Prevalence and associated factors of glaucoma in the Russian Ural Eye and Medical Study

To assess the prevalence and associated factors of glaucoma in a Russian population. The population-based Ural Eye and Medical Study included 5899 (mean age 59.0 ± 10.7 years; range 40–94 years). Glaucomatous optic neuropathy was diagnosed using International Society of Geographical and Epidemiological Ophthalmology (ISGEO) criteria. Among 5545 participants with assessable optic disc photographs, 246 individuals [4.4%; 95% confidence interval (CI) 3.9, 5.0] had glaucoma, with open-angle glaucoma (OAG) in 177 individuals (3.2%; 95% CI 2.7, 3.7) and angle-closure glaucoma (ACG) in 69 individuals (1.2; 95% CI 1.0, 1.5), with IOP > 21 mmHg in 79 (32.1%) patients, and with 80 (32.5%) patients on glaucoma therapy. Glaucoma prevalence increased from 3/485 (0.6%; 95% CI 0.0, 1.3) in the age group of 40–45 years to 33/165 (20.0%; 95% CI 13.8, 26.2) in the group aged 80 + years. Higher OAG prevalence correlated with older age [odds ratio (OR) 1.07; 95% CI 1.04, 1.09; P < 0.001], longer axial length (OR 1.36; 95% CI 1.17, 1.58; P < 0.001), higher intraocular pressure (IOP) (OR 1.18; 95% CI 1.13, 1.23; P < 0.001), higher stage of lens pseudoexfoliation (OR 1.26; 95% CI 1.08, 1.47; P = 0.004) and lower diastolic blood pressure (OR 0.98; 95% CI 0.96, 0.99; P = 0.035). Higher ACG prevalence correlated with older age (OR 1.07; 95% CI 1.03, 1.11; P < 0.001), narrower anterior chamber angle (OR 0.81; 95% CI 0.77, 0.86; P < 0.001), and higher IOP (OR 1.30; 95% CI 1.23, 1.38; P < 0.001). Glaucoma caused moderate to severe vision impairment (MSVI) in 9 (4.9%; 95% CI 1.8, 8.1) out of 184 individuals with MSVI (OAG, n = 7; ACG, n = 2), and blindness in one (9.1%) of 11 blind individuals. In this population from Russia, two thirds of glaucoma patients were not on therapy, and in two thirds of the glaucoma patients IOP was ≤ 21 mmHg. Otherwise, glaucoma prevalence, OAG-to-ACG ratio, and glaucoma associations did not differ markedly from Caucasian and East Asian populations.

Scientific Reports | (2020) 10:20307 | https://doi.org/10.1038/s41598-020-77344-z www.nature.com/scientificreports/ the degrees of cortical cataract and subcapsular cataract were measured as the percentages area of opacity. The presence of cortical cataract and subcapsular cataract was defined by the presence of any cortical or subcapsular opacity, respectively. We additionally took photographs of the optic nerve head and macula (VISUCAM 500, Carl Zeiss Meditec AG, Jena, Germany) and spectral-domain optical coherence tomographic (OCT) images (RS-3000, NIDEK co., Ltd., Aichi Japan). The latter served for measurement of the peripapillary retinal nerve fiber layer thickness, neuroretinal rim width, and thickness of the retina. The degree of fundus tessellation was examined on the fundus photographs 15 . We defined age-related macular degeneration (AMD) as suggested by the recent Beckman Initiative for Macular Research Classification Committee 16 . For the definition of glaucoma we applied criteria recommended by the ISGEO (International Society of Geographical and Epidemiological Ophthalmology) 17 . The anterior segment images taken with the Pentacam camera were used to differentiate between open-angle glaucoma and angle-closure glaucoma. The anterior chamber angle was considered to be closed if the peripheral iris had direct contact with the peripheral cornea. We defined diabetes mellitus by a glucose concentration of ≥ 7.0 mmol/L or a self-reported history of a physician-based diagnosis of diabetes mellitus or a history of drug treatment for diabetes. Arterial hypertension was defined using the criteria defined by the American College of Cardiology/American Heart Association 18 . All examinations and the interview were conducted in the Ufa Eye Research Institute in Ufa.
