Prevalence of Open-Angle Glaucoma in a Rural South Indian Population

Abstract

purpose. To determine the prevalence of primary open-angle glaucoma (POAG) and the associated risk factors in a rural population in southern India.

methods. Subjects aged 40 years or more (n = 3934) underwent a complete ophthalmic examination. Glaucoma was diagnosed according to the International Society of Geographical and Epidemiologic Ophthalmology classification.

results. Complete data were available for 3924 subjects (response rate, 81.75%). In eyes with normal suprathreshold visual fields, the mean intraocular pressure was 14.29 ± 3.32 mm Hg (97.5th and 99.5th percentiles, 21 and 25 mm Hg, respectively). The mean vertical cup-to-disc ratio was 0.39 ± 0.17 (97.5th and 99.5th percentiles, 0.7 and 0.8, respectively). Sixty-four subjects had definite POAG (1.62%, 9.5% CI 1.42–1.82); 30 were men and 34 were women. Subjects with POAG (59.85 ± 10.43 years) were older (P < 0.001) than the study population (53.78 ± 10.71 years). In only one (1.5%) person was POAG diagnosed before the study. Two (3.12%) subjects were blind due to POAG; 21 (32.81%) subjects had a presenting IOP >21 mm Hg, and 43 (67.19%) had an IOP <21 mm Hg. The mean central corneal thickness in subjects with POAG (502.82 ± 35.29 μm) was not different from that of the normal study population (505.93 ± 31.11 μm). No association was found with diabetes mellitus, systemic hypertension, gender, and myopia. Increasing IOP (per mm Hg) was associated with the disease (OR 1.12; 95% CI, 1.08–1.16). The odds for POAG increased with advancing age after adjustment for gender.

conclusions. The prevalence of POAG in this population was 1.62%. The prevalence increased with age, and 98.5% were not aware of the disease.

Glaucoma is the second leading cause of visual loss in the world. ¹ Recent population-based studies in southern India have reported various prevalence rates of primary open-angle glaucoma (POAG). ^{2 3 4} The Vellore Eye Survey (VES) ² reported a prevalence of 0.41% for POAG in the 30 to 60 year age group, whereas the Andhra Pradesh Eye Diseases Study (APEDS) ³ estimated the prevalence of POAG in the urban population to be 2.56% in those aged 40 years and older. The prevalence of POAG in the Aravind Comprehensive Eye Survey (ACES) ⁴ was 1.2%. One of the causes of these variations could be the different definitions used by each of the studies. The International Society of Geographical and Epidemiologic Ophthalmology (ISGEO) suggested a new classification for the diagnosis of glaucoma in population-based prevalence surveys. ⁵ In this classification, glaucoma is diagnosed on the grounds of both structural and functional evidence of glaucomatous optic neuropathy. We started the Chennai Glaucoma Study (CGS) in 2001 to study the prevalence of glaucoma in rural and urban populations of the southern Indian state of Tamil Nadu. The rural division of the study was completed in January 2003. In this article, we report the prevalence of POAG, according to the definitions suggested by the ISGEO, and the possible associated risk factors in a rural population. 

Methods

The methods and design of the CGS are described in detail elsewhere. ⁶ The details of the study design relevant to this article are given herein. This study was approved by the Institutional Ethics Review Board. 

Study Population and Field Work

The rural study area comprised 27 contiguous villages spread over Thiruvallur and Kancheepuram districts of Tamil Nadu. Of the residents, 49.8% were men. The literacy levels among the men and women in the rural study area were 30.04% and 17.08%, respectively. Of the employed residents, most were agricultural laborers (24%), followed by cultivators (11.18%), household workers (3.46%), trade and commerce employees (1.48%), construction workers (4.1%), and those in other services (3.86%); 51.92% were nonworkers. 

The total population of these villages was 22,000 persons, according to the report of the 1991 Census of India. Of these, 22% were older than 40 years with an age distribution similar to that reported by the Indian census. Of those aged ≥40 years, 4840 subjects were thought to reside in this area. A nonrandom sample of 4800 people aged ≥40 years was enumerated from this defined population. Trained social workers performed the enumeration by door-to-door survey. During enumeration, demographic information was collected by household questionnaire. All the eligible subjects were allotted a unique nine-digit identification number and were invited to come to the base hospital for detailed ophthalmic examination. 

