 0.31 Gender/sex (females) n (%) 100 (42.91) 115 (46.93) BSF (mg/dL) 112.43±5.76 98.7±8.98 0.84 HbA1c (%) 5.93±2.97 5.57±1.65 0.03 HDL-C (mg/dL) 43.00±1.95 46.54±2.43 0.02 LDL-C (mg/dL) 96.40±7.96 87.78±9.83 0.07 TGL-C (mg/dL) 221±37.89 176±27.89 0.00 TC (mg/dL) 178.50±27.66 169.76±26.76 0.21 SBP (mm Hg) 147.63±35.64 126.78±23.34 0.04 DBP (mm Hg) 90.56±16.64 81.24±8.90 0.04 BMI (kg/m2) 24.23±2.34 23.95±2.12 0.18 Smoking n (%) 143 (61.37) 52 (21.22) 0.02 Bold indicates significant p Values. Comparison between patients and controls for patients without diabetes: phase I: values are presented as mean±SD.

0.21 SBP (mm Hg) 147.63±35.64 126.78±23.34 0.04 DBP (mm Hg) 90.56±16.64 81.24±8.90 0.04 BMI (kg/m2) 24.23±2.34 23.95±2.12 0.18 Smoking n (%) 143 (61.37) 52 (21.22) 0.02 Bold indicates significant p Values. Comparison between patients and controls for patients without diabetes: phase I: values are presented as mean±SD. BMI, body mass index; BSF, blood sugar fasting; DBP, diastolic blood pressure; HbA1c, glycated haemoglobin; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure; TC, total cholesterol; TGL-C, triglyceride cholesterol. The ischaemic stroke group had significantly higher HbA1c levels (5.9±2.9%) compared to HbA1c levels of the age and gender-matched controls (5.5±1.6%) (p<0.05) (table 1). Mean values for BSF, BMI, TC and LDL-C were not statistically significantly different between the two groups (p>0.05). Mean HDL-C level was significantly lower and mean TGL-C level was significantly higher in patients with ischaemic stroke compared to the control individuals (p<0.05; table 1). The mean SBP and DBP values were significantly higher in patients with ischaemic stroke compared to control individuals (p<0.05; table 1). History of smoking was significant (p<0.05; table 1).

n TGL-C level was significantly higher in patients with ischaemic stroke compared to the control individuals (p<0.05; table 1). The mean SBP and DBP values were significantly higher in patients with ischaemic stroke compared to control individuals (p<0.05; table 1). History of smoking was significant (p<0.05; table 1). When logistic regression analysis was performed with ischaemic stroke being the dependent variable, HDL-C, TGL-C, SBP and DBP, smoking and increasing HbA1c were the significant predictors for ischaemic stroke (p<0.05; table 2). After adjusting for confounding variables, patients in the highest HbA1c quartile showed 7.8-fold increased odds for ischaemic stroke compared with those in the lowest HbA1c quartile (p<0.05; table 2). Table 2 Logistic regression analysis for patients with stroke (patients) as dependent variable: patients without diabetes: phase I

When logistic regression analysis was performed with ischaemic stroke being the dependent variable, HDL-C, TGL-C, SBP and DBP, smoking and increasing HbA1c were the significant predictors for ischaemic stroke (p<0.05; table 2). After adjusting for confounding variables, patients in the highest HbA1c quartile showed 7.8-fold increased odds for ischaemic stroke compared with those in the lowest HbA1c quartile (p<0.05; table 2). Table 2 Logistic regression analysis for patients with stroke (patients) as dependent variable: patients without diabetes: phase I Variables OR CI (95%) p Value Age (years) 1.23 0.76 to 1.14 0.22 Gender/sex (males) 1(reference) – – Gender/sex (females) 1.22 0.77 to 1.43 0.34 BSF (mg/dL) 1.25 0.89 to 1.76 0.27 HDL-C (mg/dL) 0.78 0.87 to 0.99 0.00 LDL-C (mg/dL) 0.88 0.78 to 1.45 0.12 TGL-C (mg/dL) 1.91 0.90 to 1.43 0.00 TC (mg/dL) 0.88 0.78 to 1.98 0.54 SBP (mm Hg) 1.45 1.56 to 1.97 0.01 DBP (mm Hg) 1.33 0.87 to 1.56 0.03 BMI (kg/m2) 1.12 0.97 to 1.24 0.12 Smoking 28.32 11.14 to 62.31 0.02 HbA1c ≤5.3 (%) 1(reference) – – >5.3 HbA1c ≤5.6 (%) 1.67 0.67 to 5.44 0.11 >5.6 HbA1c ≤6.0 (%) 3.46 1.63 to 9.87 0.04 HbA1c >6.0 (%) 7.89 3.24 to 19.59 0.00 Bold indicates significant p Values. Logistic regression analysis for patients with stroke as dependent variable: phase I. BMI, body mass index; BSF, blood sugar fasting; DBP, diastolic blood pressure; HbA1c, glycated haemoglobin; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure; TC, total cholesterol; TGL-C, triglyceride cholesterol.

Variables OR CI (95%) p Value Age (years) 1.23 0.76 to 1.14 0.22 Gender/sex (males) 1(reference) – – Gender/sex (females) 1.22 0.77 to 1.43 0.34 BSF (mg/dL) 1.25 0.89 to 1.76 0.27 HDL-C (mg/dL) 0.78 0.87 to 0.99 0.00 LDL-C (mg/dL) 0.88 0.78 to 1.45 0.12 TGL-C (mg/dL) 1.91 0.90 to 1.43 0.00 TC (mg/dL) 0.88 0.78 to 1.98 0.54 SBP (mm Hg) 1.45 1.56 to 1.97 0.01 DBP (mm Hg) 1.33 0.87 to 1.56 0.03 BMI (kg/m2) 1.12 0.97 to 1.24 0.12 Smoking 28.32 11.14 to 62.31 0.02 HbA1c ≤5.3 (%) 1(reference) – – >5.3 HbA1c ≤5.6 (%) 1.67 0.67 to 5.44 0.11 >5.6 HbA1c ≤6.0 (%) 3.46 1.63 to 9.87 0.04 HbA1c >6.0 (%) 7.89 3.24 to 19.59 0.00 Bold indicates significant p Values. Logistic regression analysis for patients with stroke as dependent variable: phase I. BMI, body mass index; BSF, blood sugar fasting; DBP, diastolic blood pressure; HbA1c, glycated haemoglobin; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure; TC, total cholesterol; TGL-C, triglyceride cholesterol. Phase II Phase II comprised of 40 individuals with diabetes and 40 without diabetes with stroke. Mean age was 55.6±2.2 (range: 19–95) years in diabetes group and 51.7±1.4 (range: 21–96) in without diabetes group. Among the individuals with diabetes, 22 (55%) were males while 18 (45%) were females. Among individuals without diabetes, there were 23 males (57.5%) and 17 females (42.5%). Mean age and gender were not statistically significantly different between the two groups (p>0.05; table 3).

ge: 21–96) in without diabetes group. Among the individuals with diabetes, 22 (55%) were males while 18 (45%) were females. Among individuals without diabetes, there were 23 males (57.5%) and 17 females (42.5%). Mean age and gender were not statistically significantly different between the two groups (p>0.05; table 3). Table 3 Comparison between individuals with diabetes and without diabetes with stroke: phase II Variables Individuals without diabetes Individuals with diabetes p Value Age (years) 51.75±1.48 55.65±2.28 0.78 Gender/sex (males) n (%) 23 (57.50) 22 (55.00) 0.54 Gender/sex (females) n (%) 17 (42.50) 18 (45.00) BSF (mg/dL) 119.56±6.45 134.27±5.78 0.34 HbA1c (%) 6.11±2.31 7.65±2.15 0.00 HDL-C (mg/dL) 47.33±2.32 44.00±5.19 0.67 LDL-C (mg/dL) 96.43±6.46 98.43±8.85 0.93 TGL-C (mg/dL) 165.43±32.44 189.12±25.56 0.55 TC (mg/dL) 166.50±24.56 179.56±25.62 0.35 SBP (mm Hg) 161.00±50.28 172.55±18.29 0.13 DBP (mm Hg) 91.11±19.96 95.87±7.40 0.16 BMI (kg/m2) 24.21±2.34 25.12±13 0.32 Smoking n (%) 16 (40.00) 19 (47.50) 0.11 Bold indicates significant p Values. Comparison between individuals with diabetes and without diabetes with stroke: phase II: values are presented as mean±SD. BMI, body mass index; BSF, blood sugar fasting; DBP, diastolic blood pressure; HbA1c, glycated haemoglobin; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure; TC, total cholesterol; TGL-C, triglyceride cholesterol.

Variables Individuals without diabetes Individuals with diabetes p Value Age (years) 51.75±1.48 55.65±2.28 0.78 Gender/sex (males) n (%) 23 (57.50) 22 (55.00) 0.54 Gender/sex (females) n (%) 17 (42.50) 18 (45.00) BSF (mg/dL) 119.56±6.45 134.27±5.78 0.34 HbA1c (%) 6.11±2.31 7.65±2.15 0.00 HDL-C (mg/dL) 47.33±2.32 44.00±5.19 0.67 LDL-C (mg/dL) 96.43±6.46 98.43±8.85 0.93 TGL-C (mg/dL) 165.43±32.44 189.12±25.56 0.55 TC (mg/dL) 166.50±24.56 179.56±25.62 0.35 SBP (mm Hg) 161.00±50.28 172.55±18.29 0.13 DBP (mm Hg) 91.11±19.96 95.87±7.40 0.16 BMI (kg/m2) 24.21±2.34 25.12±13 0.32 Smoking n (%) 16 (40.00) 19 (47.50) 0.11 Bold indicates significant p Values. Comparison between individuals with diabetes and without diabetes with stroke: phase II: values are presented as mean±SD. BMI, body mass index; BSF, blood sugar fasting; DBP, diastolic blood pressure; HbA1c, glycated haemoglobin; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure; TC, total cholesterol; TGL-C, triglyceride cholesterol. When HbA1c values were compared between individuals with diabetes and individuals without diabetes with stroke, mean HbA1c values were significantly higher in diabetes group (7.6±2.1 vs 6.1±2.3) (p<0.05) but other parameters were not statistically significantly different (p>0.05; table 3).

BMI, body mass index; BSF, blood sugar fasting; DBP, diastolic blood pressure; HbA1c, glycated haemoglobin; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure; TC, total cholesterol; TGL-C, triglyceride cholesterol. When HbA1c values were compared between individuals with diabetes and individuals without diabetes with stroke, mean HbA1c values were significantly higher in diabetes group (7.6±2.1 vs 6.1±2.3) (p<0.05) but other parameters were not statistically significantly different (p>0.05; table 3). Discussion Diabetes mellitus is a metabolic disease prevalent throughout the world and its burden has been increasing with the passage of time.9 10 The diversity of this disease varies greatly throughout the world in respect to cut-off values, prevalence and associations.10 11 The prevalence, incidence and cut-off values for diagnosis differ between Asians and non-Asians. According to international literature, prevalence of diabetes has recently increased from 3.5–6.0% to 9.9–13.1% in Pakistan, that is, in South East Asia.3 The prevalence of prediabetes has shown a similar increasing trend with current value around 9.20%.3 The current prevalence of diabetes and prediabetes in India is between 5.3–13.6% and 8.1–14.6%, respectively.3 Studies have shown that in China the current prevalence of diabetes is around 9.7% while that of prediabetes is around 15.5%.3

prediabetes has shown a similar increasing trend with current value around 9.20%.3 The current prevalence of diabetes and prediabetes in India is between 5.3–13.6% and 8.1–14.6%, respectively.3 Studies have shown that in China the current prevalence of diabetes is around 9.7% while that of prediabetes is around 15.5%.3 HbA1c level is an indicator of mean glucose control in the previous 60–90 days in patients with or without diabetes mellitus. HbA1c level is a good diagnostic tool but is also helpful for prognosis and to monitor therapeutic effects of drugs in diabetes. Well controlled and steady blood glucose levels can help prevent cardiovascular complications in patients with diabetes mellitus.4 Although blacks and whites differ in their HbA1c values, the trend for increased risk of stroke with higher HbA1c values is comparable between the two.8–11 Asian countries comprise almost two-third of the world's total mortality due to stroke.8 10 11 The role of HbA1c in the prediction of ischaemic stroke in individuals without diabetes is underestimated. While little data from around the world is available for review, data from Asia is scarce and from Pakistan is almost non-existent. The primary goal of this study was to analyse the role of HbA1c in the risk prediction of ischaemic stroke in non-diabetic Pakistani population.

viduals without diabetes is underestimated. While little data from around the world is available for review, data from Asia is scarce and from Pakistan is almost non-existent. The primary goal of this study was to analyse the role of HbA1c in the risk prediction of ischaemic stroke in non-diabetic Pakistani population. Current study showed that patients with ischaemic stroke had significantly higher mean HbA1c level (5.9±2.9% vs 5.5±1.6%, p<0.05) compared to their age and gender frequency matched controls in patients without diabetes. The OR for ischaemic stroke increased with the quartile of HbA1c. The highest quartile of HbA1c (>6%) showed nearly eightfold increased odds of having ischaemic stroke compared with the lowest quartile (≤5.3%) (p<0.05) after adjusting for confounding variables. The increased odds with increased HbA1c quartile from first to the fourth suggested that higher HbA1c is associated with increased risk for ischaemic stroke even in individuals without diabetes. Fasting blood glucose on the other hand was not found to be significantly associated with baseline variables (p<0.05) suggestive of its poor predictive power.

HbA1c quartile from first to the fourth suggested that higher HbA1c is associated with increased risk for ischaemic stroke even in individuals without diabetes. Fasting blood glucose on the other hand was not found to be significantly associated with baseline variables (p<0.05) suggestive of its poor predictive power. Comparing our results to the internationally available literature, we found that authors worldwide have shown variable results for association of HbA1c with cardiovascular disease. A study from USA on a cohort of 26 563 patients concluded that HbA1c cannot predict cardiovascular disease (CVD) but only diabetes in patients without diabetes.8 A study by Park et al on 1239 individuals without diabetes demonstrated baseline HbA1as a better predictor of cardiovascular mortality. They compared HbA1c values to fasting plasma glucose and postchallenge plasma glucose levels. According to them, relative hazard for those in the highest quintile of HbA1c (≥6.7%) compared with lower levels was 2.3 for fatal CVD (p=0.00) and 2.4 for ischaemic heart disease (p=0.02).9 In another prospective study, a 1% increase in HbA1c was associated with a relative risk of death from any cause with mortality of 1.2 (p<0.00) in men and 1.2 (p<0.00) in women. They further concluded that 15% of the deaths occurred in persons with diabetes (HbA1c≥7% by their criteria) but 72% occurred in persons with HbA1c concentrations between 5% and 6.9%. The risk for CVD and total mortality associated with increasing HbA1c concentrations increased continuously through their cohort. Most of events noted by them occurred in population having moderately elevated HbA1c values. Therefore, they suggested the need for prospective randomised trials to reduce HbA1c concentrations in persons without diabetes and to see the effect of former on future outcomes regarding stroke.10

uously through their cohort. Most of events noted by them occurred in population having moderately elevated HbA1c values. Therefore, they suggested the need for prospective randomised trials to reduce HbA1c concentrations in persons without diabetes and to see the effect of former on future outcomes regarding stroke.10 A study carried out over a period of more than a decade in individuals without diabetes concluded that increasing HbA1c concentrations are associated almost linearly with increased risk of diabetes, CVD and mortality. The results were significantly profound for HbA1c values compared to fasting glucose levels.11 Myint et al carried out a study on a population of more than 10 000 patients and found that 164 incident strokes were identified over 88 652 person-years. After adjustment for confounding variables (age, sex and cardiovascular risk factors), the relative risks for stroke with HbA1c concentrations 5–5.4%, 5.5–6.9% and ≥7% were 0.7, 0.8 and 2.8, respectively, compared with those with HbA1c <5%. In contrast to previous studies suggesting a linear relationship between blood glucose level and risk of coronary artery disease and stroke, a threshold relationship was advocated by Myint et al.12

ith HbA1c concentrations 5–5.4%, 5.5–6.9% and ≥7% were 0.7, 0.8 and 2.8, respectively, compared with those with HbA1c <5%. In contrast to previous studies suggesting a linear relationship between blood glucose level and risk of coronary artery disease and stroke, a threshold relationship was advocated by Myint et al.12 The famous Hoorn study suggested that when adjusted for gender, hypertension, dyslipidemia and smoking, HbA1c was not significantly associated with the all-cause and cardiovascular mortality.5 Similar results were reported by Blake et al.13 Hjalmarsson et al on the other hand found that HbA1c was a good predictor of acute (p=0.01) and long-term mortality (p=0.02). Furthermore, HbA1c >6% was significantly correlated with severity of acute stroke (p=0.04). It was suggested that it predicted a worse functional outcome at 12 months (p=0.02). The authors reported that poor glycaemic control estimated via baseline HbA1c prior to stroke is an independent risk factor for poor survival. They further concluded that higher baseline HbA1c was a significant biochemical marker for increased stroke severity and unfavourable long-term functional outcome.14

s (p=0.02). The authors reported that poor glycaemic control estimated via baseline HbA1c prior to stroke is an independent risk factor for poor survival. They further concluded that higher baseline HbA1c was a significant biochemical marker for increased stroke severity and unfavourable long-term functional outcome.14 A Mexican study concluded that diabetes and prediabetes are highly prevalent in hospitalised patients with ischaemic stroke. They suggested introducing the routine screening for diabetes and prediabetes via HbA1c testing in all patients with ischaemic stroke.15 Based on their suggestions and our findings, we also advocate the same. Similarly, a study from India suggested that prediction of future risk and prevention strategies for ischaemic stroke could be conceived by utilising HbA1c levels in both diabetic and non-diabetic population.16

all patients with ischaemic stroke.15 Based on their suggestions and our findings, we also advocate the same. Similarly, a study from India suggested that prediction of future risk and prevention strategies for ischaemic stroke could be conceived by utilising HbA1c levels in both diabetic and non-diabetic population.16 Among diabetics, HbA1c has been associated with cardiovascular diseases including coronary artery disease as well as stroke. Zhao et al in their study suggested an association between HbA1c and the risk of stroke among women with type 2 diabetes. At different levels of HbA1c, the trend of association according to them was as follows: for HbA1c <6.0%, 6.0–6.9%, 7.0–7.9%, 8.0–8.9%, 9.0–9.9% and ≥10.0%, trend was 0.9, 1.0, 1.0, 1.1, 1.1 and 1.2 (p=0.66) for men and 1.0, 1.0, 1.0, 1.1, 1.3 and 1.4 (p=0.00) for women, respectively.17 A recent study from Israel reported that HbA1c is directly associated with risk of future stroke. Moreover, they proposed the idea that HbA1c per se improves the predictive accuracy for stroke in patients with diabetes with atrial fibrillation.18

n and 1.0, 1.0, 1.0, 1.1, 1.3 and 1.4 (p=0.00) for women, respectively.17 A recent study from Israel reported that HbA1c is directly associated with risk of future stroke. Moreover, they proposed the idea that HbA1c per se improves the predictive accuracy for stroke in patients with diabetes with atrial fibrillation.18 In our study, mean values for total cholesterol were not statistically significantly different between individuals without diabetes with and without stroke. On the other hand, mean HDL-C level was significantly lower while mean TGL-C level was significantly higher in patients with ischaemic stroke compared to control patients. A similar result was seen when evaluating differences in mean BP levels. The mean SBP and DBP values were significantly higher in patients with ischaemic stroke. On logistic regression analysis, HDL-C, TGL-C and SBP and DBP were the significant predictors for ischaemic stroke. This means that the same patients with higher HbA1c levels who are at risk of future stroke, also have higher mean TGL-C and BP and a low HDL-C. This may suggest that the coexistence of these multiple risk factors might be related and contribute concomitantly towards the risk of stroke.

predictors for ischaemic stroke. This means that the same patients with higher HbA1c levels who are at risk of future stroke, also have higher mean TGL-C and BP and a low HDL-C. This may suggest that the coexistence of these multiple risk factors might be related and contribute concomitantly towards the risk of stroke. When HbA1c values were compared between individuals with diabetes and without diabetes with stroke, mean HbA1c values were significantly higher in the diabetes group (7.6±2.1 vs 6.1±2.3) (p<0.05) but other parameters were not statistically significantly different (p>0.05) (phase II) (table 3). The second part (phase II) findings were meant to point out that even other risk factors did not differ, HbA1c in individuals with diabetes was significantly higher than individuals without diabetes though individuals without diabetes were found to have high HbA1c value as a predictor of stroke. This could signify that diabetics having higher HbA1c than those of individuals without diabetes could be at greater risk of stroke (perhaps >eightfolds which was found for individuals without diabetes) and that higher HbA1c itself above a certain threshold could signify many fold increase in the risk of stroke with increasing HbA1c values. This may suggest that there might be a certain threshold value of HbA1c up to which the trend for other risk factors (namely BP and lipid profile) significantly affects the future risk of stroke, but above it, HbA1c alone might be a more powerful independent predictor. This however needs to be validated via conducting large population-based prospective studies adopting a model based on incremental associations of HbA1c with other risk factors for stroke.

profile) significantly affects the future risk of stroke, but above it, HbA1c alone might be a more powerful independent predictor. This however needs to be validated via conducting large population-based prospective studies adopting a model based on incremental associations of HbA1c with other risk factors for stroke. Limitations Like any other study, our study too had its limitations. First, we selected patients who had a stroke from a single centre registry but our centre being the largest specialised unit in the country renders services to nearly one-fifth of the population of Pakistan. Moreover, baseline characteristics of patients who had a stroke in this study were comparable to those of nationwide stroke data, so the groups could be representative of patients who had a stroke in Pakistan to some extent. The second limitation was that the diagnosis of diabetes was based on history, previous health records, levels of fasting glucose and HbA1c but glucose challenge test was not performed as a part of this study. Although recent guidelines recommend that fasting glucose and HbA1c levels are sufficient to diagnose diabetes, there is still a possibility that individuals with diabetes could have been included in phase I.4

rds, levels of fasting glucose and HbA1c but glucose challenge test was not performed as a part of this study. Although recent guidelines recommend that fasting glucose and HbA1c levels are sufficient to diagnose diabetes, there is still a possibility that individuals with diabetes could have been included in phase I.4 In spite of these limitations, our study presented meaningful results. It was the first to report association of increasing HbA1c with risk for ischaemic stroke in Pakistani adults without diabetes. Moreover, it identified a cut-off value for HbA1c above which the risk of stroke increases even in individuals without diabetes in our setup. We therefore suggest the need of routine HbA1c testing in all patients with ischaemic stroke for secondary prevention and routine screening in hospital registry for patients with high cardiovascular risk for primary prevention. By conducting more sophisticated prospective studies on larger population cohorts, the association between level of HbA1c and risk of ischaemic stroke can be more clearly defined in Pakistani population without diabetes.19–21

eening in hospital registry for patients with high cardiovascular risk for primary prevention. By conducting more sophisticated prospective studies on larger population cohorts, the association between level of HbA1c and risk of ischaemic stroke can be more clearly defined in Pakistani population without diabetes.19–21 Conclusions Higher HbA1c indicated a significantly increased risk for ischaemic stroke after adjusting for other confounding variables in Pakistani population without diabetes. HbA1c quartile >6% showed eightfold increase in risk of stroke compared to quartile ≤5.3%. An HbA1c value above 5.6% (prediabetic range) predicted future risk for ischaemic stroke and efforts to maintain glucose level within the normal range (≤5.6%) especially in individuals with high cardiovascular risk are important. We suggest the need for routine HbA1c testing in all patients with ischaemic stroke for secondary prevention and routine screening in hospital registry for patients with high cardiovascular risk for primary prevention. Funding: This work has been carried out without any grants or funds. It has been completed without external financial support and the expenses whatsoever for the purpose of this study have been contributed solely by the author's themselves and no one else. Competing interests: None declared. Patient consent: Obtained. Ethics approval: Hospital Ethics Committee copy submitted. Provenance and peer review: Commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

Introduction Dual antiplatelet therapy with aspirin and clopidogrel is commonly used in patients with coronary artery disease or post intra-arterial stenting. Prior to the publication of the results from the Clopidogrel in High-Risk Patients with Acute Nondisabling Cerebrovascular Events (CHANCE) trial, routine use of dual antiplatelet therapy was not supported by any evidence and, furthermore, risk of haemorrhagic events outweighed the benefit.1 2 Since publication of results of the CHANCE trial, the finding of the benefit of dual antiplatelet therapy for stroke prevention in patients with minor stroke or high-risk transient ischaemic attack (TIA) has been incorporated into the 2014 American Heart Association (AHA)/American Stroke Association (ASA) stroke prevention guidelines.3 Of 5170 patients enrolled in the CHANCE trial, 101 patients had haemorrhagic events. The overall rate of haemorrhagic complication in CHANCE was 2%, which showed no statistical differences between the two treatment groups. The purpose of this manuscript is to provide a post hoc analysis of all haemorrhagic events in patients who participated in the CHANCE trial.

l, 101 patients had haemorrhagic events. The overall rate of haemorrhagic complication in CHANCE was 2%, which showed no statistical differences between the two treatment groups. The purpose of this manuscript is to provide a post hoc analysis of all haemorrhagic events in patients who participated in the CHANCE trial. Methods Details of the CHANCE trial have been published elsewhere. Briefly, CHANCE was a prospective, multicentre, double-blind, randomised, placebo-controlled trial conducted at 114 centres in China. The trial compared the combination therapy of clopidogrel and aspirin (clopidogrel at an initial dose of 300 mg, followed by 75 mg/day for 90 days, plus aspirin at a dose of 75 mg/day for the first 21 days) versus placebo plus aspirin (75 mg/day for 90 days). Of 5170 patients enrolled who were 40 years or older and able to start the study drug within 24 hours after the onset of a minor ischaemic stroke (defined by a score of 3 or less at the time of randomisation on the National Institutes of Health Stroke Scale), or high-risk TIA (defined as a score of ≥4 at the time of randomisation on the ABCD2), 2584 patients were randomised to the clopidogrel–aspirin group and 2586 to the aspirin group. All patients received open-label aspirin at a clinician-determined dose of 75–300 mg on the first day and had a 90-day follow-up visit in the clinic. The primary efficacy end point has been reported. The primary safety end point was any moderate-to-severe haemorrhagic event according to the definition used in the Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries (GUSTO) trials.4 In GUSTO, severe haemorrhage was defined as fatal or intracranial haemorrhage (ICH) or other haemorrhage causing haemodynamic compromise that required blood or fluid replacement, inotropic support, or surgical intervention. Moderate haemorrhage was defined as bleeding that required transfusion of blood but did not lead to haemodynamic compromise requiring intervention. In this analysis, we used any bleeding as the primary outcome. Brain scans of all ICHs were reviewed and compared with the baseline scans. The size and location of haematoma were described. Haemorrhagic stroke was defined as a haemorrhage that took place in the area where the fresh ischaemic infarction was. All other haemorrhagic events were tabulated and described.

imary outcome. Brain scans of all ICHs were reviewed and compared with the baseline scans. The size and location of haematoma were described. Haemorrhagic stroke was defined as a haemorrhage that took place in the area where the fresh ischaemic infarction was. All other haemorrhagic events were tabulated and described. Statistical methods We compared the baseline characteristics of all patients with and without haemorrhage, and those with minor stroke or TIA. Proportions were used for categorical variables and medians with IQRs were used for continuous variables. A non-parametric Kruskal-Wallis test was used to compare group differences for nominal variables and χ2 tests were used for dichotomous variables. Differences in the rate of any bleeding during the 90-day follow-up period were assessed using Cox proportional hazards regression. Backward selection was used to determine factors associated with any bleeding. Whether the treatment effect differed in stroke subtypes (TIA or minor stroke) was assessed by testing the treatment-by-stroke subtype interaction effect with the use of Cox models. Kaplan-Meier survival curves were also used to illustrate such differences. We used logistic regression to examine whether any bleeding was associated with worsening of patients’ functional outcome, measured by deterioration on the modified Rankin Score (mRS). To evaluate the impact of missing values of mRS, sensitivity analysis was also performed assuming all the missing values of mRS change (mRS at visit3–mRS at visit2) either as ≤0 or ≥1. All tests were two-sided, and a p value of 0.05 was defined as being statistically significant. All statistical analyses were performed using SAS software, V.9.4 (SAS Institute, Cary, North Carolina, USA).

was also performed assuming all the missing values of mRS change (mRS at visit3–mRS at visit2) either as ≤0 or ≥1. All tests were two-sided, and a p value of 0.05 was defined as being statistically significant. All statistical analyses were performed using SAS software, V.9.4 (SAS Institute, Cary, North Carolina, USA). Results Among 5170 patients enrolled within 24 hours after onset of a minor ischaemic stroke or high-risk TIA, a total of 101 (2%) haemorrhagic events were reported, from 50 different hospitals. Table 1 shows baseline characteristics of all patients recruited to CHANCE, randomised to dual or mono antiplatelet groups—those with versus those without haemorrhages. With only a total of 101 haemorrhagic events in the CHANCE trial, there was no difference in the overall rate of ICHs between the two treatment groups. However, univariate analysis showed an increased risk of haemorrhagic events in patients with minor strokes but not in those with TIAs (p=0.03, for the interaction effect). The corresponding Kaplan-Meier survival risk curves demonstrated findings consistent with the analyses using Cox models and showed that most bleeding events occurred in the first 30 days (figure 1). In addition, older age, male gender, and history of aspirin and proton pump inhibitor (PPI) usage were associated with increased risk of haemorrhage regardless of treatment group (survival figure). On the contrary, patients with high body mass index (BMI) had lower risk of haemorrhagic events. Multivariable regression analysis showed that history of aspirin usage and concomitant usage of PPI could predict increased risk of haemorrhage independently. Furthermore, sensitivity test showed that, in patients who entered into CHANCE, diagnosis of minor stroke and experiencing a haemorrhagic event were likely associated with worsening of functional outcome (high mRS, table 2). This worsening of functional outcome was not observed in patients diagnosed with high-risk TIAs.

urthermore, sensitivity test showed that, in patients who entered into CHANCE, diagnosis of minor stroke and experiencing a haemorrhagic event were likely associated with worsening of functional outcome (high mRS, table 2). This worsening of functional outcome was not observed in patients diagnosed with high-risk TIAs. Figure 1 Kaplan-Meier survival curves demonstrate cumulative hemorrhagic events by treatment assignment for TIA and minor stroke. TIA, transient ischaemic attack. Table 1 Baseline characteristics of patients on dual antiplatelet therapy—with versus without haemorrhages

urthermore, sensitivity test showed that, in patients who entered into CHANCE, diagnosis of minor stroke and experiencing a haemorrhagic event were likely associated with worsening of functional outcome (high mRS, table 2). This worsening of functional outcome was not observed in patients diagnosed with high-risk TIAs. Figure 1 Kaplan-Meier survival curves demonstrate cumulative hemorrhagic events by treatment assignment for TIA and minor stroke. TIA, transient ischaemic attack. Table 1 Baseline characteristics of patients on dual antiplatelet therapy—with versus without haemorrhages Summary Minor stroke TIA No bleed (n=5069) Any bleed (n=101) p Value No bleed Any bleed p Value No bleed Any bleed p Value All 2524 (49.8%) 60 (59.4%) 0.0557 1821 (49.8%) 46 (66.7%) 0.0055 703 (49.8%) 14 (43.8%) 0.5019 Age 62.3 (54.6–71.2) 64.5 (57.5–74.0) 0.0114 62.0 (54.7–71.2) 65.8 (56.3–76.4) 0.0212 62.5 (54.6–71.5) 62.1 (59.9–71.5) 0.3213 Gender 3364 (66.4%) 56 (55.4%) 0.0217 2455 (67.1%) 41 (59.4%) 0.1761 909 (64.3%) 15 (46.9%) 0.0420 BMI 24.5 (22.8–26.5) 23.9 (22.0–25.4) 0.0056 24.5 (22.7–26.4) 23.7 (22.0–25.6) 0.0162 24.6 (22.9–26.6) 24.0 (22.1–25.1) 0.1643 SBP 150 (136–161) 150 (133–160) 0.7271 150 (140–165) 150 (133–163) 0.5608 145 (130–j160) 146 (135–160) 0.6827 DBP 90 (80–100) 87 (80–99) 0.4787 90 (80–100) 90 (80–100) 0.4947 87 (80–95) 82.5 (80–97.5) 0.8785 Medical history Stroke 1019 (20.1%) 14 (13.9%) 0.1204 769 (21.0%) 12 (17.4%) 0.4615 250 (17.7%) 2 (6.3%) 0.0916 TIA 170 (3.4%) 4 (4.0%) 0.7378 70 (1.9%) 2 (2.9%) 0.5565 100 (7.1%) 2 (6.3%) 0.8566 MI 94 (1.9%) 2 (2.0%) 0.9261 72 (2.0%) 2 (2.9%) 0.5837 22 (1.6%) 0 (0.0%) 0.4769 Hypertension 3334 (65.8%) 65 (64.4%) 0.7665 2352 (64.3%) 44 (63.8%) 0.9227 982 (69.5%) 21 (65.6%) 0.6383 Diabetes 1069 (21.1%) 24 (23.8%) 0.5147 749 (20.5%) 16 (23.2%) 0.5821 320 (22.6%) 8 (25.0%) 0.7533 Hyperlipidaemia 564 (11.1%) 9 (8.9%) 0.4825 370 (10.1%) 4 (5.8%) 0.2365 194 (13.7%) 5 (15.6%) 0.7583 Smoking 2186 (43.1%) 35 (34.7%) 0.0886 1606 (43.9%) 27 (39.1%) 0.4262 580 (41.0%) 8 (25.0%) 0.0677 ETOH use 1565 (30.9%) 35 (34.7%) 0.4159 1140 (31.2%) 27 (39.1%) 0.1585 425 (30.1%) 8 (25.0%) 0.5352 mRS at discharge 4174 (82.3%) 89 (88.1%) 0.2710 2942 (80.5%) 61 (88.4%) 0.1587 1232 (87.2%) 28 (87.5%) 0.7475 NIHSSS at discharge 1485 (29.3%) 35 (34.7%) 0.6061 381 (10.4%) 9 (13.0%) 0.6131 1104 (78.1%) 26 (81.3%) 0.2900 History of aspirin use 565 (11.1%) 19 (18.8%) 0.0160 355 (9.7%) 14 (20.3%) 0.0036 210 (14.9%) 5 (15.6%) 0.9045 Taking proton pump inhibitors 39 (0.8%) 7 (6.9%) <0.0001 30 (0.8%) 6 (8.7%) <0.0001 9 (0.6%) 1 (3.1%) 0.0932 BMI, body mass index; DBP, diastolic blood pressure; ETOH, alcohol; MI, myocardial infarction; mRS, modified Rankin Score; NIHSSS, NIH Stroke Scale Score; SBP, systolic blood pressure; TIA, transien

5.6%) 0.9045 Taking proton pump inhibitors 39 (0.8%) 7 (6.9%) <0.0001 30 (0.8%) 6 (8.7%) <0.0001 9 (0.6%) 1 (3.1%) 0.0932 BMI, body mass index; DBP, diastolic blood pressure; ETOH, alcohol; MI, myocardial infarction; mRS, modified Rankin Score; NIHSSS, NIH Stroke Scale Score; SBP, systolic blood pressure; TIA, transien t ischaemic attack. Table 2 Worsening of mRS once a haemorrhagic event has taken place Population Covariate No bleeding Any bleeding OR (95% CI) Minor stroke Analysis 1 197/3280 (6.0%) 10/43 (23.3%) 4.29 (2.03 to 9.03) Analysis 2 197/3656 (5.4%) 10/69 (14.5%) 2.77 (1.38 to 5.55) Analysis 3 573/3656 (15.7%) 36/69 (52.2%) 5.30 (3.21 to 8.75) TIA Analysis 1 60/1278 (4.7%) 2/28 (7.1%) 1.76 (0.40 to 7.80) Analysis 2 60/1413 (4.2%) 2/32 (6.3%) 1.55 (0.35 to 6.80) Analysis 3 195/1413 (13.8%) 6/32 (18.8%) 1.49 (0.60 to 3.69) Overall Analysis 1 257/4558 (5.6%) 12/71 (16.9%) 3.19 (1.69 to 6.01) Analysis 2 257/5069 (5.1%) 12/101 (11.9%) 2.36 (1.28 to 4.37) Analysis 3 768/5069 (15.2%) 42/101 (41.6%) 3.70 (2.45 to 5.60) Analysis 1: using data with mRS change information available. Analysis 2: sensitivity analysis, assuming the missing values of mRS change as 0 (≤0). Analysis 3: sensitivity analysis, assuming the missing values of mRS change as 1 (≥1). mRS change: mRS at visit 3−mRS at visit 2. All estimates were adjusted by age, gender, BMI, aspirin usage before randomisation, PPI usage. BMI, body mass index; mRS, modified Rankin Score; PPI, proton pump inhibitor; TIA, transient ischaemic attack.

Analysis 3: sensitivity analysis, assuming the missing values of mRS change as 1 (≥1). mRS change: mRS at visit 3−mRS at visit 2. All estimates were adjusted by age, gender, BMI, aspirin usage before randomisation, PPI usage. BMI, body mass index; mRS, modified Rankin Score; PPI, proton pump inhibitor; TIA, transient ischaemic attack. Detailed analysis showed that the clopidogrel–aspirin group had 60 (2.3%) cases and aspirin group had 41 (1.6%, p=0.09) cases of haemorrhagic events. Table 3 summarises all non-ICH events; the timeline to the first haemorrhagic event is shown in figure 1. The sum of haemorrhagic events from the groups in minor stroke or TIA showed that moderate or severe haemorrhage occurred in 7 patients (0.3%) in the clopidogrel–aspirin group and 8 (0.3%) in the aspirin group (p=0.73). There were totally 35 (0.7%) cases of ICHs, 20 (0.4%) in the clopidogrel–aspirin group and 16 (0.3%) in the aspirin group. These included 12 cases of microhaemorrhages (figure 1), 10 cases of intracerebral haematoma, 8 cases of haemorrhagic transformation and 5 cases of haemorrhagic infarctions. Among them, 8 (0.3%) cases in each group were symptomatic. Other common haemorrhagic events included 24 (0.5%) cases of skin bruises/petechia; 13 (0.3%) gastrointestinal (GI) haemorrhages; 9 (0.2%) gum haemorrhages; 8 (0.2%) intraocular haemorrhages; and 1 each of vaginal, oral, puncture site and upper respiratory bleeding. Details are listed in table 3. In addition, figure 1 shows the timing of the first haemorrhagic event in patients enrolled in each of the four treatment groups.

nal (GI) haemorrhages; 9 (0.2%) gum haemorrhages; 8 (0.2%) intraocular haemorrhages; and 1 each of vaginal, oral, puncture site and upper respiratory bleeding. Details are listed in table 3. In addition, figure 1 shows the timing of the first haemorrhagic event in patients enrolled in each of the four treatment groups. Table 3 Analysis of non-intracranial haemorrhagic events Minor stroke TIA Overall Covariate Aspirin (n=1858) Clopidogrel–aspirin (n=1867) Aspirin (n=728) Clopidogrel–aspirin (n=717) Aspirin (n=2586) Clopidogrel–aspirin (n=2584) Total 23 (1.24%) 46 (2.46%) 18 (2.47%) 14 (1.95%) 41 (1.59%) 60 (2.32%) Epistaxis 1 (0.05%) 3 (0.16%) 2 (0.27%) 1 (0.14%) 3 (0.12%) 4 (0.15%) Gastrointestinal bleeding 4 (0.22%) 6 (0.32%) 1 (0.14%) 2 (0.28%) 5 (0.19%) 8 (0.31%) Gum bleeding 1 (0.05%) 4 (0.21%) 4 (0.55%) 0 (0.00%) 5 (0.19%) 4 (0.15%) Haemoptysis 0 (0.0%) 0 (0.0%) 0 (0.00%) 1 (0.14%) 0 (0.00%) 1 (0.04%) Intraocular haemorrhage 1 (0.05%) 4 (0.21%) 2 (0.27%) 1 (0.14%) 3 (0.12%) 5 (0.19%) Intracranial haemorrhage 12 (0.65%) 17 (0.91%) 4 (0.55%) 3 (0.42%) 16 (0.62%) 20 (0.77%) Oral haemorrhage 1 (0.05%) 0 (0.00%) 0 (0.0%) 0 (0.0%) 1 (0.04%) 0 (0.00%) Puncture site bleeding 0 (0.0%) 0 (0.0%) 1 (0.14%) 0 (0.00%) 1 (0.04%) 0 (0.00%) Skin bruises 3 (0.16%) 11 (0.59%) 4 (0.55%) 6 (0.84%) 7 (0.27%) 17 (0.66%) Vaginal bleeding 0 (0.00%) 1 (0.05%) 0 (0.0%) 0 (0.0%) 0 (0.00%) 1 (0.04%) TIA, transient ischaemic attack.

(0.05%) 0 (0.00%) 0 (0.0%) 0 (0.0%) 1 (0.04%) 0 (0.00%) Puncture site bleeding 0 (0.0%) 0 (0.0%) 1 (0.14%) 0 (0.00%) 1 (0.04%) 0 (0.00%) Skin bruises 3 (0.16%) 11 (0.59%) 4 (0.55%) 6 (0.84%) 7 (0.27%) 17 (0.66%) Vaginal bleeding 0 (0.00%) 1 (0.05%) 0 (0.0%) 0 (0.0%) 0 (0.00%) 1 (0.04%) TIA, transient ischaemic attack. Discussions Based on the results of the CHANCE trial, the current AHA secondary stroke prevention guideline has recommended dual antiplatelet therapy for 21 days in patients with minor strokes or high-risk TIAs. While efficacy of dual antiplatelet therapy for secondary coronary event or stroke prevention has been proven, the issue of elevated haemorrhagic events has always been of concern. Through a literature search (PubMed, MEDLINE, Google Scholar, EMBASE), eight clinical trials were found that had, prior to the publication of the results of the CHANCE trial, tested either the combination of clopidogrel and aspirin, cilostazol and aspirin versus clopidogrel alone, or aspirin alone, for coronary artery disease or stroke/TIA prevention. The overall rate of haemorrhagic complications ranged from 1.8% to 8.1%. Please see table 4 for details.5–13 Table 4 Summary of the haemorrhagic risk of eight trials that tested dual versus single antiplatelet agent

rget. Of note, in addition to SENPs, three new SUMO proteases (DeSI-1, DeSI-2 and USPL1) have been identified.14 15 They appear to have de-conjugation capability on only a few particular substrates, however. Indeed, in contrast to SENPs, silencing these enzymes has no obvious effect on global SUMOylation in cells.14 15 SUMOylation in human diseases SUMOylation regulates almost all major cellular processes including gene expression, DNA damage repair, RNA processing and quality control of newly synthesised proteins—all of which are essential for maintaining cellular homoeostasis. Numerous studies have now provided evidence that links SUMOylation to the pathophysiology of many diseases including cancer, heart diseases and brain ischaemia.

Discussions Based on the results of the CHANCE trial, the current AHA secondary stroke prevention guideline has recommended dual antiplatelet therapy for 21 days in patients with minor strokes or high-risk TIAs. While efficacy of dual antiplatelet therapy for secondary coronary event or stroke prevention has been proven, the issue of elevated haemorrhagic events has always been of concern. Through a literature search (PubMed, MEDLINE, Google Scholar, EMBASE), eight clinical trials were found that had, prior to the publication of the results of the CHANCE trial, tested either the combination of clopidogrel and aspirin, cilostazol and aspirin versus clopidogrel alone, or aspirin alone, for coronary artery disease or stroke/TIA prevention. The overall rate of haemorrhagic complications ranged from 1.8% to 8.1%. Please see table 4 for details.5–13 Table 4 Summary of the haemorrhagic risk of eight trials that tested dual versus single antiplatelet agent Trials Antiplatelet agent Major or moderate haemorrhage Haemorrhagic event Haemorrhagic complication Minor haemorrhage CURE Clopidogrel 300 mg load+ various doses of ASA (%) for 3–12 months 3.7 2.1 ASA 75–325 mg (%) for 3–12 months 2.7 1.8 SPS3 Clopidogrel 75 mg+ ASA 325 mg for 8 years (%) 2.1 ASA 325 mg for 8 years (%) 1.1 CHARISMA Clopidogrel 75 mg+ASA 75–162 mg for 28 months 2.1 ASA 75–162 mg for 28 months 1.3 MATCH Clopidogrel 75 mg+ ASA75 mg (%) for 18 months   8.1 Clopidogrel 75 mg for 18 months   3.5 CLAIR Clopidogrel 300 loading, then 75 mg+ ASA 75–160 mg for 7 days   2 cases ASA 75–160 mg for 7 days   none Korean Cilostazol 100 mg twice daily+ ASA 75–160 mg for 7 months (%) 0.9 Clopidogrel 75 mg+ ASA 75–160 mg for 7 months (%) 2.6 SAMMPRIS Recent stroke or TIA (within 30 days) attributable to severe stenosis (70–99%) of a major intracranial artery, clopidogrel 75 mg+ASA 325 mg for 90 days 1.8 CARESS Clopidogrel 300 mg loading followed by 75 mg for 7 days   2 of 52 cases ASA 75 mg for 7 days   none CHANCE Clopidogrel 300 mg load followed by 75 mg for 90 days +ASA 75 mg for the initial 21 days (%) 2.3 ASA 75 mg for the initial 21 days (%) 1.6 ASA, American Stroke Association; CARESS, Clopidogrel and Aspirin for Reduction of Emboli in Symptomatic Carotid Stenosis; CHANCE, Clopidogrel in High-Risk Patients with Acute Nondisabling Ischemic Cerebrovascular Events; CHARISMA, Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance; CURE, Clopidogrel in the Unstable Angina to Prevent Recurrent Events; SPS3, Secondary Prevention of Small Subcortical Strokes; TIA, transient ischaemic attack.

Acute Nondisabling Ischemic Cerebrovascular Events; CHARISMA, Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance; CURE, Clopidogrel in the Unstable Angina to Prevent Recurrent Events; SPS3, Secondary Prevention of Small Subcortical Strokes; TIA, transient ischaemic attack. There are a few possible explanations for why patients of older age, males and those with a history of aspirin or PPI usage would have increased risk of haemorrhage regardless of treatment group. First, both aspirin and clopidogrel would increase the risk of haemorrhagic event used either alone or in combination. This increased risk of haemorrhage was mainly from GI bleeding (table 3). A slightly increased risk of haemorrhage was seen in the Clopidogrel versus Aspirin in Patients at Risk of Ischaemic Events (CAPRIE) trial. CAPRIE compared aspirin 325 mg to clopidogrel 75 mg, and the rate of haemorrhagic event was 2.69% in patients on aspirin 325 mg daily and 2.19% in patients on clopidogrel 75 mg daily. In the aspirin group, 1.55% had some bleeding disorder, 0.43% had ICHs and 0.71% had GI bleeding. In the clopidogrel group, 1.38% had some bleeding disorder, 0.32% had ICHs and 0.49% had GI bleeding.14 In the CHANCE trial, it is possible that these patients may already have had different degree of gastritis or peptic ulcer disease prior to enrolment and therefore were at risk after entering into the trial regardless of the treatment group. Those who had GI bleeding were perhaps not compliant with the PPI treatment and, therefore, the gastritis was not adequately treated.

may already have had different degree of gastritis or peptic ulcer disease prior to enrolment and therefore were at risk after entering into the trial regardless of the treatment group. Those who had GI bleeding were perhaps not compliant with the PPI treatment and, therefore, the gastritis was not adequately treated. Second, the interaction between aspirin and clopidogrel may play a role in potentiating haemorrhagic risk. Salicylic acid is extensively bound to plasma albumin, and causes displacement of other drugs from plasma protein.15 Clopidogrel is converted to a minor (10–15%) active thiol metabolite and major inactive carboxyl metabolite by hepatic cytochrome P450 enzymes. Both metabolites of clopidogrel are extensively protein bound.16 It is possible that salicylic acid may displace the active metabolite of clopidogrel from its protein binding site, which would increase its potency and lead to enhanced inhibition of platelet aggregation.17 Such aspirin interaction with clopidogrel might not be as prominent when aspirin is combined with cilostazol. As discussed above, cilostazol plus aspirin had less chance of haemorrhagic event (0.9%) versus clopidogrel plus aspirin (2.9%).14

increase its potency and lead to enhanced inhibition of platelet aggregation.17 Such aspirin interaction with clopidogrel might not be as prominent when aspirin is combined with cilostazol. As discussed above, cilostazol plus aspirin had less chance of haemorrhagic event (0.9%) versus clopidogrel plus aspirin (2.9%).14 Third, it has been reported that higher doses of aspirin correlate to higher chance of haemorrhagic event. With Clopidogrel in the Unstable Angina to Prevent Recurrent Events (CURE) trial, major haemorrhagic event rates in the dual therapy group were dose-dependent on the aspirin: <100 mg=2.6%; 100 to 200 mg=3.5%; >200 mg=4.9%. Major bleeding event rates for aspirin alone were dose dependent on aspirin too: <100 mg=2.0%; 100 to 200 mg=2.3%; >200 mg=4.0%. However, it was unclear if giving a loading dose of clopidogrel in addition to aspirin could cause more haemorrhagic events in dual therapy. Since neither the platelet function nor aggregation of these large antiplatelet trials was actually tested, it is possible that some patients or section of the population are very sensitive to antiplatelet agents and developing haemorrhagic events.

dition to aspirin could cause more haemorrhagic events in dual therapy. Since neither the platelet function nor aggregation of these large antiplatelet trials was actually tested, it is possible that some patients or section of the population are very sensitive to antiplatelet agents and developing haemorrhagic events. Last, advanced age alone has been related to an increased haemorrhagic event rate. Our analysis confirmed the findings from the CURE trial and another trial adding clopidogrel to aspirin in 45 852 patients with acute myocardial infarction: the Clopidogrel and Metoprolol in Myocardial Infarction Trial (COMMIT).18 In the CURE trial, major haemorrhage event rates for dual antiplatelet therapy by age were: <65 years=2.5%, ≥65 to <75 years=4.1%, ≥75 years=5.9%. Major bleeding event rates for aspirin alone by age were: <65 years=2.1%, ≥65 to <75 years=3.1%, ≥75 years=3.6%. A similar trend was seen in the COMMIT trial. Haemorrhagic rates for dual antiplatelet therapy by age were: <60 years=0.3%, ≥60 to <70 years=0.7%, ≥70 years=0.8%. Haemorrhagic rates for aspirin alone by age were: <60 years=0.4%, ≥60 to <70 years=0.6%, ≥70 years=0.7%. There was also likely an interaction between the rate of haemorrhagic events and severity of stroke plus duration of dual therapy as seen in the aspirin and clopidogrel compared with clopidogrel alone after recent ischaemic stroke or TIA in high-risk patients (MATCH) trial, but not in the CHANCE trial. The CHANCE trial proved that those patients having TIAs and minor strokes would benefit from 21 days of dual therapy in secondary stroke prevention without increased risk of haemorrhage. In other trials, as summarised above, such benefit was offset by increased rate of haemorrhage if patients with all kinds of strokes were included and the duration of therapy was longer.

As and minor strokes would benefit from 21 days of dual therapy in secondary stroke prevention without increased risk of haemorrhage. In other trials, as summarised above, such benefit was offset by increased rate of haemorrhage if patients with all kinds of strokes were included and the duration of therapy was longer. It is perceivable that patients with high BMI would have less drug–drug interaction or drug displacement. Therefore, it is likely that protein binding alterations in the active thiol metabolite could explain a significant amount of interindividual variability associated with clopidogrel.19 Higher BMI with higher protein binding of ASA and less displacement of thiol metabolite could possibly explain why higher BMI had less haemorrhagic events. That is also why taking aspirin or clopidogrel separately would have a lower rate of haemorrhagic events.14 Taking both together likely had an additive effect that potentiated the risk of developing haemorrhage. The mechanism accounting for the trend of more haemorrhagic events in males is unclear. It is contrary to the published results stating the female gender has increased haemorrhagic events in acute coronary syndrome. More research is needed to examine this phenomenon.20

at potentiated the risk of developing haemorrhage. The mechanism accounting for the trend of more haemorrhagic events in males is unclear. It is contrary to the published results stating the female gender has increased haemorrhagic events in acute coronary syndrome. More research is needed to examine this phenomenon.20 It is difficult to explain why those patients who entered into the CHANCE trial—with a diagnosis of minor stroke but not high-risk TIA, and who experienced a haemorrhagic event—had worsening of functional outcome. It could be related to the use of mRS to assess functional outcome. Since patients with TIA had no neurological deficit at baseline, a non-ICH event would not cause any neurological deficit, while in patients with minor stroke, their baseline mRS was either at one or two. This worsening therefore was a reflection of the difference in the baseline mRS of patients with either TIA or minor stroke at enrolment.

A had no neurological deficit at baseline, a non-ICH event would not cause any neurological deficit, while in patients with minor stroke, their baseline mRS was either at one or two. This worsening therefore was a reflection of the difference in the baseline mRS of patients with either TIA or minor stroke at enrolment. The assessment of clinical significance of haemorrhagic events by GUSTO classification may have its limitations. The use of the GUSTO classification can be very subjective. The classification was not specific for type of haemorrhage but, rather, for severity. A more detailed description of the type of haemorrhage may be useful, as was provided in this analysis. A minor or moderate haemorrhagic event can evolve into a severe event if bleeding continues. For example, GI bleeding could be classified as a minor event but, if a large amount of blood loss continues, it could be classified as a major bleeding event. On the other hand, an intraocular haemorrhagic event may not be classified as a severe bleeding event but certainly could be very disabling. Our analysis of the 101 patients with haemorrhagic events in the CHANCE trial showed that even short-term dual therapy would increase the risk of future haemorrhagic events (not ICHs) in patients with a diagnosis of minor strokes but not TIAs. These events were likely minor, as classified by GUSTO classification, but could be clinically important, such as in the event of intraocular haemorrhage, GI bleeding or skin bruises. It is unclear why in CHANCE patients with higher BMI had lower rate of haemorrhagic events.

The assessment of clinical significance of haemorrhagic events by GUSTO classification may have its limitations. The use of the GUSTO classification can be very subjective. The classification was not specific for type of haemorrhage but, rather, for severity. A more detailed description of the type of haemorrhage may be useful, as was provided in this analysis. A minor or moderate haemorrhagic event can evolve into a severe event if bleeding continues. For example, GI bleeding could be classified as a minor event but, if a large amount of blood loss continues, it could be classified as a major bleeding event. On the other hand, an intraocular haemorrhagic event may not be classified as a severe bleeding event but certainly could be very disabling. Our analysis of the 101 patients with haemorrhagic events in the CHANCE trial showed that even short-term dual therapy would increase the risk of future haemorrhagic events (not ICHs) in patients with a diagnosis of minor strokes but not TIAs. These events were likely minor, as classified by GUSTO classification, but could be clinically important, such as in the event of intraocular haemorrhage, GI bleeding or skin bruises. It is unclear why in CHANCE patients with higher BMI had lower rate of haemorrhagic events. Conclusion Aspirin plus clopidogrel therapy in the CHANCE trial did not increase the risk of ICH. However, dual antiplatelet therapy demonstrated a trend of developing other types of haemorrhagic events. Such a trend could be potentiated if the patient is an older male with a minor stroke and has been on aspirin and/or PPI in the past. When considering dual antiplatelet therapy for secondary stroke prevention in patients with minor stroke or TIA, the lower the dose of aspirin, likely the less chance of having a haemorrhagic event. In future designing of antiplatelet clinical trials that test antiplatelet agents, a more detailed description of the types of haemorrhage and testing of platelet aggregation may provide us with better understanding of the pharmacological and biological impact of these agents, assisting clinicians in selecting antiplatelet drugs for their patients with stroke.

al trials that test antiplatelet agents, a more detailed description of the types of haemorrhage and testing of platelet aggregation may provide us with better understanding of the pharmacological and biological impact of these agents, assisting clinicians in selecting antiplatelet drugs for their patients with stroke. The authors would like to express appreciation to all participating institutes and clinicians. This study was sponsored by grants (2008ZX09312-008, 200902004, 2011BAI08B02, 2012ZX09303 and 2013BAI09B03) from the Ministry of Science and Technology of the People's Republic of China (PRC). In addition, the study was supported by the Excellence in Young Investigator Projects (81322019, 81301015 and 81371274) of the Ministry of Science and Technology of the PRC and the Beijing Institute for Brain Disorders (BIBD-PXM2013_014226_07_000084). Twitter: Follow Yilong Wang at @yilong Funding: Ministry of Science and Technology of China; Beijing Institute for Brain Disorders. Competing interests: None declared. Patient consent: Obtained. Ethics approval: Tiantan Hospital Ethics Committee. Provenance and peer review: Not commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

Introduction Stroke is one of the leading causes of disability and mortality in the ageing population worldwide.1 Ischaemia or hypoxic brain injury often results in irreversible brain damage, mediated by multiple mechanisms including oxidative stress, inflammatory response, apoptosis, neuronal death, etc.2–4 Priapism is a persistent, painful penile erection, which can last for hours in the absence of sexual stimulation. It is found in 40% of patients with sickle cell disease, because of impaired drainage of the corpora cavernosa.5 6 Priapism often leads to erectile tissue fibrosis and ultimately results in erectile dysfunction.7 Several research groups studied the molecular mechanisms of priapism to obtain a better understanding of this disorder both experimentally and clinically. They found that opiorphin-induced priapism involves the activation of the polyamine synthetic pathway.8 Increased adenosine also contributes to penile fibrosis.9 Brain ischaemia-induced priapism is not well understood. Till now there are only a few case reports on brain injury induced priapism in clinic. Takaku et al10 reported a 35-year-old man had three attacks of subarachnoid haemorrhage in three consecutive years and suffered priapism with intolerable pain after the last attack of subarachnoid haemorrhage. Monga et al11 reported a 53-year-old man with an old temporal lobe infarction involving the anteromedial basal portion increase in libido and coital frequency with a tendency towards priapism. However, there is no animal model to mimic clinical priapism at present. In our group, we have consistently established middle cerebral artery occlusion (MCAO) models in mice as a stroke model.12–16 In this study, we report our observation that permanent MCAO (pMCAO) can induce priapism in adult mice. We explored the causative relationship between focal cerebral ischaemia and priapism occurrence. Priapism may serve as an indication for severe stroke damage in animal models.

) models in mice as a stroke model.12–16 In this study, we report our observation that permanent MCAO (pMCAO) can induce priapism in adult mice. We explored the causative relationship between focal cerebral ischaemia and priapism occurrence. Priapism may serve as an indication for severe stroke damage in animal models. Materials and methods Animal stroke model Animal procedures for the use of laboratory animals were approved by the Institutional Animal Care and Use Committee of Shanghai Jiao Tong University. Twenty-four adult male CD-1 mice (Shanghai SLAC Laboratory Animal CO) weighing 25–32 g were used. Focal cerebral ischaemia was conducted by pMCAO using the suture model as described previously.17 Briefly, mice were anaesthetised with 4% isoflurane and maintained with 1.5% isoflurane in an oxygen/air mixture using a gas anaesthesia mask in a stereotaxic frame (RWD Life Science, Shenzhen, China) and the body temperature was maintained at 37°C by using a heating pad (RWD Life Science, Shenzhen, China). A midline incision was made on the neck under an operating microscope (Leica, Germany). External carotid artery (ECA), common carotid artery (CCA) and internal carotid artery (ICA) were isolated. After temporarily clamping the CCA, a 6–0 silicon-coated round-tip monofilament nylon suture was introduced into the ICA through the ECA until slight resistance was felt, indicating a complete middle cerebral artery (MCA) occlusion. Then cerebral blood flow was measured by laser Doppler flowmetry (Moor Instruments, UK) to confirm a successful MCA occlusion that has regional blood flow of <15% of baseline blood flow.

troduced into the ICA through the ECA until slight resistance was felt, indicating a complete middle cerebral artery (MCA) occlusion. Then cerebral blood flow was measured by laser Doppler flowmetry (Moor Instruments, UK) to confirm a successful MCA occlusion that has regional blood flow of <15% of baseline blood flow. TTC staining and H&E staining After sacrificing the animals, the brains were sectioned with brain matrix, and stained with a 2% solution of 2,3,5-triphenyltetrazolium (TTC) dye (Sigma, T8877–100G) at room temperature for 20 min as described previously.18 Then stained sections were immersed in 4% phosphate-buffered paraformaldehyde and imaged with a digital camera. The infarct area was measured as previously described.19 Then the brains were embedded in paraffin. Coronal sections were processed and stained with H&E (Beyotime Institute of Biotechnology). Using NIH Image 1.63 software, the ischaemic lesion area was calculated as the difference between the area of the non-ischaemic hemisphere and the normal area in the ischaemic hemisphere. Infarct volume was calculated by multiplying the infarct area by the thickness of the section. Priapism recording The mice were observed every day after the pMCAO procedure. The time of priapism occurrence and the recovery of flaccidity (disappearance of priapism) were both recorded. Photographs were taken using a Canon digital camera (Canon EOS 450D).

TTC staining and H&E staining After sacrificing the animals, the brains were sectioned with brain matrix, and stained with a 2% solution of 2,3,5-triphenyltetrazolium (TTC) dye (Sigma, T8877–100G) at room temperature for 20 min as described previously.18 Then stained sections were immersed in 4% phosphate-buffered paraformaldehyde and imaged with a digital camera. The infarct area was measured as previously described.19 Then the brains were embedded in paraffin. Coronal sections were processed and stained with H&E (Beyotime Institute of Biotechnology). Using NIH Image 1.63 software, the ischaemic lesion area was calculated as the difference between the area of the non-ischaemic hemisphere and the normal area in the ischaemic hemisphere. Infarct volume was calculated by multiplying the infarct area by the thickness of the section. Priapism recording The mice were observed every day after the pMCAO procedure. The time of priapism occurrence and the recovery of flaccidity (disappearance of priapism) were both recorded. Photographs were taken using a Canon digital camera (Canon EOS 450D). Statistical analysis The values were given as mean±SE. The significance of difference was calculated by one-way analysis of variance, followed by Student-Newman-Keuls post hoc comparisons. p Values <0.05 were considered significant.

