Number of People Blind or Visually Impaired by Glaucoma Worldwide and in World Regions 1990 – 2010: A Meta-Analysis

Objective To assess the number of individuals visually impaired or blind due to glaucoma and to examine regional differences and temporal changes in this parameter for the period from 1990 to 2012. Methods As part of the Global Burden of Diseases (GBD) Study 2010, we performed a systematic literature review for the period from 1980 to 2012. We primarily identified 14,908 relevant manuscripts, out of which 243 high-quality, population-based studies remained after review by an expert panel that involved application of selection criteria that dwelt on population representativeness and clarity of visual acuity methods used. Sixty-six specified the proportion attributable to glaucoma. The software tool DisMod-MR (Disease Modeling–Metaregression) of the GBD was used to calculate fraction of vision impairment due to glaucoma. Results In 2010, 2.1 million (95% Uncertainty Interval (UI):1.9,2.6) people were blind, and 4.2 (95% UI:3.7,5.8) million were visually impaired due to glaucoma. Glaucoma caused worldwide 6.6% (95% UI:5.9,7.9) of all blindness in 2010 and 2.2% (95% UI:2.0,2.8) of all moderate and severe visual impairment (MSVI). These figures were lower in regions with younger populations (<5% in South Asia) than in high-income regions with relatively old populations (>10%). From 1990 to 2010, the number of blind or visually impaired due to glaucoma increased by 0.8 million (95%UI:0.7, 1.1) or 62% and by 2.3 million (95%UI:2.1,3.5) or 83%, respectively. Percentage of global blindness caused by glaucoma increased between 1990 and 2010 from 4.4% (4.0,5.1) to 6.6%. Age-standardized prevalence of glaucoma related blindness and MSVI did not differ markedly between world regions nor between women. Significance By 2010, one out of 15 blind people was blind due to glaucoma, and one of 45 visually impaired people was visually impaired, highlighting the increasing global burden of glaucoma.


Methods
As part of the Global Burden of Diseases (GBD) Study 2010, we performed a systematic literature review for the period from 1980 to 2012. We primarily identified 14,908 relevant manuscripts, out of which 243 high-quality, population-based studies remained after review by an expert panel that involved application of selection criteria that dwelt on population representativeness and clarity of visual acuity methods used. Sixty-six specified the proportion attributable to glaucoma. The software tool DisMod-MR (Disease Modeling-Metaregression) of the GBD was used to calculate fraction of vision impairment due to glaucoma.

Introduction
Previous population-based investigations have shown that glaucoma is one of the most common, and thus most important, causes for vision loss worldwide . Previous estimations of global burden of glaucoma were based on meta-analyses which did not include studies, as far as available, from all regions of the world, which did not include all available populationbased studies, which did not assess a change during the last 2 decades, or which mostly reported on the prevalence of the disease [1][2][3]. Many of the population-based glaucoma studies did not report on the number of people blind or visually impaired due glaucoma. For public health purposes, however, the number of patients functionally affected is more important than the number of patients with any stage of the disease. For the individual patient and thus for the society, the burden of a disease is more important than just the presence of a disease including its early stages. We therefore conducted this meta-analysis of all available population-based studies performed worldwide within the last two decades to estimate the number of people affected by blindness (defined as presenting visual acuity <3/60) and moderately to severe visual impairment (MSVI; presenting visual acuity <6/18, !3/60) due to glaucoma, to assess changes in that figures during the period from 1990 to 2010, to examine regional differences in the prevalence of glaucoma related blindness and MSVI, and finally to compare the number of blind and visually impaired people with glaucoma with the number of people blind and visually impaired due to other diseases.

