Original research

Population-based survey of the prevalence and types of glaucoma in Bangladesh

Abstract

Background To delineate the epidemiological landscape of glaucoma using a population-based sample representative of Bangladesh.

Methods Using multistage stratified cluster random sampling, households were selected to identify individuals ≥35 years across all 8 divisions of Bangladesh. Sampling frames were derived from the 2011 national census. Fifty-eight study examination sites were set up for comprehensive eye evaluations, including intraocular pressure, gonioscopy and visual field testing when indicated. International Society for Geographic and Epidemiological Ophthalmology definitions were used to define glaucoma and glaucoma suspect cases.

Results One hundred forty clusters (89 rural and 51 urban) were randomly selected, and 13 791 residential households were visited. We invited 17 002 individuals ≥35 years for on-site examination, of which 12 000 (71%) complied, with a male-to-female ratio of 1:1. The prevalence of glaucoma was 3.2% (95% CI 2.79% to 3.64%), and glaucoma suspect was 10.1% (95% CI 9.05% to 11.12%). The majority (78%) had primary open-angle glaucoma (POAG), while angle closure was seen in 16%. Of the POAG, 83% (n=251) were normal-tension glaucoma. Multivariable logistic regression showed increasing age (OR=1.01 for every 5-year increment, 95% CI 1 to 1.01) and male gender (OR=1.43, 95% CI 1.15 to 1.77) to be associated with an increased risk of glaucoma.

Conclusions The prevalence of glaucoma in Bangladesh is 3.2% in ≥35-year-old individuals with older men most at risk. Extrapolating the results, we estimate about 2 million patients with glaucoma. Though normal-tension variety was the most common type, caution should be exercised in generalising these results to other populations.

What is already known on this topic

  • The prevalence of glaucoma in the Dhaka division of Bangladesh was 1.5% in 2004, but there is no country-wide study on this.

What this study adds

  • We employed a robust study methodology through a nationwide, population-based survey and found a glaucoma prevalence of 3.2% and a glaucoma suspect prevalence of 10.1%.

How this study might affect research, practice or policy

  • Glaucoma is a major public health problem in Bangladesh and mandates a national health programme backed by the central government so that appropriate screening modalities can be devised to detect at-risk populations.

Introduction

Glaucoma is the second most common cause of blindness worldwide and is the leading cause of irreversible blindness, accounting for 8% of total blindness.1 Since glaucomatous damage is irreversible and asymptomatic in the early stages, appropriate detection and monitoring are essential. With a population of 165 million, Bangladesh has a potentially significant glaucoma burden.2 A previous glaucoma survey in Bangladesh carried out in 2004 estimated the prevalence of definite glaucoma to be 2.1% and speculated that approximately 586 000 people in the country would have glaucoma out of a population of approximately 26 million adults aged 40 years and older.3 However, this was limited to the Dhaka division alone with a smaller sample size.

Prevalence data for glaucoma in the Southeast Asian region are mainly from India and vary from 2.6% to 3.5% in individuals 40 years and older.4–8 In Bangladesh, a study done in 2013 found glaucoma to be responsible for 0.8% of all blindness, behind cataracts (22%) and diabetic retinopathy (1.7%).9 Nevertheless, there is a lack of population-based studies assessing the prevalence of glaucoma in Bangladesh. In addition to the prevalence, the entire epidemiological profile of glaucoma in Bangladesh is unclear, including the types of glaucoma, baseline damage of patients, demographic risk factors for glaucoma and general awareness of glaucoma among the population. Knowledge of these factors will help us advocate better for early glaucoma diagnosis and implement treatment strategies in Bangladesh.

The Bangladesh Glaucoma Society is a formally registered society of glaucoma specialists practising in Bangladesh and has a current strength of 106 members. Due to the need for more recent data about glaucoma in the country, we undertook this survey to outline the epidemiology of glaucoma using a large, nationally representative population-based sample. In this report, we present the prevalence of glaucoma, its common types in the population and risk factors in Bangladeshis 35 years and older. To the best of our knowledge, this is the first population-based survey to estimate the prevalence and epidemiology of glaucoma in Bangladesh.

