Browse the corpus

4 and tumour necrosis factor-α (TNF-α),12 15 have been shown to be associated with AMD; however, the results from different groups are inconsistent. The fact that systemic inflammatory markers are not strongly related with AMD suggests that local low-grade inflammation is more likely to be involved in its pathogenesis. Over the past few decades, there has been an increasing interest in the role of omega-3 (ω-3) polyunsaturated fatty acids (PUFAs) in inflammation. Evidence from preclinical and clinical studies has proven the effectiveness of ω-3 PUFAs against heart disease, cancer, diabetes and neurological and autoimmune diseases.16 To date, several studies have focused on the therapeutic role of ω-3 PUFAs, which are considered anti-inflammatory molecules. The resolution of inflammation is an active process primarily driven by a new family of mediators, termed resolvins, derived from the ω-3 PUFAs eicosapentaenoic acid (EPA, C20:5 ω-3) and docosahexaenoic acid (DHA, C22:6 ω-3).17 Among the major mediators of the inflammatory response is the generation of pro-inflammatory eicosanoids generated from the omega-6 (ω-6) PUFA arachidonic acid (AA, C20:4 ω-6). These mediators include pro-inflammatory prostaglandins (eg, PGE2) and leukotrienes (eg, LTB4), which can act as mediators for leucocyte chemotaxis and inflammatory cytokine production. The balance between the pro-inflammatory and anti-inflammatory molecules plays a key role in disease progression and the resolution of an inflammatory response.

Key messages What is already known about the subject? Age-related macular degeneration (AMD) is the leading cause of visual impairment and blindness in the elderly. Currently, there are no guidelines for the first-line treatment of dry AMD; however, several anti-oxidants, vitamins and zinc may reduce its progression. Several studies have demonstrated that omega-3 (ω-3) fatty acids may have a protective role in inflammatory-associated, ischaemia-associated, light-associated, oxygen-associated and age-associated pathology of the vascular and neural retina. Recently, Epitropoulos et al assessed the effect of oral ω-3 fatty acids (1680 mg eicosapentaenoic acid (EPA) and 560 mg docosahexaenoic acid (DHA), for 12 weeks) in a multi-centre, placebo-controlled, double-masked study in patients with dry eyes. The authors reported a significant improvement in several parameters, including tear osmolarity and tear break-up time. In addition, Georgiou and Prokopiou reported the results from an observational study, where patients with dry AMD were supplemented with EPA and DHA for up to 6 months (5–7.5 g/day EPA and DHA, arachidonic acid (AA)/EPA<2) and demonstrated significant improvement in vision (≥15 letters gain). What are the new findings?

Several studies have demonstrated that omega-3 (ω-3) fatty acids may have a protective role in inflammatory-associated, ischaemia-associated, light-associated, oxygen-associated and age-associated pathology of the vascular and neural retina. Recently, Epitropoulos et al assessed the effect of oral ω-3 fatty acids (1680 mg eicosapentaenoic acid (EPA) and 560 mg docosahexaenoic acid (DHA), for 12 weeks) in a multi-centre, placebo-controlled, double-masked study in patients with dry eyes. The authors reported a significant improvement in several parameters, including tear osmolarity and tear break-up time. In addition, Georgiou and Prokopiou reported the results from an observational study, where patients with dry AMD were supplemented with EPA and DHA for up to 6 months (5–7.5 g/day EPA and DHA, arachidonic acid (AA)/EPA<2) and demonstrated significant improvement in vision (≥15 letters gain). What are the new findings? In this study, we investigated the effects of ω-3 fatty acids alone or in combination with omega-6 (ω-6) in the CCL2−/− model of AMD while monitoring the blood levels of EPA and AA (AA/EPA=1–1.5). Supplementation with ω-3+ω-6 or ω-3 alone (when AA/EPA=1–1.5) suggests a protective mechanism in the CCL2−/− animal model of dry AMD, with a more beneficial effect when ω-3 are used alone. Retinal photoreceptors’ layer demonstrated a highly significant increase in the ω-3-treated group compared with the untreated. Inflammatory mediators showed decreased levels at the mRNA and protein level in the treatment groups. Our findings indicated that inflammation is not the only determining factor; perhaps a regenerative process might be involved following administration of ω-3 fatty acids.

ω-3-treated group compared with the untreated. Inflammatory mediators showed decreased levels at the mRNA and protein level in the treatment groups. Our findings indicated that inflammation is not the only determining factor; perhaps a regenerative process might be involved following administration of ω-3 fatty acids. How might these results change the focus of research or clinical practice? These findings are of great importance since supplementation with ω-3 fatty acids at the right dosage (blood AA/EPA=1–1.5) could be beneficial in patients with AMD and perhaps other ocular pathologies which involve photoreceptor damage. The benefits will have a positive impact on the quality of life of patients suffering from such conditions.

eat importance since supplementation with ω-3 fatty acids at the right dosage (blood AA/EPA=1–1.5) could be beneficial in patients with AMD and perhaps other ocular pathologies which involve photoreceptor damage. The benefits will have a positive impact on the quality of life of patients suffering from such conditions. Introduction Age-related macular degeneration (AMD) is the leading cause of visual impairment and blindness in the elderly; an estimated 200 million people suffer from AMD worldwide. By the year 2040, the number of these individuals is estimated to increase by 50%.1 2 Currently, there is no definitive pathogenesis model for AMD, but several different aetiologies have been proposed. Ageing is one of the most common contributing factors in AMD due to the accumulation of oxidised lipoproteins and free radicals in the retina and choroid. This accumulation in turn results in oxidative stress and a decrease in the number of retinal pigment epithelium (RPE) cells and photoreceptors.3 4 Genetic predisposition, as with multiple pathologies, plays a role in the development of AMD. Although there are several genetic associations, the most studied are certain polymorphic loci, in particular, those related to inflammatory genes such as complement factor H and certain complement components (eg, C3 and C2).5 6 Environmental factors, including smoking, sunlight exposure, high-fat diet, obesity and diabetes, are all associated with the development and progression of AMD.7

ertain polymorphic loci, in particular, those related to inflammatory genes such as complement factor H and certain complement components (eg, C3 and C2).5 6 Environmental factors, including smoking, sunlight exposure, high-fat diet, obesity and diabetes, are all associated with the development and progression of AMD.7 Furthermore, a tissue-adaptive response, recently described as para-inflammation, in which the innate immune system mounts a low-grade inflammatory response to restore tissue homeostasis, has been implicated in the pathogenesis of AMD.8 9 Inflammatory-related proteins, including C-reactive protein,10–13 interleukin-6 (IL-6)12 14 and tumour necrosis factor-α (TNF-α),12 15 have been shown to be associated with AMD; however, the results from different groups are inconsistent. The fact that systemic inflammatory markers are not strongly related with AMD suggests that local low-grade inflammation is more likely to be involved in its pathogenesis.

o-inflammatory prostaglandins (eg, PGE2) and leukotrienes (eg, LTB4), which can act as mediators for leucocyte chemotaxis and inflammatory cytokine production. The balance between the pro-inflammatory and anti-inflammatory molecules plays a key role in disease progression and the resolution of an inflammatory response. Several studies have demonstrated that ω-3 PUFAs may have a protective role in inflammatory-associated, ischaemia-associated, light-associated, oxygen-associated and age-associated pathology of the vascular and neural retinas.18 Administration of PUFAs was previously found to have a promising effect in several animal models of macular degeneration.19–23 Currently, there are no guidelines for the first-line treatment of dry AMD; however, several anti-oxidants, vitamins and zinc may reduce its progression according to the Age-related Eye Disease Study (AREDS).24 Following the AREDS, an additional study was performed, the AREDS2, which was a multi-centre 5-year randomised trial designed to examine the effects of oral supplementation of macular xanthophylls (10 mg lutein and 2 mg zeaxanthin) and/or ω-3 PUFAs (EPA 650 mg and DHA 350 mg) on the progression to advanced AMD. Overall, there was no additional benefit from adding the ω-3 PUFAs or a mixture of lutein and zeaxanthin to the formulation. Although the addition of ω-3 to the AREDS formulation was not shown to be beneficial, it is believed that higher doses of EPA and DHA may have a desirable effect.25

e progression to advanced AMD. Overall, there was no additional benefit from adding the ω-3 PUFAs or a mixture of lutein and zeaxanthin to the formulation. Although the addition of ω-3 to the AREDS formulation was not shown to be beneficial, it is believed that higher doses of EPA and DHA may have a desirable effect.25 Therefore, further studies were conducted or are still ongoing to investigate the possible mechanisms of action of PUFAs and to examine any positive effects on disease progression. Recently, Epitropoulos et al assessed the effect of oral ω-3 PUFAs (1680 mg EPA and 560 mg DHA, for 12 weeks) in a multi-centre, placebo-controlled, double-masked study in patients with dry eyes. The authors reported a significant improvement in several parameters, including tear osmolarity and tear break-up time.26 In addition, Georgiou and Prokopiou reported the results from an observational study, where patients with dry AMD were supplemented with EPA and DHA for up to 6 months (5–7.5 g/day EPA and DHA, AA/EPA <2) and demonstrated significant improvement in vision (≥15 letters gain).27

tear osmolarity and tear break-up time.26 In addition, Georgiou and Prokopiou reported the results from an observational study, where patients with dry AMD were supplemented with EPA and DHA for up to 6 months (5–7.5 g/day EPA and DHA, AA/EPA <2) and demonstrated significant improvement in vision (≥15 letters gain).27 The importance of particular fatty acids, such as the blood levels of EPA and AA, has recently been emphasised. In particular, the Japan EPA Lipid Intervention Study established the clinical efficacy of EPA for cardiovascular disease (CVD), in which higher levels of EPA but not DHA were found to be associated with a lower incidence of major coronary events. The risk of coronary events was significantly reduced when the ratio of EPA to AA (EPA/AA) was >0.75.28 In addition, the ratio of PGI3 and PGI2 (which both reduce cardiac ischaemic injury and arteriosclerosis and promote angiogenesis) to thromboxane A2 was determined to have a linear relationship with the EPA/AA ratio. Thus, the effects of EPA in reducing the risk of CVD could be mediated by the biological action of PGI3 in addition to the hypo-triglyceridemic action of EPA.29 A number of studies have confirmed the association of EPA/AA both in coronary diseases and in diabetes and hyperlipidaemia.30 31

ionship with the EPA/AA ratio. Thus, the effects of EPA in reducing the risk of CVD could be mediated by the biological action of PGI3 in addition to the hypo-triglyceridemic action of EPA.29 A number of studies have confirmed the association of EPA/AA both in coronary diseases and in diabetes and hyperlipidaemia.30 31 Lower levels of particular PUFAs in either circulating blood or the retina are associated with some retinopathies. In another study, eyes from AMD donors exhibited significantly decreased levels of very long-chain PUFAs and high ω-6/ω-3 ratios.32 Therefore, examining systemic biomarkers (including the levels of AA and EPA) when undertaking therapeutic trials could be a good indication of disease progression and treatment success. However, further studies are required to better establish the relationship between the level of certain fatty acids and the progression of ocular pathologies.

emic biomarkers (including the levels of AA and EPA) when undertaking therapeutic trials could be a good indication of disease progression and treatment success. However, further studies are required to better establish the relationship between the level of certain fatty acids and the progression of ocular pathologies. In this study, we investigated the effects of ω-3 PUFAs in the CCL2−/− model of AMD while monitoring the blood levels of EPA and AA (AA/EPA=1–1.5). To date, no ideal animal model has been established that fully recapitulates the human features of AMD; however, the CCL2−/− model shares certain common characteristics. There have been conflicting reports between different studies, some suggesting that most of the previously described hallmark features of AMD in the CCL2−/− model can be explained by normal ageing.33 Others indicated that this model develops drusen and other features of AMD, such as the accumulation of lipofuscin in RPE cells, progressive outer retinal degeneration and geographic/RPE atrophy.34 Materials and methods Animals Age-matched (9 months old) CCL2−/− (B6.129S4-Ccl2tm1Rol/J, Jacksons Laboratory, USA) and C57BL/6 mice (3 months old and 9 months old, considered wild type, CY/EXP/101 breeding colony) of both sexes were used in this study and kept under standard housing conditions with a 12-hour dark/light cycle and with food and water provided ad libitum. All experiments were performed in accordance with the animal care protocols approved by the Cyprus Government’s Chief Veterinary Officer according to EU guidelines.

both sexes were used in this study and kept under standard housing conditions with a 12-hour dark/light cycle and with food and water provided ad libitum. All experiments were performed in accordance with the animal care protocols approved by the Cyprus Government’s Chief Veterinary Officer according to EU guidelines. Study design The animals were randomly allocated to four different groups (n=22–24 in each group): (A) wild type untreated C57BL/6, (B) age-matched untreated CCL2−/−, (C) age-matched CCL2−/− treated with ω-3+ω-6 (100 mg EPA+50 mg DHA +30 mg γ-linolenic acid ((GLA), KD Pharma, Germany and Ophthalmos Omega 3, Cyprus) and (D) age-matched CCL2−/− treated with ω-3 only (130 mg EPA, KD Pharma, Germany and Ophthalmos Omega 3, Cyprus). The animals in the treatment groups were administered fish oil daily by gavage for 3 months. The oral dosing total volumes did not exceed 10 mL/kg. The general health and appearance of the animals were assessed daily, and body weight measurements were recorded weekly. The mice were euthanised at the end of the study by cervical dislocation.

atment groups were administered fish oil daily by gavage for 3 months. The oral dosing total volumes did not exceed 10 mL/kg. The general health and appearance of the animals were assessed daily, and body weight measurements were recorded weekly. The mice were euthanised at the end of the study by cervical dislocation. Fatty acid analysis To determine the AA/EPA ratio, following long-term treatment, blood samples (n=20–24) were collected on D0, D30 and D90, prior to treatment, processed for fatty acids separation and analysed using gas chromatography (GC). In addition, after completion of the study (D90), retinas (n=4) were collected from each group for the same purpose. Blood samples were collected on a Whatman filter paper and stored at −20°C. In each sample, NaOH/MeOH was added and heated for 1 min at 85°C, and then 14% BF3/MeOH was added and heated for 7 min at 85°C. Once the samples reached room temperature, 0.3 g of NaCl was added to ensure complete migration of the total fatty acid methyl ester fraction to the organic phase, followed by 1 mL of n-hexane. When the samples were separated into the organic and aqueous phases, the organic phase was collected and this was repeated three times. To remove any impurities, 0.9 g Na2SO4 was added to the samples and then centrifuged at 3300 r.p.m. for 5 min. Supernatants were collected and dried using an analytical evaporator at ~45°C under a nitrogen stream. Once dried, the samples were redissolved in n-hexane and analysed using a GC flame ionisation detector (FID).

remove any impurities, 0.9 g Na2SO4 was added to the samples and then centrifuged at 3300 r.p.m. for 5 min. Supernatants were collected and dried using an analytical evaporator at ~45°C under a nitrogen stream. Once dried, the samples were redissolved in n-hexane and analysed using a GC flame ionisation detector (FID). Gas chromatography flame ionisation detector An Agilent GC-6890 system was equipped with an FID. An H2 flow rate of 35 mL/min and an air flow rate of 350 mL/min were used. The flow rate of carrier gas (He) was set at 2.5 mL/min. The temperatures of the injection port and detector were set at 280°C and 300°C, respectively. The oven temperature was programmed to initiate at 160°C for 3 min, and then the temperature was raised to 200°C at a rate of 20°C/min, held there for 4 min and finally increased to 250°C at a rate of 5°C/min and held there for 23 min. The injection volume was 1 µL in the split-less injection mode. A capillary column (DP-23 fused-silica capillary, 30 m×0.25 mm I.D. x 0.25 µm film thickness; Supelco, Bellefonte, Pennsylvania, USA) was employed.

0°C/min, held there for 4 min and finally increased to 250°C at a rate of 5°C/min and held there for 23 min. The injection volume was 1 µL in the split-less injection mode. A capillary column (DP-23 fused-silica capillary, 30 m×0.25 mm I.D. x 0.25 µm film thickness; Supelco, Bellefonte, Pennsylvania, USA) was employed. Immunofluorescence staining To assess the effect of treatment on the retinal photoreceptors (ie, the outer nuclear layer (ONL)), eyes from each group (n=5) were used. An additional group of younger (3-month-old) mice served as an additional control for comparison purposes. Eyes were enucleated at the end of the study (D90) and fixed with 2% paraformaldehyde for 2 hours, then transferred to a 30% sucrose solution in PBS for 24 hours at 4°C and embedded in optimum cutting temperature media (Sakura Finetek, Torrance, California, USA). Whole eyes were cross-sectioned and sections were obtained throughout the eye, including inferior pole, the optic nerve and superior pole. A total of 20 sections were taken, 28 µm apart, throughout the whole eye, including peripheral and central retinal areas. Briefly, 14-µm-thick cryosections were blocked with 10% donkey serum in 1% Triton X-100/PBS for 1 hour. Sections were then incubated with rabbit polyclonal anti-cone arrestin (AB15282, Chemicon, Temecula, California, USA; dilution 1:10 000) and mouse monoclonal anti-rhodopsin (clone 4D2, Chemicon; dilution 1:500) overnight at 4°C. After washing, sections were incubated for 2 hours with Alexa Fluor 594 donkey anti-rabbit IgG (1:200, Invitrogen) or Alexa Fluor 488 donkey anti-mouse IgG (1:200, Invitrogen). TO-PRO-3 iodide (T3605, Thermo Fisher Scientific, Massachusetts, USA) was used for nuclear staining. Samples were cover-slipped with Dako fluorescence mounting medium (Dako, Denmark) and examined by Leica TCS SP5 confocal microscopy (Leica, Germany). Measurements of the ONL were made at different fields around the entire retinal section (centre, middle and periphery). ImageJ software (NIH, http://imagej.nih.gov/ij) was used to calculate the ONL thickness, and the mean ONL thickness of the entire retina was compared among the different groups. The density of the photoreceptors was estimated as the number of photoreceptors nuclei over the retinal area, using ImageJ software.

Health state utility values were estimated based on a weighted average of two papers (see table 1): Brown et al who provided time trade-off values for a range of visual acuities associated with DR8 and Fong et al who reported the number of people with a range of visual acuity for different stages of DR.9 Data from the better seeing eye were used, because quality of life data depends mainly on the better seeing eye. For patients with macular oedema, utility values from a study by Smith et al 10 were used. This methodology was similar to a paper by Ting et al 11 who developed a Markov model of a novel DR prognostic device for DR progression.5 Table 1 Health state utility values for the base-case analysis Health state Usual care arm Intervention arm Moderate NPDR/severe NPDR 0.7915 0.7915 Severe NPDR and CSDMO 0.7365 0.7365 Early PDR/high-risk PDR/severe PDR 0.7047 0.7047 Early PDR and CSDMO/high-risk PDR and CSDMO/severe PDR and CSDMO 0.6930 0.6930 Severe NPDR PT – 0.7915 Severe NPDR and CSDMO PT – 0.7365 Early PDR PT – 0.7047 Early PDR and CSDMO PT – 0.6930 High-risk PDR PT/severe PDR PT 0.7047 0.7047 High-risk PDR and CSDMO PT/severe PDR and CSDMO PT 0.6930 0.6930 Severe visual loss/blindness 0.6218 0.6218 CSDMO, clinically significant diabetic macular oedema; NPDR, non-proliferative diabetic retinopathy; PDR, proliferative diabetic retinopathy; PT, post-treatment.

riphery). ImageJ software (NIH, http://imagej.nih.gov/ij) was used to calculate the ONL thickness, and the mean ONL thickness of the entire retina was compared among the different groups. The density of the photoreceptors was estimated as the number of photoreceptors nuclei over the retinal area, using ImageJ software. Quantitative reverse transcription PCR On D90, retinas (n=4–5) from each group were excised, and tissue RNA was extracted using a NucleoSpin RNA/Protein kit (Macherey-Nagel, USA). The quantity and quality of RNA were determined spectrophotometrically (Nanodrop Technologies, Montchanin, DE). The same amount of total RNA was used for reverse transcription using a ProtoScript(R) II First-Strand cDNA Synthesis Kit according to the standard protocol of the supplier (New England Biolabs, UK). The quantitative PCR amplifications were performed on a ViiA 7 Real-Time PCR System (Applied Biosystems, USA) using Fast SYBR Green Master Mix (Thermo Fisher Scientific, USA) in a reaction volume of 20 µL. Relative quantification analysis was carried out on ViiA seven software. The primer sets for the genes whose differential expression was analysed by Quantitative reverse transcription PCR (qRT-PCR) are shown in table 1. Expression levels of each target gene were normalised to those of β-actin mRNA. Data are reported as the mean relative expression levels ± SE of the mean (SEM). Table 1 Primer sequences of genes confirmed by quantitative reverse transcription PCR

Quantitative reverse transcription PCR On D90, retinas (n=4–5) from each group were excised, and tissue RNA was extracted using a NucleoSpin RNA/Protein kit (Macherey-Nagel, USA). The quantity and quality of RNA were determined spectrophotometrically (Nanodrop Technologies, Montchanin, DE). The same amount of total RNA was used for reverse transcription using a ProtoScript(R) II First-Strand cDNA Synthesis Kit according to the standard protocol of the supplier (New England Biolabs, UK). The quantitative PCR amplifications were performed on a ViiA 7 Real-Time PCR System (Applied Biosystems, USA) using Fast SYBR Green Master Mix (Thermo Fisher Scientific, USA) in a reaction volume of 20 µL. Relative quantification analysis was carried out on ViiA seven software. The primer sets for the genes whose differential expression was analysed by Quantitative reverse transcription PCR (qRT-PCR) are shown in table 1. Expression levels of each target gene were normalised to those of β-actin mRNA. Data are reported as the mean relative expression levels ± SE of the mean (SEM). Table 1 Primer sequences of genes confirmed by quantitative reverse transcription PCR Gene Forward Reverse TLR3 TGCAGAAGATTCAAGGTACATCA CAAACAGAGTGCATGGTTTAGTT NF-κB ACGAGGCAGCACATAGATGAΑ AGGATGTCTCCACACCACTGTC IL-18 GCCGACTTCACTGTACAACC TCTGGTCTGGGGTTCACTGG β-Actin GCCTTCCTTCTTGGGTATGG CGGATGTCAACGTCACACTT Western blot analysis Retinal protein extracts were obtained from dissected retinas in each group (n=4–5) using a NucleoSpin RNA/Protein kit (Macherey-Nagel, USA). Twenty micrograms of total protein was detected using Qubit protein assay kit (Thermo Fisher Scientific, Massachusetts, USA) and β-actin was used as a loading control. The antibodies used against inflammatory mediators were as follows: 1) rabbit anti-IL-18 polyclonal antibody (5180 R-100, Biovision; dilution 1:1000), 2) goat anti-IL-17 polyclonal antibody (E-19, sc-6077, Santa Cruz Biotechnology, USA, dilution 1:200) and 3) goat anti-TNF-α polyclonal antibody (L-19, sc-1351, Santa Cruz Biotechnology, USA, dilution 1:200) followed by peroxidase-labelled secondary antibodies — either goat anti-rabbit IgG-HRP antibody (7074, Cell Signalling Technology; dilution 1:2,000) or donkey anti-goat IgG-HRP antibody (sc-2020, Santa Cruz Biotechnology, USA, dilution 1:2000). Proteins were detected using the Enhanced ChemiLuminescence Plus Blotting Detection system (Amersham Biosciences, Buckinghamshire, UK) and were visualised on a Chemidoc XRS+ imager (BioRad, USA). All immunoblots were re-probed with anti-β-actin antibody (D6A8, Cell Signalling Technology; dilution 1:1000) to confirm that equal amounts of protein were loaded on the membrane. The band density was defined using ImageJ Software. The results were expressed as the percentage density ratio of the protein of interest over the loading control (β-actin). All results were normalised against the untreated samples.

dilution 1:1000) to confirm that equal amounts of protein were loaded on the membrane. The band density was defined using ImageJ Software. The results were expressed as the percentage density ratio of the protein of interest over the loading control (β-actin). All results were normalised against the untreated samples. Statistical analysis All statistical analyses were performed using GraphPad Prism (V.5) statistical software. Data are expressed as the mean±SEM; the two-tailed Student’s t-test or one-way ANOVA with post hoc test: Tukey’s test was used to evaluate the differences among treated groups. Statistical significance was set at p<0.05. Results General observations Animals’ general appearance was monitored daily and nothing abnormal was noted. Body weight measurements were recorded weekly, where no significant difference was found over time in any of the groups.

Statistical analysis All statistical analyses were performed using GraphPad Prism (V.5) statistical software. Data are expressed as the mean±SEM; the two-tailed Student’s t-test or one-way ANOVA with post hoc test: Tukey’s test was used to evaluate the differences among treated groups. Statistical significance was set at p<0.05. Results General observations Animals’ general appearance was monitored daily and nothing abnormal was noted. Body weight measurements were recorded weekly, where no significant difference was found over time in any of the groups. Blood fatty acid analysis For the duration of the study, blood fatty acids were monitored, as shown in figure 1 and table 2, in terms of percentage of total fatty acids. Samples were taken on D0, D30 and D90 prior to treatment. The levels of GLA could not be accurately detected; for that reason, dihomo-GLA (DGLA, a metabolite of GLA) was used for the purposes of this study. The percentage of the ω-6 fatty acids, DGLA and AA, as well as the percentage of the ω-3 fatty acids EPA and DHA in blood, was determined using a gas chromatographic technique (figure 1). The AA/EPA ratio was also calculated in each group, as shown in figure 1D. On D0, there was significantly lower levels of DGLA (p<0.01) and EPA (p<0.05) in the C57BL/6 mice compared with those in the CCL2−/− untreated mice. The general trend observed was that, on D30, the level of EPA significantly increased (p<0.0001) in both treatment groups, whereas the level of AA was reduced (p<0.0001) compared with CCL2−/− untreated and wild type groups. In particular, on D30, the blood AA/EPA ratio was decreased (p<0.0001) in both treatment groups (1.10±0.02 for both) compared with the CCL2−/− untreated (4.71±0.15) and wild type (4.74±0.15) groups. Higher DHA levels (p<0.0001) were detected on D30 in the ω-3+ω-6-treated group (4.43±0.14) and in the ω-3-treated group (4.42±0.15) than in the CCL2−/− untreated (3.18±0.09) and wild type (3.15±0.09) groups. Interestingly, in the ω-3+ω-6-treated group, the DGLA levels showed an increase (p<0.001) on D90 (1.21±0.05) compared with D0 (1.05±0.04) but did not show a significant difference on D30. By contrast, on D30 of treatment, DGLA levels were lower (p<0.0001) in the ω-3-treated group (0.89±0.01) than on D0 (0.96±0.01). At the end of the study, on D90, there were no fluctuations in the levels of AA, EPA and DHA, and the AA/EPA ratio remained approximately constant in all groups compared with D30.

nce on D30. By contrast, on D30 of treatment, DGLA levels were lower (p<0.0001) in the ω-3-treated group (0.89±0.01) than on D0 (0.96±0.01). At the end of the study, on D90, there were no fluctuations in the levels of AA, EPA and DHA, and the AA/EPA ratio remained approximately constant in all groups compared with D30. Figure 1 Fatty acid analysis from blood samples collected on (A) D0, (B) D30 and (C) D90 from C57BL/6 untreated group, CCL2−/− untreated group, CCL2−/− ω-3+ω-6-treated group and CCL2−/− ω-3-treated group. (D) The AA/EPA ratio is shown in each group for D0, D30 and D90. The proportion of DGLA, AA, EPA and DHA is expressed as the mean±SEM (n=22–24). Table 2 Fatty acid analysis from blood samples collected on D0, D30 and D90 from groups A, C57BL/6 untreated; B, CCL2−/− untreated; C, CCL2−/− treated with ω-3+ω-6 and D, CCL2−/− treated with ω-3 DGLA AA EPA DHA AA/EPA D0 A 1.05 ±0.04 7.04 ±0.10 1.59±0.04 3.24±0.09 4.48±0.13 B 1.23±0.04 7.08±0.09 1.62±0.05 3.29±0.09 4.38±0.14 C 1.05±0.04 6.99±0.09 1.48±0.04 3.16±0.08 4.80±0.14 D 0.96±0.01 7.26±0.17 1.39±0.09 3.77±0.06 5.52±0.23 D30 A 1.07±0.04 6.97±0.10 1.50±0.05 3.15±0.09 4.74±0.15 B 1.07±0.04 7.00±0.10 1.51±0.04 3.18±0.09 4.71±0.14 C 1.13±0.05 4.99±0.16 4.57±0.16 4.43±0.14 1.10±0.20 D 0.89±0.01 5.01±0.17 4.56±0.17 4.42±0.15 1.10±0.02 D90 A 1.02±0.04 6.91±0.10 1.44±0.05 3.09±0.09 4.90±0.16 B 1.17±0.04 7.12±0.10 1.60±0.05 3.27±0.09 4.71±0.14 C 1.21±0.05 5.07±0.16 4.65±0.16 4.51±0.14 1.10±0.02 D 0.97±0.01 5.02±0.17 4.64±0.17 4.61±0.18 1.10±0.04 Gas chromatography was used to obtain the results.

±0.20 D 0.89±0.01 5.01±0.17 4.56±0.17 4.42±0.15 1.10±0.02 D90 A 1.02±0.04 6.91±0.10 1.44±0.05 3.09±0.09 4.90±0.16 B 1.17±0.04 7.12±0.10 1.60±0.05 3.27±0.09 4.71±0.14 C 1.21±0.05 5.07±0.16 4.65±0.16 4.51±0.14 1.10±0.02 D 0.97±0.01 5.02±0.17 4.64±0.17 4.61±0.18 1.10±0.04 Gas chromatography was used to obtain the results. Data are expressed as the mean±SEM (n=22–24). Fatty acid analysis from blood samples collected on D0, D30 and D90 from groups A, C57BL/6 untreated animals; B, CCL2−/− untreated; C, CCL2−/− treated with ω-3+ω-6; D, CCL2−/− treated with ω-3. GC was used to obtain the results. Data are expressed as the mean±SEM (n=22–24).

±0.20 D 0.89±0.01 5.01±0.17 4.56±0.17 4.42±0.15 1.10±0.02 D90 A 1.02±0.04 6.91±0.10 1.44±0.05 3.09±0.09 4.90±0.16 B 1.17±0.04 7.12±0.10 1.60±0.05 3.27±0.09 4.71±0.14 C 1.21±0.05 5.07±0.16 4.65±0.16 4.51±0.14 1.10±0.02 D 0.97±0.01 5.02±0.17 4.64±0.17 4.61±0.18 1.10±0.04 Gas chromatography was used to obtain the results. Data are expressed as the mean±SEM (n=22–24). Fatty acid analysis from blood samples collected on D0, D30 and D90 from groups A, C57BL/6 untreated animals; B, CCL2−/− untreated; C, CCL2−/− treated with ω-3+ω-6; D, CCL2−/− treated with ω-3. GC was used to obtain the results. Data are expressed as the mean±SEM (n=22–24). Retinal fatty acid analysis In addition, retinas from each group were collected on D90, and the percentage of retinal fatty acids, including DGLA, AA, EPA and DHA, was also determined using GC, as shown in figure 2 and table 3. The level of retinal DGLA in the ω-3+ω-6-treated group (0.89±0.01) was significantly higher (p<0.05) than that in the CCL2−/− untreated group (0.64±0.10). The DGLA level in the ω-3-treated group (0.60±0.07) was lower (p<0.01) than that in the ω-3+ω-6-treated group. The AA content was significantly reduced in both ω-3 (4.29±0.10, p<0.05) and ω-3+ω-6 (4.39±0.1, p<0.01) treated groups (compared with the CCL2−/− untreated group (4.93±0.30)). Interestingly, there was an increase in EPA retinal levels in the ω-3+ω-6-treated group (1.42±0.03, p<0.001) and in the ω-3-treated group (1.55±0.03, p<0.001), compared with the CCL2−/− untreated (0.87±0.06) or wild type (0.99±0.05) groups. As expected, retinal DHA levels were the highest of all fatty acids, at approximately 25-fold. An increase (p<0.05) in DHA was observed in the ω-3+ω-6-treated group (27.73±0.36) compared with the CCL2−/− untreated (25.30±0.77) and wild type (25.09±0.87) groups, whereas in the ω-3-treated group (23.70±0.51), there was a substantial reduction (p<0.001) compared with the ω-3+ω-6-treated group.

pproximately 25-fold. An increase (p<0.05) in DHA was observed in the ω-3+ω-6-treated group (27.73±0.36) compared with the CCL2−/− untreated (25.30±0.77) and wild type (25.09±0.87) groups, whereas in the ω-3-treated group (23.70±0.51), there was a substantial reduction (p<0.001) compared with the ω-3+ω-6-treated group. Figure 2 Fatty acid analysis from retina samples collected on D90 from groups A, C57BL/6 untreated; B, CCL2−/− untreated; C, CCL2−/− treated with ω-3+ω-6; and D, CCL2−/− treated with ω-3. The proportion of DGLA, AA, EPA and DHA is expressed as the mean±SEM (n=4). Table 3 Fatty acid analysis from retina samples collected on D90 from groups A, C57BL/6 untreated animals; B, CCL2−/− untreated; C, CCL2−/− treated with ω-3+ω-6; D, CCL2−/− treated with ω-3 DGLA AA EPA DHA D90 A 0.65 ± 0.06 5.11 ± 0.35 0.99 ± 0.05 25.09 ± 0.87 B 0.64 ± 0.10 4.93 ± 0.30 0.87 ± 0.06 25.30 ± 0.77 C 0.89 ± 0.01 4.39 ± 0.11 1.42 ± 0.03 27.73 ± 0.36 D 0.60 ± 0.07 4.29 ± 0.10 1.55 ± 0.03 23.70 ± 0.51 Gas chromatography was used to obtain the results. Data are expressed as the mean±SEM (n=4).

Table 3 Fatty acid analysis from retina samples collected on D90 from groups A, C57BL/6 untreated animals; B, CCL2−/− untreated; C, CCL2−/− treated with ω-3+ω-6; D, CCL2−/− treated with ω-3 DGLA AA EPA DHA D90 A 0.65 ± 0.06 5.11 ± 0.35 0.99 ± 0.05 25.09 ± 0.87 B 0.64 ± 0.10 4.93 ± 0.30 0.87 ± 0.06 25.30 ± 0.77 C 0.89 ± 0.01 4.39 ± 0.11 1.42 ± 0.03 27.73 ± 0.36 D 0.60 ± 0.07 4.29 ± 0.10 1.55 ± 0.03 23.70 ± 0.51 Gas chromatography was used to obtain the results. Data are expressed as the mean±SEM (n=4). Immunofluorescence staining The ONL thickness was examined in control and treated groups, as presented in figure 3. No difference in the ONL thickness was observed between the aged wild type (12 months old) and the aged-matched CCL2−/− animals (71.1±4.2 µm and 65.6±3.0 µm, respectively). Previous studies have reported that the CCL2−/− model demonstrates confluent areas of visible atrophy after 16 months of age34; therefore, no changes were expected to be seen at this point. By contrast, the younger C57BL/6 mice (3 months old) exhibited significantly (p<0.05) greater ONL thickness (98.0±3.9 µm) compared with both wild type and CCL2−/− animals. Both treatment groups, ω-3+ω-6 and ω-3 only, exhibited significantly larger ONL thicknesses than the untreated CCL2−/− animals, with a more pronounced effect in the ω-3-treated group (90.0±7.8 µm (p<0.05) and 125.6±9.8 µm (p<0.001), respectively, compared with 65.6±3.0 µm). Interestingly, the ONL thickness was even greater in the ω-3-treated group than in the young animals (125.6±9.8 µm compared with 98.0±3.9 µm, respectively, p<0.05). Quantification of rhodopsin and cone arrestin could not accurately be performed from the immunostaining; for that reason, the ONL thickness was the only measurement that was carried out in the histological analysis.

-treated group than in the young animals (125.6±9.8 µm compared with 98.0±3.9 µm, respectively, p<0.05). Quantification of rhodopsin and cone arrestin could not accurately be performed from the immunostaining; for that reason, the ONL thickness was the only measurement that was carried out in the histological analysis. Figure 3 (1) Representative ocular photomicrographs of groups A, C57BL/6 untreated; B, CCL2−/− untreated; C, CCL2−/− treated with ω-3+ω-6; and D, CCL2−/− treated with ω-3. Cryosections were stained for cone arrestin (red), rhodopsin (green) and TO-PRO-3 iodide (blue) for nuclear staining. Samples were examined by a Leica TCS SP5 confocal microscope (original magnification 100×, scale bar=50 µm). Arrow indicates retinal ONL. (2) Measurements of the ONL were performed at different fields around the entire retinal section (centre, middle and periphery). ImageJ software was used to calculate the ONL thickness, and the mean ONL thickness of the entire retina was compared among the different groups. Data represent the mean±SEM (n=5). In order to exclude the possibility that the increase in ONL was due to nuclei swelling or spacing, the density of photoreceptors was calculated by dividing the number of nuclei over the retinal area (figure 4). There was no significant difference determined in any of the groups (0.19±0.03 in C57BL/6 untreated group, 0.27±0.09 in CCL2−/− untreated group, 0.24±0.05 in ω-3+ω-6-treated group and 0.26±0.07 in ω-3-treated group).

ity of photoreceptors was calculated by dividing the number of nuclei over the retinal area (figure 4). There was no significant difference determined in any of the groups (0.19±0.03 in C57BL/6 untreated group, 0.27±0.09 in CCL2−/− untreated group, 0.24±0.05 in ω-3+ω-6-treated group and 0.26±0.07 in ω-3-treated group). Figure 4 Photoreceptors’ density as estimated by ImageJ among the different groups (photoreceptors nuclei number/retinal area). Data represent the mean±SEM (n=3). qRT-PCR analysis The response of treatment with regard to the regulation of inflammatory mediators, in particular, the mRNA levels of TLR3, NF-κB and IL-18 in retinal tissues, was examined using qRT-PCR analysis, as shown in figure 5. The expression levels of TLR3 were lower, although not significantly, in both treatment groups than in the untreated group, with a more evident effect in the ω-3-treated group. A significant (p<0.05) reduction in the expression levels of NF-κB was observed in both treatment groups compared with the CCL2−/− untreated group. However, there was no apparent difference in the expression levels of IL-18 in either treatment groups and those in the CCL2−/− untreated group, the only notable effect being an increased expression in the wild type animals.

evels of NF-κB was observed in both treatment groups compared with the CCL2−/− untreated group. However, there was no apparent difference in the expression levels of IL-18 in either treatment groups and those in the CCL2−/− untreated group, the only notable effect being an increased expression in the wild type animals. Figure 5 Gene transcripts of (A) TLR3, (B) NF-κB and (C) IL-18 in the retinas of C57BL/6 untreated, CCL2−/− untreated, CCL2−/− treated with ω-3+ω-6 and CCL2−/− treated with ω-3. Lower levels in TLR3 and NF-κB mRNA were detected by quantitative reverse transcription PCR in the ocular tissue of both treatment groups. The graph is plotted as the mean±SEM (n=4–5). Western blot analysis To further investigate the effect of treatment on the inflammatory response, retinal tissues were evaluated to identify any difference in the protein levels of TNF-α, IL-18 and IL-17, as shown in figure 6. A minor reduction was observed in the TNF-α levels in both treatment groups compared with the untreated group; however, no effect was noted in the IL-17 levels. Interestingly, a significant (p<0.001) 4-fold reduction in the IL-18 levels was observed in the ω-3-treated group compared with the CCL2−/− untreated group but not in the ω-3+ω-6-treated group. In addition, the IL-18 protein level in the ω-3-treated group was even lower than that in the wild type group (p<0.01).

evels. Interestingly, a significant (p<0.001) 4-fold reduction in the IL-18 levels was observed in the ω-3-treated group compared with the CCL2−/− untreated group but not in the ω-3+ω-6-treated group. In addition, the IL-18 protein level in the ω-3-treated group was even lower than that in the wild type group (p<0.01). Figure 6 Western blot analysis of (A) TNF-α, (B) IL-17 and (C) IL-18 protein levels in the retinas of C57BL/6 untreated,  CCL2−/− untreated, CCL2−/− treated with ω-3+ω-6 and CCL2−/− treated with ω-3. A significant reduction in the protein levels of IL-18 was observed in the ω-3 treatment group only. The graph is plotted as the mean±SEM (n=4-5). Discussion AMD is one of the leading causes of blindness worldwide. Although it is a common condition, no treatment to date has provided evidence of disease regression. However, several studies have focused on the importance of para-inflammation in the pathogenesis of AMD8; therefore, aiming to stabilise the disease using anti-inflammatory approaches is considered a reasonable strategy. ω-3 PUFAs have been used in numerous studies, both in preclinical and clinical settings, demonstrating their effectiveness against heart disease, cancer, diabetes and neurological and autoimmune diseases.16 Moreover, emphasis has been given in the treatment of ocular pathologies using ω-3 PUFAs.

a reasonable strategy. ω-3 PUFAs have been used in numerous studies, both in preclinical and clinical settings, demonstrating their effectiveness against heart disease, cancer, diabetes and neurological and autoimmune diseases.16 Moreover, emphasis has been given in the treatment of ocular pathologies using ω-3 PUFAs. The aim of this study was to investigate the effect of either ω-3+ω-6 PUFAs or ω-3 only in a CCL2−/− animal model that shares some of the clinical features of human AMD,34 while monitoring the AA/EPA ratio (~1–1.5). Supplementation with PUFAs served more as a preventive measure than as a therapeutic regime because the CCL2−/− model does not demonstrate signs of photoreceptor loss before the age of 16 months.34 The study was terminated when the animals were 12 months of age; therefore, there was no significant difference in the retinal ONL of CCL2−/− and the wild type group. However, a significant increase in the ONL thickness was observed in both treatment groups, with a more pronounced effect in the ω-3-treated group. A recent study by Georgiou et al demonstrated the neuroprotective effects of ω-3 in a rat model of anterior ischaemic optic neuropathy, through different actions, that is, preventing apoptosis of retinal ganglion cells, decreasing inflammatory cell infiltration and regulating macrophage polarisation to decrease the cytokine-induced injury of the optic nerve.35

ted the neuroprotective effects of ω-3 in a rat model of anterior ischaemic optic neuropathy, through different actions, that is, preventing apoptosis of retinal ganglion cells, decreasing inflammatory cell infiltration and regulating macrophage polarisation to decrease the cytokine-induced injury of the optic nerve.35 Other studies, in particular, Tuo et al, used the double-knockout Ccl2−/−/Cx3cr1−/− mice to investigate the effect of a high ω-3 PUFA diet, which included 1.9 wt % of each EPA and DHA, 0.66 wt % of α-linolenic acid and 0.4 wt % of docosapentaenoic acid. The ω-6/ω-3 ratio was 2.9, and the ω-6 source was linoleic acid only. Specifically, animals that ingested a high ω-3 diet up to 8 months of age showed progression of retinal lesions and minor photoreceptor loss relative to the low ω-3 diet group. This effect was suggested to occur through a reduction in the AA metabolism, as demonstrated by the decreased pro-inflammatory derivatives (PGE2 and LTB4).22 High levels of dietary ω-3 PUFAs may result in the incorporation of EPA into cell membrane phospholipids at the expense of AA, leading to less substrate being available for eicosanoid synthesis.36

. Furthermore, the vertical SD value of this group further decreased from the second 20 s period (2.41) to the third 20 s period (1.65), although this did not reach the level of the statistical significance. To confirm the present finding, it was necessary to further investigate by increasing the number of cases of MG. Usefulness of eye tracking in evaluating fixation stability in patients with MG Current clinical evaluations of ocular symptoms in patients with MG with diplopia (initial diplopia and residual diplopia following treatment) are subjective. A published report on the diagnosis and evaluation of MG-associated diplopia describes methods for confirming acquired strabismus caused by MG. These include comparing ocular deviation in the Hess screen test before and after edrophonium infusion.7 Electromyography demonstrates the fatigue phenomena electrophysiologically and the edrophonium test shows the pharmacological condition of the fatigue phenomenon. However, these tests are diagnostic invasive methods (ie, electromyography is painful26; edrophonium has effect on cardiovascular system27), and therefore, ophthalmologists avoid repeated use of these tests if alternative, non-invasive methods are available. The ice pack test, which involves cooling the orbital cavity through the application of ice over closed eyelids for 2 min, is less invasive for diagnosing myasthenic ptosis.28 In contrast, the ice pack test for diagnosing myasthenic diplopia is more invasive because it requires a 5 min or longer cooling time because extraocular muscles are located in the deeper portion of the orbital cavity.29 30 Unfortunately, the ice pack test becomes increasingly uncomfortable for the patient when cooling times are longer than 2 min.28 Compared with these clinical tests, the present quantification method using the eye-tracking system can non-invasively evaluate gaze stability in patients with MG almost without discomforts. As an improvement in gaze stability indicates an improvement in the MG disease state,3–6 eye tracking is especially suitable for evaluating treatment efficacy in patients with MG.

through a reduction in the AA metabolism, as demonstrated by the decreased pro-inflammatory derivatives (PGE2 and LTB4).22 High levels of dietary ω-3 PUFAs may result in the incorporation of EPA into cell membrane phospholipids at the expense of AA, leading to less substrate being available for eicosanoid synthesis.36 Furthermore, Ramkumar et al reported the effect of the AREDS2 formulation in the same Ccl2−/−/Cx3cr1−/− model.20 This formulation included ω-3 fatty acids (54.9 mmol EPA/kg diet and 25.2 mmol DHA/kg diet), 17.6 mmol lutein/kg diet and 1.76 mmol zeaxanthin/kg diet. After 3 months of treatment, histological analysis demonstrated that the ONL thickness was greater in the high ω-3-treated group than that in the Ccl2−/−/Cx3cr1−/− control-diet group. Additional studies reported the effect of ω-3 supplementation on the ONL in a light-induced retinal degeneration model, whereby animals treated with fish oil and anti-oxidants preserved their retinal structure and function.37 Our findings indicated a superior effect with regard to photoreceptor preservation because the ω-3-treated group exhibited an even greater ONL thickness than that of younger mice (3 months old). There was no difference in the number of photoreceptors per area when a comparison was performed between different groups, indicating that the increased ONL thickness was due to an increased number of cells and not due to any swelling effect. However, additional studies need to investigate this phenomenon, using 5-bromo-2-deoxyuridine to test for the existence and nature of any proliferating cells.

n was performed between different groups, indicating that the increased ONL thickness was due to an increased number of cells and not due to any swelling effect. However, additional studies need to investigate this phenomenon, using 5-bromo-2-deoxyuridine to test for the existence and nature of any proliferating cells. It has been hypothesised that the positive effects of ω-3 PUFAs (eg, EPA and DHA) is due to the formation of their metabolites, which possess anti-inflammatory properties. The production of pro-inflammatory eicosanoids is generated from the ω-6 PUFAs (eg, AA and GLA), which can act as mediators for leucocyte chemotaxis and inflammatory cytokine production. The balance between the pro-inflammatory and anti-inflammatory molecules is thought to be of great significance in disease progression. Therefore, in this study, we investigated the levels of fatty acids in blood and retinal tissues; in particular, the levels of DGLA, AA, EPA and DHA were monitored during treatment. The findings showed reduced pro-inflammatory and increased anti-inflammatory fatty acids in both blood and retina in the treatment groups. DHA is abundantly expressed in the photoreceptors, and vital retinal functions depend on its existence38; this influence was confirmed by its high retinal levels compared with the rest of the fatty acids. By contrast, EPA levels, although they were higher in the treated groups, were approximately 25-fold lower than those of DHA. This observation is explained by EPA’s greater metabolism by β-oxidation in the brain, which implies a shorter half-life, and by its low level of recycling into brain phospholipids.39 The gene expression and protein levels of some pro-inflammatory mediators were examined to find a correlation between the levels of fatty acids and the stage of the inflammatory response. A decrease in the transcript levels of retinal TLR3 and NF-κB was observed in the treatment groups, with a more evident effect in the ω-3 group. TLR3 is associated with microglia cell activation, which is the brain's principal resident immune cell population, central to the inflammatory response and activated by distress signals released from neighbouring cells.40 On microglia activation, through TLR3, production of ILs41 and other cytokines is stimulated.

oup. TLR3 is associated with microglia cell activation, which is the brain's principal resident immune cell population, central to the inflammatory response and activated by distress signals released from neighbouring cells.40 On microglia activation, through TLR3, production of ILs41 and other cytokines is stimulated. Regulation of the transcription of inflammatory target genes (eg, TNF-α, IL-6 and IL-1) can take place via NF-κB activation.42 Therefore, reduction of TLR3 and NF-κB mRNA levels may indicate a reduced profile of microglia activation in the retina and subsequent decrease in the production of inflammatory cytokines. Microglia morphology was also examined in our study groups, but the differentiation between resting, activated and rod microglia was limited (data not shown). No changes were observed in the gene expression of IL-18, in contrast to the protein levels, in which a significant decrease was observed in the ω-3 group, suggesting that a post-translational modification might have occurred, although further studies are required in order to establish this. Doyle et al reported that drusen isolated from donor AMD eyes activates the NLRP3 inflammasome, causing secretion of IL-18 and IL-1β.43 In addition, IL-18 was found to attenuate choroidal neovascularisation in a wet AMD model,44 which demonstrated its importance in disease progression. TNF-α protein levels showed a minor decrease in both treatment groups (with a more profound effect in the ω-3 group); however, IL-17 did not exhibit any changes, possibly due to its lack of direct involvement in the CCL2−/− model.

arisation in a wet AMD model,44 which demonstrated its importance in disease progression. TNF-α protein levels showed a minor decrease in both treatment groups (with a more profound effect in the ω-3 group); however, IL-17 did not exhibit any changes, possibly due to its lack of direct involvement in the CCL2−/− model. Previous studies have demonstrated that, in the Ccl2−/−/Cx3cr1−/− model, lower ocular TNF-α and IL-6 transcript levels are observed in animals fed with high ω-3 diet, suggesting that reactive mediators of ω-3 PUFAs may regulate differential gene expression.22 Additional studies have confirmed this connection by demonstrating that the retinal expression of TNF-α, Cox-2, IL-1β, VEGF and iNos is much lower in a high ω-3 treated group than in a control.20 Connor et al evaluated the therapeutic effect of ω-3 PUFAs on hypoxia-induced pathological neovascularisation in a mouse model of oxygen-induced retinopathy. The results suggested that, by increasing the level of ω-3 PUFAs either by dietary or genetic means (using a Fat-1 transgenic model that converts ω-6 to ω-3 PUFAs), there was a reduced hypoxic stimulus for neovascularisation. This effect was mediated through the bioactive metabolites neuroprotectinD1, resolvinD1 and resolvinE1, through reduction in the TNF-α expression.45

ω-3 PUFAs either by dietary or genetic means (using a Fat-1 transgenic model that converts ω-6 to ω-3 PUFAs), there was a reduced hypoxic stimulus for neovascularisation. This effect was mediated through the bioactive metabolites neuroprotectinD1, resolvinD1 and resolvinE1, through reduction in the TNF-α expression.45 It is evident from numerous preclinical studies and several clinical and observational studies that there is a clear association of ω-3 PUFAs and ocular pathologies, in particular, with AMD.27 46 It is important to note that dietary consumption of fish plays a key role in disease progression. A meta-analysis review examined the relationship of fish consumption and its relation with AMD. A linear association was found between the dose of fish consumption and the risk of AMD, indicating that the more fish consumption a person might have, the lower the risk for AMD.47 Our findings suggest that treatment with ω-3 PUFAs has an effect on the resolution of inflammation, but this effect is not the only determinant factor for disease regression. There was a greater response in the ω-3-treated group than in the ω-3+ω-6-treated group with regard to their anti-inflammatory action; however, both types of treatments were sufficient to increase the retinal ONL thickness. Although there is no direct explanation for the pronounced effect of ω-3 treatment with regard to the number of photoreceptors, it is hypothesised that the mode of action of ω-3 PUFAs may involve a stimulatory effect on endogenous retinal stem cells, such as the Muller glial cells, the ciliary pigment epithelial cells and the RPE cells.48 Further work is encouraged to establish a better understanding of this effect. Nonetheless, supplementation with ω-3 PUFAs has been demonstrated to be beneficial for dry AMD regression when AA/EPA=1–1.5 in this mouse model, and this could be considered a potential therapeutic regimen for patients with dry AMD.

cells.48 Further work is encouraged to establish a better understanding of this effect. Nonetheless, supplementation with ω-3 PUFAs has been demonstrated to be beneficial for dry AMD regression when AA/EPA=1–1.5 in this mouse model, and this could be considered a potential therapeutic regimen for patients with dry AMD. Acknowledgements: The authors would like to thank Professor Heping Xu and Dr. Jose Manuel Romero (School of Medicine, Dentistry and Biochemical Sciences, Queen’s University Belfast, Ireland) for kindly providing training and support with regard to the experimental procedures. In addition, the authors would like to thank Mr Neoklis Makrides for his support throughout the study and Ms Andrea Christofides for her assistance in Western blotting. Contributors: EP planned and conducted the study, analysed the data and reported the results. PK, MK and AN were also involved in the experimental part of the study. GP helped with data analysis. CD and SM supported the study with guidance and provision of equipment. TG was responsible for the overall content of the study and provided guidance and support. Competing interests: TG has a patent for the treatment of eye diseases with omega-3 fatty acids. Provenance and peer review: Not commissioned; externally peer reviewed.

Key messages Panretinal photocoagulation (PRP) is usually administered at the high-risk proliferative diabetic retinopathy (HR-PDR) stage to preserve vision. Administering PRP at an earlier stage of retinopathy (severe non-proliferative diabetic retinopathy) could be more cost-effective than delaying PRP until the HR-PDR stage. A high-quality research trial of laser treatment at an earlier stage (before HR-PDR develops) in combination with antivascular endothelial growth factor drugs in reducing adverse effects is needed. Introduction Diabetic retinopathy (DR) is a major cause of vision loss in the working age population and people with diabetes are 25 times more likely than the general population to go blind.1 Retinopathy can progress through various stages, starting with background (mild) non-proliferative diabetic retinopathy (NPDR), to moderate and severe NPDR, and in some patients to the most serious and sight-threatening form known as proliferative diabetic retinopathy (PDR). If left untreated, this can lead to blindness.

tinopathy can progress through various stages, starting with background (mild) non-proliferative diabetic retinopathy (NPDR), to moderate and severe NPDR, and in some patients to the most serious and sight-threatening form known as proliferative diabetic retinopathy (PDR). If left untreated, this can lead to blindness. A recent review concluded that rates of progression to PDR have fallen over recent times because of earlier identification and treatment of retinopathy, and improved control of blood glucose and blood pressure.2 Nevertheless, DR remains common. A study in Liverpool reported prevalences of any DR and PDR to be 46% and 4% in type 1 diabetes, and 25% and 0.5% in type 2 diabetes, respectively, although the prevalence will vary with mean duration of diabetes, with higher proportions of those with long duration having DR.3 However, a more recent study by Yau et al found that prevalence for PDR was much higher at 7%.4

nd PDR to be 46% and 4% in type 1 diabetes, and 25% and 0.5% in type 2 diabetes, respectively, although the prevalence will vary with mean duration of diabetes, with higher proportions of those with long duration having DR.3 However, a more recent study by Yau et al found that prevalence for PDR was much higher at 7%.4 Scatter or panretinal laser photocoagulation (PRP) is the standard treatment for DR, and in current practice it is administered mostly when DR reaches the high-risk (HR) PDR stage. The aim of PRP treatment is to preserve central vision. However, there are adverse events associated with laser such as visual field defects, retinal fibrosis and epiretinal membrane formation. The ability to drive can also be affected. These harms must be balanced against the benefits of earlier treatment at stages where the risk of blindness is lower. Trials comparing modern lasers with conventional methods have reported little difference in efficacy, but have fewer adverse effects.5 In recent years, antivascular endothelial growth factor (anti-VEGF) drugs have become available, and are used for conditions including diabetic macular oedema (DMO). They have to be injected into the eye, but could be used to reduce the risk of DMO associated with PRP.5 This paper assesses whether offering PRP treatment to patients with severe NPDR (intervention) could be cost-effective compared with delaying treatment until the HR-PDR stage ensues (usual care).

Scatter or panretinal laser photocoagulation (PRP) is the standard treatment for DR, and in current practice it is administered mostly when DR reaches the high-risk (HR) PDR stage. The aim of PRP treatment is to preserve central vision. However, there are adverse events associated with laser such as visual field defects, retinal fibrosis and epiretinal membrane formation. The ability to drive can also be affected. These harms must be balanced against the benefits of earlier treatment at stages where the risk of blindness is lower. Trials comparing modern lasers with conventional methods have reported little difference in efficacy, but have fewer adverse effects.5 In recent years, antivascular endothelial growth factor (anti-VEGF) drugs have become available, and are used for conditions including diabetic macular oedema (DMO). They have to be injected into the eye, but could be used to reduce the risk of DMO associated with PRP.5 This paper assesses whether offering PRP treatment to patients with severe NPDR (intervention) could be cost-effective compared with delaying treatment until the HR-PDR stage ensues (usual care). Methods Base-case analysis A literature search was undertaken for cost-effectiveness studies of the use of PRP and/or anti-VEGF medication for patients with moderate or severe NPDR. However, we could not identify any appropriate studies.5 Therefore, we developed a de novo Markov model in Microsoft Excel to assess the cost-effectiveness of early versus delayed PRP. The different clinical pathways for patients were obtained from information from the Early Treatment Diabetic Retinopathy Study (ETDRS)6 and from expert opinion.5 The model starts with a hypothetical cohort of 1000 patients with diabetes with a starting age of 50 years presenting with moderate NPDR at an ophthalmology clinic. Fifty years was chosen for the starting age, as this was the mean age of patients with DR.5 As DR is a bilateral disease, we have assumed that the model is a two-eye model and that the severity level is the same in each eye. The model assumes that people can progress through all stages of DR, outlined below, but DMO or clinically significant DMO (CSDMO) can occur at any stage:

e mean age of patients with DR.5 As DR is a bilateral disease, we have assumed that the model is a two-eye model and that the severity level is the same in each eye. The model assumes that people can progress through all stages of DR, outlined below, but DMO or clinically significant DMO (CSDMO) can occur at any stage: Moderate NPDR>severe NPDR>early PDR>HR-PDR>severe PDR>severe visual loss/blindness Online supplementary appendix figures 1–4 show the model structures including the different health states for the intervention and usual care arms. Patients receive treatment and at the end of the cycle they move to the corresponding post-treatment health state. After treatment, patients can progress to a more severe health state, regress to earlier stages, remain as they are or die. The key health states in the model are shown in online supplementary appendix table 1. A 30-year time horizon was adopted with each model cycle length set to 6 months7 and transitions between each health state occur at the end of each cycle. A National Health Service (NHS) and personal social services perspective was chosen; a standard annual discount rate of 3.5% was applied to both costs and outcomes. All costs are reported in (£) pounds sterling in 2012/2013 prices. Health outcomes were measured in quality-adjusted life years (QALYs). Results are expressed as incremental cost per QALY gained (incremental cost-effectiveness ratio; ICER). 10.1136/bmjophth-2016-000021.supp1Supplementary file 1

Online supplementary appendix figures 1–4 show the model structures including the different health states for the intervention and usual care arms. Patients receive treatment and at the end of the cycle they move to the corresponding post-treatment health state. After treatment, patients can progress to a more severe health state, regress to earlier stages, remain as they are or die. The key health states in the model are shown in online supplementary appendix table 1. A 30-year time horizon was adopted with each model cycle length set to 6 months7 and transitions between each health state occur at the end of each cycle. A National Health Service (NHS) and personal social services perspective was chosen; a standard annual discount rate of 3.5% was applied to both costs and outcomes. All costs are reported in (£) pounds sterling in 2012/2013 prices. Health outcomes were measured in quality-adjusted life years (QALYs). Results are expressed as incremental cost per QALY gained (incremental cost-effectiveness ratio; ICER). 10.1136/bmjophth-2016-000021.supp1Supplementary file 1 Model inputs Transition probabilities were derived based on information from ETDRS,6 as this was the main source of data for the effects of administering PRP at the severe NPDR or early PDR stages rather than waiting until HR-PDR develops (see Royle et al for further detailed information on these transition probabilities).5 The key transition probabilities are summarised in online supplementary appendix table 2.

s the main source of data for the effects of administering PRP at the severe NPDR or early PDR stages rather than waiting until HR-PDR develops (see Royle et al for further detailed information on these transition probabilities).5 The key transition probabilities are summarised in online supplementary appendix table 2. Health state utility values were estimated based on a weighted average of two papers (see table 1): Brown et al who provided time trade-off values for a range of visual acuities associated with DR8 and Fong et al who reported the number of people with a range of visual acuity for different stages of DR.9 Data from the better seeing eye were used, because quality of life data depends mainly on the better seeing eye. For patients with macular oedema, utility values from a study by Smith et al 10 were used. This methodology was similar to a paper by Ting et al 11 who developed a Markov model of a novel DR prognostic device for DR progression.5 Table 1 Health state utility values for the base-case analysis

antification method using the eye-tracking system can non-invasively evaluate gaze stability in patients with MG almost without discomforts. As an improvement in gaze stability indicates an improvement in the MG disease state,3–6 eye tracking is especially suitable for evaluating treatment efficacy in patients with MG. Although all the patients met the diagnostic criteria of MG3 in the present study, not all patients have undergone all tests. This is one of the limitations of this study, making it impossible to further analyse relationships between the degree of fixation instability of the patients with MG and their disease states. Ocular symptoms associated with MG strongly contribute to the quality of life decline observed in patients with MG.31 Therefore, evaluating diplopia is important, even after treatment has been initiated. The currently used Hess screen and APCT measure ocular motility and position at only one time point. Therefore, they do not accurately evaluate fatigability over time. Evaluating gaze stability using an eye-tracking system is beneficial to diagnosis and evaluation of treatment efficacy in patients with MG because it is non-invasive, quantitative and accurately measures eye position over a long enough period of time to evaluate the muscle fatigability that is characteristic of MG.

.6930 High-risk PDR PT/severe PDR PT 0.7047 0.7047 High-risk PDR and CSDMO PT/severe PDR and CSDMO PT 0.6930 0.6930 Severe visual loss/blindness 0.6218 0.6218 CSDMO, clinically significant diabetic macular oedema; NPDR, non-proliferative diabetic retinopathy; PDR, proliferative diabetic retinopathy; PT, post-treatment. Resource use information was based on information from the Royal College of Ophthalmologists guidelines7 and from our own experience (NL). During each 6-month cycle, the number of ophthalmology and monitoring visits was conservatively estimated as follows: for patients with moderate or severe NPDR they would have one visit, for patients with early PDR they would average 1.5 visits, for patients with HR-PDR or severe PDR they would have two visits, and for patients with severe vision loss/blindness they would have half a visit (ie, one visit per year). For patients receiving PRP treatment, we have assumed that both eyes will be treated at the same time and PRP treatment will be given over two sessions to reduce the risk of DMO. Patients who also have DMO will receive focal laser first for both eyes, and an optical coherence tomography test would also be undertaken. The majority of unit costs were obtained from the NHS reference costs database12 (see table 2). Table 2 Unit costs for the base-case analysis Resource use (HRG code) National average unit cost Source Ophthalmology clinic visit (WF01B) £106 12 Monitoring clinic visit (WF01A) £80 12 PRP laser (OP BZ22B) £131 12 Focal laser (OP BZ22B) £131 12 Optical coherence tomography (OP BZ23Z) £117 12

Resource use information was based on information from the Royal College of Ophthalmologists guidelines7 and from our own experience (NL). During each 6-month cycle, the number of ophthalmology and monitoring visits was conservatively estimated as follows: for patients with moderate or severe NPDR they would have one visit, for patients with early PDR they would average 1.5 visits, for patients with HR-PDR or severe PDR they would have two visits, and for patients with severe vision loss/blindness they would have half a visit (ie, one visit per year). For patients receiving PRP treatment, we have assumed that both eyes will be treated at the same time and PRP treatment will be given over two sessions to reduce the risk of DMO. Patients who also have DMO will receive focal laser first for both eyes, and an optical coherence tomography test would also be undertaken. The majority of unit costs were obtained from the NHS reference costs database12 (see table 2). Table 2 Unit costs for the base-case analysis Resource use (HRG code) National average unit cost Source Ophthalmology clinic visit (WF01B) £106 12 Monitoring clinic visit (WF01A) £80 12 PRP laser (OP BZ22B) £131 12 Focal laser (OP BZ22B) £131 12 Optical coherence tomography (OP BZ23Z) £117 12 Vitrectomy surgery (DC BZ22B) £989 12 Annual cost of blindness* £1483 26 *Excludes residential (home) care. HRF, Healthcare Resource Group; PRP, panretinal photocoagulation.

Resource use (HRG code) National average unit cost Source Ophthalmology clinic visit (WF01B) £106 12 Monitoring clinic visit (WF01A) £80 12 PRP laser (OP BZ22B) £131 12 Focal laser (OP BZ22B) £131 12 Optical coherence tomography (OP BZ23Z) £117 12 Vitrectomy surgery (DC BZ22B) £989 12 Annual cost of blindness* £1483 26 *Excludes residential (home) care. HRF, Healthcare Resource Group; PRP, panretinal photocoagulation. Some patients who receive PRP may also develop complications. We have assumed for one cycle only that a proportion of people who receive PRP will develop macular oedema6 or less often, vitreous haemorrhage (in patients who have severe PDR/severe PDR and CSDMO).13 For this cycle, we have included the appropriate treatment cost and a disutility value of −0.03.14 Age-specific mortality rates used in the model were derived from the UK general population lifetime tables from the Office of National Statistics.15 People with diabetes have about twofold higher mortality than people without diabetes16 17—mortality multiplier of 2.194 was applied in the model; and people with diabetes and advanced retinopathy have higher mortality than people with diabetes but no advanced retinopathy.18 19 We therefore applied four mortality multipliers depending on severity: moderate NPDR=1.118, severe NPDR=1.422, mild PDR=0.992 and moderate/high PDR=1.705.

f 2.194 was applied in the model; and people with diabetes and advanced retinopathy have higher mortality than people with diabetes but no advanced retinopathy.18 19 We therefore applied four mortality multipliers depending on severity: moderate NPDR=1.118, severe NPDR=1.422, mild PDR=0.992 and moderate/high PDR=1.705. Sensitivity analysis To take account of uncertainties in the parameters used in the model and to illustrate sampling uncertainty, we ran the model deterministically and probabilistically with 1000 iterations. These bootstrapped iterations were plotted onto cost-effectiveness planes and they were also used to calculate the cost-effectiveness acceptability curves (CEACs). CEACs were presented using a willingness-to-pay (WTP) threshold from £0 to £50 000. For the probabilistic analysis, the gamma distribution was used for costs and the beta distribution was used for utility values. Various sensitivity analyses were conducted by altering base-case inputs to the model; including a one-way sensitivity analysis (tornado diagram) where a number of key variables were varied holding others constant to understand what was influencing the net monetary benefit of PRP treatment.

me. Evaluating gaze stability using an eye-tracking system is beneficial to diagnosis and evaluation of treatment efficacy in patients with MG because it is non-invasive, quantitative and accurately measures eye position over a long enough period of time to evaluate the muscle fatigability that is characteristic of MG. Contributors: MM: collected patient data, performed statistical analyses, wrote the manuscript, contributed to study concept and design. AH: contributed to study concept and design, collected patient data. KF, KK: collected patient data. RT: critically evaluated the manuscript and contributed to study concept and design. Competing interests: None declared. Patient consent: Obtained. Ethics approval: This study was reviewed and approved by the Institutional Review Board of the University of Toyama (permit number: 27-159). Provenance and peer review: Not commissioned; externally peer reviewed.

was used for utility values. Various sensitivity analyses were conducted by altering base-case inputs to the model; including a one-way sensitivity analysis (tornado diagram) where a number of key variables were varied holding others constant to understand what was influencing the net monetary benefit of PRP treatment. Results In the base-case analysis, if PRP treatment were delayed until HR-PDR developed, this was more costly and less effective than if PRP treatment were administered to patients with severe NPDR (see table 3). Treating earlier with PRP laser meant that fewer people in the intervention arm compared with the usual care arm progressed to more advanced stages of DR. The uncertainty in the results is shown in the cost-effectiveness plane (figure 1A). If a decision maker is willing to pay between £20 000 and £30 000 per QALY, there is a 60% probability that early PRP is more cost-effective than usual care (figure 1B). Table 3 Base-case cost-effectiveness results (discounted) Usual care (PRP at HR-PDR) Intervention (early PRP at severe NPDR) Deterministic Total mean costs (£) £3853 £2753 Total mean QALYs 7.8236 7.9572 Incremental costs (£)/QALYs −£1101/0.1337 ICER (cost per QALY gained) Dominated Probabilistic Total mean costs (£) £3858 £2746 Total mean QALYs 7.8332 7.9624 Incremental costs (£)/QALYs −£1112/0.1292 ICER (cost per QALY gained) Dominated table 4. HR-PDR, high-risk proliferative diabetic retinopathy; ICER, incremental cost-effectiveness ratio; NPDR, non-proliferative diabetic retinopathy; PRP, panretinal photocoagulation; QALYs, quality-adjusted life years.

Total mean costs (£) £3858 £2746 Total mean QALYs 7.8332 7.9624 Incremental costs (£)/QALYs −£1112/0.1292 ICER (cost per QALY gained) Dominated table 4. HR-PDR, high-risk proliferative diabetic retinopathy; ICER, incremental cost-effectiveness ratio; NPDR, non-proliferative diabetic retinopathy; PRP, panretinal photocoagulation; QALYs, quality-adjusted life years. Figure 1 (A) Cost-effectiveness plane—usual care (usual care) versus intervention (early PRP). (B) Cost-effectiveness acceptability curve—usual care (usual care) versus intervention (early PRP). INT, intervention; PRP, panretinal photocoagulation; QALYs, quality-adjusted life years; UC, usual care. Overall, the results from the deterministic sensitivity analyses (Table 4) are in line with the base-case analysis where the intervention dominates usual care. That is, early PRP is still cheaper and more effective than delayed PRP. When we assumed that patients would receive PRP treatment in the intervention arm in the early PDR stages as opposed to the severe NPDR stages, the cost and QALY differences between the two arms were negligible. This may be due to the source of progression data used in the model, as the ETDRS did not report results separately for patients with severe NPDR and early PDR. Table 4 Deterministic sensitivity analysis cost-effectiveness results Usual care (usual care) Intervention (early PRP) Patients in the intervention arm receive PRP treatment at the early PDR or early PDR and CSDMO instead of the severe NPDR or severe NPDR or CSDMO stage.

Overall, the results from the deterministic sensitivity analyses (Table 4) are in line with the base-case analysis where the intervention dominates usual care. That is, early PRP is still cheaper and more effective than delayed PRP. When we assumed that patients would receive PRP treatment in the intervention arm in the early PDR stages as opposed to the severe NPDR stages, the cost and QALY differences between the two arms were negligible. This may be due to the source of progression data used in the model, as the ETDRS did not report results separately for patients with severe NPDR and early PDR. Table 4 Deterministic sensitivity analysis cost-effectiveness results Usual care (usual care) Intervention (early PRP) Patients in the intervention arm receive PRP treatment at the early PDR or early PDR and CSDMO instead of the severe NPDR or severe NPDR or CSDMO stage. Total mean costs (£) £3853 £3725 Total mean QALYs 7.8236 7.8645 ICER (cost per QALY gained) Dominated PRP treatment is administered in one sitting (two laser treatments for two eyes) or in four sittings (eight laser treatments for two eyes) instead of two sittings (four laser treatments for two eyes). We have assumed that the risk of DMO remains the same. One sitting Total mean costs (£) £3762 £2452 Total mean QALYs 7.8236 7.9572 ICER (cost per QALY gained) Two sittings are dominated by one sitting Four sittings

PRP treatment is administered in one sitting (two laser treatments for two eyes) or in four sittings (eight laser treatments for two eyes) instead of two sittings (four laser treatments for two eyes). We have assumed that the risk of DMO remains the same. One sitting Total mean costs (£) £3762 £2452 Total mean QALYs 7.8236 7.9572 ICER (cost per QALY gained) Two sittings are dominated by one sitting Four sittings Total mean costs (£) £4035 £3353 Total mean QALYs 7.8236 7.9572 ICER (cost per QALY gained) Two sittings are dominated by four sittings. CSDMO, clinically significant diabetic macular oedema; DMO, diabetic macular oedema; ICER, incremental cost-effectiveness ratio; NPDR, non-proliferative diabetic retinopathy; PDR, proliferative diabetic retinopathy; PRP, panretinal photocoagulation; QALYs, quality-adjusted life years. Figure 2 is centred around the net monetary benefit ratio of £3774 for intervention when compared with usual care with a WTP threshold set at £20 000. The most important variables influencing the net monetary benefit are the utility values for severe vision loss and early PDR. For the intervention arm, increasing the utility value for severe vision loss by 50% (utility value is capped at 1.00), the net monetary benefit ratio falls to £2964, whereas reducing the utility value for severe vision loss by 50% the net monetary benefit ratio increases to £4584. Other influences include the costs of the focal and PRP lasers.

increasing the utility value for severe vision loss by 50% (utility value is capped at 1.00), the net monetary benefit ratio falls to £2964, whereas reducing the utility value for severe vision loss by 50% the net monetary benefit ratio increases to £4584. Other influences include the costs of the focal and PRP lasers. Figure 2 Tornado diagram for net monetary benefit usual care (usual care) versus intervention (early PRP). NPDR, non-proliferative diabetic retinopathy; OCT, optical coherence tomography; PDR, proliferative diabetic retinopathy; PRP, panretinal photocoagulation; PT, post-treatment; WTP, willingness to pay. Discussion A thorough search for cost-effectiveness studies of the use of PRP and/or anti-VEGF medication for patients with severe NPDR or early PDR found no economic evaluations or modelling-based studies with which to compare our results. In the present study, we built a Markov model to assess whether offering PRP treatment to patients with severe NPDR (intervention) is cost-effective compared with delaying treatment until the HR-PDR stages (usual care). We found that treating earlier at the severe NPDR stage was more cost-effective than delaying treatment until HR-PDR stages. To check the inherent uncertainty in the base-case results, various sensitivity analyses were undertaken and the majority of results were in line with the base-case analyses.

stages (usual care). We found that treating earlier at the severe NPDR stage was more cost-effective than delaying treatment until HR-PDR stages. To check the inherent uncertainty in the base-case results, various sensitivity analyses were undertaken and the majority of results were in line with the base-case analyses. The strengths of the study are: (1) to our knowledge, this is the first model that compares the cost-effectiveness of using PRP treatment at an earlier stage of DR (severe NPDR) as opposed to current practice (HR-PDR); and (2) it contains more detailed health states differentiated by the different severity levels for DR.

The strengths of the study are: (1) to our knowledge, this is the first model that compares the cost-effectiveness of using PRP treatment at an earlier stage of DR (severe NPDR) as opposed to current practice (HR-PDR); and (2) it contains more detailed health states differentiated by the different severity levels for DR. However, the evaluation has various shortcomings. First and most notably is the use of progression data mainly from the ETDRS trial.6 More recent studies have been conducted but they have not addressed the issue of early versus delayed PRP (timing). Due to improved treatments, better blood glucose control and population screening, monitoring and intervention, there has been a reduction in the incidence of severe vision loss/blindness from DR.2 4 Second, even though we conducted a thorough search of the literature, we did not identify any studies with health state utility values by the detailed severity levels that we have in our model or data on disutilities associated with progressing through all the different stages of DR, or data on the disutilities after PRP, especially following the use of modern laser delivery devices. We had to rely heavily on two papers9 10 to characterise the different visual acuity levels into health states and link them to the utility values for patients with DR as reported in Brown et al.8 The key limitation from this methodology was the amalgamation of DR stages, that is, we have the same utility value for a patient with early PDR as someone who has severe PDR,9 and we did not know whether macular oedema was clinically significant or not, and whether there was any visual impairment.10

d in Brown et al.8 The key limitation from this methodology was the amalgamation of DR stages, that is, we have the same utility value for a patient with early PDR as someone who has severe PDR,9 and we did not know whether macular oedema was clinically significant or not, and whether there was any visual impairment.10 Third, we have assumed that the cost of PRP is for a generic PRP laser machine as we did not have costs for PRP with all types of laser machines. Also, we have assumed that the unit costs of PRP and focal laser are the same, even though PRP takes much longer and requires more sessions than focal laser. Even if we had more accurate costs by carrying out ‘bottom-up costing’, we do not believe that this would have made a difference to the ICER. Fourth, PRP destroys retinal tissue, and this may lead to symptoms due to the loss of function of the burned areas, including peripheral visual field defects, reduced night vision and decreased contrast sensitivity. A systematic review by Fong and colleagues reported that visual field defects could occur in up to 50% of treated patients, depending on the intensity of PRP and the level of testing using different isopters.20 They also noted that after PRP, 38% of people reported worsened night driving and 60% worsened dark adaptation. Preti and colleagues reported deterioration in contrast sensitivity after PRP, but giving intravitreal bevacizumab prevented this deterioration.21 The model will not have explicitly captured the costs and effects of any of these adverse visual field effects.

rted worsened night driving and 60% worsened dark adaptation. Preti and colleagues reported deterioration in contrast sensitivity after PRP, but giving intravitreal bevacizumab prevented this deterioration.21 The model will not have explicitly captured the costs and effects of any of these adverse visual field effects. One problem with studies of adverse effects of PRP is that the data come from studies in patients with HR-PDR. If PRP was given at the severe NPDR stage, it is likely possible to reduce the amount of treatment given (compared with what is required to treat HR-PDR) with an expected reduction in adverse effects. In the model, we have assumed that a proportion of people develop adverse events after laser treatment such as DMO and vitreous haemorrhage. However, these complication rates were based on the ETDRS studies6 which may no longer be applicable given that the development of CSDMO following PRP when performed with current protocols and with modern laser technologies appears to be less frequent.5 In the economic model, we have not included a cost or disutility for pain after PRP (although costs for a simple analgesic such as paracetamol are negligible), nor for any adverse events (such as scotomas) due to focal laser.22 However, we believe that this would not alter the magnitude and direction of the ICER because adverse effects would occur whether PRP was early or deferred. Another limitation is that PRP might be administered to some people with NPDR who would not have progressed, so we might treat more people not all of whom would benefit.

r, we believe that this would not alter the magnitude and direction of the ICER because adverse effects would occur whether PRP was early or deferred. Another limitation is that PRP might be administered to some people with NPDR who would not have progressed, so we might treat more people not all of whom would benefit. Trials have been done or are being undertaken comparing anti-VEGF treatment such as ranibizumab, compared with PRP for the treatment of PDR. At 1 year, ranibizumab treatment was non-inferior to PRP, but presumably at a much higher cost because of the number of injections required, making the cost-effectiveness doubtful.23 We are not aware of any trials of anti-VEGF therapy versus PRP in severe NPDR. Finally, we have not differentiated whether the patient had type 1 or type 2 diabetes. The ETDRS reported better results with early intervention (PRP) in type 2 diabetes. So, the key uncertainties are:differences in effectiveness and adverse events between the argon laser-based PRP used in ETDRS and the new laser technologies used today; the quality of life effects using a finer gradation of DR severity; the disutilities of PRP; whether progression is now reduced or slower than was seen in ETDRS; differences between type 1 and type 2 diabetes; whether earlier PRP would be given at severe NPDR or early PDR stage.