Out of 7328 eligible individuals, the study population consisted of 5899 individuals (2580 (43.7%) men) with a mean age of 59.0 ± 10.7 years (range: 40-94 years) who had agreed to participate. The participation rate was 80.5%. The study population did not differ significantly in the gender and age distribution from the Russian population as explored in the census carried out in 2010 19 . Stratified by ethnicity, the study population included 1185 (20.1%) Russians, 2439 (41.3%) Tartars, 1061 (18.0%) Bashkirs, 587 (10%) Chuvash, 21 (0.4%) Mari, 104 (1.8%) individuals of other ethnicities, and 502 (8.5%) individuals did not indicate their ethnic background. The proportion of non-Russians on the total study population was higher than the proportion of non-Russians on the total population of Russia.
For the present study, we included all participants for whom information about the presence of glaucoma was available. Applying a software package (SPSS/Windows, 25.0, IBM-SPSS, Chicago, IL, USA), we first calculated mean values (and 95% confidence intervals (CI)) of the prevalence of glaucoma as a whole and separated into open-angle glaucoma and angle-closure glaucoma 4,5 . In a following binary univariate regression analyses, we tested relationships between the prevalence of the glaucomas and systemic and ocular parameters. We eventually performed a multivariable binary regression analysis to assess associations between the glaucoma prevalence and all those variables which were significantly (P ≤ 0.05) correlated with the glaucoma prevalence in the univariate analyses. Odds ratios (OR) and their 95% CI were determined.

Results
Information about the presence of glaucoma was available for 5545 (94.0%) individuals with a mean age of 58.5 ± 10.5 years (range 40-94 years) and a mean axial length of 23.3 ± 1.1 mm (range 19.78-32.87 mm) ( Table 1). The group of participants with information about glaucoma as compared with the group of individuals without assessment of the presence of glaucoma was significantly (P < 0.001) younger (58.5 ± 10.5 years versus 65.8 ± 11.5 years) and showed a significantly (P < 0.001) higher proportion of women (3152/5545 or 56.8% versus 167/354 or 47.2%), while both groups did not differ significantly (P = 0.45) in axial length (23.3 ± 1.1 mm versus 23.3 ± 1.3 mm).
All study participants had fundus photographs for the assessment of the optic nerve head, while about 20% of all eyes with an advanced stage of glaucomatous optic neuropathy, as indicated by a vertical cup/disc diameter ratio of ≥ 0.9 or an inter-eye asymmetry in the vertical cup/disc diameter ratio of ≥ 0.3, did not have reliable visual field tests. Measurements of the peripapillary retinal nerve fiber layer thickness were available for 5040  (Fig. 2). The prevalence of angle-closure glaucoma increased from 0/485 (0.0%) in the age group of 40-< 45 years to 13/873 (1.5%; 95% CI 0.7, 2.3) in the age group of 60-< 65 years, and to 10/165 (6.1%; 95% CI 2.4, 9.7) in the age group of 80 + years (Table 2; Fig. 3). Among the 246 patients with glaucoma, 80 (32.5%) patients were on glaucoma therapy. IOP was higher than 21 mmHg in 79 (32.1%; 95% CI 26.4, 38.2) patients.
