Browse the corpus

Key messages The paper describes a survey sent to all members of the British Thoracic Society to find out about attitudes and knowledge around lung volume reduction surgery, showing that there is uncertainty and overestimation of the risks associated with the procedure. The response rate was low, but that itself may reflect a lack of engagement with lung volume reduction surgery as a treatment for people with emphysema. Although the low response rate impacts on the precision of some of the estimates the responders are likely to have been better informed and results from a larger group are unlikely to have presented a ‘better’ picture. Background Chronic obstructive pulmonary disease (COPD) is a major cause of disability and mortality in the UK and is now the third most frequent cause of death worldwide1 2 as a consequence of the ongoing epidemic of tobacco addiction.3 The pathological processes involved, destruction of small airways and lung parenchyma as well as narrowing of larger airways, are poorly responsive to medical therapies and many patients remain severely disabled despite optimum medical therapy.4 Only a handful of treatments including smoking cessation, long-term oxygen therapy in selected patients and lung volume reduction surgery (LVRS) have been shown to improve prognosis.5–8

arger airways, are poorly responsive to medical therapies and many patients remain severely disabled despite optimum medical therapy.4 Only a handful of treatments including smoking cessation, long-term oxygen therapy in selected patients and lung volume reduction surgery (LVRS) have been shown to improve prognosis.5–8 LVRS was described as a palliative treatment for emphysema by Brantigan and Muller in 1957.9 The aim of LVRS is to resect the most emphysematous portion of the lung. This allows healthier, less compliant areas to be ventilated more effectively, reduces operating lung volumes and thus improves chest wall and respiratory muscle mechanics. In the US National Emphysema Treatment Trial (NETT),5 6 patients with an upper lobe predominant pattern of emphysema and a low preoperative exercise capacity gained the maximum benefit and had increased survival, which long-term economic analysis suggests is achieved at an acceptable cost per quality-adjusted life year.10 The NETT study also identified a group of patients with an excess risk of surgical mortality: those with a forced expiratory volume 1 s (FEV1) <20% predicted and either homogeneous disease or a carbon monoxide transfer factor <20% predicted. The mortality rate for patients excluding this high-risk group was 5.2% at 90 days. In the NETT trial, complications included major respiratory or cardiac complication in 29.8% and 20% of patients, respectively. At 1 month, 28.1% of patients were still hospitalised or in a long-term care facility. However, audit of current practice suggests that mortality and morbidity are significantly lower.11

days. In the NETT trial, complications included major respiratory or cardiac complication in 29.8% and 20% of patients, respectively. At 1 month, 28.1% of patients were still hospitalised or in a long-term care facility. However, audit of current practice suggests that mortality and morbidity are significantly lower.11 Data from the UK Society of Cardiothoracic Surgery (SCTS) register, http://www.scts.org/professionals/audit_outcomes.aspx show that only 96 procedures in 2009–2010 and 90 in 2010–2011 were recorded. The likely pool of eligible patients is significantly larger than this12 and there is therefore concern that patients who may benefit are not being considered for treatment. Some authors have suggested that physicians are deterred from referral as LVRS is perceived as too complicated with limited patient benefit and a substantial risk of complications.5 To improve understanding of the obstacles limiting provision of LVRS to suitable patients with COPD, we undertook a survey of members of the British Thoracic Society (BTS). Methods A survey including questions about clinicians’ attitudes to and knowledge of LVRS was designed in collaboration with the British Thoracic Society Professional and Organisational Standards of Care Committee of the BTS. The survey is available in the online supplementary material. The survey was completed electronically and a link to complete it sent out to 2498 BTS members in the monthly e-newsletter on two occasions.