Finally, we have not differentiated whether the patient had type 1 or type 2 diabetes. The ETDRS reported better results with early intervention (PRP) in type 2 diabetes. So, the key uncertainties are:differences in effectiveness and adverse events between the argon laser-based PRP used in ETDRS and the new laser technologies used today; the quality of life effects using a finer gradation of DR severity; the disutilities of PRP; whether progression is now reduced or slower than was seen in ETDRS; differences between type 1 and type 2 diabetes; whether earlier PRP would be given at severe NPDR or early PDR stage. Our analysis suggests that earlier PRP could be cost-effective, but there are too many uncertainties to make a firm recommendation. Even though PRP might be cost-effective for the overall population for whom the treatment is intended, there may be occasional individuals for whom the treatment would be harmful in the short term. For example, a patient with good visual acuity but severe NPDR might lose the capability to work or drive due to peripheral visual loss caused by PRP.

t-effective for the overall population for whom the treatment is intended, there may be occasional individuals for whom the treatment would be harmful in the short term. For example, a patient with good visual acuity but severe NPDR might lose the capability to work or drive due to peripheral visual loss caused by PRP. We suggest that a trial is needed that compares early versus deferred PRP with modern laser methods and devices, and also assesses the role of anti-VEGF or steroid drugs in combination with PRP in reducing new DMO or as adjuvant in those with pre-existing DMO. An economic evaluation alongside the trial would collect accurate cost estimates for the different regimens and also collect detailed health-related quality of life measures to enable calculation of QALYs. As generic-based measures such as EQ-5D are said to be insensitive to the changes in visual impairment due to DR progression,24 the trial should also include a disease-specific measure such as NEI-VFQ-25 questionnaire.25 Conclusions Our economic model suggested that PRP administered at the severe NPDR stage might be cost-effective. However, given the limitations of the evidence on current treatments, these results should be interpreted with caution. A trial of early versus deferred laser therapy (with or without anti-VEGF medication) is needed to fully answer the question.

ed that PRP administered at the severe NPDR stage might be cost-effective. However, given the limitations of the evidence on current treatments, these results should be interpreted with caution. A trial of early versus deferred laser therapy (with or without anti-VEGF medication) is needed to fully answer the question. This project was funded by the NIHR Health Technology Assessment programme (HTA: 12/70/01) and the full report was published as part of the HTA monograph series: Royle P, Mistry H, Auguste P, Shyangdan D, Freeman K, Lois N, et al. Pan-retinal photocoagulation and other forms of laser treatment and drug therapies for non-proliferative diabetic retinopathy: systematic review and economic evaluation. Health Technology Assessment.2015;19(51). Contributors: HM developed the economic model and carried out the cost-effectiveness analysis. PA helped to develop the economic model and assisted with the cost-effectiveness analysis. NL provided expert ophthalmological advice for the economic model. NW provided leadership and expert advice. All authors were involved in reviewing, commenting, and editing drafts of the manuscript and all approved the final manuscript. Funding: This research was commissioned as part of the NIHR Health Technology Assessment programme. Competing interests: None declared. Provenance and peer review: Not commissioned; externally peer reviewed. Data sharing statement: Data available on request from the authors.

Key messages What is already known about this subject? Patients with keratoconus have numerous biases in the diagnosis (lack of univariate or multivariate indices with 100% sensitivity and specificity for clinical definitions of keratoconus, grading or staging (index specificity is affected by the corneal topographer) and progression (even for Belin-Ambrosio Enhanced Ectasia Display, the normal measurement noise is required to be known to apply these parameters). Thus far, there is no consistent or clear definition of keratoconus, which affects the scientific power of studies. Consequently, this heterogenic disease requires numerous thorough studies with long-term follow-up. What are the new findings? Evaluations about visual quality indexes, such as aberrometry, imply that the sensory system readapted to diffraction-limited ocular optics once toric implantable collamer lens (TICL) implantation occurred in patients with keratoconus. Selection of maximum tolerable myopic spherical components for patients with longer axial length with relevance to probable myopic regression postoperatively, and considering other warning signs, can yield better refractive result postoperatively. TICL implantation as a second procedure after crosslinking, in addition to refractive correction, improved contrast sensitivity in patients with keratoconus.

Selection of maximum tolerable myopic spherical components for patients with longer axial length with relevance to probable myopic regression postoperatively, and considering other warning signs, can yield better refractive result postoperatively. TICL implantation as a second procedure after crosslinking, in addition to refractive correction, improved contrast sensitivity in patients with keratoconus. How might these results change the focus of research or clinical practice? The clinical outcomes of the current study demonstrate the safety, efficacy and predictability of TICL in stable keratoconus. In addition, the use of this surgical method has had a profound effect on the quality of life of these patients. With all of this in mind, we believe that the experience gained in this case is worth sharing with those of our colleagues.

demonstrate the safety, efficacy and predictability of TICL in stable keratoconus. In addition, the use of this surgical method has had a profound effect on the quality of life of these patients. With all of this in mind, we believe that the experience gained in this case is worth sharing with those of our colleagues. Introduction Keratoconus is a cone-shaped protrusion of the cornea that is derived from the Greek words kerato (cornea) and konos (cone). It is a progressive, non-inflammatory and bilateral thinning of the centreof the cornea and is found to be the most widely seen type of corneal ectasia.1 2 However, the exact mechanism by which it manifests in terms of progression, genetic heterogeneity and phenotypic diversity is not known, thereby resulting in a series of diverse diagnostic and treatment methods.2 3 Essential to each treatment is timing and applying the appropriate intervention method to each patient.4 The implantation of a phakic intraocular lens into the posterior chamber, as demonstrated by clinical observations, can be a desirable alternative to visual defects resulting from refractive errors in the state of keratoconus.5 The STAAR Surgical Visian ICL is an intraocular implant manufactured from a propriety hydroxylethylmethacrylate/porcine collagen-based biocompatible polymer material. The Visian ICL contains an ultraviolet (UV) absorber made from a UV-absorbing material. The Visian ICL has a plate haptic configuration with a central convex/concave optical zone and fuses a forward vault to minimise contact of the Visian ICL with the anterior capsule of the crystalline lens. The Visian ICL has been designed to be set completely inside the back chamber straight behind the iris and in front of the anterior capsule of the human crystalline lens, and when accurately situated, the lens works as a refractive component to progress vision.6

an ICL with the anterior capsule of the crystalline lens. The Visian ICL has been designed to be set completely inside the back chamber straight behind the iris and in front of the anterior capsule of the human crystalline lens, and when accurately situated, the lens works as a refractive component to progress vision.6 The toric implantable collamer lens (TICL) is used in adults aged 21–40 years for correcting myopia (−3.0 to −23.0 D) and astigmatism (≤6 D) in refraction, with stable refraction and anterior chamber depth (ACD) ≥3 mm. It is shown to be highly effective in preserving and improving best-corrected vision and preoperative values such as safety and stability. Nevertheless, issues that are debatable include the effects of TICL and the readaptation of the sensory system to diffraction-limited ocular optics after TICL implantation in patients with keratoconus.5 A comprehensive literature review of PubMed and Web of Science revealed that this investigation is the first one to study the long-term effectiveness of TICL on aberrations and contrast sensitivity in Iranian patients with stable keratoconus in January 2011.

optics after TICL implantation in patients with keratoconus.5 A comprehensive literature review of PubMed and Web of Science revealed that this investigation is the first one to study the long-term effectiveness of TICL on aberrations and contrast sensitivity in Iranian patients with stable keratoconus in January 2011. Methods Study design and subjects Twenty-four patients (40 eyes) age 25–38 years with normal systemic history and no physical signs of ocular disease other than keratoconic eyes were recruited in this study and treated with a sequential corneal collagen crosslinking (CXL) implantable collamer lens (ICL) (with at least a 12-month interval) procedure at the Negah Eye Hospital in Tehran, Iran (figure 1. These cases were selected using a non-random consecutive sampling method. The tenets of the Declaration of Helsinki were followed. Before starting the study, we obtained ethical approval from the Tehran University of Medical Sciences, and all subjects then signed an informed consent form (https://clinicaltrials.gov/ct2/show/NCT02833649). Figure 1 Eligibility assessment procedure. CDVA, corrected distance visual acuity; CXL, corneal collagen crosslinking; TICL, toric implantable collamer lens. Patient enrolment criteria During 6 months after CXL, the refraction was considered to be stable if there was a change in refraction of six subjective refractions within ±0.50 D of spherical equivalent.

Figure 1 Eligibility assessment procedure. CDVA, corrected distance visual acuity; CXL, corneal collagen crosslinking; TICL, toric implantable collamer lens. Patient enrolment criteria During 6 months after CXL, the refraction was considered to be stable if there was a change in refraction of six subjective refractions within ±0.50 D of spherical equivalent. Inclusion criteria We considered the following as inclusion criteria: best corrected distance visual acuity (CDVA) of +0.4 logarithm of the minimum angle of resolution (log MAR) of 20/40 or better, K max ≤55, intraocular pressure (IOP) <20 mm Hg, clear cornea, normal ACD of at least 3 mm to the endothelium width of angle greater than 30°, a pupil diameter of less than 6.25 mm and a preoperative endothelial cell count related to age. Contact lens use was discontinued for at least 3 weeks for rigid lenses and 1 week for soft lenses before any intervention.

mm Hg, clear cornea, normal ACD of at least 3 mm to the endothelium width of angle greater than 30°, a pupil diameter of less than 6.25 mm and a preoperative endothelial cell count related to age. Contact lens use was discontinued for at least 3 weeks for rigid lenses and 1 week for soft lenses before any intervention. Exclusion criteria Exclusion criteria included presentation of autoimmune diseases and other ophthalmic problems, except keratoconus, such as retinal degeneration, corneal scar or opacification, uveitis, cataract, glaucoma, diabetic retinopathy, central endothelial cell count of less than 2000 cells/mm2 by specular microscopy (SP-8800; Konan, Nishinomiya, Japan), central corneal thickness of less than 450 µm (measured by optical pachymetry (Pentacam-HR, Oculus Optikgerate, Wetzlar, Germany)) and ACD <3 mm from the endothelium to the anterior capsule measured by Orbscan IIZ (Orbscan, Bausch and Lomb, Rochester, New York, USA). The classification of keratoconus into four stages was based on the Amsler-Krumeich criteria7; in this practical classification, visual acuity, corneal power, thickness, transparency and astigmatism are considered. We calculate the size and axis of lens by STAAR nomogram (figure 1).

and Lomb, Rochester, New York, USA). The classification of keratoconus into four stages was based on the Amsler-Krumeich criteria7; in this practical classification, visual acuity, corneal power, thickness, transparency and astigmatism are considered. We calculate the size and axis of lens by STAAR nomogram (figure 1). Surgical procedure CXL procedure With the patient positioned under the operating microscope, an eyelid speculum was placed, and with a blunt spatula, the central 9 mm corneal epithelium was removed. The procedure has been illustrated based on Wollensak’s procedure; earlier on, tetracaine 1% and chloramphenicol 0.5% were instilled after informed consent was obtained. By using an epithelial spatula (Malosa Medical, Elland, UK), a 9 mm part of the central epithelium was removed. Five drops of riboflavin 0.1% in dextran 20% (Streuli Pharma, Uznach, Switzerland) were instilled and then reapplied after 5 min. After a period of 10 min, the eye was exposed to ultraviolet A (UVA). UVA exposure was performed for 30 min with 370 nm UVA radiation at 3 mW/cm2 with a beam diameter of 8 mm. During the procedure, riboflavin 0.1% drops were administered every 3 min; if the patient reported discomfort, 1% tetracaine drops were administered. Focus of the UVA beam over the axial cornea was monitored constantly. Neither intraoperative pachymetry nor slit-lamp examination was performed, because this was not part of the original treatment protocol.8 At the end of the procedure, the eye surface was washed with balanced salt solution, two drops of levofloxacin were instilled and a bandage soft contact lens was placed.

constantly. Neither intraoperative pachymetry nor slit-lamp examination was performed, because this was not part of the original treatment protocol.8 At the end of the procedure, the eye surface was washed with balanced salt solution, two drops of levofloxacin were instilled and a bandage soft contact lens was placed. Patients used one drop of 0.3% ciprofloxacin four times per day for 7 days and 0.1% fluorometholone five times per day, with the dosage gradually tapering over 6 weeks postoperatively. After 4 days, the bandage soft contact lens was removed when epithelial healing was confirmed in slit-lamp examination. At 1 and 6 months postoperatively, complete evaluation was performed, and uncorrected distance visual acuity (UDVA), CDVA, refraction and anterior/posterior topography were included. The criteria for the progression of keratoconus were based on the following: (1) increase in the steepest K readings of at least 1.00 D in 1 year, as documented by corneal topography and/or in the Pentacam-HR (according to back surface progression indexes) and (2) deterioration of CDVA.7 Throughout the 6-month follow-up period, no sign of any further progression of keratoconus was recorded.

ng: (1) increase in the steepest K readings of at least 1.00 D in 1 year, as documented by corneal topography and/or in the Pentacam-HR (according to back surface progression indexes) and (2) deterioration of CDVA.7 Throughout the 6-month follow-up period, no sign of any further progression of keratoconus was recorded. According to the results of the autorefractometer refraction and K readings, which are generally not precise in keratoconus and after CXL, all patients received a spectacle at least 6 months before ICL implantation. During follow-up after CXL, CDVA was obtained by focus curve for spherical component and rotating the astigmatism trial axis (Snellen chart). In spectacle administration, the maximum tolerable prescription of spherical equivalent was considered. ICL power was ascertained using the product given by the manufacturer, which was decided on the basis of the horizontal white to white (W–W) distance measured by Orbscan and with a calliper and VuMAX UBM (Sonomed, New York, USA). In addition, Sulcus to Sulcus (S–S) was measured by VuMAX UBM and Quantel Medical’s Linear 50 MHz UBM Probe. A minor clinical modification of ACD was performed by subtracting no more than 0.2 mm whenever corneal anterior bulging was advanced. Axial length measurement was recorded with LensStar9 (Haag-Streit, Koeniz, Switzerland).

, Sulcus to Sulcus (S–S) was measured by VuMAX UBM and Quantel Medical’s Linear 50 MHz UBM Probe. A minor clinical modification of ACD was performed by subtracting no more than 0.2 mm whenever corneal anterior bulging was advanced. Axial length measurement was recorded with LensStar9 (Haag-Streit, Koeniz, Switzerland). Implantable collamer lens insertion procedure At least 12 months after CXL, the TICL implantation was performed. The pupil was dilated with cyclopentolate; to control cyclotorsion, the cornea was marked at four time periods (3, 6, 9 and 12 hours) by the surgeon (FD) through the slid-lamp examination in the upright surgical position. A 3.2 mm clear corneal temporal incision was made (regardless of the astigmatism axis) during the administration of stand-by anaesthesia. The anterior chamber was filled with sodium hyaluronate 1% hydroxypropylmethylcellulose. In the posterior chamber, the ICL was inserted through the incision using the injector cartridge given by the manufacturer, with consideration of the marks on the ICL (right superior and left inferior) to avoid the lens from being upside down. After alignment of the TICL and the proper intended axis in the sulcus, the remaining viscoelastic material was thoroughly removed from the anterior chamber with balanced salt solution. Eye drops containing 0.1% betamethasone and 0.3% ciprofloxacin eye drops were used four times per day for 10 days and then slowly tapered over 3 weeks.

f the TICL and the proper intended axis in the sulcus, the remaining viscoelastic material was thoroughly removed from the anterior chamber with balanced salt solution. Eye drops containing 0.1% betamethasone and 0.3% ciprofloxacin eye drops were used four times per day for 10 days and then slowly tapered over 3 weeks. Study outcomes and patient follow-up Postoperative examinations were conducted at a regular follow-up programme (baseline and at 1, 2, 4, 6 and 12 months and every 6 months thereafter to 4 years) between December 2011 and July 2015. The main outcomes parameters for this study were manifest and cycloplegic refractions, uncorrected distance visual acuity (UCDVA) and CDVA. We evaluated the following: anterior and posterior segments evaluation with dilated fundus examination, operative and postoperative complications, endothelial cell count measured on the central part of the cornea by specular microscopy (SP-8800, IOP with Goldman applanation tonometry and non-contact tonometer Topcon CT-1P. Vault height was measured subjectively (slit-lamp examination) and objectively with ultrasound biomicroscope (UBM). Evaluation of diagnostic technologies Previous studies had concluded that aberrations are dynamic in nature. To evaluate tear film irregularity due to dry eye and fatigue and to analyse intersession repeatability, one experienced examiner (AS) measured the eyes five times successively. Measurements were rechecked by the same examiner from another set of eyes in two consecutive sessions 1 week apart to account for intersession reproducibility.10

lm irregularity due to dry eye and fatigue and to analyse intersession repeatability, one experienced examiner (AS) measured the eyes five times successively. Measurements were rechecked by the same examiner from another set of eyes in two consecutive sessions 1 week apart to account for intersession reproducibility.10 Since we did not have access to the program Assort software (Assort) for vector analysis, we had to use the program Dr Peyman Calculator (http://www.drpeyman.ir/Ophthalmology_Calculator.htm), and graphical displays were performed using Microsoft Excel 2010 (Microsoft Corporation, Redmond, Washington, USA).

lm irregularity due to dry eye and fatigue and to analyse intersession repeatability, one experienced examiner (AS) measured the eyes five times successively. Measurements were rechecked by the same examiner from another set of eyes in two consecutive sessions 1 week apart to account for intersession reproducibility.10 Since we did not have access to the program Assort software (Assort) for vector analysis, we had to use the program Dr Peyman Calculator (http://www.drpeyman.ir/Ophthalmology_Calculator.htm), and graphical displays were performed using Microsoft Excel 2010 (Microsoft Corporation, Redmond, Washington, USA). The tests for contrast sensitivity were performed under mesopic conditions for illumination (3 cd/m2) and photopic (85 cd/m2) using the MediWorks C901 Acuity Chart (Shanghai MediWorks Precision Instruments, Shanghai, China). The tests were performed with best spectacle correction before the operation and without correction after the operation, using a light level of 3 cd/m2 after 10 min of dark adaptation at a distance of 5.5 m. Testing was performed at 1, 3, 5, 6, 12 and 18 cycles per degree (c/d). The defocus curve was also obtained to evaluate the range of functional vision. Corneal, internal and ocular higher-order aberrations (HOAs) were measured. After evaluation, the best objective focus using the Optical Quality Analysis System (the HOA-derived modulation transfer function (MTF) and the root mean square of HOAs) was determined for a 6.0 mm pupil with the ray-tracing aberrometer (iTrace; Tracey Technologies, Houston, Texas, USA). Participants fixated on a near-infrared point light source during the measurements. The room illumination was 42 lx (digital lux metre, LX 1010 B). MTFs were measured for six spatial frequencies (5, 10, 15, 20, 25 and 30 C/D).

mm pupil with the ray-tracing aberrometer (iTrace; Tracey Technologies, Houston, Texas, USA). Participants fixated on a near-infrared point light source during the measurements. The room illumination was 42 lx (digital lux metre, LX 1010 B). MTFs were measured for six spatial frequencies (5, 10, 15, 20, 25 and 30 C/D). Statistical analysis To statistically analyse the results, we used the SPSS software (SPSS Statistics for Windows, V.23.0, 2013; IBM). The non-parametric Wilcoxon signed-rank test was applied to determine the significant differences between the objective results before and after the implantation of TICL, such as contrast sensitivity and the log MAR visual acuity defocus curve. Given that these factors had normal distribution, we report the mean and SD for them. Normal variables were reported as mean and SD, and we sat the median if distributions were skewed. We considered 5% level to find the statistically significant differences in our analysis.

and the log MAR visual acuity defocus curve. Given that these factors had normal distribution, we report the mean and SD for them. Normal variables were reported as mean and SD, and we sat the median if distributions were skewed. We considered 5% level to find the statistically significant differences in our analysis. Results Patient population A summary of patient demographics is provided in table 1. The mean spherical error was −5.06±3.96 D (range: 0.00 to −18.00 D), and the cylindrical error was −3.57±1.56 D (range: −1.25 to −8.00 D). Patients at the time of surgery were aged 30.57±4.69 years (range 25–38 years). TICL was performed at the clinical investigational site from January 2011 to May 2012 in this group. Patients were followed up seven times after surgery at 1 month, 3 and 6 months and then 1, 2, 3 and 4 additional years. All patients had a preoperative uncorrected visual acuity (UCVA) worse than 20/40 with 95% having unaided acuity limited to counting of fingers. At 4 years, postoperative UCVA was better than or equal to preoperative CDVA in 92.50% (37/40) of eyes, and UCVA was increased by ≥1 lines in 25 eyes (table 2). The preoperative CDVA and postoperative UCVA at 4 years after TICL surgery were compared in figure 2. At 4 years postoperatively, 82.5% of eyes were within ±0.50 D, and 97.05% were within ±1.0 D of attempted correction, and the mean spherical and cylindrical manifest refractions were 0.44±0.40 D and −1.01±0.44 D, respectively. At the end of the follow-up, the mean vault was 658±54.33 µm (range, 500–711), and the IOP was 11.7±2.08 mm Hg.

ratively, 82.5% of eyes were within ±0.50 D, and 97.05% were within ±1.0 D of attempted correction, and the mean spherical and cylindrical manifest refractions were 0.44±0.40 D and −1.01±0.44 D, respectively. At the end of the follow-up, the mean vault was 658±54.33 µm (range, 500–711), and the IOP was 11.7±2.08 mm Hg. Table 1 Preoperative and postoperative demographic data of patients undergoing toric implantable collamer lens surgery after 4 years Parameter studied Outcome Refractive surgery (number of eyes) 24 patients; 40 eyes (16 patients bilateral, 8 patients unilateral) Mean age±SD 30.57±4.69 Range 25–38 Gender Male 12 (50%) Female 12 (50%) Preoperative visual acuity (log MAR) UCVA 1.28±0.37 (range: 0.3 to 1.8) CDVA 0.19±0.11 (range: −0.1 to 0.3) Preoperative refractive error (spherical equivalent) Range (−1.75 to −20.50) Mean±SD (−7.55±4.22) Postoperative visual acuity (log MAR) UCVA 0.11±0.13 CDVA 0.04±0.16 Postoperative residual refraction (D) Mean±SD 0.44±0.40 CDVA, corrected distance visual acuity; log MAR, logarithm of the minimum angle of resolution; UCVA, uncorrected visual acuity. Table 2 Manifest refraction, the toric implantable collamer lens

Parameter studied Outcome Refractive surgery (number of eyes) 24 patients; 40 eyes (16 patients bilateral, 8 patients unilateral) Mean age±SD 30.57±4.69 Range 25–38 Gender Male 12 (50%) Female 12 (50%) Preoperative visual acuity (log MAR) UCVA 1.28±0.37 (range: 0.3 to 1.8) CDVA 0.19±0.11 (range: −0.1 to 0.3) Preoperative refractive error (spherical equivalent) Range (−1.75 to −20.50) Mean±SD (−7.55±4.22) Postoperative visual acuity (log MAR) UCVA 0.11±0.13 CDVA 0.04±0.16 Postoperative residual refraction (D) Mean±SD 0.44±0.40 CDVA, corrected distance visual acuity; log MAR, logarithm of the minimum angle of resolution; UCVA, uncorrected visual acuity. Table 2 Manifest refraction, the toric implantable collamer lens Preoperative Postoperative after 4 years Cylinder n/N (%) Sphere n/N (%) Cylinder n/N (%) Sphere n/N (%) ≤−1.50  4/40 (10%) ≤−3.00 12/40 (30%) ≤−1.50 36/40 (91.1%) ±0.25 15/40 (37.5%) ≤−3.50 20/40 (50%) ≤−5.00 20/40 (50%) ≤−3.50 40/40 (100%) ±0.50 24/40 (60%) ≤−5.00 34/40 (85%) ≤−10.00 34/40 (85%) ≤−5.00 40/40 (100%) ±0.75 37/40 (92.5%) ≤−7.00 37/40 (92.5%) ≤−12.00 37/40 (92.5%) ≤−7.00 40/40 (100%) ±1.00 39/40 (97.5%) ≤−8.00 40/40 (100%) ≤−18.00 40/40 (100%) ≤−8.00 40/40 (100%) ±1.25 40/40 (100%) Mean±SD −3.57±1.56 Mean±SD −5.06±3.96 Mean±SD −1.01±0.34 Mean±SD 0.44±0.38 Figure 2 A comparison of preoperative best corrected distance visual acuity and postoperative uncorrected distance visual acuity 4 years after toric implantable collamer lens. UCVA, uncorrected visual acuity.

8.00 40/40 (100%) ±1.25 40/40 (100%) Mean±SD −3.57±1.56 Mean±SD −5.06±3.96 Mean±SD −1.01±0.34 Mean±SD 0.44±0.38 Figure 2 A comparison of preoperative best corrected distance visual acuity and postoperative uncorrected distance visual acuity 4 years after toric implantable collamer lens. UCVA, uncorrected visual acuity. Safety Log MAR CDVA was −0.11±0.11, −0.13±0.15, −0.13±0.15, −0.13±0.14 and −0.14±0.13 at 1 and 3 months and 1, 3 and 4 years after surgery with the ICL, respectively. We found a significant difference between preoperative CDVA ICL and all other follow-up (p<0.05, Wilcoxon signed-rank test). The safety index (mean postoperative CDVA/mean preoperative CDVA) was 0.45±0.56, 0.74±0.87, 0.71±0.92, 0.77±0.74 and 0.77±0.62 at 1 and 3 months and 1, 3 and 4 years after surgery with the ICL, respectively. Treatment effectiveness UCVA at 4 years postoperatively for the entire population was 20/20 or better in 45% of eyes and 20/40 or better in 100% of eyes. The uncorrected visual results in this ‘all eyes’ group must be interpreted in the context: only 82.5% of these eyes had CDVA of 20/20 or better at the baseline. At 4 years, postoperative UCVA was equal to or better than preoperative CDVA in 95% (38/40) of eyes. Figure 2 compares the preoperative CDVA and postoperative UCVA at 4 years after TICL surgery.

is ‘all eyes’ group must be interpreted in the context: only 82.5% of these eyes had CDVA of 20/20 or better at the baseline. At 4 years, postoperative UCVA was equal to or better than preoperative CDVA in 95% (38/40) of eyes. Figure 2 compares the preoperative CDVA and postoperative UCVA at 4 years after TICL surgery. Predictability of manifest refraction (attempted vs achieved) The following outcomes are expected to provide a more accurate assessment of refraction accuracy than the postoperative mean refractive spherical equivalent (MRSE) outcomes. At 4 years postoperatively, 82.5% of eyes were within ±0.50 D, and 97.05% were within ±1.0 D of attempted correction. The differences in SE, cylinder and sphere were statistically significant between preoperative and 1 month postoperatively. These differences remained stable 6 months and 1, 3 and 4 years after operation (figure 3). Even though emmetropia was the targeted postoperative refraction in all patients, small hyperopic and myopic deviations were found after TICL implantation. Figure 3 Postoperative spherical equivalent during follow-up (1 year and 4 years).

Predictability of manifest refraction (attempted vs achieved) The following outcomes are expected to provide a more accurate assessment of refraction accuracy than the postoperative mean refractive spherical equivalent (MRSE) outcomes. At 4 years postoperatively, 82.5% of eyes were within ±0.50 D, and 97.05% were within ±1.0 D of attempted correction. The differences in SE, cylinder and sphere were statistically significant between preoperative and 1 month postoperatively. These differences remained stable 6 months and 1, 3 and 4 years after operation (figure 3). Even though emmetropia was the targeted postoperative refraction in all patients, small hyperopic and myopic deviations were found after TICL implantation. Figure 3 Postoperative spherical equivalent during follow-up (1 year and 4 years). Stability Spherical equivalent: 1 and 4 years after surgery, the mean manifest spherical equivalent was −0.07±0.5 and −0.01±0.48 D, respectively. The spherical equivalent significantly decreased from the baseline to 1 and 4 years (p<0.001, Wilcoxon signed-rank test). Changes in manifest refraction from preoperative to 1 year postoperative were −7.38±4.7 D and from preoperative to 4 years postoperative were −7.44±4.75 D. Astigmatism: 1 and 4 years after surgery, the mean manifest refractive cylinder was −1.03±0.58 and −1.01±0.48 D, respectively. Manifest astigmatism was significantly decreased from the baseline to 1 and 4 years (p<0.001, Wilcoxon signed-rank test). The change in manifest astigmatism from preoperative to 1 year postoperative was −2.77±1.71 D and from preoperative to 4 years postoperative was −2.79±1.78 D.

03±0.58 and −1.01±0.48 D, respectively. Manifest astigmatism was significantly decreased from the baseline to 1 and 4 years (p<0.001, Wilcoxon signed-rank test). The change in manifest astigmatism from preoperative to 1 year postoperative was −2.77±1.71 D and from preoperative to 4 years postoperative was −2.79±1.78 D. Defocus curve Figure 4 shows the log MAR visual acuity under defocus curve of +2, +1, 0, −1, −2, −3 and −4 D in postoperative and preoperative periods in a non-cycloplegic condition. The differences between the measurements of binocular distance corrected defocus curve in the study demonstrated significant differences in log MAR visual acuity at the defocus curve levels of +1, 0 and −1 D, but no significant difference was observed at the defocus curve levels of +2, −2, −3 and −4 D. Figure 4 Log MAR (logarithm of the minimum angle of resolution) visual acuity defocus curve of +2, +1, 0, −1, −2, −3 and −4 D in a non-cycloplegic condition in the preoperative and postoperative. Contrast sensitivity Figure 5A presents the mesopic contrast sensitivity results, which demonstrate no loss of contrast at any spatial frequency and a statistically significant improvement in contrast value at 3 and 1.5 per degree. In addition, photopic contrast sensitivity (figure 5B) demonstrated a significant improvement in contrast value at 3 per degree in similar mesopic conditions. Figure 5 (A) Contrast sensitivity under mesopic illumination (3 cd/m2). *Statistically significantly different at a level of 0.05. (B) Contrast sensitivity under photopic illumination (85 cd/m2).

Contrast sensitivity Figure 5A presents the mesopic contrast sensitivity results, which demonstrate no loss of contrast at any spatial frequency and a statistically significant improvement in contrast value at 3 and 1.5 per degree. In addition, photopic contrast sensitivity (figure 5B) demonstrated a significant improvement in contrast value at 3 per degree in similar mesopic conditions. Figure 5 (A) Contrast sensitivity under mesopic illumination (3 cd/m2). *Statistically significantly different at a level of 0.05. (B) Contrast sensitivity under photopic illumination (85 cd/m2). Aberrometry Preoperative and postoperative corneal and ocular aberrations for the current study, according to evaluation after 4 years, showed significant improvement in total aberrations after TICL implantation (figure 6). Figure 6 Total and corneal aberrations in 40 eyes with keratoconus before and after undergoing toric implantable collamer lens surgery. *Statistically significantly different at a level of 0.05. RMS, root mean square,

Aberrometry Preoperative and postoperative corneal and ocular aberrations for the current study, according to evaluation after 4 years, showed significant improvement in total aberrations after TICL implantation (figure 6). Figure 6 Total and corneal aberrations in 40 eyes with keratoconus before and after undergoing toric implantable collamer lens surgery. *Statistically significantly different at a level of 0.05. RMS, root mean square, Astigmatism analysis by Alpin method Vector analysis of refractive astigmatism shows that the surgically induced astigmatism (SIA) (3.20±1.46 D, range 0.58 to 6.46 D) was not significantly different from the target induced astigmatism (TIA) (3.14±1.42 D, range 0.72 to 6.06 D) (p=0.620), but the mean difference vector (1.04±0.47 D, range 0.00 to 1.98 D) was different from zero (p=0.00). The mean magnitude of error was positive (overcorrection) and close to 0 (0.05±0.68 D, range −1.25 to 1.80 D), and the mean correction index was close to 1 (1.04±0.29 D, range 0.5 to 1.92 D) (figure 7). Figure 7 Vectorial display of the difference vector during postoperative follow-up. Modulation transfer functions MTFs were estimated for six spatial frequencies (5, 10, 15, 20, 25 and 30 C/D) from the ray-tracing aberrometer at a pupil diameter of 6 mm. Figure 8 demonstrated a significant improvement in MTF value at 5 per degree (p=0.004). Figure 8 MTF pre-TICL and post-TICL implantation.

Figure 7 Vectorial display of the difference vector during postoperative follow-up. Modulation transfer functions MTFs were estimated for six spatial frequencies (5, 10, 15, 20, 25 and 30 C/D) from the ray-tracing aberrometer at a pupil diameter of 6 mm. Figure 8 demonstrated a significant improvement in MTF value at 5 per degree (p=0.004). Figure 8 MTF pre-TICL and post-TICL implantation. Intraocular pressure The mean values of IOP changes were 11.67±2.3 mm Hg preoperatively and 13.45±1.75, 13.6±1.78, 12.37±1.59, 11.7±2.08 and 11.27±1.83 mm Hg at 1, 2 and 3 months and at 2 and 4 years, respectively. The mean values were determined using Goldman applanation tonometry and non-contact tonometer Topcon CT-1P at the postoperative follow-up periods. No significant difference was observed in the IOP change at the follow-up period. Endothelial cell substudy The mean endothelial cell count (ECC) changed from 2426.58±107.64 cells/mm2 (range 2210 to 2577 cells/mm2) preoperatively to 2357.36±105.79 cells/mm2 (range 2146.6 to 2548 cells/mm2) after 4 years postoperative. At the end of follow-up, the mean ECC loss was ≤5%. Vaulting the TICL Table 3 demonstrates the change in vault between consecutive measurements in different times. Subjective and objective vaults were stable after TICL implantation. Table 3 Comparison of vault between each pair

Endothelial cell substudy The mean endothelial cell count (ECC) changed from 2426.58±107.64 cells/mm2 (range 2210 to 2577 cells/mm2) preoperatively to 2357.36±105.79 cells/mm2 (range 2146.6 to 2548 cells/mm2) after 4 years postoperative. At the end of follow-up, the mean ECC loss was ≤5%. Vaulting the TICL Table 3 demonstrates the change in vault between consecutive measurements in different times. Subjective and objective vaults were stable after TICL implantation. Table 3 Comparison of vault between each pair Value Change p Value Mean±SD Range Mean±SD Range Vault.MONTH1 561.18±43.86 (450, 650) Vault.M2 644.08±45.45 (540, 712)  82.91±55.22 (−75, 230)  0.042 Vault.M6 672.49±41.79 (557, 725) 111.32±54.19 (−58, 240.3) <0.001 Vault.Y1 658.45±98.22 (109, 720)  97.28±100.15 (−427, 239.3) <0.001 Vault.Y4 641.02±136.31 (66, 721)  79.84±143.02 (−537, 240) <0.001 Vault.M6_Vault.M2  28.41±18.54 (−2, 62.14) <0.001 Vault.Y1_Vault.M2  14.37±101.69 (−603, 70) <0.001 Vault.Y4_Vault.M2 −3.06±135.85 (−602, 71) <0.001 Vault.Y1_Vault.M6 −14.04±95.23 (−601, 10) >0.999 Vault.Y4_Vault.M6 −31.48±131.81 (−600, 12) >0.999 Vault.Y4_Vault.Y1 −17.43±94.02 (−589, 2) >0.999 All p values are based on Bonferroni-adjusted comparison between Vault MONTH1 with each other period. Range: (minimum, maximum). Discussion This study demonstrated the visual outcomes of TICL implantation in stable keratoconus with a long-term follow-up. Our observations were similar to those of other studies regarding the support safety, efficacy, predictability and stability of this procedure in patients with stable keratoconus (table 4).

Range: (minimum, maximum). Discussion This study demonstrated the visual outcomes of TICL implantation in stable keratoconus with a long-term follow-up. Our observations were similar to those of other studies regarding the support safety, efficacy, predictability and stability of this procedure in patients with stable keratoconus (table 4). Table 4 Peer reviewed publications on the use of TICL for refractive correction in corneal ectasia

Discussion This study demonstrated the visual outcomes of TICL implantation in stable keratoconus with a long-term follow-up. Our observations were similar to those of other studies regarding the support safety, efficacy, predictability and stability of this procedure in patients with stable keratoconus (table 4). Table 4 Peer reviewed publications on the use of TICL for refractive correction in corneal ectasia Author(s) year, reference*PubMed, Scopus and science Number of eyes, follow-up for each procedure Preoperative data Postoperative data Complications Final outcomes Kamiya et al, J Refract Surg, 200844 2 eyes (2 patients), 12 months Case 1: −10.00–6.00×100 Case 2: −8.00–2.00×100 Case 1: +0.50–1.00×90 Case 2: −0.25–1.25×100 – – Alfonso et al, J Refract Surg, 200845 25 eyes (16 patients), 12 months Sph: −3.00 to −18.00 D Cyl: −0.50 to −3.00 D UCDVA: 0.17±0.19 CDVA: 0.12±0.12 – SI: 1.05 EI: 0.98 Alfonso et al, J Cataract Refract Surg, 201046 30 eyes (21 patients), 12 months Sph: −5.38±3.26 D Cyl: −3.48±1.24 D Mean UDVA: 0.81±0.20 Mean CDVA: 0.83±0.18 – SI: 1.16 EI: 1.07 Kamiya et al, Graefes Arch Clin Exp Ophthalmol, 201147 27 eyes (14 patients), 6 months SER: −10.11±2.46 D Cyl: −3.03±1.58 D UCDVA: −0.09±0.16 CDVA: −0.15±0.09 – SI: 1.12±0.18 EI: 1.01±0.25 Hashemian et al, Clin Exp Optom, 20135 22 eyes (14 patients), 6 months Sph: −4.98±2.63 D Cyl: −2.77±0.99 D Sph: −0.33±0.51 D Cyl: −1.23±0.65 D UCVA: 0.76±0.23 – SI: 1.40±0.32 EI: 1.24±0.34 Kymionis et al, Ophthalmic Surg Lasers Imaging, 201148 1 patient, 3 months UDVA: C.F CDVA: 0.7 log MAR UCVA: 0.3 CDVA: 0.2 – – Fadlallah et al, J Refract Surg, 20139 9 16 eyes (10 patients), 6 months after CXL, 6 months after TICL Mean UDVA: 1.67±0.49 CDVA: 0.15±0.06 log MAR SE: −7.24±3.53 D Mean UDVA: 0.17±0.06 CDVA: 0.12±0.04 SE: −0.89±0.76 D – SI: 1.08±0.13 EI: 0.97±0.08 Coskunseven et al, Am J Ophthalmol, 200749 3 eyes (2 patients).

.2 – – Fadlallah et al, J Refract Surg, 20139 9 16 eyes (10 patients), 6 months after CXL, 6 months after TICL Mean UDVA: 1.67±0.49 CDVA: 0.15±0.06 log MAR SE: −7.24±3.53 D Mean UDVA: 0.17±0.06 CDVA: 0.12±0.04 SE: −0.89±0.76 D – SI: 1.08±0.13 EI: 0.97±0.08 Coskunseven et al, Am J Ophthalmol, 200749 3 eyes (2 patients). TICL minimum 6 months post-Intacs, 12 months SE: −18.50±2.61 D SE: 0.42 D – – Fernandez-Vega L et al, Invest Ophthalmol Vis, 201050 18 eyes (14 patients), 6 months UDVA: 0.1±0.05 (decimal) CDVA: 0.7±0.19 (decimal) UDVA: 0.6±0.19 CDVA: 0.8±0.14 – – Navas et al, Int Ophthalmol, 201251 11 eyes (8 patients), 38.18±18.7 months SE: −10.52±5.88 D UDVA: 1.31±0.37 log MAR CDVA: 0.28±0.14 log MAR SE: −0.68±0.45 D UDVA: 0.14±0.04 CDVA: 0.16±0.08 – SI: 1.28 EI: 0.88 Coskunseven et al, J Cataract Refract Surg, 201352 14 eyes (9 patients), 12 months UDVA: 0.01 (decimal) CDVA: 0.14 (decimal)SE: −16.40±3.56 D UDVA: 0.44 CDVA: 0.57 SE: −0.80±1.02 D – – Iovieno et al, Eur J Ophthalmol, 201353 7 eyes (7 patients), 12.8±8.8 months UCDVA: 1.18±0.4 SE: −8.09±3.77 D UCDVA: 0.2±0.1 SE: −0.33±0.54 – – Kurian et al, J Cataract Refract Surg, 201254 10 eyes (7 patients), 6 months Mean SE: −7.21±2.25 Mean SE: −0.44±1.21 D – EI: 0.72 SI: 1.13 Ali M et al, Cornea, 201455 29 eyes (16 patients), 3 months CH: 9.2±1.4 mm Hg CRF: 8.4±1.6 mm Hg CH: 8.8±1.3 mm Hg CRF: 8.8±1.6 mm Hg – – Alió et al, J Cataract Refract Surg, 20146 20 eyes received an iris claw phakic IOL, and 28 eyes a PC phakic IOL, 36 months Mean SE: −9.31±4.20 (decimal) UDVA: 0.08±0.09 CDVA: 0.77±0.2 Mean SE: −0.46±0.88 (decimal) UDVA: 0.71±0.26 CDVA: 0.87±0.98 – – Antonios et al, J Ophthalmol, 201556 30 eyes (19 patients), 2 years.

Alió et al, J Cataract Refract Surg, 20146 20 eyes received an iris claw phakic IOL, and 28 eyes a PC phakic IOL, 36 months Mean SE: −9.31±4.20 (decimal) UDVA: 0.08±0.09 CDVA: 0.77±0.2 Mean SE: −0.46±0.88 (decimal) UDVA: 0.71±0.26 CDVA: 0.87±0.98 – – Antonios et al, J Ophthalmol, 201556 30 eyes (19 patients), 2 years. UDVA: 1.57±0.56 CDVA: 0.17±0.08 UDVA: 0.17±0.06 CDVA: 0.11±0.05 – SI: 1.08±0.13 EI: 0.97±0.08 Kamiya et al, Br J Ophthalmol, 20152 21 eyes (11 patients), 3 years SE: −9.70±2.33 D UCVA: 1.46±0.15 log MAR UDVA: −0.06±0.11 CDVA: −0.12±0.09 – – Dirani et al, Eur J Ophthalmol, 201457 11 eyes (7 patients), 6 months UCVA: 1.47±0.38 log MAR CDVA: 0.50±0.22 log MAR UCVA: 0.27±0.20 CDVA: 0.19±0.11 – – Kummelil et al, SYMPOSIUM:KCN, 201358 10 eyes (7 patients), 6 months Mean SE: −7.21±2.25 D Mean SE: −0.44±1.21 D – EI: 0.72 SI: 1.13 Park et al, Korean J Ophthalmol, 201359 Case report, 37-year-old man, 20 months Refraction: −12.0–3.5×30 Refraction: −1.75×180 – – Camoriano et al, J Cataract Refract Surg, 201260 10 eyes (5 patients), 3 years Mean SE: −6.71±0.9 D CDVA: 0±0.03 log MAR SE: −0.58±0.1 D CDVA: −0.04±0.03 1 eye requiring removal and replacement of the toric phakic IOL – Emara et al, J Cataract Refract Surg, 201561 11 eyes (9 myopic, 2 hyperopic), 16.8 months (in myopic eyes) Mean SE: −11.07 D, SE: +8.75 D, CDVA: 20/171 for myopia CDVA: 20/130 for hyperopia Mean SE: −1.40 D CDVA: 20/51 2 hyperopic eyes – Prakash et al, BMJ, 2015.62 Case report, 35-year-old man Right UCVA: 4/60 CDVA: 6/18 Refraction: +0.50–0.50×85, UDVA: 6/12, CDVA: 6/9 – – Alfonso et al,

ropic), 16.8 months (in myopic eyes) Mean SE: −11.07 D, SE: +8.75 D, CDVA: 20/171 for myopia CDVA: 20/130 for hyperopia Mean SE: −1.40 D CDVA: 20/51 2 hyperopic eyes – Prakash et al, BMJ, 2015.62 Case report, 35-year-old man Right UCVA: 4/60 CDVA: 6/18 Refraction: +0.50–0.50×85, UDVA: 6/12, CDVA: 6/9 – – Alfonso et al, J Cataract Refract Surg, 201163 40 eyes (31 patients), 6 months UDVA (decimal): 0.11±0.05 CDVA: 0.56±0.23 Sph: −7.56±7.27 Cyl: −4.19±2.08 UDVA (decimal): 0.50±0.27 CDVA: 0.73±0.20 SE: −1.19±1.33 D – EI: 0.88 SI: 1.28 Shaheen et al, Cornea, 20141 16 eyes (n: 11), 3 years UDVA: 0.63 CDVA: 0.56 (decimal) UDVA: 0.88 CDVA: 0.89 – – CDVA, corrected distance visual acuity; CF, Count Fingers; CH, corneal hysteresis; CRF, corneal resistance factor; CXL, corneal collagen crosslinking; Cyl, cylinder; EI, efficacy index; log MAR, logarithm of the minimum angle of resolution; IOL, intra ocular lens; MCDVA, mean corrected distance visual acuity; MUDVA, mean uncorrected distance visual acuity; MSE, manifest spherical equivalent; PC, posterior chamber; SE, spherical equivalent; SER, spherical equivalent refraction; SI, safety index; Sph, sphere; TICL, toric implantable collamer lens; UCDVA, uncorrected distance visual acuity; UCVA, uncorrected visual acuity; UDVA, uncorrected distance visual acuity.

l acuity; MSE, manifest spherical equivalent; PC, posterior chamber; SE, spherical equivalent; SER, spherical equivalent refraction; SI, safety index; Sph, sphere; TICL, toric implantable collamer lens; UCDVA, uncorrected distance visual acuity; UCVA, uncorrected visual acuity; UDVA, uncorrected distance visual acuity. This technique is currently undergoing an approval procedure from the Food and Drug Administration in the USA. These findings in conjunction with excellent results for refractive indications that affect the quality of life1 enable TICL to be the first recommended phakic IOL approved in the USA for patient younger than 60 years.11 In our study, similar to the report of Gonzalez-Lopez, even amblyopic eyes (two cases) demonstrated significant improvement in UDVA and CDVA.12 The visual acuity test is gradually becoming the gold standard for the assessment of vision, providing solely a restricted quantity of data under artificial conditions. Contrast sensitivity testing presented a variety of visual performance data under genuine conditions.13 14 This motivated us to make progress in this area of medicine. To the best of our knowledge (PubMed and Web of Science), this is the first study of TICL in patients with stable keratoconus (not limited to mild and moderate) with a long follow-up that focuses on visual quality indexes (contrast sensitivity, MTF, defocus curve and aberrations) in January 2011.

this area of medicine. To the best of our knowledge (PubMed and Web of Science), this is the first study of TICL in patients with stable keratoconus (not limited to mild and moderate) with a long follow-up that focuses on visual quality indexes (contrast sensitivity, MTF, defocus curve and aberrations) in January 2011. The amount of aberrations in the eye is related to factors such as age, refraction, severity of keratoconus and even techniques of evaluation.15–22 In the current study, after 4 years of follow-up evaluation of aberration by ray-tracing technology, there was a significant improvement in total aberrations after TICL implantation in comparison with previous surgery, and also MTFs were appraised for six spatial frequencies (5, 10, 15, 20, 25 and 30 c/d) from the ray-tracing aberrometer at a pupil diameter of 6 mm improvement in MTF value at 5 per degree (*p=0.004). An asymmetrically blurred retinal image exerted by the higher-order aberrations in keratoconic eyes with TICL seems to compensate through mechanisms such as the neural visual system and other related components that aid in improving long-term visual experience.23–26 No loss in contrast sensitivity was seen at any spatial frequency. Mesopic contrast showed a statistically significant improvement in value at 3 and 1.5 per degree, and the photopic contrast sensitivity showed a significant improvement in contrast value at 3 per degree similar to mesopic conditions.

The amount of aberrations in the eye is related to factors such as age, refraction, severity of keratoconus and even techniques of evaluation.15–22 In the current study, after 4 years of follow-up evaluation of aberration by ray-tracing technology, there was a significant improvement in total aberrations after TICL implantation in comparison with previous surgery, and also MTFs were appraised for six spatial frequencies (5, 10, 15, 20, 25 and 30 c/d) from the ray-tracing aberrometer at a pupil diameter of 6 mm improvement in MTF value at 5 per degree (*p=0.004). An asymmetrically blurred retinal image exerted by the higher-order aberrations in keratoconic eyes with TICL seems to compensate through mechanisms such as the neural visual system and other related components that aid in improving long-term visual experience.23–26 No loss in contrast sensitivity was seen at any spatial frequency. Mesopic contrast showed a statistically significant improvement in value at 3 and 1.5 per degree, and the photopic contrast sensitivity showed a significant improvement in contrast value at 3 per degree similar to mesopic conditions. There is no completely perfect test for contrast sensitivity.14 We chose this method, because it is user-friendly, time-saving and available, and it reduces the examiners’ error. However, despite all the considerations, the test outcome was influenced by many other factors that improved contrast sensitivity, besides refractive correction of secondary procedure, one-time CXL with riboflavin and UVA, improve contrast sensitivity and aberrations.27 28 We had three patients with preoperative astigmatism greater than six (table 2). A few patients had better tolerance to myopic defocus curve (−1), which seems related to the residual refraction in these patients. Owing to the Alpin analysis of the astigmatism changes, we used the vector technique to assess the magnitude and axis refractive astigmatism variety with the surgery. The mean angle of error indicated that the mean angle of the SIA vector was 0.19±9.88° counter-clockwise to the TIA vector. If the treatment is 100% effective, this vector would be 0. The torque measure of astigmatic change was induced by SIA owing to misalignments of the surgery. In our study, a torque vector of 0.2 D was acquired. The result is close to 0 and shows a trace of astigmatic change induced by SIA.

-clockwise to the TIA vector. If the treatment is 100% effective, this vector would be 0. The torque measure of astigmatic change was induced by SIA owing to misalignments of the surgery. In our study, a torque vector of 0.2 D was acquired. The result is close to 0 and shows a trace of astigmatic change induced by SIA. The major probable complications after ICL implantation are cataract formation,29 acute increase in IOP and night vision disturbance.30 Anterior segment anatomy evaluation with new advanced technologies and attention of the surgeon to warning signs before surgery allow the selection of the proper size and decrease probable complications.31 We think that before surgery, older patients and patients with shallower ACD and larger WTW (White to White) should be made aware of the probability of complication after this procedure32. However,cataract surgery in keratoconus leads to revitalisation of visual acuity, especially by different means such as toric and toric multifocal lenses.33 Possible risk factors for night vision disturbances after ICL are WTW diameter of the cornea, difference between the optic zone diameter and the mesopic pupil size, halo and toricity of the ICL and glare.34 The preoperative and postoperative screening consisted of a complete ophthalmic examination (figure 1).

lenses.33 Possible risk factors for night vision disturbances after ICL are WTW diameter of the cornea, difference between the optic zone diameter and the mesopic pupil size, halo and toricity of the ICL and glare.34 The preoperative and postoperative screening consisted of a complete ophthalmic examination (figure 1). The ICL implantation is as a feasible approach with less encroachment in visual performance, because it does not change the curvature ratios between the anterior and posterior corneas.35 In this respect, although some approaches may show slightly better outcomes for UCVA and refractive predictability, in a study by Alfonso, TICL implantation showed reliable results similar to those of bioptics. A single procedure with TICL implantation may avoid the potential complications for alternative second surgical procedures.36 A trend toward decrement of corneal transplantation for keratoconus comparing two different periods was reported by some studies.37 38 It is a promise that seems related to contemporary management modalities in earlier detection of progression7 and treatments of keratoconus. Divorce commercially available riboflavin (more potent riboflavin with fewer cytotoxicity), better protocols and techniques of crosslinking ‘the procedure in halting the progression of the disease’39–42 can enable the correction of visual defects in patients with keratoconus, thus TICL implantation becomes a perfect refractive surgical correction technique in the future.43

riboflavin with fewer cytotoxicity), better protocols and techniques of crosslinking ‘the procedure in halting the progression of the disease’39–42 can enable the correction of visual defects in patients with keratoconus, thus TICL implantation becomes a perfect refractive surgical correction technique in the future.43 We acknowledge the professional manuscript services of Elsevier. We would like to acknowledge Dr Nima Jalali for his kind help. We would also like to thank Dr Shahin Yazdani, Professor of Ophthalmology at Shahid Beheshti University of Medical Sciences and Health Services. Contributors: AS, FD and SN: collected patient data and wrote the manuscript. FD and FN: concept and study design. FD: critical evaluation of the manuscript. FD, ABR and MG: clinical evaluation for severity of addictions. CA: statistical analysis. Competing interests: None declared. Patient consent: Obtained. Ethics approval: Tehran University of Medical Sciences. Provenance and peer review: Not commissioned; externally peer reviewed.

Key messages What is already known about this subject? Subjects with myasthenia gravis (MG) often present with ocular symptoms, including diplopia and ptosis. Because these symptoms greatly impact patient quality of life, knowing how treatments affect them is important. Current clinical evaluations of MG diplopia are subjective and poor quantification of variability caused by MG fatigability, which make it difficult to study diplopia and evaluate treatment efficacy. What are the new findings? Eye-tracking systems can be used to objectively quantify fixation stability. Using these measurements, it was found that fixation of upward gaze was less stable in the vertical eye position than in the horizontal eye position. Additionally, MG subjects with diplopia have worse fixation stability than both normal subjects and those with a superior oblique palsy. Interestingly, fluctuation in upward gaze fixation seemed to decrease over time in patients with MG. How might these results change the focus of research or clinical practice? Gaze instability caused by fatigability associated with MG can be more accurately evaluated by quantifying fixation with eye tracking and considering how fixation changes over time.

What are the new findings? Eye-tracking systems can be used to objectively quantify fixation stability. Using these measurements, it was found that fixation of upward gaze was less stable in the vertical eye position than in the horizontal eye position. Additionally, MG subjects with diplopia have worse fixation stability than both normal subjects and those with a superior oblique palsy. Interestingly, fluctuation in upward gaze fixation seemed to decrease over time in patients with MG. How might these results change the focus of research or clinical practice? Gaze instability caused by fatigability associated with MG can be more accurately evaluated by quantifying fixation with eye tracking and considering how fixation changes over time. Introduction Approximately 60% of patients with myasthenia gravis (MG) have diplopia at the time of diagnosis.1 Therefore, the majority of patients with MG with diplopia are first examined by an ophthalmologist. Zambelis et al 2 reported that 38% of patients with isolated ptosis and/or diplopia who are referred for electrophysiological evaluation have MG. Furthermore, they stated that the presence of both diplopia and ptosis was more likely due to MG than any other disease. If MG is suspected, additional examinations are needed to obtain a definitive MG diagnosis. These examinations included ptosis evaluation before and after an edrophonium test and an ice pack test.3 Additionally, the presence of acetylcholine receptor antibodies or muscle-specific receptor tyrosine kinase antibodies in the serum, ‘waning’ during repetitive nerve stimulation and ‘increased jitter’ on a single-fibre electromyography can confirm the presence of MG.3

nd after an edrophonium test and an ice pack test.3 Additionally, the presence of acetylcholine receptor antibodies or muscle-specific receptor tyrosine kinase antibodies in the serum, ‘waning’ during repetitive nerve stimulation and ‘increased jitter’ on a single-fibre electromyography can confirm the presence of MG.3 A quantitative MG score and an MG composite scale are used to evaluate disease severity. For these measures, diplopia is evaluated by asking the patient to gaze in one direction. The time from gaze to subjectively determined diplopia appearance is measured.4 5 The 1 min upper gaze test is used to identify ptosis and diplopia and to determine whether the MG fatigue phenomenon is present.6 However, these evaluations are subjective with the patient condition determined using physician observation or patient symptom reporting. In clinical observation, ophthalmologists infer the presence of diplopia and its severity by measuring ocular alignment and motion using the alternate prism cover test (APCT) and the Hess screen test, respectively.7 These tests, however, are not objective and have poor reproducibility in patients with MG because ocular alignment is quantified using only one data point (ocular position). This is problematic because ocular alignment and range of ocular motion change over time in patients with MG7 because their muscles easily fatigue which is a characteristic of MG. Therefore, objective and reproducible evaluations of gaze instability caused by fatigability in patients with MG are needed and should include sequential eye position data obtained with a reasonably high sampling frequency (more than 200 Hz).8–10 There are several studies that have examined saccadic eye movements of patients with MG using electro-oculography or infrared scleral reflection or scleral search coil11–14 and reported that the peak velocity of saccade in patients with MG was similar to or greater than that of normal subjects.11 12 These studies, however, did not evaluate gaze stability of patients with MG by measuring fixation position, although fixation methods are more widely used and meaningful clinical tests (eg, the Quantitative MG score, MG composite scale) for the gaze stability of patients with MG.

ater than that of normal subjects.11 12 These studies, however, did not evaluate gaze stability of patients with MG by measuring fixation position, although fixation methods are more widely used and meaningful clinical tests (eg, the Quantitative MG score, MG composite scale) for the gaze stability of patients with MG. An eye tracker is capable of measuring ocular position continuously for a long enough period with relatively high time resolution. Upward is the most sensitive gaze direction for detecting the ocular MG symptoms (ie, diplopia and ptosis) because extraocular muscle weakness is known to appear most prominently in the elevator muscles of patients with MG.6 15 16 Therefore, in the present study, fixation stability in upward gaze was quantified by using an eye tracker in subjects with and without MG. Materials and methods All study subjects provided written informed consent to participate in the study and all study conduct adhered to the tenets of the Declaration of Helsinki.

An eye tracker is capable of measuring ocular position continuously for a long enough period with relatively high time resolution. Upward is the most sensitive gaze direction for detecting the ocular MG symptoms (ie, diplopia and ptosis) because extraocular muscle weakness is known to appear most prominently in the elevator muscles of patients with MG.6 15 16 Therefore, in the present study, fixation stability in upward gaze was quantified by using an eye tracker in subjects with and without MG. Materials and methods All study subjects provided written informed consent to participate in the study and all study conduct adhered to the tenets of the Declaration of Helsinki. Participants Four groups of volunteers (normal subjects, patients with MG with diplopia, patients with MG without diplopia and patients with superior oblique (SO) palsy) were recruited for the study; the SO palsy group was included to determine whether the presence of diplopia itself affects the fixation stability. The diagnosis of MG was made based on clinical (fluctuating symptoms with easy fatigability and recovery after rest), immunological and neurophysiological findings.3 Diagnostic tests included anti-acetylcholine receptor antibodies, anti-muscle-specific tyrosine kinase antibodies, edrophonium test and decremental muscle response to a train of low-frequency repetitive nerve stimuli. Participants in all groups did not have dyscoria and could clearly watch the target when vision was uncorrected or corrected using soft contact lenses. Patients whose best corrected decimal visual acuity was less than 1.0 (Snellen: 20/20) were not included because of the possible disturbance in visual fixation caused by lowered visual acuity. Patients were not excluded because of age, MG type or disease duration, except for those who could not fixate the target of upward from the beginning.

rected decimal visual acuity was less than 1.0 (Snellen: 20/20) were not included because of the possible disturbance in visual fixation caused by lowered visual acuity. Patients were not excluded because of age, MG type or disease duration, except for those who could not fixate the target of upward from the beginning. Set-up for eye position monitoring The horizontal (X) and vertical (Y) positions of both eyes were recorded using the ViewPoint EyeTracker system (Arrington Research, Scottsdale, AZ, USA) at a sampling rate of 220 Hz, as was previously done by Crossland et al.8–11 The ViewPoint system consists of a display for visual stimulation (Diamondcrysta, RDT222WM-S, Mitsubishi, Tokyo, Japan, resolution was 1680×1050 pixels, and refresh rate was 60 Hz) and two infrared cameras that were mounted on the head positioner. Each subject sat in a chair facing the display set at a distance of 40 cm from the subject, with horizontal eye level even with the centre of the display, the primary position (figure 1). The subject’s head was stabilised against a head and chin rest. The cameras sent images of the eye to a computer via a USB cable. The eye tracker determined eye position using the dark pupil technique and eye-tracking data were stored on a hard disk for offline data analysis. Prior to the fixation stability testing described in the next section, a 16-point grid (4×4 matrix) calibration and a subsequent validation procedure was performed for each subject using software supplied by the eye-tracker manufacturer (ViewPoint EyeTracker Software User Guide).

a hard disk for offline data analysis. Prior to the fixation stability testing described in the next section, a 16-point grid (4×4 matrix) calibration and a subsequent validation procedure was performed for each subject using software supplied by the eye-tracker manufacturer (ViewPoint EyeTracker Software User Guide). Figure 1 The fixation target was positioned 20° above each subject’s horizontal plane (primary position) to induce an upward gaze and was displayed on a computer monitor at a distance of 40 cm. CCD, charge-coupled device. Assessing fixation stability The fixation target consisted of a red inner circle (0.2° in diameter) superimposed at the centre of a black circle (1.5° in diameter) and a white cross was located in the centre of the circular target (figure 1), as previously reported by Thaler et al.17 The target was positioned 20° above each subject’s horizontal plane (primary position) to induce an upward gaze and was displayed on a computer monitor. Subjects were instructed to keep their gaze fixed on the target for 1 min with both eyes. During examinations, patients with ptosis were examined with their eyelids mildly lifted using a surgical tape (to the extent that eyes did not dry) so that the eye tracker could detect the pupil of the subject even when gazing at 20° upward, and therefore, the eyelid descent itself was not a problem.

n with both eyes. During examinations, patients with ptosis were examined with their eyelids mildly lifted using a surgical tape (to the extent that eyes did not dry) so that the eye tracker could detect the pupil of the subject even when gazing at 20° upward, and therefore, the eyelid descent itself was not a problem. Data analysis and statistical methods Data obtained immediately after initiating target presentation were discarded because these data included saccadic eye movements related to beginning fixations. As this period of eye movements varied from subject to subject, the data that accompanied the eye movements were discarded manually by visual inspection of the eye position chart in each subject. Data that contained artefact noise, related to blinking or erroneous pupil image capture, were also excluded after manual confirmation of artefact presence using stored data. For blinks and erroneous pupil image capture, the data obtained 100 ms before and after the onset of these events were removed from analysis.18 Data were selected on the dominant eye in all patients. The dominant eye was determined using the Hole-in-Card test.19 Approximately 5% of the overall data were excluded due to blinks and noisy recording. Fixation stability was first quantified using bivariate contour ellipse areas (BCEA) encompassing fixation points as an index of whole fixation stability, with the data divided into three 20 s fractions (ie, 0–20 s, 20–40 s and 40–60 s). The BCEA was calculated using the following formula8: (1) BCEA=2kπσHσV(1−ρ2)1/2

Data analysis and statistical methods Data obtained immediately after initiating target presentation were discarded because these data included saccadic eye movements related to beginning fixations. As this period of eye movements varied from subject to subject, the data that accompanied the eye movements were discarded manually by visual inspection of the eye position chart in each subject. Data that contained artefact noise, related to blinking or erroneous pupil image capture, were also excluded after manual confirmation of artefact presence using stored data. For blinks and erroneous pupil image capture, the data obtained 100 ms before and after the onset of these events were removed from analysis.18 Data were selected on the dominant eye in all patients. The dominant eye was determined using the Hole-in-Card test.19 Approximately 5% of the overall data were excluded due to blinks and noisy recording. Fixation stability was first quantified using bivariate contour ellipse areas (BCEA) encompassing fixation points as an index of whole fixation stability, with the data divided into three 20 s fractions (ie, 0–20 s, 20–40 s and 40–60 s). The BCEA was calculated using the following formula8: (1) BCEA=2kπσHσV(1−ρ2)1/2 where σH and σV are the SDs of the horizontal and vertical eye positions, respectively, and ρ is their Pearson product-moment correlation. The k is dependent on the probability area chosen and is calculated using the following equation: (2) P=1−e−k

The BCEA was calculated using the following formula8: (1) BCEA=2kπσHσV(1−ρ2)1/2 where σH and σV are the SDs of the horizontal and vertical eye positions, respectively, and ρ is their Pearson product-moment correlation. The k is dependent on the probability area chosen and is calculated using the following equation: (2) P=1−e−k where e is the base of the natural logarithm. For this study, fixation data were calculated using P values of 0.68 (ie, 1 SD), which led to k values of 1.14. Then the SDs of the X and Y eye positions were calculated as indices of detailed (individual direction) stability with the data divided into the three 20 s fractions. Although BCEA and SDs are not independent measures, both parameters were used, because BCEA is presently one of the standard indexes for evaluation of whole fixation stability as described in previous studies.8 10 20 A repeated measures analysis of variance (ANOVA) was performed for the data on 68% BCEA and SDs with groups as a between-subjects variable and the 20 s fractions as a within-subjects variable. If a significant difference was identified, pairwise comparisons were performed using a two-tailed t-test with Bonferroni correction. All statistical analyses were performed using the JMP Pro software package (V.11.2.0, SAS Institute, Cary, NC, USA). P value less than 0.05 was considered statistically significant.

ble. If a significant difference was identified, pairwise comparisons were performed using a two-tailed t-test with Bonferroni correction. All statistical analyses were performed using the JMP Pro software package (V.11.2.0, SAS Institute, Cary, NC, USA). P value less than 0.05 was considered statistically significant. Results Study subjects A total of 21 normal, healthy subjects (mean age=48.9±18.6 years, nine women) who had no strabismus or oculomotor abnormalities were included in the study as a control group. The MG group consisted of 10 patients with MG (mean age=54.7±16.5 years, six women) with diplopia (n=5) and without diplopia (n=5) at disease presentation. These patients were diagnosed with MG based on the presence of antibodies in their serum (eight patients were acetylcholine receptor antibody positive, one was muscle-specific kinase antibody positive and one was seronegative), edrophonium test results (eight of nine patients tested positive and one patient was not tested) and repetitive nerve stimulation results (five of eight patients tested positive and two patients were not tested). The primary ocular deviations of the patients with MG without diplopia measured by APCT were orthophoria or slightly exophoria. In the patients with MG with diplopia, two patients exhibited exotropia: one was 45 prism dioptres (PD) exotropic, the other was both 20 PD exotropic and 3 PD left hypertrophic. Another patient exhibited 14 PD right hypertropia, one exhibited 16 PD exophoria and one was orthophoric but presented with disturbances of adduction and abduction in his right eye. The SO palsy group consisted of six patients with SO palsy with diplopia (mean age=53.5±16.3 years, four women). Patients with SO palsy in this study were unilateral palsy. This group consisted of patients with acquired SO palsy and patients with decompensated congenital SO palsy. They had awareness of diplopia in the upward gaze. Their ocular deviation measured by the synoptophore in the upward gaze was 3.5±5.22° of the horizontal deviation, 4.83±5.24° of the vertical deviation and 3.17±3.44° of the cyclodeviation.

ents with acquired SO palsy and patients with decompensated congenital SO palsy. They had awareness of diplopia in the upward gaze. Their ocular deviation measured by the synoptophore in the upward gaze was 3.5±5.22° of the horizontal deviation, 4.83±5.24° of the vertical deviation and 3.17±3.44° of the cyclodeviation. There were no statistically significant differences in subject ages among groups (ANOVA; F (3, 33)=0.4, P=0.75). Upward gaze stability based on the measure of BCEAs A repeated measures ANOVA test revealed that there was a significant main effect of study groups for the 68% BCEA (F (3, 33)=1.46, P<0.0001). The 68% BCEA was significantly larger in the MG groups (both with diplopia and without diplopia) than in the normal group (two-tailed t-test with Bonferroni correction: Ps<0.05; figure 2). In contrast, there were no significant main effects of 20 s fractions or interaction of groups×fractions (F (2, 66)=1.32, P=0.27 or F (6, 66)=1.02, P=0.42, respectively; table 1). Table 1 Temporal fixation stability. Values are mean±SD of 68% BCEA (the upper table), the horizontal (X) and vertical (Y) fluctuations (the lower table) in the normal (Normal) patients, patients with MG without diplopia (MG diplopia−), patients with MG with diplopia (MG diplopia+) and SO palsy groups during the first (0–20 s), second (20–40 s) and third (40–60 s) fractions in the fixation period Normal (n=21) MG diplopia− (n=5) MG diplopia+ (n=5) SO palsy (n=6) 68% BCEA (°)2

Table 1 Temporal fixation stability. Values are mean±SD of 68% BCEA (the upper table), the horizontal (X) and vertical (Y) fluctuations (the lower table) in the normal (Normal) patients, patients with MG without diplopia (MG diplopia−), patients with MG with diplopia (MG diplopia+) and SO palsy groups during the first (0–20 s), second (20–40 s) and third (40–60 s) fractions in the fixation period Normal (n=21) MG diplopia− (n=5) MG diplopia+ (n=5) SO palsy (n=6) 68% BCEA (°)2 Time from fixation start 0–20 s 3.21±2.89 3.9±1.78 13.97±8.56 6.67±4.49 20–40 s 2.14±1.54 4.12±2.67 10.2±3.75 5.92±4.96 40–60 s 2.02±2.18 4.7±4.19 7.11±7.95 7.06±8.99 SD (°) X Y X Y X Y X Y Time from fixation start 0–20 s 0.49±0.27 1.02±0.62 0.5±0.22 1.22±0.29 1.21±0.61 4.87*±3.16 0.74±0.28 1.58±0.98 20–40 s 0.41±0.22 0.82±0.34 0.43±0.25 1.59±0.62 0.78±0.4 2.41±0.36 0.69±0.31 1.41±0.64 40–60 s 0.36±0.19 0.73±0.34 0.47±0.25 1.31±0.67 0.68±0.32 1.65±0.83 0.71±0.22 1.49±1.25 *Two-tailed t-test with Bonferroni correction P<0.05 versus the third fraction of MG diplopia+. BCEA, bivariate contour ellipse area; MG, myasthenia gravis; SO, superior oblique.

Time from fixation start 0–20 s 3.21±2.89 3.9±1.78 13.97±8.56 6.67±4.49 20–40 s 2.14±1.54 4.12±2.67 10.2±3.75 5.92±4.96 40–60 s 2.02±2.18 4.7±4.19 7.11±7.95 7.06±8.99 SD (°) X Y X Y X Y X Y Time from fixation start 0–20 s 0.49±0.27 1.02±0.62 0.5±0.22 1.22±0.29 1.21±0.61 4.87*±3.16 0.74±0.28 1.58±0.98 20–40 s 0.41±0.22 0.82±0.34 0.43±0.25 1.59±0.62 0.78±0.4 2.41±0.36 0.69±0.31 1.41±0.64 40–60 s 0.36±0.19 0.73±0.34 0.47±0.25 1.31±0.67 0.68±0.32 1.65±0.83 0.71±0.22 1.49±1.25 *Two-tailed t-test with Bonferroni correction P<0.05 versus the third fraction of MG diplopia+. BCEA, bivariate contour ellipse area; MG, myasthenia gravis; SO, superior oblique. Figure 2 Results of 1 min fixation stability testing for normal subjects (white bars), patients with myasthenia gravis (MG) with diplopia (hatched bars), patients with MG without diplopia (black bars) and patients with superior oblique (SO) palsy (grey bars). The 68% bivariate contour ellipse area (BCEA) was significantly larger in the MG with diplopia and MG without diplopia groups than in the normal group (P<0.05). The SD of X position was significantly larger in the MG with diplopia group than in the normal group (P<0.05). The SD of Y position was significantly larger in the MG with diplopia group than other groups and in the MG without diplopia group than in the normal group (P<0.05). Means are shown and error bars indicate SDs. *Indicates P<0.05.

was significantly larger in the MG with diplopia group than in the normal group (P<0.05). The SD of Y position was significantly larger in the MG with diplopia group than other groups and in the MG without diplopia group than in the normal group (P<0.05). Means are shown and error bars indicate SDs. *Indicates P<0.05. Typical fixation scatter plots are shown for each group in figure 3. By observation, the variation in the normal group was smaller (figure 3A) than in all other groups (figure 3B–E). Individual variation was markedly greater in the MG group (figure 3B,C) than in other groups, particularly in the vertical direction. A patient with MG with diplopia was examined before and after MG treatment, which was accompanied with amelioration of the disease state as observed in the clinical symptoms, such as diplopia and ptosis. After the treatment, the BCEAs were markedly decreased from those of the pretreatment values (figure 3E). Figure 3 Typical fixation testing results in a normal subject (A), a patient with myasthenia gravis (MG) without diplopia (B), a patient with MG with diplopia (C) and a patient with superior oblique (SO) palsy (D). Results are also compared in a patient with MG before (E-a) and 2 months after (E-b) initiating MG treatment. Each dot on all plots represents eye position, as measured by the eye tracker. Blue and red ellipses represent the 68% and 95%, respectively, of the bivariate contour ellipse areas (BCEA) encompassing the fixation points.

also compared in a patient with MG before (E-a) and 2 months after (E-b) initiating MG treatment. Each dot on all plots represents eye position, as measured by the eye tracker. Blue and red ellipses represent the 68% and 95%, respectively, of the bivariate contour ellipse areas (BCEA) encompassing the fixation points. Upward gaze stability based on the measure of vertical and horizontal SDs A repeated measures ANOVA test revealed that there was a significant main effect of study groups for SD in both the X and Y eye positions (for X, F (3, 33)=8.90, P=0.0002; for Y, F (3, 33)=17.1, P<0.0001). The SD of X position was significantly larger in the MG with diplopia group than in the normal group (two-tailed t-test with Bonferroni correction, P<0.05; figure 2). The SD of Y position was significantly larger in the MG with diplopia group than the other groups including the MG without diplopia group (two-tailed t-test with Bonferroni correction, all Ps<0.05); the SD of Y position in the MG without diplopia group was also significantly larger than the normal group (two-tailed t-test with Bonferroni correction, P<0.05; figure 2).

he MG with diplopia group than the other groups including the MG without diplopia group (two-tailed t-test with Bonferroni correction, all Ps<0.05); the SD of Y position in the MG without diplopia group was also significantly larger than the normal group (two-tailed t-test with Bonferroni correction, P<0.05; figure 2). For the SDs of X position, there was a main effect of fractions (F (2, 66)=4.81, P=0.011) but no interaction of groups×fractions (F (6, 66)=1.64, P=0.15; table 1). For the SD of Y eye position, there were both main effect of fractions and interaction of groups×fractions (F (2, 66)=7.88, P=0.0009 and F (6, 66)=5.52, P=0.0001, respectively; table 1). The SD of Y eye position in the MG with diplopia group was significantly larger during the first 20 s period than that in the third 20 s period (two-tailed t-test with Bonferroni correction, P<0.05; table 1).

interaction of groups×fractions (F (2, 66)=7.88, P=0.0009 and F (6, 66)=5.52, P=0.0001, respectively; table 1). The SD of Y eye position in the MG with diplopia group was significantly larger during the first 20 s period than that in the third 20 s period (two-tailed t-test with Bonferroni correction, P<0.05; table 1). Discussion Directional selectivity of fixation stability during the upward gaze This study showed that fixation fluctuation during the upward gaze was larger in the vertical direction than in the horizontal direction in all subjects. The upper gaze was used in the current study because it has been shown to be suitable for detecting the MG fatigue phenomenon.6 To the best of our knowledge, this is the first study to examine upward gaze stability in healthy subjects and patients with MG. Therefore, the difference in fixation stability between the vertical and horizontal directions observed in the present study could be related to the direction of the gaze, although, to get solid evidence for this relation, additional tests using gazes in left and right directions are necessary. The directional selectivity of fixation stability observed in the present study may stem from the peripheral mechanism needed to maintain the upward gaze. Anatomical factors could play a role: muscle bulk is greater for the superior rectus (SR) muscle than for the other recti21; the SR muscle load is greater than for the other vertically acting muscle in rabbits22; the SR has fewer muscle spindles than the other recti.23 During upward fixation, the SR and the inferior oblique (IO) contract. Persistent contraction of the SR and the IO may produce fatigue in these extraocular muscles, which could lead to larger fluctuation of eye position in the vertical direction. Indeed, this speculation is consistent with the present observation that the fluctuation in vertical direction was much larger in patients with MG than normal subjects.

f the SR and the IO may produce fatigue in these extraocular muscles, which could lead to larger fluctuation of eye position in the vertical direction. Indeed, this speculation is consistent with the present observation that the fluctuation in vertical direction was much larger in patients with MG than normal subjects. Fixation instability in subjects with MG The present study showed that the upward gaze in patients with MG was more unstable in the vertical position than in normal subjects or patients with SO palsy. Vertical position instability was especially high in the MG with diplopia group. Cleary et al have demonstrated that in patients with MG, weakness of the elevator muscles (ie, the SR and the IO) is significantly greater than that of other extraocular muscles.15 Almog et al also showed that the IO was involved more often than any other extraocular muscle in patients with MG.16 Such a trend of extraocular muscle dysfunction (more severe and frequent impairment of the elevator muscles) in patients with MG, which was the reason for using upward gaze in the present testing, may underlie the fixation instability of patients with MG in upward gaze.

y other extraocular muscle in patients with MG.16 Such a trend of extraocular muscle dysfunction (more severe and frequent impairment of the elevator muscles) in patients with MG, which was the reason for using upward gaze in the present testing, may underlie the fixation instability of patients with MG in upward gaze. In spite of the well-known ocular symptoms of patients with MG, no reports on gaze stability in patients with MG were identified in the academic literature. However, some studies that evaluated central fixation stability using BCEAs in patients with amblyopia and macular degeneration found that visual fixation was unstable in eyes with visual impairments.8 10 20 Patients with strabismus have unstable fixation in the deviated eye.24 Diplopia may also have unstable fixation because of retinal image fusion attempts. However, diplopia is not the underlying cause of the fixation instability observed here in MG subjects because SO palsy subjects (who also had diplopia) did not have the same degree of instability. In addition, the fixation in patients with MG without diplopia was more stable than that in patients with MG with diplopia, but not similar to normal subject. Therefore, it is likely that the marked fixation fluctuation in MG subjects with diplopia was caused by MG fatigability and not by diplopia.

gree of instability. In addition, the fixation in patients with MG without diplopia was more stable than that in patients with MG with diplopia, but not similar to normal subject. Therefore, it is likely that the marked fixation fluctuation in MG subjects with diplopia was caused by MG fatigability and not by diplopia. The present study showed no significant changes in BCEA and SD of Y over time (determined by comparing three consecutive 20 s periods) in normal subjects, in MG subjects without diplopia and in SO palsy subjects. In contrast, patients with MG with diplopia had a tendency to show increased BCEA and SD of Y during the first 20 s period in MG subjects with diplopia. This result was surprising because in the present study it was initially expected that easily fatigued muscles in MG subjects would cause fixation to become more unstable over time. Not yet fully understood, but it may be that extraocular muscles, which at least partially functioned during the first 20 s period, attempted to correct eye position drift towards the upward gaze. However, as muscles became fatigued, ocular position corrections would have become smaller and less frequent, resulting in a smaller fluctuation of eye positions during the second and third 20 s periods compared with those during the first 20 s period. That is, the unexpected result may have been caused by the decrease in eye position correction ability of the patients with MG with diplopia along with the lapse of time. In support of this theory, Feldon et al have reported that patients with MG exhibit commonly substantial instability of eye positions immediately after saccadic movements (they called this ‘wavering fixation’).25 They commented that wavering fixation was the expression of varying levels of tonic innervation to agonist and antagonist resulting from fatigue that caused the eye to drift. Also, in the present study, the vertical eye position in some of the participants of MG group drifted from the target position (data not shown). Furthermore, the vertical SD value of this group further decreased from the second 20 s period (2.41) to the third 20 s period (1.65), although this did not reach the level of the statistical significance. To confirm the present finding, it was necessary to further investigate by increasing the number of cases of MG.