Priapism recording The mice were observed every day after the pMCAO procedure. The time of priapism occurrence and the recovery of flaccidity (disappearance of priapism) were both recorded. Photographs were taken using a Canon digital camera (Canon EOS 450D). Statistical analysis The values were given as mean±SE. The significance of difference was calculated by one-way analysis of variance, followed by Student-Newman-Keuls post hoc comparisons. p Values <0.05 were considered significant. Results Priapism occurred in mice after pMCAO We conducted pMCAO procedure in 24 mice and found priapism occurrence in 20 of these animals. After the surgery, the duration of priapism was recorded in nine mice up to 14 days (eight mice were sacrificed for TTC staining and seven mice died after the surgery). Six mice displayed sustained penile erection from the second day after pMCAO (figure 1). The erection began to be relieved from the 7th day to the 14th day postoperation (table 1). Table 1 A summary of priapism condition in nine mice after permanent middle cerebral artery occlusion (pMCAO) Number Erection start (days) Erection disappear (days) 1 2 14 2 2 14 3 2 7 4 / / 5 / / 6 1 10 7 2 9 8 / / 9 1 11 The first line means nine mice were recorded after pMCAO (number 1–9), the middle and the last line displayed when priapism had occurred and disappeared after surgery. ‘/’ Indicates no priapism was observed.

Number Erection start (days) Erection disappear (days) 1 2 14 2 2 14 3 2 7 4 / / 5 / / 6 1 10 7 2 9 8 / / 9 1 11 The first line means nine mice were recorded after pMCAO (number 1–9), the middle and the last line displayed when priapism had occurred and disappeared after surgery. ‘/’ Indicates no priapism was observed. Figure 1 Adult mice were found to display priapism after permanent middle cerebral artery occlusion (pMCAO). Photographs showed that adult mouse did not get erection before pMCAO (A), or 1 day after pMCAO (B). Persistent erection started on the 2nd day (C), maintained through the 3rd (D) to the 13th day (N) and disappeared on the 14th day postoperation (O).

ay priapism after permanent middle cerebral artery occlusion (pMCAO). Photographs showed that adult mouse did not get erection before pMCAO (A), or 1 day after pMCAO (B). Persistent erection started on the 2nd day (C), maintained through the 3rd (D) to the 13th day (N) and disappeared on the 14th day postoperation (O). The difference of brain infarct volume between priapism and non-priapism ischaemic mice To compare brain infarct volume in the mice with and without priapism, the brain tissues were processed by TTC staining and H&E staining, and the infarct area was recorded, in parallel to animals with sham operation (no ischaemia and priapism). Mice after pMCAO had brain ischaemia as demonstrated by TTC staining (left and right columns of figure 2). However, the infract volume are significantly different in mice with priapism from those without priapism, consistent with H&E results (figure 3). We found that in the stroke mice with priapism, the infarct area was significantly larger than those displaying no priapism (figure 4). Furthermore, the specific infarct area was also different between them. In the ischaemic mice with priapism, the hippocampus and hypothalamus were damaged by ischaemia; however, in the non-priapism mice with ischaemia, the hippocampus area and the hypothalamus area were normal (figures 2 and 3).

no priapism (figure 4). Furthermore, the specific infarct area was also different between them. In the ischaemic mice with priapism, the hippocampus and hypothalamus were damaged by ischaemia; however, in the non-priapism mice with ischaemia, the hippocampus area and the hypothalamus area were normal (figures 2 and 3). Figure 2 Priapism closely correlated with infarct area. Photographs showed that the sham group did not get priapism and infarct in the brain (middle column, B–K); ischaemic mice that also displayed priapism were found to have extensive infarct affecting both hypothalamus (arrowhead) and hippocampus (arrow) areas (right column, C–L); in contrast, ischaemic mice that did not display priapism were found to have non-infarct hypothalamus and hippocampus areas (left column, A–J). All mice were sacrificed 3 days after permanent middle cerebral artery occlusion, n=4 for each group. Figure 3 Priapism occurrence was related with different infarct area. H&E staining showed that the sham group did not show priapism, and the nuclei of cells in these brains were normal (B and E); in the group of ischaemic mice with priapism (C and F), the nuclei of the cells in the hippocampus (arrow head) and hypothalamus (white arrow) displayed shrinkage; however, in the group of ischaemic mice without priapism (A and D), the hippocampus (arrowhead) and hypothalamus (white arrow) showed normal nuclei size; photographs A, C, D, F were taken from a, c, d and f, respectively. Bar=20 µm, n=4 for each group.

us (arrow head) and hypothalamus (white arrow) displayed shrinkage; however, in the group of ischaemic mice without priapism (A and D), the hippocampus (arrowhead) and hypothalamus (white arrow) showed normal nuclei size; photographs A, C, D, F were taken from a, c, d and f, respectively. Bar=20 µm, n=4 for each group. Figure 4 The infarct volume was significantly larger in stroke mice with priapism than those without priapism. Bar graph showed significant difference of infarct volume between the stroke priapism mice and stroke non-priapism mice. **p<0.01, n=4 for each group. Discussion In this study, we demonstrated that: (1) pMCAO could induce a high occurrence of priapism in mice (80%); (2) ischaemia-induced erection often occurs 2 days after pMCAO (67%) and disappears between the 7th and 14th day after pMCAO; and (3) hypothalamus and hippocampus damage may contribute to the occurrence of priapism. Our data suggested that cerebral ischaemia is closely related with priapism occurrence.

m in mice (80%); (2) ischaemia-induced erection often occurs 2 days after pMCAO (67%) and disappears between the 7th and 14th day after pMCAO; and (3) hypothalamus and hippocampus damage may contribute to the occurrence of priapism. Our data suggested that cerebral ischaemia is closely related with priapism occurrence. Priapism is defined as prolonged penile erection without sexual stimulation, and it often induces irreversible damage to erectile tissues and results in erection dysfunction. To the best of our knowledge, we first reported that pMCAO can induce priapism in rodents in this study. Our data showed that most ischaemic mice displayed priapism 2 days after pMCAO, and the detumescence time varied from the 7th day to the 14th day. As our TTC staining result revealed, there were some differences between the two groups (with and without priapism in stroke mice). The infarct volume in the mice with priapism was much larger, and the infarct area included the hypothalamus and hippocampus. However, these areas in the non-priapism mice were normal, suggesting that priapism is closely related to the lesion site of stroke, especially related to hippocampus and hypothalamus. We propose these differences are caused by posterior communicating artery (PComA) variation, which supplies blood to the hippocampus and hypothalamus regions.20 Our results might also indicate that hypothalamus and hippocampus injury are responsible for priapism. Indeed, these results are consistent with some previous studies. Some researchers have demonstrated that stimulation of brain can induce penile erection, and hippocampus and hypothalamus are important erection mediators.21 Melis et al22 reported that microinjection of oxytocin into the paraventricular nucleus of the hypothalamus or into the CA1 field of the hippocampus induced penile erection in male rats. Arrow suggested that penile erection was associated with brain region activation, including the claustrum, left caudate and putamen, right middle occipital/middle temporal gyri, bilateral cingulated gyrus, right sensorimotor, premotor regions and right hypothalamus.23 Other studies showed that anaesthesia could also induce priapism. For instance, Senthilkumaran et al24 reported that propofol induced priapism in a 25-year-old man, Ravindran et al25 reported two cases of priapism induced by 0.5% mg/kg ketamine and 1.5 mg of physostigmine. However, in these clinical cases, the reasons why priapism responds to anaesthesia are not explored.

e priapism. For instance, Senthilkumaran et al24 reported that propofol induced priapism in a 25-year-old man, Ravindran et al25 reported two cases of priapism induced by 0.5% mg/kg ketamine and 1.5 mg of physostigmine. However, in these clinical cases, the reasons why priapism responds to anaesthesia are not explored. To further investigate the association between the brain region and priapism, functional MRI may be required in follow-up studies. Besides the damage to the specific brain region, we suggest that priapism can be related to some chemicals induced by stroke, such as nitric oxide (NO), adenosine and so on. These chemicals are closely related with stroke, and they also play important roles in erection. NO is reported to be an important factor in stroke26 and penile erection.27 It is closely associated with stroke and shows beneficial actions,28 including enhancing angiogenesis via the synthesis of vascular endothelial growth factor and cyclic guanosine monophosphate after stroke.29 NO is also an important mediator in penile erection. Azadzoi et al30 stated that neuronal nitric oxide synthase (nNOS), endothelial nitric oxide synthase (eNOS) and inducible nitric oxide synthase (iNOS) levels in cavernosum were upregulated after atherosclerosis-induced ischaemia, and NO-mediated smooth muscle relaxation in a rabbit. However, we peritoneally injected the NOS inhibitor LNMMA (Beyotime, S0011) into the stroke mice, priapism did not disappear (data not shown). We cannot fully exclude the possible involvement of NO in causing priapism by this simple inhibitor study though. Further experiments are needed to clarify its role in the phenomenon reported here. Adenosine is an inhibitory modulator of brain activity with neuroprotective properties,31 and recent studies have reported that adenosine contributed to animal priapism.9 Intracavernous injection of adenosine resulted in tumescence and penile erection.32 Adenosine accumulated in the penis coupled with its receptor A2BR, contributed to priapism,33 and ophylline, an adenosine receptor antagonist, inhibited adenosine-induced penile tumescence.34 However, whether adenosine is a potential agent which mediates both stroke and priapism needs to be further discussed.

ection.32 Adenosine accumulated in the penis coupled with its receptor A2BR, contributed to priapism,33 and ophylline, an adenosine receptor antagonist, inhibited adenosine-induced penile tumescence.34 However, whether adenosine is a potential agent which mediates both stroke and priapism needs to be further discussed. Recent studies showed that in corpus cavernosum ischaemia-reperfusion mouse, many oxidative injury parameters like superoxide dismutase (SOD), catalase, malondialdehyde (MDA) and NO in the tissues were all upregulated, and melatonin, which also played neuroprotective actions in stroke,35 could attenuate the activity of SOD and the levels of MDA and NO.36 These conclusions strongly suggested that melatonin may be involved in the stroke-induced priapism pathway. If these chemicals are indeed associated with stroke and priapism, drugs used to cure priapism may also interfere in signalling pathways in stroke and some of them could possibly show beneficial effects to the ischaemic brain. This line of study might also be worth exploring.

ved in the stroke-induced priapism pathway. If these chemicals are indeed associated with stroke and priapism, drugs used to cure priapism may also interfere in signalling pathways in stroke and some of them could possibly show beneficial effects to the ischaemic brain. This line of study might also be worth exploring. In conclusion, this investigation has demonstrated the correlation between stroke and priapism. Most mice got sustained erection after pMCAO though their hypothalamus and hippocampus area were damaged. It is conceivable that priapism may serve as a physiological marker for adult stroke mouse. Since the patency of PComA plays an important role in the infarct results of the murine pMCAO models,37 an effective evaluation method is urgently needed to reduce the fluctuation of the infarct area among different individual animals.38 It is likely that we can use priapism as a physiological marker to identify stable pMCAO models more conveniently, which may facilitate the study of stroke. Conclusion Our study demonstrated that the occurrence of priapism is related to the infarct region: priapism is found only in mice with ischaemic injury extending to the hypothalamus and the hippocampus regions. Our results suggested priapism may be used as a deep brain injury marker for evaluating brain injury in mice after pMCAO. The authors thank Caibin Sheng for his help taking the photographs, Minjie Shen for editorial assistance and the staff of Neuroscience and Neuroengineering Center for their collaborative support.

Conclusion Our study demonstrated that the occurrence of priapism is related to the infarct region: priapism is found only in mice with ischaemic injury extending to the hypothalamus and the hippocampus regions. Our results suggested priapism may be used as a deep brain injury marker for evaluating brain injury in mice after pMCAO. The authors thank Caibin Sheng for his help taking the photographs, Minjie Shen for editorial assistance and the staff of Neuroscience and Neuroengineering Center for their collaborative support. Contributors: YT, FY and BC participated in the experimental design and data analysis. They prepared animal ischaemia model, analysed the data and drafted the manuscript. WX and YW participated in the experimental design and manuscript discussion. G-YY took care of all aspects, including participating in the study design, supervising tissue assays, analysing the results, and organising and finalising the manuscript. Funding: The study is supported by the National Natural Science Foundation of China, 81471178(GYY), U1232205 (GYY) and 81371305 (YW). Competing interests: None declared. Provenance and peer review: Commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

Introduction Pneumonia is a common complication in patients with stroke.1 It occurs in about 9% of patients with stroke, and 28% of patients with stroke in the intensive care unit.2 A variety of factors, such as older age, dysphagia and higher stroke severity score, which are assessed by the National Institutes of Health Stroke Scale (NIHSS), are associated with stroke-associated pneumonia (SAP).3 SAP potentially increases the risk of a poor and fatal outcome,4 and increases the economic burden on patients with stroke.5 Although a preventive antibiotic cannot improve the prognosis of patients with stroke, it is proved to be effective to decrease the incidence of SAP.6 Oral care is also proved to be effective to reduce the incidence of SAP.3 Recognising patients with stroke at high risk of pneumonia and providing preventive intervention can be helpful to reduce the incidence of SAP. Several risk scoring models were reported to improve the recognition of the high risk of pneumonia in patients with stroke, including Kwon's score,7 the PANTHER-IS score,8 Chumbler's score,9 A2DS210 and AISAPS.11 There are many advantages to the wide use of these risk models, such as assisting clinicians to identify patients with stroke with high risk of pneumonia and facilitating the selection of patients in clinic trials. However, these models have not been applied extensively to patients in different cohorts from various races and areas. They need to be externally validated before they can be used in clinical practice.

identify patients with stroke with high risk of pneumonia and facilitating the selection of patients in clinic trials. However, these models have not been applied extensively to patients in different cohorts from various races and areas. They need to be externally validated before they can be used in clinical practice. We conducted this study to determine whether these risk models could effectively predict SAP during hospitalisation in Chinese patients with IS. Since the risk scores of ischaemic stroke-associated pneumonia (like A2DS2 and AISAPS) are different from that of haemorrhagic stroke (like intracerebral haemorrhage-associated pneumonia score (ICH-APSs)),12 we only focused on patients with ischaemic stroke to control potential confounders.

in Chinese patients with IS. Since the risk scores of ischaemic stroke-associated pneumonia (like A2DS2 and AISAPS) are different from that of haemorrhagic stroke (like intracerebral haemorrhage-associated pneumonia score (ICH-APSs)),12 we only focused on patients with ischaemic stroke to control potential confounders. Methods Study population We identified all consecutive patients with IS in West China hospital between January 2011 and September 2013. Inclusion criteria: aged≥18; patients with IS who were admitted to hospital within 30 days of onset; ischaemic stroke was diagnosed based on the WHO criteria; IS was confirmed by brain CT or MRI. Exclusion criteria: patients with transient ischaemic attack (TIA); pneumonia happened before admission. Severity of stroke was assessed by NIHSS.13 Consciousness was evaluated by Glasgow Coma Scale.14 The subtype of IS was classified based on the Oxfordshire Community Stroke Project criteria: partial anterior circulation infarct, total anterior circulation infarct, lacunar infarction and posterior circulation infarct.15 Dysphagia was evaluated by a neurological clinician with the water drinking test. Prestroke dependence was defined as modified Rankin Scale ≥3.

Oxfordshire Community Stroke Project criteria: partial anterior circulation infarct, total anterior circulation infarct, lacunar infarction and posterior circulation infarct.15 Dysphagia was evaluated by a neurological clinician with the water drinking test. Prestroke dependence was defined as modified Rankin Scale ≥3. Outcome measures The primary outcome was the presence of SAP in hospitalised patients with IS. SAP was diagnosed by the treating physician according to the modified criteria of US Centers for Disease Control and Prevention criteria for hospital-acquired pneumonia. Pneumonia was diagnosed when at least one of the first and one of the latter criteria were fulfilled: (A) abnormal respiratory examination, pulmonary infiltrates in chest X-rays; (B) productive cough with purulent sputum, microbiological cultures from the lower respiratory tract or blood cultures, leucocytosis and elevation of C reactive protein.16

t least one of the first and one of the latter criteria were fulfilled: (A) abnormal respiratory examination, pulmonary infiltrates in chest X-rays; (B) productive cough with purulent sputum, microbiological cultures from the lower respiratory tract or blood cultures, leucocytosis and elevation of C reactive protein.16 Risk prediction scores of SAP We indentified five risk scoring models of SAP in total. Kwon's score was derived from 286 patients with stroke in a single hospital, as was the PANTHER-IS score (derived from 223 patients with IS). Chumbler et al developed their risk model from 925 patients in multicentre hospitals. A2DS2 was derived from 15 335 patients in multicentre communities and hospitals, as was AISAPS (derived from 8820). We excluded Kwon's score because it involved the variable (mechanical ventilation) which we thought would be a medical intervention rather than the original character of the patient. We excluded the PANTHER-IS score because it focused on patients with IS in the middle cerebral artery territory in the neurological intensive care unit, which would be not suitable to common patients with stroke. Statistical analysis Given the non-normal distribution of data, median values with 25th and 75th centiles were calculated for continuous variables and comparisons between groups were performed using the Nonparametric test. Categorical variables were described as counts and percentages, while comparisons between two groups were made with the χ2 test.

non-normal distribution of data, median values with 25th and 75th centiles were calculated for continuous variables and comparisons between groups were performed using the Nonparametric test. Categorical variables were described as counts and percentages, while comparisons between two groups were made with the χ2 test. The model fit was measured with the Hosmer-Lemeshow goodness-of-fit statistic. The discriminatory ability of each risk model was evaluated using the area under the receiver operating characteristic curve (c statistic). The c statistic for each pair of scores was compared with the DeLong method. We considered a p value <0.05 (2-tailed) to be statistically significant. All analyses were performed using MedCalc. Results Between January 2011 and September 2013, 1814 patients with IS were consecutively referred to the hospital. Of them, 140 patients were with TIA, 2 patients younger than 18 years, 104 patients with stroke onset more than 1 month or with pneumonia prior to admission. A total of 1569 patients were finally included in the present study. The median age was 65 years (IQR, 25–75), and 60% were male. Among all participants, 240 (15.3%) patients acquired pneumonia during hospitalisation (table 1). Table 1 Baseline characteristics of patients with and without stroke-associated pneumonia

Results Between January 2011 and September 2013, 1814 patients with IS were consecutively referred to the hospital. Of them, 140 patients were with TIA, 2 patients younger than 18 years, 104 patients with stroke onset more than 1 month or with pneumonia prior to admission. A total of 1569 patients were finally included in the present study. The median age was 65 years (IQR, 25–75), and 60% were male. Among all participants, 240 (15.3%) patients acquired pneumonia during hospitalisation (table 1). Table 1 Baseline characteristics of patients with and without stroke-associated pneumonia Pneumonia group Non-pneumonia group p Value Age, median (IQR), year 70.0 (61.0, 78.0) 64.0 (53.0, 73.0) <0.001* Sex 0.205 Male, n, % 134 (55.8) 800 (60.2) Female, n, % 106 (44.2) 529 (39.8) Hypertension, n, % 141 (58.8) 764 (57.5) 0.715 Diabetes mellitus, n, % 71 (29.6) 371 (27.9) 0.597 Hyperlipaemia, n, % 16 (38.1) 142 (10.7) 0.057 History of stroke or TIA, n, % 42 (17.5) 169 (12.7) 0.046* Valvular heart disease, n, % 2 (0.8) 7 (0.5) 0.563 Coronary heart disease, n,% 24 (10.0) 58 (4.4) <0.001* Atrial fibrillation, n, % 72 (30.0) 109 (8.2) <0.001* Congestive heart failure, n, % 3 (1.3) 1 (<0.1) <0.001* COPD, n, % 15 (6.3) 16 (1.2) <0.001* History of pneumonia, n,% 17 (7.1) 50 (3.8) 0.019* Current smoking, n, % 33 (13.8) 148 (11.1) 0.243 Excess alcohol consumption, n, % 25 (10.4) 156 (11.7) 0.555 Found down at onset, n, % 42 (17.5) 80 (6.0) <0.001* Dyaphasia, n, % 130 (9.8) 272 (20.5) <0.001* NIHSS, median (IQR) score 9.5 (5, 14.0) 3.0 (1.0, 7.0) <0.001* GCS, median (IQR) score 14.0 (11.0, 15.0) 15.0 (14.0,15.0) <0.001* OCSP <0.001* Lacunar infarction, n, % 7 (2.9) 66 (4.9) Total anterior circulation infarct, n, % 39 (16.3) 61 (4.6) Partial anterior circulation infarct, n, % 165 (68.8) 939 (70.7) Posterior circulation infarct, n, % 29 (12.1) 263 (19.8) Glu, median (IQR) mmol/L 6.73 (5.20, 8.60) 5.64 (4.87, 7.13) <0.001* WCC, median (IQR) 1012/L 8.34 (6.55, 10.71) 6.69 (5.44, 8.22) <0.001* *p<0.05.

erior circulation infarct, n, % 39 (16.3) 61 (4.6) Partial anterior circulation infarct, n, % 165 (68.8) 939 (70.7) Posterior circulation infarct, n, % 29 (12.1) 263 (19.8) Glu, median (IQR) mmol/L 6.73 (5.20, 8.60) 5.64 (4.87, 7.13) <0.001* WCC, median (IQR) 1012/L 8.34 (6.55, 10.71) 6.69 (5.44, 8.22) <0.001* *p<0.05. COPD, chronic obstructive pulmonary disease; GCS, Glasgow Coma Scale; NIHSS, National Institutes of Health Stroke Scale; OCSP, Oxfordshire Community Stroke Project; TIA, transient ischaemic attack; WCC, white cell count. Table 2 demonstrated the items of each risk scoring model included in the study. The abilities of these three risk models of predicting pneumonia are summarised in table 3. The discriminatory abilities of all models were described as the area under the receiver operating characteristic curve (AUROC), which is shown in figure 1. The AUROC of all models was 0.659 (p <0.001, Chumbler's score), 0.728 (p <0.001, A2DS2) and 0.758 (p<0.001, AISAPS), respectively. The differences in AUROC (ΔAUROC) between any two risk models were significant (p<0.05). The calibrations of all models assessed by the Hosmer-Lemeshow test were not significant (p>0.05). Although the incidence of SAP was underpredicted by all of these risk models, the predicted incidence of AISAPS was closer to the observed incidence (table 4). Table 2 The content of each risk model

Table 2 demonstrated the items of each risk scoring model included in the study. The abilities of these three risk models of predicting pneumonia are summarised in table 3. The discriminatory abilities of all models were described as the area under the receiver operating characteristic curve (AUROC), which is shown in figure 1. The AUROC of all models was 0.659 (p <0.001, Chumbler's score), 0.728 (p <0.001, A2DS2) and 0.758 (p<0.001, AISAPS), respectively. The differences in AUROC (ΔAUROC) between any two risk models were significant (p<0.05). The calibrations of all models assessed by the Hosmer-Lemeshow test were not significant (p>0.05). Although the incidence of SAP was underpredicted by all of these risk models, the predicted incidence of AISAPS was closer to the observed incidence (table 4). Table 2 The content of each risk model Number Chumbler's score (5 items) A2DS2 (5 items) AISASP (11 items) 1 age age age 2 dysphagia dysphagia dysphagia 3 NIHSS NIHSS NIHSS 4 found down at onset atrial fibrillation atrial fibrillation 5 history of pneumonia male GCS 6 – – congestive heart failure 7 – – COPD 8 – – current smoker 9 – – prestroke dependence 10 – – OCSP subtype 11 – – admission glucose COPD, chronic obstructive pulmonary disease; GCS, Glasgow Coma Scale; OCSP, Oxfordshire Community Stroke Project; NIHSS, National Institutes of Health Stroke Scale. Table 3 Comparison of various risk models for stroke-associated pneumonia

Number Chumbler's score (5 items) A2DS2 (5 items) AISASP (11 items) 1 age age age 2 dysphagia dysphagia dysphagia 3 NIHSS NIHSS NIHSS 4 found down at onset atrial fibrillation atrial fibrillation 5 history of pneumonia male GCS 6 – – congestive heart failure 7 – – COPD 8 – – current smoker 9 – – prestroke dependence 10 – – OCSP subtype 11 – – admission glucose COPD, chronic obstructive pulmonary disease; GCS, Glasgow Coma Scale; OCSP, Oxfordshire Community Stroke Project; NIHSS, National Institutes of Health Stroke Scale. Table 3 Comparison of various risk models for stroke-associated pneumonia Risk models OR p Value Hosmer-Lemeshow test AUROC p Value Chumbler's score 1.8138 <0.0001 p=0.6243 0.659 <0.001 A2DS2 1.4248 <0.0001 p=0.0953 0.728 <0.001 AISAPS 1.2365 <0.0001 p=0.8927 0.758 <0.001 AUROC,area under the receiver operating characteristic curve. Table 4 Observed and predicted rate of stroke-associated pneumonia Incidence of SAP (%) Observed 15.3 Predicted by Chumbler's score 10.7 A2DS2 10.8 AISAPS 11.7 SAP,stroke-associated pneumonia. Figure 1 The AUROC of each risk model. (A). The AUROC of Chumbler's score; (B). The AUROC of A2DS2; (C). The AUROC of AISAPS; (D). The comparison of AUROC of three risk models. AUROC, area under the receiver operating characteristic curve.

Incidence of SAP (%) Observed 15.3 Predicted by Chumbler's score 10.7 A2DS2 10.8 AISAPS 11.7 SAP,stroke-associated pneumonia. Figure 1 The AUROC of each risk model. (A). The AUROC of Chumbler's score; (B). The AUROC of A2DS2; (C). The AUROC of AISAPS; (D). The comparison of AUROC of three risk models. AUROC, area under the receiver operating characteristic curve. The optimal cut-off value of risk models for SAP is displayed in table 5. The optimal cut-off value was 1 for Chumbler's score, whose sensitivity was 71.49%, specificity was 52.93%, positive predictive value (PPV) was 21.66% and negative predictive value (NPV) was 91.07%. The optimal cut-off value was 3 for A2DS2, whose sensitivity was 69.83%, specificity was 65.56%, PPV was 26.96% and NPV was 92.27%. The optimal cut-off value was 6 for AISAPS, whose sensitivity was 76.45%, specificity was 62.56%, PPV was 27.19% and NPV was 93.59%. Table 5 Optimal cut-off value of risk models for stroke-associated pneumonia Y-index Cut-off point Sensitivity (%) Specificity (%) PPV NPV (%) Chumbler's score 0.2442 1 71.49 52.93 21.66 91.07 A2DS2 0.3540 3 69.83 65.56 26.96 92.27 AISAPS 0.3900 6 76.45 62.56 27.19 93.59 NPV, negative predictive value; PPV, positive predictive value.

Table 5 Optimal cut-off value of risk models for stroke-associated pneumonia Y-index Cut-off point Sensitivity (%) Specificity (%) PPV NPV (%) Chumbler's score 0.2442 1 71.49 52.93 21.66 91.07 A2DS2 0.3540 3 69.83 65.56 26.96 92.27 AISAPS 0.3900 6 76.45 62.56 27.19 93.59 NPV, negative predictive value; PPV, positive predictive value. Discussion We found that the Chumbler score, A2DS2 and AISAPS models could reasonably predict pneumonia after IS in the Chinese population. Among them, the A2DS2 and AISAPS models (whose AUROC was higher than 0.7) had acceptable discriminatory abilities. The discriminatory ability of every risk model was lower in this external validation cohort than in their derivation cohort, which was consistent with the AISAPS study. The AISAPS study developed the AISAPS model to predict pneumonia after IS, as well as compared it with other risk models. They found Chumbler's score, A2DS2 and AISAPS model had acceptable discriminatory abilities in derivation and validation cohort, except Kwon's score only had acceptable discriminatory ability in derivation cohort. Chumbler's score had acceptable discriminatory ability in the validation cohort, which differed from this study. In addition, the AISAPS study did not report the difference in other risk models except AISAPS.

ion and validation cohort, except Kwon's score only had acceptable discriminatory ability in derivation cohort. Chumbler's score had acceptable discriminatory ability in the validation cohort, which differed from this study. In addition, the AISAPS study did not report the difference in other risk models except AISAPS. We found that AISAPS had the highest discriminatory ability, while Chumbler's score had the lowest discriminatory ability, which is consistent with the AISAPS study. The AISAPS model may be the most effective one for Chinese patients with IS. Further study is needed to explore the predictive performance of the AISAPS model in other races and regions. We also analysed the independent risk factors of SAP in our cohort to investigate the reasons why these risks models had different predictive performances in this study. We found that age (OR=1.035, 95% CI 1.021 to 1.049), atrial fibrillation (OR=2.733, 95% CI 1.837 to 4.067), chronic obstructive pulmonary disease history (OR=5.006, 95% CI 2.143 to 11.693), NIHSS score (OR=1.124, 95% CI 1.092 to 1.157), dysphagia (OR=2.908, 95% CI 1.092 to 1.157) and the count of first time white cell count during hospitalisation (OR=1.169, 95% CI 1.116 to 1.224) were independent risk factors for SAP. The independent risk factors in our cohort had the biggest overlap with the items of AISASP, and smallest overlap with Chumbler's score. This may explain the differences of predictive performances between these risk models.

during hospitalisation (OR=1.169, 95% CI 1.116 to 1.224) were independent risk factors for SAP. The independent risk factors in our cohort had the biggest overlap with the items of AISASP, and smallest overlap with Chumbler's score. This may explain the differences of predictive performances between these risk models. In diverse populations, the predictive performances of risk models could be different. So they need to be externally validated before they are used in clinic. As recommended in the guideline about clinical decision rules, the rules at level 1 must be prospectively validated in different populations and have one impact analysis demonstrating change in the clinician's behaviour with beneficial consequences, so that rules at level 1 can be used in a wide variety of settings.17 There are some limitations in this study. First, our study was a retrospective study with potential confounding factors. Second, we only included patients in a single hospital, which could cause selected bias. For now, whether the clinical application of these risk models can assist in decreasing the incidence of SAP is still uncertain. Prospective studies with a large sample containing different races and a regional population are needed. Contributors: SG and ZZ drafted the manuscript. ZL and NC retrieved the data. JG and MZ performed the statistical analysis. LH critically revised the manuscript. Competing interests: None declared. Provenance and peer review: Not commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

Introduction Stroke is one of the leading causes of disability and death in developing countries, especially in China.1 2 Intravenous (IV) tissue plasminogen activator (tPA; 0.9 mg/kg) is the only proven effective medical treatment for acute ischaemic stroke (AIS) during the past 20 years.3–5 However, there is still a controversy on what the optimal doses of IV tPA should be. Many neurologists in Asia consider that lower doses of IV tPA are better for Asian patients with stroke because of the racial difference in coagulation and fibrinolysis responses.6–8 Many observational and registry studies have been conducted in order to prove the efficacy and safety of lower doses of IV tPA. Consequently, Japan is the only nation that recommends 0.6 mg/kg of IV tPA in its stroke care guideline.9 The newly published low-dose versus standard-dose tPA in the ENhanced Control of Hypertension and Thrombolysis strokE stuDy (ENCHANTED) has helped resolve the low-dose controversy to some extent but also raised more concerns.10 In order to help clinicians decipher through this rather complex issue, we performed an up-to-date overview of lower doses of IV tPA for AIS.

tPA in the ENhanced Control of Hypertension and Thrombolysis strokE stuDy (ENCHANTED) has helped resolve the low-dose controversy to some extent but also raised more concerns.10 In order to help clinicians decipher through this rather complex issue, we performed an up-to-date overview of lower doses of IV tPA for AIS. Methods Using key words ‘low dose IV tPA’, ‘thrombolysis’, ‘Alteplace’ and ‘tPA’ plus ‘acute ischemic stroke’ and ‘cerebral infarction’, related literatures were searched in PubMed, EMBASE, Web of Science and MEDLINE between 1992 and May 2016 in order to identify any related clinical studies. All published trials have been included in the analysis while review type articles were excluded. Level of evidence and grades of recommendation were assigned to each of these trials based on the standard.11 Results from these trials were tabulated and compared. Results From 1992 to May 2016, 24 trials including 11 127 patients on the use of lower doses of IV tPA for AIS were identified. Doses ranged from 0.5, 0.6, 0.75 to 0.85 mg/kg. Most studies were observational, retrospective and of registry type (table 1). Only one study was a prospective randomised, open-label controlled trial. Table 1 The efficacy and safety of low-dose tPA treatment in acute stroke

Results From 1992 to May 2016, 24 trials including 11 127 patients on the use of lower doses of IV tPA for AIS were identified. Doses ranged from 0.5, 0.6, 0.75 to 0.85 mg/kg. Most studies were observational, retrospective and of registry type (table 1). Only one study was a prospective randomised, open-label controlled trial. Table 1 The efficacy and safety of low-dose tPA treatment in acute stroke Author Year Design Sample Size Racial Age mean±SD, median (IQR) Baseline NIHSS mean±SD, median (IQR) Time window (hour) Dosage (mg/kg) FO sICH (%) Mortality Level of evidence Anderson 2016 RCT 1654 1043/1651 Asian 68 (58–76) 8 (5–14) 4.5 0.6 46.8 5.9 53.2 A Morihara 2016 Retrospectively 121 Japanese 74.6±10.3 11 (6–18) 3 0.6 36.0 2.5 6.9 C 56 Japanese 75.7±11.7 12 (6.75–18) 3–4.5 0.6 23.4 3.6 8.3 C Kim 2015 Retrospective 450 Korean 69.0±12.7 13.9±7.0; 15 (8–19) 4.5 0.6 32.4 8.4 12.7 C Liao 2014 Registry 75 Chinese 62 (52–71) 10 (7–17) 4.5 0.5–0.7 41.9 0 5.4 B 131 68 (57–73) 10 (6–15) 0.7–0.85 48.0 8.7 8.66 Pan 2013 Observational 31 Chinese 63.8±9.3 8.7±4.6 3 <0.75 51.5 3 3.2 C 33 64.5±7.7 9.2±5.0 3 0.75–0.90 61.2 9.7 3 Chen 2012 Registry 105 Taiwan 67.9±12.8 13.3±6.2 3 0.7 (0.66–0.74) 41.1 4.8 7.6 B Zhou 2010 Observational 23 Chinese 69.8±8.6 12.6±6.8 4.5 0.6–0.7 34.8 4.3 17.4 C 31 72.9±8.7 12.7±5.0 4.5 0.8 38.7 3.2 16.1 Chao 2010 Retrospective 116 Taiwan 66.7±13.3 14.9±6.0 3 0.72±0.07 39.3 5.4 10 C Nguyen 2010 Prospective 48 Vietnamese 57 (18–78) 12 (5–23) 3 0.62 (0.6–0.86) 56.3 2.1 2.1 B Sharma 2010 Retrospective 48 Multiethnic in Asian 55±12 12 (10) 3 0.9, maximum 50mg 0.35 0.058 0.1 C Nakagawara 2010 Observational 7492 Japanese 72 (65–79) 15 (9–20) 3 0.6 0.33 0.044 13.2 B Mori 2010 Observational 58 Japanese 70.3±11.5 12 (5–22) 3 0.6 0.466 0 1.7 C Yamaguchi 2006 Prospective 103 Japanese 70.9±9.8 15 (5–30) 3 0.6 36.9 5.8 0.097 B Haley 1992 Pilot 8 1/8 black 74 (72–76) 14.5 (9–17) 3 0.6 NA 12.5 25 C 6 1/6 Asian 71 (67–75) 23.5 (23–24) 3 0.85 33.3 50 6 2/5 black 66.5 (66–67) 18 (12–22) 3 0.95 37.5 16.7 Level of evidence was assessed according to American Heart Association/American Stroke Association (AHA/ASA) criteria.

30) 3 0.6 36.9 5.8 0.097 B Haley 1992 Pilot 8 1/8 black 74 (72–76) 14.5 (9–17) 3 0.6 NA 12.5 25 C 6 1/6 Asian 71 (67–75) 23.5 (23–24) 3 0.85 33.3 50 6 2/5 black 66.5 (66–67) 18 (12–22) 3 0.95 37.5 16.7 Level of evidence was assessed according to American Heart Association/American Stroke Association (AHA/ASA) criteria. FO, favourable outcome at 3 months follow-up; NA, not available; NIHSS, National Institution of Health Stroke Scale; RCT, randomised controlled trial; sICH, symptomatic intracranial haemorrhage, refer to National Institute of Neurological Disorders And Stroke Recombinant Tissue Plasminogen Activator Stroke Study (NINDS) criteria; tPA, tissue plasminogen activator. Two trials combined lower doses of IV tPA with a glycoprotein IIb/IIIa inhibitor (table 2). Eight trials tested low-dose IV tPA combined with intra-arterial (IA) thrombolysis or thrombectomy. Table 2 Summary of standard dosage versus the combination of other treatment plus low-dose tPA

FO, favourable outcome at 3 months follow-up; NA, not available; NIHSS, National Institution of Health Stroke Scale; RCT, randomised controlled trial; sICH, symptomatic intracranial haemorrhage, refer to National Institute of Neurological Disorders And Stroke Recombinant Tissue Plasminogen Activator Stroke Study (NINDS) criteria; tPA, tissue plasminogen activator. Two trials combined lower doses of IV tPA with a glycoprotein IIb/IIIa inhibitor (table 2). Eight trials tested low-dose IV tPA combined with intra-arterial (IA) thrombolysis or thrombectomy. Table 2 Summary of standard dosage versus the combination of other treatment plus low-dose tPA Author Year Racial (% black) Sample Age (IQR) Baseline NIHSS (IQR) Time window (hour) Dosage (mg/kg) Other Control 0.9 mg/kg/L Favourable outcome (%)* sICH (%) Mortality Morihara 2016 Japanese 21 71.7 10 (6–16) <4.5 0.6 Opeolu39 Adeoye 2015 13 85 68 (33–86) 11 (6–31) 2 0.6 Eptifibatide 0.9 52 0 15.3 Pancioli 2013 13.9 101 71.6 (58.1–81.5) 19 <4.5  0.6 Eptifibatide 0.9 43.6 2 19.8 Pancioli 2008 NA 29 72.7 (67–77  14.0 (10–20) <3 0.3 Eptifibatide 0.9 31.0 3 28.0 40 68.0 (52–77) 13.5 (8–17) <3 0.45 Eptifibatide 0.9 30.0 0 18.0 IMS II 2007 11.1 81 64±11.5† 19 <3 0.6 IA thromolysis+EKOS system NA 33.0 9.9 16.0 *Favourable outcome indicates modified Rankin scale 0–1 at 3 months follow-up. †Only mean±SD were available. IA, intra-arterial; IMS, Interventional Management of Stroke; NA, not available; NIHSS, National Institution of Health Stroke Scale; sICH, symptomatic intracranial haemorrhage, refer to NINDS criteria; tPA, tissue plasminogen activator.

Author Year Racial (% black) Sample Age (IQR) Baseline NIHSS (IQR) Time window (hour) Dosage (mg/kg) Other Control 0.9 mg/kg/L Favourable outcome (%)* sICH (%) Mortality Morihara 2016 Japanese 21 71.7 10 (6–16) <4.5 0.6 Opeolu39 Adeoye 2015 13 85 68 (33–86) 11 (6–31) 2 0.6 Eptifibatide 0.9 52 0 15.3 Pancioli 2013 13.9 101 71.6 (58.1–81.5) 19 <4.5  0.6 Eptifibatide 0.9 43.6 2 19.8 Pancioli 2008 NA 29 72.7 (67–77  14.0 (10–20) <3 0.3 Eptifibatide 0.9 31.0 3 28.0 40 68.0 (52–77) 13.5 (8–17) <3 0.45 Eptifibatide 0.9 30.0 0 18.0 IMS II 2007 11.1 81 64±11.5† 19 <3 0.6 IA thromolysis+EKOS system NA 33.0 9.9 16.0 *Favourable outcome indicates modified Rankin scale 0–1 at 3 months follow-up. †Only mean±SD were available. IA, intra-arterial; IMS, Interventional Management of Stroke; NA, not available; NIHSS, National Institution of Health Stroke Scale; sICH, symptomatic intracranial haemorrhage, refer to NINDS criteria; tPA, tissue plasminogen activator. Early-dose escalation studies Two pilot dose escalation studies by Brott et al3 and Haley et al4 were the basis for the decision on the standard dose of IV tPA at 0.9 mg/kg.5 In the first pilot study, doses of 0.35 mg/kg (n=6), 0.60 mg/kg (n=12), 0.85 mg/kg (n=30), 0.95 mg/kg (n=25) and 1.08 mg/kg (n=1) were used within 90 min of symptom onset in patients with AIS between 18 and 80 years of age.3 The study demonstrated that patients treated with 0.85 mg/kg were more likely to achieve better neurological improvement at 24 hours compared with patients treated with 0.6 mg/kg.3 No sypmtomatic intracranial haemorrhage (sICH) was observed when doses of ≤0.85 mg/kg were used.3 The second study tested doses of 0.6 mg/kg (n=8), 0.85 mg/kg (n=6) and 0.95 mg/kg (n=6) in patients with AIS presenting within 90–180 min after symptom onset.4 Two higher dose tiers had sICH (17% in doses of ≥0.85 mg/kg).4 Efficacy at a lower dose of IV tPA in these pilot trials was not examined. The optimal doses of IV tPA were still inconclusive.3 4 The major concern about tPA remained to be the factors related to sICH.12

S presenting within 90–180 min after symptom onset.4 Two higher dose tiers had sICH (17% in doses of ≥0.85 mg/kg).4 Efficacy at a lower dose of IV tPA in these pilot trials was not examined. The optimal doses of IV tPA were still inconclusive.3 4 The major concern about tPA remained to be the factors related to sICH.12 Low-dose IV tPA studies in Asians Japan has conducted many low-dose IV tPA trials and all of them were single-arm observational studies.6 9 13 The rationale for the Japanese to decide on recommending 0.6 mg/kg instead of 0.9 mg/kg was based on a dose-range study of duteplase completed two decades ago14 which showed that the lower dose was optimal for patients with AIS up to 6 hours after onset. Since 2005, 0.6 mg/kg has been the only approved dosage given for patients with AIS within 3 hours of onset in Japan.6 This recommendation was based on the results of the Japanese thrombolysis study which suggested clinical benefits.6 Subsequent studies, the Japan Alteplase Clinical Trial 2 and Japan post-Marketing tPA Registration Study, also demonstrated both efficacy and safety with the 0.6 mg/kg dose.9 13 In the Japan Alteplase Clinical Trial 2, the dose of tPA was 0.6 mg/kg and the vascular outcome was evaluated by MR angiogram (MRA) at 6 and 24 hours after symptom onset.13 Recanalisation was noted on MRA in 51.7% of patients at 6 hours and in 69.0% of patients at 24 hours after symptom onset. In total, 46.6% of patients achieved favourable clinical outcome (modified Rankin scale (mRS) score of 0–1 at 3 months after onset) with no sICH.13 The Japan post-Marketing tPA Registration Study enrolled 7492 patients in 942 centres. The primary purpose was to investigate whether 0.6 mg/kg IV tPA could be safe and effective in routine clinical practice for Japanese patients.9 The study demonstrated 33% of favourable outcome in the efficacy analysis of 4944 patients. In the safety analysis of 7492 patients, the incidence of sICH was 3.5% within 36 hours and the overall mortality rate was 13.1%.9

0.6 mg/kg IV tPA could be safe and effective in routine clinical practice for Japanese patients.9 The study demonstrated 33% of favourable outcome in the efficacy analysis of 4944 patients. In the safety analysis of 7492 patients, the incidence of sICH was 3.5% within 36 hours and the overall mortality rate was 13.1%.9 In Singapore, a local institution used 0.9 mg/kg to calculate the total dose but capped the maximum dose at 50 mg in order to avoid sICH and reduce the cost of tPA. There was no sICH at 18 months of follow-up but the rate of functional independence was low.15 It changed its protocol and used the standard dose. This change has improved the efficacy from 35% to 59% (p=0.011), with a decreased rate of sICH from 14.5% to 1.2% (p=0.004).15 In contrast, in a small sample Vietnamese study, patients treated with low-dose IV tPA were more likely to achieve functional independence at 3 months (56.3% vs 34.2%; p=0.01) and only 2.1% of patients had sICH in the low-dose group, compared with those who received the standard dose.16 In the Korean national registry study, low-dose IV tPA was given to 450 patients (29.5%) and standard-dose tPA to 1076 patients (70.5%).17 Low-dose therapy was comparable to standard-dose therapy when measured by the 3-month mRS score (OR 0.95, 95% CI 0.68 to 1.32) and mortality (OR 0.54, 95% CI 0.35 to 0.83).18 However, the incidence of sICH in the low-dose group was 8.4%, compared with 6.4% in the standard-dose group.18

e tPA to 1076 patients (70.5%).17 Low-dose therapy was comparable to standard-dose therapy when measured by the 3-month mRS score (OR 0.95, 95% CI 0.68 to 1.32) and mortality (OR 0.54, 95% CI 0.35 to 0.83).18 However, the incidence of sICH in the low-dose group was 8.4%, compared with 6.4% in the standard-dose group.18 In the Chinese population, Taiwan Thrombolytic Therapy for Acute Ischemic Stroke demonstrated that the standard dose of IV tPA might not be optimal to treat older Chinese patients.19 Researchers used the Safe Implementation of Thrombolysis in Stroke: a Monitoring Study of Safety and Efficacy of Thrombolysis in Stroke (SITS-MOST) criteria and recruited 241 participants. Among them, 116 had low doses (0.72±0.07) and 125 received the standard dose (0.9 mg/kg), respectively. The lower dose group demonstrated a higher percentage (41.4%) of functional independence (mRS score 0–1) in all age groups.19 For patients older than 70, 53.6% in the lower dose group had functional independence versus 32.6% in the standard-dose group.19 The investigators found that patients receiving a standard dose had a twofold risk of sICH (8.0% vs 2.6%) and a higher rate of mortality within 3 months (12.8% vs 6.9%).17 Hence, there was a trend that low-dose tPA seemed to be safer in elderly patients. However, in another Taiwan-based study, Chen et al20 recruited 261 patients in a retrospective study involving two centres. Among them, 105 received 0.7 mg/kg IV tPA and 156 had 0.9 mg/kg. The study demonstrated no difference between the two dose groups on outcome at 3 months (38.4% and 41.1%, respectively; p=0.676).20 There was no difference in the occurrence of sICH in the standard-dose group compared with the low-dose group (2.6% vs 4.8%, p=0.34).20 In a larger national registry study in the Chinese population, the Thrombolysis Implementation and Monitor of Acute Ischemic Stroke in China enrolled over 900 patients. The authors concluded that standard-dose IV tPA increased the chance of achieving a favourable outcome without increasing the risk of sICH compared with low-dose IV tPA.21 Therefore, in Chinese patients with AIS treated with low-dose tPA, the rate of sICH was 4.6% (28/544) and the rate of mortality at 90 days was 8.1% (44/540). The rate of favourable outcome was 43.8% (232/530). Compared with other Asian ethnic groups, the Chinese and Korean populations had a higher incidence of sICH (4.6% and 8.4% respectively, vs 0–3.5%) without increased mortality.21

the rate of sICH was 4.6% (28/544) and the rate of mortality at 90 days was 8.1% (44/540). The rate of favourable outcome was 43.8% (232/530). Compared with other Asian ethnic groups, the Chinese and Korean populations had a higher incidence of sICH (4.6% and 8.4% respectively, vs 0–3.5%) without increased mortality.21 The combination of low dose or standard dose of IV tPA with other treatment modalities for AIS Several studies have explored the combination of low dose or standard dose of IV tPA plus IA tPA,22–29 including one with IA thrombectomy study30 and two with a glycoprotein IIb/IIIa agent.31 32 IV therapy was usually given prior to other treatments. The rationale of using 0.6 mg/kg of IV tPA prior to endovascular treatment was to avoid using a total dose higher than 0.9 mg/kg of tPA.33 In the Interventional Management of Stroke (IMS) III trial, patients randomised to endovascular treatment were administered IV tPA at 0.6 mg/kg (bolus and infusion) followed by an IA dose of 0–22 mg. The total tPA dose was kept to <0.9 mg/kg.29 The steady-state concentration of 2.5 μg/mL was achieved with an IV tPA dose of 0.75 mg/kg.27 Therefore, the effective steady state of the concentration of IV tPA may not always be the same with the standard dose or lower doses. Analysis of the relationship between total (IV+IA) dose of tPA and risk of sICH in the IMS II trials did not demonstrate the same safety threshold as seen in pilot trials with IV tPA.33 The IV dose in the combined arm was increased from 0.6 to 0.9 mg/kg and a combined maximum IV and IA total dose of 112 mg was defined as the maximum dose in the IMS III trial.29 IMS III was terminated due to futility issues and no outcome was generated.29 A systematic review of 11 trials was conducted to compare IV low-dose tPA versus standard doses before endovascular treatment.34 sICH was seen in 26 (8%) patients in the 0.6 mg/kg group compared with 10 (7%) in the 0.9 mg/kg group.33 Patients in the 0.9 mg/kg group had a higher rate of favourable outcome but a similar rate of sICH.

review of 11 trials was conducted to compare IV low-dose tPA versus standard doses before endovascular treatment.34 sICH was seen in 26 (8%) patients in the 0.6 mg/kg group compared with 10 (7%) in the 0.9 mg/kg group.33 Patients in the 0.9 mg/kg group had a higher rate of favourable outcome but a similar rate of sICH. Depending on the statistical methods used, the difference in angiographically identified recanalisation between 0.9 and 0.6 mg/kg IV tPA was rated as significant (p=0.03, events/trial syntax logistic regression) or borderline significant (p=0.07, random-effects model).33 However, a recent meta-analysis on endovascular thrombectomy after large vessel ischaemic stroke only included full-dose tPA.35 There is no direct comparison between low-dose and full-dose tPA in thrombectomy studies using a new generation of devices.

ression) or borderline significant (p=0.07, random-effects model).33 However, a recent meta-analysis on endovascular thrombectomy after large vessel ischaemic stroke only included full-dose tPA.35 There is no direct comparison between low-dose and full-dose tPA in thrombectomy studies using a new generation of devices. The ENCHANTED trial Low-dose versus standard-dose tPA in the ENCHANTED was the first multicentre and multinational randomised prospective open-label study which compared 0.9 to 0.6 mg/kg IV tPA for patients with AIS within 4.5 hours of onset.10 Patients who received the low-dose (0.6 mg/kg) IV tPA were given 15% of the total dose as IV bolus followed by 85% of the total dose as infusion over 1 hour. The control group with the standard dose of IV tPA (0.9 mg/kg) was given 10% of dose as IV bolus followed by 90% of dose as IV infusion over 1 hour.10 It was designed as a non-inferiority trial with the primary outcome to be the rate of combined disability and death (mRS 2–6) at 90 days, and secondary outcome to be the rate of sICHs. Non-inferiority was prespecified as an upper limit for non-inferiority of 1.14, which was derived from a published Cochrane meta-analysis of IV tPA.10 36 A total of 3310 patients from 111 centres in 13 countries were recruited and 3206 patients were entered into the analysis with 64% Asians. Among them, 20.6% had lacunar stroke and 14% were elderly patients older than 80. The primary outcome of the trial was not reached and the authors concluded that the trial did not show the non-inferiority of low-dose tPA to standard-dose tPA with respect to death and disability at 90 days.10 Despite the statistical failure, the trial showed that low-dose tPA was non-inferior to the standard dose in the ordinal analysis of mRS scores (unadjusted common OR 1.00; 95% CI 0.89 to 1.13; p=0.04 for non-inferiority) and less major sICH occurred in the low-dose group (1.0% vs 2.1%, p=0.01); fatal events occurred less within 7 days (0.5% vs1.5%, p=0.01). Mortality at 90 days did not differ significantly between the two groups (8.5% and 10.3%, respectively; p=0.07). However, patients receiving a low dose had a higher rate of disability measured by mRS scores between 2 and 5.

se group (1.0% vs 2.1%, p=0.01); fatal events occurred less within 7 days (0.5% vs1.5%, p=0.01). Mortality at 90 days did not differ significantly between the two groups (8.5% and 10.3%, respectively; p=0.07). However, patients receiving a low dose had a higher rate of disability measured by mRS scores between 2 and 5. Discussion There were several concerns of the ENCHANTED trials. Since it was an open-label trial, many biases could not be avoided. The publication did not clarify whether patients were enrolled consecutively, which could also bring bias. The follow-up, done by either a face-to-face in-person visit or telephone interview, would bring variable assessment and inaccurate results. There are also some concerns about the authors of the study since they had significant conflicts of interest. From the trial design point of view, low-dose IV tPA did not meet the upper limit of the prespecified non-inferiority threshold for the OR in comparison to standard-dose tPA for the primary outcome of death or disability at 90 days. The combined primary outcome of death and disability masked the higher disability rate in the low-dose group. Finally, this study did not prove non-inferiority of low-dose IV tPA and did not show superiority to the standard dose of IV tPA either.37

ard-dose tPA for the primary outcome of death or disability at 90 days. The combined primary outcome of death and disability masked the higher disability rate in the low-dose group. Finally, this study did not prove non-inferiority of low-dose IV tPA and did not show superiority to the standard dose of IV tPA either.37 Despite these concerns, ENCHANTED performed well. It was one of the trials which showed that statistical hypothesis testing would not match the clinical performances. Had the primary outcome was designed as mortality rate, then 0.6 mg/kg would be superior since this dose caused less death and less fatal ICH. Patients treated with low dose have worse outcome assessed by mRS of 1–5, which meant more moderate-to-severe disability. If Asians and non-Asians are analysed separately, non-Asians performed better when measured in the mRS range of 4–6. Furthermore, the trial showed that significantly fewer sICHs in patients received 0.6 mg/kg. Finally, the ENCHANTED trial, which was the largest trial of its kind until now, confirmed that 0.6 mg/kg IV tPA had a better safety profile but failed in non-inferior comparison in efficacy as designed.10 As we are reviewing the data, the THrombolysis for Acute Wake-up and unclear-onset Strokes trial has been ongoing. It aims to determine the efficacy and safety of IV alteplase at 0·6 mg/kg but uses MRI to select patients with AIS who presented with unclear time of symptom onset.38 Perhaps this trial can provide the needed evidence that patients with wake-up stroke could benefit from low-dose IV tPA.

t Strokes trial has been ongoing. It aims to determine the efficacy and safety of IV alteplase at 0·6 mg/kg but uses MRI to select patients with AIS who presented with unclear time of symptom onset.38 Perhaps this trial can provide the needed evidence that patients with wake-up stroke could benefit from low-dose IV tPA. On the basis of this review, some subgroups of patients with AIS might be benefited from the 0.6 mg/kg dose, for example, elderly patients, patients with relatively high blood pressure at baseline, patients with cardioembolism, patients with old ischaemic stroke, patients on multiple antiplatelet agents or anticoagulants.10 The benefit of these subgroups was supported by other studies. A Chinese population study showed that the elderly might benefit more in the low-dose group.19 In addition, the Japanese postmarket observational study with a large proportion of elderly patients with cardioembolism revealed a remarkably favourable outcome, which is nearly the same rate as that seen in another standard-dose observational study.10 Other subgroups of patient, potentially to be optimal with low dose, may include on antithrombotics with early signs of larger ischemic infarction on CT and with severe stroke (National Institution of Health Stroke Scale (NIHSS) >12). Further studies are still needed to confirm the benefit in safety in these population. In patients who need bridging therapy, IV low-dose tPA could be an option, although a majority of the newly published IV tPA plus IA thrombectomy studies used the standard dose of IV tPA.29 30 33 35 Therefore, for those with severe strokes and identified large vessel occlusions, a standard dose of IV tPA should be considered before IA thrombectomy.

therapy, IV low-dose tPA could be an option, although a majority of the newly published IV tPA plus IA thrombectomy studies used the standard dose of IV tPA.29 30 33 35 Therefore, for those with severe strokes and identified large vessel occlusions, a standard dose of IV tPA should be considered before IA thrombectomy. The ENCHANTED trial included patients within 4.5 hours of onset in both dose groups. In European Cooperative Acute Stroke Study (ECASS) III trials, those who were older than 80, with a history or stroke and diabetes and anticoagulation, had worse outcome. Perhaps in these patients now excluded based on ECASS III results, 0.6 mg/kg might be beneficial. Other studies also supported the use of a lower dose in this patient population.10 36 37 Based on this comprehensive review, it is rather clear that low-dose IV tPA can lower the incidence of sICH. However, it is unclear if a lower dose would offer the same efficacy as standard dosage in selected patients based on this review. Last but not the least, low-dose IV tPA (a single 50 mg vial of Actilyse, Boehringer Ingelheim) could become a preferred option for patients who cannot afford the full dose in certain areas of the world.

The ENCHANTED trial included patients within 4.5 hours of onset in both dose groups. In European Cooperative Acute Stroke Study (ECASS) III trials, those who were older than 80, with a history or stroke and diabetes and anticoagulation, had worse outcome. Perhaps in these patients now excluded based on ECASS III results, 0.6 mg/kg might be beneficial. Other studies also supported the use of a lower dose in this patient population.10 36 37 Based on this comprehensive review, it is rather clear that low-dose IV tPA can lower the incidence of sICH. However, it is unclear if a lower dose would offer the same efficacy as standard dosage in selected patients based on this review. Last but not the least, low-dose IV tPA (a single 50 mg vial of Actilyse, Boehringer Ingelheim) could become a preferred option for patients who cannot afford the full dose in certain areas of the world. Conclusion Although the level of evidence was not high, studies prior to the ENCHANTED trial have shown that low-dose IV tPA was associated with a lower rate of sICH and effective in treating patients with AIS.10 Despite some controversy, the ENCHANTED trial has confirmed the lower rate of sICH with 0.6 mg/kg of IV tPA.10 However, the 0.6 mg/kg dose group was associated with a high incidence of moderate-to-severe disability and did not achieve the primary outcome as predetermined statistically. Since the mortality rate in both groups was similar, the increased rate of moderate disability was partially balanced by the lower rate of sICH. Beside 0.6 mg/kg, other low doses of IV tPA have not been well studied. In many developing countries in Asia such as China, a low dose provided an attractive option since it would lower the cost and help to overcome the underuse of IV tPA. The strategy of IV tPA dosage should depend on the patient's clinical profile, financial condition and a clear understanding of the current evidence. Standard dose should be considered first, but 0.6 mg/kg may be an option in certain subgroups. The practising physicians should take into the consideration that patients who received 0.6 mg/kg IV tPA are more likely to be alive at 7 days and with a low chance of developing sICH, but with a higher chance of having moderate disability.

hould be considered first, but 0.6 mg/kg may be an option in certain subgroups. The practising physicians should take into the consideration that patients who received 0.6 mg/kg IV tPA are more likely to be alive at 7 days and with a low chance of developing sICH, but with a higher chance of having moderate disability. Contributors: YD and QD designed the study. YD and WC did the original data search and screening. XC, LY, YX and KF participated in data extraction. YD drafted the manuscript, and WC and QD revised the manuscript. Funding: QD was funded by National Nature Science of China (81571109), XC was funded by National Nature Science of China (81100861) and Sponsored by Shanghai Rising-Star Program. Competing interests: None declared. Provenance and peer review: Not commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

s important to note that despite their similarities, differences in expression levels of SUMO2 and SUMO3 can have a profound impact on physiological outcomes. For example, SUMO2 is the predominately expressed SUMO isoform in mice. Deletion of Sumo2 is embryonic lethal, while deletion of Sumo3 shows no overt phenotype.7 SUMO proteins can covalently bond to the lysine residues of target proteins (SUMOylation), and thereby regulate their activity, stability and subcellular localisation. Before SUMOylation can occur, however, SUMO-specific proteases (SENPs) must proteolyticly process all SUMO precursors to expose the C-terminal functional motif. Similar to ubiquitination, SUMOylation requires three conserved steps that are catalysed by corresponding enzymes: an E1 activating enzyme—SUMO-activating enzyme subunit 1/2 (SAE1/SAE2), the solo E2 conjugating enzyme Ubc9, and various E3 ligating enzymes. The first activation step, mediated by an E1 enzyme, is an ATP-dependent reaction, and thus could be the limiting step of SUMOylation in certain conditions, including brain ischaemia, that cause rapid ATP depletion. Ubc9 is the only E2 conjugating enzyme that has been identified. Blocking or overexpressing Ubc9 has been widely used as an effective approach to study the consequence of modulating global SUMOylation in cells or animals under normal or stress conditions.8 9 The mechanism(s) that drive substrate specificity of SUMOylation are poorly understood. The interaction between substrates and the domains of specific E3 ligating enzymes contributes to the substrate selection in some cases. However, thousands of SUMO targets have been identified by proteomics studies, which is in sharp contrast to only a handful of known SUMO ligating enzymes. This suggests that many SUMO ligating enzymes have not yet been identified, and/or that other mechanisms could control the substrate specificity of SUMOylation.