llaboration with the British Thoracic Society Professional and Organisational Standards of Care Committee of the BTS. The survey is available in the online supplementary material. The survey was completed electronically and a link to complete it sent out to 2498 BTS members in the monthly e-newsletter on two occasions. Results There were 65 respondents to the questionnaire with replies from all UK nations and 13 different regions in total. A total of 82% were consultant physicians and 11% specialist trainees in respiratory medicine (table 1). The remainder comprised two respiratory nurses, one thoracic surgeon, one physiotherapist and one clinical physiologist. Twenty two (34%) identified themselves as the COPD lead for their organisation. Table 1 Summary of question results

Results There were 65 respondents to the questionnaire with replies from all UK nations and 13 different regions in total. A total of 82% were consultant physicians and 11% specialist trainees in respiratory medicine (table 1). The remainder comprised two respiratory nurses, one thoracic surgeon, one physiotherapist and one clinical physiologist. Twenty two (34%) identified themselves as the COPD lead for their organisation. Table 1 Summary of question results Total Percentage What is your role? (n=65) Consultant physician 53 82 Consultant surgeon 1 2 Specialty trainee 7 11 Specialty trainee (other specialty) 0 0 Foundation trainee 0 0 Physiotherapist 1 2 Lung physiologist 1 2 Respiratory nurse 2 3 Where is your main place of work? (n=63) Secondary care—DGH 30 48 Secondary care—teaching hospital 33 52 Are you the COPD lead for your organisation? (n=64) Yes 22 34 No 42 66 I have referred a patient for consideration of LVRS within the past 12 months (n=63) Yes 43 68 No 18 29 Not applicable 2 3 If ‘yes’, approximately how many patients have you referred for LVRS in the past 12 months? (n=45) 1–3 35 78 4–6 6 13 More than 6 3 7 Not known 1 2 I know how to refer patients for LVRS (n=63) Yes 61 97 No 1 2 Not sure 1 2 Do you have a specific MDT meeting to discuss the management of patients with advanced COPD? (n=64) Yes 18 28 No 45 70 Not known 0 0 Not applicable 1 2 Does the thoracic surgical service you use have an MDT to discuss potential LVRS patients? (n=63) Yes 34 54 No 14 22 Not known 14 22 Not applicable 1 2 How far away is the nearest LVRS service? (estimate) (n=64) On site 17 27 Less than 20 miles away 26 41 Between 20 and 40 miles away 8 12 Over 40 miles away 12 19 Not known 1 2 What do you estimate the 30 day mortality is following LVRS? (n=63) (%) 0–5 33 52 6–10 9 14 11–15 1 2 >15 3 5 Not known 17 27 What proportion of patients do you think would still be in hospital 30 days following LVRS? (n=65) (%) 0–4 10 16 5–9 17 27 10–15 14 22 >15 6 9 Not known 18 28 Which group of patients derive the most benefit from LVRS? (n=65) Homogeneous emphysema, low exercise capacity 10 15 Homogeneous emphysema, high exercise capacity 3 5 Heterogeneous emphysema, low exercise capacity 41 63 Heterogeneous emphysema, high exercise capacity 12 18 Not known 3 3 CT scanning may be indicated for haemoptysis, recurrent exacerbations or to investigate hypoxia, looking at pulmonary arteries or for interstitial fibrosis. Excluding these specific indications, do you think a CT of the thorax is indicated routinely in patients with an FEV1 <50% predicted?

capacity 12 18 Not known 3 3 CT scanning may be indicated for haemoptysis, recurrent exacerbations or to investigate hypoxia, looking at pulmonary arteries or for interstitial fibrosis. Excluding these specific indications, do you think a CT of the thorax is indicated routinely in patients with an FEV1 <50% predicted? (n=65) Strongly agree 4 6 Agree 17 26 Neither agree nor disagree 15 23 Disagree 25 38 Strongly disagree 4 6 Various bronchoscopic techniques which are intended for lung volume reduction are being developed—which option best applies to your organisation? (n=63) We currently offer bronchoscopic lung volume reduction 8 13 We intend to offer bronchoscopic lung volume reduction 13 21 We have no plans to offer bronchoscopic lung volume reduction 37 59 Not applicable 3 5 Not known 2 3 COPD, chronic obstructive pulmonary disease; DGH, district general hospital; LVRS, lung volume reduction surgery; MDT, multidisciplinary team. Indications for LVRS, morbidity and mortality In identifying which patients would derive the most benefit from LVRS, 60% of respondents correctly identified those with heterogeneous emphysema and a low exercise capacity. Estimates for a 30-day mortality were 0–5% in 52% of respondents, 21% overestimated this and 27% did not know. A third of the respondents thought that more than 10% of patients would be in hospital 1 month postprocedure and 28% were unsure.