Methods
In a systemic literature research we used the systems of Medline, Embase and the WHO (World Health Organization) library information system to search for articles on vision loss and published in the period between 1980 and 2012. The methodology for this systematic review is described in Fig 1 as a PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flowchart with a PRISMA checklist in S1 Appendix. The search strategy is presented in S2 Appendix. Out of primarily identified 14,908 relevant manuscripts, we selected 243 high-quality, population-based studies after review by an expert panel. The latter involved application of selection criteria that were based on population representativeness and clarity of visual acuity methods used. As described in detail recently, search terms included concepts to describe "blindness", "visual impairment", "population", "eye", "survey", and a list of ocular disorders [35][36][37]. Additional unpublished data sources were found by personal  [37]. Two studies per region were available for 18 of the 21 GBD (Global Burden of Disease) Study regions, while only one study was identified for Central Europe. Eastern Europe and Central Africa did not have any study with cause-specific data. No study was identified for 126 of 191 countries. Data were extracted from published and unpublished reports into an electronic database (Microsoft Excel) by two investigators working independently with consistency checks in order to minimize data inputting errors. Extracted data included prevalence of predefined severities of vision loss by age, gender, country, region, and cause. Stratifying by age, sex, and geographical region, we estimated trends in causes of vision impairment and included an analysis of uncertainties. For geographical stratification, we used the 21 regions defined in the GBD Study [38]. As part of the statistical analysis, we first identified and accessed the data and then estimated fractions for each cause. We stratified the results by the severity of vision impairment, sex, age, and region. We finally applied the cause fractions to the prevalence of all-cause presenting vision impairment [37]. The method has been described in detail previously [37].
For the statistical analysis, the Disease Modeling-Metaregression (DisMod-MR) model from the GBD Study was used to determine the fraction of vision impairment caused by glaucoma or due to other causes mentioned above (more detailed information is available in S3 Appendix with an explanation of country and regional groupings, Table A, and a full list of citations of the studies, Table B) [37]. Briefly, the DisMod-MR model is a negative binomial regression model which included the following elements: covariates that predicted the variation in the true proportion of vision impairment from each disease; fixed effects that adjusted for definitional differences (e.g. whether the causes of presenting vs. best-corrected vision impairment were reported); a hierarchical model structure which fitted random intercepts in individual countries derived from the data observed in the country, in its region, and in other regions based on the availability and consistency of country-and region-specific data; age-specific fixed-effects which allowed for a non-linear age pattern; and a fixed effect for data on males. For the assessment of the fractions of blindness and visual impairment which were caused by glaucoma, we fitted one DisMod-MR model and used three covariates: an indicator variable which described whether the data were for blindness or for MSVI, an indicator variable describing whether the data were based on presenting visual acuity or best-corrected visual acuity measurements, and a country-level covariate which reflected the health systems access. We made two sets of the prediction for glaucoma, one for best-corrected blindness and one for best-corrected MSVI. Using the WHO reference population, we age-standardized the prevalences [39]. We also calculated the numbers of people with vision impairment and blindness caused by glaucoma. It reflected each region's population size and age structure.