Methods

This was a cross-sectional study carried out as per the tenets of the Declaration of Helsinki. Informed consent was obtained from all patients at the time of home surveys as well as clinical examination. The study was carried out from June 2021 to July 2023. All households were visited per the study plan, and socioeconomic characteristics were recorded. The eligible population (35 years and above) were registered and invited to attend a designated glaucoma screening facility at different times. They were also provided transportation to and from the site.

Sampling methods and sample size

A multistage stratified cluster random sampling was employed, stratified by place of residence (urban/rural), and sampling frames were derived from the 2011 national census.10 The survey sample was stratified and selected in two stages. Bangladesh’s 8 states (divisions) were stratified into urban and rural areas. In the first stage, from the 8 divisions, 33 (out of 64) districts were selected, which included 66 subdistricts and city corporations (out of 545) based on the probability proportion to size sampling method.11 These were called first-stage sampling units (FSU). A sampling weight was also applied to ensure the actual representation of the survey results at the national and domain levels. Therefore, it was adjusted for the distribution of urban–rural households in the survey based on the urban–rural distribution in the 2011 population census.10 In the second stage, 1–2 clusters were selected from each FSU using random sampling from villages (smallest rural units) and mohallas (smallest urban units). Following site selection for a cluster, the field manager made a map of the area, and a household listing operation was carried out to identify potential subjects aged 35 years or above. In the list, the names of the household heads were recorded.

The formula for calculating the final sample size in terms of the number of individuals while taking non-response into account was:

Display Formula

where n is the sample size in households;

Deft is the design effect (a default value of 1.7 is used for Deft if not specified);

P is the estimated proportion; 1.53%3

α is the desired relative SE; 10%

Ri is the individual response rate; 90%

d is the number of eligible individuals per household.

Based on the above information and formula, the sample size was calculated to be 21 000 eligible individuals.

A total of 140 cluster sample sites were selected (89 rural, 51 urban), with an average of 100 households per cluster. With this design, the survey selected 14 000 residential households and estimated to interview about 21 000 adult individuals. Field investigators visited all the selected households, and all eligible subjects were interviewed and motivated to participate in the eye examination. Additionally, the socioeconomic characteristics of the household were also recorded. If an individual in the household agreed to participate, the field enumerator invited him/her and selected a participant for the survey. In each division, one-third of the clusters sampled were taken from urban areas, assuming that 35% of the population lived in urban areas in each of these administrative divisions. Examination centres were set up in 58 (of the 66) subdistricts (figure 1) and were equipped to conduct a comprehensive eye evaluation, including glaucoma-specific tests and were manned by 1 ophthalmologist and at least 2 optometrists.

Figure 1
Figure 1

Showing locations of the 58 examination centres on the Bangladesh map.

Teams and quality control

The study employed 9 qualified and experienced field staff (6 field investigators, 2 field supervisors and 1 field manager) for field-level data collection, who received hands-on training with role playing, interviewing techniques and record keeping from glaucoma specialists, ophthalmologists and statisticians. Two sets of staff were involved—the advance enumeration team and the eye examination team. Field data collectors were supervised by field managers and supervisors. The supervisors led performance reviews for team members, and ophthalmic nurses and optometrists were supervised by the glaucoma specialists (designated as team leaders). Investigators made periodic visits to field sites for monitoring and performance checks.

Clinical examinations

At the time of examination, participants were asked questions on the following: history of ocular difficulties; family history of glaucoma; medical histories such as hypertension, diabetes, asthma and refractive error; knowledge and management of existing glaucoma (if any including topical medications or surgery) and information of prior clinical examination were noted before survey examination.