ctly identified those with heterogeneous emphysema and a low exercise capacity. Estimates for a 30-day mortality were 0–5% in 52% of respondents, 21% overestimated this and 27% did not know. A third of the respondents thought that more than 10% of patients would be in hospital 1 month postprocedure and 28% were unsure. Process issues around LVRS Roughly half of the respondents (52%) worked in a teaching hospital and the remainder within district general hospitals. A lung volume reduction service was available in 27% of respondents’ hospitals and a further 41% had a centre within 20 miles. When asked about the referral of patients for LVRS, 68% had referred patients for LVRS in the past 12 months and 97% said that they knew how to make a referral for LVRS. The number of referrals by each respondent is listed in table 2. The majority of respondents (70%) did not have a multidisciplinary team (MDT) to discuss such cases within their hospital, although 54% answered that an MDT was available in their local referral centre. Of those respondents who worked in a referral centre, 65% had an MDT to discuss LVRS cases. Bronchoscopic lung volume reduction was available in the hospital of 13% of respondents with a further 21% planning to offer a service in the future. Table 2 Access to investigations for lung volume reduction surgery

Process issues around LVRS Roughly half of the respondents (52%) worked in a teaching hospital and the remainder within district general hospitals. A lung volume reduction service was available in 27% of respondents’ hospitals and a further 41% had a centre within 20 miles. When asked about the referral of patients for LVRS, 68% had referred patients for LVRS in the past 12 months and 97% said that they knew how to make a referral for LVRS. The number of referrals by each respondent is listed in table 2. The majority of respondents (70%) did not have a multidisciplinary team (MDT) to discuss such cases within their hospital, although 54% answered that an MDT was available in their local referral centre. Of those respondents who worked in a referral centre, 65% had an MDT to discuss LVRS cases. Bronchoscopic lung volume reduction was available in the hospital of 13% of respondents with a further 21% planning to offer a service in the future. Table 2 Access to investigations for lung volume reduction surgery Easy to access Hard to access Unavailable Not known n Percentage n Percentage n Percentage n Percentage Plethysmographic lung volumes 53 82 2 3 10 15 0 0 Gas transfer 65 100 0 0 0 0 0 0 Quantitative perfusion scanning 36 55 10 15 10 15 9 14 Walking tests 53 82 8 12 4 6 0 0 Pulmonary rehabilitation 60 92 4 6 1 2 0 0 Access to investigations for the work-up for LVRS is detailed in table 2. When asked whether a CT scan was indicated for patients with COPD with an FEV1 <50% predicted, 32% agreed or strongly agreed while 44% disagreed or strongly disagreed.

14 Walking tests 53 82 8 12 4 6 0 0 Pulmonary rehabilitation 60 92 4 6 1 2 0 0 Access to investigations for the work-up for LVRS is detailed in table 2. When asked whether a CT scan was indicated for patients with COPD with an FEV1 <50% predicted, 32% agreed or strongly agreed while 44% disagreed or strongly disagreed. In the free text section, 18 respondents left comments. One surgeon felt that there were likely to be many more patients who may benefit but were not being recognised as potential candidates. Two physicians commented that there was a misconception regarding the benefits and risks of surgery. Another remarked that referral criteria, particularly regarding exercise capacity, were not always clear. One physician added that determining emphysema heterogeneity was rarely performed at their centre. Access to pulmonary rehabilitation was mentioned by two respondents. Three physicians commented that they were increasingly opting for bronchoscopic procedures, while another felt that there was insufficient evidence for their use. Discussion The main findings from the present survey are, first, that there are significant information needs around the indications for LVRS and the accompanying risks of morbidity and mortality. Second, there is a lack of systematic structures to evaluate patients and, third, there is a lack of consensus about the best approach for screening individuals to identify potential candidates for LVRS.