Key messages The conjunctiva is normally colonised with commensal bacteria, and constitutively expresses the NOD-like receptor pyrin domain 3 (NLRP3) inflammasome. In spite of these, the conjunctiva is not normally inflamed. In this study, we found that NLRP3 serves as a sensor capable of discriminating commensal and non-pathogenic bacteria from pathogenic bacteria in conjunctival goblet cells, and that activation of the NLRP3 inflammasome induced by pathogenic bacteria mediates secretion of both mature interleukin-1β and large secretory mucins from these cells. In situations in which NLRP3 continues to be active in the absence of bacteria, inhibition of NLRP3 could be used to treat this inflammation. Introduction The mucosal epithelium acts as both a physical and an immunological barrier, limiting access of pathogenic bacteria and their products to the underlying epithelium. The host response to pathogens relies on detection of pathogens by pattern recognition receptors, such as Toll-like receptors (TLRs) and Nod-Like receptors (NLRs).1 These receptors respond to microbial pathogen-associated molecular patterns (PAMPS), and to altered host molecules known as danger-associated molecular patterns (DAMPS). Activation of these receptors results in production of multiple proinflammatory mediators that coordinate a response ensuring removal or death of the pathogen.2 However, wet surface mucosal epithelia, such as the conjunctiva, are continuously exposed to both commensals and pathogens, and yet limit the inflammatory response.

ation of these receptors results in production of multiple proinflammatory mediators that coordinate a response ensuring removal or death of the pathogen.2 However, wet surface mucosal epithelia, such as the conjunctiva, are continuously exposed to both commensals and pathogens, and yet limit the inflammatory response. Compared with the gut and respiratory tract, the ocular surface is associated with a more limited microbiota, with <100 CFU/µL typically shed in tears, compared with 107–108 CFU/µL in saliva.3 This small population of bacteria does not usually cause ocular surface disease. In tears, commensal Staphylococcus epidermidis usually predominate, but S. aureus is also frequently present.3 S. aureus is the leading cause of ocular surface infection, independent of contact lens wear.4 5 Recent studies in the gut demonstrate that interaction between TLRs and commensal bacteria is critical for maintaining the mucosal barrier to infection.6 Cytoplasmic receptors, such as NLRs, recognise intracellular PAMPs and DAMPs released following damage to host tissue. Thus, activation of NLRs contributes to activating innate pathways of inflammation.7

that interaction between TLRs and commensal bacteria is critical for maintaining the mucosal barrier to infection.6 Cytoplasmic receptors, such as NLRs, recognise intracellular PAMPs and DAMPs released following damage to host tissue. Thus, activation of NLRs contributes to activating innate pathways of inflammation.7 Goblet cells serve as an important component of the innate immune system, and recent studies show that they have much in common with other immune cells.8 9 Inflammasomes are critical mediators of innate immunity, and are constitutively expressed in epithelial goblet cells. The NOD-like receptor pyrin domain 3 (NLRP3) inflammasome is expressed by conjunctival goblet cells,9 whereas the NLRP6 but not the NLRP3 inflammasome is expressed by intestinal goblet cells.10 Goblet cells of the small intestine and more recently the conjunctiva have been shown to be important for intestinal immune homeostasis by delivering antigens from the intestinal lumen and ocular surface, respectively, to the underlying dendritic cells.10 11

NLRP3 inflammasome is expressed by intestinal goblet cells.10 Goblet cells of the small intestine and more recently the conjunctiva have been shown to be important for intestinal immune homeostasis by delivering antigens from the intestinal lumen and ocular surface, respectively, to the underlying dendritic cells.10 11 The large gel-forming mucin MUC5AC is secreted by conjunctival goblet cells, but not the other epithelial cell type in the conjunctiva, the stratified squamous cells.12 This mucin provides a physical barrier and possesses antibacterial properties. MUC5AC expression is stimulated by inflammation.13 14 Ocular surface inflammation is triggered in part through stimulation of TLRs.15 16 All TLRs except TLR8 are present in the conjunctiva.17 TLR2, TLR3, TLR4, TLR5 and TLR9 were found to be activated by bacterial products in either the conjunctiva or cornea.16 17 We previously described constitutive expression of TLRs 1, 2 and 6 by conjunctival goblet cells.9

ed in part through stimulation of TLRs.15 16 All TLRs except TLR8 are present in the conjunctiva.17 TLR2, TLR3, TLR4, TLR5 and TLR9 were found to be activated by bacterial products in either the conjunctiva or cornea.16 17 We previously described constitutive expression of TLRs 1, 2 and 6 by conjunctival goblet cells.9 TLR signalling is the first step in the innate immune response to pathogens, triggering transcriptional induction of pro-interleukin (IL)-1β and pro-IL-18. However, NLR signalling is required for the proteolytic processing of pro-IL-1β and pro-IL-18 through formation of multiprotein complexes, known as inflammasomes.18 The best characterised NLR gene is NLRP3, and the NLRP3 inflammasome is a signalling complex that activates procaspase-1 and induces the proteolytic processing of IL-1β and IL-18. We previously demonstrated that NLRP3 is constitutively expressed by human and rat conjunctival goblet cells.9 Moreover, toxigenic S. aureus activates the NLRP3 inflammasome in rat conjunctival goblet cells to produce mature, active IL-1β.9 In spite of the colonisation of the conjunctiva with commensal bacteria, and the constitutive expression of the NLRP3 inflammasome,3 9 the conjunctiva is not normally inflamed. To understand how the mucosal barrier is differentially affected by commensal, non-pathogenic and pathogenic bacteria, and to determine how inflammation is induced, we asked whether the NLRP3 inflammasome of human conjunctival goblet cells serves as a sensor capable of discriminating non-pathogenic from pathogenic bacteria.

derstand how the mucosal barrier is differentially affected by commensal, non-pathogenic and pathogenic bacteria, and to determine how inflammation is induced, we asked whether the NLRP3 inflammasome of human conjunctival goblet cells serves as a sensor capable of discriminating non-pathogenic from pathogenic bacteria. Materials and methods Human tissue Donor conjunctival tissue was obtained from Heartland Lions Eye Bank (Kansas City, Missouri, USA) and Eversight (Ann Arbor, Michigan, USA) for research purposes. No consent was obtained as no identifying information was transferred. Tissue was placed in Optisol medium within 18 hours of death. Conjunctiva was cleaned of connective tissue before use. Conjunctiva from three different individuals was used for each type of experiment.

nn Arbor, Michigan, USA) for research purposes. No consent was obtained as no identifying information was transferred. Tissue was placed in Optisol medium within 18 hours of death. Conjunctiva was cleaned of connective tissue before use. Conjunctiva from three different individuals was used for each type of experiment. Cell culture Goblet cells were grown in organ culture from human conjunctiva, as described previously.19 Briefly, pieces of minced tissue were placed in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS), 2 mM glutamine (Lonza, Walkersville, Maryland, USA) and 100 mg/mL penicillin/streptomycin in six-well plates. After nodules of cells appeared, the tissue plug and non-goblet cells were removed. After 7 days, the goblet cells were trypsinised and plated in 6, 24 or 96-well plates. As previously published,9 20–23 cultured cells were identified as goblet cells by the following methods: (1) morphology by light microscopy in bright field that indicates numerous small clear granules and positive staining with periodic acid-Schiff’s reagent that stains the secretory product blue and magenta; (2) positive staining with the lectin Helix pomatia agglutinin (HPA) and with an antibody to the large-gel forming mucin MUC5AC that both identify the secretory product; and (3) positive staining with the cytokeratin (CK) 7 that detects keratin filaments in goblet cell bodies and negative staining with CK 4 that detects keratin in stratified squamous cells.

elix pomatia agglutinin (HPA) and with an antibody to the large-gel forming mucin MUC5AC that both identify the secretory product; and (3) positive staining with the cytokeratin (CK) 7 that detects keratin filaments in goblet cell bodies and negative staining with CK 4 that detects keratin in stratified squamous cells. Cell viability measurements To determine the number of live cells, first-passage human conjunctival goblet cells were seeded in 48-well plates. After incubation with bacteria for 1–6 hours, cultured cells were removed by tryspsin digestion. Trypan blue (Sigma-Aldrich, St. Louis, Missouri, USA) was added to trypsinised cells to a final concentration of 0.5% (w/v). The number of live cells was determined using a hemocytometer.

ells were seeded in 48-well plates. After incubation with bacteria for 1–6 hours, cultured cells were removed by tryspsin digestion. Trypan blue (Sigma-Aldrich, St. Louis, Missouri, USA) was added to trypsinised cells to a final concentration of 0.5% (w/v). The number of live cells was determined using a hemocytometer. Adhesion assay Primary human goblet cells were grown to approximately 90% confluence in 24-well plates at 37°C in 5% CO2. Cell monolayers were washed three times with RPMI containing no antibiotics or FBS. Fresh overnight bacterial cultures on blood agar were grown in tryptic soy broth to an OD600 of 0.5 and washed twice with phosphate buffered saline (PBS). Pellets were resuspended in RPMI without antibiotics and FBS, and then added to the goblet cells at 106 CFU/mL. Cells were incubated for 1 hour with bacteria at 37°C in 5% CO2 and then washed three times with PBS to remove non-attached bacteria. Dissociation of the cells was achieved with trypsin, and the resulting solution was serially diluted in PBS. Each dilution was plated on blood agar and incubated overnight at 37°C for colony counting. Adhesion was performed in triplicate and repeated three times. The methicillin-resistant Staphylococcus aureus (MRSA) strain USA300, known to bind to epithelial cells, was used as a positive control.24 25 Results are percentage of adherent bacteria on goblet cells.

r and incubated overnight at 37°C for colony counting. Adhesion was performed in triplicate and repeated three times. The methicillin-resistant Staphylococcus aureus (MRSA) strain USA300, known to bind to epithelial cells, was used as a positive control.24 25 Results are percentage of adherent bacteria on goblet cells. Bacterial culture and challenge of human conjunctival goblet cells Three different bacteria were used in this study. The S. epidermidis strain 1947 was used as a representative of commensal flora. S. epidermidis was obtained from G. Pier (Harvard Medical School, Boston, Massachusetts, USA). S. aureus strain ACL135, which was derived from S. aureus RN6390, is an isogenic attenuated strain defective in the key global regulators agr and sar,26 27 and as a result produces few S. aureus toxins. S. aureus strain ACL135 was used to control for the inflammogenic activity of all other cellular constituents (eg, peptidoglycan, teichoic acids, and so on), aside from secreted toxins in S. aureus RN6390. The toxigenic strain S. aureus RN639028 was used as a positive control. All bacteria were grown at 37°C to an OD595 nm of 0.5 (early log phase). After centrifugation at 500 × g for 10 min, the supernatant was discarded and bacteria were suspended in fresh RPMI-1640 medium with 1% FBS.

creted toxins in S. aureus RN6390. The toxigenic strain S. aureus RN639028 was used as a positive control. All bacteria were grown at 37°C to an OD595 nm of 0.5 (early log phase). After centrifugation at 500 × g for 10 min, the supernatant was discarded and bacteria were suspended in fresh RPMI-1640 medium with 1% FBS. First-passage human conjunctival goblet cells were seeded in culture plates in medium without antibiotics, 24 hours prior to infection. The three strains of bacteria were added at a multiplicity of infection (MOI) of 20 or 60, and incubated for 1 to 6 hours at 37°C in 5% CO2. The incubation time and amount of bacteria was kept low in our experiments to minimise the effect of cell death and release of intracellular components. However, non-toxigenic bacteria activation of caspase-1 showed that this strain of bacteria was active at the conditions used. Measurement of NF-κB activity Goblet cell nuclear extracts were isolated after a 4-hour infection with each of the bacteria. The cells were washed with PBS containing phosphatase inhibitor solution (Cayman Chemical Company, Ann Arbor, USA) and collected using a rubber policeman. The nuclear proteins were extracted according to the manufacturer’s instructions (Cayman Chemical Company). The nuclear fraction was assayed for NF-κB subunit activity, which was determined using an NF-κB (p65) transcription factor assay kit (Cayman Chemical Company).

Arbor, USA) and collected using a rubber policeman. The nuclear proteins were extracted according to the manufacturer’s instructions (Cayman Chemical Company). The nuclear fraction was assayed for NF-κB subunit activity, which was determined using an NF-κB (p65) transcription factor assay kit (Cayman Chemical Company). Western blotting analysis First-passage human conjunctival goblet cells were seeded in six-well plates. After incubation with the three strains of bacteria, supernatant was removed and cells were lysed in RIPA buffer (10 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% deoxycholic acid, 1% Triton X100, 0.1% SDS and 1 mM EDTA). Protein was collected for analysis by western blot. Briefly, the lysate was centrifuged at 10 000 × g for 10 min at 4°C. Sample buffer was added to the lysate, proteins separated by SDS-PAGE through 10% polyacrylamide and transferred to nitrocellulose membrane to be processed for western blot. The primary antibodies were used for western blots at the following dilutions: 1:200 NLRP3, 1:400 pro-IL1β and 1:1000β-actin. NLRP3 and pro-IL-1β antibodies were purchased from R&D Systems (Minneapolis, Minnesota, USA), and β-actin antibody from Sigma Aldrich (St. Louis). Pro-IL-1β and NLRP3 were normalised to β-actin. For detection, secondary antibodies used were donkey anti-sheep (R&D Systems) and chicken anti-goat (EMD Millipore, Billerica, Massachusetts) conjugated to horseradish peroxidase. Immunoreactive bands were visualised by the Enhanced Chemiluminescence method (Thermo Scientific, Rockford, Illinois, USA). Results were expressed as fold increase above the normalised basal level.

-sheep (R&D Systems) and chicken anti-goat (EMD Millipore, Billerica, Massachusetts) conjugated to horseradish peroxidase. Immunoreactive bands were visualised by the Enhanced Chemiluminescence method (Thermo Scientific, Rockford, Illinois, USA). Results were expressed as fold increase above the normalised basal level. Knockdown of NLRP3 First-passage human goblet cells were cultured in RPMI 1640 with 10% FBS. A mixture of four predesigned NLRP3 siRNAs (Smart Pool) were obtained from Thermo Scientific Dharmacon RNAi Technologies (Lafayette, Colorado, USA). Transfection of siRNA was performed with the DharmaFECT 1 siRNA Transfection Reagent following the manufacturer’s protocol. siRNAs constructs were added at a final concentration of 0–100 nM for NLRP3 siRNA in antibiotic-free RPMI 1640 as described previously.29 Scrambled siRNA (sc-siRNA, 100 nM) was used as a negative control. The cells transfected with siRNA were cultured for 48 hours before experimental analysis. Mature IL-1β secretion Supernatants from cells in which pro-IL-1β was measured were used to determine mature IL-1β by ELISA (R&D Systems), which was performed according to the manufacturer’s instructions. The ELISA detected mature IL-1β over a range of 1–2500 pg/mL. Values were expressed as fold increase over basal (no additions).

secretion Supernatants from cells in which pro-IL-1β was measured were used to determine mature IL-1β by ELISA (R&D Systems), which was performed according to the manufacturer’s instructions. The ELISA detected mature IL-1β over a range of 1–2500 pg/mL. Values were expressed as fold increase over basal (no additions). FLICA assay for active caspase-1 analysis Active caspase-1 was detected using a fluorescent inhibitor of caspases (FAM-FLICA Caspase 1 Assay Kit, Immunochemistry Technologies, Bloomington, Minnesota) according to the manufacturer’s instructions. First-passage human goblet cells were grown in 96-well plates. Before use media was changed to antibiotic-free RPMI media containing 1% FBS. The cells were then were stimulated with bacteria for 4 hours and 10 µL of 30X FLICA solution added. Following incubation, cells were stained with Hoechst 33 342 stain (0.5% w/v) and viewed on an inverted phase-contrast microscope equipped for epifluorescence (Eclipse TE300, Nikon, Tokyo, Japan) with a UV filter with excitation 490 nm, emission >520 nm for green fluorescence for caspase-1 positive cells and excitation 365 nm, emission 480 nm for visualisation of nuclei stained with the Hoechst dye. The total number of nuclei in four ×400 fields of view was counted, and the number of cells staining green (indicated active caspase-1) was expressed as a percentage of the total number of cells. For each experiment, an exposure time was set for the basal condition and held constant for the conditions in which bacteria were present.

er of nuclei in four ×400 fields of view was counted, and the number of cells staining green (indicated active caspase-1) was expressed as a percentage of the total number of cells. For each experiment, an exposure time was set for the basal condition and held constant for the conditions in which bacteria were present. Measurement of high-molecular-weight glycoconjugate secretion First-passage human conjunctival goblet cells were seeded on 24-well plates and grown to confluency. Cultured cells were serum-starved in RPMI-1640 supplemented with 0.5% BSA for 2 hours before use, and then stimulated with bacteria for 0.5–8 hours. Goblet cell secretion was measured using an enzyme-linked lectin assay with the lectin Ulex Europaeus Agglutinin (UEA-1). UEA-1 detects high-molecular-weight glycoconjugates, including mucins synthesised and secreted by human goblet cells. The media were collected and analysed for the amount of lectin-detectable glycoconjugates to quantify goblet cell secretion that includes mucins as previously published.21–23 30–32 The amount of glycoprotein secretion was standardised to the amount of total protein in each well as determined by the Bradford assay.33 Glycoconjugate secretion was expressed as fold increase. Statistical analysis Results are presented as mean ±SEM. Data were analysed for significance using Student’s t test. A p value of <0.05 was considered statistically significant.

Measurement of high-molecular-weight glycoconjugate secretion First-passage human conjunctival goblet cells were seeded on 24-well plates and grown to confluency. Cultured cells were serum-starved in RPMI-1640 supplemented with 0.5% BSA for 2 hours before use, and then stimulated with bacteria for 0.5–8 hours. Goblet cell secretion was measured using an enzyme-linked lectin assay with the lectin Ulex Europaeus Agglutinin (UEA-1). UEA-1 detects high-molecular-weight glycoconjugates, including mucins synthesised and secreted by human goblet cells. The media were collected and analysed for the amount of lectin-detectable glycoconjugates to quantify goblet cell secretion that includes mucins as previously published.21–23 30–32 The amount of glycoprotein secretion was standardised to the amount of total protein in each well as determined by the Bradford assay.33 Glycoconjugate secretion was expressed as fold increase. Statistical analysis Results are presented as mean ±SEM. Data were analysed for significance using Student’s t test. A p value of <0.05 was considered statistically significant. Results Comparative viability of conjunctival goblet cells and bacterial cell adhesion with commensal non-toxigenic and toxigenic bacteria Co-culture of human conjunctival goblet cells with S. epidermidis (commensal), S. aureus ACL135 (non-toxigenic) and S. aureus RN6390 (toxigenic, the positive control) was first optimised to permit maximum stimulation of epithelial cells without inducing measurable cell death (figure 1). Human goblet cells were challenged with S. epidermidis, S. aureus ACL135 or S. aureus RN6390 at MOIs of either 20 or 60, and cell viability was assessed over 6 hours by trypan blue exclusion (figure 1).9 Seeding human goblet cell cultures with S. epidermidis, S. aureus ACL135 or S. aureus RN6390 at an initial MOI of 20 resulted in >90% goblet cell viability after a 4-hour incubation in all groups (figure 1), with between 80% and 90% of goblet cells surviving after 6 hours. While no significant differences in cell viability of goblet cells exposed to any of the three strains of bacteria were noted at 1, 2, 4 and 6 hours compared with 0 hour, the 4-hour endpoint with >90% goblet cell viability in all groups was chosen for subsequent experiments.

90% of goblet cells surviving after 6 hours. While no significant differences in cell viability of goblet cells exposed to any of the three strains of bacteria were noted at 1, 2, 4 and 6 hours compared with 0 hour, the 4-hour endpoint with >90% goblet cell viability in all groups was chosen for subsequent experiments. Figure 1 Goblet cell viability with Staphylococcus epidermidis, S. aureus ACL 135 andS. aureus RN6390. Cultured human goblet cells were incubated for 0–6 hours with S. epidermidis (A), S. aureus ACL135 (B) or S. aureus RN6390 (C) at an MOI of 20 (solid circles) and 60 (open circles), and cell viability was determined using the trypan blue exclusion assay. Data are expressed as mean±SEM from three independent experiments. No significant differences were detected. MOI, multiplicity of infection. Because bacteria differ in their ability to adhere to eukaryotic cells, bacterial adherence to goblet cells was measured with all three strains of bacteria at an MOI of 20 for 4 hours (figure 2). Adherence of both S. aureus ACL135 and S. epidermidis (11.8%±3.7% and 11.1%±5.8%, respectively) was not significantly different from that of S. aureus RN6390 that was 13.2%±3.2%.

eukaryotic cells, bacterial adherence to goblet cells was measured with all three strains of bacteria at an MOI of 20 for 4 hours (figure 2). Adherence of both S. aureus ACL135 and S. epidermidis (11.8%±3.7% and 11.1%±5.8%, respectively) was not significantly different from that of S. aureus RN6390 that was 13.2%±3.2%. Figure 2 Bacterial adhesion to goblet cells with Staphylococcus epidermidis, S. aureus ACL 135 and S. aureus RN6390. Cultured human goblet cells were incubated with each strain of bacteria at a multiplicity of infection of 20 for 4 hours and rinsed to remove non-adhered bacteria. Cells were dissociated, plated on blood agar and incubated overnight. The resulting colonies were counted. Data are expressed as mean±SEM from three independent experiments.

blet cells were incubated with each strain of bacteria at a multiplicity of infection of 20 for 4 hours and rinsed to remove non-adhered bacteria. Cells were dissociated, plated on blood agar and incubated overnight. The resulting colonies were counted. Data are expressed as mean±SEM from three independent experiments. Effect of commensal, non-toxigenic and toxigenic bacteria on NF-κB activity and production of pro-IL-1β in conjunctival goblet cells Several studies demonstrate that activation of the NLRP3 inflammasome and release of mature IL-1β requires two steps: (1) a priming step mediated by TLR2, resulting in NF-κB activation and production of pro-IL-1β and NLRP3, and (2) a second stimulus that triggers the assembly of the inflammasome and the activation of caspase-1.34 35 Activation of TLR-dependent signalling components was assessed in human conjunctival goblet cells co-cultured with S. epidermidis, S. aureus ACL135 or S. aureus RN6390 for 4 hours at an MOI of 20 (figure 3). Neither S. epidermidis nor S. aureus ACL135 induced NF-κB activity (1.0±0.03-fold and 0.9±0.05-fold increase over basal, respectively). In contrast, co-culture with the toxigenic S. aureus RN6390, the positive control, resulted in a significant increase in NF-κB activity by 2.0±0.1-fold over basal (figure 3A).

gure 3). Neither S. epidermidis nor S. aureus ACL135 induced NF-κB activity (1.0±0.03-fold and 0.9±0.05-fold increase over basal, respectively). In contrast, co-culture with the toxigenic S. aureus RN6390, the positive control, resulted in a significant increase in NF-κB activity by 2.0±0.1-fold over basal (figure 3A). Figure 3 Effect of Staphylococcus epidermidis, S. aureus ACL 135 and S. aureus RN6390 on NF-κB activity and amount of pro-IL-1β. Cultured human goblet cells were cultured alone for 4 hours (basal (B)) or co-cultured for 4 hours with S. epidermidis, S. aureus ACL135 or S. aureus RN6390 at a multiplicity of infection of 20. The activity of NF-κB was measured in the cell lysates and is presented as fold increase above basal (B) in (A). The amount of pro-IL-1β was also measured by western blot. A representative blot is shown in (B). The amount of pro-IL-1β was standardised to the amount of β-actin and presented as fold increase above basal in (C). Data are expressed as mean±SEM from three independent experiments. * indicates significance from basal (B). IL, interleukin.

L-1β was also measured by western blot. A representative blot is shown in (B). The amount of pro-IL-1β was standardised to the amount of β-actin and presented as fold increase above basal in (C). Data are expressed as mean±SEM from three independent experiments. * indicates significance from basal (B). IL, interleukin. Western blot was used to assess expression of pro-IL-1β in cell lysates prepared from human conjunctival goblet cells cultured under the same conditions used to measure NF-κB activity. Pro-IL-1β was detected as a major band at 31 kDa (figure 3B). There was no significant increase in pro-IL-1β levels in conjunctival goblet cells co-cultured with S. epidermidis (1.1±0.3-fold increase over basal) or S. aureus ACL135 (1.3±0.2-fold over basal). By contrast, co-culture of human conjunctival goblet cells with the toxigenic S. aureus RN6390, the positive control, significantly increased the amount of pro-IL-1β by 2.3±0.4-fold (figure 3C). In summary, neither commensal nor non-toxigenic S. aureus induced NF-κB activation or production of pro-IL-1β components of the TLR pathway.

-culture of human conjunctival goblet cells with the toxigenic S. aureus RN6390, the positive control, significantly increased the amount of pro-IL-1β by 2.3±0.4-fold (figure 3C). In summary, neither commensal nor non-toxigenic S. aureus induced NF-κB activation or production of pro-IL-1β components of the TLR pathway. Effect of commensal, non-toxigenic and toxigenic bacteria on expression of NLRP3 in conjunctival goblet cells Western blot was used to assess NLRP3 expression in cell lysates prepared from human conjunctival goblet cells cultured (figure 4A,B) under the same conditions used to measure NF-κB activity (figure 3). When co-cultured with S. epidermidis or S. aureus ACL135 (figure 4A,B), there was no significant increase in NLRP3 expression detected in the goblet cells (0.8±0.1-fold and 0.9±0.04-fold over basal, respectively). A significant increase was only observed in goblet cells co-cultured with S. aureus RN6390 (1.5±0.2-fold over basal), the positive control. Exposure of goblet cells to commensal or non-toxigenic bacteria did not trigger increased expression of NLRP3, a component of the NLRP3 inflammasome.

old over basal, respectively). A significant increase was only observed in goblet cells co-cultured with S. aureus RN6390 (1.5±0.2-fold over basal), the positive control. Exposure of goblet cells to commensal or non-toxigenic bacteria did not trigger increased expression of NLRP3, a component of the NLRP3 inflammasome. Figure 4 Effect of Staphylococcus epidermidis, S. aureus ACL 135 and S. aureus RN6390 on NLRP3 amount. Cultured human goblet cells were co-cultured for 4 hours with S. epidermidis, S. aureus ACL135 or S. aureus RN6390 at a multiplicity of infection of 20. The amount of NLRP3 was measured by western blot. A representative blot is shown in (A). The amount of NLRP3 was standardised to the amount of β-actin and data presented as fold increase above basal in (B.). Data are expressed as mean±SEM from three independent experiments. * indicates significance from basal (B). NLRP3, NOD-like receptor pyrin domain 3.

western blot. A representative blot is shown in (A). The amount of NLRP3 was standardised to the amount of β-actin and data presented as fold increase above basal in (B.). Data are expressed as mean±SEM from three independent experiments. * indicates significance from basal (B). NLRP3, NOD-like receptor pyrin domain 3. Toxigenic S. aureus-induced production of pro-IL-1β and release of mature IL-1β is NLRP3 dependent in conjunctival goblet cells To determine whether the induction of pro-IL-1β and secretion of mature IL-1β by toxigenic S. aureus was dependent on NLRP3, we knocked down NLRP3 with a mixture of specific siRNAs. First-passage human conjunctival goblet cells were incubated for 72 hours with 20, 50 or 100 nM of NLRP3 siRNAs, or 100 nM of sc-siRNA, the negative control. Under basal conditions, NLRP3 protein was detected at the expected molecular weight of 117 kDa and incubation with sc-siRNA had minimal effects with only a 20% inhibition of NLRP3 protein levels (figure 5A). In contrast 20, 50 and 100 nM of NLRP3 siRNA decreased NLRP3 protein levels by 54%, 75% and 71%, respectively.

. Under basal conditions, NLRP3 protein was detected at the expected molecular weight of 117 kDa and incubation with sc-siRNA had minimal effects with only a 20% inhibition of NLRP3 protein levels (figure 5A). In contrast 20, 50 and 100 nM of NLRP3 siRNA decreased NLRP3 protein levels by 54%, 75% and 71%, respectively. Figure 5 Effect of NLRP3 siRNA on Staphylococcus aureus RN6390-stimulated NLRP3 amount. Cultured human goblet cells were incubated for 48 hours with NLRP3 siRNA at increasing concentrations, and the amount of NLRP3 was determined by western blot analysis. A representative blot is shown in (A). Cultured human goblet cells were incubated with and without 50 nM NLRP3 siRNA or 100 nM sc-siRNA prior to co-culture for 4 hours with S. aureus RN6390 at a multiplicity of infection of 20. Representative blot is shown in (B). The amount of NLRP3 was standardised to the amount of β-actin and data presented as fold increase above basal in (C). Data are expressed as mean±SEM from three independent experiments. # indicates significance from incubation with RN6390 alone. NLRP3, NOD-like receptor pyrin domain 3; sc-siRNA, scrambled siRNA.

in (B). The amount of NLRP3 was standardised to the amount of β-actin and data presented as fold increase above basal in (C). Data are expressed as mean±SEM from three independent experiments. # indicates significance from incubation with RN6390 alone. NLRP3, NOD-like receptor pyrin domain 3; sc-siRNA, scrambled siRNA. The addition of S. aureus RN6390 increased NLRP3 protein expression by 1.4±0.2-fold (Figure 5B and C). The use of sc-siRNA with cultured cells had no significant effect on S. aureus-induced NLRP3 expression. However, the use of 50 nM NLRP3 siRNAs in cultured goblet cells completely blocked the increase in NLRP3 protein expression and decreased it to below basal levels (figure 5B,C). In the same experiments, the expression of pro-IL-1β and secretion of mature IL-1β was determined (figure 6A-C). Under basal conditions, cultured goblet cells expressed a constitutive level of pro-IL-1β, but incubation with S. aureus RN6390 increased the amount of pro-IL-1β by 1.8±0.4-fold (figure 6A,B). S. aureus RN6390-induced expression of pro-IL-1β was completely abolished when NLRP3 was knocked down with siRNA (figure 6B).

. Under basal conditions, cultured goblet cells expressed a constitutive level of pro-IL-1β, but incubation with S. aureus RN6390 increased the amount of pro-IL-1β by 1.8±0.4-fold (figure 6A,B). S. aureus RN6390-induced expression of pro-IL-1β was completely abolished when NLRP3 was knocked down with siRNA (figure 6B). Figure 6 Effect of NLRP3 siRNA on Staphylococcus aureus RN6390-stimulated pro-IL-1β expression and mature IL-1β secretion. Cultured human goblet cells were incubated for 48 hours with 50 nM NLRP3 siRNA or 100 nM sc-siRNA prior to co-culture for 4 hours with S. aureus RN6390 at a multiplicity of infection of 20. The amount of pro-IL-1β was measured by western blot. A representative blot is shown in (A). The amount of pro-IL-1β was standardised to the amount of β-actin and data presented as fold increase over basal in (B). The amount of mature IL-1β secreted was determined by ELISA and is presented as fold increase over basal in (C). Data are expressed as mean±SEM from three independent experiments. * indicates significance from basal; # indicates significance from incubation with bacteria alone. IL, interleukin; NLRP3, NOD-like receptor pyrin domain 3; sc-siRNA, scrambled siRNA.

ermined by ELISA and is presented as fold increase over basal in (C). Data are expressed as mean±SEM from three independent experiments. * indicates significance from basal; # indicates significance from incubation with bacteria alone. IL, interleukin; NLRP3, NOD-like receptor pyrin domain 3; sc-siRNA, scrambled siRNA. Similarly to its effect on pro-IL-1β, NLRP3 knockdown also prevented the S. aureus-induced increase in the secretion of mature IL-1β. Under basal conditions, mature IL-1β was secreted and detected in the culture supernatants. Co-culture with S. aureus RN6390 increased this amount by 2.1±0.1-fold (figure 6C). Stimulation of mature IL-1 β secretion by S. aureus RN6390 was not altered by sc-siRNA, but was completely blocked by NLRP3 siRNAs. These data demonstrate in human conjunctival goblet cells that S. aureus RN6390-induced expression of pro-IL-1β and secretion of mature IL-1β is dependent on NLRP3 signalling.

(figure 6C). Stimulation of mature IL-1 β secretion by S. aureus RN6390 was not altered by sc-siRNA, but was completely blocked by NLRP3 siRNAs. These data demonstrate in human conjunctival goblet cells that S. aureus RN6390-induced expression of pro-IL-1β and secretion of mature IL-1β is dependent on NLRP3 signalling. Effect of commensal, non-toxigenic and toxigenic bacteria on activation of caspase-1 and release of mature IL-1β in conjunctival goblet cells Specific to S. aureus infections, the second step of NLRP3 inflammasome activation can be mediated by bacterial products such as peptidoglycans36 and/or bacterial toxins.37 As a measure of NLRP3 inflammasome activation, caspase-1 activity was evaluated using the FLICA probe FAM-YVAD-FMK that specifically labels active caspase-1,9 in first-passage human conjunctival goblet cells co-cultured with S. epidermidis, S. aureus ACL135 or S. aureus RN6390 at an MOI of 20 for 4 hours. There was no significant caspase-1 activity detected in goblet cells incubated with commensal S. epidermidis (figure 7A,B). Surprisingly, a fivefold increase in caspase-1 activity was observed in goblet cells co-cultured with non-toxigenic S. aureus ACL135 (5.4±1.2-fold over basal) (figure 7B). However, an even greater >15 fold increase in caspase-1 activity was observed in goblet cells co-cultured with the toxigenic S. aureus RN6390 (15.1±4.0-fold over basal).

old increase in caspase-1 activity was observed in goblet cells co-cultured with non-toxigenic S. aureus ACL135 (5.4±1.2-fold over basal) (figure 7B). However, an even greater >15 fold increase in caspase-1 activity was observed in goblet cells co-cultured with the toxigenic S. aureus RN6390 (15.1±4.0-fold over basal). Figure 7 Effect of Staphylococcus epidermidis, S. aureus ACL 135 and S. aureus RN6390 on caspase-1 activity. Cultured human goblet cells were co-cultured for 4 hours with no addition (Basal, (B)), S. epidermidis, S. aureus ACL135 or S. aureus RN6390 at a multiplicity of infection of 20. The activity of caspase-1 was measured using a FLICA assay. Representative micrographs are shown in (A). Magnification ×400. Data shown in (B) are mean±SEM from three independent experiments. * indicates significance from basal (B).

epidermidis, S. aureus ACL135 or S. aureus RN6390 at a multiplicity of infection of 20. The activity of caspase-1 was measured using a FLICA assay. Representative micrographs are shown in (A). Magnification ×400. Data shown in (B) are mean±SEM from three independent experiments. * indicates significance from basal (B). The final end product of NLRP3 inflammasome activation is the secretion of mature IL-1β. IL-1β secretion was measured by ELISA in supernatant of first-passage cultures of human conjunctival goblet cells co-cultured with S. epidermidis, S. aureus ACL135 or S. aureus RN6390 at an MOI of 20 for 4 hours (figure 8). There was no significant increase in IL-1β secretion from goblet cells co-cultured with the commensal S. epidermidis or the non-toxigenic S. aureus ACL135, when compared with basal levels (1.2±0.1-fold and 1.0±0.1-fold, respectively). However, the positive control toxigenic S. aureus RN6390 significantly increased the secretion of mature IL-1β (2.2±0.4-fold over basal). These data demonstrate that commensal and non-toxigenic bacteria cannot activate the NLRP3 inflammasome to stimulate production of mature IL-1β in human conjunctival goblet cells

y). However, the positive control toxigenic S. aureus RN6390 significantly increased the secretion of mature IL-1β (2.2±0.4-fold over basal). These data demonstrate that commensal and non-toxigenic bacteria cannot activate the NLRP3 inflammasome to stimulate production of mature IL-1β in human conjunctival goblet cells Figure 8 Effect of Staphylococcus epidermidis, S. aureus ACL 135 and S. aureus RN6390 on secretion of mature IL-1β. Cultured human goblet cells were co-cultured for 4 hours with S. epidermidis, S. aureus ACL135 or S. aureus RN6390 at a multiplicity of infection of 20. The amount of IL-1β secreted was measured by ELISA and presented as fold increase over basal. Data shown are mean±SEM from three independent experiments. * indicates significance from basal (B). IL, interleukin. Effect of commensal, non-toxigenic and toxigenic bacteria on secretion of high-molecular-weight glycoconjugates by conjunctival goblet cells Human conjunctival goblet cells were co-cultured for 0–8 hours with S. aureus RN6390 at an MOI of 60. As a positive control, cultured cells were treated for 2 hours with a known stimulant of goblet cell secretion, the cholinergic agonist, carbachol at 10−4 M.31 38 39 The toxigenic S. aureus RN6390 induced conjunctival goblet cell glycoconjugate secretion in a time-dependent manner, beginning 1 hour after exposure, becoming significant after 4 hours (1.6±0.2-fold) and peaking at 2.4±0.5-fold by 8 hours (figure 9A).

ell secretion, the cholinergic agonist, carbachol at 10−4 M.31 38 39 The toxigenic S. aureus RN6390 induced conjunctival goblet cell glycoconjugate secretion in a time-dependent manner, beginning 1 hour after exposure, becoming significant after 4 hours (1.6±0.2-fold) and peaking at 2.4±0.5-fold by 8 hours (figure 9A). Figure 9 Effect of Staphylococcus epidermidis, S. aureus ACL 135 and S. aureus RN6390 on high-molecular-weight glycoconjugate secretion. (A) Human conjunctival goblet cells were co-cultured for 0–8 hours with S. aureus RN6390 at an MOI of 60 (circles) or incubated with the cholinergic agonist carbachol (Cch, triangles) at 10−4 M for 2 hours. Glycoconjugate secretion was measured by enzyme-linked lectin assay (ELLA) and presented as fold increase over basal. (B). Cultured human goblet cells were co-cultured for 4 hours with S. epidermidis, S. aureus ACL135 or S. aureus RN6390 at an MOI of 20. Glycoconjugate secretion was measured by ELLA and presented as fold increase above basal. Data shown are mean±SEM from three independent experiments. * indicates significance from basal (B). MOI, multiplicity of infection.

o-cultured for 4 hours with S. epidermidis, S. aureus ACL135 or S. aureus RN6390 at an MOI of 20. Glycoconjugate secretion was measured by ELLA and presented as fold increase above basal. Data shown are mean±SEM from three independent experiments. * indicates significance from basal (B). MOI, multiplicity of infection. Because significant levels of glycoconjugate secretion were detected at 4 hours and to minimise the impact of bacteria on cell viability, all subsequent secretion experiments were performed at 4 hours. To determine the ability of commensal S. epidermidis and non-toxigenic S. aureus ACL135 to stimulate goblet cell secretion, human conjunctival goblet cells were incubated with each strain at an MOI of 20 for 4 hours (figure 9B). Neither S. aureus ACL135 nor S. epidermidis induced goblet cell secretion above basal levels (1.1±0.1-fold and 1.1±0.04-fold, respectively), while the positive control, toxigenic S. aureus RN6390 significantly increased goblet cell secretion by 2.0±0.1-fold. Thus, neither commensal nor non-toxigenic bacteria can stimulate conjunctival goblet cell secretion.

idis induced goblet cell secretion above basal levels (1.1±0.1-fold and 1.1±0.04-fold, respectively), while the positive control, toxigenic S. aureus RN6390 significantly increased goblet cell secretion by 2.0±0.1-fold. Thus, neither commensal nor non-toxigenic bacteria can stimulate conjunctival goblet cell secretion. Discussion Our results demonstrate that conjunctival goblet cells discriminate isogenic S. aureus strains based on expression of postexponential phase secreted products including toxins,40 as well as between pathogenic and commensal staphylococci. Of the strains tested, the commensal and non-toxigenic ones did not cause activation of the NLRP3 inflammasome and high-molecular-weight glycoprotein secretion. Only toxigenic S. aureus, the positive control, was effective. In most wet mucosal surfaces such as lung, intestine and colon, monocytes or macrophages mediate a response to toxigenic or commensal bacteria.41–44 However, it is possible that epithelial cells in these tissues are responsive to bacteria as well, but less is known about the inflammasome response of epithelial cells in the absence of immune effectors. In the skin, the keratinocytes respond to both commensal S. epidermidis 1457 and toxigenic S. aureus (113 ΔspA), by secretion of antimicrobial factors45; however, commensal bacteria induce a smaller response. While our data show no induction of the NLRP3 inflammasome or mucin secretion in conjunctival goblet cells exposed to commensal or non-toxogenic S. aureus, we cannot rule out that conjunctival goblet cells may respond to non-toxigenic and commensal bacteria when exposed to greater bacterial loads than tested here.

nse. While our data show no induction of the NLRP3 inflammasome or mucin secretion in conjunctival goblet cells exposed to commensal or non-toxogenic S. aureus, we cannot rule out that conjunctival goblet cells may respond to non-toxigenic and commensal bacteria when exposed to greater bacterial loads than tested here. NLRP3 inflammasome activation is believed to be dependent on the concomitant activation of two pathways.9 The first pathway involves activation of an TLR receptor to increase synthesis of pro-IL-1β, NLRP3 and the other components of the NLRP3 inflammasome; the second pathway involves opening of an ion channel that triggers the activation of caspase-1. Activated caspase-1 then cleaves pro-IL-1β to produce mature IL-1β. In this study, we examined the components of both pathways, including NF-κB activation, production of pro-IL-1β and NLRP3 expression characteristic of the TLR pathway, and caspase-1 activation and release of mature IL-1β components of the NLRP3 pathway. In human conjunctival goblet cells, we found that neither commensal nor non-toxigenic strains triggered the activation of NF-κB, increased expression of pro-IL-1β and NLRP3, and secretion of mature IL-1β. To our surprise, though, exposure of human conjunctival goblet cells to non-toxigenic S. aureus, but not commensal S. epidermidis, resulted in significant activation of caspase-1. However, in the absence of NF-κB activation and production of pro-IL-1β, non-toxigenic bacteria were unable to stimulate the release of mature, active IL-1β. This result suggests that there are either alternative pathways for activation of caspase-1 or additional triggers that promote secretion of mature IL-1β. Known pathways include inhibition of GAPDH and glycolysis, activation of lipid metabolic pathways and stimulation of pyroptosis,46 and it will be of interest to determine their relevance to the observations made here.

ve pathways for activation of caspase-1 or additional triggers that promote secretion of mature IL-1β. Known pathways include inhibition of GAPDH and glycolysis, activation of lipid metabolic pathways and stimulation of pyroptosis,46 and it will be of interest to determine their relevance to the observations made here. Previous studies suggest that the ocular surface is normally maintained in a non-inflamed state, despite being in constant contact with bacteria from the skin and environment.17 The findings of this study demonstrating that neither commensal nor non-toxigenic bacteria trigger NLRP3 inflammasome activation or mucin secretion in conjunctival goblet cells, are in agreement with this concept. Secretion of mature IL-1β and mucin from conjunctival goblet cells was induced only by the toxigenic strain of S. aureus, similar to those frequently causing infection

al nor non-toxigenic bacteria trigger NLRP3 inflammasome activation or mucin secretion in conjunctival goblet cells, are in agreement with this concept. Secretion of mature IL-1β and mucin from conjunctival goblet cells was induced only by the toxigenic strain of S. aureus, similar to those frequently causing infection In addition to conjunctiva, goblet cells are present in airways, the gastrointestinal tract and within the nasal epithelium. These cells are packed with secretory granules filled with large gel-forming mucins, MUC5AC, MUC5B or MUC2, which protect the epithelium by entrapping bacteria and preventing bacterial contact with the epithelium.40 47–49 Goblet cells also secrete key proteins involved in innate defence, with secreted proteins from gastrointestinal goblet cells being the most thoroughly studied.50 Proteins secreted include Ca2+ channel protein CLCA1; Fc globulin-binding protein (FCGBP), which binds and crosslinks mucus proteins; ZG16, which binds Gram-positive bacteria; AGR2, an ER residential protein; and trefoil factor 3 (TFF3). These proteins are secreted in high concentration into the mucus. We show here that toxigenic bacteria activate protective conjunctival goblet cell mucous secretion, in addition to activating the NLRP3 inflammasome. Thus, goblet cells have two different protective mechanisms against toxigenic bacteria that threaten the ocular surface. Neither of these protective mechanisms were activated by the commensal or non-toxigenic bacteria strains used.

unctival goblet cell mucous secretion, in addition to activating the NLRP3 inflammasome. Thus, goblet cells have two different protective mechanisms against toxigenic bacteria that threaten the ocular surface. Neither of these protective mechanisms were activated by the commensal or non-toxigenic bacteria strains used. In summary, we conclude that (1) conjunctival goblet cells do not respond to the presence of commensal and non-toxigenic bacteria to activate two protective functions of the goblet cells activation of the NLRP3 inflammasome and stimulation of goblet cell mucous secretion and (2) in contrast NLRP3 plays a critical role as a sensor of pathogenic bacteria in goblet cells of the conjunctiva epithelium, and that activation of the NLRP3 inflammasome mediates secretion of both mature IL-1β and large secretory mucins from human conjunctival goblet cells. MG-K and DAD contributed equally. Contributors: DL perform experiments, analyse data; RRH perform experiments, analyse data, write manuscript; PB perform experiments, analyse data, write manuscript; MSG analyse data, write manuscript; MGK analyse data, write manuscript; DAD analyse data, write manuscript. Funding: This work was supported by NIH RO1 EY022415 Competing interests: None declared. Provenance and peer review: Not commissioned; externally peer reviewed.

Key messages What is already known about this subject? Involving patients in setting the research agenda ensures research benefits those who ultimately live with a condition and prevents research waste. To date, there are no published patient involvement exercises for herpes simplex keratitis (HSK), and as such it has been difficult to ensure patient priorities are being addressed. What are the new findings? This survey would constitute the first published exploration of patients’ priorities for research in HSK. We undertook a patient involvement exercise, conducted in the West Midlands, UK. We found that top research priorities for patients were knowledge of modifiable risk factors for disease recurrence, development of accurate and rapid diagnostic tests, and more understanding of how/when treatment failure occurs. How might these results change the focus of research or clinical practice? Our narrative data give a new insight into patients’ urgent clinical needs, which should be addressed in parallel to research. Our group emphasised the need for better symptom control (during and between flare ups), rapid access to specialist ophthalmic care and high-quality patient information resources.

cal practice? Our narrative data give a new insight into patients’ urgent clinical needs, which should be addressed in parallel to research. Our group emphasised the need for better symptom control (during and between flare ups), rapid access to specialist ophthalmic care and high-quality patient information resources. Introduction Understanding patients’ perspectives is vital for directing the research agenda. Clinicians, academics and the pharmaceutical industry are all key stakeholders in driving research forward, but their priorities are not always aligned with that of patients.1 It has been argued that involving patients in research ensures the benefit to those who ultimately live with the disease and therefore prevents research waste.2 In the UK, organisations such as the James Lind Alliance and INVOLVE (a National Institute for Health Research funded advisory group), have been major driving forces in facilitating public involvement in healthcare research.3 Similarly, the Patient-Centered Outcomes Research Institute in the the United States was set up to support patient-centred research, and to ensure funding is directed at research questions critical to the patient.4

visory group), have been major driving forces in facilitating public involvement in healthcare research.3 Similarly, the Patient-Centered Outcomes Research Institute in the the United States was set up to support patient-centred research, and to ensure funding is directed at research questions critical to the patient.4 Herpes simplex keratitis (HSK) can be a painful and debilitating disease, and when severe, can take a remitting and relapsing course with gradual loss of sight over time.5 The virus is usually acquired early in life, after which it resides in the trigeminal root ganglion in a quiescent state. Years later, the virus travels to the ocular surface via the trigeminal nerve, and has the potential to damage all layers of the cornea.6 Making an initial diagnosis of HSK can be difficult due to non-specific clinical signs, as well as low sensitivity and relatively low uptake of corneal PCR assays and conjunctival swabs.7 Usually, there is a need to start empirical treatment in the absence of confirmatory tests. Repeated infections can accumulate blinding complications such as scarring, neovascularisation, persistent epithelial defects, corneal melt, neurotrophic keratitis and secondary bacterial infection.8–10 The mainstay of treatment is topical antiviral therapy for epithelial disease and/or topical steroids for stromal complications.11 Oral antiviral therapy as long-term prophylaxis has been shown to significantly reduce the risk of recurrences, but there is increasing recognition that HSV resistance can occur in up to one-third of patients on oral antiviral therapy for over a year.12

elial disease and/or topical steroids for stromal complications.11 Oral antiviral therapy as long-term prophylaxis has been shown to significantly reduce the risk of recurrences, but there is increasing recognition that HSV resistance can occur in up to one-third of patients on oral antiviral therapy for over a year.12 There are currently no published studies of the patients’ priorities in research for HSK. The Sight Loss and Vision Priority Setting Partnership produced a list of priorities for research in 12 categories of eye diseases in 2014, however, HSK was not ranked in the list of priorities for corneal and external eye disease.13 Patient-reported outcome measures (PROMs) are an assessment of health status that comes directly from the patient and are increasingly used in clinical effectiveness research, and in health policy and commissioning decisions. There is growing interest in the use of PROMs in ophthalmology, however, none have been developed specifically for HSK. Understanding of the patients perspective in ocular surface disease has focused primarily on dry eye disease (DED), and published work is centred on the development of PROMs for symptom control and quality of life (QoL).14 Assessment tools, such as the Ocular Surface Disease Index15 and Impact of Dry Eye on Everyday Life,16 allow assessment of a patient’s QoL and vision-related functioning, and the National Eye Institute Visual Function Questionnaire-25 (VFQ-25) has shown that the degree of visual impairment confers a worse QoL.17 While DED and HSK share several QoL implications, there is a wide spectrum of HSK-specific consequences (such as fear of relapse, demanding treatment regimens, neurotrophic keratitis and immunosuppressive treatment) that are not addressed by existing tools.

wn that the degree of visual impairment confers a worse QoL.17 While DED and HSK share several QoL implications, there is a wide spectrum of HSK-specific consequences (such as fear of relapse, demanding treatment regimens, neurotrophic keratitis and immunosuppressive treatment) that are not addressed by existing tools. Reynaud et al have published the only QoL study in HSK thus far, focusing on patients during quiescent disease.18 They found levels of QoL impairment in quiescent HSK to be comparable with other sight threatening diseases such as anterior uveitis, cataract, graft-versus-host disease and Sjögren-related DED. However, this study was limited by the lack of a QoL tool specific to HSK, and instead used a combination of The National Eye Institute VFQ-25,19 the Glaucoma QoL20 and the Ocular Surface Disease QoL questionnaire21) to assess the various dimensions of living with HSK. From the literature that is currently available, our understanding of the patient’s perspective in HSK is incomplete. We formed the first HSK patient participation group in the West Midlands, UK with three aims: (1) to recognise patients as stakeholders for research and clinical care in HSK, (2) to provide an opportunity for patients to steer the direction of future research and (3) to understand the patient’s perspective on living with HSK.

ed the first HSK patient participation group in the West Midlands, UK with three aims: (1) to recognise patients as stakeholders for research and clinical care in HSK, (2) to provide an opportunity for patients to steer the direction of future research and (3) to understand the patient’s perspective on living with HSK. Methods The survey took place across four regional hospitals in the West Midlands, UK. Patient participation was enrolled by five corneal specialists at clinic appointments. Patients were recruited sequentially during a 6-month time frame. During consultation, all patients with an established diagnosis of HSK were invited to participate in the survey. A literature search was conducted for publications relating to HSK in the year preceding recruitment of patients (2014). The search strategy was pragmatic in approach and deliberately targeted. Database searches were restricted to PubMed/MEDLINE, with the search term ‘Herpes Simplex Keratitis’ expanded as follows: (‘keratitis, herpetic’[MeSH Terms] OR (‘keratitis’[All Fields] AND ‘herpetic’[All Fields]) OR ‘herpetic keratitis’[All Fields] OR (‘herpes’[All Fields] AND ‘simplex’[All Fields] AND ‘keratitis’[All Fields] OR ‘herpes simplex keratitis’[All Fields]). Date inclusion was: 1 January 2014 to 31 December 2014.

ed as follows: (‘keratitis, herpetic’[MeSH Terms] OR (‘keratitis’[All Fields] AND ‘herpetic’[All Fields]) OR ‘herpetic keratitis’[All Fields] OR (‘herpes’[All Fields] AND ‘simplex’[All Fields] AND ‘keratitis’[All Fields] OR ‘herpes simplex keratitis’[All Fields]). Date inclusion was: 1 January 2014 to 31 December 2014. Publications were grouped based on their clinical relevance and classified into nine key domains (online supplementary file). The number of publications for each domain was calculated as a proportion of all research relating to HSK in that year, to serve as an indicator for the research priorities of the scientific community (table 1). 10.1136/bmjophth-2018-000177.supp1Supplementary data Table 1 Table showing prioritisation of the nine key domains: first from the literature review (52 publications in total) and then from our HSK patient survey Areas of research interest Percentage of publications in each area (Number of publications on PubMed for the year 2014) Areas of research interest by number of publications found on PubMed in 2014 (1=most publications, 7=least publications) Areas of research interest in order of importance as per HSK survey ranking (1=highest priority, 7=lowest priority) 1. Risk factors for recurrence of infection 12% (6) 4 1 2. How quickly the infection can be treated 21% (11) 2 4 3. When there is failure to treat the infection 14% (7) 3 3 4. Developing tests to guide our treatment more effectively 8% (4) 5 2 5. Uncertainties about disease resistance to treatment 5% (3) 7 5 6. The need for long-term treatment 7% (4) 6 8

Areas of research interest Percentage of publications in each area (Number of publications on PubMed for the year 2014) Areas of research interest by number of publications found on PubMed in 2014 (1=most publications, 7=least publications) Areas of research interest in order of importance as per HSK survey ranking (1=highest priority, 7=lowest priority) 1. Risk factors for recurrence of infection 12% (6) 4 1 2. How quickly the infection can be treated 21% (11) 2 4 3. When there is failure to treat the infection 14% (7) 3 3 4. Developing tests to guide our treatment more effectively 8% (4) 5 2 5. Uncertainties about disease resistance to treatment 5% (3) 7 5 6. The need for long-term treatment 7% (4) 6 8 7. Risk factors for developing infection 33% (17) 1 6 8. Impact of the disease on quality of life 0% (0) 8 7 9. The frequency of hospital visits 0% (0) 9 9 All patients in the HSK Patient Participation Group were sent a survey based on these nine domains of research. For each domain, ranking is given by the number of publications in 2014 as an indicator of where research efforts were focused. This is compared with the ranking of research priorities by our HSK Patient Participation Group survey. HSK, herpes simplex keratitis.

All patients in the HSK Patient Participation Group were sent a survey based on these nine domains of research. For each domain, ranking is given by the number of publications in 2014 as an indicator of where research efforts were focused. This is compared with the ranking of research priorities by our HSK Patient Participation Group survey. HSK, herpes simplex keratitis. We distributed the survey online via Survey Monkey (SurveyMonkey, San Mateo, California, USA) for patients who had access to the internet (at https://www.surveymonkey.com; accessed 14 August 2017), and by telephone interview for those who could not access the internet. We asked participants to rank these nine domains of interest in order of importance and share thoughts on each domain. For each patient, the order of the nine domains was randomised automatically by SurveyMonkey. There was also free-text space at the end for patients to comment on anything not covered by the nine domains. 10.1136/bmjophth-2018-000177.supp2Supplementary data Ranking results were given a weighted-average score using SurveyMonkey’s standard formula.22 Thematic analysis was undertaken by XL and GW for all narrative data using the protocol described by Braun and Clarke.23 First, coding of data and identification of potential themes was conducted separately by XL and GW, then key themes were agreed on by consensus. XL, GW and PS defined and named the identified themes and highlighted representative quotes which supported each theme.

ata using the protocol described by Braun and Clarke.23 First, coding of data and identification of potential themes was conducted separately by XL and GW, then key themes were agreed on by consensus. XL, GW and PS defined and named the identified themes and highlighted representative quotes which supported each theme. Results Priorities of the scientific community Review of the literature identified nine areas. These nine areas were reworded in plain English for patients: (1) risk factors for recurrence of infection, (2) how quickly the infection can be treated, (3) when there is failure to treat the infection, (4) developing tests to guide our treatment more effectively, (5) uncertainties about disease resistance to treatment, (6) the need for long-term treatment, (7) risk factors for developing infection, (8) impact of the disease on QoL and (9) frequency of hospital visits. The most researched areas were: how quickly the infection can be treated, risk factors for developing infection and when there is failure to treat the infection (table 1). Priorities of patients Fifty-six patients from five centres in the West Midlands agreed to take part in the survey. Participating patients ranged from mild to severe disease, with varying lengths of diagnosis. Forty per cent of participants were male and 60% were female. The mean age of participants was 57 (range 19–89) years. Socioeconomic background was categorised using Index of Multiple Deprivation (IMD) as measured by postal code. Median IMD of the group was 6 (range 1–10).24

disease, with varying lengths of diagnosis. Forty per cent of participants were male and 60% were female. The mean age of participants was 57 (range 19–89) years. Socioeconomic background was categorised using Index of Multiple Deprivation (IMD) as measured by postal code. Median IMD of the group was 6 (range 1–10).24 Thirty-seven responses to the survey were received, of which six patients completed the questions via telephone. The weighted ranking score of patients for each area of interest is shown in figure 1. The highest rated domains were: risk factors for recurrence of infection (weighted ranking score 6.16), developing tests to guide treatment more effectively (5.57) and when there is failure to treat the infection (5.35). The domain ranked least important was the frequency of hospital visits (3.97). Figure 1 Results of ranking question from HSK patient survey. Weighted-average scores for each domain are shown in order of priority. 9=most important, 1=least important. HSK, herpes simplex keratitis. Thematic analysis Using qualitative research techniques and thematic analysis, we also identified three themes from the narrative data.

Figure 1 Results of ranking question from HSK patient survey. Weighted-average scores for each domain are shown in order of priority. 9=most important, 1=least important. HSK, herpes simplex keratitis. Thematic analysis Using qualitative research techniques and thematic analysis, we also identified three themes from the narrative data. Theme 1: controlling symptoms A prominent theme was difficulty in controlling symptoms (example quotes in table 2). In many cases, this translated to the need for frequent eye-drops. Acute exacerbations require topical antiviral and steroid therapy every few hours. In patients with complicating DED or high intraocular pressure (IOP), there may be further treatment with lubricating or IOP lowering drops. Some patients describe the practical difficulties with frequent eye drops: ‘I am on fourteen drops a day for the last year, it is difficult for me to keep up,’ and some patients are dependent on others to administer their medication: ‘I rely on my wife to put in my drops because my hands don’t work well—I cannot even use a knife and fork.’ Some symptoms persist despite treatment: ‘The eye is never 80% comfortable even when well. Therefore, long-term treatment would be great.’ Table 2 Example quotes from HSK patient survey narrative data Theme Example Quotes

Theme 1: controlling symptoms A prominent theme was difficulty in controlling symptoms (example quotes in table 2). In many cases, this translated to the need for frequent eye-drops. Acute exacerbations require topical antiviral and steroid therapy every few hours. In patients with complicating DED or high intraocular pressure (IOP), there may be further treatment with lubricating or IOP lowering drops. Some patients describe the practical difficulties with frequent eye drops: ‘I am on fourteen drops a day for the last year, it is difficult for me to keep up,’ and some patients are dependent on others to administer their medication: ‘I rely on my wife to put in my drops because my hands don’t work well—I cannot even use a knife and fork.’ Some symptoms persist despite treatment: ‘The eye is never 80% comfortable even when well. Therefore, long-term treatment would be great.’ Table 2 Example quotes from HSK patient survey narrative data Theme Example Quotes Theme 1: controlling symptoms ‘I am on 14 drops a day for the last year, it is difficult to keep up’ ‘I rely on my wife to put in my drops because my hands don’t work well’ ‘No matter what treatment you have, the eye is never 80% comfortable’ ‘Sometimes so painful, can’t tolerate light and causes severe headaches’ ‘When I have had the infection in my eye, immediate treatment helps, but my eye is weak for up to a year afterwards’ ‘I have to take drops every day for the last 4 years, it seems unlikely I will ever be free of them’ ‘Worse bit is putting the drops in’ ‘Need to get quicker pain relief’ ‘Anything which would make treatment more effective especially drops, rather than ointment’

nt helps, but my eye is weak for up to a year afterwards’ ‘I have to take drops every day for the last 4 years, it seems unlikely I will ever be free of them’ ‘Worse bit is putting the drops in’ ‘Need to get quicker pain relief’ ‘Anything which would make treatment more effective especially drops, rather than ointment’ Theme 2: access to the ophthalmologist ‘Early identification for front line non-specialist professionals, for example, Opticians, general practitioners’ ‘Direct access to ophthalmology department without need to involve the general practitioner’ ‘The biggest failure in the system is the inability of general practitioners to recognise it and their misdiagnosis’ ‘Hospital access by request should a worry arise unexpectedly’ ‘Diagnosis needs to be a lot quicker—more specialists need to be assigned’ Theme 3: the need for more information

access to the ophthalmologist ‘Early identification for front line non-specialist professionals, for example, Opticians, general practitioners’ ‘Direct access to ophthalmology department without need to involve the general practitioner’ ‘The biggest failure in the system is the inability of general practitioners to recognise it and their misdiagnosis’ ‘Hospital access by request should a worry arise unexpectedly’ ‘Diagnosis needs to be a lot quicker—more specialists need to be assigned’ Theme 3: the need for more information Questions we have answers to: patient education and modifiable risk factors. ‘How can patients self-identify (recurrence of infection)?’ ‘What are the risk factors and why?’ ‘I’ve never explored or had it explained to me why I got the infection’ ‘Are there any lifestyle factors that could be avoided to prevent recurrence?’ ‘People need to ensure they are not causing or exacerbating risk factors.’ ‘The possibility of resistance to treatment and how to deal with it should be analysed, and the information made available to patients.’ ‘I would like to know how to advise others to prevent them suffering the same infection.’ ‘I think it is also important to support the patient emotionally.’ ‘Is it something you catch or is it something already in your system?’ Questions we are still looking into: Setting the research agenda

Questions we have answers to: patient education and modifiable risk factors. ‘How can patients self-identify (recurrence of infection)?’ ‘What are the risk factors and why?’ ‘I’ve never explored or had it explained to me why I got the infection’ ‘Are there any lifestyle factors that could be avoided to prevent recurrence?’ ‘People need to ensure they are not causing or exacerbating risk factors.’ ‘The possibility of resistance to treatment and how to deal with it should be analysed, and the information made available to patients.’ ‘I would like to know how to advise others to prevent them suffering the same infection.’ ‘I think it is also important to support the patient emotionally.’ ‘Is it something you catch or is it something already in your system?’ Questions we are still looking into: Setting the research agenda ‘Can it go from one eye to the other?’ ‘If long term treatment will reduce a recurrence happening again?’ ‘(what is the) likelihood of complications, the effectiveness at preventing further recurrences and the necessary duration of treatment?’ ‘What are the risks of long term treatment?’ ‘What can be done to improve outcomes where there has been a late diagnosis and damage has been done?’ ‘Why has (failure to treat the infection) taken place?’ ‘How quickly does catastrophic blindness happen?’ ‘Establishing whether prolonged treatment causes resistance and therefore optimal treatment duration.’ Quotes are grouped into three themes: controlling symptoms, access to the ophthalmologist and the need for more information. HSK, herpes simplex keratitis.

‘Can it go from one eye to the other?’ ‘If long term treatment will reduce a recurrence happening again?’ ‘(what is the) likelihood of complications, the effectiveness at preventing further recurrences and the necessary duration of treatment?’ ‘What are the risks of long term treatment?’ ‘What can be done to improve outcomes where there has been a late diagnosis and damage has been done?’ ‘Why has (failure to treat the infection) taken place?’ ‘How quickly does catastrophic blindness happen?’ ‘Establishing whether prolonged treatment causes resistance and therefore optimal treatment duration.’ Quotes are grouped into three themes: controlling symptoms, access to the ophthalmologist and the need for more information. HSK, herpes simplex keratitis. Theme 2: access to the specialist Participants found it difficult to gain rapid access to the specialist (table 2). In the UK, patients may not have direct access to an ophthalmologist, without referral from a family/general practitioner. One patient commented that ‘the time difference between a general practitioner referral and a consultation appointment is important, and pathology of the disease is not always fully understood at primary care level.’ Others pointed out that ‘diagnosis needs to be a lot quicker—more specialists need to be assigned,’ and that HSK ‘needs to be spotted by normal ophthalmologists.’ It seems that some patients are experiencing significant delays to diagnosis and treatment, which has the potential to cause irreversible damage. One patient suggested developing ‘a treatment pack which can be kept on standby by the patient for instant treatment of flare ups,’ and others asked the question ‘is self-diagnosis acceptable?’

patients are experiencing significant delays to diagnosis and treatment, which has the potential to cause irreversible damage. One patient suggested developing ‘a treatment pack which can be kept on standby by the patient for instant treatment of flare ups,’ and others asked the question ‘is self-diagnosis acceptable?’ Theme 3: the need for more information In asking the patient for their priorities, we received many questions in return (table 2). We have categorised these into two groups: questions we have the answer to (which should be made widely available as patient information) and questions without clear answers (which should form the basis of setting the research agenda). Questions ranged from ‘can it go from one eye to the other?’ and ‘what the risk factors are and why?’ to more challenging ones such as ‘why has (failure to treat the infection) taken place?’ and ‘what can be done to improve outcomes when there has been a late diagnosis and damage has been done?’ Some patients pointed out the need for information early on: ‘I’ve never had it explained to me why I got the infection’ and ‘how quickly does catastrophic blindness happen?’ Many questions also centred around lifestyle changes and modifiable risk factors: ‘Are there any lifestyle factors that could be avoided to prevent recurrence?’ as well as ways in which patients can play a more active role in managing their disease: ‘How can patients self-identify (recurrence of infection)?’

s happen?’ Many questions also centred around lifestyle changes and modifiable risk factors: ‘Are there any lifestyle factors that could be avoided to prevent recurrence?’ as well as ways in which patients can play a more active role in managing their disease: ‘How can patients self-identify (recurrence of infection)?’ Discussion With a large clinical and economic impact, significant research efforts are directed towards HSK—including the development of better diagnostic tools, treatment strategies and vaccination.25–28 To our knowledge, this study is the first published report of patient priorities for HSK research. Through this qualitative exercise, we have begun to identify what is most important for the patient. Our literature search highlighted disparities between the research priorities of patients and the scientific community.

To our knowledge, this study is the first published report of patient priorities for HSK research. Through this qualitative exercise, we have begun to identify what is most important for the patient. Our literature search highlighted disparities between the research priorities of patients and the scientific community. The top priorities for research were: risk factors for recurrence of infection; developing tests to guide treatment more effectively and failure to treat the infection. Significant efforts have already been made to understand the viral and host factors influencing infection and reactivation, but this remains poorly understood. We know that certain triggers such as hormonal changes, fever, psychological stress and ultraviolet light exposure may induce reactivation, but the underlying mechanisms remain unclear. Improving diagnosis and monitoring with the use of novel imaging techniques has been an area of rapid growth in recent years and continues to expand. Newer techniques such as anterior segment optical coherence tomography and in vivo confocal microscopy, as well as automated imaging analysis platforms are providing more accurate ways of visualising and quantifying disease.28–30 Treatment failure results in irreversible scarring which may require corneal grafting to preserve vision. The complicating factor in HSK is that the insult of surgery itself may trigger viral reactivation and cause graft failure. Research efforts have focused on understanding the mechanisms of graft rejection and effectiveness of antiviral treatment to reduce the risk of graft failure.31 Surprisingly, frequency of hospital visits was ranked lowest in priority. One patient commented that ‘regular and predictable consultations not only provide reassurance, but a better chance of correct treatment,’ while another asked ‘would less time between visits help catch it earlier?’ It is worth considering whether more routine reviews improve outcomes, or if it is a source of reassurance for patients. This is especially important, as anxiety impairs QoL in patients with HSK even while in remission.18

of correct treatment,’ while another asked ‘would less time between visits help catch it earlier?’ It is worth considering whether more routine reviews improve outcomes, or if it is a source of reassurance for patients. This is especially important, as anxiety impairs QoL in patients with HSK even while in remission.18 An unintended outcome of this study was the extent to which patients also used the survey to draw attention to their priorities for clinical care. They highlighted significant areas of unmet need such as poor access to specialists during times of acute infection. Patients asked if it was acceptable to ‘self-diagnose’ acute infection and use ‘rescue treatment packs’ at home. In chronic obstructive pulmonary disease, National Institute for Health and Care Excellence guidelines recommends the use of rescue packs containing steroids and antibiotics kept at home for acute exacerbations.32 A similar set-up in HSK may benefit patients in acute infection, however, we need strategies that enable them to do so safely. An area requiring attention is patient education. High-quality information provision is an intervention which has been shown to positively impact patients’ experiences and health behaviours. As such, providing accessible information is now firmly embedded in health policy.33 Currently, there is little in the way of published literature and digital resources for HSK. Efforts should concentrate on patient education that is effective, engaging and accessible for the wider population.

es and health behaviours. As such, providing accessible information is now firmly embedded in health policy.33 Currently, there is little in the way of published literature and digital resources for HSK. Efforts should concentrate on patient education that is effective, engaging and accessible for the wider population. This study represents the first open invitation for patients with HSK to express what is most important to them, as well as the first examination of whether current research is aligned with the priorities of patients. Despite being a leading cause of infectious corneal blindness, HSK did not feature in the top priorities ranked in the corneal and external diseases category of the Sight Loss and Vision Priority Setting Partnership Survey. From the disparity of research priorities demonstrated by our study, we feel that a more in-depth priority setting exercise for HSK patients would be beneficial in directing future research.

priorities ranked in the corneal and external diseases category of the Sight Loss and Vision Priority Setting Partnership Survey. From the disparity of research priorities demonstrated by our study, we feel that a more in-depth priority setting exercise for HSK patients would be beneficial in directing future research. Our study has several limitations. Our initial scope of the literature which informed the 9 domains of research interest only included studies in the year preceding our survey. This was not intended to be an exhaustive review of the literature, but rather an indication of where research efforts within HSK were focused at the time. Nineteen patients who originally agreed to take part did not complete the survey and it is unclear whether the drop-out rate may have introduced bias to our findings. For this initial qualitative scoping exercise, we have not collected baseline characteristics of participants, therefore, further evaluation is required to determine the needs and priorities for different patient groups. Our study suggests that elderly patients have concerns specific to coping with the demands of frequent drops. With an ageing population, it is vital that clinicians consider the elderly patient’s physical and cognitive limitations, and the feasibility of their management plans.34 Furthermore, centre and patient variation in treatment regimen may exist, including, for example, the need to treat concurrent elevation in IOP, which may have exacerbated the need for drop frequency. More in-depth interviews among different patient groups are required to explore the full spectrum of patient priorities. Our geographical reach was limited to the West Midlands, UK, where guidelines and structure of service provision differ from other areas. It is, however, an area of ethnic and socioeconomic diversity, with the largest non-white regional population outside of London.35 Further investigation encompassing the wider cohort is needed.

hical reach was limited to the West Midlands, UK, where guidelines and structure of service provision differ from other areas. It is, however, an area of ethnic and socioeconomic diversity, with the largest non-white regional population outside of London.35 Further investigation encompassing the wider cohort is needed. This study has highlighted three issues in our current approach to HSK. First, the research agenda has so far been set without published knowledge on what is important for the patient. This can lead to a disparity between priorities of patients and the scientific community. We have identified several areas of key importance to patients for future research to bridge this gap. Second, there is a lack of easily accessible patient information on HSK. This hampers their ability to make informed decisions relating to their own clinical care and limits their ability to contribute towards research. Third, clinical service priorities are of equal importance to patients as research. Clinicians and researchers should be aware that a patient’s urgent desire for better clinical care today may outweigh the uncertain benefits of research tomorrow, even if the latter might lead to a cure. To understand patients’ true priorities, the enquiry must be open and unbiased.

importance to patients as research. Clinicians and researchers should be aware that a patient’s urgent desire for better clinical care today may outweigh the uncertain benefits of research tomorrow, even if the latter might lead to a cure. To understand patients’ true priorities, the enquiry must be open and unbiased. Our survey is the first attempt to engage patients with HSK in a dialogue about setting research priorities. Patients are eager to fulfil their role as key stakeholders for research, therefore, it is vital that patients are provided with a platform to voice their priorities and establish partnerships with researchers. We intend to expand our patient participation group for this purpose and continue to share our findings with the scientific and clinical community. Contributors: XL and GW: project conceptualisation, methodology, design, patient recruitment, data analysis and manuscript drafting. SK, PM, AP and MQ: are responsible for clinical care of participants, patient recruitment and manuscript review. KS: patient representative, analysis of results and manuscript review. AKD and PS: study design, providing expertise in qualitative analysis, review of manuscript. Funding: XL and AKD receive funding from the Wellcome Trust (grant number 200141/Z/15/Z). Competing interests: None declared. Patient consent for publication: Not required. Ethics approval: The survey design was agreed by the National Institute of Health Research (NIHR) Research Design Service West Midlands and approved by individual hospital’s Research and Development departments.

Funding: XL and AKD receive funding from the Wellcome Trust (grant number 200141/Z/15/Z). Competing interests: None declared. Patient consent for publication: Not required. Ethics approval: The survey design was agreed by the National Institute of Health Research (NIHR) Research Design Service West Midlands and approved by individual hospital’s Research and Development departments. Provenance and peer review: Not commissioned; externally peer reviewed. Author note: This paper is dedicated to the memory of our wonderful colleague, Vinette Cross, who recently passed away.

Current knowledge about the eye microbiome The ocular surface (OS) microbiome is an understudied topic, compared with other host-associated environments. While the Human Microbiome Project initially studied five main body areas—the skin, the gastrointestinal tract, the urogenital tract, the oral and the nasal mucosa1—an emerging area of research is focusing on the eye and the microbiota of the OS.2

an understudied topic, compared with other host-associated environments. While the Human Microbiome Project initially studied five main body areas—the skin, the gastrointestinal tract, the urogenital tract, the oral and the nasal mucosa1—an emerging area of research is focusing on the eye and the microbiota of the OS.2 Recent studies demonstrated that OS hosts a number of commensal microorganisms.3 Earlier culture-based surveys suggested that the OS are colonised by microbial communities dominated by Gram-positive Firmicutes, in particular, species belonging to the Staphylococcus, Streptococcus, Corynebacterium and Propionibacterium.4 A screening including approximately 1000 16S rRNA reads revealed that the diversity of healthy conjunctiva was higher than previously thought.5 Other recent studies based on traditional microbiological techniques have examined the microbiota of the OS,6–11 although a more comprehensive analysis of microbial diversity of OS has been hindered by the limitations of conventional cultivation techniques.12–14 More recent screening of OS-associated microbiome, using molecular metagenomic techniques, extended further the knowledge about OS microbial diversity.2 5 15 16 Shestopalov and colleagues estimated using real-time PCR that in 1 ng of extracted DNA, the number of bacterial genomes (ie, bacterial richness) was on average 79.8 and 729 in the conjunctiva and cornea, respectively. Significant amounts (22 over 55) were detected in the eye for the first time.17 Dong et al detected 59 distinct bacterial genera using a 16S rDNA gene pyrosequencing approach on the OS of four healthy individuals15 (figure 1). Despite the low number of individuals examined, this is one of the first studies focusing on the bacterial diversity of the OS microbiome. Healthy OS microbiome is dominated by Proteobacteria, Actinobacteria and Firmicutes. The most common taxa at the genus level were P seudomonas, Propionibacterium, Bradyrhizobium, Corynebacterium, Acinetobacter, Brevundimonas, Staphylococci, Aquabacterium, Sphingomonas, Streptococcus, Streptophyta and Methylobacterium (figure 1). This is in general agreement with the previous studies, although many false positives may derive from contamination.4

e P seudomonas, Propionibacterium, Bradyrhizobium, Corynebacterium, Acinetobacter, Brevundimonas, Staphylococci, Aquabacterium, Sphingomonas, Streptococcus, Streptophyta and Methylobacterium (figure 1). This is in general agreement with the previous studies, although many false positives may derive from contamination.4 Figure 1 Pie charts displaying the relative proportions of taxa at the phylum level and at the genus level in healthy eye microbiome (according to Dong et al 15 [15]). Microbial invasion into the OS compromises corneal clarity and causes inflammation in blinding conditions like keratitis, endophthalmitis and trachoma. Herpes simplex type 1, hepatitis B and C viruses can be detected using PCR in the tears of asymptomatic volunteers,18 thus suggesting that simply focusing solely on the bacterial constituents of the OS may result in an incomplete understanding of the OS microbiome. In recent time, Zhou et al 19 showed how the changes in the conjunctival microbiome occur in trachomatous disease compared with normal controls. Wen et al 20 showed how the microbiome of healthy OS is shaped by age and sex and how the ocular microbiome of house finches changed during experimentally induced mycoplasmal conjunctivitis.21 Study of microbiome in OS disease has significant potential to improve the diagnosis, treatment and management of potential blinding diseases.