An optometrist recorded the best corrected visual acuity at the examination site using a Snellen distance vision chart at 6 m. All slit lamp (Haag-Streit, Haag-Streit, Bern, Switzerland) examinations were performed by the ophthalmologist. The intraocular pressure (IOP) was measured by Goldmann applanation tonometry following corneal anaesthesia by oxybuprocaine hydrochloride. Gonioscopy was carried out on all eyes of all participants by Goldmann four mirror gonio lens, and the angle of the anterior chamber was graded by Shaffer’s angle grading system. Angles were defined as either open or narrow, depending on whether less than one-third of posterior (usually pigmented) trabecular meshwork circumference was seen. Central corneal thickness (CCT) (Ocuscan, Alcon Laboratories, USA) was measured in all patients, and CCT-corrected IOP was used for analysis.

The optic disc was examined after dilating each pupil (if angles graded as ‘open’) with one drop of a mixture of tropicamide 0.8% and phenylephrine 5% (Aristopharma, Dhaka, Bangladesh). Disc was examined through an undilated pupil if angles were judged narrow or closed. A +90 D lens (Volk) was used for slit lamp biomicroscopy. The ratio of the longest vertical diameter of the cup to the longest vertical diameter of the disc was estimated as the vertical cup disc ratio (VCDR) for each eye. Any asymmetry of the VCDRs between the two eyes, the narrowest part of the neuroretinal rim, any notching at the cup margin, disc pallor and disc haemorrhage were noted. A red-free light was used to check for any defect in the nerve fibre layer. Threshold visual field testing was carried with Humphrey visual field analyzer (ZEISS, USA). An automated threshold-related single stimulus suprathreshold program was used to check 68 points in the central 25° in each eye for those who suspected to have glaucoma based on the diagnostic criteria given below.

Criteria for glaucoma diagnosis

International Society for Geographic and Epidemiological Ophthalmology (ISGEO) definitions were used, with three levels of evidence.12

Category 1

VCDR ≥0.7 in either eye or asymmetry of 0.2 or more between eyes and a visual field defect consistent with glaucoma.

Category 2

VCDR ≥0.9 in either eye or right-left asymmetry of 0.3 or more, and a dependable field test result could not be done.

Category 3

IOP ˃ 26 mm Hg and visual acuity worse than 3/60, or evidence of previous glaucoma filtering surgery, when optic disc could not be examined because of media opacity.

Criteria for glaucoma suspect diagnosis

  • Disc suspect: met category 1 (but not category 2) disc criteria, but no definite field defects.

  • Field suspect: definite field defects, but not meeting category 1 disc criteria.

  • Optic disc margin haemorrhages.

  • IOP >97.5 percentile.

  • Occludable drainage angle (defined as one in which the pigmented/posterior trabecular meshwork could be seen for less than 120o of the angle circumference).

Visual field tests were judged to be reliable if fixation losses ≤20%, false positives ≤25% and false negatives ≤20%.

Type of glaucoma

Glaucoma was classified as primary and secondary glaucoma. Primary glaucoma was classified as primary open-angle glaucoma (POAG) or primary angle-closure glaucoma according to angle morphology viewed by gonioscopy. IOP less than 21 mm Hg in patients fulfilling other POAG criteria were considered normal-tension glaucoma (NTG). Glaucoma was classified as secondary where there was an identifiable underlying cause such as AC angle neovascularisation, exfoliation, pigment dispersion, trauma, surgical procedure or uveitis.

Patient and public involvement

While it was not possible to have public involvement in the design and conduct of the study, public representatives will be involved in disseminating our findings, so that the general audience is better informed about the extent and impact of glaucoma in our country.

Data analysis

Data entry was done using the Census and Survey Processing System software and exported to SPSS V.23 (SPSS) for analysis. The prevalence of glaucoma and glaucoma suspects was estimated and presented along with their 95% CIs.