significant information needs around the indications for LVRS and the accompanying risks of morbidity and mortality. Second, there is a lack of systematic structures to evaluate patients and, third, there is a lack of consensus about the best approach for screening individuals to identify potential candidates for LVRS. LVRS for selected patients is recommended in national and international guidelines for the management of COPD.13 14 This survey suggests that the majority of those surveyed know how to refer patients for LVRS and had done so within the last year. A significant proportion of respondents were clinical leads for COPD in their organisation and the majority had referred a patient for potential LVRS in the last year, suggesting that the respondents were likely to represent the more engaged clinicians in this area. However, survey responses suggest that published historical data from the late 1990s and early 21st century inform many individuals’ assessment of risk. In current surgical practice with largely unilateral, thoracoscopic approaches for LVRS, the morbidity and mortality are significantly lower than reported in the NETT trial5 with no deaths within 90 days and only 6% of patients in hospital at 30 days reported in one recent series.11

any individuals’ assessment of risk. In current surgical practice with largely unilateral, thoracoscopic approaches for LVRS, the morbidity and mortality are significantly lower than reported in the NETT trial5 with no deaths within 90 days and only 6% of patients in hospital at 30 days reported in one recent series.11 Most hospitals appear to have easy access to appropriate investigations, but there was no consensus on when a CT of the thorax is indicated. It has been proposed that a routine assessment of the pattern of emphysema by CT scan, as well as gas transfer measurement, should be considered in all patients with COPD with Medical Research Council dyspnoea scores of 4 or 5 and an FEV1<50%, unless there are obvious comorbidities precluding surgery, with review by an MDT including chest physicians, surgeons and radiologists, as is already the case for the management of lung cancer.15 Clearly, this requires decisions about the appropriate allocation of resources to ensure best value. The cost per quality-adjusted life year of LVRS in the NETT study in upper lobe predominant emphysema was estimated to be $48 000 for low exercise capacity and $40 000 for high exercise capacity patients at 10 years.10 The true cost in current practice is likely to be considerably lower than this as the costs are driven by early surgical morbidity and mortality, which are lower now.11 16

pper lobe predominant emphysema was estimated to be $48 000 for low exercise capacity and $40 000 for high exercise capacity patients at 10 years.10 The true cost in current practice is likely to be considerably lower than this as the costs are driven by early surgical morbidity and mortality, which are lower now.11 16 The response rate to the survey was low, which impacts on the precision of the findings. We have no data as to the reasons for non-response, and can therefore only speculate. However, as a self-selecting group, respondents may be expected to have been more interested in LVRS, so it is unlikely that responses from a larger sample would have produced ‘better’ results. Clearly, people may have been too busy to respond, but non-response itself may also represent a general lack of engagement with lung volume reduction strategies in COPD. The overestimation of mortality risk by half of the respondents and the overestimation of hospital stay by 60% of the respondents may contribute to this disengagement. A further consequence of this is that low referral rates have made the development of bronchoscopic approaches,17–20 intended to deliver lung volume reduction either more safely or in different emphysema phenotypes,21 22 more difficult.