Introduction Spontaneous cervical artery dissection (CAD) accounts for only ∼2% of all ischaemic strokes, but it accounts for 10–25% of ischaemic strokes in young adult patients.1 CAD can cause ischaemic symptoms typically through two mechanisms, which are thromboembolic and haemodynamic compromises.2 3 The early and reliable diagnosis of CAD is highly important for treatment decision-making in the era of precision medicine.4 5 Conventional angiography has long been the gold standard for the diagnosis of CAD.6 However, this method is invasive and does not have the advantage of demonstrating mural haematoma.7 With the advent and development of MRI techniques, like MR angiography (MRA) combined with an axial two-dimensional (2D) fat-saturated spin echo T1-weighted sequence has shown an evident advantage in the diagnosis of CAD.1 8–10 However, the axial 2D T1-weighted sequence has several limitations. It is time-consuming to cover all arteries and shows poor performance for the arteries with a tortuous course.11 12 The three-dimensional (3D) isotropic T1-weighted spin echo sequence can overcome the limitations of a 2D sequence.13 14 Some researchers have found that a 3D black blood T1-weighted sequence of volumetric isotropic turbo spin echo acquisition (VISTA) offers similar or more information than a 2D T1-weighted spin echo sequence.15–17 We aimed to investigate the value of a 3D T1 VISTA sequence at 3.0 T for the diagnosis of CAD. We also investigated the limitations of 3D T1 VISTA in the diagnosis of CAD.

quence of volumetric isotropic turbo spin echo acquisition (VISTA) offers similar or more information than a 2D T1-weighted spin echo sequence.15–17 We aimed to investigate the value of a 3D T1 VISTA sequence at 3.0 T for the diagnosis of CAD. We also investigated the limitations of 3D T1 VISTA in the diagnosis of CAD. Methods In reporting the current study, the standards for reporting diagnostic accuracy studies were followed.18 This study was not publicly registered. Participants Between September 2014 and February 2016, we prospectively included consecutive patients who were suspected as having an acute CAD with the following symptoms or signs: Horner's syndrome, unusual neck pain and/or headache, cranial nerve palsy and tinnitus. The CAD was considered acute if the duration of symptoms was ≤30 days. Patients who had the above signs or symptoms, particularly if in combination and/or associated with a cerebral or retinal ischaemia, were highly suspected as having a CAD. Patients who were suspected as CAD by imaging tools of MRA or ultrasound were also included. This study was approved by the institutional review board of the first hospital of Jilin University. Written consent forms were obtained from all patients.

a cerebral or retinal ischaemia, were highly suspected as having a CAD. Patients who were suspected as CAD by imaging tools of MRA or ultrasound were also included. This study was approved by the institutional review board of the first hospital of Jilin University. Written consent forms were obtained from all patients. Imaging protocols A 3D T1 VISTA examination was performed on all patients who were included in this study. At least one examination of angiography including MRA, CT angiography (CTA), and digital subtraction angiography (DSA) was performed on each patient. A 3D T1 VISTA protocol was performed using a 3.0 Tesla scanner (Philips Ingenia, Eindhoven, The Netherlands) with a standard 8-channel head/neck coil. The following parameters were used for the 3D T1 VISTA sequence: an oblique coronal plane acquisition, spectral adiabatic inversion recovery fat saturation mode, repetition time (TR)/echo time (TE)=350 ms/19 ms, field of view=280×199×120 cm3, 400×284 matrix, variable refocusing flip angle, slice interval=0; voxel size=0.7×0.7×0.7 cm3, oversample factor=1.5, and number of excitations=2. The acquisition time was 3 min and 38 s.

al adiabatic inversion recovery fat saturation mode, repetition time (TR)/echo time (TE)=350 ms/19 ms, field of view=280×199×120 cm3, 400×284 matrix, variable refocusing flip angle, slice interval=0; voxel size=0.7×0.7×0.7 cm3, oversample factor=1.5, and number of excitations=2. The acquisition time was 3 min and 38 s. Data analysis The analysis of 3D T1 VISTA results was performed by two experienced neuroradiologists who were blinded to all of the patient information and the final diagnosis. Images along the short and long axes of the arteries could be reconstructed at the workstation. For 3D T1 VISTA, the diagnosis of the dissection was based on the presence of the following features: intramural high signal, particularly if it is a semilunar hyperintense signal; intimal flap and/or double lumen; and aneurysmal dilation.14 The final diagnosis of the dissection was based on the clinical history, physical examination, and all of the imaging tests by two experienced neurologists. For conventional imaging tests, the imaging signs for diagnosis of CAD including double lumen, intimal flap, pearl and string sign, string sign, and tapered occlusion.15 Disagreement between the two observers were resolved be consensus.

physical examination, and all of the imaging tests by two experienced neurologists. For conventional imaging tests, the imaging signs for diagnosis of CAD including double lumen, intimal flap, pearl and string sign, string sign, and tapered occlusion.15 Disagreement between the two observers were resolved be consensus. SPSS V.19.0 (IBM, West Grove, Pennsylvania, USA) was used to perform the analysis. Since no gold standard for the diagnosis of CAD was available, the final diagnosis results were chosen to be the reference standard to calculate the sensitivity and specificity (including the corresponding 95% CI) for 3D T1 VISTA. Interobserver agreement for 3D T1 VISTA was examined by using the κ-coefficient of agreement.19 Fisher's exact test was used for count data. The level of statistical significance was set at p<0.05. Results A total of 46 patients were included in this study. 3D T1 VISTA examination was performed on each of them. The final diagnosis of the dissection was made for 21 patients (6 females) (table 1). There were eight patients with single carotid artery dissection and nine patients with single vertebral artery dissection. One patient had bilateral carotid artery dissections and bilateral vertebral artery dissections. Three patients had bilateral vertebral artery dissections. One patient with vertebral artery dissection presented with neck pain accompanied by right-sided weakness and left-sided numbness (Brown-Séquard syndrome) (patient number 11). The median age of patients with dissection was 38 years (range 27–74 years old).

ons. Three patients had bilateral vertebral artery dissections. One patient with vertebral artery dissection presented with neck pain accompanied by right-sided weakness and left-sided numbness (Brown-Séquard syndrome) (patient number 11). The median age of patients with dissection was 38 years (range 27–74 years old). Table 1 Characteristics of patients with a final diagnosis of dissection

ons. Three patients had bilateral vertebral artery dissections. One patient with vertebral artery dissection presented with neck pain accompanied by right-sided weakness and left-sided numbness (Brown-Séquard syndrome) (patient number 11). The median age of patients with dissection was 38 years (range 27–74 years old). Table 1 Characteristics of patients with a final diagnosis of dissection Patient Age Symptoms Acute infarction Location of dissection 3D T1 VISTA Occlusion 1 40s Headache and transient motor weakness on the right side Yes Right VA At 9 days of onset No 2 50s Neck pain and dizziness No Right VA At 15 days of onset No 3 40s Transient motor weakness on the left side Yes Right ICA At 12 days of onset Yes 4 70s Transient failure of vision and speech No Left ICA At 13 days of onset No 5 30s Hemiplegia and Horner's syndrome Yes Right ICA At 16 days of onset Yes 6 30s Neck pain and dizziness Yes Bilateral VAs At 7 days of onset Right (yes), left (no) 7 30s Neck pain and hemiplegia No Left ICA At 6 days of onset Yes 8 30s Neck pain and Horner's syndrome Yes Left ICA At 15 days of onset No 9 50s Dizziness Yes Left VA At 7 days of onset Yes 10 30s Aphasia and blurred vision Yes Left ICA At 8 days of onset No 11 40s Neck pain and Brown-Séquard syndrome Yes* Right VA At 6 days of onset No 12 20s Transient dizziness, double vision, and motor weakness on the right side No Bilateral VAs At 11 days of onset Right (yes), left (no) 13 30s Headache and transient dizziness and motor weakness on the left side No Bilateral VAs At 6 days of onset No 14 20s Dizziness and double vision Yes Left VA At 3 days of onset No 15 30s Aphasia and hemiplegia Yes Left ICA At 3 days of onset Yes 16 50s Asymptomatic No Left VA NA No 17 20s Dizziness and hemiplegia Yes Left VA At 22 and 56 days of onset Yes 18 40s Facial paralysis and aphasia Yes Bilateral VAs and bilateral ICAs At 4 days of onset No 19 70s Dysarthria and hemiplegia Yes Left VA At 3 days of onset No 20 30s Aphasia and hemiplegia Yes Left ICA At 26 days of onset Yes 21 40s Dizziness, headache, and hemiplegia Yes Left VA At 5 days and 23 days of onset Yes The diagnosis of dissection was not detected by T1 VISTA for patient number 20.

days of onset No 19 70s Dysarthria and hemiplegia Yes Left VA At 3 days of onset No 20 30s Aphasia and hemiplegia Yes Left ICA At 26 days of onset Yes 21 40s Dizziness, headache, and hemiplegia Yes Left VA At 5 days and 23 days of onset Yes The diagnosis of dissection was not detected by T1 VISTA for patient number 20. *Cervical spinal cord ischaemia was identified by MRI. 3D, three-dimensional; ICA, internal carotid artery; NA, not applicable; VA, vertebral artery; VISTA, volumetric isotropic turbo spin echo acquisition. The median time between disease onset and the examination with 3D T1 VISTA was 11 days with a range of 3–26 days for patients with dissection. Diagnosis of dissection was made for 20 patients after assessing the 3D T1 VISTA data alone by two researchers. Among these 20 patients, the diagnosis of dissection was not made for five (including 3 patients with DSA) of them before the examination with T1 VISTA. The sensitivity and specificity for 3D T1 VISTA were 95.2% (95% CI, 76.2% to 99.9%) and 100% (95% CI, 86.3% to 100%), respectively. The agreement between the two researchers for T1 VISTA for diagnosis of CAD was very good (k=0.91). Patient number 1 presented with headache and transient motor weakness on the right side (table 1). Duplex ultrasonography and a subsequently CTA showed stenosis and possible dissection of the right vertebral artery, which was confirmed by 3D T1 VISTA (figure 1). T1 VISTA images of the other 18 patients all showed typical crescent-shaped hyperintense signals surrounding the lumen (patients number 2–19).

de (table 1). Duplex ultrasonography and a subsequently CTA showed stenosis and possible dissection of the right vertebral artery, which was confirmed by 3D T1 VISTA (figure 1). T1 VISTA images of the other 18 patients all showed typical crescent-shaped hyperintense signals surrounding the lumen (patients number 2–19). Figure 1 Patient number 1 with right vertebral artery dissection. (A) CT angiography shows stenosis of the right vertebral artery. (B) Curved planar reconstruction of three-dimensional (3D) T1 volumetric isotropic turbo spin echo acquisition (VISTA) images shows the arterial wall hyperintensity of the right vertebral artery. (C) Axial reconstructions of 3D T1 VISTA images show the crescent-like hyperintensity of the wall haematoma. Patient number 20 presented with aphasia and motor weakness on the right side. Duplex ultrasonography of the left internal carotid artery at 15 days of onset showed a double lumen and intima, which suggested the aetiology was dissection. However, there was no typical crescent-shaped hyperintense signal on VISTA at 26 days of onset, one of the two researchers insisted that the definite diagnosis of dissection could not be made by T1 VISTA alone (figure 2).

t 15 days of onset showed a double lumen and intima, which suggested the aetiology was dissection. However, there was no typical crescent-shaped hyperintense signal on VISTA at 26 days of onset, one of the two researchers insisted that the definite diagnosis of dissection could not be made by T1 VISTA alone (figure 2). Figure 2 Patient number 20 with a left internal carotid artery dissection. (A) Coronal of three-dimensional (3D) T1 volumetric isotropic turbo spin echo acquisition (VISTA) image shows hyperintensity of the left internal carotid artery at 26 days of onset. (B) Axial reconstructions of 3D T1 VISTA images show the occluded left internal carotid artery without typical crescent-like hyperintensity. (C) Ultrasound of the left internal carotid artery at 15 days of onset shows a double lumen and intima. Patient number 21 presented with dizziness, headache, and hemiplegia. 3D T1 VISTA at 5 days of onset showed high-signal intensity of the left vertebral artery (figure 3). One of the reconstructed images along the short-axis of left vertebral artery showed a non-typical crescent-shaped hyperintense signal. One researcher suggested that it was unreliable to make a diagnosis of dissection based on the images of T1 VISTA. However, a repeat scan of T1 VISTA at 23 days of onset showed the recanalisation of occlusion and a residual intramural haematoma, which confirmed the diagnosis of dissection.

t-shaped hyperintense signal. One researcher suggested that it was unreliable to make a diagnosis of dissection based on the images of T1 VISTA. However, a repeat scan of T1 VISTA at 23 days of onset showed the recanalisation of occlusion and a residual intramural haematoma, which confirmed the diagnosis of dissection. Figure 3 Patient number 21 with a left vertebral artery dissection. (A–C) Coronal of three-dimensional (3D) T1 volumetric isotropic turbo spin echo acquisition (VISTA) image, curved planar reconstruction of 3D T1 VISTA images and short-axis view of the left vertebral artery at 5 days of onset show high-signal intensity of the left vertebral artery. (D–F) A repeat scan at 23 days of onset shows the recanalisation of the occlusion and the residual intramural haematoma. For patients without artery occlusion, all of them had a definite conclusion with or without dissection by VISTA (n=29). However, for 17 patients with artery occlusion, the possibility of dissection could not be excluded for six of them by VISTA (0/29 vs 6/17; p=0.001). For these six patients, the aetiology has not been determined for five of them (table 2), even after assessing all of the data, including T1 VISTA and DSA. 3D T1 VISTA showed hyperintense signals in the occluded artery segment without typical features of dissection. One of the patients was <30 years old without any risk factors of cerebrovascular diseases (patient number 25). The follow-up T1 VISTA for two of these five patients showed no recanalisation of the occluded arteries.

ISTA showed hyperintense signals in the occluded artery segment without typical features of dissection. One of the patients was <30 years old without any risk factors of cerebrovascular diseases (patient number 25). The follow-up T1 VISTA for two of these five patients showed no recanalisation of the occluded arteries. Table 2 Characteristics of patients without a definitive diagnosis of with or without dissection Patient Age Symptoms Acute infarction Location of lesion 3D T1 VISTA Occlusion 22 40s Dizziness and hemiplegia Yes Bilateral VAs At 18 days of onset Yes 23 50s Hemiplegia Yes Left ICA At 25 and 48 days of onset Yes 24 60s Dizziness and double vision Yes Right VA At 9 days of onset Yes 25 20s Dizziness and hemiplegia Yes Left VA At 19 days of onset Yes 26 60s Hemiplegia Yes Right VA At 15 days of onset Yes 3D, three-dimensional; ICA, internal carotid artery; VA, vertebral artery; VISTA, volumetric isotropic turbo spin echo acquisition. Discussion This study showed that the 3D T1 VISTA sequence at 3.0 T is useful in the diagnosis of spontaneous CAD. It can show intramural haematoma of the cervical artery clearly with a single acquisition and acceptable scan time. However, for some patients with totally acute occlusion of the artery without typical crescent-shaped hyperintense signals, a diagnosis of the aetiology of artery occlusion may be difficult by VISTA alone.

t can show intramural haematoma of the cervical artery clearly with a single acquisition and acceptable scan time. However, for some patients with totally acute occlusion of the artery without typical crescent-shaped hyperintense signals, a diagnosis of the aetiology of artery occlusion may be difficult by VISTA alone. Several studies have investigated the efficiency of 3D black blood T1 sequences using variable refocusing flip-angle turbo-spin-echo imaging in the diagnosis of CAD. All of them showed that a 3D black blood T1 sequence is more useful than conventional imaging tools. Takemoto et al17 first reported the value of T1 VISTA sequence at 1.5 T in the diagnosis of CAD. They concluded that a 3D black blood T1 sequence can improve the assessment of intramural haematoma in vertebral artery dissection compared with 2D spin-echo T1-weighted images and time-of-flight MRA (TOF-MRA).16 Another study showed that abnormal vessel enhancement was recognised in 15 of 15 patients with vertebral artery dissection on contrast-enhanced T1 VISTA images. There are two other similar 3D black blood T1sequences that are commercially available (T1 CUBE and T1 SPACE from GE Healthcare, Milwaukee, Wisconsin and Siemens, Erlangen, Germany, respectively).11 20 Studies showed that these two 3D black blood T1 sequences at 1.5 T or 3.0 T may also be a substitute for 2D T1 sequences in the diagnosis of CAD.11 20 The inter-rater and intrarater agreements were good for 3D black blood T1 sequences.11 14

hcare, Milwaukee, Wisconsin and Siemens, Erlangen, Germany, respectively).11 20 Studies showed that these two 3D black blood T1 sequences at 1.5 T or 3.0 T may also be a substitute for 2D T1 sequences in the diagnosis of CAD.11 20 The inter-rater and intrarater agreements were good for 3D black blood T1 sequences.11 14 There are several advantages of 3D black blood T1 sequences in the diagnosis of CAD compared with traditional imaging tools. First, this method can reveal the vessel wall of intracranial and extracranial arteries in a single acquisition with good image quality, good dark blood contrast, isotropic voxels, and a relatively short scan time. Second, the method can show the mural haematoma as high-signal intensity clearly. The mural haematoma is not easily identified from the source image of MRA and CTA and cannot be revealed by vascular images such as MRA, CTA, and DSA. Third, because of its isotropic volume acquisition, it can obtain the images of multiplanar reformation. Multiplanar reformation images can show the features of dissection at different angles and planes, which is especially helpful for an artery with tortuous course.

vealed by vascular images such as MRA, CTA, and DSA. Third, because of its isotropic volume acquisition, it can obtain the images of multiplanar reformation. Multiplanar reformation images can show the features of dissection at different angles and planes, which is especially helpful for an artery with tortuous course. A recently published study showed that 3D simultaneous non-contrast angiography and intraplaque haemorrhage (SNAP) imaging can provide non-contrast MRA and vessel wall images simultaneously in a single acquisition with a shorter scanning time.21 This study demonstrated excellent agreement with multisequence MRI in evaluating luminal stenosis and intramural haematoma in patients with craniocervical artery dissection. It seems that a 3D SNAP sequence may be a better choice than the aforementioned 3D black blood sequences because it can provide MRA and vessel wall images simultaneously in a single scan. However, this technique has not been widely investigated and is not commercially available.

s with craniocervical artery dissection. It seems that a 3D SNAP sequence may be a better choice than the aforementioned 3D black blood sequences because it can provide MRA and vessel wall images simultaneously in a single scan. However, this technique has not been widely investigated and is not commercially available. We also investigated the limitations of 3D T1 VISTA in the diagnosis of CAD by analysing the patients for whom a definite conclusion was not reached, with or without dissection by VISTA. Disagreement between the two readers for T1 VISTA existed for two patients who had acute artery occlusion. For one patient, the diagnosis of dissection was made by comprehensive analysis of ultrasound scan, T1 VISTA, and the other information. For the other patient, the diagnosis of dissection was made after a second scan with T1 VISTA. For five patients with acute artery occlusion, the aetiology has not been determined even after assessing all of the data including VISTA and DSA. All of this suggested that, for some patients with acute artery occlusion in whom even the aetiology may be dissection, the diagnosis of dissection cannot be made by T1 VISTA if there was no typical imaging features of dissection.

ogy has not been determined even after assessing all of the data including VISTA and DSA. All of this suggested that, for some patients with acute artery occlusion in whom even the aetiology may be dissection, the diagnosis of dissection cannot be made by T1 VISTA if there was no typical imaging features of dissection. In this study, patients without artery occlusion all had a definite conclusion with or without dissection. However, for patients with artery occlusion, the possibility of dissection may not be excluded, especially for patients with occlusion of the vertebral artery because of its smaller size.22 This suggested that the difference between intramural haematoma and intraluminal thrombus may be difficult with 3D T1 VISTA even at 3.0 T for some patients. For this group of patients, a follow-up scan, a contrast-enhanced scan, or an optimal VISTA technique may be helpful (patient number 16).20

of its smaller size.22 This suggested that the difference between intramural haematoma and intraluminal thrombus may be difficult with 3D T1 VISTA even at 3.0 T for some patients. For this group of patients, a follow-up scan, a contrast-enhanced scan, or an optimal VISTA technique may be helpful (patient number 16).20 There were several limitations in this study. First, the sample size was small and the contrast-enhanced imaging was not performed. Contrast-enhanced 3D T1 VISTA might be useful in manifesting subtle structure abnormalities, assessing vessel inflammatory reaction, and distinguishing intramural haematoma from intraluminal thrombus. Second, because previous studies already showed that a 3D black blood sequence can provide similar information and may be a substitute for 2D sequence, the 2D T1 black blood sequence was not performed for the patients. Third, we did not perform a follow-up 3D T1 VISTA scan for certain patients. Finally, all of the information including T1 VISTA were used to make the reference standard, which may lead to inaccurate results of the sensitivity and specificity for T1 VISTA.

e 2D T1 black blood sequence was not performed for the patients. Third, we did not perform a follow-up 3D T1 VISTA scan for certain patients. Finally, all of the information including T1 VISTA were used to make the reference standard, which may lead to inaccurate results of the sensitivity and specificity for T1 VISTA. Conclusions In conclusion, this study showed that a 3D T1 VISTA sequence at 3.0 T is useful in the diagnosis of spontaneous CAD. This sequence can show intramural haematoma of the cervical artery clearly with a single acquisition and acceptable scan time. However, for some patients with totally acute occlusion of the artery without typical features of dissection, the unequivocal distinction between intramural haematoma and intraluminal thrombus may still be difficult with T1 VISTA alone. Future studies should investigate whether a repeat scan or an optimal VISTA technique would be useful for making a definite diagnosis. The authors would like to thank Elsevier Language Editing Services for English language editing. Contributors: P-PN and YY contributed to the conception and design of the study, analysis and interpretation of the data, and drafting of the manuscript. P-PN performed the statistical analysis. H-WZ and YL contributed to the analysis of data and drafting of the manuscript. YL, Z-NG and HJ contributed to the design of the study, the acquisition of data, and drafting of the manuscript. All the authors approved the publication of the study. Funding: This work was supported by Changbai Mountain Scholars, Jilin Provincial government to YY.

Contributors: P-PN and YY contributed to the conception and design of the study, analysis and interpretation of the data, and drafting of the manuscript. P-PN performed the statistical analysis. H-WZ and YL contributed to the analysis of data and drafting of the manuscript. YL, Z-NG and HJ contributed to the design of the study, the acquisition of data, and drafting of the manuscript. All the authors approved the publication of the study. Funding: This work was supported by Changbai Mountain Scholars, Jilin Provincial government to YY. Competing interests: None declared. Patient consent: Obtained. Ethics approval: The Institutional Review Board of the First Hospital of Jilin University. Provenance and peer review: Not commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

es. However, thousands of SUMO targets have been identified by proteomics studies, which is in sharp contrast to only a handful of known SUMO ligating enzymes. This suggests that many SUMO ligating enzymes have not yet been identified, and/or that other mechanisms could control the substrate specificity of SUMOylation. SUMO modification is a dynamic process involving SUMOylation and de-SUMOylation. The de-SUMOylation to remove SUMO from targets is primarily mediated by SENPs. Six SENPs (SENP1-3 and 5–7) have been identified in mammalian cells (reviewed in Hickey et al).10 SENPs are cysteine proteases that play two major roles in the SUMO pathway: (1) as endopeptidases to remove the C-terminal extension and expose the di-glycine (GG) motif of SUMO precursors, and (2) as isopeptidases to de-conjugate SUMOylated proteins. SENPs show distinct SUMO preferences for endopeptidase and isopeptidase activities. Using purified proteins, Mendes et al11 found that SENP1 and SENP2 are the most potent of the SENPs for endopeptidase and isopeptidase activity in all SUMO isoforms. However, studies on SENP mutant mice revealed that SENP1 acts to de-SUMOylate primarily SUMO1-conjugated proteins, while SENP2 preferentially de-conjugates SUMO2/3-conjugated proteins.12 13 This in vivo information is crucial when targeting SENPs for drug discovery, since the compounds identified eventually will be tested in animals. For example, if the goal is to identify potent compounds that increase mainly the levels of SUMO2/3-conjugated proteins in animals, SENP2, among all SENPs, should be considered as a screening target. Of note, in addition to SENPs, three new SUMO proteases (DeSI-1, DeSI-2 and USPL1) have been identified.14 15 They appear to have de-conjugation capability on only a few particular substrates, however. Indeed, in contrast to SENPs, silencing these enzymes has no obvious effect on global SUMOylation in cells.14 15

Introduction The brain is particularly vulnerable to even a short period of insufficient blood supply, because of its lack of energy reserves. Ischaemic insult in the brain often impairs neurological functions, and is associated with major cardiac surgery and various pathological states of high clinical significance including cardiac arrest and ischaemic stroke. Each year, almost 800 000 people in the USA alone suffer devastating consequences of stroke,1 and this problem is even more serious in developing countries.2 Stroke-induced disability is a great burden for patients and their families, and presents significant challenges for healthcare systems today. Since age is a key risk factor for stroke, these problems will become even more pronounced in the near future as the median age of our population continues to increase. Despite decades of efforts to develop strategies that protect the brain against ischaemia-induced damage, therapeutic hypothermia remains the only neuroprotective option in the clinic. No pharmacological neuroprotectant has yet been shown to improve outcomes for patients with brain ischaemia. Therefore, novel neuroprotective strategies are still urgently needed to increase resistance of the brain to ischaemic insult.

mage, therapeutic hypothermia remains the only neuroprotective option in the clinic. No pharmacological neuroprotectant has yet been shown to improve outcomes for patients with brain ischaemia. Therefore, novel neuroprotective strategies are still urgently needed to increase resistance of the brain to ischaemic insult. Brain ischaemia is a complex pathological process that involves many signalling pathways. Traditionally, single signalling pathways or functional processes have been explored as targets for therapeutic strategies that provide neuroprotection. However, all clinical trials based on this work have yielded disappointing results in the clinical setting.3 4 An approach that targets a multifunctional cellular process that coordinates many signalling pathways/processes in brain ischaemia may be a better strategy.4 Mounting evidence indicates that small ubiquitin-like modifier (SUMO) conjugation (SUMOylation), a post-translational modification, is a promising candidate, and as such, a novel target with tremendous neuroprotective potential in brain ischaemia. In this review, we first provide a brief overview of the SUMO pathway, then summarise the current findings on the role of SUMOylation in human diseases, especially brain ischaemia, and finally highlight recent progress in drug discovery that targets global SUMOylation with a major focus on applications in brain iscahemia.

this review, we first provide a brief overview of the SUMO pathway, then summarise the current findings on the role of SUMOylation in human diseases, especially brain ischaemia, and finally highlight recent progress in drug discovery that targets global SUMOylation with a major focus on applications in brain iscahemia. The SUMO pathway SUMO is a group of small proteins consisting of about 100 amino acids.5 6 In mammalian cells, three SUMO isoforms have been discovered: SUMO1, 2 and 3. SUMO2 and SUMO3 are almost identical, and are thus normally referred to as SUMO2/3. SUMO1, however, shows only 50% sequence homology with SUMO2/3. It is important to note that despite their similarities, differences in expression levels of SUMO2 and SUMO3 can have a profound impact on physiological outcomes. For example, SUMO2 is the predominately expressed SUMO isoform in mice. Deletion of Sumo2 is embryonic lethal, while deletion of Sumo3 shows no overt phenotype.7

pression, DNA damage repair, RNA processing and quality control of newly synthesised proteins—all of which are essential for maintaining cellular homoeostasis. Numerous studies have now provided evidence that links SUMOylation to the pathophysiology of many diseases including cancer, heart diseases and brain ischaemia. Cancer Among all SUMO-related diseases investigated to date, cancer has been the most extensively studied. Considering the role of SUMOylation in maintaining cell function under stress/unfavourable conditions, it is not surprising that substantial evidence indicates a positive association between SUMOylation and cancer cell growth, tumourigenesis, metastasis and poor prognosis.16 For example, we found that levels of both Ubc9 and global SUMOylation are significantly increased in human brain tumour samples, and are most pronounced in glioblastoma multiforme tumours (GBM), the deadliest malignant primary brain tumours.17 Given that GBM are known for high resistance to standard radiotherapy and chemotherapy, and that SUMOylation is a key component in DNA damage repair processes, it is intriguing to speculate that inhibition of global SUMOylation may increase the sensitivity of glioblastomas to radiotherapy or chemotherapy. Importantly, Bossis et al18 demonstrated that pharmacological inhibition of the SUMO pathway can overcome chemoresistance in some leukaemia cell lines. Targeting SUMOylation in cancer therapy has now attracted considerable interest.16

lation may increase the sensitivity of glioblastomas to radiotherapy or chemotherapy. Importantly, Bossis et al18 demonstrated that pharmacological inhibition of the SUMO pathway can overcome chemoresistance in some leukaemia cell lines. Targeting SUMOylation in cancer therapy has now attracted considerable interest.16 Heart diseases SUMO homoeostasis must be finely balanced for normal heart development and function. An imbalance in SUMOylation/de-SUMOylation contributes to cardiovascular heart diseases (for detailed review, see Mendler et al).19 For example, SUMOylation of the sarcoplasmic/endoplasmic reticulum (ER) Ca2+ ATPase (SERCA2a) positively controls its activity and stability, which is critical for heart function. In mice with heart failure, the levels of SUMO1 decrease in the heart. Consequently, SERCA2a loses its SUMOylation state, which reduces its ATP-binding affinity and ATPase activity. Notably, the pathological state of the failing heart can be substantially alleviated by virus-mediated SUMO1 gene delivery.20 Similarly, in a swine model of ischaemic heart failure, viral transfer of the SUMO1 gene significantly improves cardiac function.21 Although these studies indicate that SUMOylation-based strategies have great potential for treating heart diseases, published findings in this field are not consistent, and thus a better understanding of the role of SUMOylation in heart diseases is still needed.19

e SUMO1 gene significantly improves cardiac function.21 Although these studies indicate that SUMOylation-based strategies have great potential for treating heart diseases, published findings in this field are not consistent, and thus a better understanding of the role of SUMOylation in heart diseases is still needed.19 Brain ischaemia A massive increase in global SUMOylation in the brain was first observed in hibernating squirrels.8 In this study, the Hallenbeck group used hibernation as a natural ischaemia tolerance model to identify endogenous neuroprotective pathways, and found a marked increase in the levels of both SUMO1-conjugated and SUMO2/3-conjugated proteins in the brain during the torpor phase of hibernation. During torpor, the body temperature is ∼5°C, and heart rate and cerebral blood flow are dramatically reduced to otherwise lethal levels. The authors therefore hypothesised that increased global SUMOylation is a molecular mechanism underlying this ischaemia tolerance during torpor, that is, hypothermia-mediated neuroprotection. To support this hypothesis, they showed that preconditioning (a short period of oxygen/glucose deprivation (OGD)) with hypothermia increases global SUMOylation in SHSY5Y cells, which in turn, contributes to cell protection against severe OGD, an in vitro ischaemia model. Since therapeutic hypothermia is currently used in the clinic to protect the brain from ischaemic damage, our group set out to determine whether SUMOylation also plays a role in neuroprotection conferred by therapeutic hypothermia. We demonstrated, for the first time, that when animals are exposed to various degrees of hypothermia ranging from 18°C to 30°C, both deep and moderate hypothermia activate the SUMOylation pathway, and induce nuclear accumulation of SUMO2/3-conjugated proteins.22 23 Therefore, we proposed that SUMOylation contributes to the neuroprotection afforded by therapeutic hypothermia. This notion has been substantiated by the finding that hypothermic treatment has no additional protective effect on stroke outcome in Ubc9 transgenic mice, which sustain high levels of global SUMOylation.24 Together, these results strongly support a major role for increased global SUMOylation in hypothermia-mediated brain protection against ischaemic insults. However, to definitively prove this hypothesis, genetically modified SUMO mice and pharmacological modulators of the SUMOylation pathway may be required in future studies.

er, these results strongly support a major role for increased global SUMOylation in hypothermia-mediated brain protection against ischaemic insults. However, to definitively prove this hypothesis, genetically modified SUMO mice and pharmacological modulators of the SUMOylation pathway may be required in future studies. If this hypothesis is eventually validated, pharmacologically increasing global SUMOylation could be a promising clinical strategy for protecting the brain against ischaemic damage associated with various surgical procedures, and thereby avoid the adverse effects related to therapeutic hypothermia treatment.

er, these results strongly support a major role for increased global SUMOylation in hypothermia-mediated brain protection against ischaemic insults. However, to definitively prove this hypothesis, genetically modified SUMO mice and pharmacological modulators of the SUMOylation pathway may be required in future studies. If this hypothesis is eventually validated, pharmacologically increasing global SUMOylation could be a promising clinical strategy for protecting the brain against ischaemic damage associated with various surgical procedures, and thereby avoid the adverse effects related to therapeutic hypothermia treatment. The first evidence that established a link between SUMOylation and a pathological state of brain ischaemia was reported by our group. We found that transient forebrain ischaemia rapidly induces a massive increase in the levels of SUMO2/3-conjugated proteins during reperfusion.25 This finding initiated many follow-up studies on SUMOylation in various brain ischaemia animal models.9 24 26–30 In one of these studies, we showed that after transient focal cerebral ischaemia induced by middle cerebral artery occlusion (MCAO), an ischaemic stroke model, activation of SUMO2/3 conjugation is most pronounced in neurons located at the border of the tissue supplied by the MCA where it is characterised by translocation of SUMO-conjugated proteins to the nucleus. Surprisingly, even a short, non-damaging period of MCA occlusion is sufficient to massively activate this process.27 These data suggest that postischaemic activation of SUMO2/3 conjugation is a protective stress response in neurons. This notion is corroborated by results from cell culture and animal studies.8 9 31–33 For example, under normal conditions, blocking SUMO2/3 conjugation in primary neurons does not trigger cell damage. However, after OGD, it significantly increases cell death.33 Transgenic mice overexpressing Ubc9 (the only SUMO E2 conjugating enzyme) exhibit increased levels of SUMO1-conjugated and SUMO2/3-conjugated proteins, and show improved outcome after stroke.9 These data demonstrate the neuroprotective role of SUMO conjugation after brain ischaemia. To more directly characterise the function of each SUMO in brain ischaemia, individual SUMO knockout mice are needed. All of these mice have recently become available,7 34 except SUMO2 conditional knockout mice.7

e after stroke.9 These data demonstrate the neuroprotective role of SUMO conjugation after brain ischaemia. To more directly characterise the function of each SUMO in brain ischaemia, individual SUMO knockout mice are needed. All of these mice have recently become available,7 34 except SUMO2 conditional knockout mice.7 Potential mechanisms underlying SUMOylation-mediated neuroprotection in brain ischaemia Thousands of SUMO target proteins have been identified. The increase in global SUMOylation observed in brain ischaemia is expected to affect many proteins and their related pathways/processes. Thus, it could be difficult, if not impossible, to pinpoint one, or even a few SUMO targets or pathways that play a key role in SUMOylation-mediated neuroprotection. However, recent progress in this field provided valuable information about potential mechanisms, including crosstalk between SUMOylation and ubiquitination, and the change in SUMOylation status of proteins that are crucial to the pathogenesis of brain ischaemia.

play a key role in SUMOylation-mediated neuroprotection. However, recent progress in this field provided valuable information about potential mechanisms, including crosstalk between SUMOylation and ubiquitination, and the change in SUMOylation status of proteins that are crucial to the pathogenesis of brain ischaemia. The crosstalk between SUMOylation and ubiquitination has been reported in several cell culture proteomics studies. One study demonstrated that SUMOylation-activated ubiquitination is essential for quality control of newly synthesised proteins and misfolded proteins.35 Interestingly, the accumulation of misfolded proteins in the ER causes ER stress, which is a hallmark of brain ischaemia. ER stress-mediated cell death is a major factor in ischaemic brain damage. Therefore, SUMOylation-dependent ubiquitination may be involved in defining the fate and function of postischaemic cells. Indeed, a recent proteomics study, using a novel SUMO transgenic mouse model with forebrain-specific expression of epitope-tagged SUMO1–3, revealed, for the first time, the extensive in vivo crosstalk between ubiquitination and SUMOylation after transient forebrain ischaemia.26

fate and function of postischaemic cells. Indeed, a recent proteomics study, using a novel SUMO transgenic mouse model with forebrain-specific expression of epitope-tagged SUMO1–3, revealed, for the first time, the extensive in vivo crosstalk between ubiquitination and SUMOylation after transient forebrain ischaemia.26 Identification and characterisation of individual SUMO targets involved in brain ischaemia is just beginning. Many important proteins in ischaemia-related signalling pathways and neuronal function are SUMO targets. These include hypoxia inducible factor 1, heat shock factor 1, IκBα and potassium channel proteins.36 However, the extent to which SUMOylation of these proteins is involved in the response to brain ischaemia is largely unknown. A recent proteomics study on brain samples collected after forebrain ischaemia, has generated the first data set of in vivo SUMO targets regulated by ischaemia. This data set presents a valuable resource for future studies. In this study, western blot analysis showed that glucocorticoid receptor (GR) is the most prominent SUMO target induced by ischaemia. This is an important discovery and warrants further study, because SUMOylation critically modulates GR activity, and the GR pathway contributes to brain ischaemia outcome. To date, only two studies have reported findings on the role of SUMOylation of a particular protein in cell death after ischaemic stress. In an in vitro OGD study, Guo et al37 reported that SUMOylation of Drp1 ensures that Drp1 remains in the cytoplasm and thus, suppresses Drp1-mediated release of cytochrome c from mitochondria, thereby preventing cell death. A recent in vivo MCAO study provided some evidence suggesting that NCX3 SUMOylation is involved in SUMO-mediated beneficial effects during ischaemic preconditioning.28

ensures that Drp1 remains in the cytoplasm and thus, suppresses Drp1-mediated release of cytochrome c from mitochondria, thereby preventing cell death. A recent in vivo MCAO study provided some evidence suggesting that NCX3 SUMOylation is involved in SUMO-mediated beneficial effects during ischaemic preconditioning.28 Drug discovery targeting SUMOylation for neuroprotection As discussed above, an increase in global SUMOylation is believed to be an endogenous neuroprotective mechanism whereby brain cells become more resistant to a transient episode of low blood flow. Considering the multifactorial effects of SUMOylation, and the continuing failure of clinic trials that target a single pathway/process to neuroprotection in stroke, we believe that SUMOylation is a promising therapeutic target for brain ischaemia. In fact, while the protective mechanisms of SUMOylation in brain ischaemia continue to be unravelled, considerable effort has already been directed towards identifying compounds that pharmacologically increase global SUMOylation, and testing the feasibility of this strategy for neuroprotection.

get for brain ischaemia. In fact, while the protective mechanisms of SUMOylation in brain ischaemia continue to be unravelled, considerable effort has already been directed towards identifying compounds that pharmacologically increase global SUMOylation, and testing the feasibility of this strategy for neuroprotection. In contrast to the many years of research on drug discovery in the ubiquitin field,38 targeting SUMOylation for therapeutic purposes is still in its infancy. SUMOylation and ubiquitination are highly homologous as they both require three enzymatic steps to conjugate their respective modifiers to lysine residues on target proteins, and one enzymatic step for de-conjugation. Therefore, advances in the ubiquitin field will inform the development of potential SUMOylation-based therapeutic interventions.

tion are highly homologous as they both require three enzymatic steps to conjugate their respective modifiers to lysine residues on target proteins, and one enzymatic step for de-conjugation. Therefore, advances in the ubiquitin field will inform the development of potential SUMOylation-based therapeutic interventions. In 2003, the Food and Drugs Administration (FDA) approved bortezomib (Velcade, Millennium Pharmaceuticals), the first successful anticancer drug that targets the ubiquitin–proteasome system.38 The implications of this approval are significant because it not only confirms that the ubiquitin pathway is a valid drug target, but it also demonstrates that targeting a complex network that modulates many proteins is a feasible approach. Recently, de-ubiquitylating enzymes (DUBs), which are critical regulators of the ubiquitin–proteasome system, have emerged as potential drug targets.39 Many drug discovery studies that target the isopeptidase activity of DUBs, have reported impressive results.39 Further, Reverdy et al40 have shown that it is feasible to identify compounds with high selectivity for a specific DUB. In the SUMO pathway, SENPs are the counterparts of DUBs.

ged as potential drug targets.39 Many drug discovery studies that target the isopeptidase activity of DUBs, have reported impressive results.39 Further, Reverdy et al40 have shown that it is feasible to identify compounds with high selectivity for a specific DUB. In the SUMO pathway, SENPs are the counterparts of DUBs. The remarkable progress in the ubiquitin field has inspired confidence that SUMOylation is also a promising target for drug discovery. Since boosting global SUMOylation is believed to be neuroprotective in brain ischaemia, we summarise here only the recent discoveries of compounds that increase SUMOylation. Increasing global SUMOylation requires balance between conjugation and de-conjugation processes. Thus, compounds that activate the SUMO E1 or E2 enzyme, or inhibit SENP de-conjugation enzymes are expected to achieve this goal. A recent study used SUMO E1 and E2 as targets to screen over 100 000 compounds from large National Cancer Institute (NCI) libraries.41 The most promising compound identified was N106, which can increase SUMO conjugation in cells through direct activation of the SUMO E1 enzyme. Notably, treatment with N106 shows beneficial effects on cardiac function in a mouse model of heart failure. If this compound is blood–brain barrier permeable, it may also be applicable in brain ischaemia.

identified was N106, which can increase SUMO conjugation in cells through direct activation of the SUMO E1 enzyme. Notably, treatment with N106 shows beneficial effects on cardiac function in a mouse model of heart failure. If this compound is blood–brain barrier permeable, it may also be applicable in brain ischaemia. However, it is generally believed that inhibition of enzymes by small molecules is much less challenging than activation. Therefore, perhaps the best strategy to identify new classes of compounds that increase SUMOylation is to search for compounds that block de-SUMOylation by inhibiting the isopeptidase activity of SENPs. Indeed, major efforts in this field have been put into screening for specific SENP inhibitors. To date, several classes of SENP inhibitors have been discovered (table 1). Of note, in most of those studies, SENP1 has been used as the screening target due to its preference for SUMO1-conjugated proteins and its key role in cancer development. Table 1 Summary of SENP inhibitors identified by screening

However, it is generally believed that inhibition of enzymes by small molecules is much less challenging than activation. Therefore, perhaps the best strategy to identify new classes of compounds that increase SUMOylation is to search for compounds that block de-SUMOylation by inhibiting the isopeptidase activity of SENPs. Indeed, major efforts in this field have been put into screening for specific SENP inhibitors. To date, several classes of SENP inhibitors have been discovered (table 1). Of note, in most of those studies, SENP1 has been used as the screening target due to its preference for SUMO1-conjugated proteins and its key role in cancer development. Table 1 Summary of SENP inhibitors identified by screening Screening target Screening strategy Screening library Representative hit IC50 Bioactivity in cells Reference SENP1 Benzodiazepine-based design – Compound 38 9.2 µM Not tested 42 SENP1 Cleavage of fluorogenic substrate SUMO1-AMC In-house compound library GN6958 29.6 µM Not tested 43 SENP1 Virtual screening SPECS database SI2 1.29 µM Confirmed 44 PfSENP1 SUMO-procleavage assay with SDS-PAGE detection Cysteine protease inhibitors library JCP-666 17.9 µM Not tested 45 SENP1/2 Virtual screening DTP NSC5068 µM range Confirmed 46 SENP2 Combination of virtual screening and FRET-based assay Namiki-shoji library 1,2,5-oxadiazoles <10 µM Not tested 47 SENPs Cleavage of fluorogenic AFC-based substrates In-house compound library VEA499 µM range No activity 48 DTP, Developmental Therapeutics Program; FRET, fluorescence resonance energy transfer; IC50, half-maximum inhibitory concentration; PfSENP, SENP of Plasmodium falciparum; SENP, SUMO-specific protease; SUMO, small ubiquitin-like modifier.

enic AFC-based substrates In-house compound library VEA499 µM range No activity 48 DTP, Developmental Therapeutics Program; FRET, fluorescence resonance energy transfer; IC50, half-maximum inhibitory concentration; PfSENP, SENP of Plasmodium falciparum; SENP, SUMO-specific protease; SUMO, small ubiquitin-like modifier. Notably, only two virtual screening studies provide evidence that the inhibitors identified can increase global SUMOylation in cells,44 46 and their effectiveness in animals is not yet known. Further, most SENP inhibitors reported to date appear to be suitable only as analytical tools because they lack drug-like properties. Thus, there is still a great need to identify new classes of specific SENP inhibitors. For this, conventional high throughput screening (HTS) is more promising than virtual screening. Currently, there are a few biochemical assays that can be adapted to screen for SENP inhibitors in an HTS system.49 Most of these assays use an SENP catalytic fragment and artificial peptide-based fluorogenic substrates such as SUMO1-7-amino-4-methylcoumarin (AMC), because this kind of assay is easy to set up, and the fluorescence signal generated by a leaving group, for example, AMC, can be quantitatively measured. Fluorescence resonance energy transfer (FRET)-based assays have also been developed. The substrate used in these assays is a SUMO precursor flanked by two fluorescent proteins, which produce a FRET signal if the substrate is intact and not cleaved by SENP. A few HTS studies that used these assay methods have been carried out, and the data are deposited in the PubChem BioAssay database (http://pubchem.ncbi.nlm.nih.gov/). Surprisingly, no hits from these screens have yet been reported in research publications.

FRET signal if the substrate is intact and not cleaved by SENP. A few HTS studies that used these assay methods have been carried out, and the data are deposited in the PubChem BioAssay database (http://pubchem.ncbi.nlm.nih.gov/). Surprisingly, no hits from these screens have yet been reported in research publications. Here, we propose that physiologically relevant assay substrates, that is, SUMO-conjugated proteins, be used to search for SENP inhibitors in future HTS projects, rather than the artificial substrates aforementioned. Indeed, the data from a nuclear magnetic resonance analysis of the inhibition mechanism of a compound indicated that the inhibitory effect could vary depending on whether an artificial substrate or a physiological substrate is used.46 Although two groups have established FRET-based assays using a physiological substrate—SUMO1-conjugated RanGAP1—neither assay is designed for HTS.50 51 Our group has developed an AlphaScreen-based and HTS-compatible assay that is designed to screen for inhibitors of the isopeptidase activity of SENP.52 This assay uses a SUMO-conjugated protein as substrate, and is robust and reliable.49 This platform, therefore, can be used in future HTS studies to screen for new classes of inhibitors targeting SENP isopeptidase activity. Such future efforts may have a great impact on research in neuroprotection because the new drug-like SENP inhibitors identified are expected to have great potential to boost global SUMOylation in the brain, thereby increasing the brain's resistance to ischaemic injury, and improving outcomes following an ischaemic event.

ch future efforts may have a great impact on research in neuroprotection because the new drug-like SENP inhibitors identified are expected to have great potential to boost global SUMOylation in the brain, thereby increasing the brain's resistance to ischaemic injury, and improving outcomes following an ischaemic event. Of note, a recent study took a different approach to increase SUMOylation, and identified specific inhibitors of miR-182 or miR-183.53 The rationale for this study is that levels of the miR-200 family and the miR-182 family decrease in squirrel brain during the torpor phase when global SUMOylation is increased, and further, that inhibiting these microRNAs increases global SUMOylation in cells.53 In this screening, the authors designed a luciferase reporter assay, and screened three small libraries. A dozen compounds were identified, and interestingly, most of them could increase global SUMOylation and provide protective effects for cells in OGD experiments. Since the screening libraries used are collections of pharmacologically active compounds that are in clinical use, clinical testing or preclinical testing, the repurposing potential of these compounds in animal models of brain ischaemia is high.

l SUMOylation and provide protective effects for cells in OGD experiments. Since the screening libraries used are collections of pharmacologically active compounds that are in clinical use, clinical testing or preclinical testing, the repurposing potential of these compounds in animal models of brain ischaemia is high. Conclusions and perspectives A substantial amount of evidence now supports the notion that increased global SUMOylation is protective in postischaemic cells. Notably, a recent study showed that the flavonoid quercetin protects cells against OGD. This protection is largely mediated by increasing global SUMOylation via the inhibitory effect of quercetin on SENPs.54 This study by Lee and colleagues demonstrated, for the first time, that increasing global SUMOylation via an SENP inhibitor is neuroprotective in cell culture. Taken together, we believe that the research in this field has laid the foundation for launching drug discovery efforts that target SUMOylation to promote neuroprotection in brain ischaemia. This will require a close collaboration between experimental neurology groups and drug discovery groups. Future HTS projects are expected to use physiologically relevant substrates to screen a variety of large libraries. Further, it may be best to search for pan SENP inhibitors rather than SENP-specific inhibitors, in order to increase the probability of success in testing this novel neuroprotective strategy in brain ischaemia. Therefore, the catalytic C-terminal fragment of an SENP, which contains the conserved cysteine protease domain, should be used as the screening target rather than a full-length SENP protein because data suggest that in cells, the N-terminal regions of SENPs determine substrate specificity.55 As more efforts are invested in this field, we expect that data from a proof-of-concept animal experiment will become available soon, and will validate this new exciting target for providing neuroprotection in the brain that is under ischaemic insult.

, the N-terminal regions of SENPs determine substrate specificity.55 As more efforts are invested in this field, we expect that data from a proof-of-concept animal experiment will become available soon, and will validate this new exciting target for providing neuroprotection in the brain that is under ischaemic insult. The authors thank Kathy Gage for her excellent editorial contributions. Contributors: WY conceived and drafted the review with critical input from HS and HW. Funding: This study was partly supported by American Heart Association grant number 12SDG11950003 and National Institutes of Health (NIH) grant number NS081299. Competing interests: None declared. Provenance and peer review: Not commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

Introduction Leucoaraiosis (LA), also referred to as white matter hyperintensities (WMHs), are usually seen as punctate, patchy or confluent hyperintense areas on T2-weighted or fluid-attenuated inversion recovery (FLAIR) MRI scans in the elderly population. The prevalence of LA is >80% between 60 and 70 years, and approaching 100% between 80 and 90 years.1 LA is often asymptomatic in its early stages, yet its evolution to more advanced stages may lead to substantial neurological dysfunction including dementia, depression, gait disturbance, stroke and death.2 3 Despite its clinical significance, the pathogenic mechanisms underlying LA have not been fully elucidated.

ars.1 LA is often asymptomatic in its early stages, yet its evolution to more advanced stages may lead to substantial neurological dysfunction including dementia, depression, gait disturbance, stroke and death.2 3 Despite its clinical significance, the pathogenic mechanisms underlying LA have not been fully elucidated. The alterations of white matter (WM) architecture in LA are suggested to be continuous from WMHs to its neighbourhood ‘normal-appearing white matter (NAWM)’ on conventional MRI sequences.4 Clinically, the disruption of NAWM integrity reflected by diffusion tensor imaging (DTI) parameters correlated stronger with psychomotor dysfunction than did WMHs load.5 Accumulating evidences advocate the idea that WMHs may represent merely a ‘tip of the iceberg’ of the true extent of pathophysiological changes underlying global WM. In virtue of imaging technological innovation, multimodal imaging modalities have been applied to explore the haemodynamic and microstructural changes of NAWM during the past decades. Undoubtedly, these efforts could deepen our understanding about the intrinsic biological mechanisms underlying the dynamic evolution process of WMHs, which is prerequisite for adopting effective measures to prevent or retard progression of LA in its infant stages. In this review, we thus summarise current knowledge about NAWM of LA mainly acquired from multimodal imaging studies, and attempt to give options for future work.

ing the dynamic evolution process of WMHs, which is prerequisite for adopting effective measures to prevent or retard progression of LA in its infant stages. In this review, we thus summarise current knowledge about NAWM of LA mainly acquired from multimodal imaging studies, and attempt to give options for future work. Haemodynamic disturbance in NAWM and its relationship with progression of LA Chronic hypoperfusion secondary to arteriosclerosis of small perforating arteries has long been proposed as the pathogenesis of LA. Supporting this assumption, haemodynamic studies have shown that perfusion reserve, cerebrovascular reactivity and dynamic autoregulation function were compromised in patients with LA.6 7 Direct evidence of reduced perfusion in WM of LA has been acquired from studies using positron emission tomography (PET), single-photon emission CT (SPECT) and xenon-CT. However, it is uncertain whether hypoperfusion is a primary pathogenic mechanism or simply a secondary response to downregulation of cerebral blood flow (CBF) for reduced metabolic demand in patients with LA. Researchers assumed that hypoperfusion might already exist in NAWM preceding its evolution to visible WMHs on T2-weighted or FLAIR images; in case of WMHs, expansion was primarily due to chronic ischaemia.

econdary response to downregulation of cerebral blood flow (CBF) for reduced metabolic demand in patients with LA. Researchers assumed that hypoperfusion might already exist in NAWM preceding its evolution to visible WMHs on T2-weighted or FLAIR images; in case of WMHs, expansion was primarily due to chronic ischaemia. In 2002, O'Sullivan et al8 first demonstrated that CBF of periventricular NAWM in patients with ischaemic LA was significantly reduced, compared with age-matched controls, using quantitative perfusion MRI. When the analysis was confined to patients with hypertension, both CBF of periventricular and centrum semiovale NAWM were obviously reduced. In addition, they found that CBF was lower in periventricular NAWM than that in centrum semiovale NAWM both in patients and controls, which was consistent with the phenomenon that WMHs tends to emerge earlier in periventricular WM than in centrum semiovale. Subsequent studies further demonstrated that the degree of CBF reduction in NAWM was associated with the severity of LA.9

ar NAWM than that in centrum semiovale NAWM both in patients and controls, which was consistent with the phenomenon that WMHs tends to emerge earlier in periventricular WM than in centrum semiovale. Subsequent studies further demonstrated that the degree of CBF reduction in NAWM was associated with the severity of LA.9 Recently, a prospective study revealed that a low baseline CBF of NAWM was an independent predictor of subsequent development of WMHs in patients with minor stroke/transient ischaemic attack (TIA) after 18-month observation, and the odds of having new WMHs was reduced by 0.61 for each increase of 1 mL/100 g/min with baseline CBF of NAWM.10 Similarly, another contemporary study found that NAWM transformed into new WMH at follow-up had a significantly lower baseline CBF than that of the remaining NAWM.11 Furthermore, mean CBF in NAWM surrounding the index WMHs was lower than that in the total NAWM, extending ∼12 mm from the edge of WMHs. This is consistent with the pattern of LA progression, mainly expanding from the edge of index WMHs other than those emerging in more distal NAWM. Nevertheless, a large sample study, which enrolled 575 patients with manifest arterial diseases, failed to establish the association between global baseline CBF and progression of WMHs after 3.9 years follow-up.12 The main distinction is that this study quantified mean CBF of the whole brain other than CBF of NAWM as used in other two studies. Actually, a recent study confirmed that the association of WMHs severity with intracranial atherosclerotic stenosis (ICAS) was stronger than that with extracranial atherosclerotic stenosis (ECAS).13 The hemodynamic insufficiency in periventricular border zone would occur more easily in ICAS than ECAS. Taking these findings together, it is thus conceivable to infer that progression of WMHs may be mainly driven by regional rather than global hypoperfusion.

an that with extracranial atherosclerotic stenosis (ECAS).13 The hemodynamic insufficiency in periventricular border zone would occur more easily in ICAS than ECAS. Taking these findings together, it is thus conceivable to infer that progression of WMHs may be mainly driven by regional rather than global hypoperfusion. Disruption of microstructures in NAWM Substantial evidences have linked LA with cognitive dysfunction in the elderly, particularly in domains of information processing speed and executive function.2 However, the association between the burden of LA and severity of cognitive impairment was weak. One explanation is that the underlying subtle microstructural changes in NAWM may also contribute to cognitive impairment, which was supported by the findings from diffusion weighted imaging (DWI) and DTI studies.5

nction.2 However, the association between the burden of LA and severity of cognitive impairment was weak. One explanation is that the underlying subtle microstructural changes in NAWM may also contribute to cognitive impairment, which was supported by the findings from diffusion weighted imaging (DWI) and DTI studies.5 DWI findings in NAWM DWI could reflect the freedom of water molecules movement, which is restricted by myelin, axon and other cell structures. Any disruption to these structures may lead to elevated apparent diffusion coefficient (ADC). The ADC values of the whole brain or WMHs were found to be positively correlated with LA extension.14 Combining DWI with MR spectroscopy (MRS), Firbank et al15 found that ADC as well as the metabolite ratios of N-acetylaspartate/creatine and N-acetylaspartate/choline of NAWM was significantly correlated with total WMHs volume. MRS could quantify the levels of N-acetylaspartate (NAA), which would be decreased due to the damage of neurons and their associated axons. On the basis of ADC histogram analysis, the peak height of the ADC histogram of normal-appearing brain tissue (NABT) was independently associated with multiple cognitive domains of patients with LA after adjusting for WMHs volume and brain atrophy.16 A prospective study also demonstrated that baseline ADC metrics of NABT outperformed the same metrics of WMHs lesions in predicting deterioration of cognitive and functional outcomes after 3-year follow-up.17

ciated with multiple cognitive domains of patients with LA after adjusting for WMHs volume and brain atrophy.16 A prospective study also demonstrated that baseline ADC metrics of NABT outperformed the same metrics of WMHs lesions in predicting deterioration of cognitive and functional outcomes after 3-year follow-up.17 DTI findings in NAWM DTI can non-invasively provide quantitative information about disruption of microscopic architecture underlying WM of LA. DTI measures the diffusion in at least six non-collinear directions, providing a three-dimensional representation of water motion. The most commonly used parameters are fractional anisotropy (FA) and mean diffusivity (MD). FA reflects the directionality of diffusion, while MD is a measure of diffusion averaged in all spatial directions. DTI has been extensively used to characterise the subtle pathological changes underlying NAWM, which may be invisible on conventional MRI sequences. In 1999, Derek et al found that FA of periventricular WM was significantly decreased in ischaemic LA compared with normal controls, while its MD was increased.18 However, this small study lacked the spatial resolution to establish whether these changes were confined to WMHs or involved with NAWM. Subsequent studies further demonstrated increased MD and decreased FA in NAWM of patients with LA compared with healthy controls,19 and the changes of FA and MD were quantitatively associated with the severity of WMHs.20

spatial resolution to establish whether these changes were confined to WMHs or involved with NAWM. Subsequent studies further demonstrated increased MD and decreased FA in NAWM of patients with LA compared with healthy controls,19 and the changes of FA and MD were quantitatively associated with the severity of WMHs.20 Maillard et al21 found that FA in different regions of NAWM was inversely correlated with the number and proximity of WMHs in its vicinity, suggesting a ‘penumbra’ surrounding index WMHs (as illustrated in figure 1). The WM integrity of the penumbra may be more subtly disrupted and expected to be more vulnerable than other NAWM. Consistent with this assumption, prospective studies showed that 80% of new WMHs had been emerging as direct extensions of pre-existing lesions other than isolated new lesions,22 and lower baseline FA in NAWM independently predicted greater risk of transition to WMHs.23 However, Maillard et al4 recently demonstrated that the losses of WM integrity over time in the ‘WMHs penumbra’ were similar in magnitude to other NAWM as well as WMHs themselves. This finding suggested that different regions of WM lie along a continuum of injury in patients with LA, and the ‘WMHs penumbra’ appeared to have no distinctive temporal course of deterioration. Whether the conception of the ‘WMHs penumbra’ is of clinical significance has to be tested in future interventional studies.