Microbial invasion into the OS compromises corneal clarity and causes inflammation in blinding conditions like keratitis, endophthalmitis and trachoma. Herpes simplex type 1, hepatitis B and C viruses can be detected using PCR in the tears of asymptomatic volunteers,18 thus suggesting that simply focusing solely on the bacterial constituents of the OS may result in an incomplete understanding of the OS microbiome. In recent time, Zhou et al 19 showed how the changes in the conjunctival microbiome occur in trachomatous disease compared with normal controls. Wen et al 20 showed how the microbiome of healthy OS is shaped by age and sex and how the ocular microbiome of house finches changed during experimentally induced mycoplasmal conjunctivitis.21 Study of microbiome in OS disease has significant potential to improve the diagnosis, treatment and management of potential blinding diseases. Background and evolution of metagenomics The non-culturability of microbes dates back to 1898 by Heinrich Winterberg, formalised as the ‘great plate count anomaly’ by Staley and Konopka in 1985.22 Several theories have been proposed to explain the non-culturability of microbes: (1) the cultural media used have been developed to grow microorganisms usually involved in human or animal diseases, missing a diversity of other microorganisms; (2) some bacteria may require long incubation time to form visible colonies; (3) physiological features (ie, obligate symbiosis with other species or strong quorum sensing signalling processes) may hinder the cultivation of some species; and (4) some bacteria may require a specific combination of nutritional features and aerobic requirements. The pioneering work by Woese23 in 1985 identified the 16S rRNA gene as an evolutionary chronometer for bacterial phylogeny. This gene has three unique features: (A) the ubiquity in the bacterial kingdom, (B) the structure of the gene itself, made of both variable and conserved regions (this is due to the secondary structure of the transcribed RNA, made of stretches and loops), and (C) the low (if any) amount of horizontal gene transfer.

ny. This gene has three unique features: (A) the ubiquity in the bacterial kingdom, (B) the structure of the gene itself, made of both variable and conserved regions (this is due to the secondary structure of the transcribed RNA, made of stretches and loops), and (C) the low (if any) amount of horizontal gene transfer. Pace et al in 1985 had the idea that 16S rRNA gene isolated from the environmental samples can directly be cloned.24 In 1991, Schmidt et al 25 successfully cloned 16S rRNA gene sequences from marine picoplankton with the use of bacteriophage lambda vector.26 In 1998, the term metagenome was introduced.27 In the last two decades metagenomic analyses have been performed on majority of the natural environments, for example, soils,28 marine picoplankton,29 hot springs,30 surface water from rivers,31 glacier ice,32 Antarctic desert soil33 and gut of ruminants.26 34 Earlier, major parts of the metagenomic studies were based on low-throughput approaches, like terminal restriction fragment-length polymorphism analysis,35 denaturing gradient gel electrophoresis36 or Sanger sequencing.37 The comparison with the sequences included in curated databases like the ones of Ribosomal Database Project II,38 Greengenes39 and SILVA40 has allowed the taxonomic classification and community profiling of environmental 16S rRNA gene sequences.

lysis,35 denaturing gradient gel electrophoresis36 or Sanger sequencing.37 The comparison with the sequences included in curated databases like the ones of Ribosomal Database Project II,38 Greengenes39 and SILVA40 has allowed the taxonomic classification and community profiling of environmental 16S rRNA gene sequences. Clinical metagenomics and its potential as diagnostic tool Classical microbiological methods are able to identify only the presumed cause of ocular infection in about 40% cases. In contrast, a metagenomic approach promises to provide important detail regarding all microbiota, allowing the identification of a greater portion of previously unidentified and the so-called ‘uncultured majority’ of microorganisms,41 whether being prokaryotes, eukaryotes or viruses. Efficient next-generation sequencing (NGS) technologies have developed greatly in the last decade, along with a reduced cost, gaining interest in the scientific community in several fields (such as medicine, biotechnology, agriculture or genetics).42 The deep impact on metagenomics given by the NGS technologies allowed the study of microbes with a much higher throughput. This enabled a ‘hypothesis-free’ approach providing all the components of the microbiome. The data produced by NGS have also proved to be particularly suitable for taxonomic and functional profiling. Thus, metagenomics opened the way to explore the remarkable genetic potential of bacteria; in some cases it was used to get hints on the metabolic requirements of yet-to-be cultured bacteria.43 More often, the application of NGS to metagenomics studies aimed to profile and compare taxonomy and function of microbial communities from different sources.23 For example, comparing human samples from a disease state to samples from healthy controls allows clinicians to get a more holistic view of the quantitative shifts of specific taxa during the course of the disease.1 44 The extraction of DNA from a human-derived sample is usually the first step of the workflow, then two possible downstream analyses can be done: (1) amplify a marker gene (usually a portion of the rRNA genes) and sequence the PCR amplification product; or (2) sequencing the extracted, fragmented DNA directly (this approach is known as ‘shotgun’ metagenomics sequencing).

s usually the first step of the workflow, then two possible downstream analyses can be done: (1) amplify a marker gene (usually a portion of the rRNA genes) and sequence the PCR amplification product; or (2) sequencing the extracted, fragmented DNA directly (this approach is known as ‘shotgun’ metagenomics sequencing). There is a wealth of studies on the human microbiome, spanning several physiological conditions related to age, disease, race and many more variables. Starting from the first ‘inocula’, that is, transmission from mothers to children.45 A significant role has been attributed to the birth delivery method, with marked differences in microbiota between children born through vaginal delivery versus caesarean section.46 Different studies have debated determining the ‘core’ bacterial taxa of gut and skin.1 47 The latest analysis suggests that instead of the core taxa, homeostatic communities are defined by the presence of a core microbial gene set that encodes essential metabolic pathways.12 48

inal delivery versus caesarean section.46 Different studies have debated determining the ‘core’ bacterial taxa of gut and skin.1 47 The latest analysis suggests that instead of the core taxa, homeostatic communities are defined by the presence of a core microbial gene set that encodes essential metabolic pathways.12 48 In addition, several studies have highlighted the variability of the microbiome according to body sites, race and ethnicity as outlined by Gupta et al.49 The genetic asset of the host itself influences the microbiome; this is well studied in metabolic dysfunctions such as obesity, diabetes and inflammatory bowel disease.50 Two simultaneous projects: the European project, MetaHIT (Metagenomics of the Human Intestinal Tract—www.metahit.eu), and the American Human Microbiome Project,51 use metagenomics to facilitate the study of human intestinal microbiome. The use of drugs has also an influence on the microbiome. Bioinformatic tools are being developed and updated almost monthly for the analysis of the data52 53 and in developing integrated databases, for example, https://portal.hmpdacc.org/. They are, however, over-represented by strong biases towards samples from stool, and oral and vaginal microbiomes.

nce on the microbiome. Bioinformatic tools are being developed and updated almost monthly for the analysis of the data52 53 and in developing integrated databases, for example, https://portal.hmpdacc.org/. They are, however, over-represented by strong biases towards samples from stool, and oral and vaginal microbiomes. Selecting the test The two methods (shotgun and marker-based metagenomics) can be used in different instances: the marker-based approach is used to get the taxonomic profiles of the community under study, whereas shotgun approach gives wider information on function and an extended phylogenetic breadth.54 For both methods, there are pros and cons: marker-based studies are well suited for analysis of large number of samples, that is, multiple patients, longitudinal studies, and so on, and are cheaper; however, there are well-known amplification biases and the amount of information is limited to the taxonomy.55 On the other hand, shotgun metagenomics is usually more expensive. It may miss low-abundant species and when host-associated metagenomes are studied, most of the reads derive from the host genome, especially when studying sites with low bacterial biomass. It offers, however, increased resolution, enabling the possibility to discover new microbial genes and genomes as well as a more specific taxonomic and functional classification of sequences (in some cases). Importantly, shotgun metagenomics allows the simultaneous study of viruses, bacteriophages, archaea and eukaryotes.56 Sample collection and storage methods are critical for most metagenomic studies: they are often arbitrary and rely on the common practices developed in single laboratories or even by single researchers.57 However, in some cases, such as the study of the human faecal microbiome, there are well-established standard procedures.58

ollection and storage methods are critical for most metagenomic studies: they are often arbitrary and rely on the common practices developed in single laboratories or even by single researchers.57 However, in some cases, such as the study of the human faecal microbiome, there are well-established standard procedures.58 A standardised protocol for sample collection, handling and storage for metagenomic studies in ophthalmology is still under development (data not shown). In addition, as all low biomass samples, corneal surfaces are particularly vulnerable to external contaminations, which could also derive from the reagent kits,59 therefore, a proper experimental design should include a number of blank controls and the use of ultrapure reagents to minimise this risk. Several significant efforts to unravel bacterial identity with a resolution as high as the level of strain have already been published.60 The integration of the metaomics (collective name that stands for metagenomics, metatranscriptomics, metaproteomics, and so on) with information such as clinical history, dietary information and genetic background of the patient may be useful in the implementation of mechanistic models explaining the microbiome structure and function.61 Biomarker discovery needs a high number of replicates; one pipeline developed for this task is LEfSe62 which relies on the linear discriminant analysis of effect size. It detects consistent abundance patterns among features (that can be either taxa or coding genes) in a multidimensional data set such as a species-per-sample metagenomic matrix. It is highly scalable and it has proved to achieve a discrete performance in reducing the false-positive detection, although as explicitly admitted by the developer, the false-negative rate is slightly higher. Other pipelines are also available for biomarker discoveries,63 however, a benchmark among them is beyond the scope of this review. Last, but not least, the complex tasks described above require high computational power and specific expertise in the field of biostatistics and informatics.64

ive rate is slightly higher. Other pipelines are also available for biomarker discoveries,63 however, a benchmark among them is beyond the scope of this review. Last, but not least, the complex tasks described above require high computational power and specific expertise in the field of biostatistics and informatics.64 From bench to bedside: clinical applications Clinical applications in OS Doan and Pinsky in 2016 showed how the healthy OS has a unique microbiome with viral and bacterial communities. In their study, quantitative 16S PCR resulted in 0.1 bacterial 16S rDNA/human actin copy on the OS compared with 10 16S rDNA/human actin copies of facial skin and higher bacterial diversity on the OS.65 Many ocular infections are acquired on the OS and the diagnosis of causative pathogen can be challenging.66 67 In postoperative endophthalmitis, the pathogen identification is between 50% and 70%68 69 and cultures failed in approximately 33% of cases.70 In a recent study on patients with uveitis, Doan et al found that herpes simplex virus type 1 (HSV-1), Cryptococcus neoformans and Toxoplasma gondii were associated with the disease.71 Other than Rubeola RNA virus, even Ebola RNA virus was detected in the ocular fluid after resolution of viraemia.72 A recent attempt provided a proof of concept for the use of metagenomics as diagnostic tool, and developed specific bioinformatic pipelines to differentiate pathogenic agents and antibiotic resistance genes with a higher resolution.73

even Ebola RNA virus was detected in the ocular fluid after resolution of viraemia.72 A recent attempt provided a proof of concept for the use of metagenomics as diagnostic tool, and developed specific bioinformatic pipelines to differentiate pathogenic agents and antibiotic resistance genes with a higher resolution.73 Contact lens Shin et al showed that wearing contact lenses makes ocular conjunctiva more similar to the skin microbiota.16 Lee et al 74 studied blepharitis using different sampling from eyelashes and tears showing increased Staphylococcus, Streptophyta, Corynebacterium and Enhydrobacter. Studies suggested how contact lenses could function as a medium for skin bacteria to come to OS.75 76 Zhang et al found how slight microbe variability was found between orthokeratology lens wearers with soft contact lenses wearers and in non-wearers.77

ed Staphylococcus, Streptophyta, Corynebacterium and Enhydrobacter. Studies suggested how contact lenses could function as a medium for skin bacteria to come to OS.75 76 Zhang et al found how slight microbe variability was found between orthokeratology lens wearers with soft contact lenses wearers and in non-wearers.77 Connection between gut and eye microbiome It is well acknowledged that gut microbiome influences the communication between the enteric nervous system and the central nervous system (known as gut-brain axis).78 Likewise, OS microbiome shows connection to the gut microbiome. Considering that the eye is the site of inflammatory diseases like uveitis, scleritis and Mooren’s corneal ulcer, it is possible that these autoimmune reactions are associated with dysbiosis in the gut.79 de Paiva et al found that ‘the severity of Sjögren Syndrome (SS) ocular and systemic disease was inversely correlated with microbial diversity.’80 SS is marked by a dysbiotic intestinal microbiome driven by low relative abundance of commensal bacteria and high relative abundance of potential pathogenic genera. This is associated with worse ocular mucosal disease in a mouse model of SS and in patients with SS. The lowest diversity of stool microbiota was found in subjects with the most severe keratoconjunctivitis sicca and combined systemic and ocular disease. Such result is in agreement with other findings in which a disease state correlates with the low diversity of the microbiome in a specific compartment, such as inflammatory bowel disease (where Clostridium difficile dominates the microbiome) or the pulmonary microbiome during cystic fibrosis.81 Animal models of experimental autoimmune uveitis have shown significant attenuation of this disease following administration of oral antibiotics that altered the intestinal microbiota.82 Further evidence strongly suggests that the homeostatic microbiome plays a protective role in preventing colonisation of pathogenic species. It was demonstrated that oral administration of antibiotics reduced the severity of uveitis in mice with experimentally induced autoimmune uveitis.83 84 The recent cases of persistent infection with Ebola virus,72 and possibly Zika virus,85 explain the urgency to develop better diagnostics for uveitis. These cases, with important public health consequences, highlight the eye’s role as a potential reservoir for infectious agents.

entally induced autoimmune uveitis.83 84 The recent cases of persistent infection with Ebola virus,72 and possibly Zika virus,85 explain the urgency to develop better diagnostics for uveitis. These cases, with important public health consequences, highlight the eye’s role as a potential reservoir for infectious agents. In addition, epigenetic mechanisms may cooperate with microbiota to initiate ocular inflammation.86

entally induced autoimmune uveitis.83 84 The recent cases of persistent infection with Ebola virus,72 and possibly Zika virus,85 explain the urgency to develop better diagnostics for uveitis. These cases, with important public health consequences, highlight the eye’s role as a potential reservoir for infectious agents. In addition, epigenetic mechanisms may cooperate with microbiota to initiate ocular inflammation.86 Discussions and conclusion Detection of intraocular infections relies heavily on molecular diagnostics. In ophthalmology, the most widely available molecular diagnostic panel to detect infections includes separate pathogen-directed PCRs: HSV, varicella zoster virus, cytomegalovirus and T. gondii. Not surprisingly, more than 50% of all presumed corneal infections fail to have a pathogen isolated.87 88 NGS has offered clear advantages to make a definite diagnosis compared with conventional diagnostic methods and advances in metagenomics have made the use of NGS more useful for clinical disgnostics.71 Metagenomic deep sequencing (MDS) has the potential to improve diagnostic yield; it can theoretically detect all pathogens in a clinical sample89 using an unbiased and hypothesis-free approach. Wilson et al showed that NGS protocols could be completed in less than 48 hours90 with a significant advantage compared with many of the culture-dependent assays. Improving our understanding of the composition and function of a normal ocular microbiome would be a good starting point for a targeted therapy and the development of probiotic products. Host-microbe and microbe-microbe interactions on the OS indicate the beneficial function of the microbiota and the understanding of these principles by the clinicians could possibly guide appropriate use of topical and systemic antibiotics.91

rting point for a targeted therapy and the development of probiotic products. Host-microbe and microbe-microbe interactions on the OS indicate the beneficial function of the microbiota and the understanding of these principles by the clinicians could possibly guide appropriate use of topical and systemic antibiotics.91 MDS could be used in patients with difficult-to-diagnose infections, as a front-line diagnostic tool. In difficult-to-culture samples, metagenomic shotgun is a promising test for the identification of microbial keratitis and undiagnosed encephalitis.92 93 Together with uveitis-related disorders, the microbiome study can potentially and dramatically change the management and treatment of these diseases.

diagnostic tool. In difficult-to-culture samples, metagenomic shotgun is a promising test for the identification of microbial keratitis and undiagnosed encephalitis.92 93 Together with uveitis-related disorders, the microbiome study can potentially and dramatically change the management and treatment of these diseases. MDS may supplement or replace numerous and expensive diagnostic assays and procedures currently employed and improve patient outcomes. This approach will allow a far more comprehensive characterisation of the aetiology of infections and also complement the current diagnostic paradigm in ophthalmology. In the near future, a full genetic approach to eye infections is not far away (figure 2) where the samples will be collected, sequenced for specific targets and a specific antimicrobial used to treat the disease. With a certain clinical impact, the ophthalmologists will have precise quantification, multilocus sequence typing of single species and genetic sequence of the microbiota of ocular samples without the problems associated with tests based on conventional cultures. Hence, we foresee that metagenomics could further advance the field of ophthalmology especially in diagnosis and target-specific treatments.

n, multilocus sequence typing of single species and genetic sequence of the microbiota of ocular samples without the problems associated with tests based on conventional cultures. Hence, we foresee that metagenomics could further advance the field of ophthalmology especially in diagnosis and target-specific treatments. Figure 2 The process towards genetic-guided treatment in the field of ophthalmology. The figure indicates different procedures of sample collection, nucleic acid extraction, sample preparation, sequencing, bioinformatics, analysis and report writing, indication to the eye surgeon for specific drug usage for specific microorganism and metagenomic-guided eye treatment on the patient. Being highly specific and cost-effective, metagenomics could be potentially used in ophthalmology in the near future. Contributors: Concept and design of the review: DB, VR, SBK, TS, LN, AF, PG, DP, AE, SF. Drafting the manuscript: DB, VR, SBK, AE, SF. Critical revision of the manuscript: DB, VR, SBK, TS, LN, AF, PG, DP, AE, SF. Supervision: SBK, SF, VR. Final approval: DB, VR, SBK, TS, LN, AF, PG, DP, AE, SF. Competing interests: None declared. Patient consent: Not required. Provenance and peer review: Not commissioned; externally peer reviewed.

Key messages What is already known about this subject? The pathobiology of ocular surface complications in Stevens-Johnson syndrome (SJS) with severe ocular complications has been not well known, so we performed comprehensive gene expression analysis of the conjunctival epithelium, to further investigate about it. What are the new findings? We found that the downregulation of PIGR, HEPACAM2 and ADH1C and upregulation of teneurin-2 expression contribute to the pathobiology of the ocular surface in patients with SJS in the chronic stage. How might these results change the focus of research or clinical practice? Our present finding might contribute to resolve the mechanism of ocular complications of SJS.

What are the new findings? We found that the downregulation of PIGR, HEPACAM2 and ADH1C and upregulation of teneurin-2 expression contribute to the pathobiology of the ocular surface in patients with SJS in the chronic stage. How might these results change the focus of research or clinical practice? Our present finding might contribute to resolve the mechanism of ocular complications of SJS. Introduction Stevens-Johnson syndrome (SJS) is an acute inflammatory vesiculobullous reaction of the skin and mucosa such as the ocular surface, oral cavity and genitals. In patients with extensive skin detachment and a poor prognosis, the condition is called toxic epidermal necrolysis (TEN). Severe ocular complications (SOC) appear in about half of patients with SJS/TEN diagnosed by dermatologists.1 Cold medicines, including multi-ingredient cold medications and non-steroidal anti-inflammatory drugs (NSAID), were the main causative drugs of SJS/TEN with SOC in all patients with SJS and TEN. In fact, about 80% of our patients developed SJS/TEN with SOC after taking cold medicines a few to several days before disease onset.2–5 In the acute stage, patients with SJS/TEN with SOC manifest severe conjunctivitis with corneal and conjunctival erosion and a pseudomembrane in addition to skin eruption and erosion. Despite healing of the skin lesions, in the chronic stage of SJS/TEN with SOC, ocular surface inflammation persists6 as do ocular surface complications including conjunctival invasion into the cornea, symblepharon, ankyloblepharon and dry eye.3 7 It is not easy for ophthalmologists to render a differential diagnosis of SJS or TEN when patients present in the chronic stage because the vesiculobullous skin lesion expressed in the acute stage has healed by the chronic stage. Diagnosis of SJS/TEN in ophthalmology was based on a confirmed history of acute-onset high fever, serious mucocutaneous illness with skin eruptions and involvement of at least two mucosal sites including the ocular surface.2 4 5 8–10 Thus, ophthalmologists tend to report both SJS and TEN with SOC as ‘SJS’ in a broad sense.3

osis of SJS/TEN in ophthalmology was based on a confirmed history of acute-onset high fever, serious mucocutaneous illness with skin eruptions and involvement of at least two mucosal sites including the ocular surface.2 4 5 8–10 Thus, ophthalmologists tend to report both SJS and TEN with SOC as ‘SJS’ in a broad sense.3 Conjunctival invasion into the cornea results in severe visual disturbance in patients with SJS with SOC in the chronic stage. Immunohistological differences in the characteristics of the conjunctival epithelium of patients with SJS with SOC and controls have been documented. We reported that EP311 and EP412 were markedly downregulated in the conjunctival epithelium of patients with SJS with SOC. To further investigate the pathology of ocular surface complications in SJS with SOC in the chronic stage, we performed comprehensive gene expression analysis of the conjunctival epithelium using GeneChip oligonucleotide microarrays (Affymetrix, Santa Clara, CA). We confirmed the downregulation and upregulation of transcripts of interest by quantitative real-time PCR (RT-PCR) assay and assessed the expression of protein with significant transcript upregulation by immunohistochemical methods.

al epithelium using GeneChip oligonucleotide microarrays (Affymetrix, Santa Clara, CA). We confirmed the downregulation and upregulation of transcripts of interest by quantitative real-time PCR (RT-PCR) assay and assessed the expression of protein with significant transcript upregulation by immunohistochemical methods. Materials and methods Patient and public involvement This study was approved by the Institutional Review Board of Kyoto Prefectural University of Medicine. All experimental procedures were conducted in accordance with the tenets of the Declaration of Helsinki. Written informed consent was obtained from all patients after they were given a detailed explanation of the purpose of the research and the experimental protocols. Conjunctival tissue was obtained during ocular surface reconstruction or conjunctivochalasis surgery from the patients with their written informed consents, but patients did not involve in the recruitment to and conduct of the study. Our results will disseminate to study participants by paper and our web pages. There is no patient adviser in this study. Human conjunctival epithelium For GeneChip analysis and quantitative RT-PCR, in vivo human conjunctival epithelium was harvested from conjunctival tissue obtained during ocular surface reconstruction or conjunctivochalasis surgery by overnight immersion at 4°C in 1.0 U/mL purified dispase (Roche Diagnostic, Basel, Switzerland).13

nctival epithelium For GeneChip analysis and quantitative RT-PCR, in vivo human conjunctival epithelium was harvested from conjunctival tissue obtained during ocular surface reconstruction or conjunctivochalasis surgery by overnight immersion at 4°C in 1.0 U/mL purified dispase (Roche Diagnostic, Basel, Switzerland).13 Gene expression analysis Gene expression profiles were investigated using a high-density oligonucleotide probe array (GeneChip, Human Gene 1.0 ST array (Affymetrix)). Total RNA was extracted with the Qiagen RNeasy Kit (Qiagen, Valencia, CA). We used approximately 764 885 probe sets covering more than 28 869 genes. Throughout the process we followed Affymetrix instructions. Scanned microarray images were obtained on a GeneChip Scanner 3000 7G (Affymetrix) using the default settings. Images were visually inspected to detect hybridisation artefacts. Quantitative RT-PCR Total RNA was isolated using the RNeasy Mini Kit according to the manufacturer’s instructions. For the RT reaction we used ReverTra Ace (TOYOBO, Japan). Quantitative RT-PCR assays were performed on a StepOnePlus instrument (Applied Biosystems) according to the manufacturer’s instructions. The primers and probes were purchased from Applied Biosystems. Quantification data were normalised to the expression of the housekeeping gene GAPDH. Immunohistochemistry Human conjunctival tissue samples were from patients with SJS/TEN undergoing ocular surface reconstruction. The controls were nearly normal conjunctival tissue samples from operated patients with conjunctivochalasis.11–14

Quantitative RT-PCR assays were performed on a StepOnePlus instrument (Applied Biosystems) according to the manufacturer’s instructions. The primers and probes were purchased from Applied Biosystems. Quantification data were normalised to the expression of the housekeeping gene GAPDH. Immunohistochemistry Human conjunctival tissue samples were from patients with SJS/TEN undergoing ocular surface reconstruction. The controls were nearly normal conjunctival tissue samples from operated patients with conjunctivochalasis.11–14 Conjunctival tissue sections mounted on slides were fixed for 10 min at 4°C with 4% PFA/0.1M phosphate-buffered saline (PBS), incubated overnight in a moist chamber at 4°C with sheep anti-human teneurin-2 polyclonal antibody (AF4578; R&D, MN, USA) or isotype control sheep IgG (R&D), and then washed in PBS without Ca and Mg [PBS(−)]. Alexa Fluor 488 donkey anti-sheep IgG (H+L) (Molecular Probes, Eugene, OR) was applied for 1 hour at room temperature. After washing the slides, antifade mounting medium with 4′,6-diamidino-2-phenylindole was applied (Vectashield; Vector Laboratories, Burlingame, CA). Data analysis For microarray analysis we used the analysis of variance (ANOVA) p value to record significant differences between patients with SJS and the controls. Data from quantitative RT-PCR assays were expressed as the mean±SE and evaluated by the Student’s t-test using Microsoft Excel.

Conjunctival tissue sections mounted on slides were fixed for 10 min at 4°C with 4% PFA/0.1M phosphate-buffered saline (PBS), incubated overnight in a moist chamber at 4°C with sheep anti-human teneurin-2 polyclonal antibody (AF4578; R&D, MN, USA) or isotype control sheep IgG (R&D), and then washed in PBS without Ca and Mg [PBS(−)]. Alexa Fluor 488 donkey anti-sheep IgG (H+L) (Molecular Probes, Eugene, OR) was applied for 1 hour at room temperature. After washing the slides, antifade mounting medium with 4′,6-diamidino-2-phenylindole was applied (Vectashield; Vector Laboratories, Burlingame, CA). Data analysis For microarray analysis we used the analysis of variance (ANOVA) p value to record significant differences between patients with SJS and the controls. Data from quantitative RT-PCR assays were expressed as the mean±SE and evaluated by the Student’s t-test using Microsoft Excel. Results GeneChip analysis: comparison of transcripts downregulated by less than one-tenth and upregulated more than 50-fold in conjunctival epithelium from patients with SJS and the controls We subjected conjunctival epithelium from three patients with SJS with SOC and three patients with conjunctival chalasis to gene expression analysis by microarray. We found that 49 transcripts were downregulated by less than one-tenth and that there was a significant difference between the patients with SJS and the controls (ANOVA p<0.05) (online supplementary table 1a). Transcripts downregulated by less than one-fiftieth were MUC7, PIGR, HEPACAM2, ADH1C and SMR3A (table 1a). As shown in online supplementary table 1b, 62 transcripts were upregulated more than 10-fold; there was a significant difference between samples from the patients with SJS and the controls (ANOVA p<0.05). The 14 transcripts upregulated more than 50-fold were SERPINB4, KRT1, KRTDAP, S100A7, SBSN, KLK6, SERPINB12, PNLIPRP3, CASP14, ODZ2, CA2, CRCT1, CWH43 and FLG (table 1b); 6 of these 14 transcripts were related to cell differentiation or epidermis development in biological aspect.

th SJS and the controls (ANOVA p<0.05). The 14 transcripts upregulated more than 50-fold were SERPINB4, KRT1, KRTDAP, S100A7, SBSN, KLK6, SERPINB12, PNLIPRP3, CASP14, ODZ2, CA2, CRCT1, CWH43 and FLG (table 1b); 6 of these 14 transcripts were related to cell differentiation or epidermis development in biological aspect. 10.1136/bmjophth-2018-000254.supp1Supplementary data Table 1 Transcripts regulated in the conjunctival epithelium of patients with SJS

th SJS and the controls (ANOVA p<0.05). The 14 transcripts upregulated more than 50-fold were SERPINB4, KRT1, KRTDAP, S100A7, SBSN, KLK6, SERPINB12, PNLIPRP3, CASP14, ODZ2, CA2, CRCT1, CWH43 and FLG (table 1b); 6 of these 14 transcripts were related to cell differentiation or epidermis development in biological aspect. 10.1136/bmjophth-2018-000254.supp1Supplementary data Table 1 Transcripts regulated in the conjunctival epithelium of patients with SJS Fold change ANOVA P value* Gene accession Gene symbol Gene description Transcripts down-regulated by less than one -fiftieth in the conjunctival epithelium of patients with SJS −86.97 0.005 NM_001145006 MUC7 Mucin 7, secreted −67.01 0.019 NM_002644 PIGR Polymeric immunoglobulin receptor −59.17 0.021 NM_001039372 HEPACAM2 HEPACAM family member 2 −53.46 0.019 NM_000669 ADH1C Alcohol dehydrogenase 1C (class I), gamma polypeptide −53.25 0.000 NM_012390 SMR3A Submaxillary gland androgen regulated protein 3A Transcripts upregulated more than 50-fold in the conjunctival epithelium of patients with SJS 193.83 0.0023 NM_002974 SERPINB4 Serpin peptidase inhibitor, clade B, member 4 173.26 0.0001 NM_006121 KRT1 Keratin 1 150.7 0.0007 NM_207392 KRTDAP Keratinocyte differentiation-associated protein 123.93 0.0007 NM_002963 S100A7 S100 calcium-binding protein A7 121.69 0.0019 NM_001166034 SBSN Suprabasin 96.09 0.0046 NM_002774 KLK6 Kallikrein-related peptidase 6 91.02 0.0025 NM_080474 SERPINB12 Serpin peptidase inhibitor, clade B, member 12 89.35 0.0009 NM_001011709 PNLIPRP3 Pancreatic lipase-related protein 3 85.8 0.0008 NM_012114 CASP14 Caspase 14, apoptosis-related cysteine peptidase 68.67 0.0077 AK302302 ODZ2 Odz, odd Oz/ten-m homologue 2 53.44 0.0146 NM_000067 CA2 Carbonic anhydrase II 53.04 0.0188 NM_019060 CRCT1 Cysteine-rich C-terminal 1 52.76 0.0004 NM_025087 CWH43 Cell wall biogenesis 43 C-terminal homologue 51.27 0.0043 NM_002016 FLG Filaggrin *Difference from the control.

ysteine peptidase 68.67 0.0077 AK302302 ODZ2 Odz, odd Oz/ten-m homologue 2 53.44 0.0146 NM_000067 CA2 Carbonic anhydrase II 53.04 0.0188 NM_019060 CRCT1 Cysteine-rich C-terminal 1 52.76 0.0004 NM_025087 CWH43 Cell wall biogenesis 43 C-terminal homologue 51.27 0.0043 NM_002016 FLG Filaggrin *Difference from the control. Quantitative RT-PCR analysis: comparison of downregulated and upregulated transcripts in the conjunctival epithelium from patients with SJS and the controls We subjected conjunctival epithelium from 11 patients with SJS and 26 patients with conjunctival chalasis to quantitative RT-PCR assay to confirm the less than one-fiftieth downregulation of five transcripts and the more than 50-fold upregulation of 14 transcripts.

val epithelium from patients with SJS and the controls We subjected conjunctival epithelium from 11 patients with SJS and 26 patients with conjunctival chalasis to quantitative RT-PCR assay to confirm the less than one-fiftieth downregulation of five transcripts and the more than 50-fold upregulation of 14 transcripts. In patients with SJS, five transcripts tended to be downregulated; however, there was no significant difference between SJS and the controls with respect to MUC7 and SMR3A. Also, although 14 transcripts tended to be upregulated in SJS, there was no significant difference between patients with SJS and the controls with respect to 13 (SERPINB4, KRT1, KRTDAP, S100A7, SBSN, KLK6, SERPINB12, PNLIPRP3, CASP14, CA2, CRCT1, CWH43 and FLG). We attribute these findings to interindividual differences (online supplementary figures 1a,b). We concluded that the gene expression of PIGR (polymeric immunoglobulin receptor), HEPACAM2 (hepatic and glial cell adhesion molecule family member 2) and ADH1C (alcohol dehydrogenase 1C (class I), gamma polypeptide) was significantly downregulated and that ODZ2 (odd Oz/ten-m homologue 2), also called teneurin-2, was significantly upregulated in the conjunctival epithelium of patients with SJS with SOC (figure 1a,b). 10.1136/bmjophth-2018-000254.supp2Supplementary data

In patients with SJS, five transcripts tended to be downregulated; however, there was no significant difference between SJS and the controls with respect to MUC7 and SMR3A. Also, although 14 transcripts tended to be upregulated in SJS, there was no significant difference between patients with SJS and the controls with respect to 13 (SERPINB4, KRT1, KRTDAP, S100A7, SBSN, KLK6, SERPINB12, PNLIPRP3, CASP14, CA2, CRCT1, CWH43 and FLG). We attribute these findings to interindividual differences (online supplementary figures 1a,b). We concluded that the gene expression of PIGR (polymeric immunoglobulin receptor), HEPACAM2 (hepatic and glial cell adhesion molecule family member 2) and ADH1C (alcohol dehydrogenase 1C (class I), gamma polypeptide) was significantly downregulated and that ODZ2 (odd Oz/ten-m homologue 2), also called teneurin-2, was significantly upregulated in the conjunctival epithelium of patients with SJS with SOC (figure 1a,b). 10.1136/bmjophth-2018-000254.supp2Supplementary data Figure 1 Quantitative real-time PCR (RT-PCR). (A) mRNA expression of the significantly downregulated genes in patients with SJS with severe ocular complications (SOC) compared with the control. (B) mRNA expression of significantly upregulated genes in patients with SJS with SOC compared with the control. Quantification data were normalised to the expression of the housekeeping gene GAPDH. The Y axis shows the increase in specific mRNA over the control samples. Data are the mean±SEM (controls, n=26; SJS, n=11). *p<0.05, **p<0.005, ***p<0.0005. SJS, Stevens-Johnson syndrome.

s with SJS with SOC compared with the control. Quantification data were normalised to the expression of the housekeeping gene GAPDH. The Y axis shows the increase in specific mRNA over the control samples. Data are the mean±SEM (controls, n=26; SJS, n=11). *p<0.05, **p<0.005, ***p<0.0005. SJS, Stevens-Johnson syndrome. Protein expression of teneurin-2 in the conjunctival epithelium We focused on the upregulated transcript, ‘teneurin-2’ and examined the protein expression. Our immunohistochemical studies revealed the expression of teneurin-2 protein on human conjunctival epithelium. Our findings suggest that despite individual differences, the degree of expression might be higher in patients with SJS with SOC than the controls (figure 2). Figure 2 Teneurin-2 protein expression in conjunctival epithelium of patients with SJS with severe ocular complications (SOC). Immunohistochemistry detected teneurin-2 protein in the conjunctival epithelium of patients with Stevens-Johnson syndrome/toxic epidermal necrolysis (SJS/TEN). Bar: 50 µm. Discussion Our study of the pathology of ocular surface complications in patients with SJS with SOC in the chronic stage showed that in their conjunctival epithelium, the gene expression of PIGR, HEPACAM2 and ADH1C was significantly downregulated. The expression of ODZ2, also called teneurin-2, was significantly upregulated. We first document that teneurin-2 protein can be expressed in the human conjunctival epithelium.

the chronic stage showed that in their conjunctival epithelium, the gene expression of PIGR, HEPACAM2 and ADH1C was significantly downregulated. The expression of ODZ2, also called teneurin-2, was significantly upregulated. We first document that teneurin-2 protein can be expressed in the human conjunctival epithelium. PIGR, an epithelial glycoprotein that interacts with secretory IgA, is critical for regulating the secretory IgA level by transporting locally produced IgA.15 Because secretory IgA plays an important role in protecting mucosal surfaces such as the ocular surface against pathogens and antigens,15 the downregulated expression of the PIGR transcript in the conjunctival epithelium of patients with SJS with SOC may compromise immune protection of the ocular surface. Elsewhere we postulated an association between a disordered innate immune response and SJS with SOC.3 This hypothesis was based on our observation of an association between the onset of SJS/TEN with SOC and microbial infections: many patients exhibited prodromata, including non-specific fever, coryza and sore throat, ailments that closely mimic upper respiratory tract infections of viral or mycoplasma origin that are commonly treated with antibiotics and NSAIDs.2–5 8 In addition, our patients with SJS presented with opportunistic infection of the ocular surface by bacteria, especially methicillin-resistant Staphylococcus aureus and S. epidermis; their rate of detection was higher on the ocular surface of patients with SJS with SOC than patients with other devastating ocular surface disorders.3 We posit that in patients with SJS with SOC, abnormalities in innate immunity are responsible for opportunistic bacterial infections of the ocular surface.3 Moreover, these patients presented with persistent inflammation of the ocular surface which harbours commensal bacteria.3 Consequently, we think that the downregulation of PIGR results in compromised immune protection of the ocular surface and the anomalous innate immunity seen in patients with SJS with SOC.

surface.3 Moreover, these patients presented with persistent inflammation of the ocular surface which harbours commensal bacteria.3 Consequently, we think that the downregulation of PIGR results in compromised immune protection of the ocular surface and the anomalous innate immunity seen in patients with SJS with SOC. HEPACAM2 mediates the cellular-extracellular matrix and cell-cell interactions. This mediation is critical for the formation and maintenance of the cellular architecture and for normal biological processes, including the regulation of cell adhesion, proliferation, apoptosis, migration and differentiation.16 However, the clinical importance of HEPACAM2 and of ADH1C, which plays an important role in alcohol metabolism, on the ocular surface remains unclear.

e of the cellular architecture and for normal biological processes, including the regulation of cell adhesion, proliferation, apoptosis, migration and differentiation.16 However, the clinical importance of HEPACAM2 and of ADH1C, which plays an important role in alcohol metabolism, on the ocular surface remains unclear. Teneurin-2 is a member of the teneurin family, teneurin 1–4. Teneurins were initially described as ten-a and the pair-rule gene ten-m/odz in Drosophila. Teneurins belong to a novel class of signalling molecules that function both at the cell surface as type II transmembrane receptors and, after the release of the intracellular domain, as transcriptional regulators.17 The nuclear localisation of its intracellular domain has been observed in vitro in mammalian cells. The teneurin function appears to be required for a fundamentally important signalling mechanism conserved in invertebrates and vertebrates; it impacts many processes that rely on cell-cell contact throughout development.17 Although the clinical importance of teneurin-2 on the ocular surface remains unclear, we found that teneurin-2 protein is expressed in human conjunctival epithelium and that it may be upregulated in the conjunctival epithelium of patients with SJS. To the best of our knowledge, this is the first report of teneurin-2 expression on the human ocular surface. In summary, we suggest that the downregulation of PIGR, HEPACAM2 and ADH1C, and the upregulation of teneurin-2 expression contribute to the pathology involving the ocular surface of patients with SJS in the chronic stage.

Teneurin-2 is a member of the teneurin family, teneurin 1–4. Teneurins were initially described as ten-a and the pair-rule gene ten-m/odz in Drosophila. Teneurins belong to a novel class of signalling molecules that function both at the cell surface as type II transmembrane receptors and, after the release of the intracellular domain, as transcriptional regulators.17 The nuclear localisation of its intracellular domain has been observed in vitro in mammalian cells. The teneurin function appears to be required for a fundamentally important signalling mechanism conserved in invertebrates and vertebrates; it impacts many processes that rely on cell-cell contact throughout development.17 Although the clinical importance of teneurin-2 on the ocular surface remains unclear, we found that teneurin-2 protein is expressed in human conjunctival epithelium and that it may be upregulated in the conjunctival epithelium of patients with SJS. To the best of our knowledge, this is the first report of teneurin-2 expression on the human ocular surface. In summary, we suggest that the downregulation of PIGR, HEPACAM2 and ADH1C, and the upregulation of teneurin-2 expression contribute to the pathology involving the ocular surface of patients with SJS in the chronic stage. Contributors: MU planned the study, performed the experiments, and wrote and submitted the manuscript. CS and NY took the sample. HN, SO and KM performed the experiments. YN and SK reviewed the manuscript.

In summary, we suggest that the downregulation of PIGR, HEPACAM2 and ADH1C, and the upregulation of teneurin-2 expression contribute to the pathology involving the ocular surface of patients with SJS in the chronic stage. Contributors: MU planned the study, performed the experiments, and wrote and submitted the manuscript. CS and NY took the sample. HN, SO and KM performed the experiments. YN and SK reviewed the manuscript. Funding: This work was supported by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of the Japanese government, and by the JSPS Core-to-Core Program, A Advanced Research Networks, and also partly supported by Grants-in-Aid for Scientific Research from the Japanese Ministry of Health, Labor and Welfare. Competing interests: None declared. Ethics approval: The study was approved by the Institutional Review Board of Kyoto Prefectural University of Medicine. Provenance and peer review: Not commissioned; externally peer reviewed.

Key messages What is already known about this subject? A lack of eye care providers in low-resource settings contributes to a large burden of uncorrected refractive errors. Autorefractors are conventionally considered too expensive and inaccurate to significantly to improve refractive eye care capacity in these settings. What are the new findings? Eyeglass prescriptions can be accurately measured by a minimally trained technician using a low-cost wavefront autorefractor in rural India. Data from 708 participants indicate a marginal difference in both prescription preference and resulting visual acuity between eyeglasses derived from subjective refraction versus autorefraction. Among the 438 participants 40 years old and younger, there was no statistically significant difference in the preferences for eyeglasses derived from subjective refraction versus autorefraction. How might these results change the focus of research or clinical practice? These results suggest that eyeglasses prescribed objectively by a wavefront autorefractor may be a feasible approach to increasing eyeglass accessibility in low-resource settings. Introduction Over one billion people worldwide suffer from poor vision that could be corrected with a pair of prescription eyeglasses.1–3 These uncorrected refractive errors (UREs) are a major cause of lost productivity, limited access to education and reduced quality of life.

How might these results change the focus of research or clinical practice? These results suggest that eyeglasses prescribed objectively by a wavefront autorefractor may be a feasible approach to increasing eyeglass accessibility in low-resource settings. Introduction Over one billion people worldwide suffer from poor vision that could be corrected with a pair of prescription eyeglasses.1–3 These uncorrected refractive errors (UREs) are a major cause of lost productivity, limited access to education and reduced quality of life. The prevalence of UREs is generally highest in low-resource settings, due in part to the severe shortage of eye care professionals.2 4 There are several national and international efforts to increase eye care capacities by task-shifting the eyeglass prescription procedure to mid-level personnel called ‘refractionists’.4–6 However, these dedicated eye care workers still require several years of training and practice to become proficient,7 and it is difficult to retain these skilled workers in poor, rural and remote areas.8 There is a need to deskill the refraction process to reduce the training required for refractionist, increase their efficiency and improve the quality of their prescriptions.

veral years of training and practice to become proficient,7 and it is difficult to retain these skilled workers in poor, rural and remote areas.8 There is a need to deskill the refraction process to reduce the training required for refractionist, increase their efficiency and improve the quality of their prescriptions. Autorefractors are commonly used in high-resource settings to obtain a prescription that is used as a starting point for subjective refraction, reducing the overall time required for a refraction. However, autorefractors are conventionally considered too inaccurate to provide prescriptions without subjective refinement.9–12 Previous research comparing patient tolerance and acceptance of eyeglasses has found that approximately twice as many people preferred prescriptions from subjective refraction compared with prescriptions directly from an autorefractor, even after 3 weeks of habituating to the prescribed eyeglasses.9 10 A more recent study found a smaller gap in preferences using modern autorefractors on a young adult, non-presbyopic population—in this group, 41% more patients preferred prescriptions from subjective refraction compared with objective methods.12 Sophisticated autorefractors based on wavefront aberrometry have been explored for accurate prescriptions, enabled by algorithms incorporating both high-order and low-order aberrations and advanced quality metrics.13 14

more patients preferred prescriptions from subjective refraction compared with objective methods.12 Sophisticated autorefractors based on wavefront aberrometry have been explored for accurate prescriptions, enabled by algorithms incorporating both high-order and low-order aberrations and advanced quality metrics.13 14 Despite concerns over accuracy of objective refraction, several groups have developed systems with the goal of augmenting or even substituting for eye care providers in low-resource settings. Some of these approaches include the focometer,15 16 adjustable lenses,15 17 photorefraction,18 inverse Shack-Hartmann systems19 and simplified wavefront aberrometers.20 21 Previous work has assessed the accuracy of objective autorefractors relative to subjective refraction or conventional commercial autorefractors, but these studies have limited applicability to practical use in low-resource settings because (1) they tested a small population size and age range, (2) participants were highly educated (eg, optometry students), (3) the device was operated by highly trained eye care provider or engineer, (4) the test site was a controlled laboratory without examination time constraints, and/or (5) they excluded patients with comorbidities such as cataracts, keratoconus and conjunctivitis.

icipants were highly educated (eg, optometry students), (3) the device was operated by highly trained eye care provider or engineer, (4) the test site was a controlled laboratory without examination time constraints, and/or (5) they excluded patients with comorbidities such as cataracts, keratoconus and conjunctivitis. We recently introduced an aberrometer that uses low-cost components and calculates a prescription from dynamic wavefront measurements captured from a short video. Measurements from a previous study found that spherical error from this aberrometer agreed within 0.25 dioptres (D) of subjective refraction in 74% of eyes, compared with 49% agreement of the same eyes measured with a Grand Seiko WR-5100K commercial autorefractor.20 This prototype is currently under commercial development for low-resource markets (by PlenOptika, USA and Aurolab, India). The goal of this study was to assess the prescription quality from this device under realistic constraints for applicability in low-resource environments. Specifically, we evaluated the performance of this aberrometer when operated by a minimally trained technician in a low-resource setting on a large population of patients registered for routine eye examinations at either a major eye hospital or a satellite vision centre. Methods Participants Study objectives and procedures were explained in the local dialect and verbal informed consent was obtained. Written consent was obtained from additional participants for permission to publish photographs depicting them using the autorefractor.

We recently introduced an aberrometer that uses low-cost components and calculates a prescription from dynamic wavefront measurements captured from a short video. Measurements from a previous study found that spherical error from this aberrometer agreed within 0.25 dioptres (D) of subjective refraction in 74% of eyes, compared with 49% agreement of the same eyes measured with a Grand Seiko WR-5100K commercial autorefractor.20 This prototype is currently under commercial development for low-resource markets (by PlenOptika, USA and Aurolab, India). The goal of this study was to assess the prescription quality from this device under realistic constraints for applicability in low-resource environments. Specifically, we evaluated the performance of this aberrometer when operated by a minimally trained technician in a low-resource setting on a large population of patients registered for routine eye examinations at either a major eye hospital or a satellite vision centre. Methods Participants Study objectives and procedures were explained in the local dialect and verbal informed consent was obtained. Written consent was obtained from additional participants for permission to publish photographs depicting them using the autorefractor. Subjects were recruited from consecutive patients visiting the general ophthalmology unit of Aravind Eye Hospital in Madurai or a rural satellite vision centre in Thiruppuvanam. Inclusion criteria were that patients were between the ages of 15 and 70 years and within the refractive error range of the autorefractor (spherical equivalent of −6 D to +10 D), as determined by subjective refraction. Exclusion criteria included presence of mature cataract, any prior eye surgery, any major eye illnesses, and use of systemic or ocular drugs which may affect vision. The study was completed during the summer of 2015.

e of the autorefractor (spherical equivalent of −6 D to +10 D), as determined by subjective refraction. Exclusion criteria included presence of mature cataract, any prior eye surgery, any major eye illnesses, and use of systemic or ocular drugs which may affect vision. The study was completed during the summer of 2015. Subjective refraction procedure Patients who completed a standard-of-care refraction and met the study eligibility criteria were recruited for the study. This included streak retinoscopy and subjective refraction by an experienced refractionist. Refractions at the Aravind base hospital also included measurements by a standard commercial autorefractor before the subjective refraction. Subjective refraction was performed using a trial lens set and a digital visual acuity (VA) chart (Aurolab Aurochart) placed 3 m away from the participant. Autorefractor procedure A technician with experience in coordinating eye research studies but no training in refraction or clinical optometry was trained to use the prototype autorefractor in two 2-hour sessions, followed by 4 hours of practice refractions with the goal of consistently administering verbal instructions to the participants. All participants were tested by this technician. The autorefractor was calibrated at the beginning of the study. No recalibration was performed throughout the 3-month study duration, which included daily packing, unpacking and transportation. Every autorefractor measurement was performed directly after standard-of-care subjective refraction at a second station in a different room.

tor was calibrated at the beginning of the study. No recalibration was performed throughout the 3-month study duration, which included daily packing, unpacking and transportation. Every autorefractor measurement was performed directly after standard-of-care subjective refraction at a second station in a different room. Participants were instructed to hold the autorefractor to their face, rest their elbows on a table for support and look through the device at a back-lit VA chart placed 3 m away (figure 1). The technician adjusted the interpupillary distance wheel on the autorefractor and manually adjusted the pitch of the device until the participant could see a red spot coming from the autorefractor. When the participant saw a bright red spot within, the technician turned on the VA chart and began recording a 10 s video of wavefront measurements with the autorefractor. The participant was instructed to blink whenever desired and to look at the VA chart during the video. After the video was acquired, the device was flipped upside down to measure the opposite eye and the procedure was repeated. The participant was then measured two additional times for a total of three measurements of each eye. After the first interpupillary distance adjustment was made, typically no further adjustments were necessary. The device computed the median of the three measurements and displayed this prescription in the same format as subjective refraction on a companion laptop.

l times for a total of three measurements of each eye. After the first interpupillary distance adjustment was made, typically no further adjustments were necessary. The device computed the median of the three measurements and displayed this prescription in the same format as subjective refraction on a companion laptop. Figure 1 Testing procedure for the wavefront autorefractor. Participants looked through the open-view wavefront autorefractor at a distant back-lit visual acuity chart, while three 10 s videos of wavefront images were recorded by the device. The autorefractor was flipped over to measure the opposite eye. After repeating three times, the system displayed the autorefractor eyeglass prescription. Prescription quality assessment Sphere, cylinder and axis values were transcribed from the subjective refraction and autorefractor measurements to an electronic database, which randomly assigned them to prescriptions ‘A’ or ‘B’. The participant was escorted to a third station for VA measurement and preference survey by an experienced refractionist that was not involved in either prior refraction. This refractionist measured the VA of each eye using trial lenses set to each prescription pair in a randomised sequence, using a digital VA chart placed 3 m from the participant. The refractionist then asked the participant which prescription they preferred: A, B or no preference. VA and preference results were entered into an electronic database that used a de-identified numeric code to track each participant.

in a randomised sequence, using a digital VA chart placed 3 m from the participant. The refractionist then asked the participant which prescription they preferred: A, B or no preference. VA and preference results were entered into an electronic database that used a de-identified numeric code to track each participant. Statistical analysis For statistical comparison, prescriptions were converted to power vector parameters of spherical equivalent (M), vertical Jackson cross cylinder (J0) and oblique Jackson cross cylinder (J45) for subjective refraction (MSR, J0,SR, J45,SR) and autorefraction (MAR, J0,AR, J45,AR). Given that subjective refraction has significant interoptometrist and intraoptometrist variation,22 we performed a Bland-Altman analysis to assess correlation, bias and outliers between the two measurements for each power vector component. We computed the 95% limit of agreement between the two measurements using the approximation of the average difference ± (1.96 × SD) of the differences. We tested the effect of age in the disagreement between subjective refraction and autorefraction by fitting a linear model with slope, m, and intercept, b, to the absolute difference of power vector measurements of the right eye of each patient, and assessed the influence of gender and testing site as covariates. We also compared the average anisometropia measured by subjective refraction and autorefraction by comparing the difference in power vectors measured in the right and left eye of each patient.

ower vector measurements of the right eye of each patient, and assessed the influence of gender and testing site as covariates. We also compared the average anisometropia measured by subjective refraction and autorefraction by comparing the difference in power vectors measured in the right and left eye of each patient. All VA measurements were converted to logarithm of the minimum angle of resolution (logMAR) units for statistical comparison. VAs from uncorrected vision (VAUC), correction by autorefractor-determined prescription (VAAR) and correction by subjective refraction-determined prescription (VASR) were compared using a box and whisker plot of results from the right eyes only to avoid the influence of isometropia on the independence of the samples. Differences between mean values were assessed with a Wilcoxon signed-rank test with a significance level of 0.05. The participant survey for prescription preference was evaluated using a z test of proportion with a significance level of 0.05. Both VA and prescription preference results were analysed for the entire population and within two age groups partitioned by the estimated age of onset of presbyopia of 40 years of age.23

05. The participant survey for prescription preference was evaluated using a z test of proportion with a significance level of 0.05. Both VA and prescription preference results were analysed for the entire population and within two age groups partitioned by the estimated age of onset of presbyopia of 40 years of age.23 Results Participants We enrolled 506 participants from the base hospital and 202 participants from the vision centre. All 708 participants successfully received a testable prescription from both the prototype autorefractor and the subjective refraction. Within our study population, 220 participants had presbyopia, 89 participants had at least one immature cataract, 21 participants had conjunctivitis, and 1 participant had keratoconus. The mean±SD age of participants was 35±13 years, 438 participants were 15–40 years of age, 270 participants were 41–70 years of age, and 413 participants were female. A summary of the patient population, including a description of the distribution of refractive errors by gender, is presented in table 1. Table 1 Participant demographics and average refractive error measurements Gender n (%) Age±SD (range) (years) Subjective refraction Objective refraction

Results Participants We enrolled 506 participants from the base hospital and 202 participants from the vision centre. All 708 participants successfully received a testable prescription from both the prototype autorefractor and the subjective refraction. Within our study population, 220 participants had presbyopia, 89 participants had at least one immature cataract, 21 participants had conjunctivitis, and 1 participant had keratoconus. The mean±SD age of participants was 35±13 years, 438 participants were 15–40 years of age, 270 participants were 41–70 years of age, and 413 participants were female. A summary of the patient population, including a description of the distribution of refractive errors by gender, is presented in table 1. Table 1 Participant demographics and average refractive error measurements Gender n (%) Age±SD (range) (years) Subjective refraction Objective refraction M (D) (5%–95%) J0(D) (5%–95%) J45(D) (5%–95%) M (D) (5%–95%) J0(D) (5%–95%) J45(D) (5%–95%) Female 413 (58) 34±12 (15–65) −0.48 (−3.75 to 1.50) −0.02 (−0.50 to 0.42) 0.00 (−0.13 to 0.16) −0.55 (−4.36 to 1.50) 0.00 (−0.38 to 0.49) 0.04 (0.00–0.13) Male 296 (42) 37±14 (15–70) −0.59 (−3.50 to 1.00) −0.03 (−0.50 to 0.38) 0.00 (−0.16 to 0.22) −0.71 (−4.38 to 1.25) −0.03 (−0.49 to 0.37) 0.04 (0.00–0.15) 5%, 5th percentile; 95%, 95th percentile; D, dioptres; J0, vertical Jackson cross cylinder; J45, oblique Jackson cross cylinder; M, spherical equivalent.

(0.00–0.13) Male 296 (42) 37±14 (15–70) −0.59 (−3.50 to 1.00) −0.03 (−0.50 to 0.38) 0.00 (−0.16 to 0.22) −0.71 (−4.38 to 1.25) −0.03 (−0.49 to 0.37) 0.04 (0.00–0.15) 5%, 5th percentile; 95%, 95th percentile; D, dioptres; J0, vertical Jackson cross cylinder; J45, oblique Jackson cross cylinder; M, spherical equivalent. Patient and public involvement Patient participants were not involved in defining the research questions, developing the outcome measures, designing the study or in the recruitment. We conducted two pilot studies involving real patients, and their responses and feedback were used to modify the data collection protocol. The results of the study will be made available to the participants in the form of summary (written in lay language). Prescription agreement We observed a strong correlation between prescriptions from subjective refraction and the autorefractor, with Pearson linear correlation coefficients of r=0.94, r=0.83 and r=0.40 for M, J0 and J45, respectively (r2 values of 0.83, 0.70 and 0.16) (figure 2). The smaller correlation coefficient for J45 was likely influenced by the small range of values in the study population. The SD of J45, measured by subjective refraction, was only 0.12 D, compared with 1.46 D and 0.30 D for M and J0, respectively. In the correlation plot for figure 2A, one measurement (−3.75, –8.25) falls outside of the viewable range.

nt for J45 was likely influenced by the small range of values in the study population. The SD of J45, measured by subjective refraction, was only 0.12 D, compared with 1.46 D and 0.30 D for M and J0, respectively. In the correlation plot for figure 2A, one measurement (−3.75, –8.25) falls outside of the viewable range. Figure 2 Correlation and Bland-Altman plots of power vectors measured by autorefractor versus subjective refraction. Correlation (left) and Bland-Altman (right) plots comparing agreement of prescriptions measured by subjective refraction and the prototype autorefractor for (A) spherical equivalent, M, (B) vertical Jackson cross cylinder, J0, and (C) oblique Jackson cross cylinder, J45. AR, autorefractor prescription; SR, subjective refraction prescription; D, dioptre.

(right) plots comparing agreement of prescriptions measured by subjective refraction and the prototype autorefractor for (A) spherical equivalent, M, (B) vertical Jackson cross cylinder, J0, and (C) oblique Jackson cross cylinder, J45. AR, autorefractor prescription; SR, subjective refraction prescription; D, dioptre. From Bland-Altman analysis, we observed a bias between the subjective refraction and autorefractor measurements of −0.09 D, 0.01 D and 0.04 D for M, J0 and J45, respectively (figure 2), with the autorefractor reporting more myopic spherical equivalent values on average than subjective refraction. There was also a trend for larger magnitude measurements of both myopia and hyperopia by the autorefractor. A linear fit to the Bland-Altman data has a slope of 0.16 and an R of 0.36 (line not shown), signalling either a general undercorrection from subjective refraction or an overestimation of refractive error power measurement by the autorefractor. The 95% limits of agreement between the two methods were −1.47 D to 1.30 D, −0.35 D to 0.36 D, and −0.19 D to 0.27 D for M, J0 and J45, respectively. In the Bland-Altman plot for figure 2A, three measurements ((−6.00, –4.50), (−3.31, –6.63) and (−0.94, –4.88)) fall outside of the viewable range. Analysing the absolute difference of the subjective refraction versus autorefraction power vector measurements, we observed weak relationships with age, gender and site of measurement (table 2). The slope of linear fits to these data was in all cases less than or equal to 0.003 D per year of age, indicating a maximum change in average discrepancy between the two measurement techniques of 0.09 D in patient populations separated by a 30-year age difference, which is smaller than the 0.25 D rounding increments of the two approaches. For each power vector component, the average discrepancy between refractive approaches was within 0.08 D when comparing male versus female participants, and within 0.04 D when comparing refraction at Aravind Eye Hospital versus the satellite vision centre. The average difference between refraction of the right and left eye of each patient was 0.02 D, 0.00 D and 0.00 D for M, J0 and J45, respectively, for both subjective refraction and autorefraction, indicating no systematic difference in measurements between the two eyes for the two approaches.

satellite vision centre. The average difference between refraction of the right and left eye of each patient was 0.02 D, 0.00 D and 0.00 D for M, J0 and J45, respectively, for both subjective refraction and autorefraction, indicating no systematic difference in measurements between the two eyes for the two approaches. Table 2 Trends in absolute difference of subjective refraction versus autorefraction of right eye measurements All right eyes Female Male Hospital Vision centre Avg (D) m (D/year) b (D) Avg (D) m (D/year) b (D) Avg (D) m (D/year) b (D) Avg (D) m (D/year) b (D) Avg (D) m (D/year) b (D) |MSR -MAR| 0.482 0.001 0.43 0.448 0.001 0.43 0.529 0.002 0.47 0.492 0.001 0.46 0.455 0.002 0.39 |J0,SR-J0,AR| 0.110 0.002 0.04 0.111 0.003 0.02 0.109 0.001 0.06 0.116 0.001 0.07 0.096 0.004 −0.03 |J45,SR-J45,SR| 0.063 0.001 0.05 0.061 0.001 0.03 0.067 0.000 0.06 0.077 0.000 0.07 0.028 0.000 0.02 AR, autorefractor prescription; Avg, average; D, dioptre; J0, vertical Jackson cross cylinder; J45, oblique Jackson cross cylinder; M, spherical equivalent; SR, subjective refraction prescription; b, intercept of linear fit; m, slope of linear fit.

.05 0.061 0.001 0.03 0.067 0.000 0.06 0.077 0.000 0.07 0.028 0.000 0.02 AR, autorefractor prescription; Avg, average; D, dioptre; J0, vertical Jackson cross cylinder; J45, oblique Jackson cross cylinder; M, spherical equivalent; SR, subjective refraction prescription; b, intercept of linear fit; m, slope of linear fit. Visual acuity We measured a mean±SD of 0.30±0.37, –0.02±0.14 and −0.04±0.11 logMAR units for VAUC, VAAR and VASR, respectively. VA distributions for the whole study population as well as the age-grouped populations are shown in figure 3. VA was better after correction from both refraction methods (p<0.01) for all study groups. VASR was also better than VAAR (p<0.01) for all study groups, by margins of 0.01, 0.04 and 0.02 logMAR units for the younger, older and all age groups, respectively. Figure 3 Box plot of visual acuity before and after correction. Visual acuity of the right eyes without correction (VAUC), with trial lenses set to the autorefractor-determined prescription (VAAR) and with trial lenses set to the subjective refraction-determined prescription (VASR). There was a statistically significant difference (p<0.01) between average visual acuity measurements among all combinations within each age group (indicated by *). logMAR, logarithm of the minimum angle of resolution.

cription (VAAR) and with trial lenses set to the subjective refraction-determined prescription (VASR). There was a statistically significant difference (p<0.01) between average visual acuity measurements among all combinations within each age group (indicated by *). logMAR, logarithm of the minimum angle of resolution. Prescription preference Overall, 25% of participants had no preference of eyeglasses, 42% preferred prescriptions from subjective refraction, and 33% preferred prescriptions from the autorefractor (table 3). The entire population and the older groups preferred subjective refraction prescriptions more often than autorefractor prescriptions (p<0.01). Within the 342 participants in the younger group that had a preference, there was no statistically significant difference in prescription preference (49% preferred autorefractor prescriptions, 51% preferred subjective refraction prescriptions; p=0.52). Table 3 Participant preference of trial lens prescriptions with masked origin Age group Participants, n (%) P value SR vs AR preference All No preference Preferred SR Preferred AR 15–40 438 (61.9) 96 (21.9) 174 (39.7) 168 (38.4) 0.52 41–70 270 (38.1) 82 (30.4) 123 (45.6) 65 (24.1) <0.01 All 708 (100.0) 178 (25.1) 297 (41.9) 233 (32.9) <0.01 AR, autorefractor prescription; SR, subjective refraction prescription.

ipants, n (%) P value SR vs AR preference All No preference Preferred SR Preferred AR 15–40 438 (61.9) 96 (21.9) 174 (39.7) 168 (38.4) 0.52 41–70 270 (38.1) 82 (30.4) 123 (45.6) 65 (24.1) <0.01 All 708 (100.0) 178 (25.1) 297 (41.9) 233 (32.9) <0.01 AR, autorefractor prescription; SR, subjective refraction prescription. Discussion This study found smaller differences in VA and preference of prescriptions obtained from autorefraction compared with subjective refraction than previous work.9–12 There are several differences to our study design and autorefractor that may contribute to this result. The refractionists used in our study specialise in high-volume refractive eye exams and have less training than optometrists or ophthalmologists used in other studies. Our study used a 3 m refraction distance since it is the standard of care within the Aravind system, but the convention of most eye exams is a 6 m or 20-foot distance. Our study was also conducted on an Indian population in a low-resource setting, which could have systematic differences in VA preferences and compliance to subjective refraction instructions. The autorefractor tested in our study is significantly different from previous studies. It is an open-view wavefront aberrometer that analyses wavefront data from three 10 s videos of measurements (typically 240 wavefronts), rather than a single snapshot or the average of several images. Lastly, this autorefractor prototype had a spherical equivalent refractive error range of −6 D to +10 D, and only patients within this range were recruited for this study. Therefore, the conclusion of these results is relevant only to patients within this range tested.

single snapshot or the average of several images. Lastly, this autorefractor prototype had a spherical equivalent refractive error range of −6 D to +10 D, and only patients within this range were recruited for this study. Therefore, the conclusion of these results is relevant only to patients within this range tested. This study is, to the best of our knowledge, the first that identifies a population (patients 40 years old and younger) that exhibits no statistically significant difference between preferences of prescriptions derived from an autorefractor compared with subjective refraction. The difference in preference between the two age groups may be due to several physiological parameters that vary with age. While patients with mature cataracts were excluded from this study, 6 patients (1.4%) in the younger group were noted to have at least one immature cataract, while 83 patients (30.7%) in the older group were noted to have at least one immature cataract. Pupil size was not directly measured here, but is known to decrease significantly with age.24 Both opacities in the lens and a small pupil make the projection of the wavefront beacon on the retina and the measurement of the emerging wavefront more difficult. The older group is also expected to have smaller accommodative amplitude. Closed-view wavefront autorefractors are known to cause instrument-induced myopia, leading to an overestimation of myopia.25 However, the system evaluated here is open-view and the observed trend was of greater autorefractor prescription preference in the population expected to have larger accommodation amplitude. Lastly, the technological literacy and compliance to both the subjective refraction and autorefraction procedures may differ between the age groups, both of which could influence the quality of the prescriptions from each method. We observed that participant age, gender and site were weakly related to the difference between refractive errors measured by subjective refraction and autorefraction, indicating that the increasing preference for subjective refraction-determined prescriptions with older age is likely of multifactorial origin and cannot be explained by simple systematic differences in the prescriptions from the two approaches.

en refractive errors measured by subjective refraction and autorefraction, indicating that the increasing preference for subjective refraction-determined prescriptions with older age is likely of multifactorial origin and cannot be explained by simple systematic differences in the prescriptions from the two approaches. In this study, we only surveyed participants for nominal prescription preference. Future work assessing the qualitative strength of preference and satisfaction of each prescription with ordinal surveys is under way and will provide more insight into differences in perceived quality of the prescriptions. We also assessed VA and preference immediately after the eye examination with trial lenses, but assessing prescription quality after several weeks of habituation to the prescription eyeglasses will improve the understanding of factors influencing long-term patient satisfaction. Lastly, a new version of the prototype autorefractor evaluated in this study is currently being commercialised with a larger refractive range, improved ergonomics and is targeted to be cost-effective for low-resource settings.

glasses will improve the understanding of factors influencing long-term patient satisfaction. Lastly, a new version of the prototype autorefractor evaluated in this study is currently being commercialised with a larger refractive range, improved ergonomics and is targeted to be cost-effective for low-resource settings. Participants using eyeglasses prescribed by the autorefractor operated by a non-clinical, minimally trained technician achieved a VA that was only approximately one letter worse than using eyeglasses prescribed by an experienced refractionist. Moreover, although participants preferred subjective refraction prescriptions in aggregate, participants 40 years of age and younger had no statistically significant difference in their preference. Given the minimal training required to use the autorefractor tested here and the marginal difference in prescription quality by the refractionist compared with the autorefractor, wavefront-based objective prescriptions may be a viable substitute for subjective refraction in low-resource settings. NJD and SRD contributed equally. Contributors: NJD and SRD contributed equally as co-first authors. NJD, SRD, DL, EL, SJ and TDR designed the study. NJD, SRD, DL and EL contributed to the technical design of the autorefractor prototype. NJD analysed the data and prepared the manuscript. SRD and DL trained the autorefraction technician and organised the study facilities. SRD and SJ oversaw the recruitment and data collection. DL and EL contributed to the data analysis.

, SRD, DL and EL contributed to the technical design of the autorefractor prototype. NJD analysed the data and prepared the manuscript. SRD and DL trained the autorefraction technician and organised the study facilities. SRD and SJ oversaw the recruitment and data collection. DL and EL contributed to the data analysis. Funding: This work was financially supported by a grant from the United States–India Science & Technology Endowment Fund (USISTEF) awarded jointly to PlenOptika and Aurolab. Financial support for the initial autorefractor development is gratefully acknowledged from Comunidad de Madrid through the Madrid-MIT M+Visión Consortium. These funding sources had no role in the design or execution of the study. Research relating to the autorefractor reported in this publication was partially supported by the National Eye Institute of the National Institutes of Health under Award Number R43EY025452 and R44EY025452. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Competing interests: NJD, SRD, DL and EL are inventors on patents relating to the autorefractor used in this study and have a financial interest in PlenOptika. SRD and DL are employees of PlenOptika. NJD and EL are technical advisors and collaborators of PlenOptika. Patient consent for publication: Obtained. Ethics approval: Aravind Eye Care System Institutional Review Board (Project Code RES2014065PDT) approved the study protocol. Provenance and peer review: Not commissioned; externally peer reviewed.

Key messages What is already known about this subject? Optical coherence tomography (OCT) is widely used in older patients with glaucoma in whom normative ranges have been established and patterns of disease are well described. What are the new findings? Using hand-held OCT (HH-OCT) in a younger group of patients (mean age=4.64 years), our novel findings in primary congenital glaucoma were a particularly increased cup depth (vs primary open-angle glaucoma), a reduction of the macula pit width and a non-detectable external limiting membrane. HH-OCT in children with glaucoma confirmed many similar findings to those in adults (thinned retinal nerve fibre, thinned ganglion cell layer, enlarged cup diameter, smaller rim areas and smaller rim volumes). We identify particular threshold parameters to optimise recognition of glaucomatous nerve changes within the studied population. How might these results change the focus of research or clinical practice? Sensitivity and specificity of optic nerve changes for a clinical diagnosis of glaucoma suggest that HH-OCT has the potential to be a useful tool to detect disease in young children and is optimised by using the threshold values which this work suggests. This is groundwork for a larger prospective study powered to evaluate childhood glaucoma diagnosis and disease control.

How might these results change the focus of research or clinical practice? Sensitivity and specificity of optic nerve changes for a clinical diagnosis of glaucoma suggest that HH-OCT has the potential to be a useful tool to detect disease in young children and is optimised by using the threshold values which this work suggests. This is groundwork for a larger prospective study powered to evaluate childhood glaucoma diagnosis and disease control. Introduction Glaucoma is an important cause of preventable visual loss in childhood.1–3 Causes for reduced vision include corneal scarring, anisometropic amblyopia, buphthalmia and optic neuropathy.4 5 Primary congenital glaucoma (PCG) is one of the most common types of childhood glaucoma comprising 32%4–47%2 of cases and is usually diagnosed in infancy.6 Obstruction of aqueous outflow causes raised intraocular pressure and optic neuropathy.5 Assessment of children with glaucoma is challenging because of limited co-operation, media opacity and examination equipment designed for adults rather than paediatric use. While in adults optical coherence tomography (OCT) imaging is standard for assessing glaucoma, infants and young children have not benefited from this technology as they cannot co-operate with standard table-mounted devices. To date, limited literature about the use of OCT in PCG is available.

ther than paediatric use. While in adults optical coherence tomography (OCT) imaging is standard for assessing glaucoma, infants and young children have not benefited from this technology as they cannot co-operate with standard table-mounted devices. To date, limited literature about the use of OCT in PCG is available. Retinal nerve fibre layer (RNFL), ganglion cell layer (GCL) thinning and enlarged cup with smaller rim areas/volumes of the optic nerve (ON) were found in a cohort of children of an average age of 10.1±3.6 years after glaucoma surgery on OCT.7 Cup:disc ratio (CDR) can be reversible in paediatric glaucoma after surgery. In contrast, RNFL thinning progressed even after treatment and visual prognosis depended mainly on preoperative RNFL thickness.8 Retinal changes including cystoid macular oedema, atrophy, pigment epithelial detachment, choroidal folds and/or disruption of the inner/outer photoreceptor segments were found in 13.2% of children with different types of glaucoma over 12 years of age.9 The aim of this study was to investigate in PCG (1) the feasibility of scanning the ON and central retina with hand-held OCT (HH-OCT) without sedation or anaesthesia; (2) the characteristics of ON changes in comparison with adult open-angle glaucoma (POAG), and for both groups, comparison with matched controls; (3) the sensitivity and specificity of ON parameters for diagnosis; and (4) changes of the size and shape of the foveal pit and individual retinal layers.

ion or anaesthesia; (2) the characteristics of ON changes in comparison with adult open-angle glaucoma (POAG), and for both groups, comparison with matched controls; (3) the sensitivity and specificity of ON parameters for diagnosis; and (4) changes of the size and shape of the foveal pit and individual retinal layers. Methods In this observational cross-sectional study, we investigated 20 patients with PCG (10 females and 10 males; mean age±SD=4.64±2.79 years), and 20 age-matched, gender-matched and ethnicity-matched healthy controls (9 females and 11 males; mean age 4.73±2.81 years).

ion or anaesthesia; (2) the characteristics of ON changes in comparison with adult open-angle glaucoma (POAG), and for both groups, comparison with matched controls; (3) the sensitivity and specificity of ON parameters for diagnosis; and (4) changes of the size and shape of the foveal pit and individual retinal layers. Methods In this observational cross-sectional study, we investigated 20 patients with PCG (10 females and 10 males; mean age±SD=4.64±2.79 years), and 20 age-matched, gender-matched and ethnicity-matched healthy controls (9 females and 11 males; mean age 4.73±2.81 years). All patients were imaged between November 2015 and November 2016. Twenty-four patients were recruited from paediatric glaucoma clinics at Birmingham Children’s Hospital and University Hospitals of Leicester, UK. All eligible sequential patients were included as part of this study. PCG was confirmed based on the presence of clinical features documented in the clinical record (reference test). The ninth Consensus Report of the World Glaucoma Association (http://www.worldglaucoma.org/consensus-9/) criteria were used for PCG diagnosis. Inclusion criteria for the study were two or more of the following (in the absence of other ocular or systemic causes): elevated intraocular pressure (IOP), ON cupping, corneal changes of glaucoma, increasing axial length or myopia. Only clinically confirmed PCG was included. Secondary glaucomas were excluded including all in which structural abnormalities of the anterior segment were found (except isolated goniodysgenesis; a feature of PCG). Those children for whom cooperation or media opacity prevented adequate image acquisition were excluded. Healthy children were recruited from Leicester nurseries and schools and matched by age, gender, ethnicity and clinical/scanning protocol.

rior segment were found (except isolated goniodysgenesis; a feature of PCG). Those children for whom cooperation or media opacity prevented adequate image acquisition were excluded. Healthy children were recruited from Leicester nurseries and schools and matched by age, gender, ethnicity and clinical/scanning protocol. To compare ON changes with POAG, we included data from 10 patients with POAG (all patients had visual field defect with ON cupping and RNFL and neuroretinal rim thinning on OCT; https://wga.one/wga/consensus-10/) (two females, eight males, mean age 66.8±6.94 years). POAG cases were recruited from glaucoma clinics at Leicester Royal Infirmary during 12 months from November 2015. POAG controls were recruited from hospital staff and the general public and were matched for age, gender, ethnicity and refractive error. Ophthalmic examination included best-corrected visual acuity, cycloplegic refraction, slit-lamp and fundus examination and measurement of horizontal corneal diameter and IOP (i-care TAO1i, Finland/Goldman tonometry if possible). Demographic characteristics and clinical data of participants are shown in table 1. Table 1 Clinical and demographic characteristics of children with primary congenital and adults with open-angle glaucoma at inclusion in study

Ophthalmic examination included best-corrected visual acuity, cycloplegic refraction, slit-lamp and fundus examination and measurement of horizontal corneal diameter and IOP (i-care TAO1i, Finland/Goldman tonometry if possible). Demographic characteristics and clinical data of participants are shown in table 1. Table 1 Clinical and demographic characteristics of children with primary congenital and adults with open-angle glaucoma at inclusion in study ID Age (years) Sex Ethnicity Affected eye Vision (logMAR) Refraction (SE) IOP Corneal diameter (mm) Cup:disc ratio Surgical treatment RE LE RE LE RE LE RE LE RE LE A. Primary congenital glaucoma 1 10.3 M A BE 0.00 −0.06 Plano Plano 16 18 12 12 0.4 0.3 G(BE) 2 9.3 M A BE 1.30 0.08 −9.00 −1.00 20 20 12.5 11 0.4 0.3 G(BE×2) 3 0.8 F C BE 0.98 (PL) FF (PL) −2.75 −1.75 22 20 14 10 0.3 0.2 G(BE), TRAB(LE) 4 4.3 M C BE 0.13 0.13 −0.25 −0.50 15 21 12 12 0.35 0.5 G(BE), TRAB(RE) 5 2.1 F C BE 1.10 (CC) 1.10 (CC) +1.00 Plano 14 16 14 14 0.1 0.1 G(BE), BT(RE) 6 3.7 F A BE 0.40 (CC) 0.40 (CC) +0.50 −0.50 20 20 13 13 0.8 0.8 G(BE) 7 5.8 F A BE 0.43 0.70 −6.25 −1.00 14 12 14 13 0.5 0.5 TRAB(BE) 8 8.0 F A BE 0.22 0.64 −1.50 1.00 12 12 11 13 0.5 0.5 TRAB(BE×2RE) 9 3.6 F A BE 0.20 0.20 +1.75 −2.75 6 11 14 13 0.5 0.5 G(BE) 10 1.5 M C RE 0.80 (CC) 0.70 (CC) −0.25 +2.25 16 14 12 11 0.5 0.5 G(RE×2) 11 0.4 M C BE 0.86 (PL) 0.64 (PL) −1.00 −1.50 29 25 12 13 0.5 0.5 Listed for surgery 12 5.7 M C BE 0.15 0.13 −1.00 −1.00 8 8 13 13 0.7 0.4 G(BE) 13 1.6 M C LE 0.10 (PL) 0.70 (PL) −1.00 Plano 14 12 10 13 0.3 0.6 G(LE×2) 14 3.6 M C RE 0.95 0.30 −11.00 −0.25 18 8 13 10 0.8 0.3 G(RE) 15 2.9 F A LE 0.35 1.06 1.25 −3.50 18 20 10 12 0.2 0.45 G(LE) 16 5.0 F A RE 1.83 −0.05 −3.00 +0.75 14 16 12 10 0.2 0.2 G(RE×2), CL(RE) 17 2.3 F C BE 0.51 0.51 −1.25 1.50 23 12 23 12 0.2 0.2 G(RE) 18 5.1 M C LE 0.08 LP +0.25 −7.00 28 12 9 12 0.2 0.9 G(LE×3) 19 7.2 M C BE 1.15 0.08 +2.25 +2.25 30 18 12 12.5 0.8 0.9 G(BE), TRAB(BE) 20 6.7 F C BE 0.10 0.10 −1.00 −1.00 25 23 14 13 0.4 0.9 G(BE) 21 2.1 M C BE 0.8 0.7 −0.50 +1.25 14 16 10 12 0.4 0.4 G(BE) 22 0.8 M A LE FF FF +1.00 −3.00 12 20 10 12 0.4 0.7 G(LE), TRAB(LE) 23 4.1 F A BE 0.4 0.4 Plano Plano 16 16 12 12 0.5 0.5 G(BE) 24 2.0 M A BE FF FF – – 19 20 12 11 – – Listed for surgery B.