Hospital Based Clinical Examination

Written, informed consent was obtained from all subjects who responded, and the study was performed in accordance with the tenets of Declaration of Helsinki. Detailed history was elicited pertaining to medical and ophthalmic problems. Two ophthalmologists (glaucoma specialists) and two optometrists who received additional training in the procedures of the study, performed the ophthalmic examination. The optometrists tested visual acuity and performed pachymetry and conducted visual field testing, and the ophthalmologists performed all other examinations. 

We measured the presenting and best corrected visual acuity using logarithm of minimum angle of resolution (log MAR) 4-m charts (Light House Low Vision Products, New York, NY). ⁶ Streak retinoscopy (Beta 200; Heine Optotechnik GmbH & Co. KG, Hersching, Germany) and subjective refraction were performed on all subjects. The best corrected visual acuity was ascertained and recorded. Refraction data are based on subjective refraction values. Emmetropia was defined as a spherical equivalent between −0.50 DS and +0.50 DS. Myopia was defined as spherical equivalent less than −0.50 DS and hyperopia as spherical equivalent greater than +0.50 DS. 

Automated visual fields were performed for all the subjects with best corrected visual acuity of 4/16 (logarithm of the minimum angle of resolution [logMAR] 0.6) or better, using frequency doubling perimetry (FDP; Carl Zeiss Meditec, Inc. Dublin, CA). All eligible subjects underwent C-20-1 screening (if the results were unreliable or abnormal, the test was repeated), and the N-30 threshold test. The reliability criteria for the C-20-1 screening test were no fixation or false-positive errors; and for the threshold N-30 test were <20% fixation errors and <33% false-positive and <33% false-negative errors. Visual fields with no depressed points to any level of sensitivity were considered to be normal. The central corneal thickness (CCT) was measured with the ultrasonic pachymeter (model 550; DGH Technology Inc., Exton, PA) before any contact procedure or pupillary dilation. The ocular surface was anesthetized with 0.5% proparacaine eye drops (Sunways, Mumbai, India). The measurement was made in automode with the subject supine, while he or she fixated on a distant target. The probe tip was placed perpendicular to the central cornea and applanated. Ten readings were obtained, and an average of these readings was recorded in micrometers. 

External examination and pupillary evaluation were performed with a flashlight. Slit lamp biomicroscopy was performed, and peripheral anterior chamber depth was graded according to the van Herick system. ⁷ Intraocular pressure (IOP) was recorded with a Goldmann applanation tonometer (model AT 030; Carl Zeiss Meditec) with the patient under topical anesthesia induced by proparacaine 0.5% and with fluorescein staining of the tear film. The right eye was measured first, and one reliable measurement was recorded for each eye. The instrument was calibrated on the first working day of each week. Gonioscopy was performed on all subjects in dim ambient illumination with a shortened slit that does not fall on the pupil. A four-mirror Sussmann lens (Volk Optical Inc., Mentor, OH) was used. The angle was graded according to the Shaffer system ⁸ and the peripheral iris contour, degree of trabecular meshwork pigmentation, and other angle abnormalities were recorded. An angle was considered occludable if the pigmented trabecular meshwork was not visible in >180° of angle in dim illumination. If the angle was occludable, indentation gonioscopy was performed, and the presence or absence of peripheral anterior synechiae was recorded. The agreement regarding grade of occludability between the two ophthalmologists was high (κ = 0.87). Laser iridotomy was performed in subjects with occludable angles after obtaining their consent. 

All subjects with open angles on gonioscopy had pupillary dilation with 1% tropicamide and 5% phenylephrine (Unimed Technologies, Halol, India). If phenylephrine was contraindicated, 1% homatropine (Warren Pharmaceuticals, Mumbai, India) was used. Subjects with occludable angles had dilation after laser iridotomy. Grading of lens opacification was performed at the slit lamp using the Lens Opacities Classification System II (LOCS II) ⁹ with a minimum pupillary dilation of 6 mm. Lenticular opacities were graded by comparison with the standard set of photographs. Stereoscopic evaluation of the optic nerve head was performed with a +78-D lens at the slit lamp. The vertical and horizontal cup-to-disc ratios (CDRs) were measured and recorded. The presence of any notching, splinter hemorrhages, and peripapillary atrophy was documented. The agreement between the two ophthalmologists regarding the CDR (0.7 or more) was high (κ = 0.87). A detailed retinal examination was performed using a binocular indirect ophthalmoscope and a +20-D lens. 