This finding suggested that different regions of WM lie along a continuum of injury in patients with LA, and the ‘WMHs penumbra’ appeared to have no distinctive temporal course of deterioration. Whether the conception of the ‘WMHs penumbra’ is of clinical significance has to be tested in future interventional studies. Figure 1 (A) Axial T2-weighted fluid-attenuated inversion recovery image; (B) coloured version of image (A): white matter hyperintensities (WMHs; red areas); penumbra surrounding index WMHs in normal-appearing white matter (NAWM; blue areas); other regions of NAWM (green areas). Disruption of the blood–brain barrier (BBB) in NAWM BBB failure and endothelial dysfunction are also proposed to be important pathophysiological mechanisms for WMHs development. Increased permeability of BBB may give rise to leakage of serum components into and through the walls of cerebral small vessels leading to lipohyalinosis, perivascular oedema, demyelination, loss of neurons and gliosis.24

dysfunction are also proposed to be important pathophysiological mechanisms for WMHs development. Increased permeability of BBB may give rise to leakage of serum components into and through the walls of cerebral small vessels leading to lipohyalinosis, perivascular oedema, demyelination, loss of neurons and gliosis.24 In the early dates, the permeability of BBB was mainly assessed on the basis of the cerebrospinal fluid (CSF)/serum ratio for albumin, which was expected to be elevated in proportion to the severity of endothelium damage of cerebral small vessels. Obviously, this method is not suitable for screening or dynamic monitoring of BBB permeability in general populations for its invasive and non-localisation properties. Dynamic contrast-enhanced MRI (DCE-MRI) is a new method being applied to assess subtle BBB disruption in humans, and with considerable reliability. On the basis of DCE-MRI, signal changes consistent with increased BBB permeability were detected in NAWM and WMHs in patients with LA compared with healthy controls.25 26 Recently, a study among patients with Binswanger disease demonstrated that abnormal WM permeability (WMP) was more frequently detected in NAWM than in WMHs, with the highest frequency in a 4 mm width ring surrounding the WMHs, which coincided with the aforementioned finding that the surrounding NAWM was vulnerable, leading to progression of WMHs.27

ts with Binswanger disease demonstrated that abnormal WM permeability (WMP) was more frequently detected in NAWM than in WMHs, with the highest frequency in a 4 mm width ring surrounding the WMHs, which coincided with the aforementioned finding that the surrounding NAWM was vulnerable, leading to progression of WMHs.27 Histopathological and postmortem imaging findings in NAWM Autopsy studies confirmed that the underlying pathological changes of LA extended into the NAWM. Among six AD patients with LA, neuropathological examination revealed much more extensive changes in WM than did parallel postmortem MRI, with a mean of 54% larger abnormal areas.28 The pathological lesions not identified on T2-weighted MRI represented minor changes with a lower intensity of myelin staining. A histopathological study also demonstrated that afferent microvascular density was significantly lower both in deep WMHs lesions and NAWM of patients with LA than in those of healthy subjects.29 These differences were most prominent in patients who died before 60, indicating that global microvascular rarefaction in LA is not completely age-dependent. Interpretation of these results should be done cautiously, as they usually represent a late-stage condition of tissue damage rather than the initial pathological alterations of LA.

were most prominent in patients who died before 60, indicating that global microvascular rarefaction in LA is not completely age-dependent. Interpretation of these results should be done cautiously, as they usually represent a late-stage condition of tissue damage rather than the initial pathological alterations of LA. Current evaluation of LA progression and its limitations Progression of LA has been linked to cognitive worsening,30 increased risk of depression3 and mortality.31 The reported prevalence and speed of LA progression differ greatly in various cohorts due to the different duration of observation and definition of progression. In the Austrian Stroke Prevention Study (ASPS) with healthy volunteers aged 50–75 years, participants with early confluent and confluent WMHs at baseline underwent median increases of 2·7 (IQR 0.5–5.9) cm3 and 9·3(7.1–21.0) cm3, respectively, in lesion volume over 6 years.32 In contrast, the change in WMHs volume was negligible in their counterparts without WMHs or with punctate lesions at baseline. The hospital-based multicentre, multinational Leukoaraiosis and Disability (LADIS) study showed that ∼74% of participants experienced assessable WMHs progression after 3-year follow-up, especially in the frontal and parietal lobes.33

as negligible in their counterparts without WMHs or with punctate lesions at baseline. The hospital-based multicentre, multinational Leukoaraiosis and Disability (LADIS) study showed that ∼74% of participants experienced assessable WMHs progression after 3-year follow-up, especially in the frontal and parietal lobes.33 Hitherto, progression of LA was generally assessed on the basis of changes of WMHs volume or severity grade. Long-term follow-up (usually ≥3-year) was needed to observe such an appreciable change because the evolution process of LA is pretty slow. Up to 43% of participants were reportedly lost to follow-up after 3-year observation.34 Those who completed the follow-up examinations were usually younger and had less vascular risk factors than did the whole study population, leading to an underestimation of LA progression risk. When observed thus, it is particularly not feasible to evaluate the effect of certain therapeutic interventions aiming at retardation of LA progression.

llow-up examinations were usually younger and had less vascular risk factors than did the whole study population, leading to an underestimation of LA progression risk. When observed thus, it is particularly not feasible to evaluate the effect of certain therapeutic interventions aiming at retardation of LA progression. Clinical implications of multimodal imaging studies in NAWM Promisingly, multimodal imaging modalities targeting NAWM could help us to detect and monitor subtle pathophysiological changes in early stages of LA formation. Small studies have shown that CBF could be increased by the use of sympathomimetic midodrine35 or angiotensin receptor blockers,36 as well as fitness exercises.37 However, whether these possible interventions could be applied to prevent LA progression needs further investigation in clinical trials. In the future, the use of non-invasive arterial spin-labelling (ASL) could help to find out whether progression of LA will be accelerated when cerebral perfusion declines to a certain threshold, since it could monitor the dynamic changes of CBF in WM.

event LA progression needs further investigation in clinical trials. In the future, the use of non-invasive arterial spin-labelling (ASL) could help to find out whether progression of LA will be accelerated when cerebral perfusion declines to a certain threshold, since it could monitor the dynamic changes of CBF in WM. Preliminary studies have shown that the changes of FA and MD in NAWM were detectable over a 1 year follow-up.38 In contrast, over this period, there was no change in conventional MRI parameters, including brain volume, WMHs lesion load and lacunar infarct number, as well as changes of cognitive metrics. Recently, another study conducted among community-dwelling people of similar age suggested that MD may provide the best sensitivity in discrimination between WMHs and NAWM even with a mild WMHs load, among other parameters of FA, magnetisation transfer ratio (MTR) and longitudinal relaxation time (T1). The reduction of MTR indicates changes in the underlying composition of tissue, especially the abnormalities of myelin. Moreover, substantial evidence demonstrated that the integrity of microstructure in NAWM, represented by DTI parameters, correlated much closer with cognitive performance than did the load of WMHs, lacunar infarcts or brain volume in patients with LA.5 Thus, use of DTI parameters as outcome variables would be expected to more efficiently evaluate the efficacy of therapeutic interventions designed for delaying LA progression than the simple measurement of WMHs volume or cognitive deterioration in clinical trials. To be mentioned here, there is another emerging technique, known as ‘diffusion kurtosis imaging (DKI)’, which might be even more sensitive than DTI to detect subtle microstructural changes in NAWM of LA in the near future, as indicated in relevant studies of patients with multiple sclerosis (MS).39

in clinical trials. To be mentioned here, there is another emerging technique, known as ‘diffusion kurtosis imaging (DKI)’, which might be even more sensitive than DTI to detect subtle microstructural changes in NAWM of LA in the near future, as indicated in relevant studies of patients with multiple sclerosis (MS).39 A dynamic monitor of BBB permeability may also be used to evaluate the efficacy of medications aiming at endothelium or BBB protection. For example, vitamin B supplementation was found to effectively delay WMHs progression for patients with recent stroke and severe cerebral small vessel disease.40 Currently, the proposed mechanism is that vitamin B may protect the endothelium by reducing concentrations of homocysteine. Future studies using DCE-MRI may help to confirm it by measurement of BBB permeability.

fectively delay WMHs progression for patients with recent stroke and severe cerebral small vessel disease.40 Currently, the proposed mechanism is that vitamin B may protect the endothelium by reducing concentrations of homocysteine. Future studies using DCE-MRI may help to confirm it by measurement of BBB permeability. Conclusions The clinical significance of LA has obtained increasing attention with the arrival of an ageing society. LA develops in an insidious way, while progressing persistently. The so-called ‘NAWM’ has revealed early pathophysiological abnormalities. The pathogenic role of chronic ischaemia in the development of LA was further corroborated with the findings that hypoperfusion exists in NAWM prior to its transition to WMHs. Multimodal imaging modalities could thus be appropriately applied to detect and monitor subtle microstructural and haemodynamic changes in NAWM, in terms of future interventional studies. Moreover, it is reasonable to assume that the pathophysiological changes in NAWM would be more reversible than WMHs itself, since NAWM occupies the main body of global WM and any damage to the integrity of the WM network may exert a negative impact on brain function. Therefore, preventive and therapeutic strategies targeting early changes in NAWM would be expected to be more promising and rewarding. Contributors: GZ conceived and drafted the review with guidance from ML, who made a critical revision of the manuscript.

Conclusions The clinical significance of LA has obtained increasing attention with the arrival of an ageing society. LA develops in an insidious way, while progressing persistently. The so-called ‘NAWM’ has revealed early pathophysiological abnormalities. The pathogenic role of chronic ischaemia in the development of LA was further corroborated with the findings that hypoperfusion exists in NAWM prior to its transition to WMHs. Multimodal imaging modalities could thus be appropriately applied to detect and monitor subtle microstructural and haemodynamic changes in NAWM, in terms of future interventional studies. Moreover, it is reasonable to assume that the pathophysiological changes in NAWM would be more reversible than WMHs itself, since NAWM occupies the main body of global WM and any damage to the integrity of the WM network may exert a negative impact on brain function. Therefore, preventive and therapeutic strategies targeting early changes in NAWM would be expected to be more promising and rewarding. Contributors: GZ conceived and drafted the review with guidance from ML, who made a critical revision of the manuscript. Funding: This work was supported and funded by grants from the Science Technology Department of Zhejiang Province (2013C03043-3) and the National Natural Science Foundation of China (NSFC) (81471170). Competing interests: None declared. Provenance and peer review: Commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

Doctors put drugs of which they know little into bodies of which they know less for diseases of which they know nothing at all. — Voltaire, 250 years ago Nearly 100 years ago, Sir William Osler, the Father of Modern Medicine and a medical educator, had pointed out the deficiencies of medical practice: poor understanding of history, disconnection between science and civilisation, and separation of technological progress from humanity. These three questions still exist today and are intertwined with the modern development and reformation of medicine.1

cal educator, had pointed out the deficiencies of medical practice: poor understanding of history, disconnection between science and civilisation, and separation of technological progress from humanity. These three questions still exist today and are intertwined with the modern development and reformation of medicine.1 History of evidence-based medicine (EBM) EBM is a large system of theory that contains rich connotations. As defined, it is about the medicine that is based on the evidence. Using published literature, conclusion is often drawn based on the quality of the studies and data collected. These conclusions would offer guidance to physicians in clinical practice since we believe that these results have been derived with scientific rigour. EBM offers important clinical significance and scientific value. It emphasises on the natural course of the disease, intervention of illness not often by drugs, modern diagnostic standard, clinical research strictly regulating the conflict of interest, importance of long-term medical treatment and follow-up, and encouragement of the voice of opposition and those with questions. The purpose of creating EBM by our predecessors was mainly to fix the deficiency of clinical and experimental medical models in our daily practice. EBM has overthrown many suppositions, intuitions and hypothesis, and to a great extent changed our clinical practice. The best example is the publication of various clinical guidelines from the findings derived from randomised clinical trials. Here, the key is the quality of those randomised clinical trials. In order to evaluate the validity of the trials, the following criteria are often used: (1) reasonable design; (2) truly randomised; (3) bigger the sample the better; (4) sound statistics; (5) therapeutic parameters congruent to clinical practice; (6) ample follow-up time so that the results would reflect real practice; (7) clinically significant and applicable and (8) end point design, the more ‘harder the data’, the more clinically valuable.

ised; (3) bigger the sample the better; (4) sound statistics; (5) therapeutic parameters congruent to clinical practice; (6) ample follow-up time so that the results would reflect real practice; (7) clinically significant and applicable and (8) end point design, the more ‘harder the data’, the more clinically valuable. The adverse effect of EBM when it is misused EBM has helped us learn new knowledge from published clinical trials that have been well conducted. However, objective ways of grading literature cannot help us subjectively differentiate false data. Recently, EBM has been questioned for its tendency of influencing physicians to practice incorrectly. EBM has become a ‘loaded gun,’ threatening the clinicians with potential penalty such as no-payment if they do not practice according to the so-called ‘best evidence’. While one good use of EBM is to curtail medical waste, the fact is that medical overdiagnosis and overuse are getting worse and EBM may have played a facilitating role. During the 1990s, pharmaceutical industries began to be involved in EBM. Before then, physicians could resist the temptations of ‘drug rep’ because often the drugs from these drug companies lacked clinical evidence. Gradually, the pharmaceutical industry has realised that EBM was not a threat to them but rather an opportunity. If their drug research has been published in a reputable journal, this drug would likely be recommended by the clinical science and as a result, it would bring far more profit to the manufacturer than the effort of marketing by pharmaceutical representatives. This turn of the event has had far reaching effect on the use of EBM. This process is just like a famous quote on statistics: if you query the data, data would provide the answer. Most recently, an online survey conducted by British Medical Journal (BMJ) has revealed that 75% of physicians felt that EBM was collapsing. It has asked people to be aware of those commercials that relentlessly advertise the clinical trial results.

e on statistics: if you query the data, data would provide the answer. Most recently, an online survey conducted by British Medical Journal (BMJ) has revealed that 75% of physicians felt that EBM was collapsing. It has asked people to be aware of those commercials that relentlessly advertise the clinical trial results. Without the backup of the so called ‘evidence,’ drugs would not have had a place in the guidelines. The fact is that many patients in our daily clinical practice do not fit those criteria used during clinical trials; consequently, prescribing these recommended medications could actually be harmful. Thus, directly financing and supporting of clinical trials by big pharmaceutical companies and having those results published, more or less would influence the writing of clinical guidelines. EBM itself now needs to reflect on this seemingly unavoidable influence. Although EBM itself is not a culprit, it can mislead. This negative impact could be far reaching. If this defect of ‘cave of ants’ size on a levee is not fixed in time, miles of levee might crumble soon. It is the fact that overwhelming majority of research is relying on the financial support of pharmaceutical industry and such support is actually needed and not avoidable. Pharmaceutical industry therefore has had unsubtle impact on the medical literature. Intentionally or not, pharmaceutical industry and EBM became intertwined and difficult to differentiate. Consequently, the indications for many drugs became broadened and clinicians felt that seemingly there was a drug for everyone condition.

eutical industry therefore has had unsubtle impact on the medical literature. Intentionally or not, pharmaceutical industry and EBM became intertwined and difficult to differentiate. Consequently, the indications for many drugs became broadened and clinicians felt that seemingly there was a drug for everyone condition. Scholars have reviewed the Medline and Embase and studied several published randomised placebo controlled trials and performed a meta-analysis. They have discovered that more favourable results were found from pharmaceutical industry-supported trials than those financed by non-pharmaceutical industry (device) groups. Even though many of these trials were double blinded and randomised, their results have been quite exaggerated.2 Consequently, the screening and selection process of EBM under this circumstance has been twisted into a sort of ‘healthy’ branding, enhancing overdiagnoses, overuse and become the source of various law suits and symptoms.3

of these trials were double blinded and randomised, their results have been quite exaggerated.2 Consequently, the screening and selection process of EBM under this circumstance has been twisted into a sort of ‘healthy’ branding, enhancing overdiagnoses, overuse and become the source of various law suits and symptoms.3 Then, why is it difficult to share big medical data in China? The main reason is the ‘Mahjong’ mentality and working only for us. Many practice by keeping ‘an eye on your neighbour on the left and the other on the right’. Hence, sharing data with each other is very difficult.4 The essence of scientific research is to explore the unknown world and seek truth. However, many are motivated to do research for different reasons: promotion, raising social status, obtaining fame and money. Only a few truly study the literature. Many physicians obtain their new knowledge through participating in sponsored seminars. They may not realise that even the power point slide deck used for teaching are made by the pharmaceutical industry, not put together by the teaching physician. The lecturers are often the spokespersons for the pharmaceutical industry. On the other hand, the negative results of a trial would have no commercial benefit; hence, the negative findings are hardly mentioned. That is why publishing negative results of clinical trials is as important as the positive ones.

g physician. The lecturers are often the spokespersons for the pharmaceutical industry. On the other hand, the negative results of a trial would have no commercial benefit; hence, the negative findings are hardly mentioned. That is why publishing negative results of clinical trials is as important as the positive ones. The impact of medical journals on EBM David Sackett, the ‘Father of EBM,’ once said “Half of what you'll learn in medical school will be either dead wrong or out of date within five years of your graduation; the trouble is that nobody can tell you which half…”. Research by the editors has shown that many medical literature are misleading and containing false-positive findings. No matter what kind of journal it is, close to 90% of papers need not to be published. Of 38 million already published papers, those that have been referenced for more than 200 times are <0.5%. Those that are considered classic writings and quoted >1000 times are rare. Nearly half have never been cited. Sceptics may say that the result of this analysis could only apply to the ordinary journals, not the leading ones. However, it was jaw-dropping when the elite journals were examined. The story of the 31 years old German scientist Jan Hendrik Schoen was infamous. He joined the Bell research laboratory and published 17 papers in Nature and Science within 4 years. He was considered a potential future Nobel Prize recipient. However, a scientific reporter was suspicious of his conduct and followed him for 4 years. The reporter finally provided the proof that all of his 17 papers were fabricated.5 People began to wonder why the leading journals publish papers based on fake data even though these journals have peer reviewers worldwide. Four reasons may possibly explain this phenomenon: small research project, marginal therapeutic effect, complex relationship between the researcher and funding agency, and hotly competed research topics. One such example is a large-scale population screening trial for a condition that discount the individual difference, and once completed, its results are recommended to the general population at risk.

therapeutic effect, complex relationship between the researcher and funding agency, and hotly competed research topics. One such example is a large-scale population screening trial for a condition that discount the individual difference, and once completed, its results are recommended to the general population at risk. EBM and the Era of Big Data The mission of a doctor is to heal the sick and save lives. The reality is that life itself is a sexually transmitted ‘disease’ that carries 100% mortality. On the other hand, it is true that many great medical discoveries came from close collaboration between the academia and manufacturers. However, it seems that the mission of the pharmaceutical manufacturer is financially driven and suspicious of intentionally labelling healthy people sick. In addition, it is difficult to find a specialist who has no tie with the biomedical or pharmaceutical industry. A recent American survey of 50 medical schools that own hospitals has revealed that every researcher has received on average $33 417 capital assistance from an industry. US Food and Drug Administration also heavily relies on the funding from biomedical industry and in 2010, such funding has reached a total of $1.25 billion, nearly 46% of its drug research budget.1 To great extent, billions of sales profit has created trouble in clinical research. Scientific research has been ‘stained by the stinkiness of money’, filled with false information, wrong diagnosis and confused standards. Therefore, it has generated untrustworthy survey, statistically significant but clinically meaningless research results, misleading data, and hard to differentiate false data but somehow all have been easily published.6 7 One example is the recent announcement by a major pharmaceutical company that manufactures a drug for Alzheimer's disease. The company announced that it will continue on its clinical trials by eliminating patient's daily functional ability as an outcome measure. Although the reason behind this change was that patient with early Alzheimer's disease would have mild functional impairment and 18 months of trial could not detect it, one would suspect that such change could possibly affect the fate of this new drug, which has been closely watched by the medical profession.8

measure. Although the reason behind this change was that patient with early Alzheimer's disease would have mild functional impairment and 18 months of trial could not detect it, one would suspect that such change could possibly affect the fate of this new drug, which has been closely watched by the medical profession.8 Another example that can illustrate this phenomenon was the Woman's Health Study, which enrolled 160 000 postmenopausal healthy women. The trial has found that women on hormonal replacement therapy had more breast cancer, heart disease, stroke and thrombosis than those in the control group. The price paid to treat these complications was far higher than the benefit of preventing colon cancer and hip fracture. This project was stopped 6 years into its originally planned 15 years of research. Another shocking example was the use of β-blocker in the Guidelines of European Heart Association. This recommendation was based on the fabricated research results from Holland researchers, which was possibly associated with nearly 800 000 deaths. These researchers initially found that perioperative use of β-blockers was cardioprotective in two clinical trials. This finding was incorporated into the European 2009 guideline. Nevertheless, a recent meta-analysis of 11 clinical trials of the same topic has shown that patients treated with perioperative β-blockers had 27% more deaths than those in the control group, which was equivalent to 800 000 deaths in Europe.

inical trials. This finding was incorporated into the European 2009 guideline. Nevertheless, a recent meta-analysis of 11 clinical trials of the same topic has shown that patients treated with perioperative β-blockers had 27% more deaths than those in the control group, which was equivalent to 800 000 deaths in Europe. It is undeniable that medical profession is influenced by potentially biased information from some publicised scientific publication. Although mandated by the top medical journals that all clinical trials should be registered online first, a recent Journal of the American Medical Association (JAMA) article has found inconsistency of information registered on clinicaltrials.gov and its final publication.9 The researchers selected 96 clinical trials published in 19 journals with high impact factor (>10). They have found that 70 trials were sponsored by the industries. The most popular areas of research interest included cardiovascular disease, diabetes and hyperlipidaemia (23%), cancer (21%) and infectious disease (20%). The results of these trials were published in the New England Journal of Medicine (NEJM) (24%), Lancet (19%) and JAMA (12%). Nearly 93–100% of these clinical trials published the information of cohort analysis, interventions and outcome. However, 93 of 96 trials contained at least one outcome that did not match the information registered. The inconsistency between the cohorts and intervention was about 2–23%. Ninety-one clinical trials generated 156 positive outcomes. Among them 132 (85%) described the findings in clinicaltrials.gov and journals, while only 14 were published in the website and only 10 in the journals.

t match the information registered. The inconsistency between the cohorts and intervention was about 2–23%. Ninety-one clinical trials generated 156 positive outcomes. Among them 132 (85%) described the findings in clinicaltrials.gov and journals, while only 14 were published in the website and only 10 in the journals. We know well that the essence of EBM is to combine the best external evidence, physician's personal experience and patient's wishes together. All three are needed to help a physician make the most appropriate clinical decision when treating an individual patient. Randomised and controlled studies and meta-analysis are not equivalent to the EBM. They are the reflection of external evidence. When the external evidence is lacking, the experience of a treating physician becomes very important. In this era of big data, biomedical science will have a major role in the world and the use of internet can support transparency and honesty. The explosion of large data has made the traditional research methodology obsolete. Randomisation of samples could be replaced with a complete data set. Statistics has been in use for over 100 years and perhaps one day it will be outdated. The best statistical methodology is probably the exhaustive attack method, which is to have the entire data points at once: samples=entirety. In this era of big data, we could have digitised human body and data on an individual but not a cohort. Everyone can be defined at the individual level as a single entity. The force that has the impact on this change is the internet.

tive attack method, which is to have the entire data points at once: samples=entirety. In this era of big data, we could have digitised human body and data on an individual but not a cohort. Everyone can be defined at the individual level as a single entity. The force that has the impact on this change is the internet. For the physicians today, the technology is ready but new concepts and ways of thinking are still lacking. The physicians of the future will not play the role of knowledge storage but knowledge administrator. They should interact with the patients better, provide compassionate care, consult the patient, assist in decision-making process and be a partner with the intelligent patients. By focusing on solving clinical problems, they will apply the knowledge learnt from a complete set of data to their daily clinical practice and serve the patients even better. Competing interests: None declared. Provenance and peer review: Commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

Introduction Carotid artery angioplasty and stenting (CAS) is a popular alternative to carotid endarterectomy to treat patients with carotid diseases. Hyperperfusion syndrome (HPS) is a rare but potentially a life-threatening complication after CAS with an overall incidence of 0.5–6.8%.1–3 The presence of internal carotid artery (ICA) stenosis ≥90% is a main risk factor for the development of HPS.4–7 Other important risk factors include severe contralateral ICA disease, poor collateral flow, hypertension, and recent stroke or ischaemia.8 The higher the number of risk factors present, the higher is the incidence of HPS, which ranges from 14.1% to 56%.8–13 How to prevent HPS remains controversial and there are no proven effective alternatives other than strict blood pressure (BP) control.1 14 However, it has been recently proposed that staged CAS by allowing gradual restoration of cerebral blood flow (CBF) would minimise HPS occurrence.14 We therefore have conducted a retrospective review of patients who were treated with the two-stage CAS at our centre and examined their outcome.

pressure (BP) control.1 14 However, it has been recently proposed that staged CAS by allowing gradual restoration of cerebral blood flow (CBF) would minimise HPS occurrence.14 We therefore have conducted a retrospective review of patients who were treated with the two-stage CAS at our centre and examined their outcome. Materials and methods Between October 2011 and June 2015, 908 patients with high-grade ICA stenosis had CAS. In the literature, the following criteria were used to select patients for the two-stage CAS:15 44 patients were identified to be eligible and included in this analysis. Their baseline medical condition and the presenting neurological events of each patient were recorded. Preoperative workup included brain MRI, helical CT scan, transcranial Doppler (TCD) scan of the cervical ICA and ipsilateral middle cerebral artery (MCA), digital subtraction angiography (DSA) to assess the collateral circulation status using the five-point scale proposed by ASITN/SIR,16 and CT perfusion (CTP) to assess CBF. The profile of each case has been summarised in table 1. All patients had signed the informed consent before the treatment and were included in this study. Table 1 Baseline characteristics of the 44 patients

Materials and methods Between October 2011 and June 2015, 908 patients with high-grade ICA stenosis had CAS. In the literature, the following criteria were used to select patients for the two-stage CAS:15 44 patients were identified to be eligible and included in this analysis. Their baseline medical condition and the presenting neurological events of each patient were recorded. Preoperative workup included brain MRI, helical CT scan, transcranial Doppler (TCD) scan of the cervical ICA and ipsilateral middle cerebral artery (MCA), digital subtraction angiography (DSA) to assess the collateral circulation status using the five-point scale proposed by ASITN/SIR,16 and CT perfusion (CTP) to assess CBF. The profile of each case has been summarised in table 1. All patients had signed the informed consent before the treatment and were included in this study. Table 1 Baseline characteristics of the 44 patients Age-years Median 67.4 IQR 49–79 Male sex, n (%) 34 (77.2) Stenosis ≥90%, n (%) 27 (61.3) Contralateral stenosis, n (%) ≥50 7 (15.9) ≥60 7 (15.9) ≥70 3 (7) ≥80 5 (11.4) ≥90 0 Occlusion, n (%) 2 (5) CTP Hypoperfusion, n (%) 21 (47.7) Normal, n (%) 23 (52.3) Clinical presentation Amaurosis fugax, n (%) 2 (5) Transient ischaemic attack, n (%) 23 (52.3) Ischaemic stroke, n (%) 9 (20.5) Risk factors Hypertension, n (%) 21 (47.7) Diabetes mellitus, n (%) 14 (31.8) Coronary artery diseases, n (%) 5 (11.4) Smoking, n (%) 6 (13.6) Alcohol use, n (%) 9 (20.4) Circle of Wills Normal 27 (61.4) Variation 17 (38.6) CTP, CT perfusion.

Transient ischaemic attack, n (%) 23 (52.3) Ischaemic stroke, n (%) 9 (20.5) Risk factors Hypertension, n (%) 21 (47.7) Diabetes mellitus, n (%) 14 (31.8) Coronary artery diseases, n (%) 5 (11.4) Smoking, n (%) 6 (13.6) Alcohol use, n (%) 9 (20.4) Circle of Wills Normal 27 (61.4) Variation 17 (38.6) CTP, CT perfusion. We developed the following inclusion criteria for our series: (1) patient with symptomatic carotid stenosis (defined as stroke/transient ischaemic attack of the patient within 60 days who is taking antithrombotic drugs or under other interventions on vascular risk factors); (2) DSA showed carotid stenosis ≥90%, or near occlusion according to NASCET method; (3) poor collateral flow on DSA with a grade of ≤2 base on ASITN/SIR; and (4) cerebral hypoperfusion in the vascular territory of the culprit vessel on cerebral CTP imaging before procedure. Using TCD to assess for HPS Relative CBF changes were assessed by using TCD recordings of the flow velocity of the MCA 1 day before the procedure. During this period, the patient's systolic BP was maintained at >140 mm Hg. HPS is defined as an increase in CBF of >120% over baseline on TCD. Three patients had no temporal window. Finally 41 patients completed TCD scan.

were assessed by using TCD recordings of the flow velocity of the MCA 1 day before the procedure. During this period, the patient's systolic BP was maintained at >140 mm Hg. HPS is defined as an increase in CBF of >120% over baseline on TCD. Three patients had no temporal window. Finally 41 patients completed TCD scan. Cerebral CTP All patients received CTP study 1 day presatge and poststage 1 procedure and 1 day after stage 2 procedure. Asymmetry index (%) was calculated based on the measurement of blood flow between the two cortical hemispheres by taking the ratio of CBF of the affected to unaffected hemisphere excluding any ischaemic/infarcted areas (rCBF in the affected MCA territory/rCBF in the mirror position)×100. The outcomes are shown in table 2. Imaging information before procedure from a representative case is shown in figure 1. Table 2 The CTP parameters of patients with carotid artery stenosis regarding as high risk of hyperperfusion pre-pro procedure (n=41) Preoperation After stage 1 After stage 2 P1 P2 CBF 0.85±0.05 0.98±0.06 1.01±0.07 <0.001 <0.001 CBV 0.98±0.02 0.993±0.022 0.990±0.021 0.70 0.85 MTT 1.15±0.05 1.05±0.05 1.01±0.05 <0.001 <0.001 TTP 1.20±0.06 1.04±0.06 0.99±0.06 <0.001 <0.001 Values are in mean (SD). Numbers in parentheses are the median and range. CBF, cerebral blood flow; CBV, cerebral blood volume; CTP, CT perfusion; MTT, mean transit time; P1, preoperation compared with stage 1; P2, stage 1 compared with stage 2; TTP, time to peak.

Preoperation After stage 1 After stage 2 P1 P2 CBF 0.85±0.05 0.98±0.06 1.01±0.07 <0.001 <0.001 CBV 0.98±0.02 0.993±0.022 0.990±0.021 0.70 0.85 MTT 1.15±0.05 1.05±0.05 1.01±0.05 <0.001 <0.001 TTP 1.20±0.06 1.04±0.06 0.99±0.06 <0.001 <0.001 Values are in mean (SD). Numbers in parentheses are the median and range. CBF, cerebral blood flow; CBV, cerebral blood volume; CTP, CT perfusion; MTT, mean transit time; P1, preoperation compared with stage 1; P2, stage 1 compared with stage 2; TTP, time to peak. Figure 1 Images obtained of a 56-year-old man who manifested with transient right hemiparesis due to cerebral hypoperfusion caused by near occlusion carotid artery. (A) CT angiography showing near occlusion at the ICA origin; (B) right carotid angiography before angioplasty showing near occlusion at the ICA origin; (C) CTP scan in the resting state showing that CBF is severely decreased, CBV, MTT and TTP in the right cerebral hemisphere are severely increased. CTP, CT perfusion; CBF, cerebral blood flow; CBV, cerebral blood volume; ICA, internal carotid artery; MTT, mean transit time; TTP, time to peak. Perioperative antiplatelet therapy In order to minimise the risk of thromboembolic complications, regardless if a patient had asymptomatic or symptomatic carotid stenosis, all were given aspirin (100 mg/day) and clopidogrel (75 mg/day) at least 2 and 5 days prior to the procedure, respectively. Antiplatelet therapy was continued for 3 months after the stenting, and then all were on aspirin (100 mg/day) for life.

s, regardless if a patient had asymptomatic or symptomatic carotid stenosis, all were given aspirin (100 mg/day) and clopidogrel (75 mg/day) at least 2 and 5 days prior to the procedure, respectively. Antiplatelet therapy was continued for 3 months after the stenting, and then all were on aspirin (100 mg/day) for life. Stage 1: conventional balloon angioplasty Semicompliant balloon (MONORAIL; Boston Scientific) was placed across the stenosis and inflated to a nominal pressure for 20 s (the diameter at full dilation was usually 2–3 mm). When the minimum luminal diameter exceeded 2.0 mm on angiography, the procedure was considered successful and completed. All patients had tolerated the procedure well. Imaging analysis during the procedure of stage 1 for the representative case is shown in figure 2. Figure 2 Images obtained in a 56-year-old man who manifested with transient right hemiparesis due to cerebral hypoperfusion caused by near occlusion of the carotid artery. (A) Fluoroscopy during angioplasty showing that the target ICA stenosis was dilated by the balloon (diameter 2 mm, 8 atm for 30 s); (B) carotid angiography showing that the stenosis improved from >99% to <70% after angioplasty. (C) CTP scans in the resting state showing that CBF, CBV, MTT and TTP in the right cerebral hemisphere are significantly improved, after angioplasty. CTP, CT perfusion; CBF, cerebral blood flow; CBV, cerebral blood volume; ICA, internal carotid artery; MTT, mean transit time; TTP, time to peak.

<70% after angioplasty. (C) CTP scans in the resting state showing that CBF, CBV, MTT and TTP in the right cerebral hemisphere are significantly improved, after angioplasty. CTP, CT perfusion; CBF, cerebral blood flow; CBV, cerebral blood volume; ICA, internal carotid artery; MTT, mean transit time; TTP, time to peak. Stage 2: CAS In this procedure, performed about 1 month after stage 1, an embolic protection device (Accunet, Abbott Laboratories) was used. While the distal ICA was protected with this system, the stenosis was predilated with a semicompliant balloon (Amiia or SAVVY; Cordis), followed by the deployment of a self-expanding stent (Acculink, Abbott Laboratories). When necessary, additional postdilation was performed in the same manner. A representative case is shown in figure 3. Information on imaging after stage 2 is shown in figure 3. Figure 3 Images obtained from a 56-year-old man who manifested with transient right hemiparesis due to cerebral hypoperfusion caused by near occlusion of carotid artery. (A) Carotid angiography performed just before the stage 2 procedure shows remaining stenosis of the ICA, but there is no delayed filling of the distal part of the artery. (B) Carotid angiography performed after the procedure shows that the ICA stenosis is completely dilated after balloon angioplasty and stent placement. (C) After stage 2 angioplasty shows further CBF improvement in the right cerebral hemisphere, but not hyperperfusion. CBF, cerebral blood flow; ICA, internal carotid artery.

Carotid angiography performed after the procedure shows that the ICA stenosis is completely dilated after balloon angioplasty and stent placement. (C) After stage 2 angioplasty shows further CBF improvement in the right cerebral hemisphere, but not hyperperfusion. CBF, cerebral blood flow; ICA, internal carotid artery. Postoperative BP management All patients treated with regular CAS or SAP underwent continuous BP monitoring for at least 72 hours after the procedure. The goal was to control BP at a level lower than their preprocedural baseline value and maintain it for at least 1 month. All patients' antihypertensive treatment was followed up by telephone once a week once discharged home. There was no difference in the use of BP control protocol between stages 1 and 2. Definition of HP phenomenon and HPS Hyperperfusion (HP) phenomenon was defined as TCD detected increase in CBF of >100% over baseline or an asymmetry index of more than 120% in CBF compared with the normal side on CTP. HPS was used to designate the development of clinical symptoms as a result of rapidly increasing CBF in excess of that required to meet metabolic demands.2 After the operation, all patients were observed for any new neurological symptoms. Occurrence of HPS was suspected by one of the following clinical findings: throbbing frontotemporal or periorbital headache, confusion, macular oedema, visual disturbances, seizures or focal neurological deficits. If a patient had one of these symptoms or signs, cranial CT, CT angiography and cerebral perfusion CT would be performed immediately.

cted by one of the following clinical findings: throbbing frontotemporal or periorbital headache, confusion, macular oedema, visual disturbances, seizures or focal neurological deficits. If a patient had one of these symptoms or signs, cranial CT, CT angiography and cerebral perfusion CT would be performed immediately. Statistical analysis Statistical analysis was performed by using SPSS, V.21. All results were presented as means±SD. The significance of differences was determined by paired samples t-test. A p value <0.05 was considered significant. Results Between October 2011 and June 2015, 908 patients with high-grade ICA stenosis had CAS. Among them 44 patients qualified for this analysis. Demographics, risk factors and type of carotid stenosis are listed in table 1 There was no morbidity or mortality in all 44 patients who were treated by staged carotid artery stenting. Three of the 44 patients required direct stent placement because of dissection after the stage 1 treatment. One patient experienced HPS without morbidity, the second patient had HP phenomenon, the third patient was normal. After the stage 1 treatment, the hypoperfusion condition of the culprit vascular territory in the other 41 patients resolved significantly as seen on CTP and TCD, and the clinical symptoms of these patients also resolved after stage 2 treatment. One patient had a postoperative increase of MCA flow velocity of 120% after stage 2, but without clinical symptoms.

usion condition of the culprit vascular territory in the other 41 patients resolved significantly as seen on CTP and TCD, and the clinical symptoms of these patients also resolved after stage 2 treatment. One patient had a postoperative increase of MCA flow velocity of 120% after stage 2, but without clinical symptoms. Balloon angioplasty and stent placement increased the CBF and decrease the MTT and TTP in patients with high-grade ICA stenosis or near occlusion As shown in table 2, compared with preprocedure measurement, balloon angioplasty treatment produced a significant difference in CBF (t=27.35, p<0.05), mean transit time (MTT; t=19.16, p<0.05) and time to peak (TTP; t=25.63, p<0.05) on CTP. Meanwhile, there were also significant changes between the stages 1 and 2 of CBF (t=6.14, p<0.05), MTT (t=10.19, p<0.05) and TTP (t=6.84, p<0.05) on CTP. CBV of balloon angioplasty and stent placement were similar to pre-procedure on CTP As shown in table 2, Paired-Samples t-test showed that balloon angioplasty (t=0.70, p>0.05) and stent placement (t=0.854, p<0.05) did not produce any significant change in cerebral blood volume (CBV) on CTP after the procedure. Balloon angioplasty and stent placement increased the CBF on TCD after the procedure The effect of balloon angioplasty and CAS on CBF changes were assessed by TCD 1 day preprocedure and postprocedure (table 3) paired samples t-test showed that balloon angioplasty produced significant changes in CBF after stage 1 (t=31.06, p<0.05), and there was also significant difference of CBF between the stages 1 and 2 (t=13.44, p<0.05).

angioplasty and CAS on CBF changes were assessed by TCD 1 day preprocedure and postprocedure (table 3) paired samples t-test showed that balloon angioplasty produced significant changes in CBF after stage 1 (t=31.06, p<0.05), and there was also significant difference of CBF between the stages 1 and 2 (t=13.44, p<0.05). Table 3 The TCD of patients with carotid artery stenosis regarded as having high risk of hyperperfusion preprocedure (n=41) Stage 1 Stage 2 TCD 44.44±6.43 66.41±7.41 66.41±7.41 93.78±18.81 p Value <0.001 <0.001 TCD, transcranial Doppler. Discussion Two-stage angioplasty was performed in patients with severely impaired CBF due to ICA stenosis who were judged to be at high risk of HPS. Conventional angioplasty without stent placement (stage 1) was successful in 41 patients in our institute. There was no evidence of HP phenomenon on postoperative TCD, and none of the patients manifested clinical symptoms attributable to the procedure. Only one patient had HP on postprocedural TCD during stage 2 procedure, but no clinical worsening or HP on CTP. These results suggest that HPS could be avoided in patients subjected to the two-stage carotid artery stenting in our consecutive cases, which was confirmatory to the finding by Yoshimura et al.12

procedure. Only one patient had HP on postprocedural TCD during stage 2 procedure, but no clinical worsening or HP on CTP. These results suggest that HPS could be avoided in patients subjected to the two-stage carotid artery stenting in our consecutive cases, which was confirmatory to the finding by Yoshimura et al.12 The use of small-size (usually 2 mm in diameter) balloon catheters seemed helpful and safer during the stage 1 procedure. In our cases, a smaller balloon (2 mm) was enough to open the residual lumen and reduce cerebral hypoperfusion, which was seen on TCD and CTP. The symptoms of all patients were resolved after stage 1 procedure. It was unclear how long the residual lumen stayed open. If the lesion was dilated by using a larger size balloon catheter, a wall dissection could occur in the carotid artery. More studies are needed in order to determine the appropriate balloon size to safely open up the residual lumen.

solved after stage 1 procedure. It was unclear how long the residual lumen stayed open. If the lesion was dilated by using a larger size balloon catheter, a wall dissection could occur in the carotid artery. More studies are needed in order to determine the appropriate balloon size to safely open up the residual lumen. High BP is known to be a significant risk factor for HPS after carotid intervention.5 12 Elevation of BP may increase cerebral perfusion pressure and predispose to ICH during reperfusion after carotid recanalisation. Postoperative BPs should be maintained to normal or slightly lower than normal values.17 18 However, HPS may also occur in normotensive patients or in participants with systolic pressure of <160 mm Hg, reflecting the role of impaired autoregulatory mechanisms.19–21 In our cases, the systolic BP in all of the patients was kept slightly lower than the baseline, but still there were three patients who experienced HP phenomenon on TCD, even after the systolic pressure was controlled to 20% lower than the baseline.

mm Hg, reflecting the role of impaired autoregulatory mechanisms.19–21 In our cases, the systolic BP in all of the patients was kept slightly lower than the baseline, but still there were three patients who experienced HP phenomenon on TCD, even after the systolic pressure was controlled to 20% lower than the baseline. The pathophysiology of the hyperperfusion is multifactorial, while cerebral haemodynamics and cerebral autoregulation, as previously mentioned, are individualised in each patient. The most important preoperative risk factor of developing HPS is the high-grade ipsilateral carotid stenosis plus poor collateral flow. Patients with critical ICA stenosis often present with low cerebrovascular reactivity, which consists of maximal vasodilatation of cerebral arterioles in order to maintain sufficient cerebral blood perfusion.22 After CAS, the expanded lumen size increased CBF and led to hypertension. In contrast, autoregulation mechanisms were diminished and thus caused HP in the previously hypoperfused tissue. One critical factor is that the degree of ICA stenosis was ≥90%.4–7 In those who developed ICH, the mean ICA stenosis was 95%. In this group, the incidence of ICH was 3.8%. Lacking of interhemispheric collateral blood supply via the circle of Willis may be a marker for greater risk of developing reperfusion syndrome.3 Chang et al24 reported that prestenting CBV index >0.15 and TTP index >0.22with CTP were two independent parameters that might be associated with HPS after CAS.23 Yoshimura et al12 reported that 67% of patients who underwent regular one-stage CAS had HPS after stenting, but in the two-staged angioplasty group, none of the patients showed HPS. Based on this report, we also developed the criteria for the staged CAS for our patients in order to avoid HPS. For a more accurate haemodynamic assessment, the same protocol was used for CBF measurement with TCD and CTP in all patients before and after the treatment in this study. Stage 1 procedure helped to re-establish blood flow, which might restore the impaired cerebral microvascular autoregulation during the first month. During this period, the recovered cerebral vasoreactivity could control the constriction of small cerebral artery, and dampen the pressure from blood flow during the stage 2 procedure. That was likely the explanation why there was no patient who experienced HP symptoms and signs of CHS.

gulation during the first month. During this period, the recovered cerebral vasoreactivity could control the constriction of small cerebral artery, and dampen the pressure from blood flow during the stage 2 procedure. That was likely the explanation why there was no patient who experienced HP symptoms and signs of CHS. Although this treatment was successful in the current series, there were some limitations. First, the selection of the patients could be biased since the current study only focused on patients with severely impaired CBF and CVR. Second, other factors are also known to be risk factors for HPS after CAS: severe ipsilateral stenosis, isolated hemisphere with poorly developed communicating arteries, hypertension, and use of antiplatelet or anticoagulant therapy. Third, we did not introduce CVR in this study, which might impact the power of result of this study. Instead, CTP was commonly used rather than xenon-CT, single photon emissionCT or positron emission tomography. Nevertheless, when CBF decreased, combining the findings of MTT prolongation and decreased CBV could help confirm the dilation of the cerebral blood vessels; therefore, indicating that cerebral autoregulation has been impaired.24

monly used rather than xenon-CT, single photon emissionCT or positron emission tomography. Nevertheless, when CBF decreased, combining the findings of MTT prolongation and decreased CBV could help confirm the dilation of the cerebral blood vessels; therefore, indicating that cerebral autoregulation has been impaired.24 Conclusion In patients with impaired CBF from high-grade ipsilateral ICA stenosis, two-stage CAS treatment of the stenosis could reduce the risk of developing HPS. More studies are still needed to confirm the benefit. Nevertheless, it is relatively a simple operation with the two-stage CAS in patients with high-grade ICA stenosis and the risk of developing procedure-related complication is low. Competing interests: None declared. Patient consent: Obtained. Provenance and peer review: Commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

Introduction Stroke is the second most common cause of death and the third most common cause of disability worldwide.1 2 Carotid atherosclerotic plaque is identified as one of the main sources of ischaemic stroke. Fibrous cap rupture of atherosclerotic plaque is believed to be the key event that leads to thrombus formation and clinical events. Many recent studies have focused on the identification of features of ‘high-risk’ plaques that pose increased risk of rupture, especially using non-invasive in vivo imaging techniques. Histologically, major determinants of plaques prone to rupture include plaque compositional features, such as intraplaque haemorrhage (IPH) and large lipid-rich necrotic core (LRNC).3 Carotid atherosclerotic plaque MRI (CMRI) is capable of accurately characterising plaque morphology and composition and is validated by histology.4–6 Many recent studies have focused on the use of CMRI to evaluate carotid plaque features that predict future stroke or transient ischaemic attack (TIA) in prospective settings.7–9 Recently, two systematic reviews and meta-analyses summarised the key plaque features on CMRI that are associated with cerebrovascular events.10 11

ecent studies have focused on the use of CMRI to evaluate carotid plaque features that predict future stroke or transient ischaemic attack (TIA) in prospective settings.7–9 Recently, two systematic reviews and meta-analyses summarised the key plaque features on CMRI that are associated with cerebrovascular events.10 11 Recent findings about high-risk plaques suggest the need to include vessel wall imaging as part of carotid artery examination. The current clinical diagnosis of carotid atherosclerotic plaque, however, is based on the measurement of luminal stenosis by ultrasound and/or other angiographic imaging modalities.12 13 It is well established that stenosis measurement to decide carotid plaque severity and treatment options has many limitations.14 Angiographic imaging does not provide information of the pathological changes of the vessel wall and may underestimate the severity of atherosclerotic disease due to positive remodelling.15 For example, investigators found that more than 20% of carotid arteries with lower grade stenosis (<50% stenosis) had high-risk features, such as IPH and fibrous cap rupture,16 17 and these features can also occur in 8–9% of carotid arteries with normal lumen size.18 Recent studies have also suggested that a portion of cryptogenic strokes may actually be caused by carotid atherosclerotic plaque with moderate and/or minor stenosis.19 20

igh-risk features, such as IPH and fibrous cap rupture,16 17 and these features can also occur in 8–9% of carotid arteries with normal lumen size.18 Recent studies have also suggested that a portion of cryptogenic strokes may actually be caused by carotid atherosclerotic plaque with moderate and/or minor stenosis.19 20 CMRI, even though promising, remains an experimental procedure. Researchers in different institutions have used different approaches for CMRI. But in general, the commonly accepted approach is to include ‘multiple-contrast’ image acquisitions to provide comprehensive information of plaque morphology and composition.5 21 22 Despite many studies reporting the use of CMRI in relatively small or single-centre studies, there have been limited reports of using a single standardised imaging protocol in large multicentre studies to evaluate carotid atherosclerosis in a symptomatic population.

ormation of plaque morphology and composition.5 21 22 Despite many studies reporting the use of CMRI in relatively small or single-centre studies, there have been limited reports of using a single standardised imaging protocol in large multicentre studies to evaluate carotid atherosclerosis in a symptomatic population. Stroke is the number one killer in China and there has been much attention directed to identify strokes that are caused by carotid atherosclerotic plaque.23 Findings from these studies, have been inconsistent, however. A study by Jeng et al24 reported that 13% of patients (48/367) with ischaemic stroke had severe carotid stenosis (>50% stenosis or occlusion). In contrast, Liu et al25 showed that severe carotid stenosis could be seen in 41.7% of patients who suffered stroke. More interestingly, a pilot study by Saam et al26 showed that carotid plaque composition differs between Chinese and American Caucasian symptomatic patients. In addition, the prevalence of cerebrovascular atherosclerotic disease may vary among different regions of China. A recent study has shown that the proportion of patients with severe intracranial atherosclerosis (stenosis >50%) was significantly higher in north China than in south China (50.2% vs 41.9%; p<0.0001).27 However, the difference in carotid atherosclerotic plaque characteristics between patients in south and north areas of China is unclear. Thus, an imaging study to examine the carotid atherosclerotic plaque status in patients with recent stroke may provide critical information on the prevalence of carotid plaque and their compositional features and regional differences in China. This information will help to understand the natural status of carotid atherosclerotic plaque in patients who suffered stroke and to identify the best prevention and treatment options.

ay provide critical information on the prevalence of carotid plaque and their compositional features and regional differences in China. This information will help to understand the natural status of carotid atherosclerotic plaque in patients who suffered stroke and to identify the best prevention and treatment options. Methods Carotid Atherosclerosis Risk Assessment (CARE II) study is conceived based on the needs to assess carotid atherosclerotic plaque in Chinese patients who recently suffered stroke and/or TIA using identical, state-of-the-art CMRI technique (NCT02017756). It is a cross-sectional, non-randomised, observational, multicentre study evaluating the carotid atherosclerotic plaque by CMRI in Chinese patients who suffered stroke. It is supported by funding from the Chinese government and by industrial partners. In collaboration with 13 medical centres and hospitals in China equipped with 3 T MRI scanner and with the Vascular Imaging Laboratory (VIL) of the University of Washington, this study intended to consecutively recruit 1000 patients with recent stroke or TIA and atherosclerotic plaque in at least one carotid artery, to undergo CMRI of bilateral carotid arteries and routine brain MRI. Core reading centres situated in the Center for Biomolecular Imaging Research (CBIR, Beijing, China) and VIL (Seattle, Washington, USA) conducted quantitative review of the carotid plaque, as detailed below.

otic plaque in at least one carotid artery, to undergo CMRI of bilateral carotid arteries and routine brain MRI. Core reading centres situated in the Center for Biomolecular Imaging Research (CBIR, Beijing, China) and VIL (Seattle, Washington, USA) conducted quantitative review of the carotid plaque, as detailed below. This study intended to recruit patients in different regions of China using a single standardised CMRI protocol with centralised, blinded review and analysis. IRB approvals were obtained for the entire study and for each participating institution and all study participants provided written informed consent. Study objectives The primary objective of the CARE-II study is to determine the prevalence and characteristics of high-risk features of atherosclerotic plaques in the carotid arteries in Chinese patients with recent ischaemic stroke or TIA. Main secondary objectives are to evaluate: (1) the relationship between carotid plaque features and cerebral infarcts; (2) the differences of carotid plaque patterns among different regions of China, particularly south and north areas; (3) gender specific characteristics of carotid artery atherosclerotic disease in Chinese patients who suffered stroke.

valuate: (1) the relationship between carotid plaque features and cerebral infarcts; (2) the differences of carotid plaque patterns among different regions of China, particularly south and north areas; (3) gender specific characteristics of carotid artery atherosclerotic disease in Chinese patients who suffered stroke. Targeted population This study intends to consecutively recruit 1000 patients with recent stroke or TIA (within 2 weeks after onsets of symptoms) and atherosclerotic plaque in at least one carotid artery determined by B-mode ultrasound scan (intima-media thickness ≥1.5 mm). The age of study population ranges from 18 to 80 years old. The exclusion criteria are as follows: (1) patients with evidence of cardiogenic stroke; (2) patients with haemorrhagic stroke; (3) history of radiation therapy in the neck; (4) claustrophobia; and (5) contraindication to MRI examination.

a thickness ≥1.5 mm). The age of study population ranges from 18 to 80 years old. The exclusion criteria are as follows: (1) patients with evidence of cardiogenic stroke; (2) patients with haemorrhagic stroke; (3) history of radiation therapy in the neck; (4) claustrophobia; and (5) contraindication to MRI examination. Patient recruitment and MRI A flow chart of the study protocol is presented in figure 1. Hospitals and imaging centres must have a stroke unit, defined as a multidisciplinary team which has been designated for the care of patients who suffered stroke, and it must be equipped with a 3 T whole body MRI scanner capable of running the standardised carotid imaging protocol prescribed in the study. Hospitals distributed in the following regions were selected to investigate geographical differences in carotid plaque features of patients: Northeast, North, East, South and West areas (figure 2). Participating radiologists and MRI technologists were trained on image acquisition and quality evaluation by CBIR. Phantom and human volunteer scans were conducted at each participating site and evaluated by the core laboratory in CBIR for protocol adherence and image quality before recruitment of patients was allowed. Figure 1 Flow chart of study protocol. CBIR, Center for Biomolecular Imaging Research; VIL, Vascular Imaging Laboratory. Figure 2 Regional distribution of imaging sites across China.

Patient recruitment and MRI A flow chart of the study protocol is presented in figure 1. Hospitals and imaging centres must have a stroke unit, defined as a multidisciplinary team which has been designated for the care of patients who suffered stroke, and it must be equipped with a 3 T whole body MRI scanner capable of running the standardised carotid imaging protocol prescribed in the study. Hospitals distributed in the following regions were selected to investigate geographical differences in carotid plaque features of patients: Northeast, North, East, South and West areas (figure 2). Participating radiologists and MRI technologists were trained on image acquisition and quality evaluation by CBIR. Phantom and human volunteer scans were conducted at each participating site and evaluated by the core laboratory in CBIR for protocol adherence and image quality before recruitment of patients was allowed. Figure 1 Flow chart of study protocol. CBIR, Center for Biomolecular Imaging Research; VIL, Vascular Imaging Laboratory. Figure 2 Regional distribution of imaging sites across China. All the MRI is performed on a 3.0 T MRI scanner with 8-channel phase array coil. In this multicentre study, an identical high resolution, multicontrast vessel wall imaging protocol is used for carotid plaque imaging. The imaging protocol includes three-dimensional (3D) time-of-flight (TOF), T1-weighted (T1-W) quadruple inversion recovery (QIR),28 T2-weighted (T2-W) multislice double inversion recovery (MDIR),29 and Magnetisation Prepared Gradient Recalled Echo (MP-RAGE) imaging sequences. The localisation of carotid plaque imaging is centred to the bifurcation of index carotid artery. A 3D imaging sequence of Motion sensitised driven Equilibrium prepared Rapid Gradient Echo (MERGE)30 with large longitudinal coverage is also acquired for describing the distribution of atherosclerotic plaques in different segments of extracranial carotid arteries. The imaging parameters are detailed in table 1. In addition, a standard protocol including 3D TOF MR angiography (MRA), T2-fluid-attenuated inversion recovery (FLAIR) and diffusion-weighted image (DWI) sequences was used for brain imaging. Figures 3 and 4 show examples of carotid vessel wall images acquired.

arteries. The imaging parameters are detailed in table 1. In addition, a standard protocol including 3D TOF MR angiography (MRA), T2-fluid-attenuated inversion recovery (FLAIR) and diffusion-weighted image (DWI) sequences was used for brain imaging. Figures 3 and 4 show examples of carotid vessel wall images acquired. Table 1 Imaging parameters of CMRI Standardised multicontrast imaging protocol 3D sequence TOF T1W T2W MP-RAGE MERGE Sequence FFE TSE TSE FFE FFE Black blood None QIR MDIR MSDE Repeat time, ms 20 800 4800 8.8 10 Echo time, ms 4.9 10 50 5.3 4.8 Flip angle 20° 90° 90° 15° 6° Field of view, cm 14×14 14×14 14×14 14×14 25×16×7 Matrix 256×256 256×256 256×256 256×256 356×357 Scan plane axial axial axial axial coronal Slice thickness, mm 1 2 2 1 0.7 FFE, Fast field echo; MDIR, multislice double inversion recovery; MP-RAGE, Magnetisation Prepared Gradient Recalled Echo; MSDE, Motion Sensitised Driven Equilibrium; TOF, time-of-flight; TSE, turbo spin echo; QIR, quadruple inversion recovery. Figure 3 Example of 2D multicontrast carotid vessel wall imaging protocol including TOF, T1W, T2W and MP-RAGE sequences. The lumen (*) and outer wall boundaries are well delineated on vessel wall images. A lipid-rich atherosclerotic lesion can be seen in the left common carotid artery (arrow). JV represents jugular vein. 2D, two-dimensional; MP-RAGE, Magnetisation Prepared Gradient Recalled Echo; TOF, time-of-flight; T1-W and T2-W, T1 and T2-weighted image.

ter wall boundaries are well delineated on vessel wall images. A lipid-rich atherosclerotic lesion can be seen in the left common carotid artery (arrow). JV represents jugular vein. 2D, two-dimensional; MP-RAGE, Magnetisation Prepared Gradient Recalled Echo; TOF, time-of-flight; T1-W and T2-W, T1 and T2-weighted image. Figure 4 Example of 3D MERGE vessel wall imaging with large longitudinal coverage. The middle and the bilateral panels represent original MERGE image and left and right carotid artery images after curved reconstruction, respectively. 3D, three-dimensional; MERGE, Motion sensitised driven Equilibrium prepared Rapid Gradient Echo. Clinical information collection Clinical information is acquired from the medical records prior to carotid MRI for all patients. Demographic characteristics including age, gender, height and weight were recorded. History of hypertension (defined as diastolic blood pressure ≥90 mm Hg or systolic blood pressure ≥140 mm Hg), diabetes, smoking, statin use and cardiovascular disease is collected. The levels of lipoprotein including high-density lipoprotein, low-density lipoprotein, total cholesterol and triglycerides are recorded. In addition, the index artery which is defined as carotid arteries associated with symptoms is determined when such information is available. Image interpretation All MRI acquired from the image sites are transferred to the core laboratories of Center for Biomedical Imaging Research (CBIR) at Tsinghua University, Beijing, China and Vascular Imaging Laboratory (VIL) at University of Washington, Seattle, USA.

Clinical information collection Clinical information is acquired from the medical records prior to carotid MRI for all patients. Demographic characteristics including age, gender, height and weight were recorded. History of hypertension (defined as diastolic blood pressure ≥90 mm Hg or systolic blood pressure ≥140 mm Hg), diabetes, smoking, statin use and cardiovascular disease is collected. The levels of lipoprotein including high-density lipoprotein, low-density lipoprotein, total cholesterol and triglycerides are recorded. In addition, the index artery which is defined as carotid arteries associated with symptoms is determined when such information is available. Image interpretation All MRI acquired from the image sites are transferred to the core laboratories of Center for Biomedical Imaging Research (CBIR) at Tsinghua University, Beijing, China and Vascular Imaging Laboratory (VIL) at University of Washington, Seattle, USA. Carotid artery image review The vessel wall images of bilateral carotid arteries are interpreted by trained reviewers with >3 years' experience in cardiovascular plaque imaging using custom-designed software (CASCADE; University of Washington, Seattle, USA). Each axial image is reviewed by two reviewers blinded to clinical information and brain image findings with consensus. The lumen and wall boundaries are outlined manually to measure the lumen area, wall area, total vessel area and wall thickness (WT) at each axial location. The presence or absence of calcification, LRNC, IPH and fibrous cap rupture is identified using the published criteria.5 31 The size of each plaque's compositional feature is also measured. The carotid 3D MERGE images are analysed at an MR workstation to measure the maximum WT of plaque in each segment of the carotid artery (common carotid artery, bulb, and internal carotid artery (ICA)) when present. Three-dimensional TOF MRA images are reconstructed by maximum intensity projection to measure the luminal stenosis of carotid arteries using North American Symptomatic Carotid Endarterectomy Trial (NASCET) algorithm.32

of the carotid artery (common carotid artery, bulb, and internal carotid artery (ICA)) when present. Three-dimensional TOF MRA images are reconstructed by maximum intensity projection to measure the luminal stenosis of carotid arteries using North American Symptomatic Carotid Endarterectomy Trial (NASCET) algorithm.32 Brain image review The intracranial 3D TOF MRA images are reconstructed by maximum intensity projection. The luminal stenosis is measured at each arterial segment using the following categories: 0%, 1–29%, 30–49%, 50–69%, 70–99% and 100% (occlusion) according to the WASID method.33 The segments of intracranial arteries include intracranial ICA, middle cerebral artery, anterior cerebral artery, posterior cerebral artery and basilar artery. The volume and location of acute cerebral infarcts which are hyperintense on DWI images and white matter lesions (WML) which show hyperintense on FLAIR images are evaluated using custom-designed software. The severity of WML is stratified using a previously published method.34

terior cerebral artery and basilar artery. The volume and location of acute cerebral infarcts which are hyperintense on DWI images and white matter lesions (WML) which show hyperintense on FLAIR images are evaluated using custom-designed software. The severity of WML is stratified using a previously published method.34 Statistical analysis and sample size considerations To address the primary study objective, the prevalence of plaque components and features including calcification, LRNC, IPH and fibrous cap rupture will be estimated from the entire study sample and from important subgroups defined by gender, age and geographic region. Within the same groups, the distributions of carotid vessel morphology, including wall volume, per cent wall volume, maximum WT and maximum per cent wall area and volumes of components (calcification, LRNC and IPH) will be summarised using means and SDs as well as percentiles.

ined by gender, age and geographic region. Within the same groups, the distributions of carotid vessel morphology, including wall volume, per cent wall volume, maximum WT and maximum per cent wall area and volumes of components (calcification, LRNC and IPH) will be summarised using means and SDs as well as percentiles. As part of the secondary objectives, plaque features and measurements will be compared between demographic and regional subgroups using logistic regression (for binary features) and other generalised linear models (for continuous measurements, which may be normally distributed or skewed). Multiple models will be considered, which will adjust for (1) gender and age; (2) gender, age and other traditional risk factors; and (3) gender, age, risk factors and plaque size. Similarly, multivariate logistic regression will be used to evaluate associations between the presence and absence of acute cerebral infarcts, WML or severe intracranial stenosis and carotid plaque features, while generalised linear models will be used to evaluate associations between acute infarct or WML volumes and carotid plaque features.

logistic regression will be used to evaluate associations between the presence and absence of acute cerebral infarcts, WML or severe intracranial stenosis and carotid plaque features, while generalised linear models will be used to evaluate associations between acute infarct or WML volumes and carotid plaque features. Based on the target sample size of 1000 and assuming a 10% loss rate due to missing data or image quality issues (final N=900), precision (95% CIs) for our prevalence estimates will be within ±3.3% in the full sample and ±7.7% in subgroups that are one-fifth the size of the full sample. All power calculations assumed two-sided tests with a significance level of 0.05. Calculations were performed using R (V.3.1.1, R Foundation for Statistical Computing, Vienna, Austria) and GPower (V.3.1.9.2, University of Kiel, Germany).35 Discussion This is one the first multicentre studies targeting patients who have had recent stroke and TIA using a standardised CMRI. It is likely to provide critical information of the prevalence of carotid atherosclerotic plaque and its characteristics beyond luminal narrowing. Furthermore, this study will test the feasibility of using one multicontrast imaging protocol to evaluate carotid atherosclerosis in a multicentre setting and the usefulness of this protocol. Owing to the distribution of imaging centres across China, this study may also provide useful information of the differences in prevalence in carotid atherosclerosis in different regions and plaque compositional features regional dependencies.

therosclerosis in a multicentre setting and the usefulness of this protocol. Owing to the distribution of imaging centres across China, this study may also provide useful information of the differences in prevalence in carotid atherosclerosis in different regions and plaque compositional features regional dependencies. The design incorporates centralised training, quality assurance, image review of both carotid and brain images, adds more confidence of the overall analysis. Limitations This is a cross-sectional study and the patient recruitment scheme does not rule out strokes potentially caused by intracranial atherosclerotic plaques or aortic arch disease. Although efforts are made to exclude patients with cardioembolic stroke, cardiogenic sources cannot be definitively ruled out. For example, not all patients currently undergo long-term monitoring for paroxysmal arrhythmias. Conclusion This is a cross-sectional, multicentre study to investigate the prevalence and characteristics of high-risk atherosclerotic carotid plaque in Chinese patients with stroke and TIA by using high-resolution vessel wall MRI. This trial is sufficiently powered to demonstrate the prevalence of carotid high-risk plaque and to explore regional differences in Chinese patients who suffered stroke. Contributors: XZ and CY conceived this study. RL provided technical supports. DSH conducted statistical analysis. XZ interpreted the data. XZ and DSH drafted this manuscript and CY and TSH made critical revisions. CY supervised this study.

Conclusion This is a cross-sectional, multicentre study to investigate the prevalence and characteristics of high-risk atherosclerotic carotid plaque in Chinese patients with stroke and TIA by using high-resolution vessel wall MRI. This trial is sufficiently powered to demonstrate the prevalence of carotid high-risk plaque and to explore regional differences in Chinese patients who suffered stroke. Contributors: XZ and CY conceived this study. RL provided technical supports. DSH conducted statistical analysis. XZ interpreted the data. XZ and DSH drafted this manuscript and CY and TSH made critical revisions. CY supervised this study. Funding: This study is supported by grants of Natural Science Foundation of China (81271536, 61271132 and 81361120402) and Philips Healthcare. Competing interests: None declared. Patient consent: Obtained. Ethics approval: The Institutional review Board of Tsinghua University School of Medicine. Provenance and peer review: Not commissioned; externally peer reviewed.