The prevalence of open-angle glaucoma was not significantly associated with the systemic parameters of gender, ethnicity (Russian versus non-Russian, level of education, physical activity score, family status (married www.nature.com/scientificreports/ versus unmarried, religion (Muslim versus non-Muslim, body height, body weight, body mass index, waist-hip ratio, and history of angina pectoris, arthritis, bone fractures, cancer, cardiovascular diseases including stroke, dementia, history of diarrhea, heart attack, iron-deficiency anemia, episodes with low blood pressure and hospitalization, osteoarthritis, injuries other than bone fracture, backache, headache, neck pain, thoracic spine pain, skin disease, steroid (cortisone) therapy, thyreopathy or the history of episodes of unconsciousness. It was neither associated with data on menopause (age of any last bleeding, age of last regular bleeding), serum concentrations of alanine aminotransferase, aspartate aminotransferase, bilirubin, high-density lipoproteins, low-density lipoproteins, triglycerides, cholesterol, rheumatoid factor, erythrocyte sedimentation rate, glucose, creatinine, urea, residual nitrogen, and hemoglobin, prothrombin time, the erythrocyte and leucocyte count, the intake of blood lipid lowering medication, the prevalences of diabetes mellitus, anemia and arterial hypertension, the stage of arterial hypertension, State-Trait Anxiety Inventory (STAI) score, diet parameters such number of meals taken, number of days with fruit intake or vegetable intake, type of oil used for cooking, amount of food contained whole grain, amount of self-reported salt intake, and the grade of processing of meat (weak/medium/well done), systolic, diastolic and mean blood pressure, ankle-brachial index, right side, prevalence of arterial hypertension, current smoker status, daily smoking, smoked package years, any alcohol consumed, hearing loss total score, depression score, and manual dynamometry. Nor was the prevalence of open-angle glaucoma associated with the   In the multivariable analysis, we dropped due to collinearity the parameters of prevalence of pseudoexfoliation (variance inflation factor (VIF): 12.5), anterior chamber angle and volume (VIF: 4.3), anterior chamber depth (VIF: 2.1), and prevalence of myopic maculopathy (VIF: 2.1). Due to a lack of statistical significance, we dropped the parameters of BCVA (P = 0.55), refractive error (P = 0.26), region of habitation (P = 0.16), stages of myopic maculopathy (P = 0.08). In the final model, a higher prevalence of open-angle glaucoma was associated with older age (OR 1.07; 95% CI 1.04, 1.09; P < 0.001), longer axial length (OR 1.36; 95% CI 1.17, 1.58; P < 0.001), higher IOP (OR 1.18; 95% CI 1.13, 1.23; P < 0.001), stage of lens pseudoexfoliation (OR 1.26; 95% CI 1.08, 1.47; P = 0.004), lower diastolic blood pressure (OR 0.98; 95% CI 0.96, 0.99; P = 0.035), and thinner peripapillary retinal nerve fiber layer thickness (OR 0.97; 95% CI 0.96, 0.98; P < 0.001).
The prevalence of angle-closure glaucoma was not significantly associated with the systemic parameters of gender, urban region of habitation, level of education, physical activity score, family status (married versus unmarried), religion (Muslim versus non-Muslim), body height, body weight, bod mass index, waist-hip ratio, history of angina pectoris, arthritis, cancer, cardiovascular diseases including stroke, dementia, diarrhea, heart attack, previous falls, iron-deficiency anemia, episodes with low blood pressure and hospitalization, osteoarthritis, injuries other than bone fractures, backache, neck pain, thoracic spine pain, skin disease, thyroid disorder, episodes of unconsciousness and menopause (age of any last bleeding; age of last regular bleeding). It was neither associated with the serum concentrations of alanine aminotransferase, aspartate aminotransferase, bilirubin, high-density lipoproteins, low-density lipoproteins, triglycerides, cholesterol, rheumatoid factor, glucose, www.nature.com/scientificreports/ creatinine, urea, residual nitrogen and hemoglobin, INR (international normalization ratio), the erythrocyte count, leucocyte count, intake of blood lipid lowering medication, the prevalences of diabetes mellitus, anemia and arterial hypertension, the stage of arterial hypertension, State-Trait Anxiety Inventory (STAI) score, diet parameters such as number of meals taken, number of days with fruit intake or vegetable intake, type of oil used for cooking, amount of food contained whole grain, amount of self-reported salt intake, and the grade of processing of meat (weak/medium/well done), systolic, diastolic and mean blood pressure, the ankle-brachial index, prevalence of arterial hypertension, current smoker, daily smoking, smoked package years, any alcohol consumed, hearing loss total score, depression score, and manual dynamometry. Nor was the prevalence of angle-closure glaucoma associated with the ocular parameters of refractive error, corneal refractive power, retinal thickness (total) in the fovea, prevalence and grade of pseudoexfoliation of the lens, prevalence and stage of nuclear cataract, cortical cataract and subcapsular cataract, prevalence and stage of diabetic retinopathy, prevalence and stage of myopic maculopathy, and lower BCVA.