TRAB(BE) 20 6.7 F C BE 0.10 0.10 −1.00 −1.00 25 23 14 13 0.4 0.9 G(BE) 21 2.1 M C BE 0.8 0.7 −0.50 +1.25 14 16 10 12 0.4 0.4 G(BE) 22 0.8 M A LE FF FF +1.00 −3.00 12 20 10 12 0.4 0.7 G(LE), TRAB(LE) 23 4.1 F A BE 0.4 0.4 Plano Plano 16 16 12 12 0.5 0.5 G(BE) 24 2.0 M A BE FF FF – – 19 20 12 11 – – Listed for surgery B. Healthy children (controls for primary congenital glaucoma) 1 0.4 F C – 0.4 (PL) 0.4 (PL) Plano Plano 12 12 10.5 10.5 0.3 0.3 – 2 0.5 M C – 0.3 (PL) 0.3 (PL) +2.00 +2.00 14 13 11 11 0.2 0.2 – 3 0.7 F C – 0.4 (PL) 0.4 (PL) +3.25 +3.00 11 12 10.8 10.8 0.3 0.3 – 4 0.8 F C – 0.3 (PL) 0.3 (PL) +2.50 +2.50 14 14 11 11 0.3 0.3 – 5 0.8 M A – FF FF +2.50 +2.50 12 14 11 11 0.3 0.3 – 6 10 M A – 0.3 0.3 Plano Plano 16 16 11.2 11.2 0.35 0.35 – 7 3 F C – 0.1 (CC) 0.1 (CC) +1.25 +1.25 16 17 11 11 0.2 0.2 – 8 4 F A – 0.00 0.00 +1.00 +1.00 16 16 11 11 0.3 0.3 – 9 7 F A – 0.10 0.10 Plano Plano 17 16 11 11 0.3 0.3 – 10 4 M C – 0.2 (CC) 0.2 (CC) +0.75 +0.75 18 18 11 11 0.3 0.3 – 11 6.4 M C – 0.15 0.13 Plano Plano 15 15 10 10 0.3 0.3 – 12 5.5 M C – 0.10 0.10 Plano Plano 16 18 11 11 0.3 0.3 – 13 2 F A – 0.3 (CC) 0.3 (CC) +1.25 +1.25 16 14 11 11 0.3 0.3 – 14 4 M C – 0.20 0.20 +1.00 +1.00 16 16 11 11 0.3 0.3 – 15 5 F A – 0.10 0.10 +1.25 +1.25 18 18 11 11 0.2 0.2 – 16 5 M C – 0.00 0.00 +0.50 +0.75 16 16 11 11 0.2 0.2 – 17 1.5 M C – 0.2 (PL) 0.2 (PL) +0.75 +0.75 18 18 11 11 0.3 0.3 – 18 4 M A – 0.10 0.10 +2.00 +2.00 16 16 11 11 0.35 0.35 – 19 4.3 M A – 0.00 0.00 +2.00 +2.00 19 20 11 11 0.3 0.3 – 20 3.6 F A – 0.10 0.20 +1.50 +1.75 20 20 11 11 0.4 0.4 – C. Open-angle glaucoma 1 69 F C BE 0.32 0.18 +0.25 +0.5 14 16 12 12 0.8 0.8 TRAB(RE) 2 65 M C RE 0.16 0.20 Plano Plano 20 19 12 12 0.85 0.5 – 3 76 M C BE 0.45 0.20 +1 +1 22 22 12 12 0.5 0.7 – 4 76 M C BE 0.36 0.2 Plano Plano 16 15 12 12 0.6 0.5 – 5 59 M C BE 0.10 0.10 +0.25 Plano 12 15 12 12 0.7 0.8 – 6 69 F A BE 0.10 0.10 Plano Plano 18 18 12 12 0.8 0.8 – 7 56 M C BE 0.14 0.22 Plano Plano 15 16 12 12 0.8 0.65 TRAB(BE) 8 66 M C BE 0.14 0.16 Plano Plano 13 11 12 12 0.6 0.7 – 9 72 M C BE 0.36 0.22 Plano Plano 14 14 12 12 0.7 0.8 – 10 60 M C BE 0.24 0.20 Plano Plano 16 17 12 12 0.6 0.5 – D.

Plano 12 15 12 12 0.7 0.8 – 6 69 F A BE 0.10 0.10 Plano Plano 18 18 12 12 0.8 0.8 – 7 56 M C BE 0.14 0.22 Plano Plano 15 16 12 12 0.8 0.65 TRAB(BE) 8 66 M C BE 0.14 0.16 Plano Plano 13 11 12 12 0.6 0.7 – 9 72 M C BE 0.36 0.22 Plano Plano 14 14 12 12 0.7 0.8 – 10 60 M C BE 0.24 0.20 Plano Plano 16 17 12 12 0.6 0.5 – D. Healthy adults (controls for open-angle glaucoma) 1 77 M C – 0.10 0.10 Plano +0.25 19 19 12 12 0.4 0.4 – 2 54 M C – 0.20 0.10 Plano +0.25 17 16 12 12 0.35 0.4 – 3 60 M C – 0.10 0.10 Plano +0.50 14 14 12 12 0.4 0.4 – 4 71 F A – 0.20 0.20 +0.75 Plano 12 12 12 12 0.4 0.4 – 5 66 F C – 0.20 0.10 +0.50 Plano 16 16 12 12 0.3 0.3 – 6 72 M C – 0.10 0.10 Plano Plano 12 14 12 12 0.5 0.5 – 7 57 M C – 0.10 0.20 Plano Plano 19 20 12 12 0.4 0.4 – 8 71 M C – 0.10 0.20 Plano Plano 18 18 12 12 0.3 0.4 – 9 69 M C – 0.20 0.10 Plano Plano 16 16 12 12 0.4 0.4 – 10 72 M C – 0.10 0.10 Plano Plano 20 20 12 12 0.5 0.45 – IOP was measured at the date when optical coherence tomography was done. BE, both eyes; BT, Baelvedt tube; C, Caucasian; CC, Cardiff cards (in all other participants, Kay pictures or Snellen test was used); CL, cyclodiode laser; F, female; FF, following and fixing; G, goniotomy; IOP, intraocular pressure; LE, left eye; M, male; PL, perception of light; PL, preferential looking; RE, right eye; SE, spherical equivalent; TRAB, trabeculectomy.

, Cardiff cards (in all other participants, Kay pictures or Snellen test was used); CL, cyclodiode laser; F, female; FF, following and fixing; G, goniotomy; IOP, intraocular pressure; LE, left eye; M, male; PL, perception of light; PL, preferential looking; RE, right eye; SE, spherical equivalent; TRAB, trabeculectomy. OCT imaging HH-OCT was performed without anaesthesia or sedation in outpatient clinics by research orthoptists without knowledge of the specific clinical details and diagnosis of the patients. A hand-held SD-OCT device (Envisu 2300; Leica Microsystems, Germany) 840 nm wavelength, 2.6 mm theoretic axial resolution was used. ON and macula imaging was carried out using 12 mm wide by 8 mm high scans with 80 horizontal B-scans composed of 600 A-scans and a 1.67 s acquisition time. To ensure central ON and fovea position of the B-scan, position was verified on the en-face view. Masked optic nerve head analysis was performed on one B-scan through the centre of the ON cup using a custom-written ImageJ macro (National Institutes of Health, Bethesda, Maryland, USA; available at http://rsbweb.nih.gov/ij/; accessed 3 March 2016) by SS. Cup and disc measurements are derived as shown in figure 1A from a plane of 150 µm anterior to the disc edges. Parameters included horizontal disk and cup diameter, CDR, cup area, rim width (temporal/nasal width combined), rim area (temporal/nasal combined), and nasal and temporal RNFL. Cup depth was calculated as the axial distance from the cup bottom to a line connecting the nasal and temporal rim edges (figure 1A).

sc edges. Parameters included horizontal disk and cup diameter, CDR, cup area, rim width (temporal/nasal width combined), rim area (temporal/nasal combined), and nasal and temporal RNFL. Cup depth was calculated as the axial distance from the cup bottom to a line connecting the nasal and temporal rim edges (figure 1A). Figure 1 Hand-held optical coherence tomography (HH-OCT) images of the optic nerve (ON) (A) and fovea (B) of the left eye of patient 6: (A) horizontal B-scan through the centre (deepest excavation) of the optic disc; the disc diameter was defined as an interval between the edges of Bruch’s membrane (red line), the cup diameter as the distance between the nasal and temporal internal limiting membrane (green dotted line) 150 µm anterior to the plane of the disc (blue line), the rim area consisted of the area anterior to the same plane (white dotted lines) within the disc edges (white vertical lines) and the internal limiting membrane (green dotted lines); maximal cup depth (vertical yellow line) was measured using a line perpendicular to the line between the cup diameter (blue line) and the deepest point of the cup; RNFL thickness was measured at 6° from disc margins (red dotted lines). (B) Horizontal B-scan of the fovea with labelled individual retinal layers (BM, Bruch’s membrane; ILM, inner limiting membrane; RNFL, retinal nerve fibre layer; GCL, ganglion cell layer; INL, inner nuclear layer; IPL, inner plexiform layer; IS, inner segment; ONL, outer nuclear layer; ELM, external limiting membrane; OPL, outer plexiform layer; OS, outer segment; RPE, retinal pigment epithelium). Retinal layer thickness was measured in the centre of the fovea, in the paracentral area (from 1° nasally to 1° temporally) and nasally and temporally (from 2° to 6°). The green line connecting the most prominent positions of the ILM nasally and temporally was used to define the foveal width; the yellow line indicates the foveal depth (the axial distance from the green line to the deepest point of the foveal pit); the area in blue indicates the foveal pit area. (C) Horizontal spectral domain–optical coherence tomography B-scan images of the optic nerve (top) and fovea (bottom) of patient 20 with primary congenital glaucoma in the left eye (PCG, middle column), an unaffected left eye (left column) and an eye of a healthy age-matched, gender-matched and ethnicity-matched control child (right column). On the ON scan, a larger and deeper cup is seen in PCG.

e optic nerve (top) and fovea (bottom) of patient 20 with primary congenital glaucoma in the left eye (PCG, middle column), an unaffected left eye (left column) and an eye of a healthy age-matched, gender-matched and ethnicity-matched control child (right column). On the ON scan, a larger and deeper cup is seen in PCG. On the foveal scan, the ELM is not visible in PCG while it is distinctly seen in the unaffected eye and in the healthy control. Lateral measures were converted to visual angle using a conversion factor provided by Leica Microsystems derived using a model eye (292 µm=1° angle for small angles). This addresses the problems of calibration errors of lateral distance measures caused by changing axial lengths in the developing eye and in eyes with severe refractive error present in children with glaucoma. Segmentation of the foveal region masked to PCG or control was performed using ImageJ and included the RNFL, and inner nuclear (INL) and outer plexiform (OPL) layers.10 The outer nuclear layer (ONL) and inner segment (IS), the outer segment and retinal pigment epithelium layers, the GCL and inner plexiform (IPL=ganglion cell complex) layers were combined for analysis due to poor contrast difference.11 12 Layer thicknesses were averaged in the paracentral area (from 1° nasally to 1° temporally from the foveal pit bottom) as well as in the nasal and temporal areas (from 2° to 6° from the foveal pit bottom). The inner limiting membrane on the flattened B-scan image was used to calculate foveal pit depth, width and area (figure 1B).10

esses were averaged in the paracentral area (from 1° nasally to 1° temporally from the foveal pit bottom) as well as in the nasal and temporal areas (from 2° to 6° from the foveal pit bottom). The inner limiting membrane on the flattened B-scan image was used to calculate foveal pit depth, width and area (figure 1B).10 Statistical analysis SPSS software V.16.0 was used. ON head and macular parameters were normally distributed. Separate linear mixed models were generated for each of the OCT parameters to allow inclusion of repeated measures data (ie, eye) and non-repeated measures data (ie, group and age) in the same statistical models. In each model, the OCT parameter was the outcome variable, participant-anonymised ID was included as a random factor; group and eye were included as fixed factors; and age as a covariate. Bonferroni correction was used to correct for the multiple comparisons. To compare glaucoma between children and adults, each ON parameter was expressed as percentage of the mean control value. Sensitivity and specificity of ON parameters for PCG detection were calculated using fixed thresholds for each ON parameter that demonstrated statistically significant difference. Correlations between ON and macular parameters, IOP and horizontal corneal diameter were calculated using Pearson’s correlation. P value ≤0.05 was considered statistically significant.

Sensitivity and specificity of ON parameters for PCG detection were calculated using fixed thresholds for each ON parameter that demonstrated statistically significant difference. Correlations between ON and macular parameters, IOP and horizontal corneal diameter were calculated using Pearson’s correlation. P value ≤0.05 was considered statistically significant. Results ON head morphology Four patients were excluded from analysis; in two children, the affected eye could not be assessed due to low scan quality secondary to corneal oedema, and two children did not co-operate (table 1, patients 21–24). Data from both eyes of 20 out of 24 patients with PCG were included (83.3% of eyes). Mean IOP before OCT examination was 17.68±6.52 in PCG and 22.12±4.16 in POAG. Visual inspection showed marked cupping with enlarged horizontal cup diameter and deeper cup depth in most patients (figure 1C). In PCG, ON analysis showed significantly increased CDR, cup diameter, cup area and cup disc. Rim width, rim area and nasal RNFL were significantly reduced in PCG. In contrast, in adult glaucoma, we did not find significant differences in cup depth and nasal RNFL as compared with their controls. Children with PCG had significantly deeper cup depth (165% increased from control children) than patients with POAG (117% from adult controls) (p=0.014) (table 2C and figure 2).

In PCG, ON analysis showed significantly increased CDR, cup diameter, cup area and cup disc. Rim width, rim area and nasal RNFL were significantly reduced in PCG. In contrast, in adult glaucoma, we did not find significant differences in cup depth and nasal RNFL as compared with their controls. Children with PCG had significantly deeper cup depth (165% increased from control children) than patients with POAG (117% from adult controls) (p=0.014) (table 2C and figure 2). Figure 2 Cross-sectional schematic diagrams representing mean values of optic nerve head parameters of patients with primary congenital glaucoma, adult open-angle glaucoma and matched healthy controls for each group (numeric values represent mean±SE). The upper horizontal dotted lines represent the horizontal offset (150 µm) used to determine cup diameters and the lower horizontal dotted lines indicate disc horizontal diameters. The vertical dotted lines show the margins of the rim areas. CDR, cup:disc ratio; N, nasal; RNFL, retinal nerve fibre layer; T, temporal.

. The upper horizontal dotted lines represent the horizontal offset (150 µm) used to determine cup diameters and the lower horizontal dotted lines indicate disc horizontal diameters. The vertical dotted lines show the margins of the rim areas. CDR, cup:disc ratio; N, nasal; RNFL, retinal nerve fibre layer; T, temporal. Table 2 (A) Mean (±SD) of OCT optic nerve parameters in children with PCG and age-equivalent, gender-equivalent and ethnicity-equivalent healthy controls (the two groups were compared statistically using linear mixed models); (B) mean (±SD) of optic nerve parameters in adults with POAG and age-equivalent, gender-equivalent and ethnicity-equivalent healthy controls (the two groups were compared statistically using linear mixed models); (C) comparison of optic nerve parameters relative to their control values (patient values/mean control value×100%) in children with PCG and adults with POAG

s with POAG and age-equivalent, gender-equivalent and ethnicity-equivalent healthy controls (the two groups were compared statistically using linear mixed models); (C) comparison of optic nerve parameters relative to their control values (patient values/mean control value×100%) in children with PCG and adults with POAG (A) Children with PCG compared with controls PCG Controls Linear mixed model Mean±SD Mean±SD F P value Disc diameter (°) 5.596±0.851 6.014±0.578 3.28 0.078 Cup:disc ratio 0.668±0.173 0.398±0.178 26.46 <0.000 Cup diameter (°) 3.782±1.288 2.381±1.061 15.65 <0.000 Cup area (µm) 1491.3±778.2 503.6±451.7 25.93 <0.000 Cup depth (µm, from rim edges) 781.8±196.1 473.5±144.0 34.08 <0.000 Rim width (°) 1.814±0.947 3.633±1.198 31.87 <0.000 Rim area (µm) 215.0±212.4 590.3±309.5 21.62 <0.000 Nasal RNFL at 6° from disc centre (µm) 51.1±20.0 66.1±21.1 5.29 0.028 Temporal RNFL at 6° from disc centre (µm) 59.6±17.6 60.4±13.0 0.02 0.894 Age (years) 4.45±2.82 4.73±2.81 0.14 0.711 (B) Adults with POAG compared with controls Adult glaucoma Controls T-test Mean ± SD Mean ± SD F P value Disc diameter (°) 5.536±0.612 5.180±0.513 1.47 0.175 Cup:disc ratio 0.804±0.109 0.551±0.087 8.17 0.000 Cup diameter (°) 4.470±0.876 2.866±0.612 6.18 0.000 Cup area (µm) 1636±633.4 695±175.6 5.79 0.000 Cup depth (µm, from rim edges) 668.0±148.8 572.5±90.7 1.83 0.100 Rim width (°) 1.066±0.531 2.314±0.449 8.00 0.000 Rim area (µm) 79.03±85.6 294.8±128.4 5.61 0.000 Nasal RNFL at 6° from disc centre (µm) 46.0±9.08 57.8±18.19 1.86 0.096 Temporal RNFL at 6° from disc centre (µm) 47.9±8.95 50.4±5.1 0.74 0.479 Age (years) 67.0±6.96 66.8±7.50 0.07 0.949 (C) Comparison of optic nerve parameters relative to their control values PCG Adults Linear mixed model Mean ± SD Mean ± SD F P value Disc diameter (°) 93%±14% 107%±12% 8.64 0.006 Cup:disc ratio 168%±43% 146%±20% 1.83 0.186 Cup diameter (°) 159%±54% 156%±31% 0.00 0.966 Cup area (µm) 296%±155% 235%±91% 0.26 0.611 Cup depth (µm, from rim edges) 165%±41% 117%±26% 6.65 0.014 Rim width (°) 50%±26% 46%±23% 0.14 0.708 Rim area (µm) 36%±36% 27%±29% 0.95 0.337 Nasal RNFL at 6° from disc centre (µm) 83%±42% 76%±32% 0.39 0.538 Temporal RNFL at 6° from disc centre (µm) 99%±29% 97%±17% 1.01 0.323 OCT, optical coherence tomography; PCG, primary congenital glaucoma; POAG, primary open-angle glaucoma; RNFL, retinal nerve fibre layer.

4 0.708 Rim area (µm) 36%±36% 27%±29% 0.95 0.337 Nasal RNFL at 6° from disc centre (µm) 83%±42% 76%±32% 0.39 0.538 Temporal RNFL at 6° from disc centre (µm) 99%±29% 97%±17% 1.01 0.323 OCT, optical coherence tomography; PCG, primary congenital glaucoma; POAG, primary open-angle glaucoma; RNFL, retinal nerve fibre layer. Differences in disc diameters were not significant for PCG and adult glaucoma (vs age-matched controls, table 2A and B); however, since changes deviated in opposite directions, the relative differences were significant, with PCG apparently having smaller disc diameters than controls and adults with POAG having larger disc diameters (table 2C). Optimum threshold ON HH-OCT parameters were used to generate sensitivity, specificity, positive predictive values (PPVs) and negative predictive values (NPVs) for OCT detection of clinically diagnosed PCG (figure 3). Sensitivity values for the five ON HH-OCT morphological categories CDR, cup diameter, cup depth, and rim width and rim area were respectively 81.5, 74.1, 81.5, 88.9% and 88.9%. Respective specificity values were 80.0, 70.0, 85.0, 80.0% and 75.0% (figure 3). These are presented in table 3 along with 95% CIs.

(figure 3). Sensitivity values for the five ON HH-OCT morphological categories CDR, cup diameter, cup depth, and rim width and rim area were respectively 81.5, 74.1, 81.5, 88.9% and 88.9%. Respective specificity values were 80.0, 70.0, 85.0, 80.0% and 75.0% (figure 3). These are presented in table 3 along with 95% CIs. Figure 3 (A–I) Distribution of optic nerve head parameters in patients with primary congenital glaucoma (PCG) and healthy controls. Horizontal dotted lines show optimal sensitivity and specificity thresholds for the optic nerve parameters that were highly sensitive and specific for PCG detection. BM, Bruch’s membrane; RNFL, retinal nerve fibre layer. (K) Optimal sensitivity and specificity thresholds, sensitivity and specificity analysis for optic nerve head parameters in patients with primary congenital glaucoma compared with healthy controls. Table 3 Optic nerve threshold values which optimise sensitivities, specificities, and positive and negative predictive values (with 95% CIs) Threshold Sensitivity (%) Specificity (%) Positive predictive value (%) Negative predictive value (%) Value LCI UCI Value LCI UCI Value LCI UCI Value LCI UCI Cup:disc ratio 0.5 81.5 61.2 93.7 80.0 56.3 94.3 84.6 69.2 93.1 76.19 58.5 87.9 Cup diameter (°) 2.8 74.1 53.7 88.9 70.0 45.7 88.1 76.9 62.2 87.1 66.7 49.9 80.1 Cup depth (from rim edges) 600 81.5 61.9 93.7 85.0 62.1 96.8 88.0 71.8 95.5 77.3 60.1 88.5 Rim width (°) 2.8 88.9 70.8 97.7 80.0 56.3 94.3 85.7 71.2 93.6 84.2 64.2 94.1 Rim area (µm) 400 88.9 70.8 97.7 75.0 50.9 91.3 82.8 69.0 91.2 83.3 62.6 93.7 LCI, lower CI;UCI, upper CI.

70.0 45.7 88.1 76.9 62.2 87.1 66.7 49.9 80.1 Cup depth (from rim edges) 600 81.5 61.9 93.7 85.0 62.1 96.8 88.0 71.8 95.5 77.3 60.1 88.5 Rim width (°) 2.8 88.9 70.8 97.7 80.0 56.3 94.3 85.7 71.2 93.6 84.2 64.2 94.1 Rim area (µm) 400 88.9 70.8 97.7 75.0 50.9 91.3 82.8 69.0 91.2 83.3 62.6 93.7 LCI, lower CI;UCI, upper CI. Foveal morphology On visual inspection (figure 1C bottom), the ELM was not visible in any of the patients with PCG and did not allow distinguishing between ONL and IS. Analysis of individual retinal layers of the central horizontal B-scan did not reveal any significant changes except the thickening in the INL pericentrally in the temporal retina in the PCG group as compared with controls (F=5.55; p=0.024). The foveal width in the PCG group was smaller (87%; 8.38±1.00 and 9.62±0.65, p<0.001) than in controls. Discussion This study represents the largest cohort of atients with PCG that have been imaged using HH-OCT to date. OCT was successful in 83.3% eyes of the patients with PCG without sedation or anaesthesia.

Analysis of individual retinal layers of the central horizontal B-scan did not reveal any significant changes except the thickening in the INL pericentrally in the temporal retina in the PCG group as compared with controls (F=5.55; p=0.024). The foveal width in the PCG group was smaller (87%; 8.38±1.00 and 9.62±0.65, p<0.001) than in controls. Discussion This study represents the largest cohort of atients with PCG that have been imaged using HH-OCT to date. OCT was successful in 83.3% eyes of the patients with PCG without sedation or anaesthesia. In our study, children with PCG had significantly deeper cups than patients with POAG underlining the importance of measuring cup depth in addition to CDR. Changes of collagen fibril diameter and increased elasticity of the lamina cribosa early in life have been found in animal experiments.13 Possibly increased elasticity of the lamina cribrosa in childhood leads to increased deepening of the cup in PCG. Differences in collagen elasticity have also been postulated to explain reversing of cupping after surgical reduction of IOP in PCG.14 We did not find correlation of the IOP on the day of OCT acquisition with cup depth or other OCT parameters; however, this study did not include serial measurements of particular individuals which we, like others,14 have found to show reversibility of cupping on OCT, and most of our patients with PCG had controlled IOP at the time of imaging (mean IOP was 17.68 mm Hg).

f OCT acquisition with cup depth or other OCT parameters; however, this study did not include serial measurements of particular individuals which we, like others,14 have found to show reversibility of cupping on OCT, and most of our patients with PCG had controlled IOP at the time of imaging (mean IOP was 17.68 mm Hg). Standard ON assessment of adult patients with glaucoma includes the measurement of the average RNFL thickness of the peripapillary area on OCT. In our study, we have found significantly thinner RNFL in PCG in the nasal segment only using one single horizontal B-scan. It is likely that volumetric peripapillary RNFL thickness analysis involving all quadrants around the ON is more sensitive for assessment of ON damage in PCG and POAG. RNFL thinning starts with the inferior and superior segments and then involves nasal and temporal parts in adults.15 16 Therefore, nasal and temporal segments might be suboptimal to reflect glaucomatous changes. Similarly, assessment of vertical cup diameter might show earlier changes in PCG. However, compliance with acquiring vertical scans is more difficult due to the necessity of wide opening of the upper lid. In a few studies investigating RNFL thinning in patients with juvenile glaucoma, only the difference in superior and inferior segments was statistically significant compared with controls.17 18 Due to fixation instability of young children and infants, it is difficult to obtain volumetric OCT scans covering all quadrants around the ON. A refined faster acquisition protocol with fewer B-scans has good potential to improve compliance with vertical and volumetric scans and to further increase specificity and sensitivity of OCT to detect PCG (unpublished data).

infants, it is difficult to obtain volumetric OCT scans covering all quadrants around the ON. A refined faster acquisition protocol with fewer B-scans has good potential to improve compliance with vertical and volumetric scans and to further increase specificity and sensitivity of OCT to detect PCG (unpublished data). CDR measured by OCT in adult patients with glaucoma was found to be larger as compared with clinical assessment.19 CDR was significantly larger on OCT than judged on funduscopy in our patients with PCG. This could be explained by differences in assessment such as judgement of an area on clinical examination compared with measurement of a single horizontal B-scan. Other possibilities to explain the smaller CDR on clinical examination are obscuration of the cup by vessels and/or assessment of the cup size at a deeper level on funduscopy than by using OCT where a default setting of 150 µm above the Bruch’s membrane was used. It is likely that ON parameters obtained using OCT are more reliable, at least longitudinally, than clinical judgement using ophthalmoscopy. However, as with RNFL, it would be preferable for the CDR to be assessed for the entire nerve.

by using OCT where a default setting of 150 µm above the Bruch’s membrane was used. It is likely that ON parameters obtained using OCT are more reliable, at least longitudinally, than clinical judgement using ophthalmoscopy. However, as with RNFL, it would be preferable for the CDR to be assessed for the entire nerve. Cup depth, cup diameter, CDR and rim width/area show useful sensitivity (74.1%–88.9%) and specificity (70.0%–85.0%) parameters for PCG detection. The thresholds of 600 µm, 2.5°, 0.5 and 2.8/400 µm, respectively, found in our study could be selected for screening children suspected of glaucoma with OCT. Cup depth was the most specific parameter (85%). Moreover, it is one of the easiest parameters to analyse as it can be measured using callipers available on the HH-OCT machine without the need of segmentation.

µm, respectively, found in our study could be selected for screening children suspected of glaucoma with OCT. Cup depth was the most specific parameter (85%). Moreover, it is one of the easiest parameters to analyse as it can be measured using callipers available on the HH-OCT machine without the need of segmentation. It is important to recognise the limitations of our analysis of HH-OCT as a diagnostic tool. We used HH-OCT to identify childhood glaucoma in a specific clinical cohort with disease which differs from non-selected screening; this cohort has moderate to severe surgical PCG and differs from a non-targeted screening population in prevalence and severity of disease. The parameters we describe (sensitivity, specificity, PPV, NPV) apply to this cohort. We chose to study PCG (vs controls) because here, early-onset disease allowed us to characterise ON morphology at younger ages and the aggressive nature of the disease gives a higher prevalence of those structural changes. This allows the study to establish feasibility and characterise the differences between the groups. The diagnostic accuracy of HH-OCT or any diagnostic test is contextual,20 and accuracy parameters (and optimised threshold values) would differ in studies of different cohorts. We advocate studies, where the objective, using STARD20 methodologies, is to establish whether earlier diagnosis with HH-OCT would ultimately improve outcome.

accuracy of HH-OCT or any diagnostic test is contextual,20 and accuracy parameters (and optimised threshold values) would differ in studies of different cohorts. We advocate studies, where the objective, using STARD20 methodologies, is to establish whether earlier diagnosis with HH-OCT would ultimately improve outcome. The clinical presentation of PCG differs from POAG in that the disease is usually symptomatic and the diagnosis is more often obvious to practitioners familiar with the condition, so the need for a tool for screening is less. We foresee that the primary use of HH-OCT in PCG is for disease monitoring. This study into the diagnostic use of HH-OCT informs which ON head parameters are most likely to yield useful disease monitoring information in these children and to prioritise those parameters in imaging algorithms: cup depth is promising with highest specificity and PPV of the ON parameters studied (85% and 88%, respectively). We would advocate a longitudinal cohort study to evaluate HH-OCT disease monitoring of PCG and suggest that those ON parameters presented here might inform such work. The use of HH-OCT as a diagnostic tool in childhood glaucomas will likely be greatest in non-PCG glaucomas which can present relatively insidiously, such as distinguishing GCFCs (glaucoma following cataract surgery/aphakic glaucoma) from ocular hypertension in this group.

rameters presented here might inform such work. The use of HH-OCT as a diagnostic tool in childhood glaucomas will likely be greatest in non-PCG glaucomas which can present relatively insidiously, such as distinguishing GCFCs (glaucoma following cataract surgery/aphakic glaucoma) from ocular hypertension in this group. In a recent study, outer retinal layer changes and isolated INL cysts were found in a heterogeneous group of patients with paediatric glaucoma as well as non-glaucomatous optic atrophy.9 We have not observed cysts or changes in the outer and inner retinal layers except the absence of the ELM in all patients with PCG. Thickening of the INL in our study confirmed the results of Silverstein et al.21 However, missing ELM have not been described previously and could be caused by pathological retinal development and/or photoreceptor damage due to the raised IOP in PCG. ELM loss has been proposed to be a feature of early photoreceptor damage and associated with reduced blood supply to the retina.22 23 It is possible that the ELM, which is a structure with high oxygen demand, is more vulnerable at a young age when retinal development is not complete. Therefore, hypoxia induced by increased IOP could trigger retinal damage in PCG. Another possibility causing ELM changes could be reduced scan quality due to signal loss with corneal oedema/scarring.

hich is a structure with high oxygen demand, is more vulnerable at a young age when retinal development is not complete. Therefore, hypoxia induced by increased IOP could trigger retinal damage in PCG. Another possibility causing ELM changes could be reduced scan quality due to signal loss with corneal oedema/scarring. We found significant reduction of the foveal pit width. It is reported in the literature that patients with POAG have progressive retinal thinning which is related to GCL loss.24 25 We have not found significant GCL thinning; however, statistically significant pit changes could be an early sign of GCL loss on OCT. The fovea is developing dramatically in the first years of life with inner retinal layers moving away from the fovea.12 Possibly the increase in pressure early in life changes the dynamics of foveal development. The number of patients was relatively small and ON and foveal analyses were based only on single horizontal B-scan images rather than volumetric analysis to allow fast scans in children with poor co-operation. Consequently, the described changes do not represent all the structural abnormalities in patients with PCG that could be captured by volumetric analysis. Vertical scans are probably more likely to detect early changes, as is seen in POAG. In conclusion, our study shows that HH-OCT in children with PCG is feasible in a large percentage of children without sedation or anaesthesia. ON changes in children showed increased CDR, similar to POAG; however, excavation of the cup was deeper in children.

The number of patients was relatively small and ON and foveal analyses were based only on single horizontal B-scan images rather than volumetric analysis to allow fast scans in children with poor co-operation. Consequently, the described changes do not represent all the structural abnormalities in patients with PCG that could be captured by volumetric analysis. Vertical scans are probably more likely to detect early changes, as is seen in POAG. In conclusion, our study shows that HH-OCT in children with PCG is feasible in a large percentage of children without sedation or anaesthesia. ON changes in children showed increased CDR, similar to POAG; however, excavation of the cup was deeper in children. HH-OCT has the potential to be a useful tool for detecting glaucoma in young children. These findings form the groundwork for an expanded prospective study powered to evaluate HH-OCT for both disease monitoring in PCG and disease detection in non-PCG cohorts. Contributors: All authors are justifiably credited with authorship, according to the authorship criteria. AVP contributed to the conception, design, analysis and interpretation of data, drafting of the manuscript and critical revision before final approval. SS, VS and RP were responsible for the acquisition, analysis and finally approved the manuscript. FAP contributed to the design, analysis, interpretation of data and finally approved the manuscript. JA and IG contributed to the conception, design and critical revision before final approval.

ision before final approval. SS, VS and RP were responsible for the acquisition, analysis and finally approved the manuscript. FAP contributed to the design, analysis, interpretation of data and finally approved the manuscript. JA and IG contributed to the conception, design and critical revision before final approval. Funding: Supported by a Medical Research Council Grant (MR/J004189/1 and MR/N004566/1) and the Ulverscroft Foundation. Disclaimer: The sponsors or funding organisation had no role in the design or conduct of this research. Competing interests: None declared. Patient consent for publication: Parental/guardian consent obtained. Ethics approval: The study adhered to the tenets of the Declaration of Helsinki and was approved by the local ethics committee. Provenance and peer review: Not commissioned; externally peer reviewed.

Introduction Rare diseases are challenging Rare diseases are defined by the European Commission as conditions which affect fewer than 1 in 2000 people, and in America as one which affects less than 200 000 people.1 Rare diseases affect 3.5 million people in the UK, 30 million people across Europe, 25–30 million people in the USA and 350 million people worldwide.2 To put this in perspective, 2.5 million people are living with cancer in the UK,3 and in 2015, a total of 101 200 people were diagnosed with HIV.4 Cumulatively, this is still less than the 1 in 17 people affected by a rare disease in the UK. While individual disorders are rare, collectively rare diseases pose a significant healthcare problem. These conditions are often life threatening and debilitating with ~30% of these patients with rare disease dying before they reach their fifth birthday.5 A significant issue for patients is in obtaining a diagnosis, with two in five individuals struggling to obtain an accurate and timely diagnosis, often encumbered by several misdiagnoses before the correct one is obtained.6 Even where a diagnosis is obtained, knowledge about a disease is often extremely limited by primary healthcare professionals, with treatment options even more so.

two in five individuals struggling to obtain an accurate and timely diagnosis, often encumbered by several misdiagnoses before the correct one is obtained.6 Even where a diagnosis is obtained, knowledge about a disease is often extremely limited by primary healthcare professionals, with treatment options even more so. Early diagnosis and intervention is crucial for ophthalmic diseases For many ophthalmic conditions, early diagnosis is crucial to preventing further degeneration in patient eyesight, as well as for optimal management and support of the patient. Of the approximately 285 million cases of visual impairment worldwide, including both common and rare diseases, four out of five cases are thought to be preventable. The global initiative ‘VISION 2020: The Right To Sight’ seeks to reduce avoidable blindness at 25% by 2019 by collecting evidence on visual impairment, training of eye care professionals, provision of eye care and elimination of social and economic barriers to eye care.7

out of five cases are thought to be preventable. The global initiative ‘VISION 2020: The Right To Sight’ seeks to reduce avoidable blindness at 25% by 2019 by collecting evidence on visual impairment, training of eye care professionals, provision of eye care and elimination of social and economic barriers to eye care.7 The lack of existing rare ophthalmic disease registries to provide a comprehensive and exact figure of the total number of rare eye disorders presents a challenge to researchers hoping to collate what evidence is available. However, a large number of conditions have been documented by several sources, including: The 100,000 Genomes Project ophthalmic inclusion criteria,8 the Genetic and Rare Diseases Information Center,9 the National Eye Institute,10 the Retina International and the Irish Target 5000—Gateway to Vision.11 These conditions range from rod/cone abnormalities, macular dystrophies, photoreceptors and bipolar cell abnormalities, vitreoretinopathies and hereditary choroidal diseases.12

ic and Rare Diseases Information Center,9 the National Eye Institute,10 the Retina International and the Irish Target 5000—Gateway to Vision.11 These conditions range from rod/cone abnormalities, macular dystrophies, photoreceptors and bipolar cell abnormalities, vitreoretinopathies and hereditary choroidal diseases.12 Diabetic retinopathy, a leading case of vision loss and adult blindness worldwide, is an example of a common complex ophthalmic disease where early diagnosis and treatment reduces the risk of blindness. This is evident in the success eye screening programmes and increased glycaemic control have had in reducing adult blindness caused by diabetic retinopathy in England and Wales, with inherited retinal dystrophies now taking over as the leading cause of adult blindness.13 Furthermore, eye screening programmes in Iceland have reduced diabetic retinopathy blindness from 2.4% to 0.5% between 1980 and 1994.14 Therefore, the identification of biomarkers that can help develop early diagnostic tests, stratify patients and predict disease progression is urgently required for other ophthalmic diseases, especially rare diseases, which can cause vision loss and do not have existing screening programmes. One such progressive rare ophthalmic disease is Leber congenital amaurosis (LCA), which begins in childhood and vision deteriorates as the patient ages due to a lack of early diagnosis and treatment. Treatment options are limited for the majority of rare ophthalmic diseases, but recent gene therapy trials have shown promising results in restoring vision to nearly blind patients with LCA and in other inherited retinopathies, such as choroideraemia.15

s the patient ages due to a lack of early diagnosis and treatment. Treatment options are limited for the majority of rare ophthalmic diseases, but recent gene therapy trials have shown promising results in restoring vision to nearly blind patients with LCA and in other inherited retinopathies, such as choroideraemia.15 Next-generation multiomic analysis of rare disease Multiomics is a biological approach to research which looks at multiple ‘omic’ terms, of which there are currently over 500.16 One ‘omic’ term, genomics, has been a key player in ophthalmic disease research; including the identification of disease-related genes and genomic areas in age-related macular degeneration, diabetic retinopathy and glaucoma.17 Crucially, whole-genome and whole-exome sequencing have enabled the prediction that as many as 3500 genes thought to cause rare disease (including rare ophthalmic disease) might be identified.18 Projects such as the UK 100,000 Genomes Project have been fundamental in introducing genomics into the healthcare system, with ongoing multiomics contributing to maximising diagnostic yield and improving our understanding of rare disease.8 This project aims to identify genomic variation such as a copy number variation and single nucleotide polymorphisms to provide diagnosis where patients have not received one already, improve our understanding of disease pathogenesis and identify novel treatment targets.

and improving our understanding of rare disease.8 This project aims to identify genomic variation such as a copy number variation and single nucleotide polymorphisms to provide diagnosis where patients have not received one already, improve our understanding of disease pathogenesis and identify novel treatment targets. The National Ophthalmic Disease Genotyping and Phenotyping Network (eyeGENE) is an American/Canadian research body, created by the National Eye Institute, with the central aim of improving understanding of rare ophthalmic diseases, providing clinical and molecular diagnosis to patients and ultimately identification of therapeutic targets.19 As of December 2017, eyeGENE had an encouraging 130 publications, many of which were studies researching genetic mutations in rare ophthalmic disease.20 These publications include three review articles which highlight the promise for future retinal disease research using next-generation sequencing technologies; thus identifying a need to combine classical genetic mutation research with these high-throughput techniques to generate and analyse data on ophthalmic diseases.21–23 This fits with an expansion of research characterising several human diseases, including ophthalmic disease, at the molecular level using next-generation sequencing technology and a more informative integrated multiomic analysis.12

ese high-throughput techniques to generate and analyse data on ophthalmic diseases.21–23 This fits with an expansion of research characterising several human diseases, including ophthalmic disease, at the molecular level using next-generation sequencing technology and a more informative integrated multiomic analysis.12 While genomic research has been the primary focus of ophthalmic research and indeed many other areas of human disease, several different types of ‘omic’ analyses are increasingly conducted in ophthalmic disease research. These include transcriptomics (the study of sum RNA levels),24 metabolomics (study of tissue metabolites),25 proteomics (protein analysis)26 and epigenomics (the study of variations in the DNA which are not a result of sequence-level modifications).27

analyses are increasingly conducted in ophthalmic disease research. These include transcriptomics (the study of sum RNA levels),24 metabolomics (study of tissue metabolites),25 proteomics (protein analysis)26 and epigenomics (the study of variations in the DNA which are not a result of sequence-level modifications).27 Methylation: a key epigenomic marker of human disease Methylation, an important epigenomic biomarker, is a reversible chemical modification often seen at cytosine residue in CpG dinucleotide sequences in DNA. Specifically, it occurs when DNA methyltransferases transfer a methyl group to the fifth carbon position on a cytosine from S-adenyl methionine, forming 5-methylcytosine.28 Changes in DNA methylation may result from somatic or environmental factors as well as heritable factors and these changes can persist both in the long or short term, with the potential for transgenerational inheritance. It is associated with repressive effects on gene activity; however, the function of DNA methylation has been shown to vary depending on the genomic situation.29 Methylation has been investigated as a marker of many diseases including several cancers,30 metabolic disorders such as hyperglycaemia,31 neurological conditions such as autism spectrum disorder32 and, importantly, several common eye conditions such as diabetic retinopathy,33 age-related macular degeneration34 and cataracts.35

ion has been investigated as a marker of many diseases including several cancers,30 metabolic disorders such as hyperglycaemia,31 neurological conditions such as autism spectrum disorder32 and, importantly, several common eye conditions such as diabetic retinopathy,33 age-related macular degeneration34 and cataracts.35 Therefore, in addition to the classical genetic and genomic research described above, there is a need to assess the potential for methylation as an epigenomic marker of rare ophthalmic disease. If methylation was found to be an accurate and sensitive epigenomic marker, it has the potential to contribute to early diagnosis, management and reduction of vision loss; as well as identify novel therapeutic targets to improve treatment options for patients with rare ophthalmic diseases. Aim and objectives This review aimed to determine what evidence currently exists of differential methylation in rare eye diseases by: Systematically uncovering what, if any, current literature exists relevant to methylation and rare ocular diseases. Determining the method of methylation measurement and methodological rigour of any such studies. Highlighting any existing discussion of applying differential methylation as a biomarker of rare eye disease. Methods This review was conducted using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Transparent Reporting of Systematic Reviews and Meta-Analyses Checklist,36 following our registered protocol, PROSPERO ID: CRD42018094231.37

Highlighting any existing discussion of applying differential methylation as a biomarker of rare eye disease. Methods This review was conducted using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Transparent Reporting of Systematic Reviews and Meta-Analyses Checklist,36 following our registered protocol, PROSPERO ID: CRD42018094231.37 Eligibility criteria Journal articles published before 20 August 2019 were included if they were quantitative, written in English and discussed both differential methylation and rare eye diseases. This is an updated search from the original date specified in the registered protocol, though all search terms and eligibility criteria remained the same.37 The focus of this review was non-tumorous eye diseases; therefore retinoblastoma, uveal melanoma and intraocular melanoma were excluded. Conditions were also excluded if they impacted the eye but are not primarily classed as an eye disease (eg, CHARGE (Coloboma of the eye, Heart defects, Atresia of the choanae, Retardation of Growth and development, and Ear abnormalities and deafness syndrome), Behcet’s syndrome). Fuchs endothelial corneal dystrophy (FECD) was included, as the prevalence of this condition is widely variable geographically and thus is listed as a rare disease on Orphanet, ORPHA ID: 98 974.

esia of the choanae, Retardation of Growth and development, and Ear abnormalities and deafness syndrome), Behcet’s syndrome). Fuchs endothelial corneal dystrophy (FECD) was included, as the prevalence of this condition is widely variable geographically and thus is listed as a rare disease on Orphanet, ORPHA ID: 98 974. Information sources and search terms Four online databases were searched for identification of articles: MEDLINE, EMBASE, PubMed and Cochrane Database of Systematic Reviews. Grey literature was also searched: OpenGrey and GreyLit. Search terms were generated initially for use in MEDLINE and adapted for the other databases (online supplementary table S1). The population-intervention-comparator-outcome framework was used to develop these search terms: ‘population’ being non-tumorous rare eye diseases, ‘intervention’ being measurement of methylation, no ‘comparator’ was used and ‘outcome’ being implications of differential methylation for rare eye disease research.38 As there is no existing registry of rare eye conditions, search terms were curated by referring to the Genetic and Rare Diseases Information Center, the National Eye Institute, the Retina International and the Irish Target 5000—Gateway to Vision. Additionally, rare eye conditions which met ‘The 100,000 Genomes Project’ ophthalmology inclusion criteria were also included. 10.1136/bmjophth-2019-000342.supp1Supplementary data

Information sources and search terms Four online databases were searched for identification of articles: MEDLINE, EMBASE, PubMed and Cochrane Database of Systematic Reviews. Grey literature was also searched: OpenGrey and GreyLit. Search terms were generated initially for use in MEDLINE and adapted for the other databases (online supplementary table S1). The population-intervention-comparator-outcome framework was used to develop these search terms: ‘population’ being non-tumorous rare eye diseases, ‘intervention’ being measurement of methylation, no ‘comparator’ was used and ‘outcome’ being implications of differential methylation for rare eye disease research.38 As there is no existing registry of rare eye conditions, search terms were curated by referring to the Genetic and Rare Diseases Information Center, the National Eye Institute, the Retina International and the Irish Target 5000—Gateway to Vision. Additionally, rare eye conditions which met ‘The 100,000 Genomes Project’ ophthalmology inclusion criteria were also included. 10.1136/bmjophth-2019-000342.supp1Supplementary data Study selection, data extraction and critical appraisal Searches and data extraction were last performed in duplicate by two independent personnel on 20 August 2019. Duplicate articles were first removed automatically by reference management software EndNote V.X8. Articles were then screened by publication date, title, abstract and keywords to identify relevant studies and remove any duplicates missed by the reference management software. The remaining articles were screened by reading the full text. Reference lists were screened for reverse citations and a ‘cited reference search’ was performed on Web of Science to identify any potential relevant forward citations. Data extraction was performed, and methodological rigour critically appraised using customised forms based on checklists from the Joanna Briggs Institute critical appraisal tools (online supplementary tables S2 and S3).39 Methodological rigour scoring decisions of weak, moderate and strong were based on number of participants, appropriate matching of cases to controls, use of experimental controls and addressing confounding factors. Insufficient homogenous data were available to perform a meta-analysis so a narrative synthesis of the data was undertaken, using the University of Lancaster narrative synthesis guidelines.40

ber of participants, appropriate matching of cases to controls, use of experimental controls and addressing confounding factors. Insufficient homogenous data were available to perform a meta-analysis so a narrative synthesis of the data was undertaken, using the University of Lancaster narrative synthesis guidelines.40 Results Sources identified per databases were as follows: MEDLINE (n=22), EMBASE (n=75), PubMed (n=32), Cochrane Database of Systematic Reviews (n=3), OpenGrey (n=0) and GreyLit (n=0). No further papers were identified following forward/reverse citation searching. Screening of the 132 sources via title, abstract and keywords resulted in the removal of 19 duplicates and 94 sources were found not to be relevant. This left 19 papers for full-text screening, of which five were found to not be primary studies of methylation in rare ophthalmic disease. Therefore, 14 papers met the inclusion/exclusion criteria for final inclusion in the review. This screening and selection process is summarised in figure 1. Eight of these papers were case–control studies and three were case reports, and three were in vivo interventional studies using murine models. Study characteristics extracted included research aims, methylation measurement, participant information and key results (online supplementary table S4). Methodological rigour was scored as weak for each study based on a variety of factors, including low numbers of participants, a lack of strategies to deal with confounding factors or a lack of appropriately matched experimental controls used. Countries of publication origin included the USA, Japan, Poland, the UK, China, Sweden, India, Germany, Spain, Denmark and South Korea. A marked increased was seen in the number of studies of methylation and rare ophthalmic disease in recent years, with 10/14 (71.4%) studies published within the last decade, of which 40% were published since 2018. A succinct summary of ophthalmic conditions studied, methylation measurement method and key outcomes for these studies is presented in table 1.

mber of studies of methylation and rare ophthalmic disease in recent years, with 10/14 (71.4%) studies published within the last decade, of which 40% were published since 2018. A succinct summary of ophthalmic conditions studied, methylation measurement method and key outcomes for these studies is presented in table 1. Table 1 Study characteristics summary of sources included in review

mber of studies of methylation and rare ophthalmic disease in recent years, with 10/14 (71.4%) studies published within the last decade, of which 40% were published since 2018. A succinct summary of ophthalmic conditions studied, methylation measurement method and key outcomes for these studies is presented in table 1. Table 1 Study characteristics summary of sources included in review First author, publication year and reference number Rare eye disease Methylation measurement method Participant information Outcomes Case–control studies (n=8) Ali, 200446 CFEOM1 Bisulfite sequencing of CpG dinucleotides in exon 21 of KIF21A. Ten individuals in an Indian family affected by CFEOM1, with no other clinical abnormalities. Differential methylation within exon 21 of the KIF21A gene contributed to the high mutation rate. Bulka, 201947 Retinopathy of prematurity Infinium MethylationEPIC array of bisulfite-treated DNA. Premature infants with prethreshold retinopathy of prematurity compared with unaffected premature infants. Five probes were associated with elevated methylation and disease risk. Differential methylation was identified in eight autoinflammatory genes. Chan, 201854 FECD Infinium MethylationEPIC array data reanalysed looking at 2227 miRNA probes. Corneal endothelial tissue from patients with FECD compared with age and sex- matched patients without FECD. Differential methylation was seen in miRNA genes in FECD, including hypermethylation of miR-199b-5p which negatively regulates Snail and ZEB1. Jin, 200851 RP Methylation-specific PCR of bisulfite-treated DNA. Four patients with RP compared with three unaffected patients. No differential methylation patterns were observed between healthy controls and RP participants. Kabza, 201950 Keratoconus Reduced representation bisulfite sequencing. Corneal tissue from five patients with keratoconus and five unaffected controls. 18 differentially methylated genes corresponding to expression data were identified in patients with keratoconus. Khuc, 201744 FECD HM450 DNA methylation array kit and MethyLight real-time PCR. Corneal samples from patients with FECD (n=15) undergoing corneal translation surgery and four matched normal corneal endothelium. 59% and 41% of probes displayed hypermethylation and hypomethylation, respectively, in FECD patient samples. Maeng, 201543 GCD2 ChIP microarray analysis and MeDIA-assisted CpG microarray analysis. Corneal fibroblasts of four patients with GCD2 compared with normal corneal fibroblast controls (n=3).

corneal endothelium. 59% and 41% of probes displayed hypermethylation and hypomethylation, respectively, in FECD patient samples. Maeng, 201543 GCD2 ChIP microarray analysis and MeDIA-assisted CpG microarray analysis. Corneal fibroblasts of four patients with GCD2 compared with normal corneal fibroblast controls (n=3). Epigenetic modifications by histones affect expression of TGFBIp and extracellular matrix genes in corneal fibroblasts. Porter, 201542 Brittle cornea syndrome ChIP microarray and mass spectrometry. Cells from five patients with type 2 brittle cornea syndrome compared with six unaffected controls. HP1BP3 interactions with H3K9 methylated genomic regions were lost in mutated samples with mutated PRDM5 proteins. Case report studies (n=3) Friedrich, 199345 X linked RP Methylation analysis of the XLRP2 gene using a digoxigenin-labelled M27 beta probe. XLRP2-affected family including seven obligate carrier females and six daughters of obligate carriers with variable clinical phenotypes. No correlation between differential methylation of X chromosomes and clinical severity in patients with XLRP2. García‐Hoyos, 200548 Choroideraemia Methylation-specific PCR of bisulfite DNA. One female with choroideraemia, with a balanced chromosome X and 4 translocation. Non-random methylation patterns were seen resulting in skewed X inactivation in the participant with choroideraemia. Wei, 201552 Congenital cataract MRSE-PCR was used to measure promoter methylation. Monozygotic twin sisters, one with congenital pulverulent nuclear cataract and one without. Differential methylation was not found in candidate genes between siblings. In vivo interventional studies using murine models (n=3) Farinelli, 201453 RP Immunofluorescence, MeDIP, DNA methylation microarray. Murine models of RP, known as rd1 strains, of different strains used irrespective of gender and corresponding wild-type strains. Binding sites of transcription factors differentially methylated in RP mutant strains. DNMT inhibition reduced photoreceptor cell death. Montana, 201349 RP Bisulfite DNA was PCR amplified and CpG methylated detected on bisulfite sequencing DNA methylation analysis software. Murine models of RP with germ line Nrl and Pde6b knockouts. Knockouts of Nrl and Pde6b resulted in complete reprogramming in germ line mutant mice, and Nrl/Pde6b inactivation in adulthood resulted in partial reprogramming of red to cone cell transition, alleviating RP.

fite sequencing DNA methylation analysis software. Murine models of RP with germ line Nrl and Pde6b knockouts. Knockouts of Nrl and Pde6b resulted in complete reprogramming in germ line mutant mice, and Nrl/Pde6b inactivation in adulthood resulted in partial reprogramming of red to cone cell transition, alleviating RP. Zheng, 201841 RP Histone methylation measured by TMT labelling and HPLC and mass spectrometry. Murine models of RP and age-matched wild-type controls. Treatment with the methyltransferase inhibitor DZNep is followed by reduction in outer nuclear layer cell death. CFEOM1, congenital fibrosis of extraocular muscles-1; ChIP, chromatin immunoprecipitation; CpG, cytosine-guanine dinucleotide; DNMT, DNA methyltransferase; DZNep, 3-Deazaneplanocin A; FECD, Fuchs endothelial corneal dystrophy; GCD2, granular corneal dystrophy type 2; HM450, Illumina Infinium HumanMethylation450; HPLC, high-performance liquid chromatography; MeDIA, methylated DNA isolation assay; MeDIP, methylated DNA immunoprecipitation; miRNA, microRNA; MRSE, methylation-sensitive restriction enzyme; RP, retinitis pigmentosa; TMT, tandem mass tag labelling; XLRP2, X linked retinitis pigmentosa 2. Figure 1 Adapted from the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram, summarising screening method and study selection process.

Zheng, 201841 RP Histone methylation measured by TMT labelling and HPLC and mass spectrometry. Murine models of RP and age-matched wild-type controls. Treatment with the methyltransferase inhibitor DZNep is followed by reduction in outer nuclear layer cell death. CFEOM1, congenital fibrosis of extraocular muscles-1; ChIP, chromatin immunoprecipitation; CpG, cytosine-guanine dinucleotide; DNMT, DNA methyltransferase; DZNep, 3-Deazaneplanocin A; FECD, Fuchs endothelial corneal dystrophy; GCD2, granular corneal dystrophy type 2; HM450, Illumina Infinium HumanMethylation450; HPLC, high-performance liquid chromatography; MeDIA, methylated DNA isolation assay; MeDIP, methylated DNA immunoprecipitation; miRNA, microRNA; MRSE, methylation-sensitive restriction enzyme; RP, retinitis pigmentosa; TMT, tandem mass tag labelling; XLRP2, X linked retinitis pigmentosa 2. Figure 1 Adapted from the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram, summarising screening method and study selection process. Discussion Findings of this systematic review and potentials for future research Following a comprehensive systematic review of the literature, this study found that the role of methylation in ophthalmic disease has been evaluated in histone methylation marks,41–43 gene body and promoter regions,44–53 and miRNAs,54 in the following rare ophthalmic conditions: Brittle cornea syndrome, an autosomal recessive inherited disorder characterised by corneal fragility.42 Choroideraemia, an X linked recessive disorder of retinal degeneration.48

Discussion Findings of this systematic review and potentials for future research Following a comprehensive systematic review of the literature, this study found that the role of methylation in ophthalmic disease has been evaluated in histone methylation marks,41–43 gene body and promoter regions,44–53 and miRNAs,54 in the following rare ophthalmic conditions: Brittle cornea syndrome, an autosomal recessive inherited disorder characterised by corneal fragility.42 Choroideraemia, an X linked recessive disorder of retinal degeneration.48 Congenital cataract, a disease characterised by lens opacity at birth.52 Congenital fibrosis of extraocular muscles-1 (CFEOM1), a disorder characterised by ptsosis and ophthalmoplegia.46 FECD, a late-onset progressive inherited disorder (typically autosomal dominant inheritance) leading to loss of visual acuity.44 54 Granular corneal dystrophy type 2 (GCD2), a disease characterised by granular deposits in the cornea leading to progressive vision loss.43 Keratoconus, a condition characterised by distortion of the corneal surface.50 Retinitis pigmentosa (RP), a progressive disorder leading to loss of vision through retinal cell degeneration.41 45 49 51 53 Retinopathy of prematurity, a congenital disorder of retinal cell degeneration resulting from premature birth.47

Granular corneal dystrophy type 2 (GCD2), a disease characterised by granular deposits in the cornea leading to progressive vision loss.43 Keratoconus, a condition characterised by distortion of the corneal surface.50 Retinitis pigmentosa (RP), a progressive disorder leading to loss of vision through retinal cell degeneration.41 45 49 51 53 Retinopathy of prematurity, a congenital disorder of retinal cell degeneration resulting from premature birth.47 Aberrant methylation was established as a pathogenic biomarker of ophthalmic disease in eight studies. This included: the identification of methylation as playing a role in the high mutability of exon 21 within KIF21A in CFEOM1,46 loss of H3K9me2-mediated repression in brittle cornea syndrome,42 differential methylation of autoinflammatory genes in retinopathy of prematurity (including NGPT1, BDNF, CRP, MPO, SAA1, SAA2, TNFRSF1A and TNFRSF1B),47 differential methylation of several genes corresponding to RNA sequencing gene expression analysis in keratoconus,50 hypomethylation in genes related to ion channel roles which are essential for corneal endothelium function in patients with FECD,44 differential miRNA expression in FECD cases compared with controls including the hypermethylation of miR-199b-5p which negatively regulates the corneal transcription factors Snail and ZEB1,54 altered histone methylation status associated with changes in TGFBIp expression levels in GCD243 and, finally, skewed inactivation of the X chromosome associated with a non-random DNA methylation pattern in choroideraemia.48

n of miR-199b-5p which negatively regulates the corneal transcription factors Snail and ZEB1,54 altered histone methylation status associated with changes in TGFBIp expression levels in GCD243 and, finally, skewed inactivation of the X chromosome associated with a non-random DNA methylation pattern in choroideraemia.48 DNA methylation was evaluated as a potential novel therapeutic target in two studies of RP through treatment with methyltransferase inhibitors (decitabine and 3-Deazaneplanocin A) which were found to reduce RP phenotype progression.41 53 Methylation inhibitors have been investigated in recent years as therapeutic agents for cancer treatment, due to the role of hypermethylation as an inhibitor of tumour suppressor genes.55 Decitabine is a Food and Drug Administration (FDA)-approved drug for treatment of haematological malignancies, and while 3-Deazaneplanocin A is not currently FDA approved, it has been investigated for its role in inhibiting histone methylation in acute myeloid leukaemia and are therefore examples of drug repurposing opportunities for non-cancerous rare ophthalmic disease. Although retinoblastoma and uveal/intraocular melanoma are not included in this review, epigenetics and rare tumorous ophthalmic conditions have been reviewed extensively elsewhere within the last 2 years, with extensive studies illustrating the association of hypermethylation of tumour suppressor gene promoter regions and ophthalmic tumorigenesis.56 57 Therefore, the association of methylation with ophthalmic disease, coupled with attenuated retinal degeneration following methyltransferase inhibition in the two cited articles,41 53 highlights the need to conduct further investigations into epigenetic inhibitors as therapeutic agents for both tumorous and non-tumorous ophthalmic disease. Additionally, one study identified two candidate genes (NRL and PDE6B) for targeted gene therapy RP research, with the important caveat that DNA methylation must be considered for reprogramming the transition of adult rod photoreceptor cells to cone cells; adult murine models subjected to Nrl/Pde6b inactivation showed only partial reprogramming which was insufficient to alter DNA methylation patterns, and incomplete RP symptom alleviation.49

ant caveat that DNA methylation must be considered for reprogramming the transition of adult rod photoreceptor cells to cone cells; adult murine models subjected to Nrl/Pde6b inactivation showed only partial reprogramming which was insufficient to alter DNA methylation patterns, and incomplete RP symptom alleviation.49 Two studies investigating the potential role of aberrant methylation in congenital cataracts,52 and RP,51 found no significant difference in methylation levels between cases and controls in the candidate genes selected for analysis. However, in the case report of monozygotic twins discordant for congenital cataracts, the candidate gene CRYAA was identified from previous research into common conditions such as age-related cataract; there is a need for whole epigenome analysis to identify differentially methylated genes which may differ between common and rare ophthalmic conditions.35 For the second study, methylation analysis was able to aid the exclusion of a candidate gene as having a causative role in RP, where there was a random pattern of methylation and therefore no evidence of imprinting of mutated FSCN2.51 Furthermore, the case report on a family of individuals affected by X linked RP2 did not find an association between differential methylation in the X chromosome and clinical severity.45 However, as this study was conducted in 1993 using an M27 beta probe, it would be interesting to see whether modern comprehensive methods of measuring methylation might show a different result. Additionally, as is true for the majority of epigenetic studies, a longitudinal analysis of epigenetic changes over time would be valuable to determine that differential methylation is a causative factor of disease rather than an incidental change that naturally occurs over a lifetime.

on might show a different result. Additionally, as is true for the majority of epigenetic studies, a longitudinal analysis of epigenetic changes over time would be valuable to determine that differential methylation is a causative factor of disease rather than an incidental change that naturally occurs over a lifetime. Of the articles included in this review, 71.4% studies were published within the last decade and of these 40% were published since 2018, demonstrating the increasing awareness of the importance of epigenetics in ophthalmic disease and subsequent undertaking of research. However, the rare eye diseases reported in the existing literature are a small subset of existing rare conditions. Difficulties comprising search terms exhaustive of the total list of rare eye conditions reflect the need for a registry of rare eye conditions, as exists for rare renal conditions58 and which would be invaluable in future rare ophthalmic disease research. Countries of origin for each publication varied widely from the USA, Japan, Poland, the UK, China, Sweden, India, Germany, Spain, Denmark and South Korea. Building on this, it would be useful to perform comparative methylation analysis on ethnically diverse rare ophthalmic disease populations by encouraging international collaboration.

in for each publication varied widely from the USA, Japan, Poland, the UK, China, Sweden, India, Germany, Spain, Denmark and South Korea. Building on this, it would be useful to perform comparative methylation analysis on ethnically diverse rare ophthalmic disease populations by encouraging international collaboration. Limitations of this systematic review While a randomised controlled trial (RCT) with many participants is desirable, the scarcity of patients with a given rare disease, coupled with the fact that epigenetic analysis is a developing field, makes ‘gold-standard’ RCTs difficult for researchers to achieve. Therefore, it is unsurprising that one limitation of this review is that the studies included were all scored as having weak methodological rigour based on low participant numbers, lack of addressing confounding factors or insufficient experimental controls described. This review highlights the need for a standardised approach of conducting rare disease and multiomic research, such as the standardised reporting structure that exists for genetic association studies.59 Furthermore, several of these studies reported differential methylation in animal models,41 49 53 therefore it would be helpful for future research to perform studies of these differentially methylated genes in human biopsies before differential methylation is validated as a clinically useful method of characterising rare ophthalmic disease.

these studies reported differential methylation in animal models,41 49 53 therefore it would be helpful for future research to perform studies of these differentially methylated genes in human biopsies before differential methylation is validated as a clinically useful method of characterising rare ophthalmic disease. Clinical relevance of these findings As is typical of patients with rare conditions, patients with rare ophthalmic conditions can face difficulties obtaining a diagnosis, and treatment for their condition is often limited which can have severely detrimental impacts for a patient’s vision. Therefore, there is a need for improving our understanding of these conditions, to aid diagnosis, provide prognoses or identify therapeutic targets which could correct or halt degeneration in patients with rare eye dysfunction. Projects such as the National Eye Institute’s eyeGene, Target 5000—Gateway to Vision and The 100,000 Genomes Project are important steps towards achieving this goal. Building on this momentum and the advances of next-generation sequencing technology in the last decade, this review highlights how an allied multiomic approach, including genomics and epigenomics, to characterise rare ophthalmic disease is the next step in research. In particular, epigenomics is an encouraging place to start, having been reported as associated with in common eye conditions such as diabetic retinopathy, age-related macular degeneration and acquired cataracts.60

including genomics and epigenomics, to characterise rare ophthalmic disease is the next step in research. In particular, epigenomics is an encouraging place to start, having been reported as associated with in common eye conditions such as diabetic retinopathy, age-related macular degeneration and acquired cataracts.60 Conclusions Differential methylation has been identified in a small number of rare ophthalmic disease studies, emphasising the need for further research projects like eyeGENE, Target 5000—Gateway to Vision and The 100,000 Genomes Project. Yet the limited evidence that is available is promising, and emphasises the potential for molecular characterisation of rare eye disorders using methylation. Highlights include: Identification of increased methylation in the outer nuclear layers and reduced photoreceptor cell death following treatment with a DNA methyltransferase inhibitors decitabine and 3-Deazaneplanocin A in RP models. Reprogramming of adult rod cells through Nrl inactivation as a novel therapeutic target of RP in murine models. Differential methylation of genes identified related to corneal endothelium function in FECD independent of age or gender. Identification of epigenetic histone modifications as having a disease-causing role in GCD2.

Reprogramming of adult rod cells through Nrl inactivation as a novel therapeutic target of RP in murine models. Differential methylation of genes identified related to corneal endothelium function in FECD independent of age or gender. Identification of epigenetic histone modifications as having a disease-causing role in GCD2. These projects provide a platform for future multiomic approaches to study rare ophthalmic disease and the much needed development of a standardised analysis and reporting approach. Integrating epigenomic markers such as DNA methylation with genomic analysis has the potential to aid diagnosis, limit loss of patient vision and ultimately improve patient independence and quality of life. Contributors: AJM and HM conceptualised and initiated the review. KK drafted the review. All authors contributed to the manuscript and made revisions, and have read and approved the final manuscript. Funding: KK is supported by a Department for the Economy Co-operative Awards in Science and Technology (DfE-CAST) studentship award. Funding support was provided by the Medical Research Council-Northern Ireland Executive support of the Northern Ireland Genomic Medicine Centre through Belfast Health and Social Care Trust. Competing interests: None declared. Patient consent for publication: Not required. Provenance and peer review: Not commissioned; externally peer reviewed.

Significance of the study What is already known about this subject? Quality of life (QoL) is affected in many bilateral ocular conditions such as cataract, glaucoma and trichiasis. The impact of microbial keratitis (MK) on QoL compared to unaffected individuals has not been previously reported. MK is a common cause of blindness in Sub-Saharan Africa. What are the new findings? MK severely reduces QoL in the acute phase of the disease. QoL improves with treatment and healing. Despite improvement, QoL of someone affected by MK (even with normal vision) remains lower than unaffected controls. How might these results change the focus of research or clinical practice? The focus of this study is to make the case that MK, although usually a uniocular disease severely reduces the QoL of the affected individuals. QoL is affected in many bilateral ocular conditions such as cataract, glaucoma and trichiasis. Background Microbial keratitis (MK) has been described as a ‘silent epidemic’, which leads to substantial morbidity, related to sight loss, pain and stigma.1 It is the leading cause of unilateral blindness after cataract in tropical regions, estimated at 2 million cases of monocular blindness per year.2 3 1.3 million individuals were bilaterally blind from corneal opacity globally (excluding trachoma and vitamin A deficiency), accounting for 3.2% of binocular blindness.3

.1 It is the leading cause of unilateral blindness after cataract in tropical regions, estimated at 2 million cases of monocular blindness per year.2 3 1.3 million individuals were bilaterally blind from corneal opacity globally (excluding trachoma and vitamin A deficiency), accounting for 3.2% of binocular blindness.3 Quality of life (QoL) is a very important consideration in the management of any disease and treatments should ultimately aim to maintain or restore QoL.4 Few studies from sub-Saharan Africa (SSA) have examined the effect of cataract and trachomatous trichiasis on QoL.5–7 These have generally examined both the general health-related quality of life (HRQoL) and more specific vision-related quality of life (VRQoL). However, there is no published data on the impact of MK on QoL from SSA, and very little from other World regions. In the Mycotic Ulcer Treatment Trial 1 (MUTT1) in India which compared topical natamycin to topical voriconazole for the treatment of fungal keratitis, QoL scores at 3 months were compared between the two treatment arms. However, there was no comparison group of unaffected individuals.8 There is a need to better understand how MK and its outcomes affect people, to develop improved management, counselling and support. To investigate the impact of MK on QoL, we conducted this study in South-Western Uganda. Here, we describe the QoL among patients with MK, at presentation and 3 months after presentation compared with the QoL of unaffected individuals recruited from the community who were individually matched for age, sex and location of residence.

e the impact of MK on QoL, we conducted this study in South-Western Uganda. Here, we describe the QoL among patients with MK, at presentation and 3 months after presentation compared with the QoL of unaffected individuals recruited from the community who were individually matched for age, sex and location of residence. Methods Ethical statement This study adhered to the tenets of the Declaration of Helsinki. It was approved by the London School of Hygiene & Tropical Medicine Ethics Committee (Ref 10647), Mbarara University Research Ethics Committee (Ref 10/04–16) and Uganda National Council for Science and Technology (Ref HS-2303). Written, informed consent in the local language was obtained before enrolment. If the potential participant was unable to read, the information was read to them, and they were asked to indicate their consent by application of their thumbprint, which was independently witnessed.

for Science and Technology (Ref HS-2303). Written, informed consent in the local language was obtained before enrolment. If the potential participant was unable to read, the information was read to them, and they were asked to indicate their consent by application of their thumbprint, which was independently witnessed. Study design and participants This study of the impact of MK on QoL was nested within a case-control study of MK in Uganda. We prospectively enrolled patients with MK that presented to Ruharo Eye Centre and Mbarara University and Referral Hospital Eye Centre from December 2016 to March 2018. These are the tertiary referral centres for South-Western Uganda. The case definition of MK was the presence of a corneal epithelial defect (of at least 1 mm diameter) with an underlying stromal infiltrate, associated with signs of inflammation (conjunctival hyperaemia, anterior chamber inflammatory cells, ±hypopyon).9 We excluded those not willing to participate or to return for follow-up, pregnant women, lactating mothers and those under 18 years. All the questions in the tools were responded to directly by the study participants.

with signs of inflammation (conjunctival hyperaemia, anterior chamber inflammatory cells, ±hypopyon).9 We excluded those not willing to participate or to return for follow-up, pregnant women, lactating mothers and those under 18 years. All the questions in the tools were responded to directly by the study participants. Assessment of cases At presentation, we documented basic demographic information and ophthalmic history. Presenting Logarithm of Minimum Angle of Resolution visual acuity at 2 m in a dark room was measured using the Peek Acuity smartphone application.10 Cases were examined at a slit lamp to assess the anterior segment using a structured protocol. Corneal scrape specimens were collected from the ulcer at a slit lamp or an operating microscope and samples were processed in the department of microbiology laboratory at Mbarara University of Science and Technology. Following corneal scrapping, immediate Calcofluor White staining was done in the side laboratory at the eye hospital on a fluorescence microscope (Zeiss Primostar ILED) by the attending ophthalmologist to rule out fungal Keratitis. Additional microscopy (Gram staining and KOH staining) was done in the main University microbiology laboratory and results became available within 24 hours. Agar plates and broths (Blood, Chocolate, Potato Dextrose and Brain Heart) were incubated at 35°C–37°C for bacteria for up to 7 days and at 25°C for up to 21 days for fungi. Organism identification and sensitivity testing were performed using standard microbiological techniques. Cases were treated empirically at presentation and the treatment was reviewed when the microbiology results became available. Patients with fungal keratitis were treated with Natamycin 5% eyedrops (Zonat Sunways India), those with bacterial keratitis were treated with Ofloxacin 0.3% eyedrops (Biomedica Remedies-India). They were reviewed on days 2, 7, 21 and 90 (3 months). At 3 months, the cases were followed-up in their homes for a final assessment.

with fungal keratitis were treated with Natamycin 5% eyedrops (Zonat Sunways India), those with bacterial keratitis were treated with Ofloxacin 0.3% eyedrops (Biomedica Remedies-India). They were reviewed on days 2, 7, 21 and 90 (3 months). At 3 months, the cases were followed-up in their homes for a final assessment. Control recruitment We recruited healthy community controls during the 3-month follow-up of cases in their home village. The controls were individuals without any current eye complaints and with normal vision. The controls were individually matched to the cases. They had to be living in same village as the case, be of the same gender and in the same age group (±5 years). Enrolment followed a similar approach to that previously used in study in Ethiopia.6 The research team visited the villages (typically 50–100 households), the local village head was asked to write down all individuals they thought would meet the matching criteria for a particular case in that village. One person was randomly selected from this list using a lottery method. They were approached and provided with details of the study and invited to participate if eligible. If a selected individual refused or was ineligible, another person was randomly selected and approached. Both cases and controls were asked about their social economic status, this was a self-reported question compared with their neighbours on a 5-scale level (5—‘very rich’, 4—‘rich’, 3—‘neither rich nor poor’, 2—‘poor’, 1—‘very poor’).6 7 11 In this 5-point scale, participants were asked ‘Compared with your neighbours, how do you rate your household wealth status?’

r social economic status, this was a self-reported question compared with their neighbours on a 5-scale level (5—‘very rich’, 4—‘rich’, 3—‘neither rich nor poor’, 2—‘poor’, 1—‘very poor’).6 7 11 In this 5-point scale, participants were asked ‘Compared with your neighbours, how do you rate your household wealth status?’ QoL instruments To measure QoL we used two instruments: the general health WHOQOL-BREF and the vision-related WHO/PBD-VF20. BREF stands for abbreviated form of the original WHOQOL tool. These were both initially independently translated by two translators into Runyankole, the local language. Any discrepancies were discussed with a third party and a merged final agreed version produced. Both instruments were administered to cases at presentation and at the 3-month follow-up. They were administered to the control group only once, during the 3-month assessment of the matched case.

yankole, the local language. Any discrepancies were discussed with a third party and a merged final agreed version produced. Both instruments were administered to cases at presentation and at the 3-month follow-up. They were administered to the control group only once, during the 3-month assessment of the matched case. VRQoL the WHO/PBD-VF20 tool measures VRQoL. It assesses the impact of visual impairment in several domains including mental well-being, dependency and social functioning. The WHO/PBD-VF20 consists of 20 questions divided into four subscales: ‘General Vision’ subscale (one question); ‘Visual Symptoms’ subscale (three questions); ‘General Functioning’ subscale (12 questions); and ‘Psychosocial’ subscale (four questions). It begins by asking the patient ‘Overall, how would you rate your eyesight using both eyes?’; and uses a 5-point scale answer option such as ‘very good’, ‘good’, ‘moderate’, ‘bad’, ‘very bad’. Each subsequent question also has a 5-point response option: one indicates the highest and five the lowest score.

r questions). It begins by asking the patient ‘Overall, how would you rate your eyesight using both eyes?’; and uses a 5-point scale answer option such as ‘very good’, ‘good’, ‘moderate’, ‘bad’, ‘very bad’. Each subsequent question also has a 5-point response option: one indicates the highest and five the lowest score. HRQoL the WHOQOL-BREF (WHOQOL Group, 1998) has good applicability in low and middle-income countries as it was developed simultaneously from concept across 18 countries in Africa, Asia and Latin America.12 It measures four domains of health: Physical Health, Psychological Health, Social Relationships and Environment. It asks respondents 26 questions. These include the frequency they have experienced issues and/or were able to do things (eg, feel safe, able to concentrate, enjoy life) in the past 4 weeks and how satisfied they are with certain aspects of their lives (eg, sleep, capacity for work).12 Sample size Based on the effect sizes found in previous work on cataract and trichiasis, a sample size of 215 pairs would have 80% power to detect a moderate effect of MK on QoL with an effect size of 0.27 (effect size=QoL score difference (3)/SD 11) with a Type 1 error of 5%.5 7 Analysis Data were managed in Access (Microsoft), and transferred to Stata V.14 (StataCorp) for analysis. Data were analysed using a previously described methodology, applied in other QoL studies.6 7 13

Sample size Based on the effect sizes found in previous work on cataract and trichiasis, a sample size of 215 pairs would have 80% power to detect a moderate effect of MK on QoL with an effect size of 0.27 (effect size=QoL score difference (3)/SD 11) with a Type 1 error of 5%.5 7 Analysis Data were managed in Access (Microsoft), and transferred to Stata V.14 (StataCorp) for analysis. Data were analysed using a previously described methodology, applied in other QoL studies.6 7 13 VRQoL All items were grouped, and scores added into their respective subscales: ‘General Vision’ subscale (one question); ‘Visual Symptoms’ subscale (three questions); ‘General Functioning’ subscale (12 questions); and ‘Psychosocial’ subscale (four questions). The subscale scores were then converted into a scaled value out of one hundred, using the formula: ((individual score—lowest possible score)/(highest possible score—lowest possible score)) × 100. Therefore, the person with the lowest possible VRQoL score would receive a scaled value of ‘0’ and the person with the highest possible VRQoL score receives a scaled value of ‘100’.

ue out of one hundred, using the formula: ((individual score—lowest possible score)/(highest possible score—lowest possible score)) × 100. Therefore, the person with the lowest possible VRQoL score would receive a scaled value of ‘0’ and the person with the highest possible VRQoL score receives a scaled value of ‘100’. HRQoL Data were analysed following the WHOQOL-BREF protocol.12 Three negatively framed items were reversed into a positive frame so higher scores denote higher QoL. To generate domain scores, questions were grouped into their respective domains and their scores totalled. The mean score of all items included in the domain was calculated and then multiplied by four. These scores then transformed to a 0 to 100 scale with the formula specified in the manual to allow comparison between domains made of unequal number of items.14 Psychometric property evaluation Construct validity of the VRQoL and HRQoL data was assessed through known-group difference and convergence validity using a linear regression model. Cronbach’s alpha was used to test for internal consistency and reliability of the VRQoL and HRQoL data.