Introduction Stroke is a debilitating illness rendering thousands disabled and leading a significant proportion of people to death worldwide. Its incidence in Asia, according to a rough estimate, has been increasing recently. Among Asian countries, Pakistan shares a significant burden of this devastating disease contributing towards an exponential expenditure of resources, finances, community manpower, health services and the economy as a whole.1 2 Traditional and well-known risk factors of stroke like diabetes and hypertension have also been recognised to be increasing in Asian countries. The burden of stroke is following a similar trend. According to recently published literature, Pakistan currently has an enormous proportion of its population suffering either from diabetes or hypertension or both.1 2 Unfortunately, a majority of people are unaware of their comorbid conditions. This primarily is due to a lack of awareness for routine medical checkup, availability of screening services for endemic diseases locally and ignorance on the part of the community regarding personal healthcare. Even those who have been timely diagnosed do not mostly follow the standard practice of a regular follow-up and/or compliance with medications. At the same time, poor updated knowledge of local physicians not associated with tertiary care setups contributes towards the use of obsolete and relatively less effective medical healthcare delivery.

timely diagnosed do not mostly follow the standard practice of a regular follow-up and/or compliance with medications. At the same time, poor updated knowledge of local physicians not associated with tertiary care setups contributes towards the use of obsolete and relatively less effective medical healthcare delivery. Local literature on stroke published from Pakistan is scarce. The current incidence and prevalence of stroke in Pakistan is not exactly known. Several reported case series in literature highlighting significant differences in terms of stroke epidemiology, risk factors and stroke subtypes have been published but full-text original research articles are not more than a handful. Considering a high population suffering from stroke in our country, it would be highly beneficial to collect and interpret local epidemiological data. Its consequences can range from countering physical disability and conserving millions of dollars against the social and economic consequences brought up on by stroke.2–5

onsidering a high population suffering from stroke in our country, it would be highly beneficial to collect and interpret local epidemiological data. Its consequences can range from countering physical disability and conserving millions of dollars against the social and economic consequences brought up on by stroke.2–5 The objective of this review article is to summarise the facts regarding acute ischaemic stroke and its various aspects in a third world country, that is, Pakistan, where resources are limited and healthcare system is underdeveloped. No large-scale epidemiological studies are available to determine the true incidence of stroke in Pakistan. We reviewed the available literature on stroke from Pakistan and through this article we primarily aim to present the current acute ischaemic stroke management in Pakistan in juxtaposition to that of the developed world. We also intend to highlight areas for future development and improvement in management. Literature search PubMed, MEDLINE, Medline Plus, PubMed Central and Pakmedinet search was undertaken using the keywords ‘cerebrovascular accident, acute ischemic stroke, recombinant tissue Plasminogen Activator, window period, feasibility, third world country, Pakistan, stroke management’. Full articles were reviewed; all available relevant statistics were searched and included if mentioned by the authors.

undertaken using the keywords ‘cerebrovascular accident, acute ischemic stroke, recombinant tissue Plasminogen Activator, window period, feasibility, third world country, Pakistan, stroke management’. Full articles were reviewed; all available relevant statistics were searched and included if mentioned by the authors. Epidemiology Twenty per cent of the world's population resides in South Asia. This region has the most dense concentration of cardiovascular diseases. As stroke affects the elderly exceedingly, an increase in the average lifespan has automatically led to an increased number of stroke cases reported.2 5 6 Stroke and transient ischaemic stroke are rampant in Pakistan. The risk of stroke has plumped by 100% or more in the past decade, especially in third world countries where it accounts for more than 85% of all stroke-related mortality. The mean population age of patients with stroke in developing countries like Pakistan is a decade or younger than that of their western counterparts as reported by Ali and colleagues in 2016. This leads to an increased overall lifetime burden of functional disability and consequent economic losses.1 2 5–10

d mortality. The mean population age of patients with stroke in developing countries like Pakistan is a decade or younger than that of their western counterparts as reported by Ali and colleagues in 2016. This leads to an increased overall lifetime burden of functional disability and consequent economic losses.1 2 5–10 Transient ischaemic attack (TIA) defines a subset of patients having a high risk of getting a stroke. These patients may benefit from timely intervention, most likely thrombolysis or endovascular procedures. The immediate risk of stroke is about 10% in the first 90 days after TIA, the initial 48 hours being the most critical. Papers presented from Pakistan report an almost similar statistical result. Early intervention after TIA has shown an 80% relative risk reduction in the emergence of stroke in western cohorts. While this revolutionised treatment strategy in the field of neurology has now been practised for over a decade in the western world, it is still in a stage of infancy in Pakistan with a juvenile infrastructure. Only a few centres in Pakistan offer thrombolysis. That too is offered to a very select number of cases and usually to that portion of the population who can afford its cost. This is because meeting the window period deadline is a major problem and expenses for thrombolysis are solely borne by the patient and their family. Virtually no endovascular surgical facilities are yet available.11–15

a very select number of cases and usually to that portion of the population who can afford its cost. This is because meeting the window period deadline is a major problem and expenses for thrombolysis are solely borne by the patient and their family. Virtually no endovascular surgical facilities are yet available.11–15 Risk stratification Risk factors for stroke range from common ailments like diabetes mellitus, hypertension, hyperlipidaemia, atrial fibrillation, aneurysms, arterio venous malformation (AVM) and smoking to factors observed in otherwise seemingly healthier individuals like antiphospholipid syndrome, procoagulopathies, vasculitides and Moyamoya disease. By 2020, Pakistan is expected to rank fourth among the most populous countries around the globe with respect to the burden of diabetes and every third person above the age of 45 years is anticipated to suffer from hypertension.2 3 5 9 10 15 16 The current prevalence of diabetes in Pakistan has been estimated to be around 9–10% as published by Qureshi and colleagues in 2014. Even the prediabetes prevalence is higher as compared with western countries. People aged 15 years or higher have an overall 19% prevalence of hypertension. Obesity is another major problem in South Asians and their body mass index has been estimated to have lower cut-off values for quantification of diseases like diabetes and cardiovascular diseases. Obesity has a prevalence ranging from 20% to 30% in South Asia.9 10 17 Apart from the above, various understudied and underestimated risk factors are shown in figure 1.

ans and their body mass index has been estimated to have lower cut-off values for quantification of diseases like diabetes and cardiovascular diseases. Obesity has a prevalence ranging from 20% to 30% in South Asia.9 10 17 Apart from the above, various understudied and underestimated risk factors are shown in figure 1. Figure 1 Risk factors for ischaemic stroke. An algorithm showing the risk factors for ischaemic stroke. SAH, subsarachnoid hemorrhage. Stroke referral protocols There is no systematic referral system or organised general practice (GP) in Pakistan. There is no continued medical education programme for GPs in Pakistan and no official collaboration network available for keeping in touch with tertiary care setups and teaching hospitals. While most GPs are able to identify strokes pertaining to anterior circulation, those related to posterior circulation presenting with subtle signs like homonymous hemianopsia, signs of cerebellar and brainstem dysfunction are usually missed. Similarly, owing to the unavailability of CT scan facilities in most of the areas outside major cities, stroke is usually missed or remains undiagnosed for days before the patient arrives at a tertiary care setup. Therefore, there is a dire need of a comprehensive programme for GPs with an aim to teach proper and prompt identification and referral of such cases.2–5 18–20

ilities in most of the areas outside major cities, stroke is usually missed or remains undiagnosed for days before the patient arrives at a tertiary care setup. Therefore, there is a dire need of a comprehensive programme for GPs with an aim to teach proper and prompt identification and referral of such cases.2–5 18–20 Recognition of clinical presentation Identification of stroke is encouraged to be made at the community level worldwide by introduction of acronyms like FAST, that is, Facial drooping/deviation/weakness, Arm weakness (arm paralysis/paresis/pronator drift), Speech difficulties (aphasia/dysphasia/slurring) and Time as shown in figure 2. There is no national awareness campaign or educational platform for the general public to get awareness about the signs of early stroke. Such signs are not even recognised by rescue operators or ambulance service personnel. The first person usually to identify these is the doctor at hospital setups. It is only at the level of medical specialists or neurologists that most of these patients are quickly and correctly diagnosed. This is very crucial in the current perspective of treatment of acute ischaemic strokes and TIAs as the window period for administering thrombolysis is just 4.5 hours at present. Loss of these ‘golden hours’ makes all the difference in active management of stroke.8 13 18–26

patients are quickly and correctly diagnosed. This is very crucial in the current perspective of treatment of acute ischaemic strokes and TIAs as the window period for administering thrombolysis is just 4.5 hours at present. Loss of these ‘golden hours’ makes all the difference in active management of stroke.8 13 18–26 Figure 2 Bedside clinical signs to be noted in stroke; (A) a patient who has not suffered a stroke can generally hold the arms in an extended position with eyes closed. (B) A patient with stroke will often display ‘arm drift’ (pronator drift)—one arm will remain extended when held outwards with eyes closed, but the other arm will drift or drop downwards and pronate (palm turned downwards). Stroke code The stroke code system in Pakistan is almost non-existent.

Figure 2 Bedside clinical signs to be noted in stroke; (A) a patient who has not suffered a stroke can generally hold the arms in an extended position with eyes closed. (B) A patient with stroke will often display ‘arm drift’ (pronator drift)—one arm will remain extended when held outwards with eyes closed, but the other arm will drift or drop downwards and pronate (palm turned downwards). Stroke code The stroke code system in Pakistan is almost non-existent. Local event sight protocols and transportation services The transporting facility of ambulances is rudimentary as compared with international standard protocols. While internationally in-transit neuroimaging and administration of tissue plasminogen activator (tPA) in ambulances is the next big thing, we are struggling with timely transportation of patients to the hospital. This probably is due to resource deficits including a low number of ambulances, poor road infrastructure and lack of on-call services for transport from rural suburbs to major cities.2–6 9 10 19–25 A transport facility by the name of 1122 was introduced in major cities a decade back. The staff is trained and well equipped to manage acute cardiovascular events and road traffic accidents. However, there is limited knowledge of stroke, its presentation variants and management protocols.

o major cities.2–6 9 10 19–25 A transport facility by the name of 1122 was introduced in major cities a decade back. The staff is trained and well equipped to manage acute cardiovascular events and road traffic accidents. However, there is limited knowledge of stroke, its presentation variants and management protocols. Transit times The travelling time from suburbs and even major districts to facilities designated for ‘neurology and stroke care’ is quite long. There are many reasons like unpaved roads, major traffic blocks and greater travelling distances. Using air transport facilities is a dream far from realisation in the near future.2–6 22–25

The travelling time from suburbs and even major districts to facilities designated for ‘neurology and stroke care’ is quite long. There are many reasons like unpaved roads, major traffic blocks and greater travelling distances. Using air transport facilities is a dream far from realisation in the near future.2–6 22–25 In-hospital emergency services The patient receiving policies at district-level hospitals is also not at par with international standards. There are usually no CT scan facilities except in major cities and patients are immediately referred to tertiary care setups. However, it takes the referral procedure almost an hour or 2; this time lapse combined with transport delays leads to patients usually falling out of the window period when they arrive at tertiary care setups. At tertiary setups, while immediate CT scan and stroke identification is the norm, the sophisticated code of calculating and limiting the care service provided for such patients in the form of door to CT time (target <25 min), CT to needle time (target <20 to 35 min) and door to needle time (target <45 to 60 min) is non-existent because of otherwise non-existent thrombolysis centres. Only individual efforts by certain units to minimise the duration up to needle time are practised, that too for selected cases and patients already coming with short histories, feasible for thrombolysis and mostly within the immediate vicinity of the stroke care centre. This also includes an urgent emergency laboratory workup to rule out contraindications to the above like complete blood counts, platelet count, partial thromboplastin time, international normalised ratio, electrolytes, serum creatinine levels, serum glucose, troponin levels, liver function tests and blood grouping.2–4 20 21–27

o includes an urgent emergency laboratory workup to rule out contraindications to the above like complete blood counts, platelet count, partial thromboplastin time, international normalised ratio, electrolytes, serum creatinine levels, serum glucose, troponin levels, liver function tests and blood grouping.2–4 20 21–27 According to a book recently published by Ali and colleagues from Pakistan, the mean alarm to door time was reported to be 255.88±112.63 min (4.26±1.87 hours), mean door to CT time as 37.75±20.13 min (0.62±0.33 hours) and mean alarm to drug time as 314.88±116.12 min (5.24±1.93 hours). According to them, out of a total of 100 studied patients, only 29% reached their stroke care setup within the 4.5 hours window period as shown in figure 3. According to the authors, only 29% patients were feasible for thrombolysis. If their study results are observed, one can clearly see that after the patient is received at a tertiary care setup, the triage of events is quite practical in making thrombolysis possible (door to CT time of 0.62±0.33 hours). However, the major loss of precious window time period is not within the hospital but outside the premises, that is, alarm/stroke to door time (4.26±1.87 hours). The overall result is a late alarm to drug time (5.24±1.93 hours). This highlights the fact that the untimely diagnosis and arrival of patients with stroke at stroke care centres is another major contributor towards the failure of achieving a successful thrombolysis target protocol in Pakistan. The fact that brain tissue is viable to be salvaged within a given period of time is the key concept in understanding why the time frame for delivering thrombolysis or intervention is so crucial in acute ischaemic stroke as shown in figure 4.28–40

lure of achieving a successful thrombolysis target protocol in Pakistan. The fact that brain tissue is viable to be salvaged within a given period of time is the key concept in understanding why the time frame for delivering thrombolysis or intervention is so crucial in acute ischaemic stroke as shown in figure 4.28–40 Figure 3 Graphical representation of patients within and outside window period for IV thrombolysis in a study conducted by Ali and colleagues in 2015–2016 (adapted from ref. 28). Out of the cohort of 100 patients, 29% had presented within the window period for IV thrombolysis. IV, intravenous. Figure 4 Process of tissue damage and time intervals for possible salvage in acute ischaemic stroke. Current facilities for acute ischaemic stroke in Pakistan The current stroke treatment in Pakistan can be divided into the following categories:

Figure 3 Graphical representation of patients within and outside window period for IV thrombolysis in a study conducted by Ali and colleagues in 2015–2016 (adapted from ref. 28). Out of the cohort of 100 patients, 29% had presented within the window period for IV thrombolysis. IV, intravenous. Figure 4 Process of tissue damage and time intervals for possible salvage in acute ischaemic stroke. Current facilities for acute ischaemic stroke in Pakistan The current stroke treatment in Pakistan can be divided into the following categories: Intravenous thrombolysis: availability and expenses Currently, only two centres provide the facility of ER triage of thrombolysis for acute ischaemic stroke and TIAs. These centres too have intervened in a limited number of cases so far. They receive no more than 55–65 acute ischaemic stroke cases per year as reported and the yearly intravenous (IV) thrombolysis rate is around 12–15/year. The data published on this experience are also sparse. Moreover, the cost of thrombolysis in Pakistan at these centres is between US$3000 and US$4500. The cost of IV recombinant tPA (rtPA) injection itself is around US$2250. The per capita income in Pakistan is US$450 on average. The rtPA is not authorised by the Ministry of Health in Pakistan as a registered medical product under licence of the Government of Pakistan yet and can only be arranged by individual hospitals on their own. All these issues need to be addressed separately and seriously. For this, neurologists all over the country have put forward a petition to convince the government to ponder on this fragile situation.4–6 20–26 33–38

licence of the Government of Pakistan yet and can only be arranged by individual hospitals on their own. All these issues need to be addressed separately and seriously. For this, neurologists all over the country have put forward a petition to convince the government to ponder on this fragile situation.4–6 20–26 33–38 Intra-arterial thrombolysis Up until now, no centre in Pakistan has the expertise to offer intra-arterial rtPA. Various research publications have reported variable results for this mode of therapy and interest in its advancement is growing. This modality is under research and available only in a few centres around the globe for selected patients.4 5 23–26 33–38 Endovascular surgery Endovascular surgery (EVS) is indicated for cases with a thrombus size >5 mm, which is not amenable to treatment with IV rtPA. No centre in Pakistan has the setup and expertise to offer the services of EVS. There is no trained EV surgeon/interventionist available in Pakistan at present. Those getting their training from abroad usually do not prefer to come back due to limited resources and scope of growth in their fields.24–26 33–38

h IV rtPA. No centre in Pakistan has the setup and expertise to offer the services of EVS. There is no trained EV surgeon/interventionist available in Pakistan at present. Those getting their training from abroad usually do not prefer to come back due to limited resources and scope of growth in their fields.24–26 33–38 Current treatment protocols The current targeted treatment of acute ischaemic stroke in Pakistan includes the administration of aspirin as a loading dose of 300 mg combined within first 24 hours of event in the emergency. Along with this, supportive therapy is offered in the form of lipid-lowering medication, IV fluids, control of blood pressure and serum glucose levels, proton pump inhibitors, bladder and bowel care, mouth care, nursing care and bedside physiotherapy. The current acute stroke management in Pakistan practised at large at specialised stroke units is as shown in tables 1–4.2–6 20–26 33–39 Table 1 Acute ischaemic stroke: an ER protocol (routinely followed protocols across Pakistan)

Current treatment protocols The current targeted treatment of acute ischaemic stroke in Pakistan includes the administration of aspirin as a loading dose of 300 mg combined within first 24 hours of event in the emergency. Along with this, supportive therapy is offered in the form of lipid-lowering medication, IV fluids, control of blood pressure and serum glucose levels, proton pump inhibitors, bladder and bowel care, mouth care, nursing care and bedside physiotherapy. The current acute stroke management in Pakistan practised at large at specialised stroke units is as shown in tables 1–4.2–6 20–26 33–39 Table 1 Acute ischaemic stroke: an ER protocol (routinely followed protocols across Pakistan) Acute ischaemic stroke: an ER protocol 1. Please carry out the following orders as appropriate Yes No Date and time 2. Plain CT head STAT 3. CBC and platelet count 4. prothrombin time (PT), INR 5. Electrolytes, creatinine, glucose, Trop, liver function tests (LFTs) 6. 0.9% normal saline (NS) at 100 to 120 mL/hour for 24 hours. No dextrose saline (if congestive cardiac failure (CCF), use 0.9% NS at ≤60 to 70 mL/hour) 7. If systolic BP>185 mm  Hg or diastolic BP>110 mm  Hg for ≥2 readings taken 5–10 min apart, use hypertension management protocol 8. If random blood glucose >150 mg/dL, use S/C insulin protocol 9. Metoclopramide 10 mg IV every 8 hours (for nausea/vomiting) 10. Acetaminophen 1000 mg IV every 8 hours (for headache or fever >37.5°C) 11. Aspirin 300 mg loading dose PO in ER if haemorrhagic stroke ruled out BP, blood pressure; CBC, complete blood count; INR, international normalised ratio; IV, intravenous; PO, orally; S/C, subcutaneous; Trop, troponin.

s (for nausea/vomiting) 10. Acetaminophen 1000 mg IV every 8 hours (for headache or fever >37.5°C) 11. Aspirin 300 mg loading dose PO in ER if haemorrhagic stroke ruled out BP, blood pressure; CBC, complete blood count; INR, international normalised ratio; IV, intravenous; PO, orally; S/C, subcutaneous; Trop, troponin. Table 2 Hypertension management protocol (routinely followed protocols across Pakistan) Hypertension management protocol 12. Please carry out the following orders as appropriate Yes No Date and time 13. If systolic BP>185 mm  Hg or diastolic BP>110 mm  Hg for ≥2 readings taken 5–10 min apart, use hypertension management protocol 14. If systolic BP>185 mm  Hg or diastolic BP>110 mm  Hg, treat with labetalol bolus or infusion 15. Give: IV labetalol 10–20 mg over 1–2 min 16. If still elevated after 10–15 min:The dose may be repeated every 10–15 min up to a total of no more than 300 mg/24 hours or give labetalol initial loading dose followed by infusion at 2–8 mg/min 17. Hold treatment if BP<185/110 mm  Hg BP, blood pressure; IV, intravenous. Table 3 Clinical pathway for acute ischaemic stroke (routinely followed protocols across Pakistan)

Hypertension management protocol 12. Please carry out the following orders as appropriate Yes No Date and time 13. If systolic BP>185 mm  Hg or diastolic BP>110 mm  Hg for ≥2 readings taken 5–10 min apart, use hypertension management protocol 14. If systolic BP>185 mm  Hg or diastolic BP>110 mm  Hg, treat with labetalol bolus or infusion 15. Give: IV labetalol 10–20 mg over 1–2 min 16. If still elevated after 10–15 min:The dose may be repeated every 10–15 min up to a total of no more than 300 mg/24 hours or give labetalol initial loading dose followed by infusion at 2–8 mg/min 17. Hold treatment if BP<185/110 mm  Hg BP, blood pressure; IV, intravenous. Table 3 Clinical pathway for acute ischaemic stroke (routinely followed protocols across Pakistan) Clinical pathway for acute ischaemic stroke 18. Please carry out the following orders as appropriate Yes No Date and time 19. Admit to stroke unit or ICU for 48–72 hours (at a few centres only), then shift to general ward if stable 20. National Institute of Health Stroke Scale (NIHSS)/GCS (individual centre adapted protocols may differ) once/twice daily 21. CT scan of head repeated STAT if decreased level of consciousness or new deficit 22. Monitor vital signs (BP, HR, cardiac rhythm, scores, metabolic acidemia, pulse oximeter (SaPO2) (give oxygen as indicated), Temp and respiratory rate (RR)) 23. Monitor vital signs: every 30×3; every 1 hour×every 6 hours, then every 3 hours×every 10 hours (individual centre adapted protocols may differ) 24. Start mobilisation (in bed or chair) if stable (after functional assessment) 25. Bed rest for 24 hours 26. Keep nothing per oral (NPO) until swallowing screen done (if dysphagia present, keep NPO and repeat screen in 24 hours) BP, blood pressure; GCS, Glasgow Coma Scale; ICU, intensive care unit; NPO, nothing per oral; SaPO2, scores, metabolic acidemia, pulse oximeter; Temp, temperature.

sment) 25. Bed rest for 24 hours 26. Keep nothing per oral (NPO) until swallowing screen done (if dysphagia present, keep NPO and repeat screen in 24 hours) BP, blood pressure; GCS, Glasgow Coma Scale; ICU, intensive care unit; NPO, nothing per oral; SaPO2, scores, metabolic acidemia, pulse oximeter; Temp, temperature. Table 4 Clinical stroke extended management pathway (routinely followed protocols across Pakistan) Clinical stroke extended management pathway 27. Please carry out the following orders as appropriate Yes No Date and time 28. Extracranial imaging: carotid Doppler, CTA, MRA) in 2–7 days 29. ECG in 2–7 days (trans thoracic echocardiography (TTE)/trans esophageal echocardiography (TEE) as indicated) 30. DVT prophylaxis 31. Stroke educational material for patient and family 32. Removal of Foley's catheter and start bladder management (if stable and feasible) 33. Discharged on: Antithrombotic therapy OR Anticoagulation therapy for atrial fibrillation/flutter Antihypertensives (as indicated) Statins/cholesterol lower agent (as indicated) Antidiabetic agent (if indicated) 34. Fasting lipid profile, HbA1c 35. Holter monitor (if indicated 36. Specialised workup for young stroke (like anti nuclear antibody (ANA), extractable Nuclear Antigen (eNA), etc) 37. Referral to: Rehabilitation specialist Speech therapist Dietician Medical team (if otherwise indicated or required) 38. Follow-up CTA, CT angiography; HbA1c, glycated haemoglobin; MRA, MR angiography.

34. Fasting lipid profile, HbA1c 35. Holter monitor (if indicated 36. Specialised workup for young stroke (like anti nuclear antibody (ANA), extractable Nuclear Antigen (eNA), etc) 37. Referral to: Rehabilitation specialist Speech therapist Dietician Medical team (if otherwise indicated or required) 38. Follow-up CTA, CT angiography; HbA1c, glycated haemoglobin; MRA, MR angiography. Follow-up and rehabilitation Follow-up protocols are defined at the level of tertiary care setups. However, practical implementation of these protocols is poor, especially for cases referred from smaller districts and villages. This emphasises the utmost need for establishing local stroke rehabilitation centres in the suburbs which can follow these patients and lessen the burden on the bigger stroke centres. Moreover, only a few centres are specialised in rehabilitation medicine where full-time support to stroke affectees is offered.2–5 21–26 33–37 Neuroradiological support The radiological support in the diagnosis of acute stroke is also limited. Only the tertiary care teaching units are equipped with CT and MRI machines. After office hours, MRI facility is not available. The availability of functional MRI, perfusion scans and positron emission tomography/single-photon emission CT and digital subtraction angiography (DSA) is virtually non-existent.2 3 24–26 33–37

ly the tertiary care teaching units are equipped with CT and MRI machines. After office hours, MRI facility is not available. The availability of functional MRI, perfusion scans and positron emission tomography/single-photon emission CT and digital subtraction angiography (DSA) is virtually non-existent.2 3 24–26 33–37 Stroke neurology training facility There is no neurovascular training facility in Pakistan at present, except for a single centre in Karachi with only one fellow trained per year. The total number of neurologists registered until now date in Pakistan is fewer than 200 according to the statistical records updated in 2016. More than half of these are residing abroad. So the lack of trained personnel in the face of limited resources and poor financial support culminates in a system lagging far behind in treatment of acute stroke in Pakistan.41 Role of Pakistan Stroke Society and Pakistan Society of Neurology in awareness and advocacy The Pakistan Stroke Society (PSS) was established in 2001. Its main objective is to improve prevention and care of stroke in Pakistan by increasing the public and physicians’ awareness. The PSS is affiliated with the International Stroke Society, which is an advisory body to the WHO. The society has more than 200 physician members.42

roke Society (PSS) was established in 2001. Its main objective is to improve prevention and care of stroke in Pakistan by increasing the public and physicians’ awareness. The PSS is affiliated with the International Stroke Society, which is an advisory body to the WHO. The society has more than 200 physician members.42 The society is actively working for the past 15 years holding seminars, public awareness sessions and walks with/for patients with stroke. The first International Stroke Conference was held in 2004 in Lahore followed by its successive chapters over the years. Every year, international speakers participate in an annual stroke conference from all over the world including participants from the USA, Canada, Emirates, India, Japan, Greece, China and many more. In the recent 2016, several conferences were held on neurology including an annual meeting in Quetta on stroke prevention and management. A comprehensive stroke course was held which included eight workshops and trained over 400 doctors specifically in stroke. There were also more than 20 seminars on stroke recognised countrywide and given substantial media coverage. Awareness was made by organising public awareness days like Stroke Day and National Disability Day and has become a norm on the part of PSS and Pakistan Society of Neurology (PSN). To further engage the general public, there were 8 press conferences, 20 media appearances, 6 newspaper articles and more than 1000 news items published covering various perspectives of stroke. Billboards, posters and brochures were used as well. Additionally, the PSN has recently launched a Facebook page and website. Guidelines for stroke in Pakistan have been published providing detailed national guidance under the circumstances of having limited resources. Of a total of 40 neurology centres, 18 have been approved for Fellowship of College of Physicians and Surgeons (FCPS) training by the efforts of PSS and PSN. This year, more than 30 medical doctors passed FCPS neurology, amounting to a total of 200 neurologists in the country and contributing towards handling the heavy stroke burden in Pakistan. The Pakistan Journal of Neurological Sciences is a good quality quarterly journal where contributions from members and associated faculty of PSS and PSN have been substantial.

d FCPS neurology, amounting to a total of 200 neurologists in the country and contributing towards handling the heavy stroke burden in Pakistan. The Pakistan Journal of Neurological Sciences is a good quality quarterly journal where contributions from members and associated faculty of PSS and PSN have been substantial. Currently, it is available on the web, E-commons, Higher Education Commission (HEC), Pakistan Medical and Dental Council (PMDC) and WHO-Regional Office for the Eastern Mediterranean (EMRO) and plans to be on PubMed by 2017/2018. Moreover, it is a premier journal for stroke and related neurosciences in Pakistan. Furthermore, a public awareness magazine Jehan-e-Aesab is being published since January 2016 creating countrywide awareness about stroke. It aims to promote medical writing in Urdu, and is targeting media, faculty, lay people, patients and attendants to create awareness about stroke and other neurological diseases in the local and national languages for a more effective process of communication.41 42 National stroke care guidelines43 As of 2010, the national stroke care guidelines provide a comprehensive and clear strategy for care of patients with stroke in Pakistan. The salient features of the guidelines state: Prehospital recognition and stroke triage: to decrease delays in presentation to care via a track record of admitting and managing patients with stroke.43 In-hospital management: by following strict protocols at subsequent stroke care steps. Emergency department Rapid airway, breathing and circulation management. Rapid assessment of neurological deficits.

National stroke care guidelines43 As of 2010, the national stroke care guidelines provide a comprehensive and clear strategy for care of patients with stroke in Pakistan. The salient features of the guidelines state: Prehospital recognition and stroke triage: to decrease delays in presentation to care via a track record of admitting and managing patients with stroke.43 In-hospital management: by following strict protocols at subsequent stroke care steps. Emergency department Rapid airway, breathing and circulation management. Rapid assessment of neurological deficits. Determination of time of onset of stroke symptoms. Determination of presence of risk factors. Standardised use of NIHSS stroke scale score. If rtPA is available at the facility, use of National Institute of Neurological Disorders and Stroke (NINDS) recommendations for stroke chain of survival: 10 min for the emergency physician evaluation, specialist (neurologist) assessment within 10 min, and 25 min to CT scan, allowing rtPA administration within 45 min to an hour. Imaging of the brain: an emergency non-contrast head CT. Ancillary testing at the very least should consider a 12 lead ECG, blood tests for diabetes, abnormal lipids, coagulopathy and screen for chronic renal failure. Control of blood pressure with those who receive IV rtPA requiring strict control of blood pressure to below 185/110 mm hg. The Food and Drug Administration (FDA) approved treatment protocol for thrombolysis for acute stroke presenting within 3–4.5 hours of symptom onset using IV rtPA.43 Other parameters Control of core body temperature.

Control of blood pressure with those who receive IV rtPA requiring strict control of blood pressure to below 185/110 mm hg. The Food and Drug Administration (FDA) approved treatment protocol for thrombolysis for acute stroke presenting within 3–4.5 hours of symptom onset using IV rtPA.43 Other parameters Control of core body temperature. A desired level of glucose between 80 and 140 mg/dL. Urgent anticoagulation with the goal of preventing early recurrent stroke, halting worsening or improving outcomes after ischaemic stroke is not recommended. This is regardless of the aetiology of stroke, for example, cardioembolic stroke. Anticoagulation is contraindicated within 24 hours of administration of rtPA. Antiplatelet therapy with aspirin 160–325 mg daily, given orally (or per rectum in patients who cannot swallow), and started within 48 hours of onset of presumed ischaemic stroke. Patients should receive isotonic hydration and free fluids should be avoided. Nutritional supplementation is not necessary. Evaluation for aspiration is needed prior to initiation of diet and the diet should be modified accordingly. Screening for the presence of swallowing problems with the water swallowing tests and modifications in the diet should be made accordingly. Patients who fail the initial swallowing screen should get a nasogastric tube placed for medications and feeds. Those who are having difficulty handling even their basic secretions may actually benefit from nil per oral policy in the initial 24 hours.

wing tests and modifications in the diet should be made accordingly. Patients who fail the initial swallowing screen should get a nasogastric tube placed for medications and feeds. Those who are having difficulty handling even their basic secretions may actually benefit from nil per oral policy in the initial 24 hours. All patients unable to mobilise independently should ideally be provided with a pressure relieving mattress as an alternative to a standard hospital mattress. Also, there should be standing instructions for repositioning these patients every 2 hours to avoid pressure sores. Avoidance of deep vein thrombosis in immobilised patients via frequent movements and the use of low-dose subcutaneous heparin is suggested in the acute phase. Intermittent compression devices are recommended in those where the risk of intracerebral hemorrhage (ICH) is high. Early rehabilitation. Avoid routine use of steroids. Avoid injudicious use of neuroprotection and neurotonics.43 Posthospital management: Aspirin in a dose ranging from 75 to 300 mg is efficacious in stroke prevention. Clopidogrel is marginally better at increased cost and is therefore suggested in those with concomitant peripheral vascular disease and/or intolerance to aspirin. Combination therapy in patients with stroke with aspirin and clopidogrel has higher risks of symptomatic ICH and should be used only for TIA. Dose-adjusted warfarin is suggested in an international normalised ratio 2–3 for those who have intermittent or continuous atrial fibrillation.

Posthospital management: Aspirin in a dose ranging from 75 to 300 mg is efficacious in stroke prevention. Clopidogrel is marginally better at increased cost and is therefore suggested in those with concomitant peripheral vascular disease and/or intolerance to aspirin. Combination therapy in patients with stroke with aspirin and clopidogrel has higher risks of symptomatic ICH and should be used only for TIA. Dose-adjusted warfarin is suggested in an international normalised ratio 2–3 for those who have intermittent or continuous atrial fibrillation. Goal to reduce blood pressure and any antihypertensive which will ensure compliance in the long run is preferable. Therefore, the drug should be selected depending on the availability and the financial status of the patient. Patients with ischaemic stroke or TIA with elevated cholesterol, comorbid coronary artery disease or evidence of an atherosclerotic origin should be counselled regarding lifestyle modification, dietary guidelines and statins are recommended. The recommendations for glucose control should be the same for patients with and without a prior stroke. Control of other modifiable cardiovascular risk factors in a patient with diabetes like aggressive blood pressure control and lipid control is warranted. Physicians should ask their patients repeatedly to quit smoking for primary and secondary prevention. More aggressive dietary control and regular physical activity. Dietary salt restriction.

The recommendations for glucose control should be the same for patients with and without a prior stroke. Control of other modifiable cardiovascular risk factors in a patient with diabetes like aggressive blood pressure control and lipid control is warranted. Physicians should ask their patients repeatedly to quit smoking for primary and secondary prevention. More aggressive dietary control and regular physical activity. Dietary salt restriction. Early identification and management of depression poststroke is vital to ensure early recovery and to prevent cognitive impairment. Conventional tricyclic antidepressants are contraindicated among patients with stroke due to their adverse effects. selective serotonin reuptake inhibitor (SSRIs) have a low adverse effect profile and a good efficacy, making them invaluable in patients with multiple comorbidities.43

and to prevent cognitive impairment. Conventional tricyclic antidepressants are contraindicated among patients with stroke due to their adverse effects. selective serotonin reuptake inhibitor (SSRIs) have a low adverse effect profile and a good efficacy, making them invaluable in patients with multiple comorbidities.43 Paediatric stroke in Pakistan According to internationally available literature, ischaemic stroke in children occurs in 55% of all strokes, the rest being haemorrhagic. This is in contrast to adult demographics among whom ischaemic stroke comprising 80–85% of all strokes. Delay in diagnosis is common in the paediatric population and thus opportunities for timely intervention are substantially limited. Experience of thrombolysis in paediatric stroke is non-existent. Furthermore, the risk factor account for stroke in children is more versatile than that in the adult population. Among these are central nervous system infections, cardiac diseases (congenital as well as acquired), vasculopathies (fibromuscular dysplasia and Moyamoya disease), vasculitides, sickle cell disease, coagulopathies and arterial dissection.44

nt for stroke in children is more versatile than that in the adult population. Among these are central nervous system infections, cardiac diseases (congenital as well as acquired), vasculopathies (fibromuscular dysplasia and Moyamoya disease), vasculitides, sickle cell disease, coagulopathies and arterial dissection.44 A local study from Pakistan by Chand and colleagues identified a total of 29 paediatric patients with acute ischaemic stroke over a period of 5 years. Their mean age was ∼4.5 years with 76% over infancy. Male-to-female ratio was 3:1 and mean duration of symptoms at the time of presentation was ∼5.75 hours. Seizures (mostly generalised tonic–clonic), loss of consciousness and paresis were the most common clinical presentations as observed by them. Other common features included fever, vomiting, cranial nerve palsies, delayed development, headache and blindness. They reported a positive family history of stroke among these children to be only 3.5%. Unique ECG findings particular to the paediatric population included myocarditis tetralogy of Fallot, complex congenital heart disease and ventricular septal defect. Hypercoagulopathy (protein S and/or protein C deficiency) was found in 37.5% for those having its workup done. They conclude that multiple strokes, cardiovascular diseases and age below 1 year were associated with higher mortality.44 Another observational study conducted by Siddiqui et al45 concluded that the main aetiology of stroke in children was causing strokes in 56.09% of children and the majority of children (78.26%) in this group were below 5 years.44

A local study from Pakistan by Chand and colleagues identified a total of 29 paediatric patients with acute ischaemic stroke over a period of 5 years. Their mean age was ∼4.5 years with 76% over infancy. Male-to-female ratio was 3:1 and mean duration of symptoms at the time of presentation was ∼5.75 hours. Seizures (mostly generalised tonic–clonic), loss of consciousness and paresis were the most common clinical presentations as observed by them. Other common features included fever, vomiting, cranial nerve palsies, delayed development, headache and blindness. They reported a positive family history of stroke among these children to be only 3.5%. Unique ECG findings particular to the paediatric population included myocarditis tetralogy of Fallot, complex congenital heart disease and ventricular septal defect. Hypercoagulopathy (protein S and/or protein C deficiency) was found in 37.5% for those having its workup done. They conclude that multiple strokes, cardiovascular diseases and age below 1 year were associated with higher mortality.44 Another observational study conducted by Siddiqui et al45 concluded that the main aetiology of stroke in children was causing strokes in 56.09% of children and the majority of children (78.26%) in this group were below 5 years.44 In the background of the aforementioned studies, it is clear that characteristics of stroke described for the paediatric population differ significantly from those for adults but scarce data, more so in this age group, make understanding limited.44 45 It is therefore of utmost importance that stroke should not be considered only a disease of adults, but a national database registry should be set up for children with stroke. Furthermore, stroke centres should be established separately as an integral part of a paediatric care hospital having a recognised department for paediatric neurology.

st importance that stroke should not be considered only a disease of adults, but a national database registry should be set up for children with stroke. Furthermore, stroke centres should be established separately as an integral part of a paediatric care hospital having a recognised department for paediatric neurology. Stroke in women in Pakistan There are very limited data of hospital-based studies and stroke registries about stroke in women, especially in young women in South Asia. There is poor reporting of stroke in women in South Asia, despite it being a leading cause of death in women aged above 60 years. This is probably because women with stroke or cardiac arrest in this part of the world are less likely to be taken to hospital than men with these conditions due to various reasons including level of education, socioeconomic status, overall literacy rate, local social values and community beliefs. Two studies on stroke in young women from eight Asian countries found that large-vessel thrombosis (24%), cortical vein thrombosis (CVT) (21%) and cardioembolism (19%) were the most common mechanisms of stroke in this population with a significant proportion related to pregnancy, especially postpartum.7 Genetic aspects of stroke in Pakistan Genetic aspects of stroke have been studied to a very limited extent until now in South Asia and almost none in Pakistan. While basic research has gained pace regarding local demographic data, advanced research and genetic studies are almost non-existent.7

hines. The use of higher-field MRI and SWI are known to increase CMBs detection. The variety of different techniques used in the assessment of artery atherosclerosis or CMBs in the included studies should be considered a source of heterogeneity. The accuracy of the methodology needs to be established in future studies. Conclusions In conclusion, the results of this systematic review and meta-analysis suggest that there is a relationship between large artery atherosclerosis and CMBs. Whether the occurrence of CMBs in patients with large artery atherosclerosis can predict the future risk of haemorrhagic cerebral vascular events is unclear. Longitudinal studies in larger populations are needed to confirm the impact of atherosclerosis on the bleeding-prone CMBs, which may have potential therapeutic implications. Contributors: BP designed the study, and revised the manuscript. LD collected and extracted data, and drafted the manuscript. YH collected and extracted data. Funding: This work was supported by the National Natural Science Foundation of China (grant number 81471206) and the Beijing Natural Science Foundation (grant number 7152121). Competing interests: None declared. Provenance and peer review: Not commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

Stroke in women in Pakistan There are very limited data of hospital-based studies and stroke registries about stroke in women, especially in young women in South Asia. There is poor reporting of stroke in women in South Asia, despite it being a leading cause of death in women aged above 60 years. This is probably because women with stroke or cardiac arrest in this part of the world are less likely to be taken to hospital than men with these conditions due to various reasons including level of education, socioeconomic status, overall literacy rate, local social values and community beliefs. Two studies on stroke in young women from eight Asian countries found that large-vessel thrombosis (24%), cortical vein thrombosis (CVT) (21%) and cardioembolism (19%) were the most common mechanisms of stroke in this population with a significant proportion related to pregnancy, especially postpartum.7 Genetic aspects of stroke in Pakistan Genetic aspects of stroke have been studied to a very limited extent until now in South Asia and almost none in Pakistan. While basic research has gained pace regarding local demographic data, advanced research and genetic studies are almost non-existent.7 Epidemiological studies of families, homozygous and heterozygous twins have shown a distinct genetic component in the predisposition to stroke. In this regard, three polymorphisms in the phosphodiesterase 4D (PDE4D) gene were evaluated in 200 patients with stroke compared with 250 controls from a single study in Pakistan. The study concluded that there is an association between the single nucleotide polymorphism (SNP) SNP83 and patients with ischaemic stroke.

In this regard, three polymorphisms in the phosphodiesterase 4D (PDE4D) gene were evaluated in 200 patients with stroke compared with 250 controls from a single study in Pakistan. The study concluded that there is an association between the single nucleotide polymorphism (SNP) SNP83 and patients with ischaemic stroke. An India-based study showed that SNP41, SNP56 and a novel SNP in the PDE4D gene were associated with various stroke subtypes. Some studies have shown significantly positive associations between the E-selectin gene polymorphism Ser128Arg and ischaemic stroke. Others have suggested a link between the ACE and AAD1 genes and extracranial and intracranial atherosclerosis contributing to ischaemic stroke. A study from Chennai, India showed that paraoxonase 1 (PON1) activity and PON1 Gln192Arg genotypes are associated with ischaemic stroke.7 A recent database for genetic studies has been started to gather data on such patients by individual efforts at a few centres across Pakistan but a national database registry specifically for studying genetic aspects has not yet been formulated in this regard. This needs to be done in the near future to help determine patients who might be at risk of stroke among the Pakistani population.

n such patients by individual efforts at a few centres across Pakistan but a national database registry specifically for studying genetic aspects has not yet been formulated in this regard. This needs to be done in the near future to help determine patients who might be at risk of stroke among the Pakistani population. Future perspectives While Pakistan is lagging behind countries of the developed world in stroke neurology in general and management of acute ischaemic stroke in particular, there has been some commendable progress in recent times. In Karachi, Aga Khan University Hospital has the facility to deliver state-of-the-art stroke management. It has recently been established and has treated a decent number of patients with thrombolysis in the past 2 years. In Islamabad, Shifa International Hospital has an ongoing research study in which they have successfully treated more than a dozen patients with acute ischaemic stroke in the year 2015 with IV thrombolysis. Coming to public sector hospitals, the neurology department of Pakistan Institute of Medical Sciences has put forward a proposal to the federal government for the establishment of an acute stroke unit and is also sending one of its neurologists abroad for training in vascular neurology. The future of stroke seems promising in Pakistan with the newly established Centre of Neurosciences in Lahore, which is to be inaugurated this year. A much larger scale project is on its way to start within a year or so in Islamabad, especially to deliver stroke management expertise. The establishment of new stroke centres and allocation of funds for the progress of stroke neurology in Pakistan is encouraging and that day is not far when Pakistan will be at par with the West in modern fundamentals of stroke neurology.2 4 5 20–26 39–42

in Islamabad, especially to deliver stroke management expertise. The establishment of new stroke centres and allocation of funds for the progress of stroke neurology in Pakistan is encouraging and that day is not far when Pakistan will be at par with the West in modern fundamentals of stroke neurology.2 4 5 20–26 39–42 Competing interests: None declared. Provenance and peer review: Commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

Introduction Cerebral small vessel disease (CSVD) contributes to about a fifth of all strokes worldwide, and is considered a major cause of disability and vascular cognitive impairment.1 2 Prominent manifestations of CSVD on neuroimaging include white matter hyperintensity (WMH), lacunar infarction, enlarged perivascular spaces and cerebral microbleeds (CMBs).1 2 The most common causes of CSVD are diseases that affect the cerebral perforating arteries, arterioles, capillaries and venules. However, recent studies have shown a link between CSVD and the markers of large artery atherosclerosis, including arterial stenosis and greater carotid artery intima-media thickness (cIMT).3–7 Large artery atherosclerosis is considered a systemic and chronic inflammatory disease that may lead to both cardiovascular and cerebrovascular diseases.8 Recent studies found a positive association between large artery atherosclerosis and CSVD.9 To the best of our knowledge, treatments for large artery atherosclerosis such as antihypertensive therapy are also effective in patients with CSVD.10 Although the aetiology and pathogenesis of CSVD remain unclear, large artery atherosclerosis may become a potential therapeutic target for the prevention of CVSD. Therefore, it is meaningful to clarify the relationship between CSVD and large artery atherosclerosis.

ve therapy are also effective in patients with CSVD.10 Although the aetiology and pathogenesis of CSVD remain unclear, large artery atherosclerosis may become a potential therapeutic target for the prevention of CVSD. Therefore, it is meaningful to clarify the relationship between CSVD and large artery atherosclerosis. CMBs are regarded as small areas of signal loss on T2*-weighted gradient echo (GRE) MRI sequences, pathologically representing bleeding-prone microangiopathies, including hypertensive arteriopathy and cerebral amyloid angiopathy.2 11–13 The presence of CMBs is considered a strong predictor of future stroke, particularly haemorrhagic stroke.14 15 There is evidence that large artery atherosclerosis may be associated with CMBs, but the results are controversial. Older age, hypertension and diabetes mellitus are common vascular risk factors that have an impact on both large arteries and small vessels.9 Chronic cerebral hypoperfusion could be a mechanism by which CMBs are associated with large artery atherosclerosis.16 17 To determine whether an association exists between large artery atherosclerosis and CMBs, we conducted a systematic review and meta-analysis of the literature, hoping that this would throw light on the pathogenesis and therapeutic strategy for CMBs.

m by which CMBs are associated with large artery atherosclerosis.16 17 To determine whether an association exists between large artery atherosclerosis and CMBs, we conducted a systematic review and meta-analysis of the literature, hoping that this would throw light on the pathogenesis and therapeutic strategy for CMBs. Methods Search strategy We searched PubMed, MEDLINE, Web of Science, EMBASE and the Cochrane Library to identify relevant studies published before 1 July 2016 using the following search terms: ‘microbleeds', or ‘microhaemorrhage’, or ‘gradient-echo’, or ‘susceptibility weighted imaging’ in association with ‘artery atherosclerosis’, or ‘artery stenosis’, or ‘artery plaques’, or ‘artery calcification’, or ‘carotid intima-media thickness’. Related articles and reference lists from all included articles were also searched to identify additional studies. Selection criteria Two independent reviewers selected all studies. Inclusion criteria for the studies were: (1) cross-sectional or longitudinal in design; (2) investigated the association of large artery atherosclerosis and CMBs; (3) artery atherosclerosis including: cerebral artery stenosis >50% (internal carotid artery (ICA) or common carotid artery, intracranial or extracranial arteries) measured by digital subtraction angiography, CT angiography (CTA), MR angiography or carotid duplex ultrasound (CDUS); cIMT measured by CDUS; (4) studies published in English; (5) full paper could be obtained; (6) the participants were humans. Consensus was reached through discussion.

acranial or extracranial arteries) measured by digital subtraction angiography, CT angiography (CTA), MR angiography or carotid duplex ultrasound (CDUS); cIMT measured by CDUS; (4) studies published in English; (5) full paper could be obtained; (6) the participants were humans. Consensus was reached through discussion. Quality assessment The quality assessment standards for observational studies were recommended by the Agency for Healthcare Research and Quality (AHRQ). Quality assessments of cohort studies were performed using the Newcastle Ottawa Scale (NOS), and quality of cross-sectional studies was assessed using an 11-item instrument.18 19 Two independent researchers evaluated the included studies. Any disagreements were resolved after discussion with another author. Data extraction Two authors independently went through each eligible study and extracted information on the following items: first author's name, publication year, study design, sample size, sex and age distribution, prevalence of CMBs, MRI parameters, incidence rate of stenosis >50%, mean cIMT, detection methods, definition of large artery atherosclerosis and other principal findings associated with CMBs (p<0.05). The missing data and information of included studies were obtained by contacting the authors if possible.

istribution, prevalence of CMBs, MRI parameters, incidence rate of stenosis >50%, mean cIMT, detection methods, definition of large artery atherosclerosis and other principal findings associated with CMBs (p<0.05). The missing data and information of included studies were obtained by contacting the authors if possible. Statistical analysis The association of arterial stenosis and CMBs was estimated by the OR. The association of mean cIMT and microbleeds was calculated using the standardised mean difference (SMD). A random-effects model was employed when heterogeneity was found (p<0.10 or I2 >50%); otherwise, a fixed-effects model was used. The heterogeneity among studies was assessed by the Higgins I2 statistic and Cochran's Q test. An I2 > 50% and (or) a Q test p value <0.10 indicated statistical heterogeneity. Funnel plots and Egger's linear regression test were used to evaluate publication bias. A sensitivity analysis was investigated by omitting a single study in each turn. The meta-analysis was performed using Stata V.14.1 (Stata Corporation LP, College Station, Texas, USA). All p values were two-sided and p<0.05 was considered statically significant. We prepared this report according to the Meta-analysis of Observational Studies in Epidemiology (MOOSE) proposal.20

le study in each turn. The meta-analysis was performed using Stata V.14.1 (Stata Corporation LP, College Station, Texas, USA). All p values were two-sided and p<0.05 was considered statically significant. We prepared this report according to the Meta-analysis of Observational Studies in Epidemiology (MOOSE) proposal.20 Results Selection process and study characteristics A total of 339 studies were identified. Case reports, letters and reviews were excluded. Only 42 studies were included for review of the full text based on the titles and abstracts. Of these, 34 were subsequently excluded: no stenosis>50% or mean cIMT reported (n=20), no CMBs rate reported (n=5), no study of association between large artery atherosclerosis and CMBs (n=2), no control group (n=4), not cross-sectional or longitudinal in design (n=3). Finally, eight studies including a total of 7160 participants were pooled in this meta-analysis (figure 1).21–28 Four of the eight studies reported an incidence rate of stenosis>50%,23–25 28 and four studies reported mean cIMT.21 22 26 27 Six of the included studies were cross-sectional, except that two were prospective.27 28 The main characteristics of included studies are shown in table 1. Table 1 Characteristics of included Studies

Results Selection process and study characteristics A total of 339 studies were identified. Case reports, letters and reviews were excluded. Only 42 studies were included for review of the full text based on the titles and abstracts. Of these, 34 were subsequently excluded: no stenosis>50% or mean cIMT reported (n=20), no CMBs rate reported (n=5), no study of association between large artery atherosclerosis and CMBs (n=2), no control group (n=4), not cross-sectional or longitudinal in design (n=3). Finally, eight studies including a total of 7160 participants were pooled in this meta-analysis (figure 1).21–28 Four of the eight studies reported an incidence rate of stenosis>50%,23–25 28 and four studies reported mean cIMT.21 22 26 27 Six of the included studies were cross-sectional, except that two were prospective.27 28 The main characteristics of included studies are shown in table 1. Table 1 Characteristics of included Studies Author Year Study design Population Country Patient number Sex Men (%) Age, Mean, y CMBs Large artery atherosclerosis Quality Score Other principal findings associated with CMBs MRI Prevalence n (%) Stenosis>50% n (%) Mean cIMT mm (mean±SD) Method Definition Ochi 2009 Cross-sectional Outpatients Japan 443 36 67.1 3 T, GRE 22 (5.0%) – CMB(+):0.85±0.14 CMB(−):0.81±0.15 B-mode ultrasound Far wall of the CCA 11 items (8) Age, hypertension, baPWV Miwa 2011 Cross-sectional Outpatients, age ≥45 years Japan 431 52 69.3 1.5 T, GRE 65 (15.0%) – CMB(+):1.14±0.5 CMB(−):0.99±0.5 Ultrasound Near and far walls, bilateral distal CCA, carotid bifurcation, and ICA 11 items (8) Age, fasting glucose, hsCRP, IL-6, IL-18 Chung 2014 Cross-sectional Acute ischaemic stroke or TIA South Korea 834 58 66.8 3 T, SWI 335 (40.2%) 468 (28.1%) – CTA ICA Thick or contiguous calcification ≥50% of vessel diameter and/or ≥1 cm 11 items (8) Age, hypertension, hs-CRP Peng 2014 Cross-sectional Acute ischaemic stroke China 90 61 66.4 1.5 T, GRE 30 (33.3%) 46 (51.1%) – TCD, CDUS ICA,intracranial and extracranial arteries; A visible narrowing (>50%) or significant haemodynamic changes 11 items (8) SBP, DBP Song 2015 Cross-sectional Ischaemic stroke Korea 220 60 64.0 1.5 T, GRE 46 (20.9%) 93 (42.3%) – MRA, CTA, DSA Intracranial and extracranial arteries; Stenosis≥50% 11 items (8) Age Tabara 2015 Cross-sectional Healthy middle-aged to elderly individuals Japan 1387 39 67.0 3 T, GRE 92 (6.6%) – CMB(+):0.84±0.14 CMB(−):0.79±0.14 B-mode ultrasound Far wall, bilateral carotid arteries 11 items (8) Age, hypertension, T2DM Ding 2015 Prospective Population-based, age >65 Iceland 2512 42 74.6 1.5 T, GRE 463 (18.4%) – CMB(+):0.98±0.14 CMB(−):0.96±0.14 B-mode ultrasound Near and far walls, bilateral distal CCA NOS scale (9) Age, DBP, MAP, carotid arterial strain, DC, YEM Romero 2016 Cohort Framingham Offspring Study USA 1243 47 56.9 1.5 T, GRE 101 (8.2%) 20 (1.6%) – CDUS Distal CCA, carotid artery bulb and ICA; Stenosis≥50% NOS scale (9) – baPWV, brachial-to-ankle pulse wave velocity; CCA, common carotid artery; CDUS, carotid duplex ultrasound; cIMT, carotid intima-media thickness; CTA, CT angiography; DBP, diastoli

ffspring Study USA 1243 47 56.9 1.5 T, GRE 101 (8.2%) 20 (1.6%) – CDUS Distal CCA, carotid artery bulb and ICA; Stenosis≥50% NOS scale (9) – baPWV, brachial-to-ankle pulse wave velocity; CCA, common carotid artery; CDUS, carotid duplex ultrasound; cIMT, carotid intima-media thickness; CTA, CT angiography; DBP, diastoli c blood pressure; DC, distensibility coefficient; DSA, digital subtraction angiography; GRE, gradient-recalled echo sequences; hsCRP, high-sensitivity C reactive protein; ICA, internal carotid artery; IL, interleukin; MAP, mean arterial pressure; MRA, MR angiography; NOS scale, Newcastle–Ottawa scale; SBP, systolic blood pressure; SWI, susceptibility-weighted imaging; T2DM, type 2 diabetes mellitus; TCD, transcranial Doppler; TIA, transient ischaemic attack; YEM, Young elastic modulus. Figure 1 Flow chart of the study selection process. cIMT, carotid intima-media thickness; CMBs, cerebral microbleeds. Four studies including a total of 2387 patients were pooled in the meta-analysis of the association between large artery stenosis and CMBs.23–25 28 The mean age of the study participants was 61.4 years (range, 56.9 to 66.8 years). The overall prevalence of CMBs on MRIs was 21.4% (range, 8.2–40.2%). ORs (95% CI) and adjustments are reported in table 2. Only one study demonstrated a significant relationship between cerebral large artery stenosis and CMBs presence.23 We calculated the pooled OR by most adjusted OR using a random-effects model. The combined OR was 1.95 (95% CI 1.13 to 3.36) for the association between large artery stenosis >50% and CMBs (figure 2).

e 2. Only one study demonstrated a significant relationship between cerebral large artery stenosis and CMBs presence.23 We calculated the pooled OR by most adjusted OR using a random-effects model. The combined OR was 1.95 (95% CI 1.13 to 3.36) for the association between large artery stenosis >50% and CMBs (figure 2). Table 2 ORs of studies included in the meta-analysis Author Year OR 95% CI Adjustment Chung 2014 4.22 3.19 to 5.59 Unadjusted 2.86 2.01 to 4.08 Age, hypertension, previous stroke history, hs-CRP, total cholesterol, homocysteine and ICA calcification Peng 2014 2.11 0.86 to 5.19 Unadjusted Song 2015 1.06 0.55 to 2.05 Unadjusted Romero 2016 1.38 0.38 to 5.06 Age, sex, time to MRI 1.16 0.30 to 4.49 Age, sex, time to MRI, diabetes, smoking, hypertension, systolic blood pressure, prevalent cardiovascular disease and statin use 1.99 0.48 to 8.27 Age, sex, time to MRI, diabetes, smoking, hypertension, systolic blood pressure, prevalent cardiovascular disease and statin use, baseline carotid IMT hsCRP, high-sensitivity C reactive protein; ICA, internal carotid artery; IMT, intima-media thickness. Figure 2 Forest plot for the association between large artery stenosis and CMBs. CMBs, cerebral microbleeds. Risk of bias and heterogeneity between studies The I2 statistics and Cochran's Q test indicate evidence of substantial heterogeneity among studies (p=0.077, I2=56.1%). Furthermore, visual inspection of the funnel plot and Egger's test (p=0.48) indicated no evidence of publication bias.

Figure 2 Forest plot for the association between large artery stenosis and CMBs. CMBs, cerebral microbleeds. Risk of bias and heterogeneity between studies The I2 statistics and Cochran's Q test indicate evidence of substantial heterogeneity among studies (p=0.077, I2=56.1%). Furthermore, visual inspection of the funnel plot and Egger's test (p=0.48) indicated no evidence of publication bias. We conducted a sensitivity analysis by excluding a single study each time to explore the robustness of the combined results. The range of the combined ORs was from 1.41 (95% CI 0.86 to 2.32) to 2.70 (95% CI 1.96 to 3.72). The result showed no significant relationship between large artery stenosis >50% and CMBs presence when excluding Chung's study,23 with an OR 1.41 (95% CI 0.86 to 2.32). The I2 statistics and Cochran's Q test indicated the reduction of heterogeneity (p=0.423; I2=0.0%). This absolute difference may be due to different detection methods. First, Chung et al23 used CTA to detect the degree of intracranial ICA stenosis; however, three other studies mainly defined arterial stenosis by ultrasound scan, which might lead to heterogeneity. Second, the incidence rate of CMBs in Chung's study was much higher than in other studies. Chung et al assessed CMBs on susceptibility-weighted imaging (SWI), which is more sensitive. This may explain part of the heterogeneity.

dies mainly defined arterial stenosis by ultrasound scan, which might lead to heterogeneity. Second, the incidence rate of CMBs in Chung's study was much higher than in other studies. Chung et al assessed CMBs on susceptibility-weighted imaging (SWI), which is more sensitive. This may explain part of the heterogeneity. Association between cIMT and CMBs Four studies including 4773 patients provided data on the association between mean cIMT and CMBs were pooled in the meta-analysis.21 22 26 27 The mean age of the study participants was 71.2 years (range, 67 to 74.6 years), with fewer participants being men (41.3%; range, 39.4–52%). The overall prevalence of CMBs on MRI was 13.5% (range, 5% to 18.4%). A fixed-effects model was applied to estimate a pooled SMD of 0.20 (95% CI 0.11 to 0.28), which demonstrated that patients with CMBs were more likely to have a higher cIMT (p<0.001; figure 3). Three of the four studies demonstrated a statistically significant association between cIMT and cerebral microbleeds. Only one study found no association between cIMT and CMBs in healthy individuals free from a history of stroke.21 Figure 3 Forest plot for the association between cIMT and CMBs. cIMT, carotid intima-media thickness; CMBs, cerebral microbleeds; SMD, standardised mean difference.

Association between cIMT and CMBs Four studies including 4773 patients provided data on the association between mean cIMT and CMBs were pooled in the meta-analysis.21 22 26 27 The mean age of the study participants was 71.2 years (range, 67 to 74.6 years), with fewer participants being men (41.3%; range, 39.4–52%). The overall prevalence of CMBs on MRI was 13.5% (range, 5% to 18.4%). A fixed-effects model was applied to estimate a pooled SMD of 0.20 (95% CI 0.11 to 0.28), which demonstrated that patients with CMBs were more likely to have a higher cIMT (p<0.001; figure 3). Three of the four studies demonstrated a statistically significant association between cIMT and cerebral microbleeds. Only one study found no association between cIMT and CMBs in healthy individuals free from a history of stroke.21 Figure 3 Forest plot for the association between cIMT and CMBs. cIMT, carotid intima-media thickness; CMBs, cerebral microbleeds; SMD, standardised mean difference. Risk of bias and heterogeneity between studies The I2 statistics and Cochran's Q test indicate low heterogeneity (p=0.263; I2=24.7%). A funnel plot and Egger's test were used to assess the publication bias. The funnel plot seemed asymmetric, and the asymmetry was detected by Egger's test (coefficient=1.59, p=0.23). This demonstrated that asymmetry of the funnel was unlikely due to publication bias.

test indicate low heterogeneity (p=0.263; I2=24.7%). A funnel plot and Egger's test were used to assess the publication bias. The funnel plot seemed asymmetric, and the asymmetry was detected by Egger's test (coefficient=1.59, p=0.23). This demonstrated that asymmetry of the funnel was unlikely due to publication bias. A sensitivity analysis was performed through excluding a single study each time. The result by excluding one study showed an increased SMD borderline (SMD 0.326, 95% CI 0.17 to 0.48), with no heterogeneity (p=0.909; I2=0.0%), because of the greater weight in this study.27 However, this did not change the outcome statistically. The analysis was consistent when using a random-effects model.

ach time. The result by excluding one study showed an increased SMD borderline (SMD 0.326, 95% CI 0.17 to 0.48), with no heterogeneity (p=0.909; I2=0.0%), because of the greater weight in this study.27 However, this did not change the outcome statistically. The analysis was consistent when using a random-effects model. Discussion CSVD is an important public health problem that has attracted increasing attention, and may coexist with large artery atherosclerosis.9 The aim of this systematic review and meta-analysis was to provide evidence that biomarkers of large artery atherosclerosis, such as greater cIMT and arterial stenosis, may serve as clinical markers of subclinical haemorrhage-prone CSVD, reflected by CMBs. Previous studies have suggested that the incidence rate of CMBs in patients with large artery atherosclerosis was 9–41.3%.24 28 29 Different populations and detection methods may lead to different incidence of CMBs. Kwon et al reported 14 of 313 patients (9.0%) with intracranial arterial stenosis (ICAS) presented with CMBs on baseline MRI in the Stenting and Aggressive Medical Management for Preventing Recurrent Stroke in Intracranial Stenosis (SAMMPRIS) trial. However, they found no evidence linking CSVD with an increased risk of stroke in patients with ICAS.29

tients (9.0%) with intracranial arterial stenosis (ICAS) presented with CMBs on baseline MRI in the Stenting and Aggressive Medical Management for Preventing Recurrent Stroke in Intracranial Stenosis (SAMMPRIS) trial. However, they found no evidence linking CSVD with an increased risk of stroke in patients with ICAS.29 Large artery atherosclerosis and CSVD may share disease mechanisms, which are likely to be mediated via common vascular risk factors, such as older age, hypertension and diabetes mellitus.29 The prevalence of CMBs gradually increased with age.30 Advanced age and long-term hypertension may cause structural changes in microvessels, represented by fibrinoid necrosis and lipohyalinosis, thus increasing the risk of rupture and bleeding.31 Histopathological analyses have found hypertensive vasculopathy was specially associated with CMBs in the thalamus, basal ganglia, brainstem and cerebellum.32 Otherwise, Qiu et al33 indicated that diabetes was associated with markers of both cerebral macrovascular and microvascular diseases. Older age, hypertension and diabetes mellitus are risk factors that may lead to large artery atherosclerosis, thus giving rise to CMBs.

alamus, basal ganglia, brainstem and cerebellum.32 Otherwise, Qiu et al33 indicated that diabetes was associated with markers of both cerebral macrovascular and microvascular diseases. Older age, hypertension and diabetes mellitus are risk factors that may lead to large artery atherosclerosis, thus giving rise to CMBs. Previous studies have found that carotid stenosis, carotid plaque and cIMT were likely to be associated with WMH and cognitive decline in elderly people, even after accounting for vascular risk factors.7 34 In addition to the common risk factors, other mechanisms may explain the correlation. Factors that may influence cerebral blood flow especially at the origin of perforator arteries may also be involved. Gregg et al16 reported that the presence of incidental cortical CMBs is associated with significant and widespread reduction in resting-state cerebral blood flow (CBF), especially in the frontal, parietal and precuneus cortices. Moreover, Hashimoto et al17 found reduced CBF in the centrum semiovale in patients with five or more CMBs compared with those with <5 CMBs in patients with CSVD. Individuals with cortical or deep or infratentorial CMBs may be exposed to chronic cerebral hypoperfusion.16 17 35 Long-term hypoperfusion could accelerate the development of age-related ultrastructural aberrations of capillaries and cause blood-brain barrier (BBB) damage.36 37 Disruption of BBB has been suggested as a main initial pathogenic mechanism in CSVD.22 24 However, another possibility is that both CMBs and hypoperfusion are markers of small vessel diseases, and that no causal relationship between them should be considered. Further studies are needed before a conclusion can be drawn.