A provisional diagnosis of suspected glaucoma was made when the subject had one or more of the following conditions: IOP ≥21 mm Hg in either eye; vertical CDR (VCDR) ≥0.7 in either eye or CDR asymmetry ≥0.2; and focal thinning, notching, or a splinter hemorrhage. All these subjects were advised threshold visual field test using the Swedish interactive threshold algorithm (SITA) standard 30-2 program (model 750; Carl Zeiss Meditec). 

Diagnostic Definitions

The distribution of VCDR and IOP was obtained from those subjects with reliable and normal suprathreshold visual field test results by using FDP. Cases of glaucoma were defined according to the ISGEO classification. ⁵ Glaucoma was classified according to three levels of evidence. A category-1 diagnosis is based on structural and functional evidence. It requires a CDR or CDR asymmetry ≥97.5th percentile of the normal population or a neuroretinal rim width reduced to ≤0.1 CDR (between 11–10 or 5–7 o’clock) with a definite visual field defect consistent with glaucoma. A glaucomatous field defect was diagnosed, on a single, reliable threshold visual field examination of the central 30° (SITA standard 30-2), if the glaucoma hemifield test (GHT) results were outside normal limits, three or more abnormal contiguous nonedge points (except the nasal horizontal meridian) were depressed to P < 5%. ¹⁰ Reliability criteria were as recommended by the instrument’s algorithm (fixation losses, <20%; false-positive and false-negative, <33%). 

Category 2 contains cases of advanced structural damage with unproven visual field loss. This includes those subjects in whom visual field testing could not be performed or yielded unreliable results, with a CDR or CDR asymmetry ≥99.5th percentile for the normal population. Last, category 3 consisted of persons with an IOP ≥99.5th percentile of the normal population, whose optic discs could not be examined because of media opacities. Definite POAG was defined as the presence of one of the three categories with an open and normal-appearing angle on gonioscopy. An open angle was defined as one where 180° or more of the pigmented trabecular meshwork was visible. Pseudoexfoliation glaucoma and pigmentary glaucoma were not included. 

Blindness was defined as a best corrected logMAR visual acuity of <2/40 (logMAR 1.3) and/or constriction of the visual field <10° from fixation in the better eye. Myopia was defined as a spherical equivalent less than −0.50 D in a phakic eye. ¹¹ Diabetes mellitus was detected based on current use of antidiabetes medication and/or random blood sugar level greater than 200 mg/dL. ¹² We defined systemic hypertension as the current use of systemic antihypertensive medications or a measured systolic blood pressure ≥140 mm Hg and/or a diastolic blood pressure ≥90 mm Hg. 

Significance was assessed at the P < 0.05 level for all parameters. Categorical variables between groups were compared by χ² test, continuous variables by t-test, and continuous variables for multiple groups by ANOVA. Multivariate analysis was performed after adjusting for age (the age group of 40–49 years was used as the reference age group) and gender. The effect of IOP measurements in either eye on the diagnosis of POAG was assessed with generalized estimation equations (GEE). Statistical analysis was performed on computer (SPSS for Windows; SPSS, Inc., Chicago, IL). Odds for POAG are presented with 95% CI. 

Results

Table 1gives the details of the study population. A total of 3934 subjects of the 4800 enumerated agreed to participate in the study (response rate, 82.0%). Case records of 10 subjects were incomplete in one or more respects, and these subjects were excluded from the analysis; complete data were therefore available for 3924 (81.8%) subjects. The mean age was 53.78 ± 10.71 years, and 55.1% were women. There were 1810 subjects with normal and reliable suprathreshold visual field test results (by FDP) in both eyes. Distribution of the vertical CDR (VCDR) and IOP was derived from these subjects. Only the right eye was used for the analysis. Figure 1illustrates the distribution of VCDR in this perimetrically normal population. The mean VCDR was 0.39 ± 0.17 with 97.5th and 99.5th percentiles of 0.7 and 0.8, respectively. The 99.5th percentile for the CDR asymmetry was 0.2. The mean IOP was 14.29 ± 3.32 mm Hg, with the 97.5th and 99.5th percentiles being 21 and 25 mm Hg, respectively. Figure 2gives the distribution of IOP. Three hundred four subjects were advised to undergo SITA standard testing, depending on optic disc and IOP criteria. Of these, 40 refused visual field testing, 117 could not perform visual fields reliably on repeated testing and were excluded, and 147 underwent SITA standard testing. The VCDR data were not available for 235 individuals, because of media haze that did not permit visualization. Gonioscopic data were not available for 38 persons. 