HRQoL Data were analysed following the WHOQOL-BREF protocol.12 Three negatively framed items were reversed into a positive frame so higher scores denote higher QoL. To generate domain scores, questions were grouped into their respective domains and their scores totalled. The mean score of all items included in the domain was calculated and then multiplied by four. These scores then transformed to a 0 to 100 scale with the formula specified in the manual to allow comparison between domains made of unequal number of items.14 Psychometric property evaluation Construct validity of the VRQoL and HRQoL data was assessed through known-group difference and convergence validity using a linear regression model. Cronbach’s alpha was used to test for internal consistency and reliability of the VRQoL and HRQoL data. Cases and controls were compared for baseline characteristics. However, we noticed that not all the pairs had been correctly matched for age because the village heads had subjectively guessed the ages of the controls. We thus adjusted for age throughout the analysis. The VRQoL and HRQoL analysis compared the cases to the controls at 3 months using a linear regression random effects model, which was adjusted for age and socioeconomic status, as these factors may confound the association between MK and QoL. A linear regression analysis was used to determine vision-related factors associated with QoL among the cases at 3 months adjusted for baseline QoL, age, sex, education and economic status.

ects model, which was adjusted for age and socioeconomic status, as these factors may confound the association between MK and QoL. A linear regression analysis was used to determine vision-related factors associated with QoL among the cases at 3 months adjusted for baseline QoL, age, sex, education and economic status. Patient and public involvement Apart from helping to provide information during the piloting and data collection phase, we did not explicitly involve patients or the public in the designing and implementation of our work Results A total of 313 MK cases presented and were enrolled. We were able to follow-up 260 cases at 3 months. It was not possible to enrol a control for 45/260 cases. Therefore, the analysis of QoL at 3 months comprises 215 pairs. The baseline characteristics were comparable among the cases and control group: median age was 47 years (IQR 35–60, total range 18–96 years), and 120 (56%) were male. (table 1) Table 1 Baseline characteristics among the 215 case–control pairs (matched on gender and village and adjusted for age)

Results A total of 313 MK cases presented and were enrolled. We were able to follow-up 260 cases at 3 months. It was not possible to enrol a control for 45/260 cases. Therefore, the analysis of QoL at 3 months comprises 215 pairs. The baseline characteristics were comparable among the cases and control group: median age was 47 years (IQR 35–60, total range 18–96 years), and 120 (56%) were male. (table 1) Table 1 Baseline characteristics among the 215 case–control pairs (matched on gender and village and adjusted for age) Exposure Cases (215) Controls (215) P value n (%) n (%) Married (yes) * 154 (72) 143 (67) 0.215 Head of household (yes) † 146 (68) 140 (65) 0.441 Education status ‡ None 60 (28) 48 (22) 0.148 Primary 110 (51) 114 (53) Secondary 31 (14) 32 (15) Tertiary 14 (7) 21 (10) Farmer (yes) ‡ 157 (73) 168 (78) 0.144 Size of the household § Small (1–4 people) 50 (23) 109 (51) 0.05 Medium (5–10 people) 115 (54) 94 (44) Large (>11 people) 50 (23) 12 (5) Self-reported wealth status ¶ Poor 36 (17) 20 (9) 0.003 Middle 158 (73) 188 (88) Upper 21 (10) 6 (3) *People who were married, or cohabiting were considered as married while those who were divorced, single or widowed were considered as not married. †Being head of the household meant people who were responsible for the overall care of the family, this was regardless of gender: among the cases and controls, 31% and 23% were female heads of households, respectively.

Exposure Cases (215) Controls (215) P value n (%) n (%) Married (yes) * 154 (72) 143 (67) 0.215 Head of household (yes) † 146 (68) 140 (65) 0.441 Education status ‡ None 60 (28) 48 (22) 0.148 Primary 110 (51) 114 (53) Secondary 31 (14) 32 (15) Tertiary 14 (7) 21 (10) Farmer (yes) ‡ 157 (73) 168 (78) 0.144 Size of the household § Small (1–4 people) 50 (23) 109 (51) 0.05 Medium (5–10 people) 115 (54) 94 (44) Large (>11 people) 50 (23) 12 (5) Self-reported wealth status ¶ Poor 36 (17) 20 (9) 0.003 Middle 158 (73) 188 (88) Upper 21 (10) 6 (3) *People who were married, or cohabiting were considered as married while those who were divorced, single or widowed were considered as not married. †Being head of the household meant people who were responsible for the overall care of the family, this was regardless of gender: among the cases and controls, 31% and 23% were female heads of households, respectively. ‡Majority of the participants had no or minimal education (primary level) which is not uncommon for a predominantly rural population in Uganda. Subsistence farming is the main occupation for this population. §Majority of the household sizes were medium to large (five people or more). This is not uncommon since most of the living in rural Uganda is largely in an extended family setting.

‡Majority of the participants had no or minimal education (primary level) which is not uncommon for a predominantly rural population in Uganda. Subsistence farming is the main occupation for this population. §Majority of the household sizes were medium to large (five people or more). This is not uncommon since most of the living in rural Uganda is largely in an extended family setting. ¶Self-reported wealth status was classified as poor (1, ‘very poor’ and 2 ‘poor’), middle (3, ‘neither poor nor rich’) upper (4, ’rich’ and 5, ‘very rich’). There was one missing value among the control group. Participants were asked to compare themselves to their neighbours and give a score of their economic status. Table 2 shows the baseline and 3-month VRQoL and HRQoL scores for the cases and the scores for the control group. The mean baseline VRQoL scores among the cases were all low (<50) except the psychosocial domain which had a score of 65 points. The most affected domain was visual symptoms, with a mean score of 20.7 (95% CI 18.8 to 22.7). The mean baseline HRQoL scores among the cases were all low (<50). At 3 months, all the case VRQoL and HRQoL scores had increased and were relatively high (between 80 and 91). Despite this increase, there was still very strong evidence (p<0.0001 in all domains) that QoL scores among MK cases at 3 months were lower than the controls, after adjusting for age, and economic status. Table 2 VRQoL and general HRQoL among cases (baseline and 3 months) and controls (215 pairs)

Table 2 shows the baseline and 3-month VRQoL and HRQoL scores for the cases and the scores for the control group. The mean baseline VRQoL scores among the cases were all low (<50) except the psychosocial domain which had a score of 65 points. The most affected domain was visual symptoms, with a mean score of 20.7 (95% CI 18.8 to 22.7). The mean baseline HRQoL scores among the cases were all low (<50). At 3 months, all the case VRQoL and HRQoL scores had increased and were relatively high (between 80 and 91). Despite this increase, there was still very strong evidence (p<0.0001 in all domains) that QoL scores among MK cases at 3 months were lower than the controls, after adjusting for age, and economic status. Table 2 VRQoL and general HRQoL among cases (baseline and 3 months) and controls (215 pairs) Domain Cases at baseline Cases at 3 Months Controls at 3 Months Adjusted mean difference at 3 months Mean (95% CI) Mean (95% CI) Mean (95% CI) Mean * (95% CI) P value † VRQoL Overall sight 32.9 (30.6 to 35.2) 86.3 (83.3 to 89.2) 98.6 (97.5 to 99.6) 11.6 (8.8 to 14.5) <0.0001 Visual symptom 20.7 (18.8 to 22.7) 88.3 (85.5 to 91.1) 99.4 (98.9 to 99.8) 10.5 (7.7 to 13.3) <0.0001 General functioning 42.8 (40.6 to 45.1) 89.0 (86.1 to 91.9) 99.6 (99.3 to 100) 9.9 (7.1 to 12.7) <0.0001 Psychosocial 65.6 (62.5 to 68.8) 90.7 (88.1 to 93.3) 99.8 (99.4 to 100) 8.5 (6.0 to 11.0) <0.0001 HRQoL General facet items Overall quality of life 40.3 (39.0 to 41.6) 86.6 (83.9 to 89.4) 97.2 (96.4 to 97.9) 10.2 (7.6 to 12.8) <0.0001 Overall Health 31.0 (29.0 to 33.0) 85.6 (82.7 to 88.5) 98.2 (97.3 to 99.1) 12.0 (9.1 to 14.9) <0.0001 Domains Physical health 28.4 (26.5 to 30.3) 86.1 (83.1 to 89.2) 98.3 (97.6 to 98.9) 11.5 (8.6 to 14.5) <0.0001 Psychological 49.2 (47.5 to 50.9) 84.4 (81.9 to 86.9) 94.0 (93.4 to 94.7) 9.1 (6.7 to 11.5) <0.0001 Social 48.5 (46.5 to 50.5) 88.2 (85.2 to 91.2) 98.5 (97.6 to 99.3) 9.9 (6.9 to 12.9) <0.0001 Environment 43 (41.7 to 44.3) 84.8 (82.0 to 87.6) 96.3 (95.2 to 97.3) 10.9 (8.2 to 13.6) <0.0001 Only the cases who had controls were included in this analysis (215 pairs).

to 86.9) 94.0 (93.4 to 94.7) 9.1 (6.7 to 11.5) <0.0001 Social 48.5 (46.5 to 50.5) 88.2 (85.2 to 91.2) 98.5 (97.6 to 99.3) 9.9 (6.9 to 12.9) <0.0001 Environment 43 (41.7 to 44.3) 84.8 (82.0 to 87.6) 96.3 (95.2 to 97.3) 10.9 (8.2 to 13.6) <0.0001 Only the cases who had controls were included in this analysis (215 pairs). *Mean difference between cases and controls adjusted for age, sex and wealth status. †Linear regression random effects model was used to test for significance of the differences among the cases and controls adjusted for age, sex and wealth status. HRQoL, health-related quality of life; VRQoL, vision-related quality of life. Table 3 shows the presenting vision, microbiology and 3-month outcomes among the 260 cases. Majority (137/260) presented with vision worse than 6/60 in the affected eye. Microbiology results were available for 226/260 participants out of which the majority (63%) showed fungal keratitis. At 3 months, 138/260 had vision of better than 6/18 in the affected eye. Vision had improved in 137 individuals, remained the same in 56 and worsened in 66 participants (sign rank p<0.0001). Table 3 Presenting vision, microbiology and 3-month outcomes for the cases (n=260)

Table 3 shows the presenting vision, microbiology and 3-month outcomes among the 260 cases. Majority (137/260) presented with vision worse than 6/60 in the affected eye. Microbiology results were available for 226/260 participants out of which the majority (63%) showed fungal keratitis. At 3 months, 138/260 had vision of better than 6/18 in the affected eye. Vision had improved in 137 individuals, remained the same in 56 and worsened in 66 participants (sign rank p<0.0001). Table 3 Presenting vision, microbiology and 3-month outcomes for the cases (n=260) Variable n/260 (%) Presenting visual acuity in the affected eye (Snellen) * >6/18 86 (33) 6/18-6/60 36 (14) <6/60 137 (53) Presenting visual acuity in the non-affected eye (Snellen) * >6/18 232 (90) 6/18-6/60 14 (5) <6/60 13 (5) Microbiology † Fungal 143 (63) Bacterial 18 (8) Mixed 13 (6) Unknown 52 (23) Visual acuity in the affected eye (Snellen) at 3 months >6/18 138 (53) 6/18-6/60 37 (14) <6/60 85 (33) Visual acuity in the non-affected eye at 3 months >6/18 229 (90) 6/18-6/60 11 (4) <6/60 15 (6) Outcome at 3 months Healed no scar 34 (12) Healed Mild scar 83 (30) Healed moderate scar 65 (24) Healed dense scar 46 (17) Eviscerated 24 (9) Not healed 20 (7) Staphyloma 4 (1) *There was one missing value (n=259).

60 85 (33) Visual acuity in the non-affected eye at 3 months >6/18 229 (90) 6/18-6/60 11 (4) <6/60 15 (6) Outcome at 3 months Healed no scar 34 (12) Healed Mild scar 83 (30) Healed moderate scar 65 (24) Healed dense scar 46 (17) Eviscerated 24 (9) Not healed 20 (7) Staphyloma 4 (1) *There was one missing value (n=259). †Corneal scrapping was performed on 226/260 participants; it was not possible to obtain corneal scrapping samples in 34 participants either due to uncooperative patient, declining consent, deep infiltrates with intact epithelium, such patients were treated based on clinical impression. In all, 52 samples returned negative, no organism detected on microscopy or culture, these were also managed based on clinical impression. To investigate whether the difference in QoL between the cases and controls was due to factors in addition to impaired vision in the MK group, a separate subgroup analysis was performed comparing only MK cases with normal vision in the affected eye (better than 6/18) to their paired controls (online supplementary table 1). It was observed that the differences in QoL was similar to that obtained when using all the cases. 10.1136/bmjophth-2019-000351.supp1Supplementary data

To investigate whether the difference in QoL between the cases and controls was due to factors in addition to impaired vision in the MK group, a separate subgroup analysis was performed comparing only MK cases with normal vision in the affected eye (better than 6/18) to their paired controls (online supplementary table 1). It was observed that the differences in QoL was similar to that obtained when using all the cases. 10.1136/bmjophth-2019-000351.supp1Supplementary data Tables 4 and 5 show factors associated with a good VRQoL and HRQoL among the cases at 3 months. This analysis was among all the 260 MK cases who were followed up at 3 months. Analysis was restricted to variables related to the disease such as vision at 3 months and whether the person had lost their eye. They were adjusted for sex, age, education, baseline QoL and socioeconomic status. Vision and eye loss were the both found to be associated with VRQoL and HRQoL. Table 4 Univariable and multivariable linear regression for factors associated with VRQoL among cases only (n=260) seen at 3 months

Tables 4 and 5 show factors associated with a good VRQoL and HRQoL among the cases at 3 months. This analysis was among all the 260 MK cases who were followed up at 3 months. Analysis was restricted to variables related to the disease such as vision at 3 months and whether the person had lost their eye. They were adjusted for sex, age, education, baseline QoL and socioeconomic status. Vision and eye loss were the both found to be associated with VRQoL and HRQoL. Table 4 Univariable and multivariable linear regression for factors associated with VRQoL among cases only (n=260) seen at 3 months Variable Overall sight Visual symptom General functioning Psychosocial Mean (95%CI) Mean (95%CI) Mean (95%CI) Mean (95%CI) Visual acuity at 3 months >6/18 89.7 (86.6 to 92.7) 90.9 (87.6 to 94.2) 93.3 (90.5 to 96.1) 93.3 (90.9 to 96.1) 6/18–6/60 78.4 (69.1 to 87.6) 82.7 (73.5 to 91.8) 82.2 (73.3 to 91.1) 83.3 (74.9 to 91.7) <6/60 77.4 (71.6 to 83.1) 82.9 (78.3 to 87.6) 81.3 (76 to 86.6) 83.8 (78.7 to 88.9) P value*† <0.0001 0.006 <0.0001 <0.0001 P value‡† 0.001 0.044 0.006 0.003 Eye removal § No 83.6 (80.7 to 86.6) 86.5 (83.5 to 89.3) 87.5 (84.6 to 90.3) 88.6 (86 to 91.2) Yes 87.5 (76.3 to 98.7) 93.1 (86.3 to 99.8) 90.8 (81.6 to 100) 90.8 (81.2 to 100) P value* 0.406 0.168 0.356 0.619 P value‡ 0.030 0.025 0.054 0.111 *P values from univariable linear regression analysis. †For visual acuity at 3 months (ordinal exposures with three categories), the p values were calculated for trend.

Variable Overall sight Visual symptom General functioning Psychosocial Mean (95%CI) Mean (95%CI) Mean (95%CI) Mean (95%CI) Visual acuity at 3 months >6/18 89.7 (86.6 to 92.7) 90.9 (87.6 to 94.2) 93.3 (90.5 to 96.1) 93.3 (90.9 to 96.1) 6/18–6/60 78.4 (69.1 to 87.6) 82.7 (73.5 to 91.8) 82.2 (73.3 to 91.1) 83.3 (74.9 to 91.7) <6/60 77.4 (71.6 to 83.1) 82.9 (78.3 to 87.6) 81.3 (76 to 86.6) 83.8 (78.7 to 88.9) P value*† <0.0001 0.006 <0.0001 <0.0001 P value‡† 0.001 0.044 0.006 0.003 Eye removal § No 83.6 (80.7 to 86.6) 86.5 (83.5 to 89.3) 87.5 (84.6 to 90.3) 88.6 (86 to 91.2) Yes 87.5 (76.3 to 98.7) 93.1 (86.3 to 99.8) 90.8 (81.6 to 100) 90.8 (81.2 to 100) P value* 0.406 0.168 0.356 0.619 P value‡ 0.030 0.025 0.054 0.111 *P values from univariable linear regression analysis. †For visual acuity at 3 months (ordinal exposures with three categories), the p values were calculated for trend. ‡P values from multivariable linear regression analysis, adjusted for age, education status, wealth category and baseline QoL. §Eye removal was a priori. VRQoL, vision-related quality of life. Table 5 Univariable and multivariable linear regression for factors associated with HRQoL among cases only (n=260) seen at 3 months

†For visual acuity at 3 months (ordinal exposures with three categories), the p values were calculated for trend. ‡P values from multivariable linear regression analysis, adjusted for age, education status, wealth category and baseline QoL. §Eye removal was a priori. VRQoL, vision-related quality of life. Table 5 Univariable and multivariable linear regression for factors associated with HRQoL among cases only (n=260) seen at 3 months Variable Overall QoL Overall health Physical health Psychological Social Environment Mean (95% CI) Mean (95% CI) Mean (95% CI) Mean (95% CI) Mean (95% CI) Mean (95% CI) Vision outcome at 3 months >6/18 88.7 (85.5 to 91.9) 88.8 (85.6 to 91.6) 89.4 (86.2 to 92.5) 87.7 (85.1 to 90.3) 90.6 (87.4 to 93.7) 87.9 (84.9 to 90.9) 6/18–6/60 82.1 (73.7 to 90.5) 82.0 (74.2 to 90.0) 81.9 (73.6 to 90.0) 80.9 (74.3 to 87.5) 84.8 (76.8 to 92.8) 78.8 (70.6 to 87) <6/60 78.4 (73.3 to 83.3) 79.3 (74.0 to 84.5) 77.7 (72.0 to 83.5) 77.5 (72.9 to 82.1) 80.5 (74.6 to 86.3) 77.1 (71.9 to 82.3) P value*† <0.0001 0.001 <0.0001 <0.0001 0.002 <0.0001 P value ‡† 0.008 0.011 0.007 0.003 0.015 0.013 Eye removal § No 84.0 (81.2 to 86.9) 84.5 (81.8 to 87.1) 84.0 (81.1 to 87) 83.0 (80.6 to 85.4) 86.4 (83.6 to 89.3) 82.5 (79.4 to 85.3) Yes 88.0 (79.2 to 96.9) 88.1 (79.2 to 96.9) 89.1 (78.9 to 99.4) 86.8 (78.9 to 94.7) 87.8 (77.8 to 98.1) 88.2 (79.7 to 96.6) P value* 0.356 0.402 0.288 0.331 0.633 0.198 P value‡ 0.038 0.031 0.028 0.041 0.122 0.009 *P values from univariable linear regression analysis.

.6 to 85.4) 86.4 (83.6 to 89.3) 82.5 (79.4 to 85.3) Yes 88.0 (79.2 to 96.9) 88.1 (79.2 to 96.9) 89.1 (78.9 to 99.4) 86.8 (78.9 to 94.7) 87.8 (77.8 to 98.1) 88.2 (79.7 to 96.6) P value* 0.356 0.402 0.288 0.331 0.633 0.198 P value‡ 0.038 0.031 0.028 0.041 0.122 0.009 *P values from univariable linear regression analysis. †For visual acuity at 3 months (ordinal exposures with three categories), the p values were calculated for trend. ‡P values from multivariable linear regression analysis, adjusted for age, sex, education status, wealth category and baseline QoL. §Eye removal was a priori. HRQoL, health-related quality of life. Validity of the data was found to be good. Satisfying the known-groups difference criteria, the cases had significantly lower VRQoL and HRQoL scores in all domains (p<0.0001) than the controls (table 2). The VRQoL data were reliable after being assessed for internal consistency with a Cronbach’s alpha: coefficients of >0.80 (visual symptom 0.90, general functioning 0.98, psychosocial 0.87). The overall HRQoL data had a Cronbach’s alpha of 0.98 (physical health 0.96, psychological 0.89, social 0.91 and environment 0.95).

he VRQoL data were reliable after being assessed for internal consistency with a Cronbach’s alpha: coefficients of >0.80 (visual symptom 0.90, general functioning 0.98, psychosocial 0.87). The overall HRQoL data had a Cronbach’s alpha of 0.98 (physical health 0.96, psychological 0.89, social 0.91 and environment 0.95). Discussion This study investigated the impact of MK on QoL in Uganda. Overall, both the VRQoL and HRQoL among MK patients at baseline was substantially reduced. The lowest scores were in the visual symptom (most had reduced vision) and physical health (most were in pain) categories. The least affected domain at baseline was psychological, which assesses ability to attend functions, feeling ashamed, feeling like a burden to others and fear of losing the other eye. In the only other published QoL study in people with MK, the MUTT1 study from India, QoL at the time of presentation was not reported.8 The QoL scores among the MK cases had improved greatly by 3 months. However, compared with the control group, there was strong evidence that MK results in a persistent reduction in QoL, with a mean difference in QoL scores of around 10 points. This effect remained evident even when MK cases with impaired vision were excluded from the analysis.

MK cases had improved greatly by 3 months. However, compared with the control group, there was strong evidence that MK results in a persistent reduction in QoL, with a mean difference in QoL scores of around 10 points. This effect remained evident even when MK cases with impaired vision were excluded from the analysis. In looking at factors associated with QoL among MK cases at 3 months, after adjusting for other factors that may affect QoL, such as age, sex, education and economic status, visual acuity was an important determinant for both VRQoL and HRQoL. There was evidence that a history of eye loss was also associated with both VRQoL and HRQoL at 3 months. It was surprising to note that the people who had undergone evisceration had generally better QoL scores compared with those who did not. At the time of eye removal, most of the eyes were too damaged and painful that the people were ‘demanding’ for eye removal. They received socket prostheses (artificial eyes) after their eye removal procedures, this could have led to a marked reduction in pain and other unpleasant symptoms after evisceration, which led to a less impaired QoL compared with others who were not eviscerated where there would have been some with ongoing pain and other symptoms like dense corneal scars.

ial eyes) after their eye removal procedures, this could have led to a marked reduction in pain and other unpleasant symptoms after evisceration, which led to a less impaired QoL compared with others who were not eviscerated where there would have been some with ongoing pain and other symptoms like dense corneal scars. Validity and reliability of QoL data These tools have been used in a number of other vision-related studies to show a difference in QoL and have been reported to be valid and reliable in studies conducted in similar settings.5–7 Although the WHO/PBD-VF20 tool was designed to assess binocular vision, it has been demonstrated to be effective in detecting differences in monocular visual impairment in the MUTT1 trial where patients randomised to natamycin had a better 3-month visual-related QoL outcome compared with patients randomised to variconazole.8 In this study, both the VRQoL and HRQoL data measured what they were intended to measure (construct validity) by demonstrating evidence of differences in the scores between groups known to be different; MK cases and healthy controls had lower and higher scores, respectively. The VRQoL data also showed that subscales correlate well with measures of impact of MK on QoL such as visual acuity where worsening in these measures is associated with lower VRQoL subscale scores (construct validity). There was evidence of higher homogeneity among the items in each VRQoL and HRQoL subscale (internal consistency) than the generally accepted criteria of >0.70. Cronbach’s alpha scores ranged from 0.80 to 0.99 across all sub scales.

g in these measures is associated with lower VRQoL subscale scores (construct validity). There was evidence of higher homogeneity among the items in each VRQoL and HRQoL subscale (internal consistency) than the generally accepted criteria of >0.70. Cronbach’s alpha scores ranged from 0.80 to 0.99 across all sub scales. Strengths/limitations This was the first study to look at impact of MK on QoL and compare these with unaffected controls. The control selection was good and matched for age, sex and village. We used validated tools which have been previously applied in similar settings and the sample size was adequate to test for differences in QoL. This study did not collect baseline QoL information from the control group because it was not logistically possible. Information on other ocular comorbidities such as presence of a cataract and posterior segment disease at 3 months was not collected as these examination were conducted at patients’ homes and it was not practical to perform a full eye examination in such settings. Conclusion This study showed that MK severely affects QoL in the acute phase. With treatment, QoL improves, with the highest QoL in cases who had little or no visual impairment at 3 months. Despite this impressive improvement, the QoL at 3 months of someone previously affected by MK (even when they have normal vision) remains lower than controls. The authors would like to appreciate Mr Gilbert Arinda, Ms. Pauline Boonabaana, Mr Martin Bukenya, Mr Bernard Beinomugisha, Mr Martin Bukenya and Ms. Allen Asiimwe for helping in data collection.

Conclusion This study showed that MK severely affects QoL in the acute phase. With treatment, QoL improves, with the highest QoL in cases who had little or no visual impairment at 3 months. Despite this impressive improvement, the QoL at 3 months of someone previously affected by MK (even when they have normal vision) remains lower than controls. The authors would like to appreciate Mr Gilbert Arinda, Ms. Pauline Boonabaana, Mr Martin Bukenya, Mr Bernard Beinomugisha, Mr Martin Bukenya and Ms. Allen Asiimwe for helping in data collection. Contributors: AS, AL, VHH, DM and MJB designed the study. All authors contributed to the conducting of the study. AS, DM and MJB analysed the results. All authors contributed to the draft of manuscript and editing of the final paper. Competing interests: None declared. Patient consent for publication: Not required. Provenance and peer review: Not commissioned; externally peer reviewed. Data availability statement: Data are available upon request.

INTRODUCTION In the literature and at our own institution, there has been debate about the use of commercially available fourth-generation fluoroquinolones versus compounded fortified antibiotics such as vancomycin and tobramycin for the initial treatment of bacterial corneal ulcers. Determining the best treatment depends on the causative organism; however, Gram stain and culture results are not available for hours to days, and initial therapy is commonly empirical. Several randomised controlled trials have shown equivalency between the fluoroquinolones and fortified antibiotics.1–5 In the recent Steroids for Corneal Ulcer Trial, no bacteria were resistant to moxifloxacin.6 Despite this, the use of fortified antibiotics remains quite common.78 Of note, most of these studies occurred in developing countries where bacterial resistance patterns are likely to be quite different from the those in developed countries such as the USA. New antibiotic agents, such as topical linezolid, and changing susceptibility patterns may have altered the empirical treatment of bacterial keratitis.9 Here we assess corneal specialists’ current practice patterns regarding initial treatment of severe central bacterial ulcers.

INTRODUCTION In the literature and at our own institution, there has been debate about the use of commercially available fourth-generation fluoroquinolones versus compounded fortified antibiotics such as vancomycin and tobramycin for the initial treatment of bacterial corneal ulcers. Determining the best treatment depends on the causative organism; however, Gram stain and culture results are not available for hours to days, and initial therapy is commonly empirical. Several randomised controlled trials have shown equivalency between the fluoroquinolones and fortified antibiotics.1–5 In the recent Steroids for Corneal Ulcer Trial, no bacteria were resistant to moxifloxacin.6 Despite this, the use of fortified antibiotics remains quite common.78 Of note, most of these studies occurred in developing countries where bacterial resistance patterns are likely to be quite different from the those in developed countries such as the USA. New antibiotic agents, such as topical linezolid, and changing susceptibility patterns may have altered the empirical treatment of bacterial keratitis.9 Here we assess corneal specialists’ current practice patterns regarding initial treatment of severe central bacterial ulcers. MATERIALS AND METHODS An international survey was distributed via email to members of the Cornea Society listserv. Participation was completely voluntary and anonymous, and no identifiers were collected including name, age or sex. The survey was initially distributed on 21 July 2014 via the internet survey tool SurveyMonkey (surveymonkey.com; Palo Alto, California), with one reminder on 26 August 2014. Data collection was closed in October of 2014.

completely voluntary and anonymous, and no identifiers were collected including name, age or sex. The survey was initially distributed on 21 July 2014 via the internet survey tool SurveyMonkey (surveymonkey.com; Palo Alto, California), with one reminder on 26 August 2014. Data collection was closed in October of 2014. The survey consisted of seven questions (figure 1). Respondents were presented with the clinical scenario of a patient with a large, central, bacterial corneal ulcer and were asked what their empirical antibiotic treatment would be, why they chose that regimen and whether they would obtain cultures prior to instituting treatment. Specific antibiotic regimens given as options included fourth-generation fluoroquinolones (such as moxifloxacin and gatifloxacin) and fortified antibiotics (including tobramycin, vancomycin and cefazolin). Respondents were also queried regarding their demographics including their geographic location, the average number of corneal ulcers seen per month as well as the number of years they had been in practice. Finally they were asked what percentage of Staphylococcus aureus in the community would have to be methicillin-resistant Staphylococcus aureus (MRSA) for them to cover MRSA empirically.

The survey consisted of seven questions (figure 1). Respondents were presented with the clinical scenario of a patient with a large, central, bacterial corneal ulcer and were asked what their empirical antibiotic treatment would be, why they chose that regimen and whether they would obtain cultures prior to instituting treatment. Specific antibiotic regimens given as options included fourth-generation fluoroquinolones (such as moxifloxacin and gatifloxacin) and fortified antibiotics (including tobramycin, vancomycin and cefazolin). Respondents were also queried regarding their demographics including their geographic location, the average number of corneal ulcers seen per month as well as the number of years they had been in practice. Finally they were asked what percentage of Staphylococcus aureus in the community would have to be methicillin-resistant Staphylococcus aureus (MRSA) for them to cover MRSA empirically. Responses were collected and analysed using descriptive statistics with Stata V.14.0. χ2 test was used to compare responses between international and US respondents, and logistic regression models were used to assess the relationship between specific concerns, such as toxicity or antibiotic resistance and choice of antibiotic. Institutional review board exemption was obtained from the University of California, San Francisco Committee on Human Research. The study adhered to the Declaration of Helsinki and all federal and state laws.

the relationship between specific concerns, such as toxicity or antibiotic resistance and choice of antibiotic. Institutional review board exemption was obtained from the University of California, San Francisco Committee on Human Research. The study adhered to the Declaration of Helsinki and all federal and state laws. RESULTS A total of 1009 cornea specialists were members at the time the survey was sent out. One hundred forty responses to the survey were received, for a response rate of 14%. Respondents were mostly from the USA (n=104), with the remaining international respondents (n=36) coming from 17 countries on six continents (table 1). Many respondents had greater than 20 years of clinical experience (n=62, 44%), followed by 10–20 years in practice (n=33, 24%), 5–10 years in practice (n=30, 21%) and 0–5 years in practice (n=15, 11%). In general, the international respondents encountered more cases of infectious keratitis per month, with 42% seeing more than five cases per month versus 24% of US respondents (p=0.05). When asked if they would culture a large central ulcer, all international respondents (n=36) and 98% (n=101) of US respondents reported that they would (two US respondents, or 2%, said they would not).

infectious keratitis per month, with 42% seeing more than five cases per month versus 24% of US respondents (p=0.05). When asked if they would culture a large central ulcer, all international respondents (n=36) and 98% (n=101) of US respondents reported that they would (two US respondents, or 2%, said they would not). Respondents were asked about their empirical treatment of bacterial keratitis, and their responses revealed marked differences in practice patterns between clinicians in the USA and in other countries, as seen in figure 2. In the USA, the majority of clinicians surveyed (n=83, 80%) chose fortified antibiotics as their initial treatment, with 55% (n=57) choosing fortified vancomycin. Sixteen per cent (n=17) of US clinicians surveyed chose to initially treat with fourth-generation fluoroquinolones alone. International respondents were twice as likely to use fluoroquinolone monotherapy (31%, n=11, p=0.07) and were less likely to use fortified vancomycin with 33% reporting it as their empirical choice (n=12, p=0.03).

(n=17) of US clinicians surveyed chose to initially treat with fourth-generation fluoroquinolones alone. International respondents were twice as likely to use fluoroquinolone monotherapy (31%, n=11, p=0.07) and were less likely to use fortified vancomycin with 33% reporting it as their empirical choice (n=12, p=0.03). Reasons for selecting their chosen initial antibiotic were largely similar across US and international sites. Clinicians reported desiring broad-spectrum antibiotic coverage most often, with 78% (n=81) of US respondents and 89% (n=32) of international respondents citing this as a reason for their antibiotic choice. Both groups also reported the following reasons with similar frequency: most effective (n=43, 41% US vs n=15, 42% internationally), availability (n=33, 32% US vs n=12, 33% internationally), synergy (n=15, 14% US vs n=6, 17% internationally) and toxicity profile (n=13, 13% US vs n=5, 14% internationally). US respondents were more concerned about covering resistant organisms, listing this as a reason for their empirical antibiotic choice more than twice as often as international respondents (n=41, 39% vs n=7, 19%, p=0.03). Forty-five per cent (n=46) of US respondents reported that their empirical antibiotic regimen covered MRSA, compared with 22% (n=8) of international respondents. Table 2 contrasts US and international clinicians’ threshold for empirical MRSA coverage, showing that international clinicians had a lower threshold for empirical MRSA coverage than did US clinicians (p<0.01).

d that their empirical antibiotic regimen covered MRSA, compared with 22% (n=8) of international respondents. Table 2 contrasts US and international clinicians’ threshold for empirical MRSA coverage, showing that international clinicians had a lower threshold for empirical MRSA coverage than did US clinicians (p<0.01). Table 3 shows the relationship between the different reasons for choosing a particular antibiotic and the likelihood of selecting fluoroquinolone monotherapy or fortified vancomycin, correcting for location (defined as US or international). Respondents who listed availability as a primary consideration had 20.86 times the odds of choosing fluoroquinolone monotherapy (p<0.001). Concern over ocular surface toxicity was also predictive of choosing fluoroquinolone monotherapy (OR=7.48, p<0.001). Not surprisingly, respondents concerned about resistance had 12.51 times the odds of choosing fortified vancomycin (p<0.001), while broad spectrum coverage was also a statistically significant predictor for selecting vancomycin (OR=7.10, p<0.001).

o predictive of choosing fluoroquinolone monotherapy (OR=7.48, p<0.001). Not surprisingly, respondents concerned about resistance had 12.51 times the odds of choosing fortified vancomycin (p<0.001), while broad spectrum coverage was also a statistically significant predictor for selecting vancomycin (OR=7.10, p<0.001). DISCUSSION In this study, we report the results of an international survey of cornea specialists regarding their empirical treatment of large central corneal ulcers. We found that practice patterns differ between US and international sites, with US clinicians more likely to use fortified antibiotics as their initial treatment. They were also approximately twice as likely to use fortified vancomycin empirically. Prior surveys have found that cornea specialists are more likely than general ophthalmologists to treat bacterial keratitis with fortified antibiotics.710 US respondents were more concerned about resistant organisms and also more likely to include coverage for MRSA in their default treatment.

ed vancomycin empirically. Prior surveys have found that cornea specialists are more likely than general ophthalmologists to treat bacterial keratitis with fortified antibiotics.710 US respondents were more concerned about resistant organisms and also more likely to include coverage for MRSA in their default treatment. Concern about the availability of antibiotics was the greatest predictor of fluoroquinolone monotherapy. Fortified antibiotics require a compounding pharmacy and can be difficult for clinicians to access. Concern over ocular toxicity was the second strongest predictor of fluoroquinolone use in our study. Fortified antibiotics and in particular fortified vancomycin are known to cause significant ocular surface toxicity, such as increasing epithelial defect size or conjunctival injection, which can sometimes be difficult to distinguish from progression of the corneal infection. Some have advocated using topical linezolid as an alternative to vancomycin, which has less toxicity but still covers resistant organisms including MRSA.9 Interestingly, no one in our survey reported using linezolid. Our observation that those concerned with broad-spectrum coverage were more likely to choose fortified antibiotics and those favouring a low toxicity profile were more likely to choose a single agent could be influenced by the availability of particular antibiotics.

ingly, no one in our survey reported using linezolid. Our observation that those concerned with broad-spectrum coverage were more likely to choose fortified antibiotics and those favouring a low toxicity profile were more likely to choose a single agent could be influenced by the availability of particular antibiotics. Fluoroquinolone antibiotics penetrate ocular tissues well and have good safety profiles with minimal toxicity to the ocular surface. They also exhibit broad coverage of bacterial organisms implicated in bacterial ulcers. The primary target of fluoroquinolones is DNA gyrase, an essential bacterial enzyme. There is reason to believe that because of this mechanism of action, they may be more susceptible to the development of resistance than other antibiotics.11 Rates of MRSA infection have been increasing in North America, and ocular isolates in the USA have been reported to be resistant to fluoroquinolones approximately 80% of the time.12–14 In our study concern for antibiotic resistance and broad spectrum coverage were highly predictive of choosing fortified vancomycin after controlling for location. The fact that international cornea specialists both have a lower threshold for MRSA coverage and are less likely to be using fortified antibiotics suggests that they consider the prevalence of MRSA to be relatively low in their population.

ghly predictive of choosing fortified vancomycin after controlling for location. The fact that international cornea specialists both have a lower threshold for MRSA coverage and are less likely to be using fortified antibiotics suggests that they consider the prevalence of MRSA to be relatively low in their population. The strengths of this study include the geographic diversity of respondents, providing both US and international perspectives and representing a broad range of experience. In addition, a large number of our respondents were experienced physicians and seeing ulcers frequently in clinical care. Limitations include the fact that we only surveyed members of the Cornea Society listserv, therefore our results may not be generalisable to all ophthalmologists. Additionally the survey did not ask respondents about their practice setting, which may be related to empirical antibiotic choice since those in academic settings may have more access to fortified antibiotics. Another limitation would be the multiple-choice format of our question about antibiotic preference because other options such as vancomycin monotherapy were not listed (in this case, we listed vancomycin only in conjunction with other antibiotics because it only covers Gram-positive bacteria). We attempted to address this by giving respondents the option to write in their own preferred antibiotic, but it is possible the question format affected responses given.

herapy were not listed (in this case, we listed vancomycin only in conjunction with other antibiotics because it only covers Gram-positive bacteria). We attempted to address this by giving respondents the option to write in their own preferred antibiotic, but it is possible the question format affected responses given. Practice patterns for the initial treatment of bacterial keratitis vary with clinicians in the USA more likely to use fortified antibiotics versus fluoroquinolone monotherapy and more concerned with resistant organisms than their international peers. These differences may be due to multiple factors, including varying prevalence of resistant organisms, lack of availability of fortified antibiotics and concerns for ocular toxicity. Determining which of these factors impact clinical outcomes, such as visual acuity, should be an important focus of future research. Individual support for this study came from K23 EY025025 (JRN), an unrestricted grant from the Peirles Foundation (JRN) and an unrestricted grant from Research to Prevent Blindness (JRN).

Practice patterns for the initial treatment of bacterial keratitis vary with clinicians in the USA more likely to use fortified antibiotics versus fluoroquinolone monotherapy and more concerned with resistant organisms than their international peers. These differences may be due to multiple factors, including varying prevalence of resistant organisms, lack of availability of fortified antibiotics and concerns for ocular toxicity. Determining which of these factors impact clinical outcomes, such as visual acuity, should be an important focus of future research. Individual support for this study came from K23 EY025025 (JRN), an unrestricted grant from the Peirles Foundation (JRN) and an unrestricted grant from Research to Prevent Blindness (JRN). Funding There are no conflicts of interest to declare. Corresponding author Dr Jennifer Rose-Nussbaumer contributed to study design and implementation, data analysis, and writing of this manuscript. Ariana Austin contributed to data analysis and writing of the manuscript. Dr Julie Schallhorn and Dr Mike Geske were involved in design and implementation of this study. Dr Mark Mannis contributed to study design and implementation as well as editing of the manuscript. Dr Tom Lietman contributed to the design of the study, data analysis, and editing of the manuscript.

ting of the manuscript. Dr Julie Schallhorn and Dr Mike Geske were involved in design and implementation of this study. Dr Mark Mannis contributed to study design and implementation as well as editing of the manuscript. Dr Tom Lietman contributed to the design of the study, data analysis, and editing of the manuscript. Contributors Corresponding author JR-N contributed to study design and implementation, data analysis and writing of this manuscript. AA contributed to data analysis and writing of the manuscript. JS and MG were involved in design and implementation of this study. MM contributed to study design and implementation as well as editing of the manuscript. TL contributed to the design of the study, data analysis and editing of the manuscript. Ethics approval University of California, San Francisco Institutional Review Board. Provenance and peer review Not commissioned; externally peer reviewed. Figure 1 Survey distributed to corneal specialists via email. Figure 2 Graph showing antibiotic choice of US and international cornea specialists for empirical treatment of bacterial keratitis. Table 1 Geographic distribution of respondentsl

Ethics approval University of California, San Francisco Institutional Review Board. Provenance and peer review Not commissioned; externally peer reviewed. Figure 1 Survey distributed to corneal specialists via email. Figure 2 Graph showing antibiotic choice of US and international cornea specialists for empirical treatment of bacterial keratitis. Table 1 Geographic distribution of respondentsl Country N=140 (%) USA* 104 (74) Northeast 11 (11) South 36 (35) Midwest 22 (21) West 33 (32) Unknown 2 (2) International 36 (26) Armenia 1 (3) Australia 2 (6) Brazil 4 (11) Canada 3 (8) Colombia 2 (6) Germany 2 (6) India 6 (17) Ireland 2 (6) Israel 2 (6) Mexico 3 (8) Paraguay 1 (3) Philippines 1 (3) South Africa 2 (6) Spain 1 (3) Sweden 1 (3) Turkey 1 (3) UK 2 (6) * USA divided into regions as defined by the US Census Bureau: Northeast (Connecticut, Maine, Massachusetts, New Hampshire, New Jersey, New York, Rhode Island, Pennsylvania and Vermont), South (Alabama, Arkansas, District of Columbia, Delaware, Florida, Georgia, Kentucky, Louisiana, Maryland, Mississippi, Oklahoma, North Carolina, South Carolina, Tennessee, Virginia, Texas and West Virginia), Midwest (Iowa, Illinois, Indiana, Kansas, Michigan, Minnesota, Missouri, Nebraska, North Dakota, Ohio, South Dakota and Wisconsin) and West (Alaska, Arizona, California, Colorado, Hawaii, Idaho, Montana, New Mexico, Nevada, Oregon, Utah, Washington and Wyoming).

Carolina, Tennessee, Virginia, Texas and West Virginia), Midwest (Iowa, Illinois, Indiana, Kansas, Michigan, Minnesota, Missouri, Nebraska, North Dakota, Ohio, South Dakota and Wisconsin) and West (Alaska, Arizona, California, Colorado, Hawaii, Idaho, Montana, New Mexico, Nevada, Oregon, Utah, Washington and Wyoming). Table 2 Threshold for empirical methicillin-resistant Staphylococcus aureus coverage (MRSA). Percentage of MRSA in the community necessary to change the default corneal ulcer treatment to include MRSA coverage. % MRSA USA, N (%) International, N (%) p Value 5%–15% 14 (14) 15 (42) p<0.001 15%–30% 30 (29) 7 (19) 30%–50% 6 (6) 5 (14) >50% 7 (7) 1 (3) Default MRSA coverage 46 (45) 8 (22) Table 3 Reasons for antibiotic of choice. Logistic regression models predicting likelihood of choosing fluoroquinolone monotherapy and fortified vancomycin, correcting for location. Reason OR p Value 95% CI lower bound Fluoroquinolone monotherapy Efficacy 1.05 0.91 0.45 to 2.46 Coverage 0.30 0.02 0.11 to 0.81 Availability 20.86 <0.001* 6.79 to 64.13 Toxicity 7.48 <0.001* 2.53 to 22.09 Synergy 0.16 0.08 0.02 to 1.23 Resistance 0.39 0.08 0.13 to 1.11 Fortified vancomycin Efficacy 1.96 0.06 0.97 to 3.92 Coverage 7.10 <0.001* 2.44 to 20.61 Availability 0.20 <0.001* 0.09 to 0.44 Toxicity 0.10 0.003* 0.02 to 0.45 Synergy 0.75 0.56 0.29 to 1.95 Resistance 12.51 <0.001* 4.97 to 31.45 * Statistically significant after Holm-Šídák correction for multiple comparisons.

o 1.11 Fortified vancomycin Efficacy 1.96 0.06 0.97 to 3.92 Coverage 7.10 <0.001* 2.44 to 20.61 Availability 0.20 <0.001* 0.09 to 0.44 Toxicity 0.10 0.003* 0.02 to 0.45 Synergy 0.75 0.56 0.29 to 1.95 Resistance 12.51 <0.001* 4.97 to 31.45 * Statistically significant after Holm-Šídák correction for multiple comparisons. Key messages There is debate over the best empirical treatment for bacterial keratitis, and there are regional variations in practice patterns of corneal specialists. This survey demonstrates that practice patterns are influenced by concern over availability and toxicity versus broad spectrum coverage and resistance. Overall, respondents in the USA were more likely to treat with fortified antibiotics than their international peers. A well-designed clinical trial on the treatment of bacterial ulcers is needed to help clinicians initiate the best treatment and ultimately reduce morbidity.

Key messages Although in the past all three solutions have been used for vitrectomy irrigating solutions, they are possibly not comparable in maintaining optimal retinal health during and after retinal surgery. Lactated Ringers appears to be an inadequate substitute for balanced salt solution-plus (BSS-Plus), although balanced salt solution (BSS) may be closer to the optimal choice of BSS-Plus. What is known about subject? This is the only study that the authors know of comparing the irrigation in vivo effect of irrigating solutions and their effect on retinal electrical function with supporting histological evidence with parameters that closely replicate actual human surgery. What are the findings? All findings are new and novel in this study. How might these results change the focus of research/clinical practice? As cost containment of medical care continues, there will be further efforts to find less expensive and accessible replacements for solutions and drugs that we use during surgery. What this study shows is that some solutions (BSS) may be comparable with the ‘gold standard’ (BSS-Plus) that we now use, and some may not be (Lactated Ringers). This is very important in protecting the patient from uninformed cost-cutting, which may be driven purely by saving money at the expense of the patient’s care.

hat this study shows is that some solutions (BSS) may be comparable with the ‘gold standard’ (BSS-Plus) that we now use, and some may not be (Lactated Ringers). This is very important in protecting the patient from uninformed cost-cutting, which may be driven purely by saving money at the expense of the patient’s care. Introduction Vitrectomy is the treatment of choice for posterior segment diseases that range from vitreous haemorrhage to malignant glaucoma. Machemer et al suggested using physiological saline for infusion in his original report.1 The metabolism of intraocular tissues like cornea, lens and retina depends on intraocular fluid composition, O2, pH, and osmolality. The best irrigating solution is important in achieving the best surgical outcome. The list of compounds in clinical or experimental settings started with normal saline. Solutions used include Lactated Ringers (LR), balanced salt solution (BSS) and balanced salt solution-plus (BSS-Plus). However, few in vivo studies compared the advantages/disadvantages of each of LR, BSS and BSS-Plus. Other solutions used in the past and experimentally investigated for use during vitreous surgery include Plasma-Lyte 1482, near physiological solution2 3 (see online supplemental tables), perfluorophenanthrene,4 perfluorotri-n-propylamine,5 polymethyl-3,3,3-trifluoropropylsiloxane-co-dimethylsiloxane,6 poly(2-hydroxyethyl acrylate),7 semifluorinated alkanes,8 silicone gel,9 methylated collagen,10 collagen/hyaluronic acid mixtures,11 hydroxypropylmethyl cellulose,12 crosslinked poly(vinyl) alcohol,13 polymethylacrylamidoglycolate methyl ester14 and crosslinked poly(1-vinyl-2-pyrrolidinone).15 However, in clinical practice, LR, BSS and BSS-Plus remain among the most commonly used solutions for vitrectomy. This study was a prospective, randomised, masked study. The aim of this study was to compare the effect of using three commonly used and commercially available types of irrigating solutions used during vitrectomy and their effect on postoperative electroretinogram and histology of operated animals compared with control eyes.

tudy was a prospective, randomised, masked study. The aim of this study was to compare the effect of using three commonly used and commercially available types of irrigating solutions used during vitrectomy and their effect on postoperative electroretinogram and histology of operated animals compared with control eyes. Animals and materials Thirty-six New Zealand White male rabbits weighing 3–5 lb were included for the purpose of this study in all surgeries. Twelve rabbits were used for testing each of the three irrigating solutions, with a total of 36 eyes. Intraocular infusion solutions compared in this study were LR, BSS (Alcon Laboratories, Fort Worth, Texas) and BSS-Plus (Alcon Laboratories). Anaesthesia Ketamine (35 mg/kg) and xylazine (10 mg/kg) were used as intramuscular injection for induction of anaesthesia, and a second intramuscular dose of ketamine (l8 mg/kg) and xylazine (5 mg/kg) was given during surgery. Topical anaesthetic and pupil dilating eye-drops were instilled over the rabbit cornea and in the conjunctival sac before surgery.

ylazine (10 mg/kg) were used as intramuscular injection for induction of anaesthesia, and a second intramuscular dose of ketamine (l8 mg/kg) and xylazine (5 mg/kg) was given during surgery. Topical anaesthetic and pupil dilating eye-drops were instilled over the rabbit cornea and in the conjunctival sac before surgery. Surgery The eye was proptosed and passed through a slit cut in a piece of sterile glove to maintain the proptosed position. A two-port vitrectomy was performed by a single surgeon (HAD). The surgeon was blinded from the type of solution. Nasal and temporal sclerotomies, using a 19-g Micro Vitreo Retinal blade, 1.5 mm from the limbus at 10 and 2 o’clock positions were used. A 20-g vitrectomy cutter probe (Alcon Laboratories) was inserted into one sclerotomy wound and a 20-g infusion light pipe was inserted into the second sclerotomy. An axial vitrectomy with approximately 60% of the vitreous was removed, subjectively estimated. No lensectomy, posterior vitreous detachment or other surgical procedures were performed in any of the rabbits. No retinal detachments occurred. The irrigating solution was allowed to be infused for 5 min, and the eye was closed using sclerotomy plugs for 15 min periods. The fluids were kept at room temperature. This cycle of infusing the eye for 5 min and keeping the fluid for 15 min was repeated three times for a total operation time of 1 hour. A standardised bottle height that allowed for retinal and optic nerve perfusion was used during vitrectomy in all surgeries. In total, approximately 60 mL of infusion fluid at room temperature was passed through the eye during surgery. The sclerotomies were then closed with 7-0 vicryl sutures and the eyes were checked for normal intraocular pressure.

t allowed for retinal and optic nerve perfusion was used during vitrectomy in all surgeries. In total, approximately 60 mL of infusion fluid at room temperature was passed through the eye during surgery. The sclerotomies were then closed with 7-0 vicryl sutures and the eyes were checked for normal intraocular pressure. Electrophysiology After the rabbits were anaesthetised, the pupils were dilated and topical anaesthetic was applied to the cornea. The animals were dark-adapted for at least 1 hour prior to recordings. The single bright flash electroretinography (ERG) was used. Many animals had preoperative ERGs, and the values of the preoperative ERGs were grouped so tightly that the 400 μV number was picked as the normal control reference value.

d to the cornea. The animals were dark-adapted for at least 1 hour prior to recordings. The single bright flash electroretinography (ERG) was used. Many animals had preoperative ERGs, and the values of the preoperative ERGs were grouped so tightly that the 400 μV number was picked as the normal control reference value. The ERG measurements were made at 2, 24 and 48 hours in all eyes. ERG measurements were also done at 6 and 72 hours postoperatively in some eyes. If the ERG had not returned to control values, additional ERGs were performed at 7 days and weekly thereafter. Dark-adapted ERGs were recorded by placing a cotton wick electrode on the cornea and a needle electrode on the skin above the eye. A ground electrode was attached to the ear. The electrodes were connected to a preamplifier with a half amplitude frequency set for a band pass of 0.1–2000 Hz. The output of the preamplifier was connected to an oscilloscope and to a computer for data storage. Stimuli were white light flashes of intensity setting of 16 for 10 min from a Grass instruments Model PS-22 photostimulator delivered at a rate of once per 3 min or longer in order to preserve the dark-adapted state between flashes. This stimulus is more than 5 log units above b-wave threshold; it was used to generate and record both a-wave and b-wave. ERGs were quantified by measuring the total ERG amplitude (a + b wave). The amplitude of the b-wave was measured from the peak of the a-wave to the peak of the b-wave. An ERG amplitude of 400 μV was considered a normal ERG as judged from preoperative normal control eye studies. None of the tested animals were diagnosed with surgical complications or an unusual postoperative course.

wave). The amplitude of the b-wave was measured from the peak of the a-wave to the peak of the b-wave. An ERG amplitude of 400 μV was considered a normal ERG as judged from preoperative normal control eye studies. None of the tested animals were diagnosed with surgical complications or an unusual postoperative course. Histology Biopsies were taken from the retina of 11 vitrectomised rabbit eyes 15 weeks postoperatively, and the retina was processed for H&E staining for histopathology. Statistical analysis Data were analysed using the Mann-Whitney U test.

wave). The amplitude of the b-wave was measured from the peak of the a-wave to the peak of the b-wave. An ERG amplitude of 400 μV was considered a normal ERG as judged from preoperative normal control eye studies. None of the tested animals were diagnosed with surgical complications or an unusual postoperative course. Histology Biopsies were taken from the retina of 11 vitrectomised rabbit eyes 15 weeks postoperatively, and the retina was processed for H&E staining for histopathology. Statistical analysis Data were analysed using the Mann-Whitney U test. Results Electrophysiology After analysing a number of ERGs (>10) from normal preoperative rabbit eyes, we observed tightly grouped values. We decided to establish a total amplitude (a-wave + b-wave) value of 400 μV as representing normal ERG amplitude. In subsequent experiments, we measured the time after surgery for the treated eyes to return to a normal (400 μV) level. Following vitrectomy and infusion with LR solution, all of these eyes had ERG values less than normal (around 36% of normal) at 2 and 24 hours (figure 1). After 48 hours and 72 hours, 1/12 (8.33%) and 3/12 (25%) of these eyes had normal ERGs, respectively. One week after surgery, 4/12 (33.3%) of these eyes had returned to normal levels. After 4 weeks, 6/12 (50%) of these eyes had normal ERGs and this number increased to 9/12 eyes (75%) at 10 weeks. In the remaining three eyes, no improvement in ERG amplitude was seen between 10 and 15 weeks, suggesting that these eyes were at their maximal improved level. In the BSS-treated and BSS-Plus-treated eyes, the ERGs recovered more rapidly after vitrectomy. At 2 hours, the amplitude of the recordings was reduced to approximately 35% of normal in all of these eyes (figure 1). A gradual increase in ERG amplitude was seen at 6, 24 and 48 hours postoperatively. At 24 hours, 4/12 (33.3%) of the BSS-irrigated and 6/12 (50%) of the BSS-Plus-irrigated eyes had normal ERGs. At 48 hours the number of rabbits with normal ERGs had increased to 8/12 (66.6%) in the BSS-irrigated group and 9/12 (75%) in the BSS-Plus-irrigated group. At 72 hours after vitrectomy, 12/12 (100%) BSS-Plus-infused eyes had ERGs with a total amplitude of 400 μV or greater, whereas 10/12 (83.3%) BSS-irrigated eyes demonstrated normal ERGs (figure 2). By 1 week, all (12/12) BSS eyes had also returned to normal. The statistical analysis of the effect of the three solutions on postoperative ERG in this report is based on total ERG (a+b wave) amplitude.

a total amplitude of 400 μV or greater, whereas 10/12 (83.3%) BSS-irrigated eyes demonstrated normal ERGs (figure 2). By 1 week, all (12/12) BSS eyes had also returned to normal. The statistical analysis of the effect of the three solutions on postoperative ERG in this report is based on total ERG (a+b wave) amplitude. Closer examination of the individual ERG recordings of the LR group revealed that the a-wave amplitude returned to a normal value in the majority of the eyes, even in the 25% of the LR-irrigated eyes where the total ERG amplitude had not returned to 80% of the normal value after 15weeks (figure 3). In conclusion, our results showed that differences between LR and BSS/BSS-Plus were statistically significant at 24, 48, 72 and 168 hours (p<0.05). No significant differences were found between BSS and BSS-Plus at any time point. However, looking at study rabbit eyes, the a-wave recordings showed that control a-waves had amplitudes around −180 μV previtrectomy. Twenty-four-hour postoperative recordings from each of the eyes had depressed a-wave values for all solutions used. Postoperative a-wave values for LR and BSS eyes were lower than for BSS-Plus eyes at 24 hours. These a-wave values increased over time back towards the control value of -180 μV in BSS-Plus and BSS eyes more rapidly than in the LR eyes. Figure 1 Temporal changes in average ERG postvitrectomy as a percentage of Normal ERG relative to the intraocular infusion fluids. L-Ringers, Latacted Ringers solution; BSS, balanced salt solution; BSS-Plus, balanced salt solution-plus; ERG, electroretinography.

Closer examination of the individual ERG recordings of the LR group revealed that the a-wave amplitude returned to a normal value in the majority of the eyes, even in the 25% of the LR-irrigated eyes where the total ERG amplitude had not returned to 80% of the normal value after 15weeks (figure 3). In conclusion, our results showed that differences between LR and BSS/BSS-Plus were statistically significant at 24, 48, 72 and 168 hours (p<0.05). No significant differences were found between BSS and BSS-Plus at any time point. However, looking at study rabbit eyes, the a-wave recordings showed that control a-waves had amplitudes around −180 μV previtrectomy. Twenty-four-hour postoperative recordings from each of the eyes had depressed a-wave values for all solutions used. Postoperative a-wave values for LR and BSS eyes were lower than for BSS-Plus eyes at 24 hours. These a-wave values increased over time back towards the control value of -180 μV in BSS-Plus and BSS eyes more rapidly than in the LR eyes. Figure 1 Temporal changes in average ERG postvitrectomy as a percentage of Normal ERG relative to the intraocular infusion fluids. L-Ringers, Latacted Ringers solution; BSS, balanced salt solution; BSS-Plus, balanced salt solution-plus; ERG, electroretinography. Figure 2 Comparing the number of rabbit eyes showing full recovery of ERG for each of LR, BSS and BSS-Plus in relation to time after surgery. L-Ringers, Latacted Ringers solution; BSS, balanced salt solution; BSS-Plus, balanced salt solution-plus; ERG, electroretinography.

Figure 1 Temporal changes in average ERG postvitrectomy as a percentage of Normal ERG relative to the intraocular infusion fluids. L-Ringers, Latacted Ringers solution; BSS, balanced salt solution; BSS-Plus, balanced salt solution-plus; ERG, electroretinography. Figure 2 Comparing the number of rabbit eyes showing full recovery of ERG for each of LR, BSS and BSS-Plus in relation to time after surgery. L-Ringers, Latacted Ringers solution; BSS, balanced salt solution; BSS-Plus, balanced salt solution-plus; ERG, electroretinography. Figure 3 Sample ERG recordings. The left graph shows an LR-infused eye showing subnormal ERG (b-wave below 200 μV) after 48 hours. The right graph shows a normal amplitude ERG after 48 hours using BSS solution. BSS, balanced salt solution; ERG, electroretinography; LR, Lactated Ringers. Histology results The study included 11 rabbit eyes: 7 LR, 2 BSS and 2 BSS-Plus eyes. The retina biopsies were taken and fixed 15 weeks after surgery. In the LR eyes, 2 had normal retinas and 5 of 7 had histological changes in the inner and outer retina that included vacuolisation of either the inner or outer retina or both (figure 4) up to complete disruption. Each of the two retinas from BSS-irrigated and BSS-Plus-irrigated eyes demonstrated no gross histological abnormalities. Figure 4 Photomicrograph of the retina of a vitrectomised rabbit eye. Lactated Ringers solution was used. There is vacuolisation within the inner and outer retina (arrows). ONL, outer nuclear layer specimens were stained with H&E.

Histology results The study included 11 rabbit eyes: 7 LR, 2 BSS and 2 BSS-Plus eyes. The retina biopsies were taken and fixed 15 weeks after surgery. In the LR eyes, 2 had normal retinas and 5 of 7 had histological changes in the inner and outer retina that included vacuolisation of either the inner or outer retina or both (figure 4) up to complete disruption. Each of the two retinas from BSS-irrigated and BSS-Plus-irrigated eyes demonstrated no gross histological abnormalities. Figure 4 Photomicrograph of the retina of a vitrectomised rabbit eye. Lactated Ringers solution was used. There is vacuolisation within the inner and outer retina (arrows). ONL, outer nuclear layer specimens were stained with H&E. Discussion During a vitrectomy operation the extracellular milieu of the retina is changed due to perfusion of vitreous cavity and exposure of the inner retina to solutions that differ in varying degrees in ionic and metabolite composition compared with the physiological vitreous. Even the outer retina in many situations, as in cases of macular holes, rhegmatogenous retinal detachments, giant retinal tears and retinectomies are exposed to intraocular irrigation solutions. Retinal function and physiology are not expected to be altered significantly for a prolonged period after vitrectomy. A return to normal function is expected to occur quickly as vitreous is mainly water. Cells, proteins and solutes compose less than 1% of the vitreous. Many researchers have investigated how intraocular irrigation solutions may affect the retina. At the biochemical level, Winkler et al and Negi et al reported that bicarbonate is essential for normal retinal function. 2 16 Manzanas et al observed that intraocular concentrations of protein increased to nearly the same extent after vitrectomy in rabbits using BSS-Plus, saline and LR without a statistical difference between each solution.17 They also observed that lactic acid (assumed to be produced mainly by the retina) increased with saline and LR but not with BSS-Plus. Ascorbic acid was highest in the vitreous if vitrectomy was performed using saline as intraocular irrigating solution. Ascorbic acid levels were also high when using BSS-Plus rather than LR if measured during the early postoperative period. Manzanas et al suggested that BSS-Plus could be more harmful to the ciliary body at that time. They speculated that oxidised glutathione, present in BSS-Plus, contributed to oxidative damage, leading to higher synthesis of ascorbic acid by the eye in order to change the oxidised to the reduced form of glutathione. At the clinical level, Saornil and Pastor demonstrated that retinal oedema on early postoperative fluorescein angiography was increased when LR or normal saline was used more than when BSS-Plus was used for vitrectomy in rabbits.

corbic acid by the eye in order to change the oxidised to the reduced form of glutathione. At the clinical level, Saornil and Pastor demonstrated that retinal oedema on early postoperative fluorescein angiography was increased when LR or normal saline was used more than when BSS-Plus was used for vitrectomy in rabbits. This oedema became more severe the longer the perfusion with the particular solution was maintained. However, they had a short follow-up and did not demonstrate how long this difference persisted.18 In our study, we investigated the effects of LR, BSS and BSS-Plus used as intraocular irrigating solutions on postoperative electrophysiology of the retina. Single bright flash (ERG) following at least 1 hour of dark adaptation was used to assess retinal function. We originally hypothesised that we would see a small difference among the three solutions, with BSS-Plus-irrigated eyes showing the fastest recovery. We were surprised at the length of time required for the LR-irrigated eyes to recover. It was even more concerning that some (25%) of the LR-infused eyes did not recover completely, even 15 weeks after surgery. Adding to the delayed recovery was the vacuolisation observed in these eyes on histological examination.

ery. We were surprised at the length of time required for the LR-irrigated eyes to recover. It was even more concerning that some (25%) of the LR-infused eyes did not recover completely, even 15 weeks after surgery. Adding to the delayed recovery was the vacuolisation observed in these eyes on histological examination. A significant decrease to 35% of a normal ERG (400 μV amplitudes) in a-wave and b-wave amplitudes was seen in all eyes when the ERG was recorded 2 hours after vitrectomy. This transient depression of ERG values noticed at 2 hours postoperatively with the three solutions can be multifactorial, with irrigating solutions, anaesthesia, or yet other factors that may not have been recorded and analysed in this report. Although there might be some differences in ERG response between rabbits and the eyes within the same rabbit, the sustained difference between test eyes and control value over our measurements minimised the significance of this phenomenon. In contrast, Moorhead et al reported that BSS-Plus-infused eyes demonstrated a smaller reduction in ERG amplitude than BSS-treated or normal saline-treated eyes. The magnitude of the decrease in ERG amplitude postoperatively was much larger in our study. This may be due to the increased time of vitrectomy (60 min vs 10 min). It may also be due to the larger surface area of retinacovered by only a thin cortical vitreous remnant which is exposed to infusion fluid current during the operation with our more complete vitrectomy technique, as compared to previous studies(single-port (Moorhead et al), vs. two-port vitrectomy in our study).

min). It may also be due to the larger surface area of retinacovered by only a thin cortical vitreous remnant which is exposed to infusion fluid current during the operation with our more complete vitrectomy technique, as compared to previous studies(single-port (Moorhead et al), vs. two-port vitrectomy in our study). In this study, the serial postoperative ERG recordings showed gradual recovery. However, many of the ERG amplitudes returned to 90% of control values at 24–48 hours when using BSS and BSS-Plus. BSS-Plus showed an advantage over BSS at 48 and 72 hours, nonetheless not statistically significant. This is in agreement with the findings of Moorhead et al. By 72 hours, 83.3% of BSS and 100% BSS-Plus eyes had ERGs of normal amplitude. At 1 week, the ERGs of all eyes of these two groups showed normal values. In contrast, 1 of 12 of the LR-infused eyes had recovered to normal within the first 48 hours. Additionally, 75% of the LR eyes had recovered when measured 10 weeks after surgery. The remaining 25% of the LR-irrigated eyes had ERGs 20%–40% less than normal amplitudes even out to 15 weeks after vitrectomy. However, the authors could not explain why the SD in recovering LR eyes showed a tighter SD at 168 hours, similar to the level at 2 hours. It was expected that the data would be more disparate at 168 hours after surgery. Moorhead et al showed that postoperative ERG b-waves were better using BSS and BSS-Plus versus saline as intraocular irrigating solutions during vitrectomy in pigmented rabbits. BSS-Plus was best, however, not with statistically significant values. They reported that at 48 hours, normal saline-infused eyes had near complete recovery, with no statistical significance at 2 days postoperatively among the three solutions. This is in contrast to reports on biochemical effects, retinal oedema, and in vitro ERG depression associated with the use of normal saline in experiments previously performed: Negi et al; Manzanas et al; and, Saornil and Pastor.16–18 This is the first report, to our knowledge, on the in vivo effects of using LR as an intraocular irrigating solution. In this study LR was compared with the other most commonly used irrigating fluids in clinical practice (BSS and BSS-Plus). We also avoided any additional surgical procedures in contrast to cryopexy 2 weeks prior to vitrectomy reported by Moorhead et al. A prolonged vitrectomy time was applied (1 hour vs 10 min), closer to the actual vitrectomy surgery time.

other most commonly used irrigating fluids in clinical practice (BSS and BSS-Plus). We also avoided any additional surgical procedures in contrast to cryopexy 2 weeks prior to vitrectomy reported by Moorhead et al. A prolonged vitrectomy time was applied (1 hour vs 10 min), closer to the actual vitrectomy surgery time. More vitreous (subjectively estimated 60% vs 30%) was removed, and two separate infusion and suction ports were used ensuring exposure of a wider area of retinal surface to the irrigating fluid solution. In Moorhead et al the investigators used a single infusion-suction-cutter probe with fewer animals.3 In our study, we also used 400 μV as an estimated normal ERG amplitude derived from preoperative eyes (recordings) as control. BSS-Plus contains bicarbonate, oxidised glutathione and glucose, and has a fixed pH of 7.4. These components were demonstrated to have a protective effect on corneal endothelium in previous studies (Benson et al and Waltman et al).19 20

imated normal ERG amplitude derived from preoperative eyes (recordings) as control. BSS-Plus contains bicarbonate, oxidised glutathione and glucose, and has a fixed pH of 7.4. These components were demonstrated to have a protective effect on corneal endothelium in previous studies (Benson et al and Waltman et al).19 20 However, attached retina is expected to be affected less by the type of intraocular irrigating fluid as diffusion through choroidal circulation remains a source of ions and metabolites. This is in contrast to corneal endothelium, whose supplies are restricted to the aqueous humour. We think that LR might have a permanent toxic effect on the retinal cells due to lower pH, higher lactic acid and lower osmolality. This might cause harmful retinal oedema, as previously reported by Manzanas et al, or some other type of damage to bipolar, Mueller or other inner retinal cells that contribute to the ERG b-wave.17 Our limited histological specimens corroborate our electrophysiological findings. This study is not free of some shortcomings: First, analysing b-wave and a-wave ratio and/or recording ERG using the International Society for Clinical Electrophysiology of Vision (ISCEV) standard would have been more informative, as we used total ERG amplitude for the analysis. It would have been better to include more serial histological examinations at different points of time. Further studies that include intraoperative ERG recordings and preoperative and postoperative retinal/choroidal optical coherence tomography recordings would also be more informative. In this report an axial vitrectomy was performed that removed around 60% of the vitreous, and there was a possibility that a peripheral layer of vitreous cortex was not removed in these rabbits, such that in this test environment any infusion fluid might only come into limited direct contact with the retina itself. In turn, the extent of the infusion fluid/retina interaction will be mediated by the residual cortical vitreous either through diffusion and/or hydration. To the best of our knowledge, the solution came into contact with the retina, although certainly not as much as if the posterior cortical vitreous were able to be removed. A much more direct contact between infusion fluid and retina in the human setting where the cortical vitreous has likely already been removed for the most part of any vitrectomy operation might lead to a more pronounced effect of the infusion solution.

much as if the posterior cortical vitreous were able to be removed. A much more direct contact between infusion fluid and retina in the human setting where the cortical vitreous has likely already been removed for the most part of any vitrectomy operation might lead to a more pronounced effect of the infusion solution. In addition, studies comparing newer irrigating solutions are needed. In conclusion, vitrectomy in a standardised experimental setting comparing three infusion solutions showed compromising effects on ERGs and retinal histology from LR solution infusion. The benefit of BSS-Plus solution over BSS solution infusion was less marked. 10.1136/bmjophth-2016-000004.supp1Supplementary file Contributors: PF: gave the idea, shared in the primary planning/set-up and data analysis, supervised, read and helped in the corrections of the manuscript. GW: supervision, expenses, shared in manuscript revision and data retrieval. MH: shared in research plan and set-up, assisted in work, performed data acquisition, data statistical analysis and manuscript revision. HAD: performed surgeries, assisted in data acquisition, and wrote and submitted the manuscript. The previous persons are listed as authors of the manuscript. Funding: This research received no specific grant from any funding agency that involves any of the authors. Competing interests: None declared. Provenance and peer review: Not commissioned; externally peer reviewed.

Key messages Glaucoma is a multifactorial disorder of the eye and the leading cause of blindness worldwide, posing a major public health challenge. As such, it is important to assess the relationship between anthropometric measures and vertical cup-to-disc ratio (VCDR). This study demonstrated different associations between VCDR and other factors, stratified by sex. These results include important findings that may help to clarify the reason for higher proportions of normal tension glaucoma around Asia, and to find novel therapeutic options in other countries for refractory glaucoma in which the visual field defect worsens despite treatment with topical and/or surgical therapy. Moreover, these results may shed light on the different mechanisms leading to glaucoma in men and women.

normal tension glaucoma around Asia, and to find novel therapeutic options in other countries for refractory glaucoma in which the visual field defect worsens despite treatment with topical and/or surgical therapy. Moreover, these results may shed light on the different mechanisms leading to glaucoma in men and women. Introduction Glaucoma, a multifactorial optic neuropathy for which the leading aetiology is high intraocular pressure (IOP), is the main cause of irreversible visual loss, which ultimately leads to blindness.1 In a large population-based study, it was found that the relative risk of glaucoma optic nerve damage was 13 times higher for an IOP of 22–29 mm Hg and 39 times higher for an IOP of more than 30 mm Hg in comparison to an IOP of less than or equal to 15 mm Hg.2 It is important in the early stages of glaucoma to detect ophthalmic structural changes, particularly in the retinal ganglion cell complex and the optic disc because the visual field defects in glaucoma are often detected only after 20% of the axons are lost.3 Also, the vertical cup-to-disc ratio (VCDR) is a very useful clinical measurement for detecting glaucoma because it increases earlier than the horizontal cup-to-disc ratio, especially in the early to moderate glaucoma stages.4

he visual field defects in glaucoma are often detected only after 20% of the axons are lost.3 Also, the vertical cup-to-disc ratio (VCDR) is a very useful clinical measurement for detecting glaucoma because it increases earlier than the horizontal cup-to-disc ratio, especially in the early to moderate glaucoma stages.4 IOP, which is believed to be a major risk factor for glaucoma and also the only modifiable one, can be influenced by central corneal thickness (CCT). Systolic blood pressure and body mass index (BMI) were also reported to have a positive correlation to IOP.5–7 Additionally, positive associations between BMI and blood pressure have been documented in several ethnicities.8 Several reports suggest that a high BMI protects against glaucoma, whereas a low BMI has a possibly greater risk factor.9–11 They hypothesised that higher circulating oestrogen and other paracrine factors, which fatty tissue produces, bind to oestrogen receptors expressed on retinal ganglion cells and may lead to neuroprotection. The interrelations between IOP, CCT, BMI, body fat percentage (BFP) and VCDR are complex. It is important to assess the relation between anthropometric measures and VCDR. However, few cross-sectional and prospective studies have been conducted that include BMI and BFP. The objective of this study is to clarify the relationship between VCDR and IOP, CCT, BMI or BFP (after adjusting for age, optic disc area, systolic blood pressure and diastolic blood pressure, all of which are influential factors in the middle-aged and elderly Japanese population.)

conducted that include BMI and BFP. The objective of this study is to clarify the relationship between VCDR and IOP, CCT, BMI or BFP (after adjusting for age, optic disc area, systolic blood pressure and diastolic blood pressure, all of which are influential factors in the middle-aged and elderly Japanese population.) Methods Study design and participants Study participants were selected from the third wave examination (May 2002–June 2004) of the National Institute for Longevity Sciences — Longitudinal Study of Aging (NILS-LSA). The NILS-LSA is a longitudinal, dynamic prospective cohort study that includes medical, physiological, nutritional and psychological examinations. A total of 2378 men and women between the age of 40 and 84 (inclusively) were randomly selected from a central region of Japan: Obu and Higashiura. Obu, which possesses the characteristics of an urban area, and Higashiura, which possesses the characteristics of a rural area, are located in the prefecture of Aichi. We selected 1428 right eyes with no surgical history (736 men, 692 women; average age, 59.3±11.7 years) for univariate analysis and 2819 eyes with no surgical history (788 men, 739 women; average age, 59.6±11.7 years) for multivariate analysis in order to remove the impacts of any ophthalmic surgery.

efecture of Aichi. We selected 1428 right eyes with no surgical history (736 men, 692 women; average age, 59.3±11.7 years) for univariate analysis and 2819 eyes with no surgical history (788 men, 739 women; average age, 59.6±11.7 years) for multivariate analysis in order to remove the impacts of any ophthalmic surgery. Ocular parameters CCT was obtained with specular microscope (SP-2000; Topcon Corporation, Tokyo, Japan). VCDR and optic disc area were assessed using the Heidelberg Retina Tomograph II (HRT; Heidelberg Engineering, Heidelberg, Germany). IOP of each eye was measured three times from 09:00 to 12:00 using a non-contact tonometer (NT-3000, Nidek Technologies, Gamagori, Japan). Other parameters Blood pressure was measured at the upper right arm in a sitting position. Height and weight were measured with a lightweight examination gown, but without shoes. BFP was measured by dual-energy X-ray absorptiometry scanners (QDR4500A; Hologic, Bedford, Massachusetts, USA). BMI was calculated by dividing the square of the height (square metres) from the weight (kilograms).

a sitting position. Height and weight were measured with a lightweight examination gown, but without shoes. BFP was measured by dual-energy X-ray absorptiometry scanners (QDR4500A; Hologic, Bedford, Massachusetts, USA). BMI was calculated by dividing the square of the height (square metres) from the weight (kilograms). Statistical analysis The data were processed and analysed using the Statistical Analysis System (SAS V.9.3). We used four age groups (40s, 50s, 60s, 70s) from the fourth decade group (40–49) to the seventh decade group (70 and older). Differences in VCDR, IOP, CCT, optic disc area, BMI, BFP, systolic blood pressure and diastolic blood pressure between men and women were tested by two-sample Student's t-test. VCDR was classified by cup-to-disc ratio (CDR) classes (CD1, VCDR <0.3; CD2, 0.3 ≤ VCDR <0.6; CD3, 0.6 ≤ VCDR <0.7; CD4, 0.7 ≤ VCDR <0.9; CD5, 0.9 ≤ VCDR).12 13 Only the right eye data of each subject were evaluated in this analysis. To analyse the association between CD1 to CD3 classes and age groups, cross-tabulation, χ2 tests and residual analysis were carried out in the order named. CD4 and CD5 groups were excluded in the analysis because χ2 tests require that there are at least five data points in the class. In addition, the stratified table including age groups and CDR classes was analysed using the Cochran-Mantel-Haenszel statistic.

sts and residual analysis were carried out in the order named. CD4 and CD5 groups were excluded in the analysis because χ2 tests require that there are at least five data points in the class. In addition, the stratified table including age groups and CDR classes was analysed using the Cochran-Mantel-Haenszel statistic. Then VCDR was analysed using a general linear mixed model, a type of statistical analysis used for repeated measurements and binomial values of a single individual. Both eye data of each subject were evaluated in this analysis. The analysis was applied using the SAS procedure PROC MIXED with RANDOM INTERCEPT/SUB=ID. Explanatory variables were IOP, CCT, BMI and BFP. The data were analysed by sex and adjusted for age, optic disc area, systolic blood pressure and diastolic blood pressure. Also, predicted VCDR was calculated using the ESTIMATE command as IOP, CCT, BMI and BFP were varied from the average to the average ± SD. We applied Bonferroni's correction for multiple comparisons under the assumption that each analysis was not independent. The level of significance in all analyses was set at p<0.05. Results

Then VCDR was analysed using a general linear mixed model, a type of statistical analysis used for repeated measurements and binomial values of a single individual. Both eye data of each subject were evaluated in this analysis. The analysis was applied using the SAS procedure PROC MIXED with RANDOM INTERCEPT/SUB=ID. Explanatory variables were IOP, CCT, BMI and BFP. The data were analysed by sex and adjusted for age, optic disc area, systolic blood pressure and diastolic blood pressure. Also, predicted VCDR was calculated using the ESTIMATE command as IOP, CCT, BMI and BFP were varied from the average to the average ± SD. We applied Bonferroni's correction for multiple comparisons under the assumption that each analysis was not independent. The level of significance in all analyses was set at p<0.05. Results Figure 1 shows the proportion of age groups in each CDR class. The highest proportions of CD1, CD2, CD3, CD4 and CD5 were 27.1% (found in the fourth decade group), 30% (in the fifth decade group), 31% (in the sixth and seventh decade groups), 54.8% (in the seventh decade group) and 100% (in the seventh decade group), respectively. In χ2 tests and residual analysis, there was a significantly lower proportion of seventh decade participants and a higher proportion of fourth and sixth decade participants in CD1. There was a significantly lower proportion of seventh decade participants and a higher proportion of fifth decade participants in CD2, and a significantly lower proportion of fourth, fifth and sixth decade participants, as well as a higher proportion of seventh decade participants in CD3. There were no significant differences between the observation values and the expected values of the fifth decade participants in CD1, and of the fourth and sixth decade participants in CD2. Greater CDR classes are linked to significantly older age groups (p<0.0001, Cochran-Mantel-Haenszel test).

de participants in CD3. There were no significant differences between the observation values and the expected values of the fifth decade participants in CD1, and of the fourth and sixth decade participants in CD2. Greater CDR classes are linked to significantly older age groups (p<0.0001, Cochran-Mantel-Haenszel test). Figure 1 The proportion of age groups in each cup-to-disc ratio (CDR) class. From the results of residual analysis, white arrowheads pointing up (△) indicate that observation value is significantly more than the expectation value, while black arrowheads pointing down (▼) indicate that observation value is significantly less than expectation value. Mean values of IOP, CCT, VCDR, optic disc area, BMI, BFP, systolic blood pressure and diastolic blood pressure were summarised by age group and gender (table 1). Table 1 Mean values ±SD of VCDR, IOP, CCT, optic disc area, BMI, systolic blood pressure, diastolic blood pressure and BFP

Figure 1 The proportion of age groups in each cup-to-disc ratio (CDR) class. From the results of residual analysis, white arrowheads pointing up (△) indicate that observation value is significantly more than the expectation value, while black arrowheads pointing down (▼) indicate that observation value is significantly less than expectation value. Mean values of IOP, CCT, VCDR, optic disc area, BMI, BFP, systolic blood pressure and diastolic blood pressure were summarised by age group and gender (table 1). Table 1 Mean values ±SD of VCDR, IOP, CCT, optic disc area, BMI, systolic blood pressure, diastolic blood pressure and BFP Age 40–49 50–59 60–69 70– Total Gender Male Female p Value Male Female p Value Male Female p Value Male Female p Value Male Female p Value Number ofsubjects 165 191 204 183 198 180 169 138 736 692 VCDR 0.316±0.181 0.283±0.178 0.0822 0.334±0.189 0.295±0.199 0.0475 0.295±0.210 0.308±0.228 0.5482 0.378±0.229 0.307±0.207 0.0051 0.330±0.205 0.298±0.202 0.003 IOP(mm Hg) 13.7±2.5 12.4±2.6 <0.0001 13.3±2.5 12.7±2.5 0.0221 12.7±2.6 12.6±2.3 0.7421 12.4±2.4 12.4±2.3 0.915 13.0±2.6 12.6±2.4 0.0004 CCT(mm) 0.52±0.03 0.51±0.03 0.0014 0.52±0.04 0.51±0.04 0.0008 0.52±0.03 0.51±0.03 0.7806 0.51±0.03 0.51±0.02 0.7169 0.52±0.03 0.51±0.03 0.0002 Discarea(mm2) 2.7±0.6 2.6±0.6 0.0807 2.7±0.5 2.7±0.5 0.9627 2.7±0.5 2.6±0.5 0.5843 2.8±0.5 2.7±0.5 0.0076 2.7±0.6 2.7±0.5 0.0145 BMI(kg/m2) 23.2±2.8 21.9±3.4 <0.0001 23.4±2.7 22.7±3.2 0.0399 23.4±2.8 22.8±3.0 0.044 22.5±2.9 23.0±3.4 0.1847 23.1±2.8 22.6±3.3 0.0006 Systolic bloodpressure (mm Hg) 117.2±15.0 108.5±16.5 <0.0001 121.3±15.0 116.0±20.1 0.0043 124.7±17.2 120.9±20.0 0.0481 125.6±18.2 128.7±22.5 0.1614 122.4±17.0 118.2±21.0 <0.0001 Diastolicbloodpressure (mm Hg) 74.0±9.5 67.0±10.5 <0.0001 76.5±10.1 71.9±11.6 <0.0001 77.4±9.6 73.5±11.0 0.0002 75.2±10.8 75.9±12.5 0.5948 75.9±10.1 71.9±11.8 <0.0001 BFP (%) 20.1±4.3 28.5±5.0 <0.0001 20.7±4.8 30.2±4.9 <0.0001 21.2±4.7 31.0±4.5 <0.0001 21.4±4.6 31.3±5.5 <0.0001 21.0±4.6 30.2±5.1 <0.0001 BFP, body fat percentage; BMI, body mass index; CCT, central corneal thickness; IOP, intraocular pressure; VCDR, vertical cup-to-disc ratio.