Disruption of BBB has been suggested as a main initial pathogenic mechanism in CSVD.22 24 However, another possibility is that both CMBs and hypoperfusion are markers of small vessel diseases, and that no causal relationship between them should be considered. Further studies are needed before a conclusion can be drawn. In this meta-analysis, we found a significant association between large artery stenosis and microbleeds in the included studies, with an OR 1.95 (95% CI 1.13 to 3.36). The results from the Framingham Heart Study demonstrated that carotid stenosis ≥25% was associated with presence of CMBs overall (OR 2.20, 95% CI 1.10 to 4.40), especially at deep and mixed locations (OR 3.60, 95% CI 1.23 to 10.5). Paradoxically, the study found that carotid stenosis ≥50% was not associated with CMBs. The study observed carotid stenosis ≥50% in only 1.5% of 1243 participants and the incidence rate of CMBs was 8.3%, which might be an underestimate.28 The evidence relating arterial stenosis to CMBs was inconsistent. This may be attributed to different inclusion criteria, different sample size, differences in demographic characteristics and different methodology for cerebral artery stenosis estimation. Some of the included studies defined arterial stenosis by ultrasonography, which may limit accuracy. In our study, we found evidence that arterial stenosis >50% was related to the incidence of CMBs. Large artery stenosis may lead to hypoperfusion, thus resulting in CMBs. Further research is needed for confirmation.

on. Some of the included studies defined arterial stenosis by ultrasonography, which may limit accuracy. In our study, we found evidence that arterial stenosis >50% was related to the incidence of CMBs. Large artery stenosis may lead to hypoperfusion, thus resulting in CMBs. Further research is needed for confirmation. There may be other potential mechanisms that cause CMBs in patients with large artery atherosclerosis. CIMT is a non-invasive ultrasound marker of early atherosclerosis, and is increasingly used as a predictor of future clinical cardiovascular events including myocardial infarction and stroke.38–40 Previous studies considered cIMT as a marker of large-artery damage rather than CSVD. The Second Manifestations of Arterial disease study showed that mean cIMT was greater in patients with large vessel disease (1.08 mm) than in those with small vessel disease (0.92 mm) (SMD 0.11 mm, 95% CI 0.05 to 0.18).41 However, new research showed that patients with CSVD had greater cIMT compared with normal participants.42 The present systematic review and meta-analysis identified four studies focusing on the relationship between cIMT and CMBs. The result demonstrated that patients with CMBs were more likely to have a greater cIMT (SMD 0.20, 95% CI 0.11 to 0.28). Ding et al27 conducted a prospective population-based cohort study that included individuals aged >65 years without dementia and found that an increase in mean cIMT as a marker of arterial atherosclerosis was associated with an increased risk of CMBs, especially in the deep and infratentorial brain regions. All these findings suggest that there is a significant relationship between higher cIMT and CMBs risk.

s aged >65 years without dementia and found that an increase in mean cIMT as a marker of arterial atherosclerosis was associated with an increased risk of CMBs, especially in the deep and infratentorial brain regions. All these findings suggest that there is a significant relationship between higher cIMT and CMBs risk. One of the included studies in this review also explored the relationship between inflammation and CMBs in addition to increased cIMT. The results indicated that higher levels of circulating inflammatory markers, such as high-sensitivity C reactive protein (hsCRP), interleukin-6 (IL-6) and IL-18 were associated with CMBs, suggesting the involvement of inflammation.22 Inflammation is also implicated in the pathogenesis and development of atherosclerosis. Several studies have demonstrated that high levels of some inflammatory cytokines are important determinants in the pathogenesis of increased cIMT. Patients with higher cIMT have increased circulating levels of fibrinogen, tumour necrosis factor α, white cell count, hsCRP and IL-6.43 44 Moreover, Chung et al45 detected underlying intracranial atheroma in 60% of patients with lacunar infarction by high-resolution MRI and found potential intraplaque inflammation, suggesting the involvement of inflammation in both cerebral artery atherosclerosis and CSVD. We speculated that the link between CMBs and large artery atherosclerosis might be inflammation. Further studies are needed to confirm these findings.

arction by high-resolution MRI and found potential intraplaque inflammation, suggesting the involvement of inflammation in both cerebral artery atherosclerosis and CSVD. We speculated that the link between CMBs and large artery atherosclerosis might be inflammation. Further studies are needed to confirm these findings. There were some limitations in our study. First, only studies published in English were included, which may introduce publication bias. Second, the participants in the cross-sectional studies were consecutive patients, which may have introduced selection bias. Third, some of the studies were subject to bias because they did not involve blinded assessment of large artery atherosclerosis or blinded identification of CMBs. Fourth, the included studies varied in many aspects, such as study population, which resulted in a wide range of CMBs incidence rate (5% to 40.2%). Fifth, only one study demonstrated a significant relationship between cerebral large artery stenosis >50% and CMBs,23 using CTA to define arterial stenosis. The included studies used CDUS or transcranial Doppler to detect arterial stenosis may introduce bias due to limited accuracy. Moreover, most of the included studies assessed CMBs using T2*GRE or 1.5 T MRI machines. The use of higher-field MRI and SWI are known to increase CMBs detection. The variety of different techniques used in the assessment of artery atherosclerosis or CMBs in the included studies should be considered a source of heterogeneity. The accuracy of the methodology needs to be established in future studies.

Introduction Homocysteine (Hcy) is a sulfur-containing amino acid. The metabolism of Hcy is influenced by folic acid and vitamin B12, deficiencies of which can lead to high Hcy levels in the blood.1 2 In 1969, McCully3 made initial observations linking plasma Hcy concentrations and arteriosclerotic vascular disease. Many subsequent studies have shown that high Hcy levels may present a risk factor for atherosclerosis;4–6 however, there is little consensus among epidemiological investigations or case–control studies.6–12 These generally involved using participants from economically developed areas,6–9 where most participants demonstrated normal or slightly elevated Hcy levels.6–11 As a result, it remains unclear if high Hcy levels are related to atherosclerosis onset. Lvliang city is one of the poorest areas of Shanxi Province, China. It has the highest incidence of neural tube defects (NTDs) in the world.13 NTDs may be related to local environmental conditions and eating habits.14 The use of folic acid has significantly reduced the incidence of NTDs in pregnant women in Lvliang city.14 Therefore, it was hypothesised that blood folic acid levels may be low while the levels of Hcy may be high among individuals aged 55 years and older. The aim of this cross-sectional study was to describe the relationship between carotid atherosclerosis, as assessed by ultrasonography, and serum concentrations of Hcy in this region. Methods This study was approved by the Research Ethics Committee of the Shanxi Medical University. Written informed consents were obtained from the patients.

Lvliang city is one of the poorest areas of Shanxi Province, China. It has the highest incidence of neural tube defects (NTDs) in the world.13 NTDs may be related to local environmental conditions and eating habits.14 The use of folic acid has significantly reduced the incidence of NTDs in pregnant women in Lvliang city.14 Therefore, it was hypothesised that blood folic acid levels may be low while the levels of Hcy may be high among individuals aged 55 years and older. The aim of this cross-sectional study was to describe the relationship between carotid atherosclerosis, as assessed by ultrasonography, and serum concentrations of Hcy in this region. Methods This study was approved by the Research Ethics Committee of the Shanxi Medical University. Written informed consents were obtained from the patients. Participants From August to November 2012, a team of researchers (physicians and senior medical students) from the First Hospital of Shanxi Medical University travelled to Lvliang, Shanxi Province of China to study villagers born before 1 January 1958 who resided at least 2 months of the year in one of the four towns located there; namely, Xiajiaying, Kangchen, Gaojiagou or Caijiaya. A population of 9286 people aged 55 years and older was provided by the household registry department in each town. Of the 3005 participants who volunteered to complete the study questionnaires, 2304 completed carotid ultrasound examinations and blood tests. Thirteen participants were excluded because of incomplete data, and the remaining 2291 were evaluated, of which 1016 were men and 1275 were women.

istry department in each town. Of the 3005 participants who volunteered to complete the study questionnaires, 2304 completed carotid ultrasound examinations and blood tests. Thirteen participants were excluded because of incomplete data, and the remaining 2291 were evaluated, of which 1016 were men and 1275 were women. A survey was carried out in the village clinics. Two days before the survey, villagers were informed of the survey through posters, radio advertising and telephone. For those volunteers who agreed to participate in the study, a face-to-face structured questionnaire was administered by the medical research team. After signing an informed consent form, participants underwent carotid ultrasound examinations and blood tests. Participants for statistical analysis included those who completed the questionnaire, carotid ultrasound examination and blood tests.

ce structured questionnaire was administered by the medical research team. After signing an informed consent form, participants underwent carotid ultrasound examinations and blood tests. Participants for statistical analysis included those who completed the questionnaire, carotid ultrasound examination and blood tests. Data collection The collected survey data included demographic information and cerebrovascular disease risk factors. Demographic characteristics included gender, date of birth, education, occupation and marital status. Cerebrovascular disease risk factors included history of smoking, hypertension and diabetes. Participants’ blood pressure (BP), height and weight were recorded. A qualified nurse took a fasting venous blood sample. Information regarding the participants’ medical history was self-reported or obtained through medical records. For the purposes of this study, history of smoking was defined as smoking continuously for 6 months or more and within 30 days prior to the survey.15 History of drinking was defined as a daily alcohol intake of more than 25 mg (80 mL of liquor, 200 mL of wine and 600 mL of beer).16

or obtained through medical records. For the purposes of this study, history of smoking was defined as smoking continuously for 6 months or more and within 30 days prior to the survey.15 History of drinking was defined as a daily alcohol intake of more than 25 mg (80 mL of liquor, 200 mL of wine and 600 mL of beer).16 For each participant, the same researcher measured the BP twice, between 08:00 and 10:00, with the participant seated and using the same calibrated sphygmomanometer. Participants were asked to rest for at least 15 min, not to smoke and to empty their bladder up to 30 min beforehand. Average values of the systolic and diastolic BP (SBP and DBP, respectively) were used for analysis. Hypertension was defined as SBP≥140 mm Hg or DBP≥90 mm Hg, according to the 2010 Chinese guidelines for the management of hypertension. Height and weight were measured indoors. Body mass index (BMI; kg/m2) was calculated as an index of obesity.16 The WHO cut-off points of BMI<25.0 for non-obese and ≥25.0 for obese adults in Asian populations were used.

0 mm Hg or DBP≥90 mm Hg, according to the 2010 Chinese guidelines for the management of hypertension. Height and weight were measured indoors. Body mass index (BMI; kg/m2) was calculated as an index of obesity.16 The WHO cut-off points of BMI<25.0 for non-obese and ≥25.0 for obese adults in Asian populations were used. Biochemical determinations Blood was drawn from the antecubital vein, to measure fasting blood glucose (FBG). Blood samples were centrifuged within 1 hour and frozen at −70°C. Blood samples were analysed at the First Hospital of Shanxi Medical University clinical laboratory. A Beckman UniCel DxC 800 Synchron Clinical System Analyzer (Beckman Coulter) was used to detect total cholesterol (TC), high-density lipoprotein (HDL) cholesterol (Immuno FS, DiaSys), triglycerides (TGs; Beckman Coulter) and uric acid (UA; Beckman Coulter). Low-density lipoprotein (LDL) cholesterol levels were calculated using the Friedewald equation for participants who had TG levels <400 mg/dL. Hcy was measured using an enzyme cycling method with a Beckman UniCel DxC 800 Synchron Clinical System Analyzer (Beckman Coulter). All the participants were divided into four groups according to the Hcy level of four quantiles.

Biochemical determinations Blood was drawn from the antecubital vein, to measure fasting blood glucose (FBG). Blood samples were centrifuged within 1 hour and frozen at −70°C. Blood samples were analysed at the First Hospital of Shanxi Medical University clinical laboratory. A Beckman UniCel DxC 800 Synchron Clinical System Analyzer (Beckman Coulter) was used to detect total cholesterol (TC), high-density lipoprotein (HDL) cholesterol (Immuno FS, DiaSys), triglycerides (TGs; Beckman Coulter) and uric acid (UA; Beckman Coulter). Low-density lipoprotein (LDL) cholesterol levels were calculated using the Friedewald equation for participants who had TG levels <400 mg/dL. Hcy was measured using an enzyme cycling method with a Beckman UniCel DxC 800 Synchron Clinical System Analyzer (Beckman Coulter). All the participants were divided into four groups according to the Hcy level of four quantiles. Measurement of carotid atherosclerosis Carotid artery ultrasound scan was performed by a qualified sonographer, using a colour Doppler ultrasound scanner (Logic E, American GE). Participants were inspected in the supine position, with the head upright and the anterior portion of the neck fully exposed. The proximal internal carotid artery (ICA), ICA bulb and distal ICA were detected in sequence. The distance from the ICA lumen-intima border to the media-outer border was measured as the intima media thickness (IMT), in longitudinal images at the diastolic phase. The proximal part of the vessel wall to a plaque was measured, where there was a plaque. According to Mannheim carotid IMT consensus17 and carotid stenosis ultrasound diagnostic criteria,18 IMT≥1.0 mm was defined as IMT thickening. IMT≥1.5 mm or above the lumen was defined as a plaque. Participants showing one or several stenotic lesions, with a stenosis rate of <50% and no change in blood flow velocity (BFV), were not recorded. A stenosis rate of more than 50% and altered BFV was recorded. We defined IMT thickening and or carotid plaque(s) as carotid atherosclerosis.

e lumen was defined as a plaque. Participants showing one or several stenotic lesions, with a stenosis rate of <50% and no change in blood flow velocity (BFV), were not recorded. A stenosis rate of more than 50% and altered BFV was recorded. We defined IMT thickening and or carotid plaque(s) as carotid atherosclerosis. Statistical analysis Data were checked, verified and recorded using Epidata software (V.3.1). Results are reported as means±SDs. Quantitative variables between groups were compared by t-test, approximate t-test and variance analysis, whereas classifications of variable rate were compared using the χ2 test. Logistic regression was used to estimate ORs and 95% CIs. Statistical significance was defined as a p value <0.05. All statistical analyses were performed using SPSS (V.13.0). Results The local housing registry department provided a list of 9286 residents. Of the 3005 participants who volunteered to complete the study questionnaires, 2304 completed carotid ultrasound examinations and blood tests. Incomplete information was provided by 13 participants. Of the 2291 participants included in the statistical analysis, 1016 were men and 1275 were women. Table 1 shows the age and sex distribution of the targeted population and participants, as well as the participation rates. Table 1 Sex and age distribution of the targeted population and participants, and participation rates

Results The local housing registry department provided a list of 9286 residents. Of the 3005 participants who volunteered to complete the study questionnaires, 2304 completed carotid ultrasound examinations and blood tests. Incomplete information was provided by 13 participants. Of the 2291 participants included in the statistical analysis, 1016 were men and 1275 were women. Table 1 shows the age and sex distribution of the targeted population and participants, as well as the participation rates. Table 1 Sex and age distribution of the targeted population and participants, and participation rates N Per cent Response rate (%) Targeted population 9286 100.0 – Female 4830 52.0 – Age 55–64 4997 53.8 – 65–74 2567 27.6 – 75+ 1722 18.6 – Participants 2291 100.0 24.7 Female 1275 55.7 26.4 Age 55–64 1302 56.8 26.1 65–74 693 30.2 27.0 75+ 296 13.0 17.2 The overall average detection rate of carotid atherosclerotic lesions was 76.3% and the detection rate of plaque was 48.6%. We observed that 84.1% of men and 70.0% of women showed atherosclerotic lesions. The characteristics of all participants involved are presented in table 2. Table 2 Research overview of study participants N=2291 Men N=1016 Women N=1275 Age (year) 64.6±7.4 65.4±7.8 64.0±7.0 Homocysteine (µmol/L) 24.7±18.0 29.3±20.8 21.0±14.5 Folic acid (nmol/L) 11.2±13.0 10.6±14.7 11.6±11.5 Atherosclerosis 1747 (76.3%) 854 (84.1%) 893 (70.0%) Thickening of IMT only 633 (27.6%) 265 (26.1%) 368 (28.9%) Plaque formation only 982 (42.9%) 514 (50.6%) 468 (36.7%) Stenosis or occlusion 132 (5.8%) 75 (7.4%) 57 (4.5%) IMT, intima media thickness.

9.3±20.8 21.0±14.5 Folic acid (nmol/L) 11.2±13.0 10.6±14.7 11.6±11.5 Atherosclerosis 1747 (76.3%) 854 (84.1%) 893 (70.0%) Thickening of IMT only 633 (27.6%) 265 (26.1%) 368 (28.9%) Plaque formation only 982 (42.9%) 514 (50.6%) 468 (36.7%) Stenosis or occlusion 132 (5.8%) 75 (7.4%) 57 (4.5%) IMT, intima media thickness. Table 3 shows the background factor classified by the presence of carotid atherosclerosis. The prevalence of men, Hcy, age, SBP, TG, UA, FBG, smoking and drinking was higher in the groups with carotid atherosclerosis. Table 4 shows the background factor classified by the presence of carotid plaque(s). The prevalence of men, Hcy, age, SBP, TG, UA, smoking, drinking and education was higher in the groups with carotid plaque(s). Table 3 Comparison of demographic characteristics in participants with or without carotid atherosclerosis

Table 3 shows the background factor classified by the presence of carotid atherosclerosis. The prevalence of men, Hcy, age, SBP, TG, UA, FBG, smoking and drinking was higher in the groups with carotid atherosclerosis. Table 4 shows the background factor classified by the presence of carotid plaque(s). The prevalence of men, Hcy, age, SBP, TG, UA, smoking, drinking and education was higher in the groups with carotid plaque(s). Table 3 Comparison of demographic characteristics in participants with or without carotid atherosclerosis Carotid atherosclerosis No N=544 Yes N=1747 p Value Age (year) 61.9±6.0 65.4±7.5 0.000 BMI (kg/m2) 24.4±3.5 23.7±3.3 0.000 SBP (mm Hg) 136.9±18.6 140.3±20.8 0.000 DBP (mm Hg) 83.6±10.7 83.3±11.3 0.602 Total cholesterol (mmol/L) 4.3±1.2 4.4±1.1 0.538 Triglycerides (mmol/L) 1.6±1.1 1.5±1.1 0.004 HDL-C (mmol/L) 1.2±0.4 1.2±0.4 0.110 LDL-C (mmol/L) 2.5±0.9 2.5±0.8 0.109 Uric acid (µmol/L) 241.3±71.3 253.8±74.0 0.001 Homocysteine (µmol/L) 22.4±16.0 25.4±18.6 0.000 FBG (mmol/L) 5.2±1.2 5.4±1.5 0.000 Hypertension (%) 213 (39.2%) 683 (39.1%) 0.980 Diabetes mellitus (%) 32 (5.9%) 142 (8.1%) 0.084 Dyslipidaemia (%) 66 (12.1%) 131 (7.5%) 0.001 Family history of stroke (%) 44 (8.1%) 117 (6.7%) 0.268 Current smoker (%) 164 (30.1%) 810 (46.4%) 0.000 Drink (%) 34 (6.3%) 167 (9.6%) 0.017 Education—junior high school (%) 110 (20.2%) 336 (19.2%) 0.611 Sex (male, %) 162 (29.8%) 854 (48.9%) 0.000 BMI, body mass index; DBP, diastolic blood pressure; FBG, fasting blood glucose; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure.

(9.6%) 0.017 Education—junior high school (%) 110 (20.2%) 336 (19.2%) 0.611 Sex (male, %) 162 (29.8%) 854 (48.9%) 0.000 BMI, body mass index; DBP, diastolic blood pressure; FBG, fasting blood glucose; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure. Table 4 Comparison of demographic characteristics in participants with or without carotid plaque(s) Carotid plaque(s) No N=1177 Yes N=1114 p Value Age (year) 62.8±6.5 66.5±7.8 0.000 BMI (kg/m2) 24.1±3.4 23.6±3.4 0.000 SBP (mmHg) 137.7±19.6 141.3±21.0 0.000 DBP (mmHg) 83.3±11.0 83.4±11.3 0.877 Total cholesterol (mmol/L) 4.3±1.1 4.4±1.1 0.012 Triglycerides (mmol/L) 1.6±1.1 1.4±1.0 0.004 HDL-C (mmol/L) 1.2±0.4 1.2±0.4 0.176 LDL-C (mmol/L) 2.4±0.8 2.6±0.9 0.000 Uric acid(µmol/L) 243.5±70.7 258.6±75.8 0.000 Homocysteine (µmol/L) 23.2±17.2 26.2±18.7 0.000 FBG (mmol/L) 5.3±1.4 5.4±1.5 0.183 Hypertension (%) 449 (38.1%) 447 (40.1%) 0.177 Diabetes mellitus (%) 71 (6.0%) 103 (9.2%) 0.002 Dyslipidaemia (%) 116 (9.9%) 81 (7.3%) 0.016 Family history of stroke (%) 93 (7.9%) 68 (6.1%) 0.055 Current smoker (%) 417 (35.4%) 557 (50.0%) 0.000 Drink (%) 91 (7.7%) 110 (9.9%) 0.041 Education—junior high school (%) 249 (21.2%) 197 (17.7%) 0.020 Sex (male, %) 427 (36.3%) 589 (52.9%) 0.000 BMI, body mass index; DBP, diastolic blood pressure; FBG, fasting blood glucose; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure.

(9.9%) 0.041 Education—junior high school (%) 249 (21.2%) 197 (17.7%) 0.020 Sex (male, %) 427 (36.3%) 589 (52.9%) 0.000 BMI, body mass index; DBP, diastolic blood pressure; FBG, fasting blood glucose; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure. Participants were divided into four groups according to their Hcy concentrations: <14.49, 14.49–19.44, 19.44–28.30 and ≥28.30 µmol/L. The prevalence rates of carotid atherosclerosis in these four groups were 70.33%, 76.47%, 78.09% and 80.14%, respectively (p=0.001). The prevalence rates of carotid plaque(s) in these four groups were 40.7%, 45.8%, 52.8% and 55.2%, respectively (p=0.000). There were differences between the groups in terms of age, SBP, TC, TG, HDL, LDL, UA, gender, smoking, diabetes, hypertension and family history of stroke (p<0.05; table 5). Table 5 Baseline characteristics and laboratory profiles by quartile of Hcy level

Participants were divided into four groups according to their Hcy concentrations: <14.49, 14.49–19.44, 19.44–28.30 and ≥28.30 µmol/L. The prevalence rates of carotid atherosclerosis in these four groups were 70.33%, 76.47%, 78.09% and 80.14%, respectively (p=0.001). The prevalence rates of carotid plaque(s) in these four groups were 40.7%, 45.8%, 52.8% and 55.2%, respectively (p=0.000). There were differences between the groups in terms of age, SBP, TC, TG, HDL, LDL, UA, gender, smoking, diabetes, hypertension and family history of stroke (p<0.05; table 5). Table 5 Baseline characteristics and laboratory profiles by quartile of Hcy level First quartile N=573 Second quartile N=578 Third quartile N=566 Fourth quartile N=574 p Value Age (year) 62.9±6.6 64.2±7.2 65.3±7.3 65.9±7.9 0.000 BMI (kg/m2) 23.7±3.6 23.8±3.2 23.9±3.5 24.0±3.3 0.477 SBP (mm Hg) 138.4±20.4 138.8±20.9 140.2±20.5 140.5±19.6 0.219 Triglycerides (mmol/L) 1.4±1.1 1.5±1.0 1.6±1.2 1.5±1.0 0.004 HDL-C (mmol/L) 1.1±0.4 1.2±0.4 1.2±0.4 1.2±0.4 0.000 LDL-C (mmol/L) 2.3±0.9 2.6±0.9 2.6±0.9 2.6±0.8 0.000 Uric acid (µmol/L) 211.5±72.5 253.4±66.5 265.4±66.8 273.2±72.5 0.000 FBG (mmol/L) 5.4±1.5 5.4±1.8 5.4±1.3 5.3±1.3 0.186 Current smoker (%) 196 (34.2%) 221 (38.2%) 245 (43.3%) 312 (54.4) 0.000 Drink (%) 39 (6.8%) 52 (9.0%) 52 (9.2%) 57 (9.9%) 0.273 Dyslipidaemia (%) 53 (9.3%) 41 (7.1%) 50 (8.8%) 53 (9.2%) 0.511 Education—junior high school (%) 116 (20.2%) 114 (19.7%) 98 (17.3%) 118 (20.6%) 0.502 Sex (male, %) 177 (30.9%) 220 (38.1%) 261 (46.1) 358 (62.4) 0.000 Atherosclerosis (%) 403 (70.3%) 442 (76.5%) 442 (78.1%) 460 (80.1%) 0.001 Plaque(s) (%) 233 (40.7%) 265 (45.8%) 299 (52.8%) 317 (55.2%) 0.000 Continuous variables are shown as mean±SD. Quantitative variables between groups were compared by t-test, approximate t-test or variance analysis, whereas classification of variable rate was compared by χ2 test.

42 (76.5%) 442 (78.1%) 460 (80.1%) 0.001 Plaque(s) (%) 233 (40.7%) 265 (45.8%) 299 (52.8%) 317 (55.2%) 0.000 Continuous variables are shown as mean±SD. Quantitative variables between groups were compared by t-test, approximate t-test or variance analysis, whereas classification of variable rate was compared by χ2 test. BMI, body mass index; DBP, diastolic blood pressure; FBG, fasting blood glucose; Hcy, homocysteine; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure. Logistic regression was used with carotid atherosclerosis as the dependent variable. Hcy quartiles were modelled with indicator variables to represent the three highest quartiles. The relative risk of the carotid atherosclerosis for each quartile as compared with the risk for the lowest quartile was estimated as the OR derived from the logistic regression coefficients. The OR of carotid atherosclerosis was higher in the second, third and fourth quartiles than in the first quartile by 1.371-fold, 1.504-fold and 1.702-fold (p<0.05), respectively. After further adjusting for age and gender, the OR of carotid atherosclerosis in the third and fourth quartiles of Hcy were 1.219 (95% CI 0.922 to 1.612) and 1.156 (95% CI 0.859 to 1.555; p>0.05; table 6), respectively. Table 6 ORs (95% CIs) for the presence of carotid atherosclerosis by quartile of Hcy level

Logistic regression was used with carotid atherosclerosis as the dependent variable. Hcy quartiles were modelled with indicator variables to represent the three highest quartiles. The relative risk of the carotid atherosclerosis for each quartile as compared with the risk for the lowest quartile was estimated as the OR derived from the logistic regression coefficients. The OR of carotid atherosclerosis was higher in the second, third and fourth quartiles than in the first quartile by 1.371-fold, 1.504-fold and 1.702-fold (p<0.05), respectively. After further adjusting for age and gender, the OR of carotid atherosclerosis in the third and fourth quartiles of Hcy were 1.219 (95% CI 0.922 to 1.612) and 1.156 (95% CI 0.859 to 1.555; p>0.05; table 6), respectively. Table 6 ORs (95% CIs) for the presence of carotid atherosclerosis by quartile of Hcy level OR (95% CI) Hcy quartile Crude Model 2 Model 3 First quartile 1.00 1.00 1.00 Second quartile 1.371 (1.054 to 1.783) 1.296 (0.992 to 1.692) 1.253 (0.957 to 1.640) p=0.019 p=0.057 p=0.101 Third quartile 1.504 (1.150 to 1.966) 1.334 (1.014 to 1.755) 1.219 (0.922 to 1.612) p=0.003 p=0.040 p=0.164 Fourth quartile 1.702 (1.296 to 2.235) 1.465 (1.107 to 1.938) 1.156 (0.859 to 1.555) p=0.000 p=0.007 p=0.339 Reference group: first quartile. Model 2: adjusted for age, model 3: adjusted for age and gender. Hcy, homocysteine.

OR (95% CI) Hcy quartile Crude Model 2 Model 3 First quartile 1.00 1.00 1.00 Second quartile 1.371 (1.054 to 1.783) 1.296 (0.992 to 1.692) 1.253 (0.957 to 1.640) p=0.019 p=0.057 p=0.101 Third quartile 1.504 (1.150 to 1.966) 1.334 (1.014 to 1.755) 1.219 (0.922 to 1.612) p=0.003 p=0.040 p=0.164 Fourth quartile 1.702 (1.296 to 2.235) 1.465 (1.107 to 1.938) 1.156 (0.859 to 1.555) p=0.000 p=0.007 p=0.339 Reference group: first quartile. Model 2: adjusted for age, model 3: adjusted for age and gender. Hcy, homocysteine. The relative risk of the carotid plaque(s) for each quartile as compared with the risk for the lowest quartile was estimated as the OR derived from the logistic regression coefficients. After controlling for other demographic variables (age, gender, current smoker, FBG, LDL and SBP) the OR of carotid plaque(s) in the third and fourth quartiles were 1.246 (95% CI 0.967 to 1.606) and 1.259 (95% CI 0.963 to 1.646; p>0.05; table 7), respectively. Table 7 ORs (95% CIs) for the presence of carotid plaque(s) by quartile of Hcy level

The relative risk of the carotid plaque(s) for each quartile as compared with the risk for the lowest quartile was estimated as the OR derived from the logistic regression coefficients. After controlling for other demographic variables (age, gender, current smoker, FBG, LDL and SBP) the OR of carotid plaque(s) in the third and fourth quartiles were 1.246 (95% CI 0.967 to 1.606) and 1.259 (95% CI 0.963 to 1.646; p>0.05; table 7), respectively. Table 7 ORs (95% CIs) for the presence of carotid plaque(s) by quartile of Hcy level OR (95% CI) Hcy quartile Crude Model 4 Model 5 Model 6 First quartile 1.00 1.00 1.00 1.00 Second quartile 1.235 (0.978 to 1.560) 1.162 (0.915 to 1.475) 1.117 (0.877 to 1.422) – p=0.076 p=0.218 p=0.369 – Third quartile 1.634 (1.292 to 2.066) 1.454 (1.142 to 1.852) 1.335 (1.044 to 1.709) 1.246 (0.967 to 1.606) p=0.000 p=0.002 p=0.022 p=0.089 Fourth quartile 1.800 (1.424 to 2.275) 1.580 (1.241 to 2.011) 1.324 (1.040 to 1.730) 1.259 (0.963 to 1.646) p=0.000 p=0.000 p=0.023 p=0.093 Reference group: first quartile. Model 4: adjusted for age, model 5: adjusted for age and gender, model 6: adjusted for age, gender, current smoker, FBG, LDL and SBP. FBG, fasting blood glucose; Hcy, homocysteine; LDL, low-density lipoprotein; SBP, systolic blood pressure.

OR (95% CI) Hcy quartile Crude Model 4 Model 5 Model 6 First quartile 1.00 1.00 1.00 1.00 Second quartile 1.235 (0.978 to 1.560) 1.162 (0.915 to 1.475) 1.117 (0.877 to 1.422) – p=0.076 p=0.218 p=0.369 – Third quartile 1.634 (1.292 to 2.066) 1.454 (1.142 to 1.852) 1.335 (1.044 to 1.709) 1.246 (0.967 to 1.606) p=0.000 p=0.002 p=0.022 p=0.089 Fourth quartile 1.800 (1.424 to 2.275) 1.580 (1.241 to 2.011) 1.324 (1.040 to 1.730) 1.259 (0.963 to 1.646) p=0.000 p=0.000 p=0.023 p=0.093 Reference group: first quartile. Model 4: adjusted for age, model 5: adjusted for age and gender, model 6: adjusted for age, gender, current smoker, FBG, LDL and SBP. FBG, fasting blood glucose; Hcy, homocysteine; LDL, low-density lipoprotein; SBP, systolic blood pressure. Discussion The average value of serum Hcy in this study was 24.7±18.0 µmol/L, which is much higher than the normal range of Hcy (5–15 µmol/L).19 The mean values of serum Hcy were 29.3±20.8 µmol/L for men and 21.0±14.5 µmol/L for women. These results are much higher than those of the general population in the western world (13 µmol/L for men, 10 µmol/L for women),19 also higher than China's coastal rural population20 and the Japanese rural general population9 (12.0 µmol/L for men, 9.6 µmol/L for women in China; 12.6 µmol/L for men, 9.8 µmol/L for women in Japan). The differences in plasma Hcy concentration described in various reports could be due to different inclusion criteria and ethnic or geographic differences. The average value for folic acid was 11.2±13.0 nmol/L, and 68.1% of folic acid levels measured was below 11 nmol/L. These data confirm our hypothesis that villagers aged 55 years and older have high levels of Hcy and low levels of folic acid.

be due to different inclusion criteria and ethnic or geographic differences. The average value for folic acid was 11.2±13.0 nmol/L, and 68.1% of folic acid levels measured was below 11 nmol/L. These data confirm our hypothesis that villagers aged 55 years and older have high levels of Hcy and low levels of folic acid. Carotid IMT and plaques are both measures of atherosclerosis, perhaps having different attributes or risk associations but still closely related.21 22 Sonographic characterisation of carotid IMT and plaque can be considered a measure of atherosclerosis quality. The overall detection rate of carotid atherosclerotic lesions was 76.3% and the detection rate of plaque was 48.6%. These detection rates are comparable to findings in a survey of the national population and related epidemiological findings.23–26 McCully3 defined the normal value for Hcy as 6–10 µmol/L for women and 8–12 µmol/L for men. The Framingham Heart Study suggested a value >14 µmol/L for participants with normal folic acid, vitamins B6 and B12.27 In most studies, hyperhomocysteinaemia was defined as ≥15 µmol/L.19 Serum Hcy values in our study were much higher than normal.19 The first quartile (serum Hcy<14.49 µmol/L) was normal and the fourth quartile (serum Hcy≥28.30 µmol/L) was even nearly onefold greater than the normal range.

ic acid, vitamins B6 and B12.27 In most studies, hyperhomocysteinaemia was defined as ≥15 µmol/L.19 Serum Hcy values in our study were much higher than normal.19 The first quartile (serum Hcy<14.49 µmol/L) was normal and the fourth quartile (serum Hcy≥28.30 µmol/L) was even nearly onefold greater than the normal range. Age is a strong risk factor of atherosclerotic diseases in Western countries28 29 and China. Cigarette smoking has been reported to be a risk factor of coronary artery disease and stroke in many studies. Furthermore, smoking is well known to be a risk factor of arteriosclerosis obliterans.30 In this study, the levels of Hcy were higher, the percentage of smokers was greater and the mean age of the participants was higher. After adjusting for age and gender, serum Hcy≥28.30 µmol/L did not increase the risk of carotid artery atherosclerosis compared with serum Hcy<14.49 µmol/L. After controlling for age, gender, current smoker and other known risk factors, elevated Hcy did not increase the risk of carotid plaque(s).

rticipants was higher. After adjusting for age and gender, serum Hcy≥28.30 µmol/L did not increase the risk of carotid artery atherosclerosis compared with serum Hcy<14.49 µmol/L. After controlling for age, gender, current smoker and other known risk factors, elevated Hcy did not increase the risk of carotid plaque(s). Hyperhomocysteinaemia has been identified as a strong predictor of cardiovascular disease, independent of classical atherothrombotic risk factors.31 However, not all disease predictors are risk factors in the strict definition. A biomarker commonly represents an early stage of the disease. Interventions aimed at optimising a biomarker may or may not be associated with reduced disease incidence depending on whether the treatment has an effect on a causal mechanism that underlies both the appearance of the biomarker as well as the occurrence of disease.32 The mechanisms by which Hcy induces atherosclerosis are largely unknown. Several biological mechanisms have been proposed to explain cardiovascular pathological changes associated with Hcy. These include (1) endothelial cell damage and impaired endothelial function; (2) dysregulation of cholesterol and TG biosynthesis; (3) stimulation of vascular smooth muscle cell proliferation; (4) thrombosis activation; and (5) activation of monocytes.5 Treatment that reduces the biomarker without affecting disease incidence is useless.

endothelial cell damage and impaired endothelial function; (2) dysregulation of cholesterol and TG biosynthesis; (3) stimulation of vascular smooth muscle cell proliferation; (4) thrombosis activation; and (5) activation of monocytes.5 Treatment that reduces the biomarker without affecting disease incidence is useless. Folic acid intake is associated with reduced risk of ischaemic stroke in some epidemiological studies but not in others.33 Most studies involving patients with established atherosclerotic vascular disease found no benefit in reducing Hcy by vitamin B complex therapy on clinical cardiovascular end points.33 A substudy of the VITAmins TO Prevent Stroke (VITATOPS) trial reported that vitamin B complex did not reduce the change in carotid IMT.10 Similarly, folic acid did not significantly affect carotid IMT in the Atherosclerosis and Folic Acid Supplementation Trial (ASFAST).11 These findings show that the role of Hcy in atherosclerosis and cardiovascular events may be overvalued. Taken together, these results and our data indicate that Hcy has no significant effect on early atherosclerosis preceding stroke and myocardial infarction.

osclerosis and Folic Acid Supplementation Trial (ASFAST).11 These findings show that the role of Hcy in atherosclerosis and cardiovascular events may be overvalued. Taken together, these results and our data indicate that Hcy has no significant effect on early atherosclerosis preceding stroke and myocardial infarction. Routine screening for hyperhomocysteinaemia among patients with a recent ischaemic stroke or transient ischaemic attack (TIA) is not indicated (class III; level of evidence C). In adults with a recent ischaemic stroke or TIA who are known to have mild-to-moderate hyperhomocysteinaemia, supplementation with folate, vitamin B6 and B12 safely reduces levels of Hcy but has not been shown to prevent stroke (class III; level of evidence B). These come from Guidelines for the Prevention of Stroke in Patients With Stroke and Transient Ischemic Attack 2014.34 The role of Hcy in stroke is declining. We are aware of the limitations of the present study. First, the response rate was low and this could lead to bias. Second, because of low awareness rate of hypertension and diabetes, the history of providing might influence the results. In conclusion, the mean levels of Hcy among villagers aged 55 years and older in a rural region of Lvliang City, China was higher than that reported in other parts of China; however, there was no significant correlation between high Hcy and carotid atherosclerosis. Elevated Hcy was found not to be an independent determinant of carotid atherosclerosis.

Hcy among villagers aged 55 years and older in a rural region of Lvliang City, China was higher than that reported in other parts of China; however, there was no significant correlation between high Hcy and carotid atherosclerosis. Elevated Hcy was found not to be an independent determinant of carotid atherosclerosis. The authors thank the villagers who took part in the survey made this study possible. They also thank local government and township health centres for their kind cooperation. The authors are grateful to members of the First Hospital of Shanxi Medical University clinical laboratory for blood testing. Contributors: YL and LW planned the study, performed data analyses and wrote the manuscript. YF helped plan the study and collected the data. WZ performed data analyses and contributed to the revision of the manuscript. XN supervised the study and contributed to the revision of the manuscript. Competing interests: None declared. Patient consent: Obtained. Ethics approval: Research Ethics Committee of the Shanxi Medical University. Provenance and peer review: Commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

Introduction The first description of obstructive sleep apnoea (OSA) was provided not by a physician or scientist, but by Charles Dickens in a series of papers titled ‘The posthumous papers of the Pickwick club’ in 1836, in which he described an obese boy who had excessive daytime somnolence, loud snoring and probably right heart failure.1 In 1889, Hill2 observed that upper airway obstruction contributed to ‘stupidity’ in children. In 1965, Gastaut et al3 in France performed nocturnal polygraphy of respiratory pauses in patients with obesity, and 10 years later, Lugaresi4 in Italy associated nocturnal apnoeas with loud snoring. The term OSAS was first defined after the 1972 Rimini conference.5 Around the same time, the first sleep disorders centre was established in Stanford University, California. Upper airway surgery was used for treatment of snoring in 1964 and was later used for treatment of OSA in 1981 in Japan.6 During the same period, nasal positive airway pressure was introduced by Sullivan for its treatment in 1981 and soon became the treatment of choice.7

established in Stanford University, California. Upper airway surgery was used for treatment of snoring in 1964 and was later used for treatment of OSA in 1981 in Japan.6 During the same period, nasal positive airway pressure was introduced by Sullivan for its treatment in 1981 and soon became the treatment of choice.7 Sleep apnoea and/or habitual snoring began to be recognised as independent risk factors for arterial hypertension (HTN),8 cardiac arrhythmias,9 coronary artery disease, myocardial infarction10 11 and ischaemic stroke12 only during late 20th century. The same research also recognised that patients with untreated sleep apnoea had higher risk of cardiovascular morbidity compared with patients with treated sleep apnoea.13 Recently, population studies have suggested that sleep apnoea may be a risk factor for vascular dementia.14 Sleep-disordered breathing (SDB) involves both obstructive and central breathing disorders, Cheyne-Stokes breathing and central sleep apnoea belonging to the latter category. This article reviews the current literature describing the inter-relationship between sleep apnoea and stroke; both direct and indirect evidences are discussed. Literature review was carried out by using PubMed search with words ‘Sleep apnea’ AND ‘stroke’, and personal article collection of the corresponding author.

y. This article reviews the current literature describing the inter-relationship between sleep apnoea and stroke; both direct and indirect evidences are discussed. Literature review was carried out by using PubMed search with words ‘Sleep apnea’ AND ‘stroke’, and personal article collection of the corresponding author. Sleep apnoea as vascular risk factor Sleep apnoea and autonomic nervous system Understanding the effects of sleep apnoea on autonomic nervous system (ANS) is important for better understanding of the subsequent sections. The body's biological clock—suprachiasmatic nucleus (SCN) has autonomous rhythmicity in its neuronal activity The autonomic body functions modulated by SCN include sympathetic–parasympathetic balance,15 hepatic glucose production and insulin sensitivity.16

important for better understanding of the subsequent sections. The body's biological clock—suprachiasmatic nucleus (SCN) has autonomous rhythmicity in its neuronal activity The autonomic body functions modulated by SCN include sympathetic–parasympathetic balance,15 hepatic glucose production and insulin sensitivity.16 During sleep, physiological changes in respiratory and cardiovascular activity are predominantly sleep-cycle dependent and mediated by autonomic control.17 During NREM, there is an increase in parasympathetic activity while during REM sleep, there is a decrease in parasympathetic activity accounting for increase in cardiovascular activity during the latter.18 Any arousal during sleep results in an increase in respiratory and cardiovascular activity.17 The intrinsic rhythmicity increases heart rate and blood pressure with tilting of sympathetic–parasympathetic balance towards the former immediately before waking, preparing the body for daily activities.19 The pathophysiological responses to OSA occur mainly in response to decrease in ion arterial oxygen tension and increase in arterial carbon dioxide tension. These provoke an increase in sympathetic nervous system activity causing peripheral vasoconstriction to divert blood flow to vital organs. At the same time, parasympathetic activity reduces myocardial activity and hence oxygen requirements. At the end of apnoeic episodes, there is an increase in blood pressure as myocardial function is restored. The vasoconstriction and changes in myocardial activity cause an increase in cardiac after load, while pulmonary vasoconstriction induced by hypoxia may contribute to right heart failure.17 20 Frequent and sustained episodes contribute to non-dipping of blood pressure at night17 and sensitisation of the hypoxic sensory response of carotid bodies which induces changes at genetic levels associated with increased oxidative stress.21 Microneurography has shown increased muscle sympathetic nervous activity at the termination of apnoeas in patients with OSA.22

on-dipping of blood pressure at night17 and sensitisation of the hypoxic sensory response of carotid bodies which induces changes at genetic levels associated with increased oxidative stress.21 Microneurography has shown increased muscle sympathetic nervous activity at the termination of apnoeas in patients with OSA.22 Use of CPAP improves sympathetic–parasympathetic balance in patients with moderate and severe sleep apnoea23 and improves heart rate variability.24 Increase in catecholamines has been found in urine and plasma in patients with OSA25 which can be lowered with therapies.26 In summary, patients with untreated sleep apnoea tend to have a heightened sympathetic activity and autonomic dysregulation which can benefit with management of OSA with either CPAP or tracheostomy. Sleep apnoea and HTN As discussed in an earlier section, there is increase in parasympathetic activity during slow-wave sleep, while sleep in general tends to have a parasympathetic predominance.27 A direct effect of autonomic dysfunction of OSA on blood pressure is an increase in its variability.28 There is evidence to suggest that this autonomic dysfunction extends into wakefulness.29

rease in parasympathetic activity during slow-wave sleep, while sleep in general tends to have a parasympathetic predominance.27 A direct effect of autonomic dysfunction of OSA on blood pressure is an increase in its variability.28 There is evidence to suggest that this autonomic dysfunction extends into wakefulness.29 In terms of its effect on nocturnal blood pressure, there is evidence of non-dipping blood pressure29 in patients with OSA. Hla et al30 studied the longitudinal association between SDB and incident non-dipping in 328 adults enrolled in the Wisconsin sleep cohort study, and found a dose–response increase in odds of systolic non-dipping with SDB. Non-dipping blood pressure itself is a well-known risk factor for cardiovascular disease. In addition to non-dipping of blood pressure, there is ample evidence to suggest that OSA contributes to daytime HTN as well. Nieto et al31 did a cross-sectional analysis on 6132 participants enrolled in the multicentre Heart Health Study and found independent correlation between increasing severity of OSA and raised systolic and diastolic HTN. Peppard et al32 found a similar dose–response relationship between severity of SDB and daytime HTN. Another large study conducted in Spain studied 1889 normotensive participants and found increase in risk of HTN in patients with OSA, which decreased in response to treatment with CPAP.33

ised systolic and diastolic HTN. Peppard et al32 found a similar dose–response relationship between severity of SDB and daytime HTN. Another large study conducted in Spain studied 1889 normotensive participants and found increase in risk of HTN in patients with OSA, which decreased in response to treatment with CPAP.33 OSA also has significant interaction with treatability of HTN. A study by Logan et al34 found a prevalence of sleep apnoea of 83% among participants with refractory HTN. Walia et al35 analysed data from the HeartBEAT study and found that patients with OSA had a fourfold OR of having resistant HTN as compared with participants with moderate OSA. Walia et al36 in their clinic-based effectiveness for blood pressure control study observed a significant reduction in blood pressure measures after treatment with CPAP in both groups containing participants with refractory and non-refractory HTN. Martinez-Garcia et al37 in the HIPARCO randomised controlled trial (RCT) studied 194 patients with refractory HTN and apnoea–hypopnoea index (AHI) ≥15. They found that CPAP treatment for 12 weeks was effective in reducing 24-hour mean and diastolic blood pressure, improving nocturnal blood pressure patterns. A meta-analysis of observational and RCTs by Iftikhar et al38 examined the effects of CPAP treatment in patients with refractory HTN and OSA. They found a favourable reduction in blood pressure with CPAP treatment in their patients. The effect size was also found to be larger than reported in patients with OSA without resistant HTN.

s of observational and RCTs by Iftikhar et al38 examined the effects of CPAP treatment in patients with refractory HTN and OSA. They found a favourable reduction in blood pressure with CPAP treatment in their patients. The effect size was also found to be larger than reported in patients with OSA without resistant HTN. In summary, OSA contributes to development of HTN, its refractoriness to medical therapy; treatment of OSA contributes to reduction in blood pressure. Sleep apnoea and atrial fibrillation Sleep apnoea has been linked to increase in prevalence, effect on treatment and aggravation of atrial fibrillation (AF) as described in this section. Prevalence of AF in patients with sleep apnoea has been found to be between 3% and 5% as compared with 1% in the general population.39 40 Mooe et al41 studied patients undergoing cardiac bypass surgery with polysomnography for the diagnosis of sleep apnoea and found both AHI and O2 desaturation index to be highly predictive of postoperative AF. The Sleep Heart Health Study found AF prevalence rates of 4.8% and 0.9% in participants with and without SDB. The same study also found the likelihood of arrhythmic events to be more common after hypopnoeic or apnoeic episodes.39 Severity of sleep apnoea has also been shown to influence the prevalence of AF.42 Under-reporting of excessive daytime sleepiness and low EDSS scores is a possible contributory factor to underdiagnosis of sleep apnoea in these patients.43 44

of arrhythmic events to be more common after hypopnoeic or apnoeic episodes.39 Severity of sleep apnoea has also been shown to influence the prevalence of AF.42 Under-reporting of excessive daytime sleepiness and low EDSS scores is a possible contributory factor to underdiagnosis of sleep apnoea in these patients.43 44 In patients undergoing management of AF who were followed up for 1 year, recurrence of AF was found to be present in 82% of the patients with untreated sleep apnoea compared with 42% in patients with treated sleep apnoea.45 In another study, procedural failure of AF ablation was predicted by presence of sleep apnoea and non-compliance with CPAP.46 The Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) registry found that participants with sleep apnoea were more likely to have more severe or disabling symptoms as compared with participants without OSA (22% vs 16%), and more likely to be receiving rhythm control therapy (35% vs 31%, respectively).47 In a reverse association, studies have found favourable effect of treatment of AF on central sleep apnoea,48 and a significant reduction in AHI in patients with SDB.49 Various effects of sleep apnoea on pathophysiology of AF have been described. Increased negative intrathoracic pressure during apnoeic episodes of OSA increase cardiac vagal output that enhances AF inducibility. This effect can be prevented by vagotomy and atropine.50 It has also been found to activate stretch-sensitive ion channels or fibrosis at anchoring regions of atria that are critical to AF induction.51 52

intrathoracic pressure during apnoeic episodes of OSA increase cardiac vagal output that enhances AF inducibility. This effect can be prevented by vagotomy and atropine.50 It has also been found to activate stretch-sensitive ion channels or fibrosis at anchoring regions of atria that are critical to AF induction.51 52 Obesity and the magnitude of nocturnal oxygen desaturation are independent risk factors for AF in individuals <65 years of age53 and OSA is a univariate predictor of AF (HR 2.18, 95% CI 1.34 to 3.54).54 So, sleep apnoea not only increases the risk of AF, it also increases the refractoriness of AF to treatment. In addition, the treatment of AF has a positive beneficial effect on central sleep apnoea. Sleep apnoea and stroke While sleep apnoea has been shown to indirectly increase the risk of stroke by its effect on vascular risk factors as aforementioned, it has also been independently associated with increased risk of stroke.

So, sleep apnoea not only increases the risk of AF, it also increases the refractoriness of AF to treatment. In addition, the treatment of AF has a positive beneficial effect on central sleep apnoea. Sleep apnoea and stroke While sleep apnoea has been shown to indirectly increase the risk of stroke by its effect on vascular risk factors as aforementioned, it has also been independently associated with increased risk of stroke. In the landmark study by Yaggi et al,55 investigators found an independent increase in risk of stroke and all-cause mortality with a HR of 2.24 in patients with AHI ≥35/hour. This risk remained elevated despite controlling for traditional stroke risk factors such as HTN, AF, smoking status, diabetes and hyperlipidaemia. A study of 394 patients aged 70–100 years old found an AHI≥30 associated with an increased risk of ischaemic stroke in an elderly non-institutionalised male population.56 In a study of 1189 patients, AHI≥20 was associated with an increase in the risk of having stroke over the next 4 years.57 The Sleep Heart Health Study helped link sleep apnoea with stroke.58 It found that men in the highest quartile of AHI (>19) had a HR of 2.86 for having stroke and even in the mild–moderate sleep apnoea category (AHI 5–25), each 1 unit increase in AHI increased the risk of stroke by 6%. In women, the same study found an increase in risk of stroke only in the severe sleep apnoea group (AHI>25). In a study by Marin et al,59 severe OSA significantly increased the risk of fatal and non-fatal cardiovascular events while CPAP treatment reduced the risk. In the Wisconsin sleep cohort study, there was a significant, high cardiovascular mortality risk with untreated SDB, independent of age, sex and body mass index (BMI).60 The American Heart Association recommends screening for OSA for stroke prevention and suggests that treatment might be reasonable.61

reduced the risk. In the Wisconsin sleep cohort study, there was a significant, high cardiovascular mortality risk with untreated SDB, independent of age, sex and body mass index (BMI).60 The American Heart Association recommends screening for OSA for stroke prevention and suggests that treatment might be reasonable.61 In conclusion, the various effects of sleep apnoea/SDB on cardiovascular risk factors and stroke as a risk factor itself are now well documented and should be routinely identified as risk factor for stroke in clinical practice. Management of sleep apnoea In this section, we will summarise the current management of sleep apnoea without going into details, which can be found in the references listed. Screening for sleep apnoea The clinical symptoms of sleep apnoea may include snoring, witnessed episodes of apnoea by family members, history of awakenings associated with a sensation of choking, nocturia, morning headaches, palpitations due to difficult to control AF, refractory HTN and excessive daytime sleepiness. While any of the above should trigger a search for sleep apnoea, the authors believe that the first screening should ideally happen in a primary care setting similar to other vascular risk factors such as HTN and obesity, prior to development of complications of sleep apnoea.

ctory HTN and excessive daytime sleepiness. While any of the above should trigger a search for sleep apnoea, the authors believe that the first screening should ideally happen in a primary care setting similar to other vascular risk factors such as HTN and obesity, prior to development of complications of sleep apnoea. The Epworth Sleepiness Scale (EPSS), Berlin apnoea questionnaire, STOP and STOP-Bang questionnaires are various clinical screening tools available. The Berlin apnoea questionnaire identifies patients at risk for OSA.62 STOP and STOP-Bang questionnaires have been studied and validated in presurgical patients.63 The EPSS quantifies subjective measures of excessive daytime sleepiness64 using a questionnaire format; the multiple sleep latency test (MSLT) can be used to objectively assess excessive daytime sleepiness.65 Another in-hospital tool that may be used for screening of sleep apnoea in patients with stroke is continuous overnight oximetry which will be discussed under ‘Sleep apnoea in the acute stroke’ subsection below. Diagnostic criteria OSA falls under the category of SDB which also includes central sleep apnoea and Cheyne-Stokes breathing. Once sleep apnoea is suspected, the diagnostic gold standard is an overnight polysomnogram.65 During polysomnography, various abnormal respiratory parameters have been defined and include the following: Apnoea—cessation of airflow ≥10 s. (with or without respiratory effort). Hypopnoea—a ≥30% reduction in airflow for at least 10 s that is accompanied by either ≥3% desaturation or an arousal.

Diagnostic criteria OSA falls under the category of SDB which also includes central sleep apnoea and Cheyne-Stokes breathing. Once sleep apnoea is suspected, the diagnostic gold standard is an overnight polysomnogram.65 During polysomnography, various abnormal respiratory parameters have been defined and include the following: Apnoea—cessation of airflow ≥10 s. (with or without respiratory effort). Hypopnoea—a ≥30% reduction in airflow for at least 10 s that is accompanied by either ≥3% desaturation or an arousal. Respiratory effort-related arousal (RERA): a partially obstructed breath that does not meet the criteria for hypopnoea but is associated with increased respiratory effort and arousal. Flow-limited breath: a partially obstructed breath identified by flattened inspiratory flow shape. Severity of OSA/hypopnoea (OSAHS) syndrome can be quantified using the following criteria: AHI—number of apnoeas and hypopnoeas per hour of sleep. Respiratory disturbance index—AHI+RERAs per hour of sleep. Mild OSAHS—AHI 5–14/hour. Moderate OSAHS—AHI 15–29/hour. Severe OSAHS—AHI≥30/hour. Therapies for sleep apnoea Treatment includes a recommendation to lose weight. Most patients who lose weight have less severe apnoea, but it is difficult to predict the amount of improvement associated with loss of a specified amount of weight. Nonetheless, weight loss via intensive lifestyle interventions should be encouraged as a treatment for mild-to-moderate OSA.

ecommendation to lose weight. Most patients who lose weight have less severe apnoea, but it is difficult to predict the amount of improvement associated with loss of a specified amount of weight. Nonetheless, weight loss via intensive lifestyle interventions should be encouraged as a treatment for mild-to-moderate OSA. Avoidance of precipitating factors, such as alcohol, smoking and hypnotic drugs is helpful. For patients in whom sleep apnoea occurs only when supine (positional OSA), training to avoid this sleeping position is often beneficial (positional therapy). The cornerstone of therapy is positive airway pressure ventilation. The air pressure splint acts to maintain airway patency during sleep and ameliorate the effects of sleep apnoea. Various PAP modes that are used include CPAP, bilevel-PAP and autoPAP. Oral appliances that can be used to ameliorate mild-to-moderate OSA include tongue retaining devices and soft palate lifting devices. Surgical therapies may be used for definitive anatomical obstructions contributing to OSA. Mandibular repositioning devices have been used for patients with mild-to-moderate OSA with PAP intolerance.66 Fully implantable hypoglossal nerve stimulating systems inducing electrical stimulation of the genioglossus muscle have been approved for nerve stimulation and prevention of pharyngeal collapse without arousing patients from sleep.67

devices have been used for patients with mild-to-moderate OSA with PAP intolerance.66 Fully implantable hypoglossal nerve stimulating systems inducing electrical stimulation of the genioglossus muscle have been approved for nerve stimulation and prevention of pharyngeal collapse without arousing patients from sleep.67 Sleep apnoea in acute stroke Epidemiology of sleep apnoea in acute stroke Respiratory changes are seen acutely after stroke and can be divided into sleep–wake cycle and SDB. The changes may vary with the location of the stroke. As mentioned in an earlier section, OSA is part of SDB which includes central sleep apnoea. About 50–70% of patients with stroke have SDB as defined by AHI≥10/hour with OSA being the most common pathology.68 Some studies indicate that during the first 5 days poststroke central sleep apnoea predominates.69

f the stroke. As mentioned in an earlier section, OSA is part of SDB which includes central sleep apnoea. About 50–70% of patients with stroke have SDB as defined by AHI≥10/hour with OSA being the most common pathology.68 Some studies indicate that during the first 5 days poststroke central sleep apnoea predominates.69 The frequency of OSA itself has been reported to be between 38% when measured as AHI>20/hour to 72% when measured as AHI>5/hour in a meta-analysis performed by Johnson and Johnson only 7% of SDB was central apnoea. Males had a higher percentage of SDB (AHI>10) than females (65% vs 48%, respectively). Patients with recurrent strokes had higher percentage of SDB than patients with first stroke (74% vs 57% respectively).70 A small study involving patients in an acute stroke rehabilitation unit demonstrated AHI >10 in 91% of the studied population with a mean AHI of 32/hour.71 Worsening of OSA may also be found after acute stroke due to impairment of respiratory muscle coordination. In some studies, presence of dysphagia was found to predict the development of OSA in patients with acute stroke,72 while another suggests that presence of prestroke leucoencephalopathy predicts a more severe OSA.73 BMI and neck circumference have also been found to predict the presence of OSA in poststroke patients.74

In some studies, presence of dysphagia was found to predict the development of OSA in patients with acute stroke,72 while another suggests that presence of prestroke leucoencephalopathy predicts a more severe OSA.73 BMI and neck circumference have also been found to predict the presence of OSA in poststroke patients.74 Screening patients with acute stroke for sleep apnoea Concentrating on minimal requirements in a stroke population, a sleep apnoea assessment would require a system that monitors nasal airflow, and thoracic and abdominal respiratory muscles. A comprehensive polysomnogram would provide additional details regarding specific SDB, but is rarely available as an inpatient service even in many US hospitals with comprehensive stroke care capabilities. One of the oldest tools available for screening patients for sleep apnoea includes a continuous overnight oximetry recording. It is the modality that we recommend to ease the unavailability of other recording methods.

ailable as an inpatient service even in many US hospitals with comprehensive stroke care capabilities. One of the oldest tools available for screening patients for sleep apnoea includes a continuous overnight oximetry recording. It is the modality that we recommend to ease the unavailability of other recording methods. Overnight oximetry has been used for a very long time and can be found in the British Thoracic Society report in 1990 on diagnosis and treatment of sleep apnoea/hypopnoea syndrome.75 A review published in Chest concluded overnight oximetry to be a cost-effective tool with substantial accuracy for screening OSA76. In a validation study, Wang and colleagues reported an accuracy of 87.33–87.77%, a sensitivity of 89.36–89.87% and area under the curve of 0.953–0.957 for the diagnosis of severe sleep apnoea in a patient population clinically suspected of having OSA. Validation parameters were slightly less impressive in the moderate-to-severe sleep apnoea category.77 Studies comparing overnight oximetry to ambulatory PSG in patients with acute stroke are needed. It is anticipated that in-hospital ambulatory PSG will be of significantly greater clinical utility in patients admitted with acute ischaemic stroke suspected of having sleep apnoea.