Continuous variables were presented as mean with SD, and group differences between continuous variables were analysed using Student’s t-test or Wilcoxon’s rank sum test for non-parametric distributions. Similarly, categorical variables were presented as proportions (n, %) and group differences were assessed using the χ2 of Fischer’s exact test. Univariate and multivariable analysis was performed to investigate the associations of gender, age, area of residence, education and occupational status with glaucoma.

Results

Out of 21 387 households, 13 791 households (64%) (5002 urban, 8789 rural), with an eligible population of 27 205 were contactable. Of these, 17 002 consented to the survey, and 12 000 attended the final eye exam at the examination spot (figure 2). The prevalence of glaucoma was 3.2% (n=386, 95% CI 2.79 to 3.64), and the glaucoma suspect was 10.1% (n=1206, 95% CI 9.05 to 11.12). The majority (>98%) did not know they had glaucoma and were not on any IOP-lowering medications.

Figure 2
Figure 2

Patient selection and recruitment.

Men and women formed approximately equal numbers among the respondents (50.4% women, 49.6% men), but men had a significantly higher proportion of glaucoma compared with women (table 1). The proportion of glaucoma increased with increasing age such that risk of glaucoma in participants >60 years old was twice that of those <40 years of age (table 1).

Table 1
|
Sociodemographic characteristics of participants with glaucoma in the study population

Though the urban population had a slightly higher prevalence of glaucoma, the difference was not significantly higher compared with the rural participants (table 1). Within the eight divisions, Dhaka contributed the highest number of participants and also accounted for the highest proportion of patients with glaucoma (table 1). Those with secondary education had a higher proportion of glaucoma compared with those with primary education. Similarly, those in the highest wealth quintile had a higher proportion of glaucoma compared with those in the middle quintile (table 1).

Most (85.7%) patients were diagnosed based on category 1 of the ISGEO criteria. Women (12.0%) had a higher proportion of glaucoma suspect cases compared with men (table 2). This was primarily driven by a higher prevalence of occludable angles in women.

Table 2
|
Percentage of glaucoma and glaucoma suspect among men and women

Primary glaucoma formed most of the patients with glaucoma (94.6%). POAG was the most common (table 3), forming 78.4% of the population, with angle closure glaucoma forming 16.2%. Within POAG, the vast majority of patients (83.3%) were diagnosed as the normal tension variety. Of the secondary glaucomas, there were approximately equal proportions of neovascular, lens-induced, uveitic and traumatic glaucomas.

Table 3
|
Proportions of the different types of glaucoma

In multivariable logistic regression analysis (table 4) adjusting for type of residence, education and wealth status, we found that increasing age (1% higher likelihood of glaucoma with every 5-year increment in age, p=0.001) and male gender (43% higher likelihood of glaucoma compared with women, p<0.0001) were the only factors that increased the likelihood of glaucoma in our study population.

Table 4
|
Risk factors for glaucoma in Bangladesh

Discussion

The current survey was the largest nationwide, population-based survey of glaucoma in Bangladesh and aimed to give precise estimates of the disease burden in the general population. Because the study clusters were randomly selected using a robust methodology, we believe the sample was nationally representative by age, gender, ethnicity, rural and urban residence and socioeconomic status, and the findings are generalisable to the whole country. We found a glaucoma prevalence of 3.2% and glaucoma—suspects of 10.1%. Extrapolating these prevalence data based on a 2022 census carried out by the Bangladesh Bureau of Statistics,2 we estimate that, in people over 35 years of age, there are almost 2 million (1 943 470) patients with glaucoma in the country and an additional 6 million may be glaucoma suspects.