There were 64 subjects with definite POAG (a prevalence of 1.62% with 95% CI, 1.42–1.82). There were 30 men and 34 women, a gender difference that was not significant (P = 0.95). Subjects with POAG were significantly older (P < 0.001) than the overall study population (mean age, 59.9 ± 10.4 years vs. 53.8 ± 10.7 years). The prevalence of POAG increased from 0.63% (95% CI, 0.44–0.83) in the age group of 40 to 49 years to 3.64% (95% CI, 1.16–6.12) in the age group of ≥80 years (Table 2) . The age- and gender-adjusted prevalence of POAG (based on provisional population totals, census of India 2001 ¹³ ) among the subjects ≥40 years of age in the rural Tamil Nadu population was 1.57% (95% CI, 1.19–1.95). 

POAG was classified as category 1 in 33 (51.6%) subjects, category 2 in 30 (46.9%) subjects, and category 3 in 1 (1.6%) subject. Two subjects (3.1%) were blind (visual field in the better eye <10° from fixation) due to POAG. Six subjects with POAG had a visual acuity of <2/40 in the better eye. In all six subjects, cataract accounted for the decrease in vision. Only one subject (1.5%) was a previously known case of POAG, the remainder (63 subjects; 98.5%) received the diagnosis in this study. 

There was no significant difference (P = 0.15) in age between subjects in level 1 (58.5 ± 9.8 years) and level 2 (61.5 ± 7.2 years). IOP also was not significantly associated (P = 0.89): level 1, 17.6 ± 3.69 mm Hg; level 2, 17.8 ± 6.65 mm Hg. A significantly larger proportion of the women (P = 0.01) had a level-1 diagnosis (M-F: 10:23) than had a level-2 diagnosis. (M-F: 19:11). 

The decade-wise distribution of the mean IOP across different age groups in the normal study population and those with POAG is shown in Table 3 . The IOP was significantly different in different age groups (one-way ANOVA, P < 0.001). The mean IOP was significantly (P < 0.001) higher among subjects with POAG than in the normal subjects (17.93 ± 5.35 mm Hg vs. 14.29 ± 3.32 mm Hg). When the influence of IOP in either eye on the diagnosis of POAG was assessed with generalized estimation equations, a statistically significant association (P < 0.05) was found between increasing IOP and the diagnosis of POAG. 

Twenty-one (32.8%) of 64 subjects with POAG had a presenting IOP of >21 mm Hg (category 1, 17 subjects; category 2, 3 subjects; category 3, 1 subject). The IOP was ≤21 mm Hg in the remaining 43 (67.19%) subjects (category 1, 16 subjects; category 2, 27 subjects). Even though a large proportion of subjects with definite POAG had an IOP ≤21 mm Hg, an increased prevalence of POAG was noted among persons with IOP >21 mm Hg (Fig. 3) . 

The mean CCT in the normal study population was 505.9 ± 31.1 μm. The mean CCT in subjects with POAG (502.8 ± 35.3 μm) was not different from the normal study group (P = 0.43). Subjects with POAG with an IOP >21 mm Hg had slightly thicker CCT (511.1 ± 26.7 μm) in comparison with subjects with POAG with IOP ≤21 mm Hg (500.2 ± 37.5 μm). This difference, however, was not statistically significant (P = 0.30). In the present study, none of the subjects with POAG had diabetes mellitus. The association of age, gender, myopia, IOP, CCT, and systemic hypertension with POAG is shown in Table 4 . After adjustment for gender, POAG was found to be significantly associated with increasing age. With the 40 to 49 age group used as a reference population, the odds ratio (OR) increased from 2. 59 (95% CI, 1.17–5.73) for the age group of 50 and 59 years to 5.26 (95% CI, 2.34–11.8) for subjects ≥70 years of age. IOP was positively associated with the diagnosis of POAG (OR, 1.12, 95% CI, 1.08–1.16). We found no association with CCT, gender, myopia, and systemic hypertension. 

Discussion

The Chennai Glaucoma Study was designed exclusively for estimating the prevalence of glaucoma. This is the first study from India to define glaucoma based on the ISGEO criteria and therefore is comparable with other studies all over the world that use the same criteria. In the present study, the prevalence of POAG was 1.62%. There was a significant increase in prevalence with age, but there was no difference in age-adjusted specific rates between genders. Of the subjects with POAG, 98.5% had undiagnosed disease before the study and 67.19% had an IOP ≤21 mm Hg at presentation. 