Overnight oximetry has been used for a very long time and can be found in the British Thoracic Society report in 1990 on diagnosis and treatment of sleep apnoea/hypopnoea syndrome.75 A review published in Chest concluded overnight oximetry to be a cost-effective tool with substantial accuracy for screening OSA76. In a validation study, Wang and colleagues reported an accuracy of 87.33–87.77%, a sensitivity of 89.36–89.87% and area under the curve of 0.953–0.957 for the diagnosis of severe sleep apnoea in a patient population clinically suspected of having OSA. Validation parameters were slightly less impressive in the moderate-to-severe sleep apnoea category.77 Studies comparing overnight oximetry to ambulatory PSG in patients with acute stroke are needed. It is anticipated that in-hospital ambulatory PSG will be of significantly greater clinical utility in patients admitted with acute ischaemic stroke suspected of having sleep apnoea. Effect of treating sleep apnoea in patients with acute stroke on outcome In addition to its effect on the cardiovascular system, an arterial blood flow steal phenomenon has been described in patients with acute stroke. The affected tissue in ischaemic stroke is supplied by maximally or nearly maximally dilated arterioles with decreased vasomotor reactivity of the blood vessels.78 Studies have found severely impaired cerebral vasoreactivity and increased arterial stiffness in patients with PSG—confirmed severe OSA,79 even during wakefulness.80 During episodes of apnoea, development of hypercapnia results in selective vasodilation of blood vessels supplying normal brain tissue, resulting in a steal phenomenon away from the ischaemic vasoparalysed region depriving it of critical oxygen. The term ‘reverse Robin Hood syndrome’ has been used for this phenomenon.81

uring episodes of apnoea, development of hypercapnia results in selective vasodilation of blood vessels supplying normal brain tissue, resulting in a steal phenomenon away from the ischaemic vasoparalysed region depriving it of critical oxygen. The term ‘reverse Robin Hood syndrome’ has been used for this phenomenon.81 Studies demonstrating the effect of positive airway pressure ventilation during acute stroke are not many and much needed. During the acute phase, SDB of variable degree has been found to be associated with an increased risk of neurological deterioration within 72 hours of acute stroke.81 This deterioration may occur even after improvement in stroke severity. In one study,81 neurological deterioration was independently predicted by sleep apnoea with an OR of 8.2. Another small study using BiPAP treatment initiated within 24 hours of acute stroke onset indicated a favourable effect.82

of acute stroke.81 This deterioration may occur even after improvement in stroke severity. In one study,81 neurological deterioration was independently predicted by sleep apnoea with an OR of 8.2. Another small study using BiPAP treatment initiated within 24 hours of acute stroke onset indicated a favourable effect.82 Various non-randomised studies suggest that CPAP may improve mortality and prevent new vascular events after ischaemic stroke in patients with OSA.83–85 Both long-term survival and functional outcomes have been shown to be affected by presence of SDB.86 87 In an RCT by Parra et al,88 neurological improvement at 1 month and length of time from stroke to first cardiovascular event were in favour of the CPAP treatment group which was started 3–6 months after acute stroke. Another RCT by Bravata et al89 showed that treatment with CPAP had a potential benefit on cardiovascular events in patients with transient ischaemic attack. The downside is that poststroke patients tolerate poorly PAP treatments.90 In summary, treatment of sleep apnoea may prevent acute neurological worsening, decrease neurological morbidity and improve long-term outcomes in patients with acute ischaemic stroke. More definitive studies are needed.

Various non-randomised studies suggest that CPAP may improve mortality and prevent new vascular events after ischaemic stroke in patients with OSA.83–85 Both long-term survival and functional outcomes have been shown to be affected by presence of SDB.86 87 In an RCT by Parra et al,88 neurological improvement at 1 month and length of time from stroke to first cardiovascular event were in favour of the CPAP treatment group which was started 3–6 months after acute stroke. Another RCT by Bravata et al89 showed that treatment with CPAP had a potential benefit on cardiovascular events in patients with transient ischaemic attack. The downside is that poststroke patients tolerate poorly PAP treatments.90 In summary, treatment of sleep apnoea may prevent acute neurological worsening, decrease neurological morbidity and improve long-term outcomes in patients with acute ischaemic stroke. More definitive studies are needed. Evolution of sleep apnoea after acute stroke SDB improves as the stroke evolves, although 50% of patients still exhibit an AHI≥10/hour after 3 months following the acute event. CSA has been reported in about 26% of patients with acute stroke which improves over time more rapidly than the obstructive events.88 Another study suggests better improvement of SDB in haemorrhagic versus ischaemic strokes.91 Improvement in stroke symptoms probably plays a role in the resolution of symptoms after the acute phase of stroke. Indication for sleep apnoea therapies including positive airway pressure ventilation may require reassessment after the acute and subacute phases of stroke.

SDB in haemorrhagic versus ischaemic strokes.91 Improvement in stroke symptoms probably plays a role in the resolution of symptoms after the acute phase of stroke. Indication for sleep apnoea therapies including positive airway pressure ventilation may require reassessment after the acute and subacute phases of stroke. Sleep apnoea and vascular dementia Increasingly, more studies are finding a link between sleep apnoea and cerebral white matter disease. Harbison et al73 found the presence of prestroke cardiovascular and white matter disease to be associated with worse SDB. They concluded that either white matter is particularly sensitive to the hypoxia and cardiovascular effect of SDB, or white matter disease contributes to exacerbation of stroke following SDB. Guarnieri et al92 found SDB to be more severe among patients with vascular dementia while no increased prevalence of the same was found in other dementia syndromes. Shin and colleagues found moderate-to-severe OSA as an independent risk factor for white matter changes in middle-aged and older individuals in Korea.14 Minoguchi et al93 studied patients with moderate-to-severe OSA and found a higher prevalence of silent brain infarcts in patients with OSA compared with control participants. The biochemical markers for silent brain infarcts were also looked at and were found to respond favourably when patients were treated with nasal CPAP.93 Another study found one out of four patients with newly diagnosed OSA had a severe and distinctive neuropsychological dysfunction mainly involving inductive and deductive thinking, and constructive ability. Some analogy with cognitive patterns of MID suggests that a mainly subcortical damage underlies this dysfunction.94 Moderate-to-severe OSA may be a risk factor for development of vascular cognitive impairment as a result of cerebral subcortical small vessel disease.95 In Yaffe et al96 study, old women with OSA >15 AHI were more likely to develop cognitive impairment. Intermittent nocturnal hypoxia in patients with moderate-to-severe OSA contributes to ischaemic damage in the cerebral periventricular territory of long penetrating terminal arteries. Ischaemic damage to the cerebral periventricular white matter partially disconnects the frontal cortex with the thalamus leading to a form of subcortical dementia characterised by apathy, decreased executive functions, poor memory and in advanced cases difficulty walking and urinary incontinence.

ing terminal arteries. Ischaemic damage to the cerebral periventricular white matter partially disconnects the frontal cortex with the thalamus leading to a form of subcortical dementia characterised by apathy, decreased executive functions, poor memory and in advanced cases difficulty walking and urinary incontinence. CPAP applications may delay onset of dementia.97 While more definitive studies are needed to establish causality of sleep apnoea in the pathogenesis of vascular dementia, there seems to be a provocative association between the two. Conclusion Sleep apnoea causes pathological increase in sympathetic activity, contributing to autonomic dysregulation. This dysautonomia probably contributes to worsening of the cardiovascular risk profile in patients with sleep apnoea, and may be responsive to treatment with positive airway pressure ventilation and other sleep apnoea therapies. The worsening of cardiovascular profile is well known to increase the risk of stroke. Early diagnosis and treatment of sleep apnoea should reduce the risk of stroke. Sleep apnoea also contributes to the refractoriness of AF, HTN and diabetes to treatment. Untreated sleep apnoea can contribute to acute, subacute and long-term neurological deterioration in patients with acute stroke. Vascular dementia is another entity that may be associated with sleep apnoea. Competing interests: None declared. Provenance and peer review: Commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

Introduction Stroke is the second leading global cause of death and the fifth leading cause of death in the USA. Stroke is also a significant cause of disability, with 39.4 million daily adjusted life years (DALYs) lost due to ischaemic stroke globally, in 2010.1 Women are disproportionately affected by stroke, in part because they live longer, but they also have worse recovery from stroke than men.2 Mechanisms of ischaemic stroke include large artery atherosclerosis, cardioembolism, small artery occlusion, stroke of other determined aetiology (ie, dissection, vasculopathy, prothrombotic disorder) and stroke of undetermined aetiology.3 Large-vessel cerebrovascular disease accounts for about 15–20% of ischaemic stroke, and internal carotid artery (ICA) stenosis accounts for about half of these.4 5

, small artery occlusion, stroke of other determined aetiology (ie, dissection, vasculopathy, prothrombotic disorder) and stroke of undetermined aetiology.3 Large-vessel cerebrovascular disease accounts for about 15–20% of ischaemic stroke, and internal carotid artery (ICA) stenosis accounts for about half of these.4 5 In order to make medical decisions regarding revascularisation for patients with carotid disease, we use clinical trial data that dichotomise carotid disease into symptomatic and asymptomatic disease and focus largely on degree of stenosis. We weigh the expected benefit of carotid revascularisation with the potential surgical risk of periprocedural stroke or other medical complication. It is known that gender is a major determinant of the long-term outcome after carotid revascularisation. Two big issues complicate our ability to make evidence-based clinical decisions for any patient, but especially for women. First, the clinical trials we use to guide our decision-making in patients with carotid disease suffer from under-representation of women. Those that do assess outcomes in women have lacked power for adequate analysis, and this limits the generalisability of results among men and women. Second, it must be noted that many of the seminal trials were conducted in an era prior to our current standard which includes potent statin, antithrombotics and aggressive cardiovascular risk factor control, including tight blood pressure control, glycemic control, tobacco cessation and exercise. This brings into question whether we have an accurate understanding of the risk of stroke in the setting of modern medical therapy and how this changes the risk–benefit calculation of carotid revascularisation, specifically in women. Should we be making decisions based on data that is >20 years old? Do we have enough data on which to base decisions on optimal treatment for women?

nding of the risk of stroke in the setting of modern medical therapy and how this changes the risk–benefit calculation of carotid revascularisation, specifically in women. Should we be making decisions based on data that is >20 years old? Do we have enough data on which to base decisions on optimal treatment for women? In this article, we will review the data on carotid revascularisation in women in symptomatic and asymptomatic disease. We will discuss the usefulness and appropriateness of optimal medical therapy. Additionally, we will discuss possible aetiologies of gender-related differences in outcomes and the potential role of advanced imaging for risk stratification. Finally, we will discuss the importance of ongoing and new clinical trials to determine the benefit and risk of carotid revascularisation in women in the current era of medical therapy.

uss possible aetiologies of gender-related differences in outcomes and the potential role of advanced imaging for risk stratification. Finally, we will discuss the importance of ongoing and new clinical trials to determine the benefit and risk of carotid revascularisation in women in the current era of medical therapy. Women and asymptomatic carotid disease Asymptomatic carotid is defined as the presence of atherosclerotic narrowing of the extracranial ICA in individuals without a history of recent ipsilateral carotid territory ischaemic stroke or transient ischaemic attack (TIA). The US Preventive Services Task Force recommends against screening for asymptomatic carotid artery stenosis in the general adult population, but nevertheless, with the availability of non-invasive diagnostic testing and ease, many patients are found to have asymptomatic disease and require guidance on management. In the prestatin era, annual rate of stroke in asymptomatic disease with >75% stenosis was 2–2.5% and 1.3% with <75% stenosis.4 The medical management of patient with asymptomatic carotid disease has significantly improved over the past 20 years, with statin therapy, aggressive BP control and focus on lifestyle modification. Thus, these already relatively low numbers are likely not representative of the current setting of medical practice. For this reason, the management of asymptomatic disease is persistently a topic of debate. More recent studies showed between 0.5% and 1% per year annual rate of stroke in asymptomatic patients.6–8 In a meta-analysis of 11 asymptomatic studies, rate of ipsilateral stroke/TIA and any stroke TIA on medical therapy was as low as 1%, which brings into question intervention for anyone with asymptomatic disease, let alone women, especially in the era of current medical therapy.6

e of stroke in asymptomatic patients.6–8 In a meta-analysis of 11 asymptomatic studies, rate of ipsilateral stroke/TIA and any stroke TIA on medical therapy was as low as 1%, which brings into question intervention for anyone with asymptomatic disease, let alone women, especially in the era of current medical therapy.6 The Asymptomatic Carotid Atherosclerosis Study (ACAS) and the Asymptomatic Carotid Surgery Trial (ACST) enrolled only 34% women. ACAS and ACST enrolled asymptomatic patients with >60% stenosis and assessed outcomes of carotid endarterectomy (CEA) +medical therapy versus medical therapy alone. ACAS showed an overall combined 5-year rate of ipsilateral stroke, perioperative stroke and death of 5.1% in CEA arm versus 11% in medical arm with absolute risk reduction (ARR) 5.1% and relative risk reduction (RRR) 53%. Post hoc subgroup analysis showed a higher risk of operative stroke or death in women versus men (3.6% vs 1.7%). Five year ARR for CEA was only 1.4% in women compared with 8% in men.9 ACST showed an overall 6.4% risk of recurrent stroke in the surgical group versus 11.8% in the medical group, with an ARR of 5-year ipsilateral stroke, perioperative stroke or death of 5.3%. Unlike ACAS, ACST had prespecified sex subgroup analysis. Women had a lower ARR at 5 years than men (4.1% vs 8.2%).10 A meta-analysis of data from ACAS and ACST1 showed a benefit of CEA for men, but not women in 5 year risk of any stroke or perioperative death (women OR 0.96, 95% CI 0.63 to 1.45 vs men OR 0.49, 95% CI 0.36 to 0.66).11 12

ad prespecified sex subgroup analysis. Women had a lower ARR at 5 years than men (4.1% vs 8.2%).10 A meta-analysis of data from ACAS and ACST1 showed a benefit of CEA for men, but not women in 5 year risk of any stroke or perioperative death (women OR 0.96, 95% CI 0.63 to 1.45 vs men OR 0.49, 95% CI 0.36 to 0.66).11 12 Preplanned analysis of sex differences in asymptomatic patients in the Carotid Revascularization Endarterectomy versus Stenting Trial (CREST) did not show significant differences in 30-day composite perioperative risk of stroke/death/myocardial infarction (MI) (3.7% in women, 3.5% in men), stroke and death (1.6% in women, 1.3% in men) or MI (2.1% in women and 2.3% in men) between men and women who underwent CEA.13 This finding is discordant with findings from older randomised control trials that suggested less benefit of CEA for women compared to men. However, statistical power for subgroup analysis by sex may have been limited by low number of women enrolled.

in women and 2.3% in men) between men and women who underwent CEA.13 This finding is discordant with findings from older randomised control trials that suggested less benefit of CEA for women compared to men. However, statistical power for subgroup analysis by sex may have been limited by low number of women enrolled. We do not currently have reliable evidence of benefit for women with asymptomatic carotid disease. Women have a higher perioperative risk, as well as lower risk of stroke without surgery. Given that more recent studies showed between 0.5% and 1% per year annual rate of stroke in asymptomatic patients overall, a lower number than in older studies, and given the inconsistencies in benefit of CEA or carotid artery stenting (CAS) in asymptomatic women, it is reasonable to consider medical therapy for women who are not enrolled in a clinical trial to assess this question. We recommend using all available risk reduction strategies in these patients, especially in light of the fact that carotid disease, even if asymptomatic, can be a marker of increased risk for CAD and MI.

s reasonable to consider medical therapy for women who are not enrolled in a clinical trial to assess this question. We recommend using all available risk reduction strategies in these patients, especially in light of the fact that carotid disease, even if asymptomatic, can be a marker of increased risk for CAD and MI. In North America, the Carotid Revascularization Endarterectomy versus Stenting Trial (CREST)-2 (clinical trials.gov identifier: NCT02089217) trial is actively enrolling patients with asymptomatic 70–99% stenosis in a National Institute of Neurological Disorders and Stroke-sponsored trial comparing intensive medical therapy (IMT) alone versus either CEA or CAS (This trial will have prespecified analysis of sex subgroups and will ideally enrol a more representative percentage of women; ideally >40%). This will help ensure clinicians have optimal information on the risk and benefit ratio of CEA and CAS for asymptomatic stenosis in women specifically. Women and symptomatic carotid disease Symptomatic carotid disease is defined as TIA or stroke ipsilateral to the carotid stenosis in the preceding 180 days. In the prestatin era, annual rate of stroke with symptomatic carotid disease with >70% stenosis was 10–15%.5

In North America, the Carotid Revascularization Endarterectomy versus Stenting Trial (CREST)-2 (clinical trials.gov identifier: NCT02089217) trial is actively enrolling patients with asymptomatic 70–99% stenosis in a National Institute of Neurological Disorders and Stroke-sponsored trial comparing intensive medical therapy (IMT) alone versus either CEA or CAS (This trial will have prespecified analysis of sex subgroups and will ideally enrol a more representative percentage of women; ideally >40%). This will help ensure clinicians have optimal information on the risk and benefit ratio of CEA and CAS for asymptomatic stenosis in women specifically. Women and symptomatic carotid disease Symptomatic carotid disease is defined as TIA or stroke ipsilateral to the carotid stenosis in the preceding 180 days. In the prestatin era, annual rate of stroke with symptomatic carotid disease with >70% stenosis was 10–15%.5 For severe disease, defined as stenosis >70%, two trials published in 1991 provided the initial basis for our current practice of surgical revascularisation with CEA in symptomatic carotid disease. The North American Symptomatic Carotid Artery Endarterectomy Trial (NASCET) and the European Carotid Surgery Trial (ECST) were published in 1991 and compared CEA alone versus CEA and medical management. It must be emphasised that in the late 1980s and early 1990s, when these trials were conducted and published, statin and optimal medical therapy as we know it today with aggressive blood pressure control, glycemic control and focus on lifestyle modification were not standard of care.

versus CEA and medical management. It must be emphasised that in the late 1980s and early 1990s, when these trials were conducted and published, statin and optimal medical therapy as we know it today with aggressive blood pressure control, glycemic control and focus on lifestyle modification were not standard of care. NASCET enrolled 30% women. Overall, NASCET showed 17% ARR (p<0.001) in ipsilateral stroke at 2 years for the CEA arm, with a number needed to treat of six.14 ECST enrolled 28% women. Overall, ECST showed 11.6% RR in the surgical arm.15 Combined analysis from NASCET and ESCT (29% women) showed that the 30-day risk of perioperative stroke or death after CEA was higher in women compared to men (8.7% vs 6.8%). Five-year ARR of ipsilateral stroke and any perioperative stroke or death with surgery was higher in men (2.8% in women and 11% in men).16 Pooled analyses from NASCET and the Aspirin and Carotid Endarterectomy (ACE) Study showed increased 30-day risk of death in women (2.3% vs 0.8%, p=0.02).17 Higher perioperative risk of stroke and death was also shown but was not significant. For stenosis ≥70%, 5-year ARR from stroke was similar for women and men (15.1% vs 17.3%). With 50–69% stenosis, CEA was not beneficial in women (ARR 3%, p=0.94), but was in men (ARR 10%, p=0.02). Women treated with medical therapy had a low risk of stroke. They only benefitted from surgery if they had additional risk factors (age >70, severe hypertension, history of MI or hemispheric event).17

1% vs 17.3%). With 50–69% stenosis, CEA was not beneficial in women (ARR 3%, p=0.94), but was in men (ARR 10%, p=0.02). Women treated with medical therapy had a low risk of stroke. They only benefitted from surgery if they had additional risk factors (age >70, severe hypertension, history of MI or hemispheric event).17 Overall, in subgroup analyses of large trials comparing CEA to medical management, women appear to derive less benefit, and this is driven by increased risk of perioperative events. For less severe symptomatic disease, in the 50–69% stenosis range, revascularisation does not appear to be beneficial. The Stenting versus Aggressive Medical Therapy for Intracranial Arterial Stenosis (SAMMPRIS) trial highlighted the superiority of IMT compared to stenting for intracranial atherosclerotic disease.18 A similar trial focused on carotid disease revascularisation versus optimal medical therapy is needed. Women need to be included in representative numbers for preplanned subgroup analyses in order to provide a valid analysis of sex interaction. Women and carotid artery stenting CAS was developed as an alternative to CEA for patients who are high surgical risk, is widely available and in practice has been performed in patients with symptomatic and asymptomatic disease. As with CEA trials, CAS versus CEA trials suffer from under-representation of women participants. Another issue is that all comparisons of CAS to medical therapy are indirect; for example, CREST did not have a medical arm to compare medical treatment versus revascularisation.

with symptomatic and asymptomatic disease. As with CEA trials, CAS versus CEA trials suffer from under-representation of women participants. Another issue is that all comparisons of CAS to medical therapy are indirect; for example, CREST did not have a medical arm to compare medical treatment versus revascularisation. Various trials have compared CEA and CAS in women with inconsistent results. Some have found that women have worse short-term outcomes with CAS than CEA. An analysis of 20 613 women undergoing carotid intervention from hospitals in the states of New York and Florida found that CAS in symptomatic women was associated with increased perioperative morbidity and mortality when compared to CEA. Combined perioperative stroke/mortality was 10.9% for CAS and 3.8% for CEA in symptomatic women. The difference was less pronounced in asymptomatic women: those who underwent CAS had 3.1% rate of perioperative death or stroke, compared to 1.7% after CEA.19 Another analysis from the state of New York hospital discharge database included 27 439 women, and 36 295 men with about 90% asymptomatic patients in both sexes, found increased periprocedural risk with CAS versus CEA in symptomatic women; those who had CAS had higher mortality (4.19% vs 0.47%, p=0.01) and higher rate of combined stroke and mortality (12.09% vs 6.05%, p=0.02).20

e database included 27 439 women, and 36 295 men with about 90% asymptomatic patients in both sexes, found increased periprocedural risk with CAS versus CEA in symptomatic women; those who had CAS had higher mortality (4.19% vs 0.47%, p=0.01) and higher rate of combined stroke and mortality (12.09% vs 6.05%, p=0.02).20 There is less evidence regarding the comparison of outcomes between men and women. Overall, trials examining CAS versus CEA were not powered to detect a difference between men and women, and thus outcomes of carotid stenting in women compared to men have not been adequately investigated. A retrospective analysis including 228 patients, with 93 women, showed no significant differences in overall 30-day periprocedural stroke rate (2.1% in women vs 4.2% in men, p=0.48), death rate (0% vs 0.7%, p>0.99) or cardiac events (3.2% vs 0.7%, p=0.3). No differences were noted in long-term survival or stroke-free survival between genders.21 In CREST, analysis of asymptomatic and symptomatic patients demonstrated higher rates of combined periprocedural end points (stroke/death/MI) after CAS in women (6.8%, n=455) versus men (4.3%, n=807).13 22 Women undergoing CAS had higher periprocedural stroke risk than those undergoing CEA (5.5% vs 1.7%, p=0.01). No significant difference was found in men. In 4 year follow-up, there were no significant sex-related differences.13

end points (stroke/death/MI) after CAS in women (6.8%, n=455) versus men (4.3%, n=807).13 22 Women undergoing CAS had higher periprocedural stroke risk than those undergoing CEA (5.5% vs 1.7%, p=0.01). No significant difference was found in men. In 4 year follow-up, there were no significant sex-related differences.13 Available evidence suggest that CAS and CEA provide similar long-term outcomes for patients with asymptomatic and symptomatic carotid occlusive disease, but the periprocedural risk of stroke and death may be higher with CAS in women. Given that this elevated risk is in comparison to revascularisation with CEA, which carries higher periprocedural risk for women than men, it is difficult to recommend CAS routinely for women.

tic and symptomatic carotid occlusive disease, but the periprocedural risk of stroke and death may be higher with CAS in women. Given that this elevated risk is in comparison to revascularisation with CEA, which carries higher periprocedural risk for women than men, it is difficult to recommend CAS routinely for women. Timing of carotid revascularisation in women Timing of surgery for symptomatic carotid stenosis is more crucial in women than in men. Post hoc analysis of ECST and NASCET showed that there was significantly less benefit in women, and no change in men, with increased time from most recent index event. For stenosis >70–99%, women had ARR of 41.7% when surgery was conducted within 2 weeks, but this dropped to 6.6% at 2–4 weeks, and after 4–12 weeks, surgery was found to be harmful, with ARR −2.2%. This is in contrast to men, who had ARR 23% at 2 weeks, 23.8% at 2–4 weeks, 18.3% at 4–12 weeks and 20.4% after 12 weeks. For moderate stenosis, women showed ARR 13.8% at 2 weeks, vs 15.2% for men. Within weeks 2–4, surgery was harmful for women, with ARR −5.7%, while men still showed some benefit, with ARR 6.8%. Thus, women benefitted most when CEA was performed within 2 weeks, with a stark decline in benefit the longer out from index event. This was attributed to a more rapid decline in risk of stroke and death over time for women in the medical group.23 24

, with ARR −5.7%, while men still showed some benefit, with ARR 6.8%. Thus, women benefitted most when CEA was performed within 2 weeks, with a stark decline in benefit the longer out from index event. This was attributed to a more rapid decline in risk of stroke and death over time for women in the medical group.23 24 In actual practice, women tend to receive CEA later than men, even in the setting of severe, symptomatic stenosis. A study from California of patients seen at 19 emergency departments found that in patients with a TIA diagnosis and carotid stenosis >70%, the median time to carotid surgery was 18 days in men and 35 days in women. Women were also older than men (74 years vs 71 years) and were more likely to present with a ABCD2 score of ≥4.25 Thus, timely CEA for symptomatic women is not often achieved in real world settings.

th a TIA diagnosis and carotid stenosis >70%, the median time to carotid surgery was 18 days in men and 35 days in women. Women were also older than men (74 years vs 71 years) and were more likely to present with a ABCD2 score of ≥4.25 Thus, timely CEA for symptomatic women is not often achieved in real world settings. Basis of gender differences in carotid disease and revascularisation outcomes The traditional risk predictors for stroke in carotid disease have been assumed to be dependent on degree and severity of vessel stenosis, but other factors such as plaque size, composition, intraplaque haemorrhage, ulceration and overall plaque stability should be considered when determining risk of stroke. This is emphasised by the observation that strokes due to carotid disease are usually atheroembolic and less likely due to reduced flow related to the stenosis. Newer pathological and imaging studies highlight that carotid plaque constitution may play a role in determining risk of stroke. Differences in plaque morphology and composition may help explain why women benefit less from carotid revascularisation and more from medical therapy than men. If women have more stable plaques that are less likely to embolise, then removal of the plaque may provide less benefit.

y a role in determining risk of stroke. Differences in plaque morphology and composition may help explain why women benefit less from carotid revascularisation and more from medical therapy than men. If women have more stable plaques that are less likely to embolise, then removal of the plaque may provide less benefit. A study of duplex analysis of carotid plaque volume showed that plaque volume was higher in men than women at a comparable degree of stenosis, and that outcome of stroke, myocardial infarction and death were predicted by plaque area and not by degree of stenosis.26 A study involving 135 women (25% asymptomatic) and 315 men (22% asymptomatic) examined CEA specimens and found that women had more stable, less inflammatory plaque that was less likely to disrupt; they had less atheromatous plaque (22% vs 40%, p>0.001), more smooth muscle (38% vs 24%, p=0.001) and less macrophage infiltration (14% vs 21%, p=0.05). Symptomatic women had the most stable plaque.27 More prevalent stable plaque in women may help explain a lower benefit CEA.

ory plaque that was less likely to disrupt; they had less atheromatous plaque (22% vs 40%, p>0.001), more smooth muscle (38% vs 24%, p=0.001) and less macrophage infiltration (14% vs 21%, p=0.05). Symptomatic women had the most stable plaque.27 More prevalent stable plaque in women may help explain a lower benefit CEA. Another study included 64 women and 67 men with ≥50% asymptomatic stenosis determined duplex ultrasound and assessed factors considered high-risk plaque features on 3 T multicontrast MRI. Men were significantly more likely to have a thin/ruptured fibrous cap (48% vs 17%, p<0.01) and lipid-rich/necrotic core (73% vs 50%, p<0.01) and showed a trend towards more haemorrhage (33% vs 17%, p=0.07).28 Another study of 763 patients in which histological analysis was performed on carotid plaques found that plaques from men had higher rates of cellularity, more inflammatory infiltrates and less calcification.29 Increased incidence of high-risk plaque features in men may contribute to explanation of greater benefit of CEA in men versus women. In the future, advanced imaging modalities may aid in patient selection for carotid revascularisation by helping to determine who is at high risk of stroke based on plaque characteristics, instead of simply basing decision on degree of stenosis.

Another study included 64 women and 67 men with ≥50% asymptomatic stenosis determined duplex ultrasound and assessed factors considered high-risk plaque features on 3 T multicontrast MRI. Men were significantly more likely to have a thin/ruptured fibrous cap (48% vs 17%, p<0.01) and lipid-rich/necrotic core (73% vs 50%, p<0.01) and showed a trend towards more haemorrhage (33% vs 17%, p=0.07).28 Another study of 763 patients in which histological analysis was performed on carotid plaques found that plaques from men had higher rates of cellularity, more inflammatory infiltrates and less calcification.29 Increased incidence of high-risk plaque features in men may contribute to explanation of greater benefit of CEA in men versus women. In the future, advanced imaging modalities may aid in patient selection for carotid revascularisation by helping to determine who is at high risk of stroke based on plaque characteristics, instead of simply basing decision on degree of stenosis. Clinical trials focused on women On the basis of the findings described above, a carotid stenosis trial focused on women is worthy of serious consideration. A trial in which all the participants are women will circumvent the longstanding lack of representation of women in carotid trials. Further, since women appear to have a lower long-term stroke rate with medical therapy, they represent an ideal group to test the efficacy of current optimal medical therapy strategies. Finally, if the trial is designed to as a ‘pragmatic’ trial, with relatively few exclusion criteria, then a broader range of women can be enrolled, including greater representation of patients age 80 years and above. Inclusion of elderly patients is especially important considering the ageing of the population in most developed countries.

l is designed to as a ‘pragmatic’ trial, with relatively few exclusion criteria, then a broader range of women can be enrolled, including greater representation of patients age 80 years and above. Inclusion of elderly patients is especially important considering the ageing of the population in most developed countries. Conclusions CEA to prevent stroke is less beneficial for women compared with men. Women with symptomatic carotid disease have a higher periprocedural risk and a lower risk of recurrent stroke on medical treatment. Current evidence for asymptomatic women demonstrates minimal benefit for women, in conjunction with known increased perioperative risk. For asymptomatic disease, it is uncertain whether anyone will derive significant benefit from revascularisation in the era of modern medical therapy and this is being investigated in the CREST-2 trial. Thus, for women with asymptomatic disease not able to be enrolled in a clinical trial, medical management can be considered, until we have more data regarding benefit of revascularisation from a trial that has sufficient representation of women. These women should receive aggressive medical therapy and lifestyle modification for stroke prevention. Women should be counselled that risk reduction benefit from CEA is less than that for men. In general, women tend to be undertreated medically, and practitioners should ensure they optimise medical therapy, regardless of whether revascularisation is pursued.

Conclusions CEA to prevent stroke is less beneficial for women compared with men. Women with symptomatic carotid disease have a higher periprocedural risk and a lower risk of recurrent stroke on medical treatment. Current evidence for asymptomatic women demonstrates minimal benefit for women, in conjunction with known increased perioperative risk. For asymptomatic disease, it is uncertain whether anyone will derive significant benefit from revascularisation in the era of modern medical therapy and this is being investigated in the CREST-2 trial. Thus, for women with asymptomatic disease not able to be enrolled in a clinical trial, medical management can be considered, until we have more data regarding benefit of revascularisation from a trial that has sufficient representation of women. These women should receive aggressive medical therapy and lifestyle modification for stroke prevention. Women should be counselled that risk reduction benefit from CEA is less than that for men. In general, women tend to be undertreated medically, and practitioners should ensure they optimise medical therapy, regardless of whether revascularisation is pursued. The periprocedural risk of stroke and death may be higher with CAS than with CEA in women. Given that this elevated risk is in comparison to revascularisation with CEA, it is difficult to recommend CAS routinely for women.

Women should be counselled that risk reduction benefit from CEA is less than that for men. In general, women tend to be undertreated medically, and practitioners should ensure they optimise medical therapy, regardless of whether revascularisation is pursued. The periprocedural risk of stroke and death may be higher with CAS than with CEA in women. Given that this elevated risk is in comparison to revascularisation with CEA, it is difficult to recommend CAS routinely for women. A possible explanation for gender-associated differences in outcomes may be that women have different carotid plaque phenotypes, with more stable plaque. More advanced vascular imaging may help identify women with high-risk plaque and aid in clinical decision-making regarding revascularisation. Representation of women in stroke clinical trials remains an issue. The fact that women have been under-represented in carotid stenosis trials has led to uncertainty about the optimal treatment for women. A carotid stenosis trial focused on women is one potential solution to this vexing clinical problem. Contributors: EM-L wrote the first draft and contributed to interpretation of data and final approval of the manuscript. SC contributed to critical revision of the manuscript, interpretation of data, conception of work and final approval of the manuscript. EM-L and SC agree to be accountable for all aspects of the work. Competing interests: None declared. Provenance and peer review: Not commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

Introduction Mechanical thrombectomy in selected patients has become the gold standard treatment for patients presenting with acute thromboembolic infarction. The publication of multiple trials documenting a significant benefit represents a major step forward in the management of patients with acute ischaemia. These trials, with the exception of the Multicentre Randomised Clinical Trial of Endovascular Treatment for Acute Ischaemic Stroke in the Netherlands (MR CLEAN)1 and Extending the Time for Thrombolysis in Emergency Neurological Deficits—Intra-Arterial (EXTAND-IA),2 focused on patients with moderate-to-severe neurological deficits with a score of at least 6–8 on the National Institutes of Health Stroke Scale (NIHSS). Even though, the inclusion criteria for MR CLEAN included people with low NIHSS (≥2) the median NIHSS score of patients across all the trials ∼15–17. The complex interplay between large vessel proximal occlusion, blood pressure, collateral status and clinical symptomatology can result in some patients with large vessel thrombus presenting with low NIHSS scores3 and it has been suggested that all patients with an NIHSS score of 2 or above should undergo angiographic imaging in order to detect at least 90% of proximal occlusions.3 Furthermore, it is known that patients with proximal occlusion can deteriorate rapidly.4 Therefore, it is important to recognise that patients with a proximal large vessel occlusion do not necessarily have a high NIHSS score.

e should undergo angiographic imaging in order to detect at least 90% of proximal occlusions.3 Furthermore, it is known that patients with proximal occlusion can deteriorate rapidly.4 Therefore, it is important to recognise that patients with a proximal large vessel occlusion do not necessarily have a high NIHSS score. We sought to determine the outcome of patients with acute ischaemic stroke with minor symptomatology and NIHSS score ≤5 with confirmed occlusions in the M1 segment of the middle cerebral artery (MCA). We present our data on the clot length, initial imaging findings, collateral status, recanalisation rate, complications rate and clinical outcome. Materials and methods From our prospectively maintained database, we retrospectively selected patients with admission NIHSS scores of ≤5 and confirmed thrombus in the M1 segment of the MCA on preoperative imaging (CT or MRI) and in whom a mechanical thrombectomy procedure was performed at our institution. All consecutive patients entered into the database between January 2008 and August 2016 who met the inclusion criteria were included in the analysis. We chose an upper limit of 5 points on the NIHSS scale because patients with low scores are under-represented in the current literature. We selected only those patients with clot in the M1 segment as these patients represented a significant proportion of the patients included in the recent stroke trials.

Materials and methods From our prospectively maintained database, we retrospectively selected patients with admission NIHSS scores of ≤5 and confirmed thrombus in the M1 segment of the MCA on preoperative imaging (CT or MRI) and in whom a mechanical thrombectomy procedure was performed at our institution. All consecutive patients entered into the database between January 2008 and August 2016 who met the inclusion criteria were included in the analysis. We chose an upper limit of 5 points on the NIHSS scale because patients with low scores are under-represented in the current literature. We selected only those patients with clot in the M1 segment as these patients represented a significant proportion of the patients included in the recent stroke trials. A neurologist examined all the patients at admission and this included the NIHSS score. Patients with suspected acute ischaemic stroke underwent dedicated imaging, either CT or MRI. Those who underwent CT imaging had an unenhanced CT scan as well as CT angiogram. Those patients who underwent MRI had axial diffusion weighted imaging (DWI), Fluid Attenuated Inversion Recovery (FLAIR), susceptibility weighted imaging (SWI) and time of flight magnetic resonance angiogram (TOF-MRA) sequences to evaluate for further treatment with either thrombolysis and/or mechanical thrombectomy. Imaging was performed from the aortic arch to the vertex. Stroke therapy The use of intravenous (IV) thrombolysis followed national and international guidelines and was performed only if the patients presented with 4.5 hours of symptom onset.

A neurologist examined all the patients at admission and this included the NIHSS score. Patients with suspected acute ischaemic stroke underwent dedicated imaging, either CT or MRI. Those who underwent CT imaging had an unenhanced CT scan as well as CT angiogram. Those patients who underwent MRI had axial diffusion weighted imaging (DWI), Fluid Attenuated Inversion Recovery (FLAIR), susceptibility weighted imaging (SWI) and time of flight magnetic resonance angiogram (TOF-MRA) sequences to evaluate for further treatment with either thrombolysis and/or mechanical thrombectomy. Imaging was performed from the aortic arch to the vertex. Stroke therapy The use of intravenous (IV) thrombolysis followed national and international guidelines and was performed only if the patients presented with 4.5 hours of symptom onset. The patients were determined to be candidates for mechanical thrombectomy after discussion between the treating neurologist and the neurointerventionist. Mechanical thrombectomy was only offered if an acute occlusion could be demonstrated on the preoperative imaging and if salvageable brain tissue was present. An upper age limit for performing mechanical thrombectomy is not used in our institution, rather the premorbid status of the patient is taken into consideration along with comorbidities and the will of the patient or family members. All patients underwent mechanical thrombectomy under general anaesthesia. All patients underwent thrombectomy with stent retrievers and either a balloon guide catheter or distal aspiration catheter.

The patients were determined to be candidates for mechanical thrombectomy after discussion between the treating neurologist and the neurointerventionist. Mechanical thrombectomy was only offered if an acute occlusion could be demonstrated on the preoperative imaging and if salvageable brain tissue was present. An upper age limit for performing mechanical thrombectomy is not used in our institution, rather the premorbid status of the patient is taken into consideration along with comorbidities and the will of the patient or family members. All patients underwent mechanical thrombectomy under general anaesthesia. All patients underwent thrombectomy with stent retrievers and either a balloon guide catheter or distal aspiration catheter. Postinterventional management All patients were admitted to the neurointensive care or stroke unit after mechanical thrombectomy. All patients had routine follow-up imaging with either CT or MRI 24–36 hours postintervention. Follow-up assessment, including modified Rankin Scale (mRS) score, was obtained at 90 days by an inpatient hospital visit or telephone interview with a neurologist.

intensive care or stroke unit after mechanical thrombectomy. All patients had routine follow-up imaging with either CT or MRI 24–36 hours postintervention. Follow-up assessment, including modified Rankin Scale (mRS) score, was obtained at 90 days by an inpatient hospital visit or telephone interview with a neurologist. Data collection Data collection included baseline demographics (age, sex), medical history (including history of diabetes mellitus, hypercholesterolaemia, hypertension, atrial fibrillation), smoking history, and time of symptom onset and NIHSS to assess severity. The ASPECTS score was calculated for all patients on axial CT or DWI MRI sequences. The location of the occlusion and the thrombus length were assessed on axial images and correlated with catheter angiography. The collateral supply was graded according to the proposal by Higashida et al5 (grade 0—no collaterals visible to the ischaemic site, grade 1—slow collaterals to the periphery of the ischaemic site with the persistence of some of the defect, grade 2—rapid collateral to the periphery of the ischaemic site with persistence of some of the defect to only a portion of the ischaemic territory, grade 3—collaterals with slow but complete angiographic blood flow of the ischaemic bed by the late venous phase, grade 4—complete and rapid collateral blood flow to the vascular bed in the entire ischaemic territory by retrograde perfusion). The number of passes to obtain the final result was recorded. The final angiographic outcome was recorded using the Thrombolysis in Cerebral Infarction (TICI) scale.6 Post-treatment repeat imaging, at 24 hours, was performed and the ASPECTS score noted. The presence of haemorrhage was also recorded.

rograde perfusion). The number of passes to obtain the final result was recorded. The final angiographic outcome was recorded using the Thrombolysis in Cerebral Infarction (TICI) scale.6 Post-treatment repeat imaging, at 24 hours, was performed and the ASPECTS score noted. The presence of haemorrhage was also recorded. Results Demographics Between January 2008 and August 2016, we identified 41 patients who underwent mechanical thrombectomy for M1 occlusion and NIHSS≤5. The baseline characteristics are outlined in table 1. The mean age of the patients was 72 years old (±12 years) and 20 patients were male. Associated conditions were common among our cohort with over 80% of patients having a medical history of hypertension and 43.9% of patients having atrial fibrillation. The main suspected cause of the stroke was cardiac embolism with extracranial atherosclerosis being the second most common cause. Table 1 Baseline and outcome characteristics of patients with low NIHSS stroke and M1 thromboembolic occlusion that underwent mechanical thrombectomy

Results Demographics Between January 2008 and August 2016, we identified 41 patients who underwent mechanical thrombectomy for M1 occlusion and NIHSS≤5. The baseline characteristics are outlined in table 1. The mean age of the patients was 72 years old (±12 years) and 20 patients were male. Associated conditions were common among our cohort with over 80% of patients having a medical history of hypertension and 43.9% of patients having atrial fibrillation. The main suspected cause of the stroke was cardiac embolism with extracranial atherosclerosis being the second most common cause. Table 1 Baseline and outcome characteristics of patients with low NIHSS stroke and M1 thromboembolic occlusion that underwent mechanical thrombectomy Characteristic Results Age (mean±SD) 72±14 Male 20 (48.7%) Associated conditions Atrial fibrillation 18 (43.9%) Diabetes mellitus 7 (17.1%) Hypertension 33 (80.5%) Hypercholesterolaemia 15 (36.6%) Smoker 6 (14.6%) Suspected cause Cardiac embolism 16 (39%) Extracranial atherosclerosis 11 (26.9%) Intracranial atherosclerosis 2 (4.9%) Dissection 2 (4.9%) Unknown 10 (24.3%) NIHSS on admission 0 3 (7.3%) 1 3 (7.3%) 2 7 (17.1%) 3 7 (17.1%) 4 13 (31.7%) 5 8 (19.5%) Preoperative MRI 28 (68.3%) Preoperative CT imaging 13 (31.7%) ASPECTS score preoperative 8.8 (6–10) Side Left 22 (53.6%) Right 19 (46.4%) Clot length (mm) 10 (4.6–23.1 mm) Mean duration of procedure (min) 119±3 (range 36–294) Mean number of thrombectomy attempts 1.8 Time from stroke onset to recanalisation (min) (n=33) 385±57 (235–727) Postoperative ASPECTS (24 hours) 8.1 (5–10) Haemorrhage postoperative 7 (2 symptomatic haemorrhage) Collateral status 0 0 1 3 (7.3%) 2 9 (21.9%) 3 23 (56.1%) 4 6 (14.6%) TICI≥2b 36 (87.8%) TICI 3 26 (63.4%) 90-day Mrs 0 20 (50%) 1 7 (17.5%) 2 3 (7.5%) 3 3 (7.5%) 4 3 (7.5%) 5 1 (2.5%) 6 3 (7.5%) NA 1 (2.5%) mRS, modified Rankin Scale; NA, not available; NIHSS, National Institutes of Health Stroke Scale; TICI, Thrombolysis in Cerebral Infarction.

.3%) 2 9 (21.9%) 3 23 (56.1%) 4 6 (14.6%) TICI≥2b 36 (87.8%) TICI 3 26 (63.4%) 90-day Mrs 0 20 (50%) 1 7 (17.5%) 2 3 (7.5%) 3 3 (7.5%) 4 3 (7.5%) 5 1 (2.5%) 6 3 (7.5%) NA 1 (2.5%) mRS, modified Rankin Scale; NA, not available; NIHSS, National Institutes of Health Stroke Scale; TICI, Thrombolysis in Cerebral Infarction. Admission status All patients had an admission NIHSS of ≤5 but ∼50% of these patients had an NIHSS of 4 or 5. Preoperative MRI was performed in ∼70% of patients and just over half of patients had a left-sided thromboembolic occlusion. In our cohort, the average clot length was 10 mm (range 4.6–23.1 mm). The average preoperative ASPECTS score was 8.8 (range 6–10). Six patients received IV alteplase prior to mechanical thrombectomy (mean average dose 67 mg). Angiographic and procedural details All patients underwent mechanical thrombectomy. In all patients, the procedure was performed via a standard right common femoral approach. Just over 70% of the patients in our cohort had a collateral status of 3 or 4 on catheter angiography performed prior to the thrombectomy. In patients with occlusion of the internal carotid artery, this was determined via angiographic runs of the contralateral ICA and vertebral arteries.

right common femoral approach. Just over 70% of the patients in our cohort had a collateral status of 3 or 4 on catheter angiography performed prior to the thrombectomy. In patients with occlusion of the internal carotid artery, this was determined via angiographic runs of the contralateral ICA and vertebral arteries. In 5 cases, a proximal balloon guide catheter was used, and in 37 cases, a distal aspiration catheter was used. A distal aspiration catheter was used in all cases that did not use a proximal balloon guide catheter. In 39 cases, stent retrievers were used with aspiration used in the two remaining cases. The pREset (phenox, Bochum, Germany) stent retriever was used in 34 cases7 and a variety of other stents used in the remaining cases. The mean number of thrombectomy attempts was 1.8 (range 1–8 attempts). Thirteen patients (31.7%) required extracranial stenting with 2 patients requiring stenting for acute dissections and 11 patients requiring carotid stents for severe (>70%) stenosis or complete occlusion of the internal carotid artery. Carotid stenting was always performed prior to the thrombectomy. The mean duration of the procedure was 119 min and the mean time from stroke onset to recanalisation was 407 min (n=34). In seven patients the time of stroke onset was unknown. A TICI 2b result or better was achieved in 87.8% of patients with TICI 3 achieved in 63.4% of patients. There were no intraprocedural complications.

In 5 cases, a proximal balloon guide catheter was used, and in 37 cases, a distal aspiration catheter was used. A distal aspiration catheter was used in all cases that did not use a proximal balloon guide catheter. In 39 cases, stent retrievers were used with aspiration used in the two remaining cases. The pREset (phenox, Bochum, Germany) stent retriever was used in 34 cases7 and a variety of other stents used in the remaining cases. The mean number of thrombectomy attempts was 1.8 (range 1–8 attempts). Thirteen patients (31.7%) required extracranial stenting with 2 patients requiring stenting for acute dissections and 11 patients requiring carotid stents for severe (>70%) stenosis or complete occlusion of the internal carotid artery. Carotid stenting was always performed prior to the thrombectomy. The mean duration of the procedure was 119 min and the mean time from stroke onset to recanalisation was 407 min (n=34). In seven patients the time of stroke onset was unknown. A TICI 2b result or better was achieved in 87.8% of patients with TICI 3 achieved in 63.4% of patients. There were no intraprocedural complications. Postprocedural clinical and angiographic All patients had follow-up imaging. The average ASPECTS score on the 24-hour scan was 8.1 (5–10). Seven patients showed signs of haemorrhage, five of which had localised clinically asymptomatic subarachnoid haemorrhage (SAH). Two patients had symptomatic haemorrhage, both of which were SAH.

al clinical and angiographic All patients had follow-up imaging. The average ASPECTS score on the 24-hour scan was 8.1 (5–10). Seven patients showed signs of haemorrhage, five of which had localised clinically asymptomatic subarachnoid haemorrhage (SAH). Two patients had symptomatic haemorrhage, both of which were SAH. At 90-day follow-up (n=40), 75% of patients had a good outcome with mRS≤2 and a moderate outcome with mRS≤3 in 82.5% of patients. There were three deaths in our cohort, one of which was secondary to a severe lower respiratory tract infection and the other secondary to the stroke.

al clinical and angiographic All patients had follow-up imaging. The average ASPECTS score on the 24-hour scan was 8.1 (5–10). Seven patients showed signs of haemorrhage, five of which had localised clinically asymptomatic subarachnoid haemorrhage (SAH). Two patients had symptomatic haemorrhage, both of which were SAH. At 90-day follow-up (n=40), 75% of patients had a good outcome with mRS≤2 and a moderate outcome with mRS≤3 in 82.5% of patients. There were three deaths in our cohort, one of which was secondary to a severe lower respiratory tract infection and the other secondary to the stroke. Discussion The recent meta-analysis conducted by the HERMES collaborators confirmed the beneficial effect of mechanical thrombectomy in patients with acute large vessels thromboembolic stroke.8 This meta-analysis grouped patients from the five major trials, (MR CLEAN, ESCAPE, REVASCAT, SWIFT PRIME and EXTEND IA) carried out between 2012 and 2014 with data from 1287 patients being analysed. They concluded that endovascular thrombectomy led to significantly reduced disability compared with control at 90 days (mRS 0–2 46% for the interventional arm, and 26.5% for control population).1 2 9–11 The number needed to treat via mechanical thrombectomy to reduce disability by at least one point on the mRS for one patient was 2.6 and subgroup analysis showed persistent benefits in patients aged over 80, those with symptom onset over 300 min prior to randomisation and in those patients not receiving alteplase. It was also seen that a similar effect on disability was seen across the entire NIHSS severity range and despite the relative paucity of patients with mild strokes (≤10, n=177) included in the studies there was a trend to favour intervention. This was contrary to previous studies that have identified patients with the most severe strokes (baseline NIHSS>19) as deriving most benefit from mechanical thrombectomy.12 However, there is evidence that a significant number of patients with large vessel occlusion can present with relatively mild neurological symptoms. In the study by Maas et al,3 they demonstrated that a large number of patients with low baseline NIHSS scores (55% of patients with NIHSS scores <11) had large vessel occlusion on non-invasive imaging. Furthermore, Smith et al13 showed that 31.2% of patients (29 200 of 93 517) with acute ischaemic stroke did not receive IV thrombolysis because of mild or improving symptoms and that of this group 28.3% could not be discharged home and 1.1% died. The study by Nedeltchev et al14 showed similar findings. In this study of 90-day postdischarge outcomes in patients were not given recombinant tissue plasminogen activator (rtPA) because of mild or rapid stroke symptoms, 24.7% showed mRS≥2.

mptoms and that of this group 28.3% could not be discharged home and 1.1% died. The study by Nedeltchev et al14 showed similar findings. In this study of 90-day postdischarge outcomes in patients were not given recombinant tissue plasminogen activator (rtPA) because of mild or rapid stroke symptoms, 24.7% showed mRS≥2. In the study by Mokin et al15 the authors noted the variability seen in the outcomes of patients admitted with low baseline NIHSS scores (range 0–7). Within this range, the authors demonstrated that an increase in NIHSS score by each point demonstrated a trend towards progressively higher numbers of patients being unable to ambulate at discharge. Interestingly, there appeared to be a rapid drop in the ability to ambulate independently at discharge (mean duration of stay 6 days) for patients with NIHSS 0–2 (NIHSS=0–86%, 1–73%, 2–52%) with a second drop occurring for patient with NIHSS 5–7 (NIHSS=5–52%, 6–44%, 7–28%). There is also some evidence to suggest that right-sided infarcts may present with falsely mild symptoms and poor outcomes without thrombolytic treatment.16 Taken together, this information strongly suggests that a low admission NIHSS score is not necessarily predictive of a good functional outcome and that these patients should undergo investigation to exclude large vessel occlusion as has been suggested.15 17

d symptoms and poor outcomes without thrombolytic treatment.16 Taken together, this information strongly suggests that a low admission NIHSS score is not necessarily predictive of a good functional outcome and that these patients should undergo investigation to exclude large vessel occlusion as has been suggested.15 17 The limited symptomatology of some patients presenting with proximal large vessel occlusion is likely due to their leptomeningeal collateral status. Numerous studies have shown a correlation between collateral status and clinical outcome. For example, Miteff et al18 graded collaterals into good, moderate and poor based on the filling of the MCA branches distal to the obstruction on CT angiography. They found good collaterals in 55% of patients, moderate collaterals in 26% and poor collaterals in 18%. They found that good collateral status was associated with reduced infarction expansion as well as good outcome at 3 months (mRS≤2) in addition to the acute NIHSS, which was lower in patients with good collaterals. Interestingly of those patients with poor collaterals, none had a good outcome at 3 months. Similarly, Maas et al19 studied 134 patients with acute MCA occlusion and graded the collateral circulation on a five-point scale. They showed that patients with poor collaterals were four times more likely to clinically deteriorate in hospital as those with normal or exuberant collaterals. Other groups have also shown a relationship between collateral status and functional outcome20–22 with the recent post hoc analysis of the MR CLEAN data showing a significant modification of treatment effect by collateral status with the strongest benefit seen in patients with good collaterals (grade 3) and no effect seen in patients with absent collaterals (grade 0).23 Our findings are consistent with previous reports with ∼70% of patients having good leptomeningeal collaterals.

ignificant modification of treatment effect by collateral status with the strongest benefit seen in patients with good collaterals (grade 3) and no effect seen in patients with absent collaterals (grade 0).23 Our findings are consistent with previous reports with ∼70% of patients having good leptomeningeal collaterals. Miteff et al18 identified further interesting information in their study. They showed that those patients with good collaterals and reperfusion all achieved a good outcome. However, they also showed that in those with good collateral status but no major reperfusion, only 38% had a favourable outcome. Therefore, the presence of good collaterals alone will not prevent a poor outcome and reperfusion is still key. Furthermore, the status of the collaterals is not static and a decrease in their ability to supply the demands of the brain will result in expansion of the infarction. Campbell et al24 have previously shown that collateral circulation is not static, and in patients with good collaterals at baseline but without successful reperfusion, there was a strong correlation between collateral deterioration and infarct growth. Therefore, the question then arises as to why collaterals fail and if there is anything that can be done to prevent this? At the moment, conclusive answers to these questions remain elusive. Blood pressure appears to play an important role in cerebral collaterals with a post hoc analysis of data from the MR CLEAN trial demonstrating that the use of general anaesthesia negated the beneficial effects of thrombectomy. This was also confirmed by Ouyang et al25 and Brinjikji et al,26 although it has been noted that confounding factors are difficult to eradicate from the current data. The exact nature of this deterioration in outcome is unknown; however, the drop in blood pressure commonly seen during the induction of anaesthesia is thought by many to be a contributory cause and in two small clinical trials norepinephrine-o-phenylephrine-induced hypertension improved the outcome in patients with stroke27 28 which gives some credence to this theory. Other authors have suggested that a rise in intracranial pressure (ICP) can cause a collapse in the collateral circulation29 and in animal models ICP rise of even 5 mm Hg can cause significantly retarded flow via collateral pathways.

ved the outcome in patients with stroke27 28 which gives some credence to this theory. Other authors have suggested that a rise in intracranial pressure (ICP) can cause a collapse in the collateral circulation29 and in animal models ICP rise of even 5 mm Hg can cause significantly retarded flow via collateral pathways. This change in the flow seems to be related to as of yet unidentified cellular processes within the ischaemic penumbra as they are oedema-independent.30 Other proposed mechanisms are cerebral venous steal,31 reversed Robin Hood syndrome32 and blood pressure fluctuations secondary to autonomic dysfunction33 as well as others.34 Interestingly, with regard to the latter, a recent study by Xiong et al35 showed that 76.5% of patients with relatively mild strokes (mean NIHSS score 4.4) had severe autonomic dysfunction and that those with more severe autonomic dysfunction had poorer outcomes at 2 months. In all likelihood, it is probable that multiple processes interact to cause the deterioration and the over-riding aim should be to restore normal or near normal flow as soon as possible through whichever means is most suitable. Pfaff et al36 recently published the results from their series of 33 patients with low NIHSS (≤8) treated by mechanical thrombectomy. The median NIHSS score was 5 (IQR 4–7) and presenting median ASPECTS score 10 (IQR 9–10). Just over half of the patients had a confirmed M1 occlusion and 72.7% had grade 3 collaterals (93.9% had grade 2 or 3 collaterals). In this study, 63.6% of patients had a favourable outcome (mRS≤2) at 90 days with 90.9% having moderate clinical outcome (mRS≤3) with 9.1% mortality, one of which was caused by SAH secondary to wire perforation. While the results of this study do not differ significantly from those of Strbian et al,37 one very important finding was that the median thrombus length was 12 mm (IQR 10–12 mm) and it is known that the chance of recanalisation with thrombus lengths over 8 mm is extremely low.38 39 Our results are similar and interestingly we also showed the presence of long clots with the mean average clot length in our cohort measuring 10 mm. It is also perhaps worth considering clot length in a slightly different way. On average, ∼75% of patients in the five major mechanical thrombectomy trials achieved TICI≥2b reperfusion status. A similar probability of reperfusion after IV-tPA is seen if the clot is below 4 mm.

clot length in our cohort measuring 10 mm. It is also perhaps worth considering clot length in a slightly different way. On average, ∼75% of patients in the five major mechanical thrombectomy trials achieved TICI≥2b reperfusion status. A similar probability of reperfusion after IV-tPA is seen if the clot is below 4 mm. Therefore, one should actually consider mechanical thrombectomy in all patients in whom the clot is 4 mm or longer since at this length the probability of achieving reperfusion favours mechanical thrombectomy. In our data, all the clots were longer than 4 mm with the shortest clot measuring 4.6 mm.

clot length in our cohort measuring 10 mm. It is also perhaps worth considering clot length in a slightly different way. On average, ∼75% of patients in the five major mechanical thrombectomy trials achieved TICI≥2b reperfusion status. A similar probability of reperfusion after IV-tPA is seen if the clot is below 4 mm. Therefore, one should actually consider mechanical thrombectomy in all patients in whom the clot is 4 mm or longer since at this length the probability of achieving reperfusion favours mechanical thrombectomy. In our data, all the clots were longer than 4 mm with the shortest clot measuring 4.6 mm. Recently Haussen et al40 published their intention-to-treat study. In this study, patients with a low NIHSS≤5 and confirmed large vessel occlusion were assigned to either medical treatment with thrombolysis (69% of patients at admission) or thrombectomy (31% at admission). Of those assigned to medical treatment, 41% deteriorated and required mechanical thrombectomy despite the fact that the median NIHSS for patients in the medical treatment group was only 2 meaning that the collateral status in these patients was likely to be very good (median NIHSS=4 in the group treated immediately with mechanical thrombectomy). Unfortunately, the collateral status for each patient was not documented. Furthermore, the median time to deterioration of 5.2 hours (range 2–25 hours) which adds credence to the stance that the collateral pathways that preserve brain tissue are exhaustible and when these systems collapse propagation of the infarct will ensue. This finding is consistent with Campbell et al24 as mentioned earlier. It is also noteworthy to mention that in those patients who went on to require mechanical thrombectomy, the average time from deterioration to reperfusion was 2.7 hours. This is important since at the time of arrival in hospital, the collateral pathway is able to sustain hypoxic brain tissue. However, once the patient begins to deteriorate, the implication is that the collaterals have failed and therefore the rate of expansion of infarction is likely to be greater at this time point than earlier in the clinical course. If correct, this should warrant early intervention when the time pressure is not as acute. This group also demonstrated a mRS phase shift of −2.5 points in favour of mechanical thrombectomy even though the median NIHSS score in both groups was ≤2. There were three deaths in the medically treated group and none in the thrombectomy group.

hould warrant early intervention when the time pressure is not as acute. This group also demonstrated a mRS phase shift of −2.5 points in favour of mechanical thrombectomy even though the median NIHSS score in both groups was ≤2. There were three deaths in the medically treated group and none in the thrombectomy group. Although randomised controlled trial data do not currently exist, there is a growing body of evidence to suggest patients presenting with low NIHSS scores should be considered for mechanical thrombectomy, and that if this is considered feasible, it should be performed early. This study has several limitations. First it is retrospective in nature and we have no control group of patients with low NIHSS scores and confirmed M1 occlusions who did not receive mechanical thrombectomy to compare with. We have focused on patients with low NIHSS and anterior circulation thromboembolic stroke and therefore the study is not applicable to those patients with low NIHSS score and posterior circulation stroke. Furthermore, we excluded those patients with terminal carotid and internal carotid artery terminus (ICA-T) occlusions in this analysis and therefore the results may not be applicable to this group of patients. Conclusion Thrombectomy in patients presenting with low NIHSS scores is safe and carries a high rate of technical success and good safety profile. Although data from randomised controlled trial data is lacking, we would advocate a policy of early intervention if the presence of an occlusive thrombus is identified.

This study has several limitations. First it is retrospective in nature and we have no control group of patients with low NIHSS scores and confirmed M1 occlusions who did not receive mechanical thrombectomy to compare with. We have focused on patients with low NIHSS and anterior circulation thromboembolic stroke and therefore the study is not applicable to those patients with low NIHSS score and posterior circulation stroke. Furthermore, we excluded those patients with terminal carotid and internal carotid artery terminus (ICA-T) occlusions in this analysis and therefore the results may not be applicable to this group of patients. Conclusion Thrombectomy in patients presenting with low NIHSS scores is safe and carries a high rate of technical success and good safety profile. Although data from randomised controlled trial data is lacking, we would advocate a policy of early intervention if the presence of an occlusive thrombus is identified. Contributors: PB was involved in data collection, analysis and overall study design. HB and OG were involved in review and editing. HH was involved in data collection, study design, review and editing. MAP is the guarantor. Competing interests: MAP and PB serve as proctors and consultants for Phenox GmbH, with moderate financial compensation. HH is a co-founder and shareholder of Phenox GmbH. Provenance and peer review: Not commissioned; internally peer reviewed. Data sharing statement: No additional data are available.