Results
Blindness caused by glaucoma was found to be present in 2.1 million (95% uncertainty interval (UI): 1.9, 2.6) people, and MSVI caused by glaucoma was detected for 4.2 million (95%UI: 3.7-5.8) million individuals (Table 1). Taking into account that overall 32.4 million people were blind and 191 million people were vision impaired in 2010, glaucoma caused worldwide 6.6% (95%UI: 5.9, 7.9) of all causes for blindness in 2010 and 2.2% (95%UI: 2.0, 2.8) of all causes for MSVI (Table 1) [35]. The percentage of blindness caused by glaucoma varied from <5% in South Asia, East and West Sub-Saharan Africa, and Oceania, to 15.5% (9.5-21.9%) in Tropical Latin America. From the baseline in 1990 to 2010 the number of individuals blind due to glaucoma increased by 0.8 million (95%UI: 0.7, 1.1) and the number of individuals with MSVI due to glaucoma increased by 2.3 million (95%UI: 2.1, 3.5) (Tables 1 and 2). If only individuals with an age of 50+ years were included, the number of people blind due to glaucoma increased from 1.3 million (95%UI: 1.2, 1.6) in 1990 to 2.0 million (95%UI: 1.8, 2.5) in 2010, and the number of individuals with glaucoma related visual impairment increased from 1.9 million (95%UI: 1.5, 2.3) in 1990 to 3.8 million (95%UI: 3.3,5.3) in 2010. Compared with 1990, the percentage of global blindness caused by glaucoma increased from 4.4% (4.0, 5.1) to 6.6% (Tables 1 and 2). World regions with older populations such as the high-income regions, Southern Latin America, and Central and Eastern Europe, as compared to regions with younger populations showed a higher percentage of blindness caused by glaucoma in 1990 and in 2010 ( Table 1). The increase in the percentage of global blindness caused by glaucoma from 1990 to 2010 had taken place in all world regions without major difference between them (Tables 1 and 2). Vision Loss Due to Glaucoma Age-standardized prevalence of glaucoma related blindness was worldwide 0.1% (95%UI: 0.1, 02) in adults aged 50+ years in 2010, and the age-standardized prevalence of MSVI caused by glaucoma was worldwide 0.3% (95%UI: 0.2, 0.4) ( Table 1). Compared with 1990, the agestandardized prevalence of glaucoma-related blindness was reduced from 0.2% (95%UI: 0.1, 0.2) to 0.1% and the prevalence of glaucoma-related MSVI increased from 0.2% (95%UI: 0.2, 0.3) to 0.3% (Tables 1 and 2).

Discussion
Glaucoma was the cause for blindness in 2.1 million people or 6.6% of overall 32.4 million blind people globally in 2010, and glaucoma was the cause for MSVI in 4.2 million people or 2.2% of overall 191 million people visually impaired in 2010 [36]. These figures are lower than those reported by Quigley and Broman who forecasted in 2005 that in 2010, bilateral glaucoma related blindness would affect 8.4 million people [2]. Quigley and Broman discussed the difference between their estimate and an estimate of 4.4 million that was the most recent estimate at the time by the WHO Vision Group published by Resnikoff et al. [40]. Quigley and Broman argued that the difference was due to methodological issues given that blindness prevalence surveys often assigned the most ''treatable" disease as the primary cause of blindness. It is often assumed that cataract is more treatable than glaucoma, which leads to an underestimation of glaucoma blindness. In the recent analysis of global blindness, glaucoma ranked third together with macular degeneration (both: 6.6% of all blind people globally) after cataract (33.4% of all blind people globally) and undercorrection of refractive error (20.9%) in the list of the most common causes of global blindness [36]. Glaucoma and macular degeneration ranked first in the list of most common irreversible causes of blindness. With respect to MSVI, glaucoma ranked fourth (2.2%) after undercorrection of refractive error (52.9% of all people with MSVI globally), cataract (18.4%) and macular degeneration (2.2%) in the list of the most common causes of MSVI worldwide. These data confirm previous studies and meta-analyses which showed that glaucoma had a prominent ranking in the frequency list of causes for blindness and visual impairment. In contrast to the previous landmark study by Quigley and Broman, glaucoma was ranked third and fourth in our study as compared to being ranked second by Quigley and Broman and more recently Pascolini et al. as cause for blindness worldwide [2,41]. The numbers however also show that on a global perspective, cataract and undercorrection of refractive error are by far more prevalent as causes for blindness and MSVI. Only one out of 15 blind people was blind due to glaucoma, and only one out of 45 visually impaired people was visually impaired due to glaucoma. These figures may suggest that from a public health of view, providing adequate glasses for correction of refractive error and supplying cataract surgery to the blind and visually impaired may be at least as important as glaucoma care. In the recent meta-analysis of population-based studies by Tham and colleagues, the global prevalence of glaucoma in the population aged 40 to 80 years was 3.54% (95% Credible Intervals, 2.09, 5.82), and the number of individuals aged 40 to 80 years and affected by glaucoma worldwide was 64.3 million in the year 2013 [3]. These figures cannot directly be compared with the figures found in our investigation since Tham´s study addressed the number of individuals affected by glaucoma, independently of the stage of the disease, while our study assessed the number of individuals visually impaired or blind due to glaucoma. The percentage of blindness caused by glaucoma showed regional variations, with relatively low figures in regions with relatively young populations such as South Asia and Sub-Saharan Africa, and with relatively high figures in regions with relatively old populations such as the high-income regions (Tables 1 and 2). It was due to the dependence of the prevalence of glaucoma on age, while other causes for blindness and MSVI, namely undercorrection of refractive error and, to a lesser degree, cataract occurred also in younger groups of the population. These regional differences in the percentage of glaucoma as cause for blindness and MSVI remained mostly unchanged in the period from 1990 to 2010, since the differences in the age structure between the various world regions did not markedly change.