In the absence of reliable visual fields, glaucoma could have been overdiagnosed in category 2. Overall rates of high IOP associated with POAG were similar to the APEDS and lower than the ACES. Using IOP as a surrogate 17 (51.51%) of 33 of those in POAG category 1 had IOP >21 mm Hg, and 3 (10.0%) of 30 in category 2 had IOP >21 mm Hg. This could have caused an overestimation of disease among category 2 subjects. However, because performance of visual fields in this rural population was poor, some subjects with VCDRs between the 97.5th and the 99.5th percentiles, who could not perform reliably in visual field testing would be excluded, leading to an underdiagnosis of cases in category 1. Use of the Sussman lens to grade the angle may have resulted in some indentation, leading to an artifactual opening of the angle. This could have caused an overdiagnosis of POAG; however, the rates of angle closure are similar to those reported by investigators from India who used a Goldmann two-mirror lens for classification. ²  

Available data suggest that the prevalence of POAG varies from race to race. The reported prevalence of POAG among adult black populations ranges from 4.2% to 8.8%, ^{14 15} whereas the prevalence estimates in predominantly white adult populations range from 1.1% to 3%. ^{16 17 18 19 20} The prevalence estimates for POAG in east Asia vary from 0.5% to 2.3%. ^{21 22} The reported prevalence of POAG in India is between 0.41% and 2.56%. ^{2 3 4} Table 5summarizes the comparison between the present study and three other studies conducted in southern India. In this study, the prevalence rate of POAG (1.62%; 95% CI, 1.42–1.82) was higher than the reported prevalence from the Vellore Eye Survey (VES; 0.41%; 95% CI, 0.01–0.81). Vellore is a town situated 130 kilometers from our study area, and because of this geographical proximity, one would expect the prevalence rates to be similar. One reason for the variation observed may be the differences in the ages of the study participants. In the VES, persons examined were aged between 30 and 60 years, because the prevalence of POAG shows an increasing trend with age, the younger age profile of their population could have resulted in an underestimation of the prevalence of POAG. Another possible reason for the lower prevalence rate is the high drop out rate for visual field testing noted in the VES (51.5%). The investigators suggested that some participants with POAG were missed because of this reason, leading to underestimation of the prevalence. In addition, the definition of POAG in VES included IOP >21 mm Hg. This study and other studies ^{3 4} have shown that most people with POAG could have a presenting IOP ≤21 mm Hg. This again could have contributed to lower prevalence rates. Although the reported prevalence rates (in persons aged >40 years) in the APEDS ³ (2.56%; 95% CI, 1.22–3.91) and the ACES ⁴ (1.2%; 95% CI, 0.9–1.5) are different from that in our population, there is an overlap of confidence intervals. In the APEDS and the ACES different criteria were used to diagnose glaucoma. APEDS investigators used a combination of disc changes, IOP ≥22 mm Hg, and IOP asymmetry of 6 mm Hg as the criteria for advising the participant to undergo visual field examination. The inclusion of pseudoexfoliation glaucoma along with POAG in APEDS may explain the higher prevalence. Threshold visual fields were performed in 86.5% of subjects screened in the ACES. A visual field defect that corresponded to disc findings (a cup-to-disc ratio of ≥0.8, asymmetry >0.2, or thinnest neuroretinal rim width of 0.2) was used for diagnosis in 70.3% of cases. The remaining 30% were diagnosed on the basis of significant optic disc excavation compatible with glaucoma or end-stage glaucoma with severe central vision loss, or total optic disc cupping. 

Results from several studies have shown that the prevalence of POAG increases with age. ^{3 4 14 15 16 17 18 19} Our findings show a similar trend in the prevalence of POAG among older subjects. Subjects aged 70 years or more were five times more likely to have POAG than were those younger than 50 years. Some studies have shown higher prevalence of POAG in men, ^{4 14 19 23} but other studies have shown no gender difference in POAG prevalence. ^{3 17} The Blue Mountains Eye Study ¹⁸ reported a higher prevalence of glaucoma in women. In our study after adjusting for age, we found no difference in gender prevalence. 