on of hospital stay by 60% of the respondents may contribute to this disengagement. A further consequence of this is that low referral rates have made the development of bronchoscopic approaches,17–20 intended to deliver lung volume reduction either more safely or in different emphysema phenotypes,21 22 more difficult. Lung volume reduction has a strong evidence base in appropriately selected patients with COPD where, unlike current pharmacotherapy, it can modify the natural history of the disease.5 19 23 The survey confirms that work is needed to ensure that clinicians are aware of the risks and benefits associated with the technique in modern practice and that structures are put in place to ensure systematic evaluation of patients. The assistance of the British Thoracic Society membership is gratefully acknowledged. Contributors: NSH and SJ conceived the study. NSH and WM analysed the data. WM prepared the first draft and all authors contributed to and approved the final manuscript. NSH is the guarantor. Funding: The study was supported by the NIHR Biomedical Research Unit at Royal Brompton and Harefield NHS Foundation Trust and Imperial College, London. Competing interests: None. Provenance and peer review: Not commissioned; externally peer reviewed. Data sharing statement: No additional data are available.

Introduction Sputum production is a cardinal feature of chronic obstructive pulmonary disease (COPD), though it varies between individuals,1 with many patients remaining symptomatic despite optimal medical treatment.2 Coughing to clear sputum can be tiring and uncomfortable, and patients report embarrassment coughing and disposing of sputum in public. Therapeutic approaches include smoking cessation,3 inhaled bronchodilators and corticosteroids. The active cycle of breathing technique helps patients clear sputum by ‘huffing’ rather than coughing.4 In some individuals, carbocisteine reduces sputum viscosity, making it easier to clear and reducing exacerbations.1

lic. Therapeutic approaches include smoking cessation,3 inhaled bronchodilators and corticosteroids. The active cycle of breathing technique helps patients clear sputum by ‘huffing’ rather than coughing.4 In some individuals, carbocisteine reduces sputum viscosity, making it easier to clear and reducing exacerbations.1 Handheld oscillatory positive expiratory pressure (OPEP) devices require patients to exhale against a fluctuating resistance. The positive pressure prevents airway collapse, sustaining expiratory flow. The vibration helps to mobilise sputum, making it easier to clear. A Cochrane review5 of the use of OPEP devices in COPD suggests that they can improve short-term health status and may reduce respiratory-related admissions to hospital over the longer term, as well as improving exercise capacity. However, these conclusions are based on only a few studies in small numbers of patients. Neither Global Initiative on Obstructive Lung Disease 20171 nor joint American Thoracic Society/European Respiratory Society COPD guidelines6 make any reference to sputum clearance techniques (searched using the words ‘sputum’, ‘clearance’ and ‘physiotherapy’). The joint British Thoracic Society/Association of Chartered Physiotherapist in Respiratory Care (ACPRC) recommendation is ‘consider … OPEP for patients with stable COPD who need an airway clearance technique to assist in the removal of secretions. (Grade C)’.4 ‘Need’ for an OPEP device is not defined.

‘physiotherapy’). The joint British Thoracic Society/Association of Chartered Physiotherapist in Respiratory Care (ACPRC) recommendation is ‘consider … OPEP for patients with stable COPD who need an airway clearance technique to assist in the removal of secretions. (Grade C)’.4 ‘Need’ for an OPEP device is not defined. We therefore analysed prescribing data for the actual usage of therapies to improve sputum clearance as well as clinicians’ attitudes to the use of OPEP devices in COPD, to help to define research questions in this area. Methods English prescriptions for sputum clearance treatments English prescribing data on prescription of carbocisteine, together with that for OPEP devices for the years 2013–2015 were obtained from OpenPrescribing.net. Prescriptions for paediatric preparations of carbocisteine were excluded. We assumed that each drug item represented 1 month of treatment, multiplying item numbers by 12 to calculate treatment/years received. Clinical attitudes to use of sputum adjunct devices An online survey (using SurveyMonkey.com) was sent by email to members of the ACPRC. This included questions about awareness of sputum adjunct devices and likelihood of using them in patients with COPD with varying patterns of daily sputum production/tenacity and exacerbation frequency (online supplementary file 1). Options were presented in random order to avoid bias. 10.1136/bmjresp-2017-000226.supp1Supplementary file 1