Discussion
In our population-based study population, the prevalence of glaucoma was 4.4%, with open-angle glaucoma (3.2%) having a higher prevalence than angle-closure glaucoma (1.2%). Glaucoma prevalence increased from 0.6% in the age group of 40 to < 45 years to 20.0% in the age group of 80 + years. Factors associated with a higher prevalence of open-angle glaucoma were older age, longer axial length, higher IOP, higher stage of lens pseudoexfoliation, lower diastolic blood pressure and thinner peripapillary retinal nerve fiber layer thickness, or higher estimated trans-lamina cribrosa pressure difference, if age, IOP and blood pressure were dropped. Higher ACG prevalence correlated with older age, narrower anterior chamber angle, higher IOP and thinner RNFL. Among the 246 patients with glaucoma, 80 (32.5%) patients were on glaucoma therapy. Glaucoma caused MSVI in nine (4.9%) out of 184 individuals with MSVI (open-angle glaucoma, n = 7; angle-closure glaucoma, n = 2), and blindness in one individual (9.1%) out of 11 blind individuals. The single IOP measurement was ≤ 21 mmHg in 167 (67.9%) individuals with glaucoma. The findings on the prevalence of glaucoma of 4.4% and of open-angle glaucoma of 3.2% in our study population agree with the observations made in some, and differ from the observations made in other, previous investigations. In the preceding studies, the prevalence of open-angle glaucoma varied between 0.5% in a rural population in Mongolia, 1.6% in a Singaporean urban population, 1.6% in an Indian urban population, and 1.7%, 2.1%, 2.7%, 3.0%, 3.0%, 3.8%, 5.7%, and 7.0% in populations from Melbourne, Australia, United States, South Africa, Sydney, East Africa, South China (Liwan District, Guangzhou), Central Sweden, and Barbados, to mention only few examples  . The glaucoma prevalence of 4.4% and prevalence of open-angle glaucoma of 3.2% in the Ural Eye and Medical Study was similar to the glaucoma prevalence of 3.7% reported in the Liwan Eye Study in Guangzhou/South China and of 3.8% as found in the Beijing Eye Study 22,46 . Similar glaucoma prevalences were also reported from Japan, India, and Singapore, while the glaucoma prevalence was higher in Sub-Saharan African countries. A meta-analysis analyzing data from 50 population-based studies found a global prevalence of glaucoma of 3.54% (95% credible intervals (CrI): 2.09-5.82) for the population aged 40-80 years 47 . In that analysis, the prevalence of open-angle glaucoma was the highest in Africa (4.20%; 95% CrI: 2.08-7.35), and the prevalence of angle-closure glaucoma was the highest in Asia (1.09%; 95% CrI: 0.43-2.32) 42 . Reasons for the discrepancies between the studies in the prevalence of the glaucomas may be ethnic differences, with a  www.nature.com/scientificreports/ relatively high glaucoma prevalence in Sub-Saharan Africa, and differences in the examination technique and glaucoma definitions 17 . The associations of the prevalence of the glaucomas with other ocular and general parameters as found in our study agrees with most of the previous investigations on different study populations  . It also includes the correlation between a higher prevalence of open-angle glaucoma and a longer axial length 46,48 . In agreement with previous studies and partially in contrast to some hospital-based studies, the prevalence of open-angle glaucoma was not related with central corneal thickness in the multivariable analysis 49 .
In our study population, the ratio of the prevalence of open-angle glaucoma to primary angle-closure glaucoma of 3.2-1.2% or 2.7 to 1 was similar to the findings in other population-based investigations such as the Beijing Eye Study with a ratio of 2.6-1 46 . In other studies different ratios were found such as in the South China Liwan Eye Study a ratio of 2.1-1.5% (or 1.4:1), the Japanese Tajimi Study with a ratio of 3.9-0.6% (or 6.5:1), the Singapore Malay Eye Study with a ratio of 2.5% to 0.1% (or 25:1),and the Tanjong Pagar study with a ratio of 2.4% to 0.8% (3:1) 21,22,45,50,51 . Reasons for differences between the studies and study populations may be differences in the definition of the glaucomas, ethnically associated differences in the anterior chamber anatomy, and differences in the prevalence of pseudophakia/aphakia. Glaucoma caused MSVI in 4.9% of all individuals with MSVI and blindness in 9% of the blind individuals. These figures were comparable with the figures found in other countries. In the meta-analysis on the worldwide prevalence of MSVI and blindness by the Flaxman and colleagues, 8.49% (80% UI: 2.99-15.66) of all cases with MSVI and 2.5% (80% UI: 0.62-4.03) of all cases with blindness were caused by glaucoma 1 . Interestingly, seven out of the 9 individuals with glaucoma-related MSVI had open-angle glaucoma, while other studies reported that the prevalence of glaucoma-related blindness was higher in individuals with angle-closure glaucoma than in individuals with open-angle glaucoma 46 . The only participants with glaucoma-related blindness in our study population had angle-closure glaucoma.