The global number of glaucoma blind increased by 0.8 million in the period from 1990 to 2010, although the age-standardized global prevalence of glaucoma related blindness in adults aged 50+ years decreased from 0.2% to 0.1%. The worldwide demographic transition with increasing population size, substantial increase in the average age in most regions and falling death rates more than outweighed the decrease in the prevalence of glaucoma related blindness so that the absolute numbers increased by 0.8 million or 62% from 1990 to 2010. The global prevalence of glaucoma related MSVI increased from 0.2% to 0.3% from 1990 to 2010, leading to marked increase in the absolute number of people visually impaired by glaucoma by 2.3 million or 83% in the same period. These figures show that, despite the relatively low percentage of glaucoma related blindness and MSVI on all causes of blindness and MSVI, an intensification of measures to address the growing number of people blind or visually impaired by glaucoma appears mandatory.
Expressed in percentage points, the age-standardized global prevalence of cataract, undercorrected refractive error and trachoma showed marked declines between 1990 and 2010 as reported previously [35]. The age-standardized prevalence of glaucoma declined less (for blindness) or even increased slightly (for MSVI). Similar findings were observed for the age-standardized prevalence of macular degeneration and diabetic retinopathy [35]. These developments may indicate a shift in the relative importance of the various diseases as causes for blindness and visual impairment, with a decrease for the major causes of cataract and undercorrection refractive error, which are relatively easily, safely and cost efficiently treatable, unlike diseases such as glaucoma, macular degeneration and diabetic retinopathy for which the therapy takes considerably more time and effort with a markedly lower rate of success.
Globally and in all regions, a larger percentage of blindness and MSVI caused by cataract and macular degeneration occurs in women than in men [36]. Globally, 36% of blindness among women was caused by cataract versus 30% of blindness among men; for MSVI, the figures were 20% versus 16%, respectively. In a similar manner, macular degeneration caused 7.3% of blindness among women versus 5.5% of blindness among men [36]. The glaucoma related blindness and MSVI did not show such marked disparities by sex in our study. This differs from the predictions of Quigley and Broman who estimated that 59% of all people with glaucoma of any stage would be women in 2010.
Literature reviews published by the WHO and the WHO Prevention of Blindness and Deafness program have previously been used to make worldwide estimates of numbers of people blind or with vision impairment. The latest of these studies included literature published in the period from 2000 to 2010 [41]. That analysis was limited to three age groups, with no breakdown by sex, provision of a point estimate for 2010, or estimates for the six WHO epidemiological subregions within a more limited timeframe. Interestingly, the figures for the percentage of blindness and MSVI caused by glaucoma did not markedly differ between Pascolini and Mar-iotti´s study and our study (glaucoma related blindness: 2% versus 2.2%; glaucoma related MSVI: 8% versus 6.0%) [41].