Clinical attitudes to use of sputum adjunct devices An online survey (using SurveyMonkey.com) was sent by email to members of the ACPRC. This included questions about awareness of sputum adjunct devices and likelihood of using them in patients with COPD with varying patterns of daily sputum production/tenacity and exacerbation frequency (online supplementary file 1). Options were presented in random order to avoid bias. 10.1136/bmjresp-2017-000226.supp1Supplementary file 1 Results Prescribing data There were 1.07 million patients with COPD on general practitioner registers in England in 2015 (via www.gpcontract.co.uk and the period 2013–2015 covers a total of 3.2 million COPD patient-years. Over this period, an estimated 422 744 patient-years of treatment with carbocisteine were prescribed (table 1) and 1.1 million years treatment with tiotropium. By contrast, only 4989 sputum adjunct devices were prescribed over the 3 years (all but seven were Flutter and Acapella). The costs for these were £73 million, £483 million and £187 000, respectively. Table 1 UK prescribing data for sputum clearance medications and devices Year Carbocisteine (treatment years) Carbocisteine cost Tiotropium (treatment years) Tiotropium cost OPEP device (n) OPEP device cost 2013 102 203 £18 731 563 295 691 £129 183 487 976 £36 570 2014 152 242 £25 929 656 403 363 £175 912 591 1836 £68 791 2015 168 299 £28 316 401 409 224 £178 550 070 2177 £81 509 Total 422 744 £72 977 620 1 108 278 £483 646 148 4989 £1 86 870 Data from OpenPrescribing.net, EBM DataLab, University of Oxford, 2017 (https://openprescribing.net/).

63 295 691 £129 183 487 976 £36 570 2014 152 242 £25 929 656 403 363 £175 912 591 1836 £68 791 2015 168 299 £28 316 401 409 224 £178 550 070 2177 £81 509 Total 422 744 £72 977 620 1 108 278 £483 646 148 4989 £1 86 870 Data from OpenPrescribing.net, EBM DataLab, University of Oxford, 2017 (https://openprescribing.net/). Treatment year was calculated by dividing number of items prescribed per year by 12. For carbocisteine, paediatric preparations were excluded. OPEP, oscillatory positive expiratory pressure. Survey responses There were 116 responses (12% response rate), 72% in hospital practice, 28% based in the community (table 2). Ninety-six per cent and 95% were aware of the Acapella and Flutter, the two most commonly prescribed devices. 72.9% would usually or always advocate a device in a patient with COPD with daily, difficult to clear sputum, whereas 83.8% would rarely or never recommend one for individual-producing sputum during acute exacerbations of COPD only with 0–1 exacerbations/year (figure 1). There were significant areas of variation. 35.1% would usually or always recommend a device for individuals having >4 exacerbations per year, even if they did not produce sputum at other times. 38.7% would ‘sometimes’ recommend a device for people with morning sputum only. Figure 1 How often would responders recommend oscillatory positive pressure devices for sputum clearance in a patient with this chronic obstructive pulmonary disease phenotype?

Survey responses There were 116 responses (12% response rate), 72% in hospital practice, 28% based in the community (table 2). Ninety-six per cent and 95% were aware of the Acapella and Flutter, the two most commonly prescribed devices. 72.9% would usually or always advocate a device in a patient with COPD with daily, difficult to clear sputum, whereas 83.8% would rarely or never recommend one for individual-producing sputum during acute exacerbations of COPD only with 0–1 exacerbations/year (figure 1). There were significant areas of variation. 35.1% would usually or always recommend a device for individuals having >4 exacerbations per year, even if they did not produce sputum at other times. 38.7% would ‘sometimes’ recommend a device for people with morning sputum only. Figure 1 How often would responders recommend oscillatory positive pressure devices for sputum clearance in a patient with this chronic obstructive pulmonary disease phenotype? Table 2 Survey responses regarding the likelihood of using devices for varying clinical patterns of chronic obstructive pulmonary disease (COPD)