Among the 246 patients with glaucoma, the single IOP measurement was ≤ 21 mmHg in 167 (67.9%) individuals. This figure was similar to a fraction of 70% of the glaucomatous eyes in the Beijing Eye Study with glaucomatous optic neuropathy and an IOP of ≤ 21 mmHg 52 . It shows the limited value of a single IOP measurement Table 6. Associations (binary univariate analysis) between the prevalence of angle-closure glaucoma and ocular parameters after adjusting for age in the Ural Eye and Medical Study. www.nature.com/scientificreports/ as screening method to detect glaucoma. Simultaneously, with only one third of the glaucomatous patients being on glaucoma therapy, screening measures, in particular on risk populations, should be developed and applied to reduce the risk of glaucoma-related blindness.
There are limitations of our investigation. First, we differentiated between open-angle glaucoma and angleclosure glaucoma by anterior segment imaging and did not perform gonioscopy. Since a closed anterior chamber angle can appear upon imaging as a falsely open angle, the prevalence of angle-closure glaucoma might have been underestimated. Second, we measured the IOP by non-contact tonometry instead of applanation tonometry. Third, the group of individuals with information about the presence of glaucoma as compared with the group of individuals without assessment of the presence of glaucoma was significantly younger and showed a significantly higher proportion of women, while both groups did not differ significantly in axial length. The younger age in the individuals with glaucoma assessment may have led to an underestimation of the prevalence of glaucoma in the total study population. Fourth, we tested associations between the prevalence of glaucoma and a whole panoply of ocular and systemic parameters, what might have led to falsely high significances in the univariate analyses. In full recognition of this potential weakness, a subsequent multivariate analysis assessed these associations with additionally taking into account interdependencies between the independent parameters. If despite of it a Bonferroni correction was performed for the results of the multivariate analysis, the associations of the prevalence of open-angle glaucoma with all parameters, except for the diastolic blood pressure, and the associations of the prevalence of angle-closure glaucoma with all parameters remained significant. Fifth, information about other (more rare) types of secondary open-glaucoma other than pseudoexfoliative glaucoma, such as uveitic glaucoma, has not been available for statistical analysis. Sixth, the perimeter used in our study has not extensively been used in international scientific studies, and its parameters of perimetric defects may not be fully congruent with the perimetric indices of the internationally established perimeters of Octopus or Humphrey. The definition of glaucomatous optic neuropathy may thus not be fully comparable to the definition applied in other studies. Sixth, the comparison of the glaucoma prevalence between the various studies is hampered since the studies varied in the definitions of glaucoma used and the examination techniques applied. In addition, the overall prevalence of glaucoma in the various study populations depends on their age and sex structure, so that age-specific and sex-adjusted prevalence data should be the basis for the comparison. For many of the previous studies however, such data in tabulated form are not fully available ( Table 2).
In conclusion, in this typical ethnically mixed population from Russia with an age of 40 + years, the prevalence of glaucoma was 4.4% and increased with age, longer axial length, higher intraocular pressure, lens pseudoexfoliation, and lower diastolic blood pressure. The glaucoma prevalence was comparable with figures from Caucasian and East Asian populations, and it was lower than the figures for Sub-Saharan African populations. The ratio of open-angle glaucoma to angle-closure glaucoma of 2.7:1 and the percentage of glaucoma-related MSVI on total MSVI and of glaucoma-related blindness on total blindness roughly agree with the findings obtained in other recent studies. Two thirds of glaucoma patients were not on therapy, and in two thirds of the glaucoma patients the single IOP reading was ≤ 21 mmHg.