The Barbados Eye Studies ¹⁴ reported that 51% of their persons with POAG had not had the disease diagnosed before the study. Similar estimates have been seen in other studies: 60% in the Visual Impairment Project, ²⁰ 53% in the Rotterdam Study, ¹⁹ and 51% in the Blue Mountains Eye Study. ¹⁸ In contrast to these, studies from India have reported a very high rate of undiagnosed POAG. In our study, only one subject (1.5%) was known to have POAG, whereas the remainder (98.5%) had the disease diagnosed during the course of study. The rate of undiagnosed POAG in the APEDS ³ was 92.6% and that in the ACES ⁴ was 93%. It is very evident from all three studies that >90% of subjects with POAG received their diagnosis for the first time during the respective studies. This finding is a matter of concern from a public health point of view. This may be due to the lack of facilities for comprehensive ophthalmic examination for the Indian population or to the widespread use of inappropriate eye examination techniques. 

In our study, only 32.81% of subjects with POAG presented with an IOP >21 mm Hg, possibly because of the consideration of a single presenting IOP measurement for analysis. Although a presenting IOP>21 mm Hg is associated with a higher risk of the development of glaucoma, most of those with POAG had IOPs of <21 mm Hg. The rate of prevalence of POAG still increased with increasing IOP, however, and the mean IOP of subjects with POAG was more than that of the overall study population. These findings are similar to the other study reports. ^{3 4} Our results reconfirm that the diagnosis of POAG cannot be based only on the level of IOP, but higher IOP is an important risk factor for POAG. In the present study, the prevalence of blindness due to POAG was found to be lower than that in the other two studies from India. ^{3 4} Regarding bilateral blindness among the three studies, the APEDS reported 3 (11.1%) of 27 with blindness in an urban population, and the ACES reported a single bilaterally blind person (1/64, 1.6%) in a rural population. The report on a rural population is similar to our finding. In our study bilateral blindness due to POAG was seen only in two (3.12%) subjects. Both were classified as blind by visual field definition (visual field <10° from fixation). Six other subjects with diagnosed POAG had visual acuity <2/40. Because visual field testing was not possible in these subjects and they did not have total glaucomatous optic atrophy (except in one subject in whom the optic disc was not visualized) the visual loss was attributed to cataract. 

Some studies have shown diabetes as a risk factor for POAG. ^{18 24 25} The Baltimore Eye Survey ²⁶ has shown no relationship between diabetes and POAG. In our study, none of the subjects with POAG were found to be diabetic. The limitation of our study is that we did not perform fasting blood sugar estimates for the participants, and our definition included only random blood sugar estimates. Some studies have shown an association between systemic hypertension and POAG, ^{27 28} whereas others have not. ^{20 29} Our study did not show any association of systemic hypertension with POAG. 

Contrary to other reports, we did not find myopia to be a risk factor for POAG ¹⁰ among those with a refractive error less than −0.5 spherical equivalent. However, our definition of myopia did not exclude lenticular myopia, and this could have affected our estimates. Patients with POAG in the Rotterdam Study ³⁰ were reported to have significantly thinner CCT than were the control subjects. In The Barbados Eye Studies, ³¹ the participants with POAG had thinner corneas (520.6 ± 37.7 μm) than those classified having no glaucoma (530.0 ± 37.7 μm). In the present study, the mean CCT in subjects with POAG was not significantly different from that of the normal study population. Within the POAG group, however, subjects with an IOP >21 mm Hg had thicker corneas than did the subjects with an IOP ≤21 mm Hg, but the difference was not statistically significant. 

The mean VCDR in our normal population was 0.39 ± 0.17, with the 97.5th and 99.5th percentiles being 0.7 and 0.8, respectively. The mean VCDR in subjects with POAG with IOP ≤21 mm Hg was 0.77 ± 0.15 and in those with IOP >21 mm Hg was 0.62 ± 0.23. When we compared our distribution of VCDR with other studies in which used normal suprathreshold visual fields were used to derive the distribution of VCDR, we found a similar pattern. The 97.5th and 99.5th percentiles for VCDR in a Chinese population of Singapore ³² were 0.71 and 0.81, respectively, whereas those in urban Thailand ²² were 0.72 and 0.86, respectively. The investigators in the Rotterdam Study ³³ and a population-based study from South Africa ³⁴ reported that the 97.5th percentile for VCDR in their populations was 0.7. We suggest that in our population, a VCDR >0.7 should be viewed as suspect for glaucoma. 