Figure 1 How often would responders recommend oscillatory positive pressure devices for sputum clearance in a patient with this chronic obstructive pulmonary disease phenotype? Table 2 Survey responses regarding the likelihood of using devices for varying clinical patterns of chronic obstructive pulmonary disease (COPD) COPD clinical pattern How often would you consider the use of an adjunct device to help sputum clearance in the following situation? For this question, assume that the person has been taught active cycle of breathing techniques (%) Never Rarely Sometimes Usually Always Daily difficult to clear thick sputum 0.9 5.4 20.7 38.7 34.2 Sputum throughout the day but able to clear it 20.7 37.8 27.9 8.1 5.4 Morning sputum only 12.6 37.8 38.7 7.2 3.6 Sputum only with exacerbations (0–1 exacerbations/year) 33.3 50.5 11.7 2.7 1.8 Sputum only with exacerbations (2–3 exacerbations/year) 8.1 29.7 45.1 15.3 1.8 Sputum only with exacerbations (>4 exacerbations/year) 3.6 19.8 41.4 22.5 12.6 Asked how they would usually provide a device, 58% of respondents had devices available to give out themselves from stock, 13% were able to prescribe them, 12% advised patients to buy them, 13% did not provide them and 5% no response. Recalling actual use in the preceding year, 21.6% had used the Acapella, 14.4% the Flutter and 1.8% the positive expiratory pressure (PEP) mask for more than five patients with COPD; 48%, 59% and 79%, respectively, replied none (table 3). Asked to choose between the Acapella, Flutter and PEP, if only one were available, preferences were 69%, 24% and 6%, respectively.

had used the Acapella, 14.4% the Flutter and 1.8% the positive expiratory pressure (PEP) mask for more than five patients with COPD; 48%, 59% and 79%, respectively, replied none (table 3). Asked to choose between the Acapella, Flutter and PEP, if only one were available, preferences were 69%, 24% and 6%, respectively. Table 3 Reported frequency of use of sputum adjunct devices for patients with COPD in the preceding year No of patients treated with device None 1–2 3–5 >5 Flutter 59.5% 18.0% 8.1% 14.4% Acapella 47.8% 22.5% 8.1% 21.6% PEP mask 79.3% 15.3% 3.6% 1.8% Discussion Prescribing data reveal a striking disparity between the volume of prescription of medications for sputum clearance and use of adjunct devices; 13.2% of COPD patient-years were treated with carbocisteine, whereas the equivalent figure for sputum adjunct devices is 0.16%. Precise data on phenotypes are not available, but there is likely to be a substantial overlap between patients on carbocisteine who would benefit from OPEP devices. The recent Royal College of Physicians COPD audit demonstrated higher use of inhaled medication compared with provision of pulmonary rehabilitation and smoking cessation,7 despite evidence that the latter are higher-value therapies.8 9 Our data raise further concern about the underutilisation of non-pharmacological therapies in COPD.

l College of Physicians COPD audit demonstrated higher use of inhaled medication compared with provision of pulmonary rehabilitation and smoking cessation,7 despite evidence that the latter are higher-value therapies.8 9 Our data raise further concern about the underutilisation of non-pharmacological therapies in COPD. Although the 100-fold difference between use of carbocisteine and OPEP devices in COPD is likely to be broadly accurate, a number of issues limit the precision of this estimate. Some prescriptions of carbocisteine or adjunct devices will have been for bronchiectasis or other conditions. Also, we estimated each drug item as representing 1 month's treatment but they might have lasted more or less than a month depending on pattern of use. Adjunct devices were assumed to last 1 year. We are not able to capture use of devices by patients with COPD who obtained them directly from hospital or bought the devices themselves. However, the vast majority of patients are managed in primary care. As the data are based on total prescribing, we are unable to comment on what proportion of the patients taking carbocisteine had also received an OPEP device.