The present study forms part of a series of investigations on the prevalence and causes of vision loss in different world regions and on the number of individuals affected by MSVI and blindness caused by the major disorders of under-correction of refractive error, cataract, macular degeneration, diabetic retinopathy and glaucomatous optic neuropathy [42][43][44][45][46][47][48][49][50][51]. Applying the same statistical method (DisMod-MR) as the previous investigation, the present study addressed the number of individuals affected by glaucoma as cause for their MSVI or blindness. These figures may be of interest and help for politicians to direct financial means for the improvement of public health with respect to MSVI and blindness and to direct financial support for research in specific fields of medicine and in particular of ophthalmology. The previous studies of the series either assessed the worldwide prevalence of MSVI and blindness or examined the number of people affected by the main ocular disorders except for glaucomatous optic neuropathy.
The figure of 2.1 million individuals blind due to glaucoma and of 4.2 million individuals visually impaired due to glaucoma (representing 6.6% of all blindness and 2.2% of all MSVI worldwide) were lower than the figures of 10.8 million people blind and 35.1 million visually impaired due to cataract (representing 33.4% of all blindness and 18.4% of all MSVI worldwide), and also lower than the figures of 6.8 million people blind and 101.2 million people vision impaired due to undercorrected refractive errors (representing 20.9% of all blindness and 52.9% of all MSVI worldwide) [49,51]. The figures of glaucoma associated blindness and MSVI were similar to the numbers of 2.1 million individuals blind and 6.0 million individuals visually impaired due to macular diseases (representing 6.6% of all blindness and 3.1% of all vision impairment) [48]. They were higher than the figures of 0.8 million people were blind due to diabetic retinopathy and 3.7 million visually impaired due to diabetic retinopathy (2.6% of all blindness and 1.9% of all MSVI worldwide) [50].
The design of our study had potential limitations. First, as we discussed in our report of global prevalence of vision loss, a significant limitation was that many country-years lacked data, or there was only sub-national data available [36]. Relatively few national studies reported vision impairment for all ages and for all causes. Second, some data sources did not present the prevalence by age. By imputing age-specific cause fractions we were able to utilize this data with the assumption that the age pattern of the vision impaired in the particular study matched the modeled age pattern in the country where the study was conducted [36]. Third, the majority of population-based studies within the database that reported on vision loss due to glaucoma did not disaggregate their reported findings into glaucoma diagnostic subtypes such as open-angle glaucoma and angle-closure glaucoma, therefore it was not possible to differentiate between glaucoma subtypes in our analysis. Fourth, protocol dictated that population-based studies will report one cause as the principal cause for an individual examined in that individual study, so that causal prevalence can be calculated. In situations where multiple disorders contribute equally to visual loss, only the ''most easily preventable" or the ''most readily curable" cause is usually recorded [52]. This approach can underestimate the impact of diseases such as diabetic retinopathy and glaucoma when a study participant presents with cataract, while underestimating cataract burden when study participants also have an uncorrected refractive error [53]. Finally, the relatively small sample size of some studies meant that the confidence intervals of cause-specific prevalence estimates were relatively large. Our methods did however take sample size into account, so small sample size studies had less influence on the estimates than larger studies. Strengths of this study included the large amount of population-based data accessed and utilized and the trend analysis of causes of vision impairment and blindness, usage of non-linear age trends and modeling of data that were not reported by age, systematic quantitative analysis and reporting of uncertainty intervals. The large size of network of ophthalmologic researchers involved in first identification and then evaluation of data sources allowed access to unpublished materials and permitted us to obtain additional unpublished data from study investigators who had only published summary data, to evaluate all the major vision impairment studies, and to include only studies that met specific inclusion criteria regarding population representativeness and clear description and definition of visual acuity procedures.

Conclusion
In conclusion, in 2010, 2.1 million people were blind and 4.2 million people were visually impaired due to glaucoma. The number of people blind and visually impaired due to glaucoma increased by 0.8 million people or 62% and by 2.3 million people or 83%, respectively, in the period from 1990 to 2010. The contribution of glaucoma to total blindness and MSVI was higher in high-income regions with relatively older populations. One out of 15 blind people was blind due to glaucoma, and one out of 45 visually impaired people was visually impaired due to glaucoma.