9±12.5 0.5948 75.9±10.1 71.9±11.8 <0.0001 BFP (%) 20.1±4.3 28.5±5.0 <0.0001 20.7±4.8 30.2±4.9 <0.0001 21.2±4.7 31.0±4.5 <0.0001 21.4±4.6 31.3±5.5 <0.0001 21.0±4.6 30.2±5.1 <0.0001 BFP, body fat percentage; BMI, body mass index; CCT, central corneal thickness; IOP, intraocular pressure; VCDR, vertical cup-to-disc ratio. VCDR of men in the fifth and seventh decade groups was significantly greater than that of women in the corresponding groups. In the fourth and fifth decade groups, IOP and CCT were significantly greater for men than for women. Disc area was significantly greater for men than for women in only the seventh decade group. BMI, systolic blood pressure and diastolic blood pressure were significantly greater for men than for women in the fourth to sixth decade groups. On the other hand, BFP was significantly less for men than for women among all decade groups. The relationship between VCDR and IOP, CCT, BFP or BMI by gender was analysed in a general linear mixed model after being adjusted for age, systolic and diastolic blood pressure, and optic disc area (table 2). Table 2 Statistical results of the relationship between vertical cup-to-disc ratio (VCDR) and several explanatory valuables by sex in general linear mixed model analysis after adjusting for age, optic disc area, systolic blood pressure and diastolic blood pressure

The relationship between VCDR and IOP, CCT, BFP or BMI by gender was analysed in a general linear mixed model after being adjusted for age, systolic and diastolic blood pressure, and optic disc area (table 2). Table 2 Statistical results of the relationship between vertical cup-to-disc ratio (VCDR) and several explanatory valuables by sex in general linear mixed model analysis after adjusting for age, optic disc area, systolic blood pressure and diastolic blood pressure Sex Explanatory variables Estimated effects SE  p Value Male IOP 0.0018 0.0025 0.4768 CCT −0.1916 0.2010 0.3407 BMI −0.0059 0.0029 0.0426 BFP 0.0011 0.00172 0.5166 Female IOP 0.0125 0.0027 <0.0001 CCT −0.4229 0.2081 0.0425 BMI −0.0026 0.0028 0.3658 BFP 0.00005 0.0018 0.9798 BFP, body fat percentage;BMI,body mass index; CCT, centre corneal thickness;IOP, intraocular pressure.

ale IOP 0.0018 0.0025 0.4768 CCT −0.1916 0.2010 0.3407 BMI −0.0059 0.0029 0.0426 BFP 0.0011 0.00172 0.5166 Female IOP 0.0125 0.0027 <0.0001 CCT −0.4229 0.2081 0.0425 BMI −0.0026 0.0028 0.3658 BFP 0.00005 0.0018 0.9798 BFP, body fat percentage;BMI,body mass index; CCT, centre corneal thickness;IOP, intraocular pressure. VCDR significantly increased with a low BMI among men. Meanwhile, VCDR significantly increased with a thin CCT and a high IOP among women. No significant relationship was found between VCDR and CCT, IOP or BFP among men, as well as VCDR and BMI or BFP among women. However, only the relationship between IOP and VCDR in women would be statistically significant after applying Bonferroni's correction for multiple comparisons under the assumption that each analysis was not independent. Figure 2 shows the predicted VCDR after being adjusted for age, optic disc area and systolic and diastolic blood pressure when each of the explanatory variables varied from average to average ±SD. The predicted VCDR has a positive trend against IOP and a negative trend against CCT and BMI. We have added the analysis to see whether the significant association of greater VCDR is with BMI and also with weight and height because BMI is a ratio of weight and height. However neither height nor weight was found to be a significant determinant of greater VCDR after adjusting for age, gender and disc area in a stepwise multiple regression analysis.

ether the significant association of greater VCDR is with BMI and also with weight and height because BMI is a ratio of weight and height. However neither height nor weight was found to be a significant determinant of greater VCDR after adjusting for age, gender and disc area in a stepwise multiple regression analysis. Figure 2 Predicted vertical cup-to-disc ratio (VCDR) when each of the explanatory variables, intraocular pressure (IOP), central corneal thickness (CCT), body mass index (BMI) and body fat percentage (BFP), was varied from average to average ±SD, as calculated by linear mixed model by sex. Discussion Glaucoma is a major public health problem and the leading cause of blindness across the world and a progressive optic neuropathy characterised by both ocular morphological changes, such as optic nerve disc cupping or loss of retinal ganglion cell, and ocular functional changes, such as visual field defects that correspond with the loss of the area of the optic neuronal rim. Some of the risk factors for primary open-angle glaucoma (POAG), such as high IOP,14 advancing age,15 sex,16 family history,17 myopia,18 race,19 high systolic blood pressure, low diastolic blood pressure20 and diabetes mellitus,21 have been widely investigated and described. Considered a multifactorial disorder, several theories for the pathogenesis of glaucomatous optic neuropathy have been described as a mechanical compression,22 a vascular23 and a genetic theory.24

stolic blood pressure, low diastolic blood pressure20 and diabetes mellitus,21 have been widely investigated and described. Considered a multifactorial disorder, several theories for the pathogenesis of glaucomatous optic neuropathy have been described as a mechanical compression,22 a vascular23 and a genetic theory.24 Specific examination of the optic disc is a very important part of the assessment in patients with glaucoma. There is a push for more objective optic disc assessment, such as the HRT, ocular coherence tomography and GDx nerve fibre analyser. HRT is an established imaging device that creates both accurate and reproducible measurements of optical nerve head parameters, including VCDR, for the diagnosis and monitoring of glaucoma.25 VCDR was classified by five CDR classes, including 0.7 (VCDR >97.5th percentile), the value that was used as glaucoma diagnosis according to the International Society of Geographical and Epidemiological Ophthalmology criteria.13 There was a higher proportion of older groups and a lower proportion of younger groups in the upper CDR class. Conversely there was a lower proportion of older groups and a higher proportion of younger groups in the lower CDR class. Overall there was a significant positive relationship between CDR class and age group.

There was a higher proportion of older groups and a lower proportion of younger groups in the upper CDR class. Conversely there was a lower proportion of older groups and a higher proportion of younger groups in the lower CDR class. Overall there was a significant positive relationship between CDR class and age group. In univariate analysis, IOP and CCT were significantly greater for men than for women in the fourth and fifth decade groups. IOP is influenced by many factors, such as CCT and corneal biomechanical properties.26 CCT especially has a significant influence on IOP measurement. It has been reported that thinner corneas result in artificially lower IOP readings and that thicker corneas cause artificially high IOP readings. This is because the thick cornea has greater resistance to corneal shape distortion against external pressure than the thinner cornea, leading to the projected IOP reading being less than the real IOP in a thin CCT.27

sult in artificially lower IOP readings and that thicker corneas cause artificially high IOP readings. This is because the thick cornea has greater resistance to corneal shape distortion against external pressure than the thinner cornea, leading to the projected IOP reading being less than the real IOP in a thin CCT.27 IOP, BMI, systolic blood pressure and diastolic blood pressure were significantly greater for men than for women in the fourth and fifth decade groups. Pasquale et al reported that BMI had a significantly positive relation with IOP after adjusting for several parameters.9 Previous reports explained that increasing intraorbital fat tissue caused a high intraorbital pressure, an increase in episcleral venous pressure and a consequent decrease in outflow facility. As a result, IOP may increase. Obesity increases blood viscosity through increasing red cell count, haemoglobin and haematocrit, as well as consequently increased outflow resistance of episcleral veins, resulting in increased IOP.28 29 Elevated blood pressure increases IOP by increasing ciliary artery pressure and ultrafiltration of the aqueous humour.30 31 In addition, some studies have documented positive associations between BMI and blood pressure in several ethnicities.8

sed outflow resistance of episcleral veins, resulting in increased IOP.28 29 Elevated blood pressure increases IOP by increasing ciliary artery pressure and ultrafiltration of the aqueous humour.30 31 In addition, some studies have documented positive associations between BMI and blood pressure in several ethnicities.8 VCDR and optic disc area were significantly greater for men than for women in the seventh decade group. Several researchers reported positive association between VCDR and optic disc area in normal and glaucomatous eyes.32–34 Because several factors are intricately related to each other, multivariate analyses were performed using linear mixed model adjusted for age, optic disc area and blood pressure.

eventh decade group. Several researchers reported positive association between VCDR and optic disc area in normal and glaucomatous eyes.32–34 Because several factors are intricately related to each other, multivariate analyses were performed using linear mixed model adjusted for age, optic disc area and blood pressure. In multivariate analysis, only BMI was a significant VCDR-related factor in men after adjusting for age, optic disc area and blood pressure. Even though BMI is a number based on weight and height, neither height nor weight was found to be a significant determinant of greater VCDR after adjusting for age, gender and disc area in a stepwise multiple regression analysis. Further studies will be needed in order to explore this relationship and the meaning of BMI. In contrast, IOP and CCT were significant VCDR-related factors in women after adjusting for age, optic disc area and blood pressure. Several authors described that CCT was significantly thinner in normal tension glaucoma (NTG) than in POAG or normal subjects.35 36 At a glance, BMI and CCT were negatively related to VCDR, while IOP was positively related to VCDR (figure 2). A large epidemiological study suggested that the significant factors of elevated VCDR were increased age, being male, higher IOP, lower diastolic blood pressure, lower BMI and previous cataract surgery.37 Previous studies suggested that higher IOP was the most significant factor of elevated VCDR. However, our data suggested that the most significant factor of elevated VCDR was IOP in women, but BMI in men. For this possible reason, a large population-based epidemiological study reported that the Japanese population seems to have a higher incidence rate of NTG.38 39 NTG is a common form of POAG and considered a mechanism shared with both conditions.40 However, IOP seems to be a relatively lower potential factor in NTG than in POAG. A low BMI may be related to NTG for men. Pasquale et al 9 reported that higher BMI among women was associated with a low risk of NTG (IOP ≤21 mm Hg). Also, a population-based study in Singapore found that men generally had greater VCDR than women and that lower BMI among both sexes was associated with greater VCDR.37 These studies hypothesised that higher oestrogen levels from fatty tissue may lead to neuroprotection.

sociated with a low risk of NTG (IOP ≤21 mm Hg). Also, a population-based study in Singapore found that men generally had greater VCDR than women and that lower BMI among both sexes was associated with greater VCDR.37 These studies hypothesised that higher oestrogen levels from fatty tissue may lead to neuroprotection. On the other hand, some researchers revealed a statistically significant correlation between elevated cerebrospinal fluid pressure (CSFP) levels and higher BMI.41 42 They concluded that a depressed CSFP caused by low BMI might be a potential risk factor for NTG. We speculated that in men BMI levels are inversely related to glaucoma because BMI is inversely related to VCDR. Our hypothesis is that CSFP levels, not oestrogen levels from body fat tissue, affect VCDR since we could not find out the relationship between VCDR and BFP in our multivariate analysis. IOP may be a relatively low risk factor in NTG compared with POAG, though. Considering the possible mechanism of a translaminar pressure gradient and treatment difficulties in NTG, we believe there must be other factors contributing.

On the other hand, some researchers revealed a statistically significant correlation between elevated cerebrospinal fluid pressure (CSFP) levels and higher BMI.41 42 They concluded that a depressed CSFP caused by low BMI might be a potential risk factor for NTG. We speculated that in men BMI levels are inversely related to glaucoma because BMI is inversely related to VCDR. Our hypothesis is that CSFP levels, not oestrogen levels from body fat tissue, affect VCDR since we could not find out the relationship between VCDR and BFP in our multivariate analysis. IOP may be a relatively low risk factor in NTG compared with POAG, though. Considering the possible mechanism of a translaminar pressure gradient and treatment difficulties in NTG, we believe there must be other factors contributing. In conclusion, we observed a significant negative relationship between VCDR and BMI for men and VCDR and CCT for women, as well as a significant positive relationship between VCDR and IOP for women in a large population-based study. Male participants with higher BMI had elevated VCDR than those with lower BMI. Female participants with thinner CCT had elevated VCDR than those with thicker CCT. Also, female participants with higher IOP had elevated VCDR than those with lower IOP. These are important findings that may help to clarify the reason for higher proportions of NTG in Asia (including Japan) and to find novel therapeutic options for refractory glaucoma in which the visual field defect worsens despite treatment with topical and/or surgical therapy. Moreover, it may shed light on the different mechanisms leading to glaucoma in men and women.

the reason for higher proportions of NTG in Asia (including Japan) and to find novel therapeutic options for refractory glaucoma in which the visual field defect worsens despite treatment with topical and/or surgical therapy. Moreover, it may shed light on the different mechanisms leading to glaucoma in men and women. Acknowledgements: This research was performed in accordance with the tenets of the Declaration of Helsinki. Prior informed consent was obtained from all participants. Contributors: HF and HS: Conception and design of study. HF: Analysis and interpretation, writing the article and literature search. CT: Critical revision of the article and statistical expertise. FA: Final approval of the article. RO: Data collection, and provision of materials, patients or resources. HS: Administrative, technical or logistic support. Funding: This work was supported in part by grants from the Research Fund for Longevity Sciences from the National Center for Geriatrics and Gerontology, Japan (25-22). Competing interests: None declared. Ethics approval: The NILS-LSA was approved by the ethics committee of the National Center for Geriatrics and Gerontology. Provenance and peer review: Not commissioned; externally peer reviewed.

Key messages Studies on pre-eclampsia and retinopathy of prematurity (ROP) show variable results. Some are from earlier periods when nutritional practices, mechanical ventilation, supplemental oxygen and antenatal corticosteroid use were different from current standard of care. Some studies have included cases of maternal hypertension other than pre-eclampsia in their cohorts, there is marked variation in the gestational ages and some have substantial missing data. In a well-defined cohort from a recent era, we demonstrate pre-eclampsia is not associated with ROP, either as a protective or as risk factor. There were no missing data and high level of antenatal corticosteroid use. Unlike previous studies, chorioamnionitis was defined pathologically, and oxygen saturation targets and mechanical ventilation strategies were consistent during this period. This is a retrospective cohort study from a single centre potentially subject to confounding. Prematurity and intrauterine growth restriction, whether in pre-eclamptic or normotensive pregnancies, are major risk factors for the development of ROP.

In a well-defined cohort from a recent era, we demonstrate pre-eclampsia is not associated with ROP, either as a protective or as risk factor. There were no missing data and high level of antenatal corticosteroid use. Unlike previous studies, chorioamnionitis was defined pathologically, and oxygen saturation targets and mechanical ventilation strategies were consistent during this period. This is a retrospective cohort study from a single centre potentially subject to confounding. Prematurity and intrauterine growth restriction, whether in pre-eclamptic or normotensive pregnancies, are major risk factors for the development of ROP. Introduction Retinopathy of prematurity (ROP), a vasoproliferative disorder of the developing retina, is one of the more severe consequences of preterm birth and a major cause of childhood blindness and visual impairment in the developing and developed world.1 The disease is more common in infants less than 31 weeks’ gestation with infants of lesser gestation at higher risk and severity of ROP. In infants with a birth weight of <1500 g, the reported incidence of ROP ranges from 20% to 50% in different populations.1–3

visual impairment in the developing and developed world.1 The disease is more common in infants less than 31 weeks’ gestation with infants of lesser gestation at higher risk and severity of ROP. In infants with a birth weight of <1500 g, the reported incidence of ROP ranges from 20% to 50% in different populations.1–3 Described first as ‘retrolental fibroplasia’ in 1942, the pathogenesis of ROP is still not fully understood but involves an intricate interplay between retinal blood vessels, oxygen, angiogenic and growth factors, of which vascular endothelial growth factor (VEGF) is the most important.4 ROP is described in two phases. The first phase begins after preterm birth secondary to the hyperoxic extrauterine environment and involves suppression of VEGF, vaso-obliteration with cessation of retinal blood vessel growth and endothelial apoptosis. The second phase is proliferative and involves neovascularisation of retinal vasculature by vasoactive factors such as VEGF, produced secondary to a hypoxic and avascular retina of phase 1. The second phase begins around 32 weeks postmenstrual age but can have a wide range of onset.2 4 Prematurity continues to be the single most important risk factor for ROP, but given its developmental nature, efforts have recently focused on identifying antenatal risk factors for ROP.

ic and avascular retina of phase 1. The second phase begins around 32 weeks postmenstrual age but can have a wide range of onset.2 4 Prematurity continues to be the single most important risk factor for ROP, but given its developmental nature, efforts have recently focused on identifying antenatal risk factors for ROP. Pre-eclampsia is characterised by new-onset maternal hypertension and proteinuria at or after 20 weeks of gestation. Although the aetiology of pre-eclampsia remains unknown, evidence suggests that a hypoxic and dysfunctional placenta produces excess antiangiogenic factors, soluble fms-like tyrosine kinase 1 and soluble endoglin, resulting in decreased maternal levels of angiogenic factors, VEGF and placental growth factor. This imbalance between angiogenic and antiangiogenic factors is responsible for maternal endothelial damage and manifestations of pre-eclampsia.5 How this antiangiogenic intrauterine milieu in pre-eclampsia affects retinal vasculature development is not well understood. Given that angiogenic factors such as VEGF are involved in the pathogenesis of ROP, and pre-eclampsia is associated with lower levels of VEGF, we hypothesised that preterm infant less than 31 weeks’ gestation and\or with a birth weight of <1500 g born to mothers with pre-eclampsia would be at a lower risk of developing ROP when compared with infants born to normotensive mothers.

e pathogenesis of ROP, and pre-eclampsia is associated with lower levels of VEGF, we hypothesised that preterm infant less than 31 weeks’ gestation and\or with a birth weight of <1500 g born to mothers with pre-eclampsia would be at a lower risk of developing ROP when compared with infants born to normotensive mothers. Methods The Neonatal Intensive Care Unit (NICU) in Calgary maintains a prospectively collected electronic database of all infants admitted to the NICU. Infants less than 31 weeks gestational age (GA) and/or with a birth weight of <1500 g born to mothers with pre-eclampsia between January 2007 and June 2010 were included in the study. Their outcome was compared with infants born to the next two normotensive mothers with a ±1 week GA difference. In case of missing data, the infant’s medical record charts were reviewed. The Conjoint Health Research Ethics Board of the University of Calgary approved the study.

and June 2010 were included in the study. Their outcome was compared with infants born to the next two normotensive mothers with a ±1 week GA difference. In case of missing data, the infant’s medical record charts were reviewed. The Conjoint Health Research Ethics Board of the University of Calgary approved the study. Maternal data Pre-eclampsia was defined as systolic blood pressure ≥140 mm Hg or a diastolic level of (Korotkoff 5) ≥90 mm Hg on two or more occasions at least 4–6 hours (but not more than 7 days) apart after 20 weeks’ gestation in a woman with previously normal blood pressure. Proteinuria was defined as ≥0.3 g protein in a 24-hour urine sample. When a 24-hour urine sample was not feasible, ≥0.3 g/L protein or ≥1+ on a dipstick test strip on two random urine samples taken at least 4–6 hours apart was used as criteria for proteinuria.6Histological chorioamnionitis was defined as infiltration of polymorphonuclear leucocytes in the fetal membranes and chorionic plate.7 Antenatal steroids were considered a course if more than 12 hours had elapsed after the first dose. Neonatal data Screening for ROP in Calgary is based on published recommendations and includes infants <1500 g and/or gestation <31 weeks as well as infants with birth weight of >1500 g and gestation >30 weeks who have a severe and complicated clinical course. All infants have their first examination at 28 days of life and then at 1–2 weeks interval depending on the findings of the initial examination. All funduscopic examinations were performed by two paediatric ophthalmologists.

th weight of >1500 g and gestation >30 weeks who have a severe and complicated clinical course. All infants have their first examination at 28 days of life and then at 1–2 weeks interval depending on the findings of the initial examination. All funduscopic examinations were performed by two paediatric ophthalmologists. The severity of ROP was based on the International Classification of ROP.8 No ROP was designated as stage 0, stage 1 was a line of demarcation, stage 2 a ridge, stage 3 intravitreous neovascularisation, stage 4 partial retinal detachment and stage 5 total retinal detachment. Severe ROP was defined as >stage 2. Plus disease, an indication of severity of ROP, was defined as venous dilation and arteriolar tortuosity of the posterior retinal vessels sometimes including vascular engorgement of the iris, poor papillary dilation and vitreous haze. GA was assessed in the following order of preference: date of embryo transfer for in vitro fertilisation, first trimester fetal ultrasound, second trimester fetal ultrasound, date of the last menstrual period and postnatal assessment using the modified Ballard score.

The severity of ROP was based on the International Classification of ROP.8 No ROP was designated as stage 0, stage 1 was a line of demarcation, stage 2 a ridge, stage 3 intravitreous neovascularisation, stage 4 partial retinal detachment and stage 5 total retinal detachment. Severe ROP was defined as >stage 2. Plus disease, an indication of severity of ROP, was defined as venous dilation and arteriolar tortuosity of the posterior retinal vessels sometimes including vascular engorgement of the iris, poor papillary dilation and vitreous haze. GA was assessed in the following order of preference: date of embryo transfer for in vitro fertilisation, first trimester fetal ultrasound, second trimester fetal ultrasound, date of the last menstrual period and postnatal assessment using the modified Ballard score. Intraventricular haemorrhage (IVH) was graded on Papile’s criteria.9 The diagnosis of necrotising enterocolitis (NEC) was based on modified Bell’s criteria.9 Neonatal sepsis was the presence of single organism from either blood or cerebrospinal fluid culture. Respiratory distress syndrome (RDS) was diagnosed based on the presence of signs of respiratory distress, a typical chest X-ray and/or need for surfactant. Patent ductus arteriosus (PDA) was diagnosed based on clinical signs and echocardiography. Bronchopulmonary dysplasia (BPD) was defined as the need for supplemental oxygen or any form of ventilation including continuous positive airways pressure at 36 weeks postmenstrual age.9 Intrauterine growth restriction (IUGR) was defined as birth weight less than the 10th percentile based on the growth charts of Kramer et al.10

lmonary dysplasia (BPD) was defined as the need for supplemental oxygen or any form of ventilation including continuous positive airways pressure at 36 weeks postmenstrual age.9 Intrauterine growth restriction (IUGR) was defined as birth weight less than the 10th percentile based on the growth charts of Kramer et al.10 Exclusion criteria included infants born with any congenital malformations or chromosomal anomalies, mothers with chronic hypertension, maternal renal, cardiovascular or autoimmune disease, substance abuse, TORCH infections and infants who died before their first eye examination.

lmonary dysplasia (BPD) was defined as the need for supplemental oxygen or any form of ventilation including continuous positive airways pressure at 36 weeks postmenstrual age.9 Intrauterine growth restriction (IUGR) was defined as birth weight less than the 10th percentile based on the growth charts of Kramer et al.10 Exclusion criteria included infants born with any congenital malformations or chromosomal anomalies, mothers with chronic hypertension, maternal renal, cardiovascular or autoimmune disease, substance abuse, TORCH infections and infants who died before their first eye examination. As the distribution of the relevant variables was not normal, we chose conservative non-parametric analysis for continuous variables, using the Mann-Whitney U test. Categorical variables were compared using the χ2 or Fisher’s exact test as appropriate. To identify risk factors for the development of ROP, multivariable logistic regression with backward elimination approach was performed. Any risk factors that demonstrated associations, whether statistically significant or judged to be clinically significant, with both pre-eclampsia and ROP but were not intermediate variables, were included in the modelling process as possible confounders.11 The least significant variables were then removed until all remaining variables were significant at p value of 0.2. The p value of 0.2 was set conservatively as an entry for variables to proceed to the next step in the analysis.12The adjusted OR values and their 95% CI are reported. A p <0.05 was considered significant. Data were analysed using STATA V.13.

hen removed until all remaining variables were significant at p value of 0.2. The p value of 0.2 was set conservatively as an entry for variables to proceed to the next step in the analysis.12The adjusted OR values and their 95% CI are reported. A p <0.05 was considered significant. Data were analysed using STATA V.13. Results During the study period, there were 99 women with pre-eclampsia  and  189 women in the normotensive group who delivered infants less than 31 weeks' gestation and/or with a birth weight of <1500 g. Figure 1 shows the mothers and infants who were excluded resulting in 95 women and 97 infants in the pre-eclampsia group and 161 women and 185 infants in the normotensive group. Figure 1 Flow diagram of patients included in the study. There was no difference in the maternal age or gravidity between the two groups. Although the number of primigravidas and antenatal steroid use was higher in the pre-eclampsia group, the difference was not statistically significant. Smoking rates, number of Caucasians and diabetes were similar between the two groups. Caesarean section rates were significantly higher and pathological chorioamnionitis and multiple births were lower in the pre-eclampsia group (table 1). Table 1 Maternal characteristics

There was no difference in the maternal age or gravidity between the two groups. Although the number of primigravidas and antenatal steroid use was higher in the pre-eclampsia group, the difference was not statistically significant. Smoking rates, number of Caucasians and diabetes were similar between the two groups. Caesarean section rates were significantly higher and pathological chorioamnionitis and multiple births were lower in the pre-eclampsia group (table 1). Table 1 Maternal characteristics Normotensive group (n=161) Pre-eclampsia group (n=95) p Value Maternal age, year, median (IQR) 31 (28–35) 31 (26–35) 0.93 Gravidity, median (IQR) 2 (2) 2 (2) 0.12 Parity, median (IQR) 0 (1) 0 (1) 0.08 Smoking, n (%) 19 (12) 5 (5) 0.06 Antenatal steroid, n (%) 148 (92) 94 (99) 0.05 Caucasian, n (%) 119 (74) 71 (75) 0.50 Diabetes, n (%) 12 (7) 6 (6) 0.47 C-section, n (%) 70 (43) 88 (93) 0.00 Chorioamnionitis, n (%) 85 (53) 5 (5) 0.00 Multiple births, n (%) 22 (14) 2 (2) 0.00 C-section, caesarean section. Although the GA was significantly lower in the normotensive group, the birth weights were higher. IUGR rates were significantly higher in the pre-eclampsia group. There was no difference in SNAP-PE (Score for Neonatal Acute Physiology-Physiological Extension) scores, number of males, RDS, PDA, IVH, NEC, BPD, blood transfusion rates, duration of oxygen use and mechanical ventilation between the two groups. Sepsis rates were higher in the normotensive group. Importantly, there was no difference in ROP, including severe ROP (>stage 2) between the two groups (table 2).

tension) scores, number of males, RDS, PDA, IVH, NEC, BPD, blood transfusion rates, duration of oxygen use and mechanical ventilation between the two groups. Sepsis rates were higher in the normotensive group. Importantly, there was no difference in ROP, including severe ROP (>stage 2) between the two groups (table 2). Table 2 Neonatal characteristics and outcomes Normotensive group (n=185) Pre-eclampsia group (n=97) p Value Gestational age, weeks, median (IQR) 29 (26–30) 29 (28–30) 0.01 Birth weight (g), median (IQR) 1207 (900–1425) 1020 (785–1250) 0.00 IUGR, n (%) 6 (3) 33 (35) 0.00 SNAP-PE score, median (IQR) 18 (9–32) 19 (9–36) 0.80 Males, n (%) 104 (56) 47 (49) 0.13 Inotrope use, n (%) 51 (28) 17 (18) 0.04 RDS, n (%) 142 (77) 74 (76) 0.52 Surfactant use, n (%) 96 (52) 58 (60) 0.12 PDA, n (%) 75 (40) 39 (40) 0.53 Any IVH, n (%) 24 (13) 11 (13) 0.42 Severe IVH, n (%) 7 (4) 2 (2) 0.34 Culture-proven sepsis, n (%) 31 (17) 8 (8) 0.03 Blood transfusion, n (%) 60 (32) 19 (19) 0.15 BPD at 36 weeks, n (%) 55 (30) 23 (24) 0.35 NEC stage ≥2, n (%) 10 (5) 3 (3) 0.28 Duration of oxygen use (days), median (IQR) 13 (2–57) 8 (2–48) 0.27 Duration of ventilation (days), median (IQR) 1 (0–7) 1 (0–5) 9.36 0.68 Any ROP, n (%) 50 (27) 26 (27) 0.54 Severe ROP,* n (%) 17 (9) 4 (4) 0.09 *>Stage 2 and/or plus disease.

at 36 weeks, n (%) 55 (30) 23 (24) 0.35 NEC stage ≥2, n (%) 10 (5) 3 (3) 0.28 Duration of oxygen use (days), median (IQR) 13 (2–57) 8 (2–48) 0.27 Duration of ventilation (days), median (IQR) 1 (0–7) 1 (0–5) 9.36 0.68 Any ROP, n (%) 50 (27) 26 (27) 0.54 Severe ROP,* n (%) 17 (9) 4 (4) 0.09 *>Stage 2 and/or plus disease. BPD, bronchopulmonary dysplasia; IUGR, intrauterine growth restriction; IVH, intraventricular haemorrhage; NEC, necrotising enterocolitis; PDA, patent ductus arteriosus; RDS, respiratory distress syndrome; ROP, retinopathy of prematurity; SNAP-PE, Score for Neonatal Acute Physiology-Physiological Extension. Out of the total cohort of 282 infants, 76 (27%) developed ROP. Bivariate analysis confirmed a number of known risk factors for ROP and included GA, birth weight, SNAP-PE scores, oxygen and mechanical ventilation days, IUGR, RDS, surfactant use, PDA, IVH, NEC and BPD (table 3). There was no difference in the mode of delivery, multiple birth, chorioamnionitis and gender between the two groups. Out of 76 infants with ROP, 26 (34%) were delivered to mothers with pre-eclampsia and out of the 206 infants without ROP, 71 (34%) were born to mothers with pre-eclampsia. This was statistically non-significant, p=0.54. Table 3 Characteristics of infants with and without ROP

Bivariate analysis confirmed a number of known risk factors for ROP and included GA, birth weight, SNAP-PE scores, oxygen and mechanical ventilation days, IUGR, RDS, surfactant use, PDA, IVH, NEC and BPD (table 3). There was no difference in the mode of delivery, multiple birth, chorioamnionitis and gender between the two groups. Out of 76 infants with ROP, 26 (34%) were delivered to mothers with pre-eclampsia and out of the 206 infants without ROP, 71 (34%) were born to mothers with pre-eclampsia. This was statistically non-significant, p=0.54. Table 3 Characteristics of infants with and without ROP No ROP (n=206) ROP (n=76) p Value Gestational age (weeks), median (IQR) 29 (28–30) 26 (25–28) 0.00 Birth weight (g), median (IQR) 1250 (1027–1427) 790 (680–964) 0.00 SNAP-PE II , median (IQR) 14 (5–24) 32 (19–48) 0.00 Oxygen (days), median (IQR) 5 (12–28) 69 (31–95) 0.00 Ventilator (days), median (IQR) 1 (0–2) 11 (2–25) 0.00 Pre-eclampsia, n (%) 71 (34.5) 26 (34.2) 0.54 C-section, n (%) 124 (60.2) 51 (67.1) 0.17 Caucasian, n (%) 163 (79.1) 51 (67.1) 0.02 Multiples, n (%) 24 (11.7) 6 (7.9) 0.25 Chorioamnionitis, % 66 (32.0) 32 (42.1) 0.07 Males, n (%) 115 (55.9) 36 (47.4) 0.13 IUGR, n (%) 22 (10.7) 17 (22.4) 0.01 RDS, n (%) 145 (70.4) 71 (93.4) 0.00 Surfactant, n (%) 96 (46.6) 58 (76.3) 0.00 PDA, n (%) 60 (29.1) 54 (71.1) 0.00 Positive blood cultures, n (%) 18 (8.7) 21 (27.6) 0.00 Blood transfusion, n (%) 32 (15.5) 47 (61.8) 0.00 IVH, n (%) 16 (7.8) 19 (25.0) 0.00 Grades 3 and 4, IVH, n (%) 3 (1.5) 6 (7.9) 0.00 NEC, n (%) 4 (1.9) 9 (11.8) 0.00 BPD, n (%) 28 (13.6) 50 (65.8) 0.00 BPD, bronchopulmonary dysplasia; C-section, caesarean section; IUGR, intrauterine growth restriction; IVH, intraventricular haemorrhage; NEC, necrotising enterocolitis; PDA, patent ductus arteriosus; RDS, respiratory distress syndrome; ROP, retinopathy of prematurity; SNAP-PE, Score for Neonatal Acute Physiology-Physiological Extension.

ary dysplasia; C-section, caesarean section; IUGR, intrauterine growth restriction; IVH, intraventricular haemorrhage; NEC, necrotising enterocolitis; PDA, patent ductus arteriosus; RDS, respiratory distress syndrome; ROP, retinopathy of prematurity; SNAP-PE, Score for Neonatal Acute Physiology-Physiological Extension. On logistic regression including significant factors, pre-eclampsia was neither protective nor a risk factor for the development of ROP (OR 1.7, 95% CI 0.6 to 4.7, p=0.29). Lower GA, blood transfusion and IUGR were risk factors for the development of ROP (table 4). Table 4 Results of multivariate regression modelling risk of ROP OR (95% CI) p Value Pre-eclampsia 1.716741 0.6207951 4.74 0.29 Gestational age 0.531063 0.4113149 0.68 0.00 Blood transfusion 1.566465 1.197519 2.040 0.001 IUGR 3.517356 1.217739 10.15 0.02 IUGR, intrauterine growth restriction; ROP, retinopathy of prematurity. Analysis of data for infants ≤28 weeks’ gestation and <1000 g birth weight did not change the results.

OR (95% CI) p Value Pre-eclampsia 1.716741 0.6207951 4.74 0.29 Gestational age 0.531063 0.4113149 0.68 0.00 Blood transfusion 1.566465 1.197519 2.040 0.001 IUGR 3.517356 1.217739 10.15 0.02 IUGR, intrauterine growth restriction; ROP, retinopathy of prematurity. Analysis of data for infants ≤28 weeks’ gestation and <1000 g birth weight did not change the results. Discussion The results of our study demonstrate that pre-eclampsia is not associated with the development of ROP, either as a protective or risk factor. The results also confirmed a number of known risk factors of ROP, including GA, IUGR and blood transfusions.13 14 Our results are similar to the report by Huang et al14 who, using a large cohort of very low birthweight infants, demonstrated pre-eclampsia to be not associated with the development of ROP but found IUGR and blood transfusions to be significant risk factors. As in our study, gender, RDS, days of oxygen supplementation and mechanical ventilation were not risk factors for ROP. Other studies have also reported no association of pre-eclampsia with ROP.15–18 More recently, a meta-analysis on the association of ROP with gestational hypertensive disorders, which included studies on gestational hypertension and pre-eclampsia, demonstrated no association.19

entilation were not risk factors for ROP. Other studies have also reported no association of pre-eclampsia with ROP.15–18 More recently, a meta-analysis on the association of ROP with gestational hypertensive disorders, which included studies on gestational hypertension and pre-eclampsia, demonstrated no association.19 Reports of the association of pre-eclampsia and ROP are, however, conflicting with pre-eclampsia being reported protective and also as a risk factor for the development of ROP.20–31 Investigators reporting pre-eclampsia as protective for ROP have attributed it to the lower VEGF levels and higher levels of antiangiogenic factors in pre-eclampsia.20 While reports of maternal levels of angiogenic and antiangiogenic factors in pre-eclampsia are relatively consistent, reports of umbilical cord blood and neonatal levels are equivocal at best, with reported levels being higher, lower and the same as infants of normotensive mothers.32 33 It has also been speculated that amniotic fluid antiangiogenic factors in pre-eclampsia could access the retina through the corneal epithelium and in this way protect against ROP.20 21 The few reports of angiogenic and antiangiogenic factors in amniotic fluid in pre-eclamptic and normotensive pregnancies also show disparate results.33 34 It is also important to note that the pathophysiology of ROP is complex, and an antiangiogenic state with low levels of VEGF would be of benefit in the second phase of ROP, remote and several weeks away from the time of birth and the antiangiogenic intrauterine milieu of pre-eclampsia.4 Some authors have also suggested that intrauterine stress associated with pre-eclampsia could lead to the accelerated development of retinal blood vessels, lessening the chances of ROP.22 However, IUGR, which is also associated with intrauterine stress, is a risk factor rather than protective for the development of ROP in most if not all studies on risk factors of ROP. Although angiogenic factors such as VEGF are important for retinal vasculature development, the role these factors play in fetal retinal development in pre-eclampsia remains unclear.

ntrauterine stress, is a risk factor rather than protective for the development of ROP in most if not all studies on risk factors of ROP. Although angiogenic factors such as VEGF are important for retinal vasculature development, the role these factors play in fetal retinal development in pre-eclampsia remains unclear. An equal number of investigators have reported pre-eclampsia to be a risk factor for the development of ROP.25–29 Although the exact mechanism of how pre-eclampsia leads to ROP is not established, various reasons have been reported. Increased oxidative stress and inflammation associated with pre-eclampsia have been suggested to cause impaired retinal vascularisation resulting in ROP.25 35 However, in a recent systematic review and meta-analysis, Mitra et al36 reported chorioamnionitis was not associated with ROP. Lee et al29 using the Extremely Low Gestational Age Newborns study cohort comprising 1199 infants, reported prelabor premature rupture of membranes and placental abruption, both conditions associated with inflammation, to have a reduced risk of ROP.29 Low levels of insulin-like growth factor (IGF)-1 in pre-eclampsia have also been implicated in the increased risk of ROP.37 As with angiogenic and antiangiogenic factors, reports of reduced maternal levels of IGF-1 are consistent but reports of umbilical cord blood levels are conflicting with reports of higher umbilical cord levels of IGF-1 in pre-eclampsia.38 39

)-1 in pre-eclampsia have also been implicated in the increased risk of ROP.37 As with angiogenic and antiangiogenic factors, reports of reduced maternal levels of IGF-1 are consistent but reports of umbilical cord blood levels are conflicting with reports of higher umbilical cord levels of IGF-1 in pre-eclampsia.38 39 There are several reasons for the conflicting results on the association between pre-eclampsia and ROP. These include different definitions of pre-eclampsia and ROP, inclusion of cases of hypertension other than pre-eclampsia, different GAs, different time of eye examinations, varying rates of antenatal corticosteroid use and performing only univariate analysis rather than multiple regression. Some investigators have excluded infants who died before their first eye examination and some have not. Studies that have investigated risk factors for ROP and not pre-eclampsia specifically, generally have small numbers of infants with pre-eclamptic mothers.21 23 24 26 Importantly, some studies are from an earlier time period when oxygen use, ventilator and nutritional practices were different from current practice.23 27

tudies that have investigated risk factors for ROP and not pre-eclampsia specifically, generally have small numbers of infants with pre-eclamptic mothers.21 23 24 26 Importantly, some studies are from an earlier time period when oxygen use, ventilator and nutritional practices were different from current practice.23 27 The incidence of all ROP and severe ROP in our cohort was 27% and 6.7%, respectively, not different from that reported by the Canadian Neonatal Network (www.canadianneonatalnetwork.org). On multiple logistic regression, we did not find oxygen and mechanical ventilation to be risk factors for ROP. This may be because the NICU in Calgary has strict targets for oxygen saturations maintaining them between 85% and 95%.40 In addition, the emphasis is on non-invasive ventilation with early extubation of intubated infants. Mechanical ventilation may also be a confounding variable to oxygen supplementation. A recent study from Denmark with a large cohort also did not find oxygen duration to be a risk factor for ROP.13

etween 85% and 95%.40 In addition, the emphasis is on non-invasive ventilation with early extubation of intubated infants. Mechanical ventilation may also be a confounding variable to oxygen supplementation. A recent study from Denmark with a large cohort also did not find oxygen duration to be a risk factor for ROP.13 The strengths of our study include a well-defined cohort from a recent era during which there was no change in clinical practice with high rates of antenatal corticosteroids use, and the ophthalmologists performing the eye examination were not aware of the diagnosis of pre-eclampsia. Another strength is the histopathological basis of chorioamnionitis in our study as correlation with clinical chorioamnionitis is poor. There were no missing clinical data or missed eye examinations. We also used GA to define our cohort, thereby minimising the confounding due to IUGR. Limitation of the study is that it is a single-centre study, making generalisations difficult. However, multicentre studies, although having large sample sizes, are beset with variation in clinical practice, differing rates of ROP between centres and substantial relevant data may also be missing.15 27

ding due to IUGR. Limitation of the study is that it is a single-centre study, making generalisations difficult. However, multicentre studies, although having large sample sizes, are beset with variation in clinical practice, differing rates of ROP between centres and substantial relevant data may also be missing.15 27 In summary, contrary to our hypothesis, we did not find pre-eclampsia to be associated with ROP, either as a protective or risk factor. Although there may be several reasons, the literature on pre-eclampsia and ROP is conflicting and inconsistent, with studies reporting increased risk, protective effect and no effect. The study confirmed known risk factors for ROP, including prematurity, blood transfusions and IUGR status. Acknowledgement: The authors gratefully acknowledge the funding provided by the Alberta Children’s Hospital Foundation and Alberta Children’s Hospital Research Institute for this study. Contributors: KY made substantial contributions to the conception, design and acquisition of data, analysis and interpretation. He was also involved in drafting and revising the manuscript. BA contributed to the conception and design, wrote the first draft and also helped with data analysis and interpretation and revision of the manuscript. OS helped in design, acquisition of data and writing of the manuscript. NS was involved in the conception and design, acquisition of data and writing and critiquing of the manuscript. AE was involved in conception and design and providing critical appraisal of the manuscript. Competing interests: None declared.

Contributors: KY made substantial contributions to the conception, design and acquisition of data, analysis and interpretation. He was also involved in drafting and revising the manuscript. BA contributed to the conception and design, wrote the first draft and also helped with data analysis and interpretation and revision of the manuscript. OS helped in design, acquisition of data and writing of the manuscript. NS was involved in the conception and design, acquisition of data and writing and critiquing of the manuscript. AE was involved in conception and design and providing critical appraisal of the manuscript. Competing interests: None declared. Patient consent: Retrospective cohort study. Ethics approval: Conjoint Health Research Ethics Board, University of Calgary. Provenance and peer review: Not commissioned; externally peer reviewed.

Key messages Ocular inhibition of 11β-hydroxysteroid-dehydrogenase (11βHSD) has been suggested in the literature to decrease intraocular pressure (IOP) in patients with primary open angle glaucoma (POAG) or ocular hypertension (OHT). We did not find a clinically relevant IOP reduction of the highly selective and potent 11β-hydroxysteroid-dehydrogenase type 1 (11βHSD1) inhibitor RO5093151 in patients with POAG and OHT following oral administration. The clinical trial design included a placebo-controlled run-in period for all patients and a Bayesian decision criterion to test for success based on an absolute 5 mm Hg IOP reduction. This allowed for a relatively small sample size to conclude on lack of efficacy. High ocular RO5093151 concentrations in the rabbit following oral administration, together with the clinical observation of high unbound drug concentrations well above the IC50 value of 11βHSD1 inhibition and high systemic 11βHSD1 inhibition suggested sufficiently high ocular concentrations of RO5093151 were achieved. Inhibition of 11βHSD1 is not a promising target to reduce IOP in patients with POAG or OHT. Further studies are required to investigate the benefit of 11βHSD1 inhibition in patients with steroid-induced glaucoma.

High ocular RO5093151 concentrations in the rabbit following oral administration, together with the clinical observation of high unbound drug concentrations well above the IC50 value of 11βHSD1 inhibition and high systemic 11βHSD1 inhibition suggested sufficiently high ocular concentrations of RO5093151 were achieved. Inhibition of 11βHSD1 is not a promising target to reduce IOP in patients with POAG or OHT. Further studies are required to investigate the benefit of 11βHSD1 inhibition in patients with steroid-induced glaucoma. Introduction Elevated intraocular pressure (IOP) is a key risk factor in the development and progression of vision impairment in glaucoma with over 60 million affected individuals worldwide.1 Ocular hypotensive therapy remains the mainstay in the pharmacological treatment for primary open angle glaucoma (POAG) and ocular hypertension (OHT).2 3 The pharmacological interventions currently available mainly target the turnover of the aqueous humour by either reducing the production or increasing its outflow.2

1 Ocular hypotensive therapy remains the mainstay in the pharmacological treatment for primary open angle glaucoma (POAG) and ocular hypertension (OHT).2 3 The pharmacological interventions currently available mainly target the turnover of the aqueous humour by either reducing the production or increasing its outflow.2 The use of topical and systemic glucocorticoids can result in corticosteroid-induced glaucoma. This condition is associated with an increase in IOP, which ultimately can lead to visual field loss and blindness. Several lines of evidence implicate the involvement of cortisol in the regulation of IOP: (A) an analysis of diurnal IOP curves found in 40% of cases the highest IOP at the earliest morning measurement with 65% of the peaks occurring before noon,4 (B) raised IOP can also occur in patients with Cushing’s syndrome5 and (C) an increase in plasma cortisol levels has been reported in patients with OHT and POAG,4 indicative of an alteration of the hypothalamus–pituitary–adrenal axis in these patients.

t morning measurement with 65% of the peaks occurring before noon,4 (B) raised IOP can also occur in patients with Cushing’s syndrome5 and (C) an increase in plasma cortisol levels has been reported in patients with OHT and POAG,4 indicative of an alteration of the hypothalamus–pituitary–adrenal axis in these patients. Glucocorticoids are proposed to have effects both on aqueous humour production in non-pigmented epithelial cells of the ciliary body6–9 and on reducing outflow through the trabecular meshwork by changing its morphology.10 Cortisol levels in the eye are not merely a reflection of circulating cortisol levels but are actively controlled by the two enzymes 11β-hydroxysteroid-dehydrogenase type 1 (11βHSD1) and 11β-hydroxysteroid-dehydrogenase type 2 (11βHSD2), the former transforming cortisone into the active form cortisol, the latter facilitating the reverse. 11βHSD1 was shown to be expressed in human basal cells of the corneal epithelium and in a ciliary non-pigmented epithelial cell-line, while 11βHSD2 was restricted to the corneal endothelium.9 Therefore, increased IOP was hypothesised to be linked to increased ocular cortisol levels, which could be reduced by inhibition of 11βHSD1.11

own to be expressed in human basal cells of the corneal epithelium and in a ciliary non-pigmented epithelial cell-line, while 11βHSD2 was restricted to the corneal endothelium.9 Therefore, increased IOP was hypothesised to be linked to increased ocular cortisol levels, which could be reduced by inhibition of 11βHSD1.11 RO5093151 is a potent and highly selective inhibitor of human 11βHSD1, which showed to decrease the urinary cortisol/cortisone metabolite ratio of tetrahydrocortisol (5α and 5β)/tetrahydrocortisone (THF/THE) as surrogate of total body 11βHSD1 inhibition.12 13 Overall, RO5093151 was shown to have a good safety profile in clinical studies using doses of up to 200 mg twice daily for up to 12 weeks. Glaucoma medications are typically administered by topical administration. Based on the available evidence of the role of 11βHSD1 in IOP regulation provided above, a preclinical study was first conducted to assess the ocular exposure of RO5093151 after oral administration, followed by a clinical study testing the effect of orally administered RO5093151 on IOP in patients with OHT or POAG. Methods Exploratory study to assess ocular exposure of RO5093151 in the rabbit The animal experiments were conducted at Iris Pharma, La Gaude, France, in 2011. All animals were treated according to the European Convention (French Decree N° 2001–486, dated 6 June 2001) and to the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research (Handbook for the Use of Animals in Biomedical Research second ed., 1993).

2011. All animals were treated according to the European Convention (French Decree N° 2001–486, dated 6 June 2001) and to the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research (Handbook for the Use of Animals in Biomedical Research second ed., 1993). A single oral dose of 30 mg/kg RO5093151 was administered to 12 pigmented rabbits. At 0.5, 2, 6 and 12 hours post-dose, three rabbits at each time-point were sacrificed and samples from plasma, tears, cornea, ciliary body and aqueous humour were analysed for RO5093151 using a non-validated liquid chromatography–mass spectrometry (LC-MS)/MS method. Ocular tissues were collected for both eyes. Clinical efficacy, safety and tolerability of RO5093151 in patients with POAG or OHT: study design and population The clinical study was conducted according to the provisions of the Declaration of Helsinki, and written informed consent was obtained from each study participant prior to conducting any protocol-related procedures. The study was approved by the local independent ethics committees and health authorities. The study was an adaptive, multicentre (three sites in USA, two in Czech Republic and one in Hungary), randomised, investigator-masked (during the randomised treatment phase), subject-masked, multiple-dose, placebo-controlled, parallel study to investigate efficacy, safety, tolerability and pharmacokinetics of RO5093151 for up to 28 days in subjects with POAG or OHT. It was conducted from December 2011 to November 2012 (NCT01493271).

vestigator-masked (during the randomised treatment phase), subject-masked, multiple-dose, placebo-controlled, parallel study to investigate efficacy, safety, tolerability and pharmacokinetics of RO5093151 for up to 28 days in subjects with POAG or OHT. It was conducted from December 2011 to November 2012 (NCT01493271). Patients who had given written informed consent were screened within 8 weeks prior to the start of the lead-in phase. A wash-out phase for up to 6 weeks was conducted during the screening phase for patients on previous treatment (corticosteroids, systemic or ocular medication known to affect IOP, glaucoma medications). Screening assessments included medical history (general and ophthalmological), physical examination, height and weight, 12-lead ECG, vital signs including blood pressure and body temperature, drugs of abuse and alcohol test, blood and urine clinical laboratory safety tests (haematology, coagulation, chemistry, urinalysis and serology), pregnancy test as appropriate and a thorough ophthalmological examination (biomicroscopy and best corrected Early Treatment Diabetic Retinopathy Study (ETDRS) visual acuity, pachymetry, gonioscopy, visual field and fundus examination under dilatation).

sts (haematology, coagulation, chemistry, urinalysis and serology), pregnancy test as appropriate and a thorough ophthalmological examination (biomicroscopy and best corrected Early Treatment Diabetic Retinopathy Study (ETDRS) visual acuity, pachymetry, gonioscopy, visual field and fundus examination under dilatation). Patients started with a subject-masked 7-day run-in period (from days −7 to −1) with twice daily dosing of placebo, followed by a 7-day treatment phase with twice daily dosing of RO5093151 200 mg or placebo randomised at a ratio of 4:1 (days 1–7, at 08:00 and 18:00) using an interactive voice recording system for treatment group allocation. During the treatment phase, investigators and patients were masked, while selected members of the study team from the sponsor were unmasked to allow interim analysis. A dose of 200 mg twice daily was chosen as it was the highest daily dose tested in previously conducted phase 1/2 studies following multiple dose administrations12 13 and revealed to be safe and well tolerated with a treatment duration of at least 4 weeks. Already a 5 mg twice daily regimen was shown to inhibit the conversion of cortisone to cortisol (see urinary cortisone/cortisol metabolite assessment below) to a similar extent observed at higher doses (data not shown). High inhibition was also shown at a dose of 50 mg for hepatic and adipose 11βHSD1 (data not shown).

Already a 5 mg twice daily regimen was shown to inhibit the conversion of cortisone to cortisol (see urinary cortisone/cortisol metabolite assessment below) to a similar extent observed at higher doses (data not shown). High inhibition was also shown at a dose of 50 mg for hepatic and adipose 11βHSD1 (data not shown). Time-matched IOP assessments (Goldmann Applanation Tonometry) were performed on days −7, –1, 1 and 7 at 1-hour predose, at time of dosing (0 hour) and at 1, 2, 4, 8 and 12 hours postdose, referring to the first daily dose administered at 08:00. Additional 1-hour postdose IOP assessments were performed on days 4 (on treatment), 14 and 21 (follow-up days). Spot urine samples for assessment of the urinary ratio of 5α-tetrahydrocortisol (5αTHF)+5β-tetrahydrocortisol (THF) over tetrahydrocortisone (THE) (THF/THE) were collected on days of IOP assessment (except for day −7) also in a time-matched manner. Serum cortisol and ACTH were collected on days −1, 7 and 21 (data not shown). Sparse pharmacokinetic samples were taken on day 7; plasma concentrations of RO5093151 were assessed using a validated LC-MS/MS assay (limit of quantification (LOQ)=0.5 ng/mL).

IOP assessment (except for day −7) also in a time-matched manner. Serum cortisol and ACTH were collected on days −1, 7 and 21 (data not shown). Sparse pharmacokinetic samples were taken on day 7; plasma concentrations of RO5093151 were assessed using a validated LC-MS/MS assay (limit of quantification (LOQ)=0.5 ng/mL). Inclusion criteria included male and female patients of at least 21 years of age, inclusive, diagnosed with OHT or POAG in at least one eye, a cup-to-disk ratio ≤0.8, best corrected visual acuity of 6/30 or better, central cornea thickness of 420–620 µm and able to participate and willing to give informed consent. Patients with intraocular hypertension were required to have an IOP ≥22 mm Hg at 08:00 (±1 hour) and ≥18 mm Hg between 15:00 and 18:00 in at least one eye and ≤32 mm Hg at all time-points in both eyes at screening and on day −7. Subject with diagnosed glaucoma had to have an IOP ≥18 mm Hg at screening and IOP ≥22 mm Hg on day −7, at 8:00 (±1 hour) and at the afternoon measurement, in at least one eye. The worse eye was defined as the eye with the higher IOP at baseline (day −1, time matching 1-hour postdose). Patients with the following criteria were excluded: presence of extreme narrow angle with complete or partial closure, progressive retinal or optic nerve disease from any cause other than glaucoma, history or signs of penetrating ocular trauma, uncontrolled hypertension despite treatment, clinically significant abnormalities in laboratory test results, positive test results on hepatitis B, hepatitis C or HIV 1 and 2, kidney disease or dysfunction.

retinal or optic nerve disease from any cause other than glaucoma, history or signs of penetrating ocular trauma, uncontrolled hypertension despite treatment, clinically significant abnormalities in laboratory test results, positive test results on hepatitis B, hepatitis C or HIV 1 and 2, kidney disease or dysfunction. Statistical analysis The primary endpoint of the study was the change from baseline for IOP measured on day 7 at 1-hour postdose and baseline being the time-matched IOP on day −1. The following Bayesian decision criterion was applied to the primary endpoint when a predefined number of patients had completed day 7 assessments: Probability (ΔIOP ≥5 mm Hg)>80% based on the posterior distribution of the mean ΔIOP. The Bayesian analysis was performed three times (figure 1), after the following number of evaluable subjects on RO5093151 was achieved: n=12 (15 including placebo), n=16 (20 including placebo), n=20 (25 including placebo). If the decision criterion had not been reached with 25 evaluable patients, additional 35 evaluable patients would have been enrolled and treatment duration would have been prolonged to a total of 28 days (part 2a) (figure 1). If the decision criterion had been reached, it was planned to include the investigation of lower doses and dose response of RO5093151 (part 2b). As the clinical study was terminated after completion of part 1, no further description of part 2 is provided. Figure 1 Flowchart of clinical study design. The study was terminated after completion of part 1.

The Bayesian analysis was performed three times (figure 1), after the following number of evaluable subjects on RO5093151 was achieved: n=12 (15 including placebo), n=16 (20 including placebo), n=20 (25 including placebo). If the decision criterion had not been reached with 25 evaluable patients, additional 35 evaluable patients would have been enrolled and treatment duration would have been prolonged to a total of 28 days (part 2a) (figure 1). If the decision criterion had been reached, it was planned to include the investigation of lower doses and dose response of RO5093151 (part 2b). As the clinical study was terminated after completion of part 1, no further description of part 2 is provided. Figure 1 Flowchart of clinical study design. The study was terminated after completion of part 1. Results Exploratory study to assess ocular exposure of RO5093151 in the rabbit Following a single oral dose of 30 mg/kg RO5093151 to the rabbit, low plasma concentrations were observed, with mean quantifiable levels up to 6-hour postdose. In aqueous humour, RO5093151 levels were also quantifiable up to 6 hours, whereas the other eye tissues exhibited quantifiable RO5093151 exposures only up to 3-hour postdose (figure 2). Despite the low plasma exposure, a high distribution into the investigated ocular tissues was observed, with tissue/plasma ratios of 209, 133, 20 and 0.6 for tears, ciliary body, cornea and aqueous humour, respectively, at 2-hour postdose where highest concentrations were achieved.

nly up to 3-hour postdose (figure 2). Despite the low plasma exposure, a high distribution into the investigated ocular tissues was observed, with tissue/plasma ratios of 209, 133, 20 and 0.6 for tears, ciliary body, cornea and aqueous humour, respectively, at 2-hour postdose where highest concentrations were achieved. Figure 2 Tissue concentrations of RO5093151 following single oral administration of 30 mg/kg RO5093151 to rabbits. Each data point reflects the mean of individual concentration of ocular tissue (from both eyes) and plasma of three animals. Results of the ocular tissues from one eye at 2-hour postdose were excluded from the analysis due to implausible high concentrations as compared with the results from the other eyes at the same time-point. Concentrations below the detection limit were set to 0 for mean calculation. Clinical efficacy, safety and tolerability of RO5093151 in patients with POAG or OHT A total of 35 patients were enrolled into the study; of these, 34 patients entered the placebo lead-in phase and 32 patients were randomised to either 200 mg twice daily RO5093151 (n=26) or matching placebo (n=6). All subjects who were randomised and received treatment were included in the intention-to-treat (ITT) population for the safety analysis. Six subjects were excluded from the per-protocol (PP) analysis; two each were excluded due to non-adherent drug intake, inclusion criteria violation and closure of a site. The ‘better eye’ of one subject was excluded from the PP analysis due to non-eligibility of this eye (previous trabeculectomy).

the safety analysis. Six subjects were excluded from the per-protocol (PP) analysis; two each were excluded due to non-adherent drug intake, inclusion criteria violation and closure of a site. The ‘better eye’ of one subject was excluded from the PP analysis due to non-eligibility of this eye (previous trabeculectomy). Most of the patients were female (15/25 subjects) with a mean age of approximately 60 years. The majority of patients had POAG (approx. 70%). All subjects who participated in the study reported angular widths of >20° (wide) and were classified to Shaffer Scale scores 2–4 and reported central cornea pachymetry measurements of 454–610 µm in each qualifying eye (inclusion criterion 420–620 µm) at baseline. The mean (SD) baseline IOP on day −1 at the nominal 09:00 time-point was 24.5 (3.1) and 23.4 (1.6) mm Hg for the RO5093151 and placebo group, respectively. Safety and tolerability RO5093151 was well tolerated in all subjects. Three subjects on active treatment reported in total four adverse events (diverticulum, osteoporosis, urinary tract infection and protein present in urine), which were all judged by the investigators to be not related to drug treatment. The reported adverse events were mild in intensity and resolved with no sequelae. No subject on placebo, either during the placebo run-in phase or during the randomised treatment phase, reported any adverse event. No clinically significant changes in vital signs, clinical laboratory values, general safety ophthalmological assessments and ECG readings were noted.

ensity and resolved with no sequelae. No subject on placebo, either during the placebo run-in phase or during the randomised treatment phase, reported any adverse event. No clinically significant changes in vital signs, clinical laboratory values, general safety ophthalmological assessments and ECG readings were noted. Pharmacodynamics A high and sustained inhibition of 11βHSD1 was shown in the RO5093151-treated group by the decrease of urinary THF/THE from between 0.9 to 1 at baseline to 0.18 on days 4–7 (figure 3), with no apparent changes in the placebo-treated group. After cessation of active treatment, the THF/THE ratio gradually returned to values close to baseline within 14 days, with the ratio in the active group still slightly below the baseline value. Figure 3 Mean (+SEM) urinary (5αTHF+5βTHF)/THE ratio as measure of 11βHSD1 inhibition (PP population). Randomised treatment (RO5093151/placebo) started on day 1 and lasted until day 7. PP, per-protocol; THE, tetrahydrocortisone; THF, tetrahydrocortisol. Pharmacokinetics The mean (SD) RO5093151 trough concentrations were at 124±127 ng/mL (n=20, PP analysis), ranging from 20 to 440 ng/mL, except for one patient with undetectable concentrations.

Figure 3 Mean (+SEM) urinary (5αTHF+5βTHF)/THE ratio as measure of 11βHSD1 inhibition (PP population). Randomised treatment (RO5093151/placebo) started on day 1 and lasted until day 7. PP, per-protocol; THE, tetrahydrocortisone; THF, tetrahydrocortisol. Pharmacokinetics The mean (SD) RO5093151 trough concentrations were at 124±127 ng/mL (n=20, PP analysis), ranging from 20 to 440 ng/mL, except for one patient with undetectable concentrations. Intraocular pressure The mean baseline IOP values (day −1) in the active treatment group were slightly higher than in the placebo group. Mean IOP was highest in the morning, with a gradual decrease during the course of the day and a sometimes apparent additional peak around lunchtime (figure 4C,D). There was a gradual trend for an overall decrease in IOP with time for both the placebo and active treatment groups, while a return to baseline values was observed after treatment cessation (figure 4B).

radual decrease during the course of the day and a sometimes apparent additional peak around lunchtime (figure 4C,D). There was a gradual trend for an overall decrease in IOP with time for both the placebo and active treatment groups, while a return to baseline values was observed after treatment cessation (figure 4B). Figure 4 Effect of RO5093151 and placebo on IOP. (A) Adjusted least square mean change from baseline and 95% CI in IOP from ANCOVA model at 1-hour postdose on day 7 versus time-matched 1-hour postdose on day −1. (B) Mean (±SEM) 1-hour postdose IOP profile for RO5093151 and placebo-treated patients over the study period. (C) Mean (±SEM) IOP time profile for RO5093151-treated patients (worse eye) for days −1 and 7. (D) Mean (±SEM) IOP time profile for placebo-treated patients (worse eye) for days −1 and 7. Arrow indicates time-point of primary endpoint (1-hour postdose); clock time shown for C and D. ANCOVA, analysis of covariance; IOP, intraocular pressure.

me profile for RO5093151-treated patients (worse eye) for days −1 and 7. (D) Mean (±SEM) IOP time profile for placebo-treated patients (worse eye) for days −1 and 7. Arrow indicates time-point of primary endpoint (1-hour postdose); clock time shown for C and D. ANCOVA, analysis of covariance; IOP, intraocular pressure. In the ‘worse eye’, the adjusted least square mean change from baseline in the PP population was −2.7 mm Hg (95% CI −4.2 to –1.2) and −2.9 (95% CI −5.9 to 0.1) in the RO5093151 and placebo group, respectively, with a mean difference between RO5093151 and placebo of 0.2 mm Hg (95% CI −3.1 to 3.6), in favour of placebo. In the ‘better eye’, displaying better IOP values at baseline compared with the selected worse eye, the adjusted least square mean change from baseline was −1.5 (95% CI −2.9 to 0.0) and −3.0 (95% CI −5.8 to −0.1) in the RO5093151 and placebo group, respectively, with a mean difference between RO5093151 and placebo of 1.5 (95% CI −1.7 to 4.7), in favour of placebo (figure 4A). While there was a statistically significant IOP reduction in the change from baseline for the ‘worse eye’ in the active treatment group and for the ‘better eye’ in both active and placebo treatment groups, the ANCOVA model adjusting for baseline levels did not show statistically significant difference in change of IOP between the active treatment group and placebo for both the ‘worse eye’ and the ‘better eye’.

‘worse eye’ in the active treatment group and for the ‘better eye’ in both active and placebo treatment groups, the ANCOVA model adjusting for baseline levels did not show statistically significant difference in change of IOP between the active treatment group and placebo for both the ‘worse eye’ and the ‘better eye’. The Bayesian decision criterion applied to the primary endpoint highlighted that the probability to reach the target reduction of at least 5 mm Hg was very low even at the final analysis with 20 subjects on active treatment (p=0.003). Discussion Published results from oral administration of the non-selective 11βHSD1 and 11βHSD2 inhibitor carbenoxolone8 and the cortisol biosynthesis inhibitor metyrapone14 were promising in supporting the rationale that a selective and potent inhibition of 11βHSD1 may achieve clinically relevant reductions in IOP following oral administration (table 1). Table 1 Key trial characteristics and IOP and pharmacological effects of orally administered drugs in clinical trials to reduce cortisol levels Drug name (reference) Drug type Design and regimen Effect on IOP (mm Hg) (n)

Discussion Published results from oral administration of the non-selective 11βHSD1 and 11βHSD2 inhibitor carbenoxolone8 and the cortisol biosynthesis inhibitor metyrapone14 were promising in supporting the rationale that a selective and potent inhibition of 11βHSD1 may achieve clinically relevant reductions in IOP following oral administration (table 1). Table 1 Key trial characteristics and IOP and pharmacological effects of orally administered drugs in clinical trials to reduce cortisol levels Drug name (reference) Drug type Design and regimen Effect on IOP (mm Hg) (n) (mean±SD or 95% CI) Primary pharmacological effect (systemically) Metpyrapone14 Inhibitor of adrenal cortisol biosynthesis (CYP11B1) Two-way cross-over, single administration −4.9±3.0 (active drug) (14) −2.6±2.2 (placebo) only highest effect for statistically significant time points shown Up to about 80% plasma cortisol reduction from baseline Carbenoxolone8 11βHSD1/2 inhibitor Two-way cross-over, t.i.d. for 4 days −2.2 (active drug) (20) −0.8 (placebo) p=0.0003 Urinary THF/THE ratio:decrease from 1.09±0.25 to 0.78±0.58 (n=7) in healthy volunteers, statistically significant AZD401716 11βHSD1 inhibitor Parallel groups, twice daily or four times a day for 4 weeks −2.6 (−3.8 to −1.3) (twice daily active drug) (19) −1.9 (−3.3 to −0.6) (twice daily placebo) (16) +0.1 (−4.8 to 4.9) (four times daily placebo) (6) No data available RO5093151 (this study) 11βHSD1 inhibitor Parallel groups, twice daily for 7 days −2.7 (−4.2 to −1.2) (active drug) (20) −2.9 (−5.9 to 0.1) (placebo) (5) (worse eye) Urinary THF/THE ratio decreased from 0.90±0.34 (day 1 baseline) to 0.18±0.32 (day 7 trough) 11βHSD1, 11β-hydroxysteroid-dehydrogenase type 1; IOP, intraocularpressure; THE, tetrahydrocortisone; THF, tetrahydrocortisol.

daily for 7 days −2.7 (−4.2 to −1.2) (active drug) (20) −2.9 (−5.9 to 0.1) (placebo) (5) (worse eye) Urinary THF/THE ratio decreased from 0.90±0.34 (day 1 baseline) to 0.18±0.32 (day 7 trough) 11βHSD1, 11β-hydroxysteroid-dehydrogenase type 1; IOP, intraocularpressure; THE, tetrahydrocortisone; THF, tetrahydrocortisol. The current study was designed as an exploratory proof of concept and included only a limited number of patients on placebo. Even when discarding the change from baseline for placebo due to the low number of patients on placebo included, the IOP effect of 2.7 mm Hg in the active treatment group had a very low probability (p<0.005) to achieve an IOP reduction of 5 mm Hg at 1-hour postdose but also at any time-point assessed. This target reduction was chosen for the decision criterion as it reflects a clinically relevant IOP reduction of approximately 20%, as achieved by ocular antihypertensive medications already on the market. The timing of the primary endpoint was chosen to be in the morning, as cortisol levels are known to be highest in the morning. Maximum concentration of RO5093151 (data not shown) is expected quickly after dosing, at approximately 1-hour postdose. IOP decreased over time from day 1–7 for both the active treatment and the placebo groups (figure 4).

dpoint was chosen to be in the morning, as cortisol levels are known to be highest in the morning. Maximum concentration of RO5093151 (data not shown) is expected quickly after dosing, at approximately 1-hour postdose. IOP decreased over time from day 1–7 for both the active treatment and the placebo groups (figure 4). The specific adaptive design of the study would have enabled to adapt the study conduct based on the sequential interim analyses using a Bayesian decision criterion. If the treatment duration of 7 days had not been achieving the target IOP reduction, the study protocol would have planned a prolongation of the treatment duration from 7 to 28 days. This was based on the bipartite rationale of the mode of action of RO5093151 on reduction of aqueous humour production and reduced outflow of aqueous humour.10 However, if the analysis had shown the anticipated IOP reduction, lower doses would have been tested to assess the dose-response of 11βHSD1 inhibition on IOP reduction. However, the study was discontinued due to portfolio reprioritisation prior to extending the treatment duration to 28 days. The results of the presented study are in line with a similar type of study investigating the impact of another selective and potent oral 11βHSD1 inhibitor AZD4017 on IOP (table 1).15 16 However, AZD4017 was administered for a duration of 4 weeks, suggesting that a longer treatment duration in the present study may also not have provided a larger effect on IOP.

ine with a similar type of study investigating the impact of another selective and potent oral 11βHSD1 inhibitor AZD4017 on IOP (table 1).15 16 However, AZD4017 was administered for a duration of 4 weeks, suggesting that a longer treatment duration in the present study may also not have provided a larger effect on IOP. A weakness of the present study and clinical studies reported so far is the lack of demonstrating adequate ocular exposures of the investigational drugs following oral administration. Sampling of aqueous humour is an established methodology in clinical research, not so much for glaucoma specialists though, and could have provided highly valuable data on ocular drug exposure and ocular cortisol levels under treatment.

equate ocular exposures of the investigational drugs following oral administration. Sampling of aqueous humour is an established methodology in clinical research, not so much for glaucoma specialists though, and could have provided highly valuable data on ocular drug exposure and ocular cortisol levels under treatment. Our preclinical data suggest high drug exposure in ocular tissues following oral administration (figure 2). Aqueous humour concentrations were close to the unbound plasma concentrations, and a high ciliary body concentration indicates that RO5093151 reaches its target in the rabbit eye. In the clinical study, an oral dose of 200 mg RO5093151 twice daily was used in order to aim for a maximised ocular exposure. This dose is 40-fold above a dose level achieving a similar systemic THF/THE ratio and is still not interfering with 11βHSD212. Plasma RO5093151 trough concentrations measured on day 7 demonstrated unbound concentrations 60-fold to 1300-fold above the in vitro IC50 of 0.19 ng/mL, characterised for RO5093151 using H4IIE cells stably transfected with a cDNA encoding full-length human 11βHSD1 (data not shown). Assuming aqueous humour and ciliary body fluid concentrations being close to the unbound plasma concentrations in humans, as suggested in rabbits, a high inhibition of ocular 11βHSD1 by RO5093151 should therefore have been achieved in our study. However, it is unclear whether the blood–aqueous humour barrier limits the entry of the investigated drugs to the aqueous humour and subsequently to the ocular target tissue in humans. Furthermore, a theoretical compensatory response diminishing local cortisol reductions in ocular tissues cannot be ruled out.

ur study. However, it is unclear whether the blood–aqueous humour barrier limits the entry of the investigated drugs to the aqueous humour and subsequently to the ocular target tissue in humans. Furthermore, a theoretical compensatory response diminishing local cortisol reductions in ocular tissues cannot be ruled out. In conclusion, there was no clinically meaningful IOP reduction by a 7-day oral treatment with RO5093151. Results of this study are in line with previously reported clinical data of compounds administered orally in the same class of 11βHSD1 inhibitors. Oral administration of RO5093151 200 mg twice daily for 7 days was safe and well tolerated. Whether topical administration or a longer exposure of RO5093151 would have achieved a clinically relevant result is currently not elucidated. Acknowledgements: The authors would like to thank Dr Svetlana Spassova, Senior Medical Director at PPD, Sofia, Bulgaria, for scientific advice regarding the definition of ocular inclusion and exclusion criteria and the principal investigators Mária Ferencz (Fövárosi Önkormányzat Szent Imre Kórház, OSZMSZ Szemészeti Profil, Budapest, Hungary), Jara Hornova (Fakultni Nemocnice Kralovske Vinohrady, Oftalmologicka Klinika, Praha, Czech Republic) and Jana Kadlecová (Oční klinika FN, Hradec Králové, Czech Republic) for participating to the trial. Contributors: DS, PJ, SF-R and WCS designed the clinical study. DH designed, analysed and provided contribution to the preclinical study. CS, AP, SF-R were involved in the design and analysis of the data.

Acknowledgements: The authors would like to thank Dr Svetlana Spassova, Senior Medical Director at PPD, Sofia, Bulgaria, for scientific advice regarding the definition of ocular inclusion and exclusion criteria and the principal investigators Mária Ferencz (Fövárosi Önkormányzat Szent Imre Kórház, OSZMSZ Szemészeti Profil, Budapest, Hungary), Jara Hornova (Fakultni Nemocnice Kralovske Vinohrady, Oftalmologicka Klinika, Praha, Czech Republic) and Jana Kadlecová (Oční klinika FN, Hradec Králové, Czech Republic) for participating to the trial. Contributors: DS, PJ, SF-R and WCS designed the clinical study. DH designed, analysed and provided contribution to the preclinical study. CS, AP, SF-R were involved in the design and analysis of the data. MET and HD were responsible for data acquisition and research execution. A ll authors were involved in revising and final approval of the manuscript. Competing interests: DS, CS, AP, SF-R, DH and PJ were employees of F. Hoffmann La Roche Ltd. The study was funded by Roche Pharmaceutical Research and Early Development. Patient consent: Patient. Ethics approval: Health authority approval in the USA, Bulgaria, Hungary and Czech Republic, corresponding ethics approval by Quorum Review IRB, Ethics Committee for Multi-centre Trials, Egeszsegugyi Tudomanyos Tanacs, Eticka komise Fakultni nemocnice, respectively. Provenance and peer review: Not commissioned; externally peer reviewed.

Key messages What is already known about this subject? Blepharitis is a common chronic condition characterised by eyelid inflammation. Although the pathophysiology remains uncertain, bacterial overcolonisation is implicated with ocular inflammatory responses. Topical antibiotics and corticosteroids are currently prescribed to treat inflammatory exacerbations; however, concerns regarding drug resistance and adverse effects necessitate development of alternative therapeutic options. New Zealand Manuka honey exhibits antimicrobial and anti-inflammatory properties that suggest potential for the treatment of blepharitis. Complexation with α-cyclodextrin has previously been reported to enhance the antibacterial effects of Manuka honey. What are the new findings? In clinically applicable concentrations, cyclodextrin-complexed MGO Manuka Honey (Manuka Health New Zealand) generally exhibited greater in vitro antimicrobial effects than uncomplexed honey on bacteria commonly associated with blepharitis. In vivo safety and tolerability evaluation on rabbit eyes demonstrated that a cyclodextrin-complexed Manuka honey microemulsion (MHME), designed for topical application, did not result in significant immediate or cumulative adverse effects. How might these results change the focus of research or clinical practice? The results of this study support progress to clinical safety and tolerability trials of MHME eye cream in human subjects.