patients with COPD who obtained them directly from hospital or bought the devices themselves. However, the vast majority of patients are managed in primary care. As the data are based on total prescribing, we are unable to comment on what proportion of the patients taking carbocisteine had also received an OPEP device. A Cochrane review5 of the use of airway clearance devices in COPD suggests that they can improve short-term health status and may reduce respiratory-related admissions over the longer term, as well as improving exercise capacity. However, these conclusions are based on only a few studies in small numbers of patients. A German randomised controlled trial (RCT) in 50 patients with severe COPD found that use of an OPEP device, at least three times a day for 5 min, led to significantly fewer patients (13 vs 24) needing antibiotics in comparison with controls, with fewer requiring hospital admission (5 vs 12) (OR 0.08–0.95) though the number of hospital days was the same in both arms.10 By contrast, a 1990 6-month RCT in 47 patients with a non-oscillating device did not find any impact on exacerbations.11 One study looked at exercise capacity over a longer term. In 20 patients, 3 months of OPEP was associated with an improved six minute walk distance compared with a sham device (mean difference 11 m (95% CI 66 to 156 m)).12 A 3-week OPEP device cross-over study13 in 14 sputum producers and 13 non-producers showed an improvement in St George's Respiratory Questionnaire (SGRQ), ease of bringing up sputum (PEQ score), forced vital capacity, 6MWD and a reduced 3He MRI ventilation deficit per cent in sputum producers. Another study found that short-term use of a Flutter device can reduce airways resistance even in patients with COPD with minimal sputum production.14

iratory Questionnaire (SGRQ), ease of bringing up sputum (PEQ score), forced vital capacity, 6MWD and a reduced 3He MRI ventilation deficit per cent in sputum producers. Another study found that short-term use of a Flutter device can reduce airways resistance even in patients with COPD with minimal sputum production.14 Thus, there are data suggesting a short-term benefit, but little information to guide long-term use or patient selection. Physiotherapists surveyed were more likely to recommend adjuncts in patients with COPD with frequent exacerbations and in those with difficult to clear sputum. However, there is significant variation in their reported thresholds for treatment, reinforcing the need for trials in specific patient phenotypes to inform clearer guidelines. Survey respondents were self-selecting which may have introduced bias. Other sputum adjunct devices are becoming available; however, only seven of these were prescribed in 2015 so we did not include them. Conclusion The absence of sputum clearance devices from COPD guidelines in part reflects the dearth of long-term clinical trial data, and there are insufficient studies to recommend precise evidence-based criteria for their use. However, it is clear that far fewer devices are being prescribed than are included in the phenotypes that clinicians believe they are effective in. This suggests significant unmet need and an important disparity between actual usage and clinical consensus, highlighting the urgent need for further research regarding OPEP device usage in COPD. Twitter: Follow NS Hopkinson @COPDdoc

Conclusion The absence of sputum clearance devices from COPD guidelines in part reflects the dearth of long-term clinical trial data, and there are insufficient studies to recommend precise evidence-based criteria for their use. However, it is clear that far fewer devices are being prescribed than are included in the phenotypes that clinicians believe they are effective in. This suggests significant unmet need and an important disparity between actual usage and clinical consensus, highlighting the urgent need for further research regarding OPEP device usage in COPD. Twitter: Follow NS Hopkinson @COPDdoc Contributors: RB and NSH developed the study. AAL analysed the prescribing data. RB prepared the first draft of this paper which all authors subsequently contributed to and approved. NSH is the guarantor. Competing interests: None declared. Patient consent: Obtained. Provenance and peer review: Not commissioned; internally peer reviewed. Data sharing statement: The data sources used for this work are publically available.