Introduction Stroke is a serious brain injury that can result in permanent disability and death. The burden of stroke, including the absolute numbers of incidence and death, increased during the last decade.1 Globally, an estimated 33 million strokes occurred and 5.8 million individuals died from stroke in 2010.1 2 In addition, around 5 million stroke survivors have permanent disability.2 In 2010, the estimated total cost of stroke, including direct medical cost and indirect cost, was US$53.9 billion in the USA and €64.1 billion in Europe.3 4 Approximately 70% of strokes are ischaemic worldwide, while the proportion of ischaemic stroke varies by race/ethnicity and region.1 5 To reduce the burden associated with stroke, investigations of cost-effectiveness of available treatments for patients with stroke such as intravenous (IV) injection of recombinant plasminogen activator (rtPA) are necessary. Since the US Food and Drug Administration (FDA) approval in 1996, rtPA remains the only thrombolytic agent approved for acute ischaemic stroke in the USA.6 IV rtPA has been shown to improve health outcomes after stroke.7 8

oke such as intravenous (IV) injection of recombinant plasminogen activator (rtPA) are necessary. Since the US Food and Drug Administration (FDA) approval in 1996, rtPA remains the only thrombolytic agent approved for acute ischaemic stroke in the USA.6 IV rtPA has been shown to improve health outcomes after stroke.7 8 In the past two decades, there have been some cost-effectiveness studies on IV rtPA. For instance, Fagan et al showed that IV rtPA within 3 hours after the onset of stroke saved cost associated with stroke treatment as well as improved outcomes from stroke in their 1998 study.9 Additionally, we found three review articles on the cost-effectiveness of IV rtPA for acute ischaemic stroke.10–12 All of them reviewed studies published prior to 2008.10–12 Since the new guidelines of IV rtPA between 3 and 4.5 hours after the onset of acute ischaemic stroke from the American Heart Association/American Stroke Association (AHA/ASA) as well as similar new recommendations from other organisations in Europe or Australia were released in late 2000s and early 2010s,13–16 and the cost-effectiveness of IV rtPA for the extended time windows, within 4.5 hours after the onset of stroke, has never been examined, an up-to-date review of economic impact of IV rtPA is needed to better understand the cost-effectiveness of IV rtPA under various treatment conditions. Thus, we conducted a literature review of cost-effectiveness of IV rtPA published up to 2014.

dows, within 4.5 hours after the onset of stroke, has never been examined, an up-to-date review of economic impact of IV rtPA is needed to better understand the cost-effectiveness of IV rtPA under various treatment conditions. Thus, we conducted a literature review of cost-effectiveness of IV rtPA published up to 2014. Methods We performed a comprehensive literature search of peer-reviewed journal articles published in English between January 1995 and December 2016 by using the databases MEDLINE, EMBASE, CINAHL and Cochrane Library. We augmented the search by using Google Scholar and checking the references of the articles we obtained. The strategy used for the search included keywords in stroke and rtPA treatment including acute ischemic stroke, tissue plasminogen activator and rtPA, and keywords in cost-effectiveness analyses including cost, economic, benefit, effectiveness and ICER (incremental cost-effectiveness analysis).

e articles we obtained. The strategy used for the search included keywords in stroke and rtPA treatment including acute ischemic stroke, tissue plasminogen activator and rtPA, and keywords in cost-effectiveness analyses including cost, economic, benefit, effectiveness and ICER (incremental cost-effectiveness analysis). Figure 1 depicts the process of literature selection for this review. The initial search yielded 224 abstracts. By screening of titles and abstracts, 197 studies were excluded because they were not cost-effectiveness studies or because they were about supporting strategies to increase the usage of IV rtPA, such as telemedicine or air transportation for patients with stroke, and thus were excluded. In addition, review articles, editorial letters, abstracts and commentaries were excluded (n=8). We completed full-text review of all articles that passed the initial titles and abstracts review and finalised the set of original research articles (n=16) for this study by further excluding three studies that were not original cost-effectiveness studies. Cost-effective analysis is an economic evaluation method comparing both costs and health outcomes of alternative interventions.17 Common health outcomes used in the literature include quality-adjusted life years (QALYs), life years gained, number of cases prevented and mortality.17 QALYs, which were developed in 1960s for cost-effectiveness analyses, are measures of health considering both mortality and morbidity. QALYs are valued between 0 and 1 per year, meaning 0 as death and 1 as perfect health.18 Cost-effectiveness analysis using QALYs is also called as cost-utility analysis.19 ICER, the main estimate in a cost-effectiveness analysis, is derived by the difference in costs over the difference in health outcomes between alternative interventions. In this review, ICER is the difference in cost between IV rtPA-treated group and non-IV rtPA group, that is, incremental cost, over the differences in QALYs between them, that is, incremental QALYs.

analysis, is derived by the difference in costs over the difference in health outcomes between alternative interventions. In this review, ICER is the difference in cost between IV rtPA-treated group and non-IV rtPA group, that is, incremental cost, over the differences in QALYs between them, that is, incremental QALYs. Figure 1 Selection of studies on cost-effectiveness analysis of recombinant tissue plasminogen activator (rtPA) for acute ischaemic stroke. We analysed the literature by1: model structure and main data sources,2 study results and3 major limitations. For model structure and data sources, we examined perspective, modelling method, and intervention type, and main source of economic and clinical data. For study results, we summarised the cost-effectiveness results by various study time windows, time horizon, net-cost savings, QALYs gained and ICER. Major limitations mentioned in each study were also summarised.

mined perspective, modelling method, and intervention type, and main source of economic and clinical data. For study results, we summarised the cost-effectiveness results by various study time windows, time horizon, net-cost savings, QALYs gained and ICER. Major limitations mentioned in each study were also summarised. We used a cost-effectiveness quadrant diagram to demonstrate the costs and outcomes of an IV rtPA strategy compared with a non-rtPA strategy (figure 2). The horizontal axis represents incremental QALYs associated with IV rtPA and the vertical axis represents the incremental cost associated with IV rtPA. For instance, the negative numbers in the vertical axis means that cost for a patient who received IV rtPA were lower than the cost for a patient who did not receive IV rtPA. When an estimated ICER is located in quadrant IV (lower right), IV rtPA is a cost-saving or a dominant strategy, that is, higher QALYs with less cost. When an estimated ICER is located in quadrant I or III, the acceptance decision depends on value of the estimated ICER and an ICER threshold. In this paper, we used US$50 000/QALY as a reference threshold.20 If the estimated ICER is below the threshold, that is, located under the dotted line in figure 2, we define that IV rtPA is a cost-effective strategy and adopt the IV rtPA strategy. Figure 2 Conceptual framework of cost-effectiveness of recombinant tissue plasminogen activator (rtPA) therapy. ICER, incremental cost-effectiveness analysis; QALY, quality-adjusted life year.

We used a cost-effectiveness quadrant diagram to demonstrate the costs and outcomes of an IV rtPA strategy compared with a non-rtPA strategy (figure 2). The horizontal axis represents incremental QALYs associated with IV rtPA and the vertical axis represents the incremental cost associated with IV rtPA. For instance, the negative numbers in the vertical axis means that cost for a patient who received IV rtPA were lower than the cost for a patient who did not receive IV rtPA. When an estimated ICER is located in quadrant IV (lower right), IV rtPA is a cost-saving or a dominant strategy, that is, higher QALYs with less cost. When an estimated ICER is located in quadrant I or III, the acceptance decision depends on value of the estimated ICER and an ICER threshold. In this paper, we used US$50 000/QALY as a reference threshold.20 If the estimated ICER is below the threshold, that is, located under the dotted line in figure 2, we define that IV rtPA is a cost-effective strategy and adopt the IV rtPA strategy. Figure 2 Conceptual framework of cost-effectiveness of recombinant tissue plasminogen activator (rtPA) therapy. ICER, incremental cost-effectiveness analysis; QALY, quality-adjusted life year. To compare ICERs from different countries, we derived 2014 US dollar value from all studies, which did not report ICERs in US dollars, by using consumer price indices (CPI) from the World Bank and purchasing power parity (PPP) exchange rate in 2014 from the Organisation for Economic Cooperation and Development (OECD).21–23 The 2014 US dollar value was derived by multiplying CPI in 2014 at a study country by incremental costs from a study, divided by CPI in a study year at a study country, and divided by a PPP exchange rate (national currency of study country per US dollar) in 2014 (ncremental costs from a study×(CPI in 2014 at a study country/CPI in a study year at a study country) / PPP exchange rate). When a study reported multiple ICERs from different time periods, we included ICERs from both a short-term (1 year) and a long-term (30 years or a lifetime) time period.

ry per US dollar) in 2014 (ncremental costs from a study×(CPI in 2014 at a study country/CPI in a study year at a study country) / PPP exchange rate). When a study reported multiple ICERs from different time periods, we included ICERs from both a short-term (1 year) and a long-term (30 years or a lifetime) time period. Results Among 15 original articles reviewed, six studies were from the US,9 24–28 two from the United Kingdom (UK),29 30 two from Australia,31 32 two from China,33 34 and one each from Canada,35 New Zealand,36 Denmark,37 and Spain.38 Nine of them used the payers’ perspective or healthcare system perspective, and four studies used the societal perspective while two studies did not clearly mention it. In table 1, nine of 15 studies investigated the cost-effectiveness of IV rtPA therapy within 0–3 hours after stroke onset,9 24 26 28 30 32 35 37 38 two studies within 3–4.5 hours,25 27 three studies within 0–4.5 hours,31 34 36 and one study within 0–6 hours29 33 (figure 1). The first study that examined cost-effectiveness of IV rtPA was published in 1998, two years after the FDA approval.9 Eight out of 16 studies were published between 2011 and 2016. Among them, the five studies were the studies of IV rtPA within 3–4.5 hours or 0–4.5 hours after the onset of stroke.25 27 31 34 36 The remaining three studies published during this period were US studies looking at the 0–3 hours time window to investigate up-to-date cost-effectiveness of IV rtPA24 or state specific cost-effectiveness of IV rtPA,26 and Chinese study examining the 0–6 hours cost-effectiveness of IV rtPA.33

t of stroke.25 27 31 34 36 The remaining three studies published during this period were US studies looking at the 0–3 hours time window to investigate up-to-date cost-effectiveness of IV rtPA24 or state specific cost-effectiveness of IV rtPA,26 and Chinese study examining the 0–6 hours cost-effectiveness of IV rtPA.33 Table 1 Summary of model structure and main data sources used in the cost-effectiveness studies of rtPA for acute ischaemic stroke

t of stroke.25 27 31 34 36 The remaining three studies published during this period were US studies looking at the 0–3 hours time window to investigate up-to-date cost-effectiveness of IV rtPA24 or state specific cost-effectiveness of IV rtPA,26 and Chinese study examining the 0–6 hours cost-effectiveness of IV rtPA.33 Table 1 Summary of model structure and main data sources used in the cost-effectiveness studies of rtPA for acute ischaemic stroke Study/Year/ Country Perspective Intervention Modelling method Economic data Clinical data Data collection/ Analyses Previous literature Data collection/ Analyses Previous literature Te Ao et al 36/2015/New Zealand Health funder perspective IV rtPA use within 4.5 hours after onset Simulation model (TreeAge, Excel) No Yes Yes ARCOS III Yes Yan et al 33/2015/China Chinese public health system perspective IV rtPA use within 6 hours after onset Decision tree Yes No Yes No Boudreau et al 24/2014/USA US payers’ perspective IV rtPA use within 3 hours after onset Decision tree, and Markov model (Excel) Yes (rtPA cost) Analy$ource Yes No Yes Pan et al 34/2014/China Healthcare payers’ perspective IV rtPA use within 4.5 hours after onset Decision tree and Markov model Yes CNSR, CHSY TIMS-China No Yes TIMS-China Yes Boudreau et al 25/2013/USA Payers’ perspective IV rtPA use within 3–4.5 hours after onset Decision tree, and Markov model (Excel) Yes Medicare reimbursement Yes No Yes Kazley et al 26/2013/USA (SC) Societal perspective IV rtPA use within 3 hours after onset Markov model Yes Hospital billing, HCUP No No Yes Tan Tanny et al 31/2013/Australia Societal and healthcare perspective IV rtPA use within 4.5 hours after onset Decision analytic model (Excel), and Monte Carlo simulation Yes Royal Melbourne Hospital Yes Yes Royal Melbourne Hospital Yes Tung et al 27/2011/USA Societal perspective IV rtPA use within 3–4.5 hours after onset A decision-analytic model (TreeAge) No Yes No Yes Johnston28/2010/USA NA IV rtPA use within 3 hours after onset NA No Yes No Yes Ehlers et al 37/2007/Denmark NA IV rtPA use within 3 hours with 24 hours in house neurology coverage andMRI Decision tree with Markov model (TreeAge) Yes Aarhus Hospital and Hvidovre Hospital data Yes Yes Aarhus Stroke Register, Death Register Yes Mar et al 38/2005/Spain Societal perspective IV rtPA use within 3 hours after onset Monte Carlo simulation (4000, no modelling) Yes Sakontzen questionnaire and social service experts Yes Yes Survey from hospitals in the province of Gipuzkoa Yes Moodie et al 32/2004/Australia Healthcare perspective IV rtPA use within 3 hours after onset MORUCOS Yes NEMESIS No Yes NEMESIS No

within 3 hours after onset Monte Carlo simulation (4000, no modelling) Yes Sakontzen questionnaire and social service experts Yes Yes Survey from hospitals in the province of Gipuzkoa Yes Moodie et al 32/2004/Australia Healthcare perspective IV rtPA use within 3 hours after onset MORUCOS Yes NEMESIS No Yes NEMESIS No Sandercock et al 29/2004/UK Healthcare and personal social services perspective IV rtPA use within 6 hours after onset Decision analysis model (TreeAge) Yes Western General Hospital, Edinburgh Yes No Yes Chambers et al 30/2002/UK Healthcare and social care perspective IV rtPA use within 3 hours after onset Stroke Outcome Model (TreeAge, Excel) Yes Clinicians’ panels Yes No Yes Sinclair et al 35/2001/Canada Healthcare system perspective IV rtPA use within 3 hours after onset Decision analytic model (TreeAge), and Markov model Yes Vancouver Hospital and Health Sciences Centre Yes No Yes Fagan et al 9/1998/USA Healthcare system perspective IV rtPA use within 3 hours after onset Markov model Yes (rtPA cost) Seven Detroit area hospitals Yes No Yes ARCOS III, Auckland Regional Community Stroke Study. CNSR, China National Stroke Registry. CHSY, China Health Statistics Yearbook. HCUP, Healthcare Cost and Utilization Project. IV rtPA, intravenous recombinant tissue plasminogen activator; NA, not applicable. MORUCOS, Model of Resource Utilization, Costs, and Outcomes for Stroke. NEMESIS, North East Melbourne Stroke Incidence Study. TIMS-China study, Thrombolysis Implementation and Monitor of acute ischaemic Stroke in China study.

HCUP, Healthcare Cost and Utilization Project. IV rtPA, intravenous recombinant tissue plasminogen activator; NA, not applicable. MORUCOS, Model of Resource Utilization, Costs, and Outcomes for Stroke. NEMESIS, North East Melbourne Stroke Incidence Study. TIMS-China study, Thrombolysis Implementation and Monitor of acute ischaemic Stroke in China study. The reviewed studies used various sources of data for analyses. Main data sources were published data or literature. When published data were not available, data from hospitals or panel survey data were used.30 38 For economic data, 10 studies used both previously published literature and data from their own collection or analyses. Three studies used previously published literature data only and two studies used data from the authors’ own collection or analyses. For clinical data, only five studies used data from both sources. In addition, three studies were from a small community-based study. All studies consistently showed that IV rtPA improved QALYs (table 2), even some showing marginal improvement of QALYs. Sinclair et al 35 showed exceptionally high improvement of QALYs associated with IV rtPA (3.46 QALYs per patient). Because of the complexity of the cost-effectiveness model and multiple input sources, there could be multiple reasons of high QALYs improvement in this study. Table 2 Main findings from the cost-effectiveness studies of rtPA for acute ischaemic stroke Study/year/country Year of cost Time windows (hours)* Time horizon Incremental cost†,‡ Incremental QALYs‡ Cost per QALY§

All studies consistently showed that IV rtPA improved QALYs (table 2), even some showing marginal improvement of QALYs. Sinclair et al 35 showed exceptionally high improvement of QALYs associated with IV rtPA (3.46 QALYs per patient). Because of the complexity of the cost-effectiveness model and multiple input sources, there could be multiple reasons of high QALYs improvement in this study. Table 2 Main findings from the cost-effectiveness studies of rtPA for acute ischaemic stroke Study/year/country Year of cost Time windows (hours)* Time horizon Incremental cost†,‡ Incremental QALYs‡ Cost per QALY§ (ICER) At year of cost¶ 2014 US$ At year of cost¶ 2014 US$ Te Ao et al 36/2015/New Zealand 2010 4.5 1 year NZ$413 302 0.06 6641 5037 Lifetime NZ$4051 2965 0.61 5093 4860 Yan et al 33/2015/China 2008 6 14 days US$569 626 0.04 14 231 15 652 Boudreau et al 24/2014/USA 2013 3 Lifetime (US$25 000) ((US$42 500)- to (US$11 000)) (25 421) 0.39 (0.16–0.66) Dominant Dominant Pan et al 34/2014/China 2011 4.5 1 year US$1560 1642 0.056 27 852 29 315 30 years US$1000 1052 0.422 2380 2494 Boudreau et al 25/2013/USA 2011 3–4.5 Lifetime US$1495 (US$4637– to US$6100) 1573 0.24 (0.01–0.60) 6255 6555 Kazley et al 26/2013/USA (SC) 2010 3 6 years (US$3454) (3751) 0.425 Dominant Dominant Lifetime (US$4084) (4435) 0.692 Dominant Dominant Tan Tanny et al 31/2013/Australia 2003–2011 4.5 1 year $A55.61 40 0.04 1478 991 Tung et al 27/2011/USA 2010 3 to 4.5 Lifetime US$6050 6570 0.28 21 978 23 465 Johnston28/2010/USA 2004 3 30 years (US$6074) (7617) 0.75 Dominant Dominant Ehlers et al 37/2007/Denmark 2004–2005 3 1 st year US$3335 4042 0.06 55 591 67 370 2nd year US$433 525 0.12 3 615 4373 3rd year (US$2093) (2537) 0.16 Dominant Dominant 30 years (US$16 561) (20 073) 0.43 Dominant Dominant Mar et al 38/2005/Spain 2001 3 1 year Men: (US$7874) Women: (US$10 496) (10 531) (14 038) 0.528 0.655 Dominant Dominant Moodie et al 32/2004/Australia 1997 3 Lifetime (US$1496) (2 207) 0.61 DALYs Dominant Dominant Sandercock et al 29/2004/UK NA 6 1 year £110 ((£441) to £471) 211 0.0081 (−0.0040–0.0183) 13 581 26 018 Lifetime (£3504) ((£4436) to (£3067)) (6713) 0.0363 (−0.0332–0.0848) Dominant Dominant Chambers et al 30/2002/UK 1996 3 Lifetime (£2333) (4835) 0.155 Dominant Dominant Sinclair et al 35/2001/Canada 1999 3 Lifetime ($C3800) (4085) 3.46 Dominant Dominant Fagan et al 9/1998/USA 1996 3 30 years (US$4255) ((US$13 022) to (US$531)) (6427) 0.564 (0.003–0.850) Dominant Dominant *Timing of patient presentation after onset of ischaemic stroke symptoms.

me (£2333) (4835) 0.155 Dominant Dominant Sinclair et al 35/2001/Canada 1999 3 Lifetime ($C3800) (4085) 3.46 Dominant Dominant Fagan et al 9/1998/USA 1996 3 30 years (US$4255) ((US$13 022) to (US$531)) (6427) 0.564 (0.003–0.850) Dominant Dominant *Timing of patient presentation after onset of ischaemic stroke symptoms. †Numbers in parenthesis stands for negative sign. ‡All numbers are per patient per time horizon. 95% CIs are shown in the squared bracket. §When the IV rtPA improves QALYs and reduces cost, it is shown as ‘dominant’. IV rtPA dominates not using IV rtPA. When IV rtPA is cost-effective, ICER is calculated at year of cost. ¶All monetary values in these two columns are consistent. DALYs, Disability Adjusted Life Year; ICER: incremental cost-effectiveness ratio; IV rtPA, intravenous recombinant tissue plasminogen activator; QALY, quality-adjusted life year. The impact of IV rtPA on cost was ambiguous and varied by time window and study time horizon. In the USA, two of the six studies examined the cost-effectiveness of IV rtPA within the 3–4.5 hours time window. Use of IV rtPA within 3–4.5 hours after the onset of stroke increased costs (US$1495–US$6050) but improved QALYs (0.24–0.28) over the lifetime. The estimated ICERs (US$6255/QALY–US$21 978/QALY) showed the therapy was cost-effective using the US$50 000/QALY threshold. The remaining four studies in the USA showed that IV rtPA within 0–3 hours after onset of stroke was a dominant strategy, that is, cost saving and QALYs gained.

ALYs (0.24–0.28) over the lifetime. The estimated ICERs (US$6255/QALY–US$21 978/QALY) showed the therapy was cost-effective using the US$50 000/QALY threshold. The remaining four studies in the USA showed that IV rtPA within 0–3 hours after onset of stroke was a dominant strategy, that is, cost saving and QALYs gained. The results from non-US studies using IV rtPA within 0–3 hours were consistent with the results from the US studies. One exception, which showed an ICER marginally above an ICER threshold of US$50 000/QALY at the first year (US$55 591/QALY), is the Danish study by Ehlers et al that examined a range of time periods and showed that IV rtPA within 0–3 hours after the onset of stroke with 24 hours in-house MRI imaging and neurology coverage increased cost for the first and the second year after stroke.37 IV rtPA, however, became a dominant strategy after the third year and the 30 years estimates also indicated the IV rtPA as a dominant strategy.37 Results from three non-US studies examining the cost-effectiveness of IV rtPA within 0–4.5 hours showed that IV rtPA increased cost but was cost-effective with an ICER threshold of US$50 000/QALY. The UK study by Sandercock et al showed that IV rtPA within 6 hours of symptom onset increased cost and the ICER was £13 581/QALY (US$25 045/QALY in 2012) for the first year after the stroke, but over the lifetime the therapy was a dominant strategy.29 The Chinese study by Yan et alalso showed that IV rtPA within 6 hours increased both cost and utility and cost-effective within 14 days after the stroke.33

reased cost and the ICER was £13 581/QALY (US$25 045/QALY in 2012) for the first year after the stroke, but over the lifetime the therapy was a dominant strategy.29 The Chinese study by Yan et alalso showed that IV rtPA within 6 hours increased both cost and utility and cost-effective within 14 days after the stroke.33 All of the ICERs were located in quadrant I or IV (figure 3). Lifetime ICERs of IV rtPA within 0–3 hours or 0–4.5 hours were located in quadrant IV, and therefore using IV rtPA was a dominant strategy. The ICER of IV rtPA within 0–3 hours from Sinclair et al 35 is not shown in figure 3 because of space limitation but the ICER was located in quadrant IV. The ICERs from studies that examined IV rtPA within 3–4.5 hours were located in quadrant I and under the threshold line, thus IV rtPA was a cost-effective strategy in this scenario. The impact of IV rtPA on cost in the first year was ambiguous, but IV rtPA was still a short-term dominant or a cost-effective strategy from most studies. Figure 3 Summary of incremental cost-effectiveness ratio (ICERs) of rtPA therapy from the literature. IV rtPA, intravenous recombinant tissue plasminogen activator; QALY, quality-adjusted life year.

All of the ICERs were located in quadrant I or IV (figure 3). Lifetime ICERs of IV rtPA within 0–3 hours or 0–4.5 hours were located in quadrant IV, and therefore using IV rtPA was a dominant strategy. The ICER of IV rtPA within 0–3 hours from Sinclair et al 35 is not shown in figure 3 because of space limitation but the ICER was located in quadrant IV. The ICERs from studies that examined IV rtPA within 3–4.5 hours were located in quadrant I and under the threshold line, thus IV rtPA was a cost-effective strategy in this scenario. The impact of IV rtPA on cost in the first year was ambiguous, but IV rtPA was still a short-term dominant or a cost-effective strategy from most studies. Figure 3 Summary of incremental cost-effectiveness ratio (ICERs) of rtPA therapy from the literature. IV rtPA, intravenous recombinant tissue plasminogen activator; QALY, quality-adjusted life year. We summarised major limitations of the literature (table 3). The most common limitation was insufficient data for accurate cost-effectiveness estimates. Some studies mentioned a lack of generalisability because of data limitations.24–26 28 33 35 It was also pointed out that some studies used multiple data sources because of limited data.24 25 Lack of long-term mortality and cost data as well as insufficient up-to-date outcome and cost data were also mentioned as limitations.24 26 37 Table 3 Major limitations listed in the cost-effectiveness studies of rtPA for acute ischaemic stroke Study/year/country Limitations Yan et al 33/2015/China The medical costs did not include the cost after discharge

We summarised major limitations of the literature (table 3). The most common limitation was insufficient data for accurate cost-effectiveness estimates. Some studies mentioned a lack of generalisability because of data limitations.24–26 28 33 35 It was also pointed out that some studies used multiple data sources because of limited data.24 25 Lack of long-term mortality and cost data as well as insufficient up-to-date outcome and cost data were also mentioned as limitations.24 26 37 Table 3 Major limitations listed in the cost-effectiveness studies of rtPA for acute ischaemic stroke Study/year/country Limitations Yan et al 33/2015/China The medical costs did not include the cost after discharge The study used charges not real costs The study used data from a single hospital in China Boudreau et al 24/2014/USA The results were specific to the assumptions and the data used QALYs were derived by using multiple inconsistent studies Long-term cost, QALYs, disabilities and mortality data were limited and dated Pan et al 34/2014/China Inaccurate estimate for each component of rtPA-associated cost Informal caregiving costs were not included The study did not model changes in functional status from causes other than stroke The study used the efficacy and the utility data from studies in high-income countries Boudreau et al 25/2013/USA The results are specific to the assumptions and the data used The data are from numerous published studies including clinical trials Kazley et al 26/2013/USA (SC) The study examined only a single state

The study did not model changes in functional status from causes other than stroke The study used the efficacy and the utility data from studies in high-income countries Boudreau et al 25/2013/USA The results are specific to the assumptions and the data used The data are from numerous published studies including clinical trials Kazley et al 26/2013/USA (SC) The study examined only a single state The assumptions and data used in the study did not fully represent the clinical practise situation Data do not represent the current year The study may underestimate the benefit because of previously validated model with conservative estimates The study only considered treatment within 3 hours after stroke onset (not up to 4.5 hours) Tan Tanny et al 31/2013/Australia The study assumed that survival and quality of life would not change between 90 days and 12 months after stroke Efficacy data were drawn from analyses of studies of rtPA being given between 3 and 4.5 hours (not rtPA within 4.5 hours) Tung et al 27/2011/USA Input parameters were best estimates from previously published data The study did not model changes in functional status from causes other than stroke Johnston28/2010/USA The results depended on a single cost-utility analysis that required a number of uncertain assumptions Ehlers et al 37/2007/Denmark The lack of adequate long-term data Mar et al 38/2005/Spain The use of proxies to answer the questionnaire Chambers et al 30/2002/UK Limited published data about the cost of care for stroke survivors Indirect costs, informal care costs and quality of life of other family members were excluded from the model

Ehlers et al 37/2007/Denmark The lack of adequate long-term data Mar et al 38/2005/Spain The use of proxies to answer the questionnaire Chambers et al 30/2002/UK Limited published data about the cost of care for stroke survivors Indirect costs, informal care costs and quality of life of other family members were excluded from the model No sufficient published information on resource use, rates of recurrence or disability and mortality by age group The variability of parameter estimates is not well known Sinclair et al 35/2001/Canada Short-term hospitalisation cost based on a small sample size of 22 patients from a single centre (generalisability) There was a difficulty in determining the costs of stroke care and services in Canada on a ‘per patient basis’ The study used a point estimate without a formal quantitative estimate of its precision Fagan et al 9/1998/USA The study used a placebo group from the NINDS rtPA Stroke Trial as the source of data for some aspects of the cost analysis The protocol precluded antithrombotic therapy in the first 24 hours after stroke onset, which may affect cost and health outcomes Three studies (Te Ao et al 36 Moodie et al 32 Sandercock et al 29) did not list limitations. NINDS, National Institue of Neurologic Disorders and Stroke; rtPA, intravenous recombinant tissue plasminogen activator; QALY, quality-adjusted life year.

The protocol precluded antithrombotic therapy in the first 24 hours after stroke onset, which may affect cost and health outcomes Three studies (Te Ao et al 36 Moodie et al 32 Sandercock et al 29) did not list limitations. NINDS, National Institue of Neurologic Disorders and Stroke; rtPA, intravenous recombinant tissue plasminogen activator; QALY, quality-adjusted life year. Discussion This review investigated studies about cost-effectiveness of IV rtPA for treating patients with acute ischaemic stroke. IV rtPA within 0–3 hours after the onset of stroke was cost-saving while improving QALYs during lifetime. The finding about the cost-effectiveness of IV rtPA within 0–3 hours after the onset of stroke is consistent with previous reviews.10–12 However, the most recent review was published before AHA/ASA released the updated guidelines with extended time window. In the review, we found that IV rtPA within 0–4.5 hours or within 3–4.5 hours after the onset of stroke was cost-saving or cost-effective. Although some studies showed that IV rtPA within 0–4.5 hours or within 3–4.5 hours after the onset of stroke increased cost, it was a cost-effective strategy. The review results emphasise the importance of reducing door-to-needle time for patients with acute ischaemic stroke.

f stroke was cost-saving or cost-effective. Although some studies showed that IV rtPA within 0–4.5 hours or within 3–4.5 hours after the onset of stroke increased cost, it was a cost-effective strategy. The review results emphasise the importance of reducing door-to-needle time for patients with acute ischaemic stroke. In addition to time windows, some other factors may lead to heterogeneity in study results. For example, the study perspective affects the cost-effectiveness of IV rtPA. Healthcare payers’ perspective considered only direct medical cost, while societal perspective included both direct medical cost and indirect cost, such as productivity loss and informal caregiving costs. IV rtPA is expected to decrease indirect costs associated with stroke, while IV rtPA is known as reducing the short-term disability rate.7–9 25 Considering indirect costs could improve the ICER for IV rtPA within 3–4.5 hours after stroke or make IV rtPA a dominant strategy. Time horizon may also significantly affect the cost-effectiveness of IV rtPA. All the studies consistently concluded that IV rtPA increased short-term (1 year) cost. However, IV rtPA reduced long-term cost (lifetime or 30 years) because of lower rehabilitation and disability-associated cost among patients with IV rtPA.

Time horizon may also significantly affect the cost-effectiveness of IV rtPA. All the studies consistently concluded that IV rtPA increased short-term (1 year) cost. However, IV rtPA reduced long-term cost (lifetime or 30 years) because of lower rehabilitation and disability-associated cost among patients with IV rtPA. A main strength of reviewed studies is a timely research using the most recent available costs and outcomes from published secondary sources or primary data collection as inputs for evaluations. These inputs changed over time because of new medical technology for treating acute ischaemic stroke and updated recommendations or guidelines. After releasing the updated guidelines from AHA/ASA in 2009 and other organisations in Europe and Australia on the extended time window for IV rtPA therapy,13–16 a number of publications (n=6) have examined the extended time window in the past 6 years.

g acute ischaemic stroke and updated recommendations or guidelines. After releasing the updated guidelines from AHA/ASA in 2009 and other organisations in Europe and Australia on the extended time window for IV rtPA therapy,13–16 a number of publications (n=6) have examined the extended time window in the past 6 years. Some common limitations of the studies, however, were also observed. One of the main limitations in the studies was that indirect costs, such as productivity loss and informal caregiving cost, were usually not included in the cost analyses. The proportion of indirect costs for stroke is significant.39 A literature review showed that the median proportion of indirect costs was 32% of the total cost of stroke.39 However, most studies chose the healthcare perspective or payers’ perspective, which did not consider indirect costs. Moreover, studies using the societal perspective did not include indirect costs,26 27 or included informal caregiving cost only.38 None of the studies included productivity loss as a part of cost. When current cost-effectiveness models assumed an elderly cohort, productivity loss among stroke survivors may be negligible. However, stroke onset among young adults has been increasing40 and productivity loss could be a large burden for young stroke survivors with disabilities. For better cost-effectiveness evaluation, indirect cost should be considered as a part of cost in the analyses.

tivity loss among stroke survivors may be negligible. However, stroke onset among young adults has been increasing40 and productivity loss could be a large burden for young stroke survivors with disabilities. For better cost-effectiveness evaluation, indirect cost should be considered as a part of cost in the analyses. Next, most of lifetime and long-term effectiveness data, including QALYs of disabled stroke survivors, incidence of recurrent stroke among stroke survivors and 1-year mortality, were limited as well as outdated, although all studies tried to use the most up-to-date data available. Most studies in the 2010 still used QALYs data from the 1990s studies.24–27 Although cost data, especially long-term cost data, could hardly be free from outdated data, the reviewed studies tried to use recent cost data or at least adjusted cost to current currency value by using consumer price index (CPI) to alleviate concerns regarding outdated data. Lastly, there were some inconsistencies because of using multiple data sources. For instance, QALYs by disability status were not well-developed in the literature. Thus, QALYs of disabled and non-disabled stroke survivors were obtained from different data sources.24 25 In addition, most of the cost data were not collected within clinical trials, leading to a lack of consistency within a study.

es. For instance, QALYs by disability status were not well-developed in the literature. Thus, QALYs of disabled and non-disabled stroke survivors were obtained from different data sources.24 25 In addition, most of the cost data were not collected within clinical trials, leading to a lack of consistency within a study. Potential research areas to make up for these limitations as well as to improve the quality of research remain. Despite robust results from sensitivity analyses, developing high-quality data sources is still important for future efforts. Developing long-term follow-up trials among stroke survivors and research in long-term cost and effectiveness is most needed. Published large-scale effectiveness data from the real-world, including cost as a subcomponent, and studies which investigate those data are also needed. There are needs for indirect cost data and cost-effectiveness studies from the societal perspective to better understand societal impact of IV rtPA therapy. Concurrently, better models with multiple age cohorts would be useful to identify the impact of IV rtPA on different age cohorts. Boudreau et al 25 partly show how much the ICER could be different by age. Another future research area would be to examine the impact of the age or severity of stroke on ICER of IV rtPA treatment.

ently, better models with multiple age cohorts would be useful to identify the impact of IV rtPA on different age cohorts. Boudreau et al 25 partly show how much the ICER could be different by age. Another future research area would be to examine the impact of the age or severity of stroke on ICER of IV rtPA treatment. There were few studies in middle-income countries, likely because of a lack of infrastructure to provide IV rtPA. The incidence of stroke in the middle-income world has increased since 1970s, with 85% of stroke deaths worldwide occurring in middle-income countries.41 We found only two studies of cost-effectiveness of IV rtPA from middle-income countries.33 34 To better understand the cost-effectiveness of IV rtPA worldwide, more studies from countries in Africa, Latin America and Asia would be useful.

970s, with 85% of stroke deaths worldwide occurring in middle-income countries.41 We found only two studies of cost-effectiveness of IV rtPA from middle-income countries.33 34 To better understand the cost-effectiveness of IV rtPA worldwide, more studies from countries in Africa, Latin America and Asia would be useful. In this review, we did not include studies examining cost-effectiveness of strategies to improve the underutilisation of IV rtPA. Despite strong evidence of better clinical outcomes associated with IV rtPA, IV rtPA remains underutilised among patients with acute ischaemic stroke.42 Only 3.4%–5.2% of patients with stroke received rtPA therapy in the USA in 2009.43 Telestroke, air transport and certified stroke centres have been discussed as strategies to improve the utilisation of IV rtPA. The implementation of those strategies may improve IV rtPA utilisation but require additional costs. However, reviewed studies assumed that there were no additional costs to provide patient access to IV rtPA. Further cost-effectiveness studies including implementation costs are needed to support utilisation of IV rtPA.

The implementation of those strategies may improve IV rtPA utilisation but require additional costs. However, reviewed studies assumed that there were no additional costs to provide patient access to IV rtPA. Further cost-effectiveness studies including implementation costs are needed to support utilisation of IV rtPA. Conclusions This study found that the IV rtPA was a dominant strategy for those who received the therapy within 0–3 hours after the onset of stroke and a cost-effective strategy for those who received the therapy within 3–4.5 hours after stroke in long-term compared with traditional treatment for patients with acute ischaemic stroke without IV rtPA. This review provides considerable support for further development of interventions to promote IV rtPA use. To better evaluate cost-effectiveness of IV rtPA, establishing relevant clinical and cost data sources and developing evaluation, including programme costs, may be useful to improve the access to and use of IV rtPA. Contributors: HJ planned the project, performed review and wrote the manuscript. MGG helped plan the project, commented for medical issues and contributed to revising the manuscript. GW planned and supervised the project, and contributed to revising the manuscript. Competing interests: None declared. Provenance and peer review: Not commissioned; externally peer reviewed.

Introduction The normal bone metabolism is extremely essential for maintaining the bone mineral density and bone strength, which is mainly determined by a delicate balance between bone formation by anabolic activities of osteoblasts and bone resorption by catabolic effects of osteoclasts.1 Since bone marrow is a good source of mesenchymal stem cells which have clonogenicity and multipotential characteristics, it is the most commonly used cell therapy for bone regeneration in the orthopaedics.2 Bone marrow-derived mesenchymal stem cells (BMSCs), a kind of adult stem cells, have the same hereditary basis with other tissues.3 BMSCs can be differentiated into osteoblasts, chondrocytes, adipocytes and myoblasts.4 Mankani et al reported that transplantation of BMSC-conjugated hydroxyapatite/tricalcium phosphate particles was able to heal craniofacial bone defects for a long term.5 Quarto et al first presented the preliminary clinical data of three patients with the treatment of autologous BMSCs and hydroxyapatite porous blocks in long bone defects. Complete fusion was observed in all three patients.6 Therefore, the proliferation and differentiation of BMSCs is critical for the osteogenesis.7

m.5 Quarto et al first presented the preliminary clinical data of three patients with the treatment of autologous BMSCs and hydroxyapatite porous blocks in long bone defects. Complete fusion was observed in all three patients.6 Therefore, the proliferation and differentiation of BMSCs is critical for the osteogenesis.7 Increasing evidences proved that the inflammation could disturb the balance between bone formation and resorption resulting in osteoporosis or osteonecrosis.8 The proliferation and osteogenic differentiation of BMSCs could be also affected by inflammatory cytokines including tumour necrosis factor-alpha (TNF-α), interleukin 1 (IL-1), IL-6 and interferon γ (IFN-γ).9 TNF-α is an important inflammatory cytokine which can regulate cellular proliferation, differentiation and apoptosis by activating various signal pathways and molecular events.10 TNF-α can cause osteoclast-induced bone destruction, inhibit the osteoblastogenesis and regulate the migration of BMSCs in vivo.11 12 Lu et al demonstrated that TNF-α inhibited osteoblasts differentiation from precursor cells, reduced expression of RUNX2 and osteoblast-associated transcription factors (Osterix) and suppressed vitamin D-stimulated transcription owing to the activation of transcription factor Nuclear factor kappa-light-chain-enhancer of activated B cellI (NF-κB).13 Hess et al showed that TNF-α increased the expression of bone morphogenetic protein-2 (BMP-2) in BMSCs through NF-κB signalling pathway in early osteogenic differentiation and NF-κB stimulated key regulators of osteogenesis, such as BMP-2, RUNX2 and Osterix, resulting in enhanced mineralisation of the extracellular matrix.14 Huang et al reported that levels of Runx2, Osx and ALP were upregulated at lower concentration of TNF-α, but downregulated at higher concentration of TNF-α causing dose-dependent effects of TNF-α on BMSCs’ osteogenic differentiation.15 These findings suggested that the binding of TNF-α to its receptors resulted in activating multiple signalling pathways.

Runx2, Osx and ALP were upregulated at lower concentration of TNF-α, but downregulated at higher concentration of TNF-α causing dose-dependent effects of TNF-α on BMSCs’ osteogenic differentiation.15 These findings suggested that the binding of TNF-α to its receptors resulted in activating multiple signalling pathways. Geraniin is a kind of polyphenols extracted from under leaf pearl, which is widely used to treat hepatitis, enteritis, dysentery, nephritis and oedema.16 Recently, Okabe et al reported that geraniin could relieve pain and reduce blood pressure through inhibiting releasing TNF-α.17 Xiao et al demonstrated that geraniin inhibited RANKL-induced osteoclastogenesis in a dose-dependent manner through NF-κB and Extracellular-signal Regulated protein Kinase (ERK) signalling pathways, which might be one of the mechanisms of anti-osteoporosis effects.18 In this study, the effects of geraniin on TNF-α-induced impairments of osteogenesis in BMSCs were studied by using methyl thiazol tetrazolium (MTT) assay and lactate dehydrogenase (LDH) assay. Oil red was used to measure osteogenesis of BMSCs. Real-time PCR and western blot analysis were used to analyse the miRNA expression of RunX2 and Osx, and detect the protein expression of NF-κB/Inhibitor of nuclear factor kappa-B, alpha (IкB-α) and p38 Mitogen Activated Proteinkinase (MAPK). In summary, the aim of this study is to unveil the mechanisms of the inhibitory effects of geraniin on TNF-α-induced impairments of osteogenesis in BMSCs.

ssion of RunX2 and Osx, and detect the protein expression of NF-κB/Inhibitor of nuclear factor kappa-B, alpha (IкB-α) and p38 Mitogen Activated Proteinkinase (MAPK). In summary, the aim of this study is to unveil the mechanisms of the inhibitory effects of geraniin on TNF-α-induced impairments of osteogenesis in BMSCs. Materials and methods Isolation and culture of BMSCs All experiments were performed following the guidelines of Animal Use and Care Committee. BALB/c mice aged 4–6 weeks were aseptically raised in the clean animal room. The mice BMSCs were harvested from long bones in the aseptic condition after carefully trimming excess tissues and washing marrow cavity using Phosphate Buffered Solution at 4°C. Subsequently, BMSCs were centrifugated at 2000 rpm for 5 min prior to resuspending with 5 mL Dulbecco’s modified eagle medium (Gibco, USA) plus 10% fetal bovine serum (Gibco, USA) in 25 cm2 flasks. After 24 hours, the floating cells were removed and the adherent cells were cultured at 37°C with 5% humidified CO2. The adherent cells were passaged by trypsinisation when reaching 80%–90% confluency and subcultured at a density of 4000–5000 cells/cm2. Meanwhile, adherent cells were incubated with 0.25% trypsin and 1 mM Ethylene Diamine Tetraacetic Acid at 37°C for 5 min. BMSC were subcultured four times and incubated for 3 days after labelling with 3 µg/mL of bromodeoxyuridine.19

n when reaching 80%–90% confluency and subcultured at a density of 4000–5000 cells/cm2. Meanwhile, adherent cells were incubated with 0.25% trypsin and 1 mM Ethylene Diamine Tetraacetic Acid at 37°C for 5 min. BMSC were subcultured four times and incubated for 3 days after labelling with 3 µg/mL of bromodeoxyuridine.19 MTT assay BMSCs were seeded into a 96-well plate at a density of 2×104 cells/well, and the fresh medium was supplemented with 20 ng/mL TNF-α or 0, 1.25, 2.5, 5, 10, 20 µM geraniin for 5 days. Subsequently, 20 µL 5 mg/mL MTT (Sigma) was added into each well. After 4-hour incubation, the absorbancy of each well was measured by using an ELX800 absorbance microplate reader (Bio-Tek, USA) at 560 nm. LDH assay BMSCs were seeded into a 96-well plate at a density of 2×104 cells/well, and the fresh medium was supplemented with 20 ng/mL TNF-α or 2.5 µM geraniin for 1 day. Then, 50 µL LDH was added into each well at 37°C for 30 min protecting from light. The absorbancy was measured at 450 nm using an ELX800 absorbance microplate reader.

ere seeded into a 96-well plate at a density of 2×104 cells/well, and the fresh medium was supplemented with 20 ng/mL TNF-α or 2.5 µM geraniin for 1 day. Then, 50 µL LDH was added into each well at 37°C for 30 min protecting from light. The absorbancy was measured at 450 nm using an ELX800 absorbance microplate reader. Oil red O stain BMSCs were seeded into a 96-well plate at a density of 2×106 cells/well, and the fresh medium was supplemented with 20 ng/mL TNF-α or 2.5 µM geraniin for 1 day. BMSCs was fixed by using 5% precool paraformaldehyde for 30 min at 4°C. 0.6% (w/v) oil red was added into each well and stained for 15 min at 37°C. BMSCs were carefully washed to remove unbound dyes and then immersed with 1 mL isopropyl alcohol for 10 min at 37°C. BMSCs were observed through fluorescent microscope (D5300, Nikon, Japan) at 510 nm. Three different visual fields of control group and geraniin-treated group containing approximately 10 cells were analysed, and triplicate experiments were done at three independent time points. The mean fluorescence intensity (MFI) of the whole cell was separately analysed by Image J software.

Nikon, Japan) at 510 nm. Three different visual fields of control group and geraniin-treated group containing approximately 10 cells were analysed, and triplicate experiments were done at three independent time points. The mean fluorescence intensity (MFI) of the whole cell was separately analysed by Image J software. Real-time reverse transcription PCR BMSCs were seeded into a 96-well plate at a density of 2×106 cells/well and the fresh medium was supplemented with 20 ng/mL TNF-α or 2.5 µM geraniin for 1 day. RNeasy Mini kit (Qiagen, USA) was used to collect total RNA from each well according to the manufacturer’ s instructions. One microgram of total RNA was used to synthesise cDNA using reverse transcriptase (TaKaRa Biotechnology, Japan). Then, the expression of miRNA was detected by using Real-time PCR ABI 7500 Sequencing Detection System (Applied Biosystems, USA) and a SYBR Premix Ex Tag kit (TaKaRa Biotechnology, Japan). The primers of Runx2 were as follows: 5′-TCTTAGAACAAATTCTGCCCTTT-3′ and 5′-TGCTTTGGTCTTGAAATCACA-3′; the primers of Osx were as follows: 5′-CCTCCTCAGCTCACCTTCTC-3′ and 5′-GTTGGGAGCCCAAATAGAAA-3′; the primers of GAPDH were as follows: 5′-GAAGGTGAAGGTCGGAGTC-3′ and 5′-GAGATGGTGATGGGATTTC-3′. Real-time PCR was performed under the following conditions: 40 cycles of 95°C for 35 s and 60°C for 35 s.20

TTTGGTCTTGAAATCACA-3′; the primers of Osx were as follows: 5′-CCTCCTCAGCTCACCTTCTC-3′ and 5′-GTTGGGAGCCCAAATAGAAA-3′; the primers of GAPDH were as follows: 5′-GAAGGTGAAGGTCGGAGTC-3′ and 5′-GAGATGGTGATGGGATTTC-3′. Real-time PCR was performed under the following conditions: 40 cycles of 95°C for 35 s and 60°C for 35 s.20 Western blot analysis BMSCs were seeded into a 96-well plate at a density of 2×106 cells/well, and the fresh medium was supplemented with 20 ng/mL TNF-α or 2.5 µM geraniin for 1 day. BMSCs were lysed using radioimmunoprecipitation assay lysis buffer (Thermo Fisher Scientific, USA) for 20–30 min on ice. The protein concentration was measured using a bicinchoninic acid protein assay (Thermo Fisher Scientific, USA). Thirty microgram proteins were separated using 10%–12% SDS-PAGE (Beyotime Institute of Biotechnology, China), transferred onto polyvinylidene difluoride membranes, blocked with 5% skimmed milk in TBS–Tween (0.2% Tween-20) for 2 hour at 37°C and then incubated with primary antibodies overnight at 4°C including anti-NF-κB/p65 (1:2000, Cell Signaling Technology, USA), anti-IкB-α (1:2000, Cell Signaling Technology, USA), anti-p38 MAPK(1:2000, Cell Signaling Technology, USA) and β-actin (Beyotime Institute of Biotechnology, China). Membranes were incubated with the appropriate secondary antibodies (Beyotime Institute of Biotechnology, China) conjugated with an IRDye 800CW. Immunoblots were analysed with Image-Pro Plus 5.0 software (Media Cybernetics, USA).

, Cell Signaling Technology, USA) and β-actin (Beyotime Institute of Biotechnology, China). Membranes were incubated with the appropriate secondary antibodies (Beyotime Institute of Biotechnology, China) conjugated with an IRDye 800CW. Immunoblots were analysed with Image-Pro Plus 5.0 software (Media Cybernetics, USA). Statistical analysis All data in graphs are generated from at least three independent experiments and expressed as the mean±SD. Student’s t-test was used to determine the statistical significance, and values of p<0.05 were considered to be statistically significant. Results Identification of BMSCs Cell morphology was observed by a light microscope with staining of bromodeoxyuridine. After 24 hours of incubation, most BMSCs had already adhered to the surface of culture flasks (figure 1A). BMSCs exhibited a fibrocyte-like form with the blue nucleus and a long spindle shape after being cultured for 4 days (figure 1B). Figure 1 The morphology of BMSCs was observed by light microscope. (A) Primary BMSCs cultured for 24 hours; (B) primary BMSCs cultured for 4 days. The scale bar is 10 µm. BMSCs, bone marrow-derived mesenchymal stem cells.

Results Identification of BMSCs Cell morphology was observed by a light microscope with staining of bromodeoxyuridine. After 24 hours of incubation, most BMSCs had already adhered to the surface of culture flasks (figure 1A). BMSCs exhibited a fibrocyte-like form with the blue nucleus and a long spindle shape after being cultured for 4 days (figure 1B). Figure 1 The morphology of BMSCs was observed by light microscope. (A) Primary BMSCs cultured for 24 hours; (B) primary BMSCs cultured for 4 days. The scale bar is 10 µm. BMSCs, bone marrow-derived mesenchymal stem cells. Effect of geraniin on TNF-α-induced BMSCs cytotoxicity MTT assay was used to detect the effect of geraniin on TNF-α-induced BMSCs’ viability. After being treated with 2.5 µM geraniin for 1 day, the viability of BMSCs was significantly increased, compared with control group (p<0.01). However, the viability of BMSCs was significantly suppressed with treatment of 20 µM geraniin at 3 days (p<0.01). After being treated with 10 or 20 µM geraniin for 5 days, the BMSCs’ viability significantly decreased (p<0.01) (figure 2A). Therefore, treating with 2.5 µM geraniin for 1 day was selected as optimal condition in this study. The viability of BMSCs was significantly reduced after treating with 20 ng/mL TNF-α (p<0.01); however, 2.5 µM geraniin could significantly improve TNF-α-induced BMSCs cytotoxicity (p<0.05) (figure 2B). Moreover, LDH assay was used to investigate the effect of geraniin on TNF-α-induced BMSCs cytotoxicity. As shown in figure 3, TNF-α showed significant BMSCs cytotoxicity at the dose of 20 ng/mL. In contrast, 2.5 µM geraniin could significantly diminished TNF-α-induced BMSCs cytotoxicity (figure 3).

p<0.05) (figure 2B). Moreover, LDH assay was used to investigate the effect of geraniin on TNF-α-induced BMSCs cytotoxicity. As shown in figure 3, TNF-α showed significant BMSCs cytotoxicity at the dose of 20 ng/mL. In contrast, 2.5 µM geraniin could significantly diminished TNF-α-induced BMSCs cytotoxicity (figure 3). Figure 2 (A) The effects of geraniin with different concentrations on BMSCs’ viability; (B) effect of geraniin on TNF-α-induced BMSC viability. # p<0.01 compared with control group, *p<0.05 compared with 20 ng/mL TNF-α group. BMSCs, bone marrow-derived mesenchymal stem cells; TNF-α, tumour necrosis factor-alpha. Figure 3 Effect of geraniin on TNF-α-induced BMSCs cytotoxicity. # p<0.01 compared with control group, *p<0.05 compared with 20 ng/mL  TNF-α group. BMSCs, bone marrow-derived mesenchymal stem cells; Ger, geraniin; TNF-α, tumour necrosis factor-alpha. Effect of geraniin on osteogenesis in BMSCs To investigate the effect of geraniin on osteogenesis in BMSCs, Oil red O stain was used to detected osteogenesis of BMSCs. As shown in figure 4, the red area of control group (figure 4A) was lower than that of geraniin group (figure 4B). The MFI of BMSCs nucleus in the control group was significantly lower than that of geraniin group (p<0.05) (figure 4C). Figure 4 Effect of geraniin on osteogenesis in BMSCs. *p<0.01 compared with control group; the scale bar is 10 µm. BMSCs, bone marrow-derived mesenchymal stem cells.

Effect of geraniin on osteogenesis in BMSCs To investigate the effect of geraniin on osteogenesis in BMSCs, Oil red O stain was used to detected osteogenesis of BMSCs. As shown in figure 4, the red area of control group (figure 4A) was lower than that of geraniin group (figure 4B). The MFI of BMSCs nucleus in the control group was significantly lower than that of geraniin group (p<0.05) (figure 4C). Figure 4 Effect of geraniin on osteogenesis in BMSCs. *p<0.01 compared with control group; the scale bar is 10 µm. BMSCs, bone marrow-derived mesenchymal stem cells. The expression of miRNA Runx2 and Osx To unveil the mechanisms on the effect of geraniin on TNF-α-induced impairment of BMSCs, the expression of miRNA Runx2 and Osx were measured using real-time PCR in this study. The expression of miRNA Runx2 and Osx was significantly decreased in BMSCs after treating with 20 ng/mL TNF-α compared with control group. In contrast, the expression of miRNA Runx2 and Osx was significantly upregulated by treating with 2.5 µM geraniin (figure 5). Figure 5 The expression of Runx2 (A) and Osx (B) in BMSCs treating with TNF-α or TNF-α+Ger. # p<0.05 compared with control group, *p<0.05 compared with TNF-α group. BMSCs, bone marrow-derived mesenchymal stem cells; Ger, geraniin; TNF-α, tumour necrosis factor-alpha.

The expression of miRNA Runx2 and Osx To unveil the mechanisms on the effect of geraniin on TNF-α-induced impairment of BMSCs, the expression of miRNA Runx2 and Osx were measured using real-time PCR in this study. The expression of miRNA Runx2 and Osx was significantly decreased in BMSCs after treating with 20 ng/mL TNF-α compared with control group. In contrast, the expression of miRNA Runx2 and Osx was significantly upregulated by treating with 2.5 µM geraniin (figure 5). Figure 5 The expression of Runx2 (A) and Osx (B) in BMSCs treating with TNF-α or TNF-α+Ger. # p<0.05 compared with control group, *p<0.05 compared with TNF-α group. BMSCs, bone marrow-derived mesenchymal stem cells; Ger, geraniin; TNF-α, tumour necrosis factor-alpha. Effect of geraniin on TNF-α-activated signal pathways in BMSCs To investigate mechanisms on the effect of geraniin on TNF-α-induced impairment of BMSCs, NF-κB/p65, p-IкB-α and p38 MAPK protein expression was detected using western blot analysis. As shown in figure 6A, the expression of NF-κB/p65 and p38 MAPK protein in TNF-α-treated BMSCs was higher than that in control group; however, TNF-α significantly inhibited the expression of p-IкB-α protein in BMSCs. After treating with 2.5 µM geraniin, the expression of NF-κB/p65 and p38 MAPK significantly reduced in BMSCs. In contrast, geraniin significantly increased the expression of p-IкB-α protein in TNF-α-treated BMSCs (figure 6B-D).

ol group; however, TNF-α significantly inhibited the expression of p-IкB-α protein in BMSCs. After treating with 2.5 µM geraniin, the expression of NF-κB/p65 and p38 MAPK significantly reduced in BMSCs. In contrast, geraniin significantly increased the expression of p-IкB-α protein in TNF-α-treated BMSCs (figure 6B-D). Figure 6 Effect of geraniin on TNF-α-activated NF-κB, p-IкB-α and p38 MAPK protein expression in BMSCs, including western blot analysis (A) and statistics of NF-κB (B), p-IкB-α (C) and p38 MAPK (D) protein expression. # p<0.01 compared with control group, *p<0.05 compared with TNF-α group. BMSCs, bone marrow-derived mesenchymal stem cells; Ger, geraniin; TNF-α, tumour necrosis factor-alpha.

K protein expression in BMSCs, including western blot analysis (A) and statistics of NF-κB (B), p-IкB-α (C) and p38 MAPK (D) protein expression. # p<0.01 compared with control group, *p<0.05 compared with TNF-α group. BMSCs, bone marrow-derived mesenchymal stem cells; Ger, geraniin; TNF-α, tumour necrosis factor-alpha. Discussion BMSCs are pluripotent stem cells derived from various organs or tissues. In the normal conditions, most of BMSCs stay in the dormant state; however, under pathological conditions, the renewability of BMSCs will be activated.21 BMSCs can be differentiated into osteoblasts, chondrocytes, adipocytes and myoblasts.4 Various evidences demonstrated that BMSCs played an essential role in osteogenesis and inflammatory cytokines could disrupt the balance between bone formation and bone resorption.1 2 TNF-α is an important inflammatory cytokine regulating proliferation and differentiation of BMSCs, resulting in osteoclast-induced bone destruction and inhibiting osteogenesis.10–12 In this study, we demonstrated that 2.5 µM geraniin could decrease TNF-α-induced BMSCs cytotoxicity. Lu et al indicated that geraniin was able to prevent ovariectomy-induced osteoporosis in rats.16 Park et al suggested that soybean-treated osteoblasts suppressed TNF-α-stimulated osteoclast formation and RANKL-induced osteoclast differentiation through increasing OPG/RANKL ratio.22 Previous study showed that geraniin played a role of TNF-α inhibitor on impairments of osteogenesis.18 In this study, we found that geraniin could significantly improve expression of Runx2 and Osx miRNA in BMSCs. Velázquez-González et al also showed that geraniin possessed antinociceptive and anti-inflammatory activities.23 As we all know, Runx2 is the most important osteogenic transcriptional factor playing a determinant role in the early stage of osteogenesis and is a significant marker of early osteogenic differentiation.24 It has been proved that TNF-α can stimulate osteoclast formation via NF-κB pathway.25 Recently, Fujii et al reported that TNF-α promoted the differentiation of osteogenesis through activation of NF-κB pathways.26 In this study, geraniin significantly suppressed TNF-α-induced expression of NF-κB/p65 protein and promoted p-IкB-α protein expression in BMSCs. Similarly, Xiao et al suggested that geraniin suppressed RANKL-induced NF-κB expression.18 Moreover, we suggested that geraniin could significantly reduce the TNF-α-induced p38 MAPK protein expression in BMSCs.

ly suppressed TNF-α-induced expression of NF-κB/p65 protein and promoted p-IкB-α protein expression in BMSCs. Similarly, Xiao et al suggested that geraniin suppressed RANKL-induced NF-κB expression.18 Moreover, we suggested that geraniin could significantly reduce the TNF-α-induced p38 MAPK protein expression in BMSCs. Adhesion molecules are produced through stimulation of BMSCs by p38 MAPK signalling pathways.27 TNF-α can promote the secretion of Hepatocyte growth factor by BMSCs through p38 MAPK and PI3K/Akt pathways.28 Thus, NF-κB, ERK and c-Jun-NH2-terminal Kinase (JNK) pathways play an essential role for migration homing and adhesion of BMSCs.29 In summary, geraniin can inhibit TNF-α-induced impairments in BMSCs. Moreover, the mechanisms about protective effect of geraniin are mediated through NF-κB, IкB-α and p38 MAPK signalling pathways in BMSCs. It suggests that geraniin may be a new candidate drug for treating TNF-α-induced impairment of osteogenesis. Contributors: CL: guarantor of integrity of the entire study, study concepts and manuscript review; CL and SG: study design, statistical analysis and manuscript editing; GX: definition of intellectual content; SG and GX: experimental studies, data acquisition, data analysis and manuscript preparation. Competing interests: None declared. Provenance and peer review: Not commissioned; externally peer reviewed.

Introduction Stroke is the second leading cause of death and causes excessive loss of disability-adjusted life-years every year worldwide.1 Despite the reduction in stroke mortality, the absolute number of people with stroke, stroke survivors, related death as well as global burden had increased greatly in the past two decades.2 In this regard, the prevention of stroke by early intervention is of great importance. Since the conventional risk factors cannot fully account for the pathogenesis, it is essential to detect unknown stroke risk factors especially biomarker of artery injury for an appropriate intervention. Arterial stiffness, also known as the loss of arterial elasticity, represents the mechanical property of artery resistant to deformation.3 Compliance and distensibility, although related to arterial stiffness, are not interchangeable with arterial stiffness because they depend on the stiffness of arteries, as well as on the size and thickness of arteries.4 Arterial stiffness has been regarded as a reliable marker of arterial structural and functional alteration after abundant experimental and clinical studies.3 5 Furthermore, a growing number of studies have demonstrated the association between arterial stiffness and stroke attack.6–10 The goal of this review is to address arterial stiffness with the following aspects: the measurements, the secondary haemodynamic consequences and the predictive role, the possible pathophysiological mechanism and de-stiffening therapy for stroke prevention.

Arterial stiffness, also known as the loss of arterial elasticity, represents the mechanical property of artery resistant to deformation.3 Compliance and distensibility, although related to arterial stiffness, are not interchangeable with arterial stiffness because they depend on the stiffness of arteries, as well as on the size and thickness of arteries.4 Arterial stiffness has been regarded as a reliable marker of arterial structural and functional alteration after abundant experimental and clinical studies.3 5 Furthermore, a growing number of studies have demonstrated the association between arterial stiffness and stroke attack.6–10 The goal of this review is to address arterial stiffness with the following aspects: the measurements, the secondary haemodynamic consequences and the predictive role, the possible pathophysiological mechanism and de-stiffening therapy for stroke prevention. Measurements of artery stiffness in clinical investigation There are various parameters to present systemic and regional arterial stiffness by different invasive or non-invasive methods. Here, we mainly discuss three major measurements of arterial stiffness that are generally applied in clinical researches.