In clinically applicable concentrations, cyclodextrin-complexed MGO Manuka Honey (Manuka Health New Zealand) generally exhibited greater in vitro antimicrobial effects than uncomplexed honey on bacteria commonly associated with blepharitis. In vivo safety and tolerability evaluation on rabbit eyes demonstrated that a cyclodextrin-complexed Manuka honey microemulsion (MHME), designed for topical application, did not result in significant immediate or cumulative adverse effects. How might these results change the focus of research or clinical practice? The results of this study support progress to clinical safety and tolerability trials of MHME eye cream in human subjects. Introduction Blepharitis is a common chronic condition that predisposes the eyelids to bacterial overcolonisation and inflammation, causing significant morbidity worldwide.1 2 Clinically, the condition can be differentiated into anterior blepharitis, which affects the anterior lamella of the eyelid and the eyelashes; or posterior blepharitis, most commonly meibomian gland dysfunction, which affects the posterior lamella.3 The pathophysiological mechanisms of blepharitis remain uncertain. However, overcolonisation of the eyelid region, with bacteria such as Staphylococcus aureus, Staphylococcus epidermidis and Pseudomonas aeruginosa,4–7 has been implicated in the development and exacerbation of ocular surface inflammation.8 9 Increased bacterial lipase activity may promote tear film instability, leading to excessive tear evaporation and dry eye symptoms.8 10

teria such as Staphylococcus aureus, Staphylococcus epidermidis and Pseudomonas aeruginosa,4–7 has been implicated in the development and exacerbation of ocular surface inflammation.8 9 Increased bacterial lipase activity may promote tear film instability, leading to excessive tear evaporation and dry eye symptoms.8 10 While topical antibiotics may reduce bacterial load,11 their long-term use raises concerns with regard to drug resistance.12 Similarly, prolonged use of topical corticosteroids to manage ocular surface inflammatory episodes is contraindicated on account of the risk of side effects.1 9 As a result, the management of blepharitis remains limited largely to palliative therapies such as warm compresses and eyelid hygiene techniques.11 13 14 However, these therapeutic strategies are often perceived by patients as laborious and ineffective, and consequently are met with poor compliance.15 This highlights the need for a therapeutic intervention that attempts to address the bacterial and inflammatory aspects of the condition and is suitable for long-term application.

these therapeutic strategies are often perceived by patients as laborious and ineffective, and consequently are met with poor compliance.15 This highlights the need for a therapeutic intervention that attempts to address the bacterial and inflammatory aspects of the condition and is suitable for long-term application. The inhibitory effects on ocular microbiota and anti-inflammatory properties of honey have previously been demonstrated, and are attributed to its high osmolarity, low pH, hydrogen peroxide content and nonperoxide components, including methylglyoxal (MGO).16 New Zealand Mānuka honey (Leptospermum scoparium), in particular, shows potential as a novel therapy for the management of blepharitis, due to its high concentrations of MGO, which is more resistant to physiological inactivation by heat and catalases than antimicrobial peroxide components.17 Cyclodextrins are commonly used additives in topical formulations, which increase the stability and solubility of hydrophobic drugs in water.18 19 Manuka honey has been complexed with α-cyclodextrin to form the product MGO Manuka Honey, which has previously been reported to have greater antibacterial effects than uncomplexed honey.20

re commonly used additives in topical formulations, which increase the stability and solubility of hydrophobic drugs in water.18 19 Manuka honey has been complexed with α-cyclodextrin to form the product MGO Manuka Honey, which has previously been reported to have greater antibacterial effects than uncomplexed honey.20 This preclinical study sought to investigate the in vitro effects of cyclodextrin-complexed and uncomplexed MGO Manuka Honey on the growth of bacteria commonly associated with blepharitis, including S. aureus, S. epidermidis and P. aeruginosa. The in vitro human corneal epithelial cell viability and in vivo rabbit eye tolerability of an MGO Manuka Honey microemulsion (MHME) designed for topical periocular application were also evaluated. Materials and Methods In vitro phase Materials Cyclodextrin-complexed and uncomplexed MGO Manuka Honey containing 250, 400 and 550 mg/kg MGO (Manuka Health NZ, Auckland, New Zealand) were evaluated in the in vitro phase. Cyclodextrin-complexed MGO Manuka Honey comprises α-cyclodextrin and Manuka honey solids, while uncomplexed Manuka honey comprises honey solids and water. Cyclodextrin-complexed and uncomplexed Manuka honey, and cyclodextrin, were resuspended in Difco BHI broth (Fort Richard Laboratories, Auckland, New Zealand) containing 0.1% w/v catalase (Sigma Aldrich, New South Wales, Australia) and incubated with shaking at 37°C for up to 60 min. Catalase neutralises hydrogen peroxide so that any effect on bacterial growth observed can be attributed to MGO.21

suspended in Difco BHI broth (Fort Richard Laboratories, Auckland, New Zealand) containing 0.1% w/v catalase (Sigma Aldrich, New South Wales, Australia) and incubated with shaking at 37°C for up to 60 min. Catalase neutralises hydrogen peroxide so that any effect on bacterial growth observed can be attributed to MGO.21 Bacterial strains The effect of cyclodextrin-complexed and uncomplexed Manuka honey, and cyclodextrin on bacterial growth was tested against bacteria commonly associated with blepharitis, including S. epidermidis ATCC (American Type Culture Collection) 14990, S. aureus ATCC 6538 and P. aeruginosa ATCC 27853 (ATCC, Virginia, USA). Time course of the effect on growth of S. aureus

Bacterial strains The effect of cyclodextrin-complexed and uncomplexed Manuka honey, and cyclodextrin on bacterial growth was tested against bacteria commonly associated with blepharitis, including S. epidermidis ATCC (American Type Culture Collection) 14990, S. aureus ATCC 6538 and P. aeruginosa ATCC 27853 (ATCC, Virginia, USA). Time course of the effect on growth of S. aureus A 17.8% w/v solution of catalase-treated cyclodextrin-complexed Manuka honey and a 9.7% w/v solution of uncomplexed Manuka honey containing honey with 250, 400 and 550 mg/kg MGO, and a 9.7% w/v solution of catalase-treated cyclodextrin were prepared. Both 17.8% and 9.7% w/v solutions of cyclodextrin-complexed Manuka honey and uncomplexed Manuka honey respectively contain 8% w/v honey solids, and a 9.7% w/v solution of cyclodextrin is equivalent to that present in a 17.8% w/v solution of cyclodextrin-complexed Manuka honey. Difco BHI broth alone was used as a control. Solutions were mixed with an equal volume of bacterial culture containing S. aureus ATCC 6538 at approximately 106 CFU/mL in Difco BHI broth, giving a total volume of 200 µL. Cultures were incubated in sterile, flat-bottomed, 96-well microplates (Sarstedt, Germany) for 24 hours at 37°C, with agitation at 100 rpm. Growth was measured as absorbance of the culture at 595 nm using a µQuant microplate spectrophotometer (BioTek, USA). Minimum inhibitory and bactericidal concentrations for S. aureus, P. aeruginosa and S. epidermidis

A 17.8% w/v solution of catalase-treated cyclodextrin-complexed Manuka honey and a 9.7% w/v solution of uncomplexed Manuka honey containing honey with 250, 400 and 550 mg/kg MGO, and a 9.7% w/v solution of catalase-treated cyclodextrin were prepared. Both 17.8% and 9.7% w/v solutions of cyclodextrin-complexed Manuka honey and uncomplexed Manuka honey respectively contain 8% w/v honey solids, and a 9.7% w/v solution of cyclodextrin is equivalent to that present in a 17.8% w/v solution of cyclodextrin-complexed Manuka honey. Difco BHI broth alone was used as a control. Solutions were mixed with an equal volume of bacterial culture containing S. aureus ATCC 6538 at approximately 106 CFU/mL in Difco BHI broth, giving a total volume of 200 µL. Cultures were incubated in sterile, flat-bottomed, 96-well microplates (Sarstedt, Germany) for 24 hours at 37°C, with agitation at 100 rpm. Growth was measured as absorbance of the culture at 595 nm using a µQuant microplate spectrophotometer (BioTek, USA). Minimum inhibitory and bactericidal concentrations for S. aureus, P. aeruginosa and S. epidermidis Solutions of catalase-treated cyclodextrin-complexed Manuka honey and uncomplexed Manuka honey containing 200 mg/kg MGO were prepared from cyclodextrin-complexed Manuka honey and uncomplexed Manuka honey containing honey with 400 mg/kg MGO. Each was diluted with BHI broth to give solutions containing 25, 50, 75, 100, 125, 150 and 175 mg/kg MGO. Solutions of 4.7%, 9.5%, 14%, 19%, 24%, 28%, 33% and 38% w/v catalase-treated cyclodextrin were also prepared; these concentrations of cyclodextrin are equivalent to those present in MGO Manuka Honey containing 25, 50, 75, 100, 125, 150, 175 and 200 mg/kg MGO, respectively. BHI broth served as a control. At each concentration, 180 µL honey solution was mixed with 20 µL bacteria to give 107 CFU/well. Cultures of S. aureus ATCC 6538, P. aeruginosa ATCC 27317 and S. epidermidis ATCC 14990 were incubated in clear, flat bottomed, 96-well plates (Greiner Bio-One, Austria) for 16 hours at 37°C with agitation at 100 rpm. Growth was measured as absorbance of the culture at 600 nm using a μQuant spectrophotometer (Bio-Tech Instruments, USA). Growth inhibition was defined as an A600 nm below 0.05 after 16 hours of growth (equal to uninoculated broth control). Samples (10 µL) of minimum inhibitory concentration (MIC) cultures not exhibiting growth were inoculated onto BHI agar (Fort Richard Laboratories) and incubated for 16 hours at 37°C with 5% CO2 in air. Each dilution was tested against each bacterial species in triplicate on three separate occasions.

ted broth control). Samples (10 µL) of minimum inhibitory concentration (MIC) cultures not exhibiting growth were inoculated onto BHI agar (Fort Richard Laboratories) and incubated for 16 hours at 37°C with 5% CO2 in air. Each dilution was tested against each bacterial species in triplicate on three separate occasions. Formulation preparation and cell viability evaluation An MGO MHME was developed and subsequently manufactured by Manuka Health New Zealand for in vivo evaluation. The microemulsion was prepared by mixing glycerol and polysorbate 80 (PS 80) before adding caprylic/capric triglyceride MI12 and phosphate buffered saline (PBS). Cyclodextrin-complexed MGO Manuka Honey 400 mg/kg (Manuka Health New Zealand) was then added to achieve a final MGO concentration of 100 mg/kg (MHME). For sterilisation, formulations were gamma irradiated (IR4; Atomic Energy, Canada) and osmolarity (VAPRO Vapor Pressure Osmometer; ELITech, New Zealand) and pH (Mettler Toledo pH Meter, New Zealand) of the formulations were determined before and after gamma exposure.

hieve a final MGO concentration of 100 mg/kg (MHME). For sterilisation, formulations were gamma irradiated (IR4; Atomic Energy, Canada) and osmolarity (VAPRO Vapor Pressure Osmometer; ELITech, New Zealand) and pH (Mettler Toledo pH Meter, New Zealand) of the formulations were determined before and after gamma exposure. For the purpose of toxicity testing, human corneal epithelial cells (HCEC) were maintained in minimum essential medium (MEM) supplemented with 10% fetal bovine serum, 1 g/L glucose, 100 U/mL penicillin and 100 µg/mL streptomycin. HCEC were derived from human cadaver tissue obtained with consent, for research use, from the donor's family, and with approval from the Northern X National Ethics Committee (NTX/07/08/080/AM04). Cells were cultured at 37°C in 5% CO2-95% atmospheric air until confluent and seeded at a density of 1×104 cells/well. After 24 hours of attachment, 100 µL of diluted (in MEM) formulation excipients and MHME were added for 15 min. Cells were then washed and incubated with 100 µL of MTT solution at 37°C for 3 hours before adding 100 µL of acidified isopropanol and measuring the absorbance at 570 nm. Negative controls were cultured in medium only and cell viability was expressed as a percentage of the negative control. In vivo phase (rabbit) The in vivo phase was approved by the institutional animal ethics committee (UAAEC 001156).

For the purpose of toxicity testing, human corneal epithelial cells (HCEC) were maintained in minimum essential medium (MEM) supplemented with 10% fetal bovine serum, 1 g/L glucose, 100 U/mL penicillin and 100 µg/mL streptomycin. HCEC were derived from human cadaver tissue obtained with consent, for research use, from the donor's family, and with approval from the Northern X National Ethics Committee (NTX/07/08/080/AM04). Cells were cultured at 37°C in 5% CO2-95% atmospheric air until confluent and seeded at a density of 1×104 cells/well. After 24 hours of attachment, 100 µL of diluted (in MEM) formulation excipients and MHME were added for 15 min. Cells were then washed and incubated with 100 µL of MTT solution at 37°C for 3 hours before adding 100 µL of acidified isopropanol and measuring the absorbance at 570 nm. Negative controls were cultured in medium only and cell viability was expressed as a percentage of the negative control. In vivo phase (rabbit) The in vivo phase was approved by the institutional animal ethics committee (UAAEC 001156). Paired-eye tolerability evaluation The 5-day prospective in vivo paired-eye tolerability evaluation was conducted on six healthy male New Zealand white rabbits (weighing 2.5±0.5 kg). The rabbits were treated with simultaneous daily administration of 20 µL of diluted MHME (1:10 with saline) to the right eye, and 20 µL of saline control to the left eye, for 5 consecutive days. The solutions were instilled into the lower fornix of the conjunctival sac using a positive displacement pipette. The eyelids were manually blinked several times following each instillation to facilitate solution distribution across the ocular surface.

nd 20 µL of saline control to the left eye, for 5 consecutive days. The solutions were instilled into the lower fornix of the conjunctival sac using a positive displacement pipette. The eyelids were manually blinked several times following each instillation to facilitate solution distribution across the ocular surface. Tear film and ocular surface evaluations were conducted at baseline, and 15 min following each instillation. The tests were conducted in ascending order of invasiveness to minimise the impact on tear film physiology for subsequent tests. Tear film lipid layer grade was assessed using the handheld Tearscope (Keeler, Berkshire, UK). Tear film lipid layer grading was based on the Guillon-Keeler grading system: grade 1, open meshwork; grade 2, closed meshwork; grade 3, wave/flow; grade 4, amorphous; grade 5, coloured fringes; grade 0, non-continuous layer (absent or abnormal coloured fringes).22 23 Tear evaporation rate was measured using an Evaporimeter EP3 (ServoMed, Sweden) within a modified goggle housing. Tear film osmolarity was evaluated with a clinical osmometer (TearLab, California, USA), using 50 nL tear samples collected from the lower lid tear meniscus. Aqueous tear production was evaluated using a phenol red thread hooked over the inferior lids into the fornices of the conjunctival sacs, with the wetted length measured after 15 s.

y was evaluated with a clinical osmometer (TearLab, California, USA), using 50 nL tear samples collected from the lower lid tear meniscus. Aqueous tear production was evaluated using a phenol red thread hooked over the inferior lids into the fornices of the conjunctival sacs, with the wetted length measured after 15 s. Conjunctival hyperaemia, corneal clarity and iris appearance were examined under a Burton lamp, and the severity graded clinically from 0 (none), through 1 (mild), 2 (moderate) to 3 (severe) judged to the nearest 0.25 increment. Sodium fluorescein dye was applied to the bulbar conjunctiva in order to evaluate the localised corneal and conjunctival areas of epithelial desiccation. Conjunctival and corneal staining was graded, respectively, on a 0–5 scale with increasing confluence in each area.24 Tear film osmolarity profile evaluation On day 1 of the 5-day in vivo tolerability evaluation, osmolarity measurements of both eyes were taken at baseline, and at 1 min intervals following instillation of either diluted MHME or saline control, for 10 min.

Conjunctival hyperaemia, corneal clarity and iris appearance were examined under a Burton lamp, and the severity graded clinically from 0 (none), through 1 (mild), 2 (moderate) to 3 (severe) judged to the nearest 0.25 increment. Sodium fluorescein dye was applied to the bulbar conjunctiva in order to evaluate the localised corneal and conjunctival areas of epithelial desiccation. Conjunctival and corneal staining was graded, respectively, on a 0–5 scale with increasing confluence in each area.24 Tear film osmolarity profile evaluation On day 1 of the 5-day in vivo tolerability evaluation, osmolarity measurements of both eyes were taken at baseline, and at 1 min intervals following instillation of either diluted MHME or saline control, for 10 min. Undiluted formulation safety evaluation The MHME formulation was developed for topical application onto the periocular skin of closed eyelids, without direct contact to the ocular surface. However, to evaluate the effect of accidental application to the ocular surface, 20 µL of undiluted MHME formulation was instilled into the right eye of each rabbit, and 20 µL of saline control to the left eye simultaneously, following the conclusion of the 5-day in vivo tolerability evaluation. Ocular surface characteristics were assessed at baseline, and 30 s, 5 min and 10 min following instillation.

L of undiluted MHME formulation was instilled into the right eye of each rabbit, and 20 µL of saline control to the left eye simultaneously, following the conclusion of the 5-day in vivo tolerability evaluation. Ocular surface characteristics were assessed at baseline, and 30 s, 5 min and 10 min following instillation. Statistics Statistical analyses were performed using IBM SPSS Statistics V.22 and GraphPad Prism V.6.02. For the in vitro phase, area under the growth curve (AUC) values were calculated in GraphPad Prism V.6.02. AUCs were generated for each microplate well by plotting increasing culture absorbance (a measure of growth) against time. Comparisons of AUC values (as a relative quantity of control AUC) were conducted using Kruskal-Wallis tests, and post hoc multiplicity-adjusted Dunn's tests. For the in vivo phase, repeated measures (RM) two-way analysis of variance (ANOVA) testing was performed to test the significance of treatment, time and interaction (treatment-by-time) effects on continuous measurements, and ordinal data were converted to rank values prior to undergoing analysis. Post hoc multiplicity-adjusted Sidak's tests were conducted to examine immediate treatment effects at each time point, and cumulative treatment effects over the treatment period. All tests were two tailed and p <0.05 was considered significant. Results

For the in vivo phase, repeated measures (RM) two-way analysis of variance (ANOVA) testing was performed to test the significance of treatment, time and interaction (treatment-by-time) effects on continuous measurements, and ordinal data were converted to rank values prior to undergoing analysis. Post hoc multiplicity-adjusted Sidak's tests were conducted to examine immediate treatment effects at each time point, and cumulative treatment effects over the treatment period. All tests were two tailed and p <0.05 was considered significant. Results In vitro phase Microbiological evaluations Summary statistics of the in vitro microbiology evaluation are presented in table 1. In the assessment of time course effects on S. aureus growth, cyclodextrin-complexed Manuka honey displayed effects at least as strong as uncomplexed honey at each concentration of MGO tested. Cyclodextrin-complexed honey effected a greater reduction in AUC at 250 and 550 mg/kg MGO (both p<0.01), but not 400 mg/kg MGO (p=0.98). Although cyclodextrin alone also reduced AUC, the effect was significantly weaker than cyclodextrin-complexed or uncomplexed honey at all concentrations of MGO tested (all p<0.01). The cyclodextrin-complexed honey with MGO levels of 400 mg/kg and 550 mg/kg had greater efficacy than that containing 250 mg/kg, but no difference was found to exist between the 400 mg/kg and 550 mg/kg levels.

ntly weaker than cyclodextrin-complexed or uncomplexed honey at all concentrations of MGO tested (all p<0.01). The cyclodextrin-complexed honey with MGO levels of 400 mg/kg and 550 mg/kg had greater efficacy than that containing 250 mg/kg, but no difference was found to exist between the 400 mg/kg and 550 mg/kg levels. Table 1 Relative AUC for S. aureus ATCC 6838, MIC and MBC for S. aureus ATCC 6838, S. epidermidis 14990, and P. aeruginosa ATCC 27853 of cyclodextrin-complexed MGO Manuka Honey, uncomplexed Manuka honey and cyclodextrin. Relative AUC data are presented as mean±SD. Asterisks denote statistically significant differences (p<0.05). Cyclodextrin-complexed Manuka honey Uncomplexed Manuka honey Cyclodextrin p Relative AUC of S. aureus 0 mg/kg MGO – – 0.877±0.059 250 mg/kg MGO 0.502±0.036 0.741±0.026 – 0.001* 400 mg/kg MGO 0.387±0.012 0.381±0.011 – 0.98 550 mg/kg MGO 0.375±0.035 0.594±0.029 – 0.002* MIC (mg/kg MGO) S. aureus 125 175 >200 S. epidermidis 125 175 >200 P. aeruginosa 150 150 >200 MBC (mg/kg MGO) S. aureus 150 >200 >200 S. epidermidis 150 >200 >200 P. aeruginosa >200 >200 >200 ATCC, American Type Culture Collection; AUC, area under the growth curve; MBC, minimum bactericidal concentration; MGO, methylglyoxal; MIC, minimum inhibitory concentration. As cyclodextrin-complexed Manuka honey and uncomplexed Manuka honey with an MGO level of 250 mg/kg exhibited antibacterial activity in the time course experiment, MGO levels under 250 mg/kg were subsequently assessed for their bacteriostatic and bactericidal effects with BHI broth serving as a control.

P. aeruginosa >200 >200 >200 ATCC, American Type Culture Collection; AUC, area under the growth curve; MBC, minimum bactericidal concentration; MGO, methylglyoxal; MIC, minimum inhibitory concentration. As cyclodextrin-complexed Manuka honey and uncomplexed Manuka honey with an MGO level of 250 mg/kg exhibited antibacterial activity in the time course experiment, MGO levels under 250 mg/kg were subsequently assessed for their bacteriostatic and bactericidal effects with BHI broth serving as a control. Bacterial growth patterns were consistent between all replicates of each assay for determining MIC and minimum bactericidal concentration. Cyclodextrin-complexed Manuka honey exhibited greater inhibitory and bactericidal actions on S. aureus and S. epidermidis than uncomplexed honey, although antibacterial effects against P. aeruginosa were comparable. At the levels assayed, no inhibitory or bactericidal actions were observed with cyclodextrin alone, while uncomplexed honey did not exhibit any bactericidal effects, and cyclodextrin-complexed Manuka honey was not bactericidal towards P. aeruginosa. Taking into account the results of the AUC evaluations, as well as the commercial availability of the different concentrations of the cyclodextrin-complexed Manuka honey, the 400 mg/kg MGO concentration of cyclodextrin-complexed Manuka honey was selected for further evaluation.

Bacterial growth patterns were consistent between all replicates of each assay for determining MIC and minimum bactericidal concentration. Cyclodextrin-complexed Manuka honey exhibited greater inhibitory and bactericidal actions on S. aureus and S. epidermidis than uncomplexed honey, although antibacterial effects against P. aeruginosa were comparable. At the levels assayed, no inhibitory or bactericidal actions were observed with cyclodextrin alone, while uncomplexed honey did not exhibit any bactericidal effects, and cyclodextrin-complexed Manuka honey was not bactericidal towards P. aeruginosa. Taking into account the results of the AUC evaluations, as well as the commercial availability of the different concentrations of the cyclodextrin-complexed Manuka honey, the 400 mg/kg MGO concentration of cyclodextrin-complexed Manuka honey was selected for further evaluation. Formulation preparation and in vitro human corneal epithelial cell viability evaluation After addition of the MGO Manuka Honey complex, the clear microemulsion underwent phase transition into a liquid crystalline semisolid suitable for application to the eyelids. Gamma irradiation had no significant effect on the physical stability, pH and osmolarity (6.29±0.25 and 603.4±10.9 mOsm/kg, respectively; 1:10 in PBS).

ddition of the MGO Manuka Honey complex, the clear microemulsion underwent phase transition into a liquid crystalline semisolid suitable for application to the eyelids. Gamma irradiation had no significant effect on the physical stability, pH and osmolarity (6.29±0.25 and 603.4±10.9 mOsm/kg, respectively; 1:10 in PBS). HCEC did not show significant toxicity following 15 min application of formulation excipients or the final MHME (figure 1) with the exception of the 1:1 dilution of PS 80 (cell viability 10.95%±0.26%). However, application of a concentration of surfactant equivalent to that incorporated in the final formulation proved to be non-toxic (cell viability >95%). Figure 1 In vitro percentage cell viability following application of formulation excipients and the final MHME (1:10 dilution) to human corneal epithelial cells for 15 min. Each bar represents the mean percentage cell viability. Error bars represent the SD. CCTG, caprylic/capric triglyceride MI12; MHME, Manuka honey microemulsion; PS 80, polysorbate 80; SD, standard deviation. In vivo phase Paired-eye tolerability evaluation

Figure 1 In vitro percentage cell viability following application of formulation excipients and the final MHME (1:10 dilution) to human corneal epithelial cells for 15 min. Each bar represents the mean percentage cell viability. Error bars represent the SD. CCTG, caprylic/capric triglyceride MI12; MHME, Manuka honey microemulsion; PS 80, polysorbate 80; SD, standard deviation. In vivo phase Paired-eye tolerability evaluation Table 2 illustrates the summary statistics of the in vivo rabbit paired-eye tolerability evaluation. There were no significant differences in baseline tear film and ocular surface measurements between the eyes instilled with diluted MHME or saline control (all p>0.05). RM-ANOVA demonstrated no significant treatment, time, or treatment-by-time effects for any of parameters assessed during the 5-day period (all p>0.05). Post hoc multiplicity-adjusted Sidak's analysis showed that there were no immediate treatment effects following instillation of diluted MHME or saline control at all time points assessed (all p>0.05), and no significant cumulative treatment effects were observed during the trial period (all p>0.05). Table 2 In vivo tear film parameters and ocular surface characteristics during the 5-week trial period of the eyes of rabbits instilled with undiluted MHME (1:10) or saline control. Data are presented as median (range) or mean±SD. Asterisks denote statistically significant differences (p<0.05).

Table 2 illustrates the summary statistics of the in vivo rabbit paired-eye tolerability evaluation. There were no significant differences in baseline tear film and ocular surface measurements between the eyes instilled with diluted MHME or saline control (all p>0.05). RM-ANOVA demonstrated no significant treatment, time, or treatment-by-time effects for any of parameters assessed during the 5-day period (all p>0.05). Post hoc multiplicity-adjusted Sidak's analysis showed that there were no immediate treatment effects following instillation of diluted MHME or saline control at all time points assessed (all p>0.05), and no significant cumulative treatment effects were observed during the trial period (all p>0.05). Table 2 In vivo tear film parameters and ocular surface characteristics during the 5-week trial period of the eyes of rabbits instilled with undiluted MHME (1:10) or saline control. Data are presented as median (range) or mean±SD. Asterisks denote statistically significant differences (p<0.05). Preinstillation Postinstillation Preinstillation versus postinstillation p values MHME Saline control p MHME Saline control p MHME Saline control Tear film lipid layer grade Day 1 5 (4–5) 5 (4–5) 0.88 5 (4–5) 5 (4–5) 0.79  0.95 0.95 Day 2 4.5 (4–5) 5 (4–5) 0.88 5 (5–5) 5 (4–5) 0.16 0.16 >0.99 Day 3 5 (4–5) 4.5 (4–5) 0.88 5 (4–5) 5 (4–5) 0.79 0.95 0.95 Day 4 5 (4–5) 5 (4–5) 0.88 5 (5–5) 5 (4–5) 0.79 0.95 0.95 Day 5 5 (4–5) 5 (4–5) 0.88 5 (5–5) 5 (4–5) 0.16 0.95 >0.99 p 0.75 0.86 0.41 0.86 RM-ANOVA for treatment effect 0.92 0.67 0.18 0.92 RM-ANOVA for time effect 0.61 0.14 0.65 0.83 RM-ANOVA for interaction effect 0.38 0.82 0.77 0.79 Tear film osmolarity (mOsmol/kg) Day 1 341±25 361±236 0.87 318±14 357±23 0.09 0.13 >0.99 Day 2 353±16 356±28 >0.99 338±32 345±27 >0.99 0.80 0.83 Day 3 339±16 349±22 >0.99 325±14 334±33 >0.99 0.97 0.67 Day 4 342±31 347±16 >0.99 328±32 330±11 >0.99 0.92 0.53 Day 5 343±17 351±29 >0.99 332±34 336±22 >0.99 >0.99 0.66 p 0.41 0.81 0.45 0.16 RM-ANOVA for treatment effect 0.76 0.10 0.38 0.22 RM-ANOVA for time effect 0.18 0.47 0.12 0.18 RM-ANOVA for interaction effect 0.93 0.13 0.94 0.96 Tear evaporation rate (g/m2/hour) Day 1 22±13 24±8 0.46 24±5 29±9 0.59 0.80 0.60 Day 5 15±7 14±5 0.11 19±8 20±6 0.08 0.46 0.82 p 0.58 0.61 0.48 0.08 RM-ANOVA for treatment effect 0.98 0.08 0.13 0.16 RM-ANOVA for time effect 0.10 0.35 0.10 0.33 RM-ANOVA for interaction effect 0.44 0.31 0.68 0.81 Phenol red thread (mm) Day 1 23±3 21±4 0.82 23±5 23±4 >0.99 0.99 0.43 Day 5 20±7 21±4 0.87 23±3 23±4 >0.99 0.31 0.66 p 0.53 >0.99 0.85 0.86 RM-ANOVA for treatment effect 0.66 0.95 0.79 0.20 RM-ANOVA for time effect 0.91 0.89 0.38 >0.99 RM-ANOVA for interaction effect 0.48 0.59 0.33 0.77 Conjunctival hyperaemia grade Day 1 0.75 (0–1.5) 0.75 (0–1) 0.16 0.75 (0–1) 0.75 (0–1) >0.99 0.52 >0.99 Day 2 0.5 (0–0.5) 0.50 (0–0.75) >0.99 0.25 (0–0.5) 0.25 (0.25–1) 0.23 0.95 >0.99 Day 3 0.25 (0–0.5) 0.25 (0–0.5) >0.99 0.25 (0–0.5) 0.25 (0–0.75) >0.09 0.95 >0.99 Day 4 0.25 (0.25–0.5) 0.25 (0–0.5) >0.99 0.25 (0–0.5) 0.5 (0–1) 0.07 >0.99 0.40 Day 5 0.25 (0–0.5) 0.50 (0–0.5) >0.99 0.25 (0–0.5) 0.25 (0–0.5) >0.99 >0.99 0.92 p 0.19 0.15 0.68 0.21 RM-ANOVA for treatment effect 0.14 0.08 0.4

0.95 >0.99 Day 3 0.25 (0–0.5) 0.25 (0–0.5) >0.99 0.25 (0–0.5) 0.25 (0–0.75) >0.09 0.95 >0.99 Day 4 0.25 (0.25–0.5) 0.25 (0–0.5) >0.99 0.25 (0–0.5) 0.5 (0–1) 0.07 >0.99 0.40 Day 5 0.25 (0–0.5) 0.50 (0–0.5) >0.99 0.25 (0–0.5) 0.25 (0–0.5) >0.99 >0.99 0.92 p 0.19 0.15 0.68 0.21 RM-ANOVA for treatment effect 0.14 0.08 0.4 7 0.34 RM-ANOVA for time effect 0.80 0.17 0.93 0.64 RM-ANOVA for interaction effect 0.12 0.43 0.85 0.43 Corneal opacity grade Day 1 0 (0–0) 0 (0–0) >0.99 0 (0–0) 0 (0–0) >0.99 >0.99 >0.99 Day 2 0 (0–0) 0 (0–0) >0.99 0 (0–0) 0 (0–0) >0.99 >0.99 >0.99 Day 3 0 (0–0) 0 (0–0) >0.99 0 (0–0) 0 (0–0) >0.99 >0.99 >0.99 Day 4 0 (0–0) 0 (0–0) >0.99 0 (0–0) 0 (0–0) >0.99 >0.99 >0.99 Day 5 0 (0–0) 0 (0–0) >0.99 0 (0–0) 0 (0–0) >0.99 >0.99 >0.99 p >0.99 >0.99 >0.99 >0.99 RM-ANOVA for treatment effect >0.99 >0.99 >0.99 >0.99 RM-ANOVA for time effect >0.99 >0.99 >0.99 >0.99 RM-ANOVA for interaction effect >0.99 >0.99 >0.99 >0.99 Iris appearance grade Day 1 0 (0–0) 0 (0–0) >0.99 0 (0–0) 0 (0–0) >0.99 >0.99 >0.99 Day 2 0 (0–0) 0 (0–0) >0.99 0 (0–0) 0 (0–0) >0.99 >0.99 >0.99 Day 3 0 (0–0) 0 (0–0) >0.99 0 (0–0) 0 (0–0) >0.99 >0.99 >0.99 Day 4 0 (0–0) 0 (0–0) >0.99 0 (0–0) 0 (0–0) >0.99 >0.99 >0.99 Day 5 0 (0–0) 0 (0–0) >0.99 0 (0–0) 0 (0–0) >0.99 >0.99 >0.99 p >0.99 >0.99 >0.99 >0.99 RM-ANOVA for treatment effect >0.99 >0.99 >0.99 >0.99 RM-ANOVA for time effect >0.99 >0.99 >0.99 >0.99 RM-ANOVA for interaction effect >0.99 >0.99 >0.99 >0.99 Conjunctival sodium fluorescein staining score Day 1 0.25 (0–0.5) 0 (0–0) 0.35 0 (0–0.5) 0.25 (0–1) 0.93 0.99 0.43 Day 2 0 (0–1) 0 (0–1.5) 0.35 0 (0–1) 0.5 (0–1.5) 0.20 0.85 0.80 Day 3 0 (0–0.25) 0 (0–0.5) >0.99 0 (0–0.5) 0 (0–0.25) >0.99 0.95 >0.99 Day 4 0 (0–0) 0.25 (0–0.5) 0.35 0 (0–0.25) 0 (0–0.25) >0.99 0.99 0.99 Day 5 0 (0–0) 0 (0–0.25) >0.99 0 (0–0) 0 (0–1) >0.99 >0.99 0.99 p >0.99 >0.99 0.68 0.21 RM-ANOVA for treatment effect 0.39 0.21 0.78 0.53 RM-ANOVA for time effect 0.13 0.12 0.42 0.37 RM-ANOVA for interaction effect 0.07 0.47 0.82 0.21 Corneal sodium fluorescein staining score Day 1 0 (0–0) 0 (0–0.5) 0.33 0 (0–0) 0 (0–0.5) 0.82 >0.99 >0.99 Day 2 0 (0–0) 0 (0–0) >0.99 0 (0–0) 0 (0–0) >0.99 >0.99 >0.99 Day 3 0 (0–0) 0 (0–0) >0.99 0 (0–0) 0 (0–0.25) 0.82 >0.99 0.14 Day 4 0 (0–0) 0 (0–0.25) >0.99 0 (0–0) 0 (0–0.25) 0.82 0.17 0.14 Day 5 0 (0–0) 0 (0–0.25) >0.99 0 (0–0) 0 (0–0) >0.99 >0.99 0.77 p >0.99 0.65 0.41 0.15 RM-ANOVA for treatment effect 0.68 0.18 0.43 0.59 RM-ANOVA for time effect 0.12 0.11 0.36 0.30 RM-ANOVA for interaction effe

0) >0.99 0 (0–0) 0 (0–0.25) 0.82 >0.99 0.14 Day 4 0 (0–0) 0 (0–0.25) >0.99 0 (0–0) 0 (0–0.25) 0.82 0.17 0.14 Day 5 0 (0–0) 0 (0–0.25) >0.99 0 (0–0) 0 (0–0) >0.99 >0.99 0.77 p >0.99 0.65 0.41 0.15 RM-ANOVA for treatment effect 0.68 0.18 0.43 0.59 RM-ANOVA for time effect 0.12 0.11 0.36 0.30 RM-ANOVA for interaction effe ct 0.68 0.65 0.43 0.75 MHME, Manuka honey microemulsion; RM-ANOVA, repeated measures analysis of variance; SD, standard deviation. Tear film osmolarity profile evaluation Baseline tear film osmolarity did not differ significantly between the eyes instilled with diluted MHME or saline control (p=0.99, figure 2). RM-ANOVA showed that there were no significant treatment, time, or treatment by time interaction effects for tear film osmolarity during the 10 min period following instillation (all p>0.05). Multiplicity-adjusted Sidak's post hoc analysis did not detect significant treatment effects at any of the time points assessed (all p>0.05). Figure 2 In vivo tear film osmolarity profile during the 10 min period following instillation of diluted MHME (1:10) or saline control to rabbit eyes. Each point represents the mean tear film osmolarity at a given time point: solid circles represent eyes instilled with diluted MHME, and hollow circles represent eyes instilled with saline control. Error bars represent the SD. Post hoc multiplicity-adjusted Sidak p values of the treatment effects at each time point are provided above the error bars. MHME, Manuka honey microemulsion; SD, standard deviation.

represent eyes instilled with diluted MHME, and hollow circles represent eyes instilled with saline control. Error bars represent the SD. Post hoc multiplicity-adjusted Sidak p values of the treatment effects at each time point are provided above the error bars. MHME, Manuka honey microemulsion; SD, standard deviation. Undiluted formulation safety evaluation Summary statistics of the in vivo undiluted formulation safety evaluation are illustrated in table 3. No significant differences in baseline ocular surface characteristics were observed between the eyes instilled with diluted MHME or saline control. Table 3 In vivo ocular surface characteristics during the 10 min period following instillation of undiluted MHME (1:10) or saline control to rabbit eyes. Data are presented as median (range). Asterisks denote statistically significant differences (p<0.05).

Undiluted formulation safety evaluation Summary statistics of the in vivo undiluted formulation safety evaluation are illustrated in table 3. No significant differences in baseline ocular surface characteristics were observed between the eyes instilled with diluted MHME or saline control. Table 3 In vivo ocular surface characteristics during the 10 min period following instillation of undiluted MHME (1:10) or saline control to rabbit eyes. Data are presented as median (range). Asterisks denote statistically significant differences (p<0.05). Time MHME Saline control p Conjunctival hyperaemia grade Baseline 0.500 (0.00-0.75) 0.500 (0.00-0.75) >0.99 30 s 1.125 (1.00-1.5) 0.500 (0.25-0.75) <0.001* 5 min 1.000 (0.75-1.25) 0.625 (0.25-1.00) <0.001* 10 min 0.500 (0.50-1.00) 0.500 (0.00-0.75) 0.85 RM-ANOVA for treatment effect 0.001* RM-ANOVA for time effect 0.002* RM-ANOVA for interaction effect <0.001* Corneal opacity grade Baseline 0 (0–0) 0 (0–0) >0.99 30 s 0 (0–0) 0 (0–0) >0.99 5 min 0 (0–0) 0 (0–0) >0.99 10 min 0 (0–0) 0 (0–0) >0.99 RM-ANOVA for treatment effect >0.99 RM-ANOVA for time effect >0.99 RM-ANOVA for interaction effect >0.99 Iris appearance grade Baseline 0 (0–0) 0 (0–0) >0.99 30 s 0 (0–0) 0 (0–0) >0.99 5 min 0 (0–0) 0 (0–0) >0.99 10 min 0 (0–0) 0 (0–0) >0.99 RM-ANOVA for treatment effect >0.99 RM-ANOVA for time effect >0.99 RM-ANOVA for interaction effect >0.99 Ocular surface sodium fluorescein staining score Baseline 0 (0–0.250) 0 (0–0) 0.58 30 s 0 (0–0.250) 0 (0–0) 0.58 5 min 0 (0–0.125) 0 (0–0) 0.94 10 min 0 (0–0.050) 0 (0–0) 0.99 RM-ANOVA for treatment effect 0.36 RM-ANOVA for time effect 0.42 RM-ANOVA for interaction effect 0.42 MHME, Manuka honey microemulsion; RM-ANOVA, repeated measures analysis of variance.

g score Baseline 0 (0–0.250) 0 (0–0) 0.58 30 s 0 (0–0.250) 0 (0–0) 0.58 5 min 0 (0–0.125) 0 (0–0) 0.94 10 min 0 (0–0.050) 0 (0–0) 0.99 RM-ANOVA for treatment effect 0.36 RM-ANOVA for time effect 0.42 RM-ANOVA for interaction effect 0.42 MHME, Manuka honey microemulsion; RM-ANOVA, repeated measures analysis of variance. Significant treatment, time, and treatment-by-time interaction effects for conjunctival hyperaemia grade were observed (all p<0.01). Post hoc multiplicity-adjusted Sidak's analysis detected elevated conjunctival hyperaemia grades relative to baseline in the MHME group at 30 s and 5 min (both p<0.001), and these values were significantly greater than the saline control group (both p<0.001). At 10 min, conjunctival hyperaemia in the MHME group decreased to levels that were not significantly different from baseline (p=0.37), and there were no differences between the MHME and saline control groups (p=0.85). No significant treatment, time, and treatment-by-time interaction effects for corneal opacity, iris appearance or fluorescein staining were detected (all p>0.05).

ed to levels that were not significantly different from baseline (p=0.37), and there were no differences between the MHME and saline control groups (p=0.85). No significant treatment, time, and treatment-by-time interaction effects for corneal opacity, iris appearance or fluorescein staining were detected (all p>0.05). Discussion The in vitro phase demonstrated antimicrobial activity of both cyclodextrin-complexed and uncomplexed MGO Manuka Honey solutions against bacterial species commonly associated with blepharitis, following pretreatment with catalase. The neutralisation of hydrogen peroxide by catalase treatment indicates that the bioactivity is attributable to other constituents including MGO, which has previously been identified as a significant antimicrobial factor in Manuka honey.21 25 MGO is believed to exert its effects via inhibition of murein hydrolase which is involved in the breakdown of peptidoglycan during cellular division.26 It is conceivable that with less peptidoglycan,27 and thus less murein hydrolase, Gram-negative organisms may be less susceptible to the antimicrobial effects of MGO than Gram-positive bacteria.28 This was demonstrated by the stronger sensitivity of S. aureus and S. epidermidis towards MGO than P. aeruginosa in the current study, and is consistent with previous studies which also report discrepancies between the response to Gram-positive and Gram-negative organisms.20 29 30 Nevertheless, blepharitis is more commonly associated with overcolonisation of S. aureus and S. epidermidis,6 against which MGO Manuka Honey solutions exhibited greater activity. It is also acknowledged that laboratory-adapted bacterial strains, used in our in vitro testing, have the potential to behave differently from clinical strains.31

itis is more commonly associated with overcolonisation of S. aureus and S. epidermidis,6 against which MGO Manuka Honey solutions exhibited greater activity. It is also acknowledged that laboratory-adapted bacterial strains, used in our in vitro testing, have the potential to behave differently from clinical strains.31 Cyclodextrin-complexed Manuka honey solutions were generally shown to have stronger antimicrobial effect than uncomplexed honey, with smaller relative AUCs being observed in cultures containing cyclodextrin-complexed Manuka honey than uncomplexed honey. The magnitude of this effect paralleled the concentration of MGO. These findings are consistent with the previously reported enhancement of the antibacterial properties of Manuka honey following complexation with cyclodextrin.20 This is thought to be attributable to the slowed and sustained release of active honey constituents, and thereby the maintenance of an inhibitory concentration of these factors for an extended time.20 Complexation to cyclodextrin has been shown to reduce the release rate of many hydrophobic drugs to sustain a therapeutic concentration.18 Furthermore, in agreement with previously reported findings,20 cyclodextrin, alone, exhibited some antibacterial activity in the current study. However, these effects were significantly lower than those displayed by cyclodextrin-complexed or uncomplexed Manuka honey regardless of MGO concentration. Nevertheless, this suggests that the enhanced antimicrobial activity in cyclodextrin-complexed honey may be attributable to a combination of the complexation reaction and synergistic effects.

significantly lower than those displayed by cyclodextrin-complexed or uncomplexed Manuka honey regardless of MGO concentration. Nevertheless, this suggests that the enhanced antimicrobial activity in cyclodextrin-complexed honey may be attributable to a combination of the complexation reaction and synergistic effects. A microemulsion similar to that developed and described here has previously been reported to undergo phase transition from a clear liquid to a semisolid liquid crystal upon reduction of the glycerol and/or Tween 80 content.32 As cyclodextrins are capable of forming complexes with many excipients including glycerol,33 and Tween 80,34 it is likely that cyclodextrins from MGO Manuka Honey complexed glycerol and/or PS 80 leading to a component ratio shift and thus phase transition. This complexation behaviour is also supported by the cell viability data with death of almost 90% of cells exposed to PS 80 (1:1 dilution) for 15 min while PS 80 incorporated into the final formulation, and thus complexed and less available to the environment, caused no significant cytotoxicity. It should be noted here that, while the tested exposure time of 15 min is short, any formulation components accidentally entering the eye would be washed away quickly by reflex tearing and rapid tear turnover thus concentrations tested here are unlikely ever to be reached. Overall, formulations were found suitable for eyelid application with no anticipated significant ocular adverse effects in the event of formulation components entering the eye across the lid margins.

away quickly by reflex tearing and rapid tear turnover thus concentrations tested here are unlikely ever to be reached. Overall, formulations were found suitable for eyelid application with no anticipated significant ocular adverse effects in the event of formulation components entering the eye across the lid margins. The subsequent in vivo phase, evaluating the safety and tolerability of the MHME in rabbit eyes, showed that measurements of lipid layer grade, tear evaporation rate, tear film osmolarity, phenol red thread, fluorescein staining, conjunctival hyperaemia, corneal opacity and iris appearance grades, did not change following instillation of either diluted MGO MHME or saline control. Furthermore, there were no significant changes in these measurements during the 5-day period in both groups, and no differences between groups. This suggests that instillation of diluted MHME was not associated with tear film destabilisation, ocular surface irritation, inflammation or epithelial damage in the rabbit eye. Furthermore, the MHME formulation is intended to be used as a cream applied onto the periocular skin of closed eyelids. However, this mode of application was not possible in the rabbit model due to fur around the eyes, and thus direct ocular surface instillation of an MHME at 1:10 dilution in saline was used in the current study. The migration of periocular particles into the tear film can occur in human subjects due to the action of the muscles of Riolan,35 36 and the surface tension at the tear meniscus.37 However, it is unlikely for the concentrations that reach the tear film with periocular application in humans to exceed the levels directly instilled into the rabbit eyes in the current study. Nevertheless, despite the favourable findings of this animal study, future tolerability trials on human subjects are required to confirm the clinical safety of MHME.

trations that reach the tear film with periocular application in humans to exceed the levels directly instilled into the rabbit eyes in the current study. Nevertheless, despite the favourable findings of this animal study, future tolerability trials on human subjects are required to confirm the clinical safety of MHME. Although the MGO MHME is designed for topical application to the periocular skin of closed eyelids in human patients, the potential for accidental instillation directly onto the ocular surface cannot be discounted. Such a situation was simulated in the in vivo phase through the direct ocular instillation of undiluted MHME in the rabbits, without the immediate aqueous flushing that would be recommended as standard in clinical use following accidental instillation. Although the conjunctival hyperaemia grade was elevated at 30 s following undiluted MHME instillation, levels returned to baseline within 10 min. The results of the in vivo phase need to be interpreted cautiously in the context of the methodological limitations, which preclude the direct extrapolation of the findings to the clinical safety and tolerability in human patients. There are significant differences in the anatomy and physiology of the eyes of rabbits and humans. Nevertheless, the rabbit is one of the closest models of the human ocular surface and tear film, and preclinical animal studies are required prior to the conduct of clinical trials in human subjects.

ty in human patients. There are significant differences in the anatomy and physiology of the eyes of rabbits and humans. Nevertheless, the rabbit is one of the closest models of the human ocular surface and tear film, and preclinical animal studies are required prior to the conduct of clinical trials in human subjects. Of note, cyclodextrin-complexed Manuka honey generally exhibited stronger in vitro antimicrobial effects than uncomplexed honey on bacteria commonly associated with blepharitis. In vivo safety and tolerability evaluation of MGO MHME on rabbit eyes did not result in significant immediate or cumulative adverse effects. These findings therefore support clinical safety and tolerability evaluation of MGO MHME in human subjects. Acknowledgements: The authors are grateful to Angela Cunningham, BSc; Yen-Heng (Amy) Chen, BOptom; May Young, BOptom; Iana Pearce, BOptom; and William Perriam, BOptom, for their involvement in preliminary aspects of the work reported in this manuscript. Contributors: All authors took an active part in the design, conduct, data analysis, and publication drafting and approval. Funding: This research received unrestricted grant funding from the Faculty of Medical and Health Sciences at the University of Auckland and from Manuka Health, who themselves received grant support from the Ministry of Business, Innovation and Employment (MBIE PROP31150VCHMANUKAHEALTH). Competing interests: None declared. Ethics approval: University of Auckland Animal Ethics Committee UAAEC 001156. Provenance and peer review: Commissioned; externally peer reviewed.

Funding: This research received unrestricted grant funding from the Faculty of Medical and Health Sciences at the University of Auckland and from Manuka Health, who themselves received grant support from the Ministry of Business, Innovation and Employment (MBIE PROP31150VCHMANUKAHEALTH). Competing interests: None declared. Ethics approval: University of Auckland Animal Ethics Committee UAAEC 001156. Provenance and peer review: Commissioned; externally peer reviewed. Data sharing statement: All data relating to the study are published in the manuscript and approaches to share this can be made to the corresponding author. Queries relating to the MGO Manuka Honey microemulsion can be made to Manuka Health New Zealand.

Key messages What is already known about this subject? Glaucoma is the leading cause of irreversible blindness globally. Poor awareness and late presentation often militate against effective management of the condition. What are the new findings? Awareness does not invariably translate to adequate knowledge of glaucoma. The mass media is a potentially effective tool for disseminating information about glaucoma. How might these results change the focus of research or clinical practice? For awareness efforts to be efficient, their contents have to be tailored to suit the needs of the target audience. Screening programmes are important in creating awareness and detecting glaucoma suspects and cases , and may help alleviate the problem of late presentation. Introduction Glaucoma is a group of diseases that cause structural damage and visual field dysfunction, leading to progressive and irreversible vision loss.1 It is the second leading cause of blindness globally, accounting for 8% of blindness.2 Open angle glaucoma is the predominant type in people of African descent, in whom it has an earlier onset and more aggressive progression.1 3 In Nigeria, the prevalence is 5.02% (95% CI 4.60% to 5.47%) in people aged 40 years and older, with open-angle glaucoma and angle closure glaucoma comprising 86% and 14% of cases, respectively.4 Secondary glaucoma accounts for 8% of cases.4

Introduction Glaucoma is a group of diseases that cause structural damage and visual field dysfunction, leading to progressive and irreversible vision loss.1 It is the second leading cause of blindness globally, accounting for 8% of blindness.2 Open angle glaucoma is the predominant type in people of African descent, in whom it has an earlier onset and more aggressive progression.1 3 In Nigeria, the prevalence is 5.02% (95% CI 4.60% to 5.47%) in people aged 40 years and older, with open-angle glaucoma and angle closure glaucoma comprising 86% and 14% of cases, respectively.4 Secondary glaucoma accounts for 8% of cases.4 Open-angle glaucoma usually runs a symptomless but progressive course, affecting the peripheral vision first. Therefore, many affected remain undiagnosed, presenting only when the disease has progressed significantly to affect the central vision late in the disease process. Late presentation is therefore an important setback in glaucoma management.5–7 In a review of population-based studies by Quigley and Broman,8 the global rate of previous diagnosis of open-angle glaucoma was low at 26%, and even lower at 8% for developing countries. In Nigeria, the rate was 5.6%,4 and more than half of patients are blind in one or both eyes at presentation.5 9 10 Late presentation has been attributed to low level of awareness about glaucoma, especially in developing countries.7 11–14 Creating awareness about the disease, its insidious nature and the importance of early detection and treatment in preventing blindness is therefore key in glaucoma control.

s at presentation.5 9 10 Late presentation has been attributed to low level of awareness about glaucoma, especially in developing countries.7 11–14 Creating awareness about the disease, its insidious nature and the importance of early detection and treatment in preventing blindness is therefore key in glaucoma control. There is currently no consensus on a precise and easily implemented screening test or strategy for glaucoma detection.15 Population screening has not been found to be cost-effective,16 and opportunistic case finding may miss a majority of those with glaucoma.17 Screening targeted at high-risk groups such as blacks and people with a positive family history in first degree relatives16 and incorporation of glaucoma screening into other eye screening programmes such as cataract screening programmes have been suggested as more feasible control strategies.18 19 A study in Ibadan, Nigeria, that evaluated the effectiveness of screening programmes in early detection of glaucoma18 revealed that patients referred from outreach screening programmes were more likely to present with mild to moderate diseases than their counterparts referred from other sources. Such screening programmes may therefore be invaluable in early detection of glaucoma and may serve as feasible means of improving public awareness of the condition. Evidence necessary for planning can also be generated by examining the demographic features and visual characteristics of patients attending such outreach programmes.

ing programmes may therefore be invaluable in early detection of glaucoma and may serve as feasible means of improving public awareness of the condition. Evidence necessary for planning can also be generated by examining the demographic features and visual characteristics of patients attending such outreach programmes. The aim of this study was to determine the level of awareness and knowledge of glaucoma and the proportion of people with glaucoma among participants of a glaucoma urban outreach exercise. To the best of our knowledge, this is the first report of such a study carried out on an urban outreach population in Southeast Nigeria. Igbos, the predominant ethnic group in this region, as well as urban dwellers, have the highest prevalence of glaucoma in the country.4 The findings provide important evidence for efficient planning of comprehensive glaucoma management in the region. Materials and methods The study was a cross-sectional survey of participants of a 1-day eye outreach programme in March 2015 in an urban area in Enugu State, Southeast Nigeria. The programme was widely publicised through the mass media (radio and television), church bulletins and announcements, and posters and announcements at eye clinics of the two tertiary-level eye hospitals in the study area. The programme took place in a field beside the premises of a popular church in the study area, whereas ocular examinations were carried out in an adjacent private eye hospital.

n), church bulletins and announcements, and posters and announcements at eye clinics of the two tertiary-level eye hospitals in the study area. The programme took place in a field beside the premises of a popular church in the study area, whereas ocular examinations were carried out in an adjacent private eye hospital. The study participants comprised all consenting participants of the programme aged 18 years or older. This age cut-off was chosen to ensure that participants would be able to answer the questions correctly. Trained facilitators collected data on the first aspect of a two-part structured questionnaire, whereas the second part was completed by ophthalmologists after ocular examination. The primary outcomes of interest were awareness, knowledge and the proportion of people with glaucoma among participants. Awareness of glaucoma was assessed based on whether participants had heard about glaucoma prior to the programme. Knowledge was assessed using eight questions. A score was assigned to each question correctly answered. The level of knowledge was classified as ‘good’ (five or more correct answers) or ‘poor’ (less than five correct answers). Participants then underwent visual assessments (visual acuity (VA), intraocular pressure measurement and funduscopy). Visual fields could not be assessed due to the logistic challenges of an outreach programme. All identified glaucoma cases and suspects were appropriately counselled and referred for further evaluation and management.

en underwent visual assessments (visual acuity (VA), intraocular pressure measurement and funduscopy). Visual fields could not be assessed due to the logistic challenges of an outreach programme. All identified glaucoma cases and suspects were appropriately counselled and referred for further evaluation and management. The following criteria, adapted from the International Society of Geographical and Epidemiological Ophthalmology definitions for glaucoma in prevalence surveys20 and the normative data for Nigeria,21 were used to determine the glaucoma status of participants:Glaucoma cases Glaucoma category 2 (structural damage, visual field not done): vertical cup–disc ratio (VCDR) >0.75 (99.5th percentile) or VCDR asymmetry >0.2 (99.5th percentile) Glaucoma category 3 (no view of fundus and no visual field): VA <3/60+intraocular pressure (IOP) >28 mm Hg (99.5th percentile) OR VA <3/60+evidence of glaucoma filtering surgery or medical records showing visual glaucomatous morbidityGlaucoma category 1 (structural and typical visual field defect) was not included as visual fields were not assessed. Glaucoma suspects Disc suspects: VCDR >0.7 (97.5th percentile) or VCDR asymmetry >0.1 (97.5th percentile) IOP suspect: IOP >20 mm Hg (97.5th percentile) IOP + disc suspect: meets both disc suspect and IOP suspect criteria

Glaucoma category 3 (no view of fundus and no visual field): VA <3/60+intraocular pressure (IOP) >28 mm Hg (99.5th percentile) OR VA <3/60+evidence of glaucoma filtering surgery or medical records showing visual glaucomatous morbidityGlaucoma category 1 (structural and typical visual field defect) was not included as visual fields were not assessed. Glaucoma suspects Disc suspects: VCDR >0.7 (97.5th percentile) or VCDR asymmetry >0.1 (97.5th percentile) IOP suspect: IOP >20 mm Hg (97.5th percentile) IOP + disc suspect: meets both disc suspect and IOP suspect criteria Data were analysed using IBM Statistical Package for Social Sciences (SPSS) version V.20. For ease of analysis, the ages of participants were presented and analysed in intervals of 10, except for people aged 18–29 years and >70 years. Means and SD were calculated for quantitative variables, whereas frequencies and percentages were calculated for qualitative variables. The relationships between demographic and other features of participants, and awareness, knowledge and proportion of people with glaucoma were explored using Pearson's Chi-squareχ2 test for categorical and dichotomous variables. p Values <0.05 were considered statistically significant.

ulated for qualitative variables. The relationships between demographic and other features of participants, and awareness, knowledge and proportion of people with glaucoma were explored using Pearson's Chi-squareχ2 test for categorical and dichotomous variables. p Values <0.05 were considered statistically significant. The study abided by the guidelines of the Declaration of Helsinki and ethical approval was granted by the Ethics Review Committee of the Enugu State University of Science and Technology Teaching Hospital, Parklane, Enugu, Nigeria. Informed consent was obtained from all participants prior to their inclusion in the study and was not a prerequisite to being part of the outreach programme. Results Of the 513 participants who attended the programme, 478 (93.2%) people were aged 18 years and older and consented to participate in the study. All analyses were based on these 478 eligible participants except analyses of proportion of people with glaucoma that involved only people who were completely examined (n=469). Mean age was 50 (SD± 15.4) years with a range of 18–92 years. Participants were predominantly of the Igbo ethnic group (n=474/478, 99.2%). There were slightly more men than women (1.2:1). Two hundred and eighty-nine (60.8%) people came because of an eye problem, whereas the rest came for routine eye checks. The mass media was the most common source of information about the outreach (n=361/478, 75.5%), followed by information from family and friends (n=62/478, 13%). The sociodemographic characteristics of participants are shown in table 1.

people came because of an eye problem, whereas the rest came for routine eye checks. The mass media was the most common source of information about the outreach (n=361/478, 75.5%), followed by information from family and friends (n=62/478, 13%). The sociodemographic characteristics of participants are shown in table 1. Table 1 Sociodemographic characteristics of eligible participants at the eye outreach programme S. no Sociodemographic feature Frequency Percentage (95% CI) 1 Sex Male 264 55.2 (50.8 to 59.8) Female 214 44.8 (40.2 to 49.2) 2 Age (years) 18–29 55 11.5 (8.6 to 14.6) 30–39 60 12.6 (9.6 to 15.9) 40–49 107 22.4 (19.0 to 26.6) 50–59 118 24.7 (21.1 to 28.2) 60–69 86 18.0 (14.4 to 21.1) >70 52 10.9 (8.2 to 13.8) 3 Marital status Married 364 76.2 (72.4 to 79.7) Single 87 18.2 (15.3 to 21.5) Widowed 26 5.4 (3.3 to 7.5) Divorced/separated 1 0.2 (0.0 to 0.6) 4 Occupation Professional 26 5.4 (3.6 to 7.5) Business/self-employed 124 25.9 (22.4 to 29.9) Student 44 9.2 (6.7 to 11.7) Civil servant 95 19.9 (16.3 to 23.6) Pensioner 61 12.8 (10.0 to 15.7) Unemployed 35 7.3 (5.2 to 9.6) Farmer/labourer 93 19.5 (15.9 to 22.8) 5 Highest educational level attained None 43 9.0 (6.7 to 11.7) Primary 111 23.2 (19.5 to 27.4) Secondary 129 27.0 (23.0 to 30.8) Tertiary 181 37.9 (33.5 to 42.9) Higher (MSc, PhD, etc) 14 2.9 (1.5 to 4.4) Business/self-employed people were the most common occupational group (n=124/478, 25.9%).

er 93 19.5 (15.9 to 22.8) 5 Highest educational level attained None 43 9.0 (6.7 to 11.7) Primary 111 23.2 (19.5 to 27.4) Secondary 129 27.0 (23.0 to 30.8) Tertiary 181 37.9 (33.5 to 42.9) Higher (MSc, PhD, etc) 14 2.9 (1.5 to 4.4) Business/self-employed people were the most common occupational group (n=124/478, 25.9%). Two-hundred and twenty-two (46.4%) participants had their eyes checked for the first time at the outreach programme. Among those who had had a previous eye check (n=256/478, 53.6%), 140 (54.7%) had had their eyes checked within the preceding 1 year. Two hundred and eighty-nine (60.5%) people came because they had eye complaints, and 365 (77%) said they would still have come even if they had to pay to have their eyes checked. Awareness about glaucoma Two-hundred and ninety-three (61.3%, 95% CI 56.3% to 65.5%) participants had heard of glaucoma prior to the outreach. The mass media was the most common source of information (n=169, 57.7%), followed by information from a health worker (n=77, 26.3%) and from family members or friends (n=36, 12.4%). Only 47 (9.8%, 95% CI 7.3% to 12.6%) people were sure of a positive family history of glaucoma, of which 39 (83%) were first-degree relatives (28 parents, 9 siblings and 2 children). The relationship between awareness and demographic and other features of participants are shown in table 2. Table 2 Determinants of awareness of glaucoma among the study participants

Only 47 (9.8%, 95% CI 7.3% to 12.6%) people were sure of a positive family history of glaucoma, of which 39 (83%) were first-degree relatives (28 parents, 9 siblings and 2 children). The relationship between awareness and demographic and other features of participants are shown in table 2. Table 2 Determinants of awareness of glaucoma among the study participants S. no Feature Awareness of glaucoma Aware N (%) Not aware N (%) Total N (%) p Value 1 Age 18–29 37 (67.3) 18 (32.7) 55 (100) 0.314 30–39 35 (58.3) 25 (41.7) 60 (100) 40–49 60 (56.1) 47 (43.9) 107 (100) 50–59 79 (66.4) 40 (33.6) 119 (100) 60–69 55 (64.7) 30 (35.3) 85 (100) >70 27 (51.9) 25 (48.1) 13 (100) 2 Sex Male 165 (62.5) 99 (37.5) 264 (100) 0.549 Female 128 (59.8) 86 (40.2) 214 (100) 3 Marital status Married 220 (60.4) 144 (39.6) 364 (100) 0.786 Single 55 (63.2) 32 (36.8) 87 (100) Widowed 17 (65.4) 9 (34.6) 26 (100) Divorced/separated 1 (100) 0 (0) 1 (100) 4 Occupation Professional 17 (65.4) 9 (34.6) 26 (100) 0.360 Business/self-employed 72 (58.1) 52 (41.9) 124 (100) Student 27 (61.4) 17 (38.6) 44 (100) Civil servant 65 (68.4) 30 (31.6) 95 (100) Pensioner 42 (68.9) 19 (31.1) 61 (100) Unemployed 19 (54.3) 16 (45.7) 35 (100) Farmer/labourer 51 (54.8) 42 (45.2) 93 (100) 5 Highest educational attainment None 15 (34.9) 28 (65.1) 43 (100) <0.001 Primary 57 (51.4) 54 (48.6) 111 (100) Secondary 81 (62.8) 48 (37.2) 129 (100) Tertiary 130 (71.8) 51 (28.2) 181 (100) Higher (MSc, PhD, etc) 10 (71.4) 4 (28.6) 14 (100) 6 First eye check Yes 105 (47.3) 117 (52.7) 222 (100) <0.001 No 188 (72.3) 68 (26.6) 256 (100) 7 Family history <0.001 Yes 47 (100) 0 (0) 47 (100) No/don't know 185 (42.9) 246 (57.1) 431 (100) Awareness of glaucoma was strongly associated with higher levels of education, awareness of a positive family history of glaucoma and having had an eye check prior to the outreach programme. People who had primary education or greater were more likely to be aware of glaucoma than people who had no formal education (Xχ2 (4, n=478)=26.4, p<0.001). All the patients who were aware of a positive family history of glaucoma (n=47) had heard of glaucoma before the outreach (χ2 (1, n=478)=32.912, p<0.001). Awareness of glaucoma was higher in people who had had a previous eye check (n=188/256, 73.4%) than in people who were having their eyes checked for the first time (n=105/222, 47.3%) (χ2 (1, n=478)=34.245, p<0.001).

ly history of glaucoma (n=47) had heard of glaucoma before the outreach (χ2 (1, n=478)=32.912, p<0.001). Awareness of glaucoma was higher in people who had had a previous eye check (n=188/256, 73.4%) than in people who were having their eyes checked for the first time (n=105/222, 47.3%) (χ2 (1, n=478)=34.245, p<0.001). The mass media was the main source of information for both participants who had had previous eye checks (n=100/188, 53.2%) and those having their eyes checked for the first time (n=69 of 105, 65.7%) (χ2 (4, n=293)=10.231, p=0.037). However, 31.9% (60) of people with previous checks had heard through a health worker/hospital setting compared with 16.2% (17) of first timers who heard through the same source. Knowledge about glaucoma Participants who were aware of glaucoma (n=293/478, 61.3%) were questioned further to assess their knowledge about glaucoma using eight questions (figure 1). Questions asked bordered on effects, perceptions and management of glaucoma. Figure 1 Responses to questions on knowledge about glaucoma. Most people knew that glaucoma causes blindness (n=245/293, 83.6%) and that it can be managed by means of drugs (n=198/293, 67.6%) and surgery (n=187/293, 63.8%). One hundred and thirty-four (45.7%) people were unsure or said it was a spiritual problem or curse. More than a third (107/293, 36.5%) did not know that vision loss due to glaucoma is irreversible.

indness (n=245/293, 83.6%) and that it can be managed by means of drugs (n=198/293, 67.6%) and surgery (n=187/293, 63.8%). One hundred and thirty-four (45.7%) people were unsure or said it was a spiritual problem or curse. More than a third (107/293, 36.5%) did not know that vision loss due to glaucoma is irreversible. Knowledge was assessed using a knowledge score as described in the methodology. A ‘good’ knowledge score was recorded for 176 of the 293 people (60.1%, 95% CI 50.0% to 65.9%) that were aware of glaucoma. This accounts for 36.8% of the total study participants. The relationship between knowledge score of participants and their demographic and other features are shown in table 3. Table 3 Determinants of knowledge of glaucoma among the study participants

Knowledge was assessed using a knowledge score as described in the methodology. A ‘good’ knowledge score was recorded for 176 of the 293 people (60.1%, 95% CI 50.0% to 65.9%) that were aware of glaucoma. This accounts for 36.8% of the total study participants. The relationship between knowledge score of participants and their demographic and other features are shown in table 3. Table 3 Determinants of knowledge of glaucoma among the study participants S. no Feature Poor (%) Good (%) p Value 1 Age 0.202 18–29 11 (29.7) 26 (70.3) 37 (100) 30–39 9 (25.7) 26 (74.3) 35 (100) 40–49 28 (46.7) 32 (53.3) 60 (100) 50–59 29 (37.2) 50 (63.3) 79 (100) 60–69 26 (46.4) 29 (52.7) 55 (100) >70 14 (51.9) 13 (48.1) 27 (100) 2 Sex 0.978 Male 66 (40) 99 (60) 165 (100) Female 51 (39.8) 77 (60.2) 128 (100) 3 Marital status 0.110 Married 93 (42.3) 127 (57.7) 220 (100) Single 15 (27.3) 40 (72.7) 55 (100) Widowed 9 (52.9) 8 (47.1) 17 (100) Divorced/separated 0 (0) 1 (100) 1 (100) 4 Occupation Professional 7 (41.2) 10 (58.8) 17 (100) 0.008 Business/self-employed 28 (38.9) 44 (61.1) 72 (100) Student 5 (18.5) 22 (81.5) 27 (100) Civil servant 22 (33.8) 43 (66.2) 65 (100) Pensioner 16 (38.1) 26 (61.9) 42 (100) Unemployed 7 (36.8) 12 (63.2) 19 (100) Farmer/labourer 32 (62.7) 19 (37.3) 51 (100) 5 Highest educational attainment p<0.001 None 11 (73.3) 4 (26.7) 15 (100) Primary 31 (54.4) 26 (45.6) 57 (100) Secondary 37 (45.7) 44 (54.3) 81 (100) Tertiary 35 (26.9) 95 (73.1) 130 (100) Higher (MSc, PhD, etc) 3 (30) 7 (70) 10 (100) 6 First eye check 0.606 No 73 (38.8) 115 (61.2) 188 (100) Yes 44 (41.9) 61 (58.1) 105 (100) 7 Family history 0.566 Yes 100 (40.7) 146 (59.3) 246 (100) No 17 (36.2) 30 (63.8) 47 (100) 8 Do you have glaucoma 0.796 No or don't know 106 (39.7) 161 (60.3) 267 (100) Yes 11 (42.3) 15 (57.7) 26 (100) 9 Source of information about glaucoma 0.335 Mass media 70 (41.4) 99 (58.6) 169 (100) Social media 1 (11.1) 8 (88.9) 9 (100) Family and friends 17 (47.2) 19 (52.8) 36 (100) Health workers 28 (36.4) 49 (63.6) 77 (100) Books 1 (50) 1 (50) 2 (100) Occupation and level of education were significantly associated with knowledge of glaucoma. There was a strong association between level of education and knowledge of glaucoma (χ2 (4, n=293)=22.640, p<0.001). The higher the level of education attained, the greater the percentage of people with ‘good’ knowledge, rising from 26.7% (n=4/15) in people with no formal education to >70% in people with tertiary education and higher. About three-quarters (n=11/15, 73.3%) of the people who had no formal education had poor knowledge of glaucoma.

the level of education attained, the greater the percentage of people with ‘good’ knowledge, rising from 26.7% (n=4/15) in people with no formal education to >70% in people with tertiary education and higher. About three-quarters (n=11/15, 73.3%) of the people who had no formal education had poor knowledge of glaucoma. A similar association was also noted for occupation and knowledge about glaucoma (χ2 (6, n=293)=17.409, p=0.008). The ‘farmer/labourer’ occupation group had by far the highest proportion of people with poor knowledge of glaucoma (n=32/51, 62.7%). They also had the highest proportion of people who had no formal education (20/43, 46.5%) or only primary education (42/111, 37.8%), and this was statistically significant (χ2 (24, n=478)=196.631, p<0.001). Proportion of people with glaucoma Of the 478 eligible participants, 469 (98.1%) underwent ophthalmic examination, whereas nine people left before they could be examined and were therefore excluded from the analysis of this section. The proportion of people with glaucoma in the study population was 14.5% (68/469, 95% CI 11.7% to 17.7%), affecting 38 (14.7%) men and 30 women (14.3%). The proportion increased significantly with increasing age from about 1.9% in people younger than 30 years to 25% in people 70 years and older (χ2 (5, n=469)=26.685, p<0.001). Four hundred and forty-three (94.5%) of the 469 people examined said they did not have glaucoma, but 11.1% (49) of them had glaucoma on examination.

Proportion of people with glaucoma Of the 478 eligible participants, 469 (98.1%) underwent ophthalmic examination, whereas nine people left before they could be examined and were therefore excluded from the analysis of this section. The proportion of people with glaucoma in the study population was 14.5% (68/469, 95% CI 11.7% to 17.7%), affecting 38 (14.7%) men and 30 women (14.3%). The proportion increased significantly with increasing age from about 1.9% in people younger than 30 years to 25% in people 70 years and older (χ2 (5, n=469)=26.685, p<0.001). Four hundred and forty-three (94.5%) of the 469 people examined said they did not have glaucoma, but 11.1% (49) of them had glaucoma on examination. About a fifth (13/68, 19.1%) of glaucoma cases detected were blind with presenting VA <3/60 in the better eye. The diagnosis of glaucoma was not significantly associated with family history of the condition.Figure 2 shows the different categories of participants with respect to their glaucoma status. Figure 2 Categories of participants with respect to their glaucoma status. Discussion Glaucoma is the leading cause of irreversible blindness in Nigeria,22 and poor awareness and late presentation are often major impediments to adequate management and prevention of blindness from the condition.

About a fifth (13/68, 19.1%) of glaucoma cases detected were blind with presenting VA <3/60 in the better eye. The diagnosis of glaucoma was not significantly associated with family history of the condition.Figure 2 shows the different categories of participants with respect to their glaucoma status. Figure 2 Categories of participants with respect to their glaucoma status. Discussion Glaucoma is the leading cause of irreversible blindness in Nigeria,22 and poor awareness and late presentation are often major impediments to adequate management and prevention of blindness from the condition. There was a generally high level of awareness (n=293/478, 61.3%) of glaucoma among participants in this study compared with other outreach-based studies. In Southwest Nigeria14 and Ethiopia,13 the levels of awareness were lower at 15.8% (n=41/259) and 2.4% (n=8/340), respectively. These studies were however carried out on rural populations. The level of awareness found in this study is more similar to levels reported in studies carried out among hospital and university staff in Nigeria of 68.6%23 and 46.8%,24 respectively. In a population-based study in urban Chennai, South India, awareness level was only 13.5% (n=1926).11

arried out on rural populations. The level of awareness found in this study is more similar to levels reported in studies carried out among hospital and university staff in Nigeria of 68.6%23 and 46.8%,24 respectively. In a population-based study in urban Chennai, South India, awareness level was only 13.5% (n=1926).11 High awareness of glaucoma in this population may be attributed to accessibility to mass media, which was the major source of information about glaucoma in this study (n=168/291, 57.7%). A lot of awareness about glaucoma had been created in the study town in the preceding 5 years through the mass media especially the radio, with accompanying free eye screening exercises (such as that for this study) about three times a year. Although there is no published baseline study to which the results of this study population can be compared, results of studies in Ethiopia25 and of an Indian population in London26 indicate that awareness campaigns through the mass media can significantly increase awareness levels of glaucoma among target populations. Though ‘family members and friends’ were not the main source of information about glaucoma in this study, being aware of a family history of glaucoma was strongly associated with glaucoma awareness (p<0.001). A positive family history is one of the risk factors of glaucoma, and this finding highlights the invaluable role of glaucoma patients as agents of awareness creation among their families and other contacts.

is study, being aware of a family history of glaucoma was strongly associated with glaucoma awareness (p<0.001). A positive family history is one of the risk factors of glaucoma, and this finding highlights the invaluable role of glaucoma patients as agents of awareness creation among their families and other contacts. People who had had some formal education as opposed to no formal education were also more likely to be aware of glaucoma in this, as well as other studies.11 13 27 Awareness of the existence of a chronic, symptomless disease may be a key step towards control of that disease, but a good knowledge of the condition is pertinent for good compliance with the management plan. Knowledge about glaucoma in this study (36.8%) was better than findings in other settings.11 13 27 Higher glaucoma knowledge scores were observed among participants with higher educational attainment. This group of people may have better access to relevant information from the mass media (the most common source in this study) and other sources than their counterparts. The association between occupation and knowledge about glaucoma in this study may likely be due to the significant association between educational attainment and occupation also noted.

er access to relevant information from the mass media (the most common source in this study) and other sources than their counterparts. The association between occupation and knowledge about glaucoma in this study may likely be due to the significant association between educational attainment and occupation also noted. Furthermore, although awareness of family history and previous eye check were significantly associated with awareness in this study, this was not the case with knowledge. This underscores the importance of tailoring the content of awareness efforts to suit the needs of the intended audience. Good knowledge of glaucoma may enhance health-seeking behaviour among citizens and enable both patients and their family members cope with the demands of the life-long management of the condition. The proportion of people with glaucoma in this study population (14.5%) was higher than that in other studies in Nigeria4 28 and its environs,29 though these were prevalence studies. Participants were also mostly of the Igbo ethnic group (99.2%), which had the highest prevalence of glaucoma from the Nigerian blindness and visual impairment survey.4 The greater prevalence with increasing age noted in other studies4 28 29 was also noted in this study, but there was no significant difference in terms of sex. About one in five people (19.1%) was blind with presenting vision <3/60 in the better eye, which is comparable with results of the national survey.4

nt survey.4 The greater prevalence with increasing age noted in other studies4 28 29 was also noted in this study, but there was no significant difference in terms of sex. About one in five people (19.1%) was blind with presenting vision <3/60 in the better eye, which is comparable with results of the national survey.4 The high number of glaucoma suspects (n=89/469, 19%) and cases (n=68/469, 14.5%) identified, as well as the number of cases blind thereof (n=13/68, 19.1%), makes a strong case for the usefulness of screening programmes in early detection of glaucoma in high-risk populations such as Nigeria until better screening strategies are developed. One limitation of this study is that it was carried out on an outreach population that may undermine the generalisability of its results. Nevertheless, participants were from a varied mix of occupational, educational and demographic backgrounds, and some key findings were comparable with results of the national survey. The findings therefore provide important evidence for planning of public health measures for glaucoma control in terms of content, target population and medium. Contributors: All the authors have read and approved the final draft of the manuscript, and satisfied the conditions for authorship. All the authors agree that this work represents their honest contribution towards the writing of this manuscript. Competing interests: None declared. Provenance and peer review: Not commissioned; externally peer reviewed.

Key messages There has been a gradual shift and push in using disposable tonometer prism heads throughout eye units in the country, primarily due to concerns of cross infection and contamination. There is limited clinical evidence supporting transmission of infectious diseases by non-disposable tonometer prisms if appropriate precautions are used in addition to the significant cost savings it offers to departments using them. We believe that the perceived cost-effectiveness and safety of disposable prism is probably overestimated and that switching to non-disposable ones will result in significant cost savings to the National Health Service without compromising patient safety. Purpose Applanation tonometry is standard practice in UK eye units for measuring intraocular pressure (IOP), and the Goldmann tonometer has been widely regarded as the gold standard. These tonometers require the placement of a prism and its applanation onto the corneal surface to obtain a reading for IOP. Up to recently, non-disposable prisms were the preferred choice in the majority of eye departments in the UK.1 2 However, there have been reports of outbreaks of epidemic viral ocular infections in ophthalmic clinics worldwide.3–5 Epidemiological and clinical studies of epidemic keratoconjunctivitis in eye clinics have identified possible contamination with tonometer tips, eye drop bottles and contact with infected healthcare workers as vehicles of transmission (particularly of adenoviral serotypes).6 7

ctions in ophthalmic clinics worldwide.3–5 Epidemiological and clinical studies of epidemic keratoconjunctivitis in eye clinics have identified possible contamination with tonometer tips, eye drop bottles and contact with infected healthcare workers as vehicles of transmission (particularly of adenoviral serotypes).6 7 Insufficient disinfection of Goldmann tonometer prism heads has been postulated as a possible significant cause of transmission. In addition, improper sterilisation techniques can also increase the risk sustaining a corneal injury due to residual cleaning agent on re-usable prism heads. A recent publication by Haag-Streit (a manufacturer of Goldmann tonometers) found that the mean age of non-disposable prisms in the UK was 9.8 years, with 56% of these being damaged or scratched.8 These factors, along with the recommendations by the Medical Device Agency advice that ‘components of ophthalmic devices that touch the surface of the eye should be restricted to single patient use where practicable and where this does not compromise clinical outcome’, led to the introduction of disposable (single-use) prism heads, which were felt to be safer, as effective and equally reliable to the re-usable Goldman prisms.4 Interestingly, Haag-Streit is also the manufacturer of Tonosafe disposable prisms, one of the most popular disposable prism brands in the UK.

linical outcome’, led to the introduction of disposable (single-use) prism heads, which were felt to be safer, as effective and equally reliable to the re-usable Goldman prisms.4 Interestingly, Haag-Streit is also the manufacturer of Tonosafe disposable prisms, one of the most popular disposable prism brands in the UK. The notion of ‘perceived safety’ and evidence showing comparable accuracy and cost between both prism heads have led to a gradual but definite shift to the use of disposable prisms worldwide.3 4 8 9 The primary aim of this survey was to compare the use of disposable versus non-disposable prisms and to determine the cost–benefits of each in current UK practice.