A population-based survey from Bangladesh over two decades ago showed a slightly lower prevalence of glaucoma at 2.1%, with higher rates in men compared with women.3 However, this was confined to the Dhaka division and had a much smaller sample size (n=2347) than ours (n=12 000). The prevalence of glaucoma suspect cases was significantly higher in our study compared with the previous report as well. We believe that the larger sample representative of the entire nation is more demonstrative of the actual picture. It is also possible that the incidence of glaucoma has increased over the past two decades in Bangladesh, a worrisome trend. However, it is more likely that older men (>60 years of age) comprised only 8% of the sample in the previous study compared with 35% in ours. Given that older age and male gender are risk factors for glaucoma, this difference may be primarily responsible for the higher prevalence of glaucoma we report.

Studies carried out in neighbouring India report varied prevalence rates. The Chennai glaucoma study (2010), a population-based study on the prevalence of glaucoma in south India, reported a prevalence of 3.5% from a sample of 2532 subjects 40 years or older, similar to our results.6 In another study from south India4, Ramakrishnan et al reported a prevalence of 2.6% from a population of 5150 subjects who were 40 years or older, similar to reports by Rahaman et al, and lower than our findings.4 A more recent study (2015) in 13 591 Nigerians estimated an overall glaucoma prevalence of 5.02%. A higher prevalence of glaucoma in more recent studies compared with older ones suggests an increasing incidence and may be explained by higher life expectancy in most populations worldwide, and most studies, including ours, show increasing age to be a risk factor for glaucoma. These findings are also in accordance with the projections made by Tham et al, for a growing incidence of glaucoma by 2040.13

Men, especially older men >60 years of age, were found to be at the highest risk of glaucoma compared with younger men, and at a 43% higher risk than women of the same age. A similar finding was found by Rahman et al in Bangladesh in 2004.3 The evidence for similar gender predilections with glaucoma is more inconsistent in literature,14 given the fact that women tend to outlive men. Additionally, it appears that women are more predisposed to angle-closure glaucoma, while older men get open-angle glaucoma.14 Associations of gender with incidence of glaucoma need further investigations and probable mechanisms for these associations need to be discovered to attain credibility.

In our results, the prevalence of glaucoma was slightly higher in urban areas than in rural areas. The Chennai Glaucoma study found a much higher difference (1.6% difference).6 A study in Hyderabad in a predominantly urban population found a prevalence of 6.1%,5 against a study carried out in Tamil Nadu in a rural population which found a prevalence of 3.2%.4 Though our study did not elicit such marked differences between urban and rural populations, this requires further study.

The majority of our participants had POAG, with NTG being the predominant variety among these. A study in West Bengal,7 in an ethnically similar population, also found predominantly patients with POAG with an overall age-adjusted prevalence of 3.4%, with >90% being POAG. The high prevalence of NTG in our study is surprising and all ophthalmologists in Bangladesh should be aware of this fact. This also implies that the diagnosis of glaucoma cannot be based on IOP alone, and future screening programmes must include fundus imaging and visual field testing to enable early diagnosis of glaucoma.

The main drawback of our study was the slightly lower-than-expected response rate, where about 63% agreed to visit the examination site, and about 70% of these turned up. Although this might influence prevalence rates, we hope the overall sample of 12 000 participants is large enough to mitigate these effects. Additionally, due to budgeting constraints, optic disc imaging was not done for participants and visual fields were only done for some. Lastly, subjective patient responses about their history of glaucoma may have led to an inflation in the number of cases with NTG. Though this variety may be the most common type of glaucoma in our population, we believe the numbers are likely smaller than what we observed. Caution should be exercised while generalising these results to other populations, and future studies in the Bangladeshi population are required to see the true prevalence in our population.

Extrapolating the data from the target population, the current survey estimates there are about 2 million patients with glaucoma and nearly 6 million glaucoma suspect cases in Bangladesh, with a higher proportion of males affected. This indicates a major public health problem and mandates a national health programme backed by the government of Bangladesh so that appropriate screening modalities can be devised to detect at-risk populations. With this landmark population-based survey, appropriate resources should be devoted to properly assessing and managing glaucoma cases so as to reduce avoidable blindness in Bangladesh.