In conclusion, the overall prevalence of POAG was 1.62% in this population of rural southern India. The increase in prevalence with age is a cause of concern, as India’s adult population is expected to increase dramatically over the next few decades. Current levels of case detection in this rural population and from other reports from the country are extremely low. Case detection rates must improve significantly, to reduce ocular morbidity and minimize blindness due to glaucoma. 

 Supported by a grant from The Chennai Willingdon Corporate Foundation

 Submitted for publication December 30, 2004; revised May 26 and July 24, 2005; accepted October 14, 2005.

 Disclosure: L. Vijaya, None; R. George, None; P. G. Paul, None; M. Baskaran, None; H. Arvind, None; P. Raju, None; S. V. Ramesh, None; G. Kumaramanickavel, None; C. McCarty, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Demographics of the Study Sample

Age (y)	Responders (%)			Nonresponders (%)
		Men n (%)	Women n (%)	Men n (%)	Women n (%)
40–49	668 (75.1)	917 (84.6)	222 (24.9)	167 (15.4)
50–59	452 (77.4)	533 (82.6)	132 (22.6)	112 (17.4)
60–69	394 (84.0)	499 (85.7)	75 (16.0)	83 (14.3)
70–79	209 (88.2)	195 (87.4)	28 (11.8)	28 (12.6)
≥80	38 (77.6)	19 (70.4)	11 (22.4)	8 (29.6)
Subtotal for gender	1761 (79.0)	2163 (84.5)	468 (21.0)	398 (15.5)
Total	3924*			866

*  Ten Subjects with incomplete data were excluded from the analysis.

Figure 1.

 

The distribution of VCDR in those subjects with normal suprathreshold visual fields (right eye only).

Figure 2.

 

The distribution of intraocular pressure with normal suprathreshold visual fields (right eye only).

Table 2.

 

Prevalence of POAG by Age and Sex

Age (y)	Men			Women			All
		Number*	Prevalence (95% CI)	Number*	Prevalence (95% CI)	Number	Prevalence (95% CI)
40–49	3 (2,1,0)	0.45 (0.19–0.71)	7 (5,2,0)	0.76 (0.48–1.05)	10	0.63 (0.44–0.83)
50–59	9 (2,6,1)	1.99 (1.33–2.65)	7 (5,2,0)	1.31 (0.82–1.81)	16	1.62 (1.22–2.02)
60–69	9 (4,5,0)	2.28 (1.53–3.04)	14 (8,6,0)	2.81 (2.07–3.54)	23	2.64 (2.11–3.18)
70–79	7 (2,5,0)	3.35 (2.10–4.59)	6 (5,1,0)	3.08 (1.84–4.31)	13	3.33 (2.43–4.22)
≥80	2 (0,2,0)	5.26 (1.64–8.89)	0	—	2	3.64 (1.16–6.12)
Total	30		34		64	1.62 (1.42–1.82)

 POAG — Primary open angle glaucoma.

*  In parentheses: level 1, level 2, level 3.

Table 3.

 

Distribution of IOP across Age Groups in Persons with Normal Suprathreshold Visual Field Results and Patients with POAG

Age (y)	Normal (n = 1810)	POAG (n = 64)
40–49	14.20 (3.23)	18.80 (4.29)
50–59	14.71 (3.55)	20.06 (7.25)
60–69	14.10 (3.40)	16.91 (5.57)
70–79	13.32 (3.20)	17.61 (3.98)
≥80	11.67 (4.72)	13.50 (2.12)

 Data are expressed as the mean mm Hg ± SD. Normal: subjects with reliable, normal FDP screening test results. The difference between the different age groups was significant one-way ANOVA; P < 0.001). Linear regression for IOP versus age: r = −0.04 (P = 0.39).

Figure 3.

 

Prevalence of POAG at each IOP level (the higher IOP of the two eyes was used for analysis).

Table 4.

 

Multiple Logistic Regressions for Risk Factors for POAG in the Chennai Glaucoma Study

	n	Odds Ratio for POAG	95% CI
Age (y)
40–49	1585	1.00
50–59	985	2.59	1.17–5.73
60–69	892	4.15	1.97–8.76
≥70	462	5.26	2.34–11.80
Gender
Male	1761	1.00
Female	2163	0.98	0.58–1.62
IOP (mm Hg)	3864	1.12	1.08–1.16
CCT (μm)	3851	1.00	0.99–1.00
Myopia
Absent	2214	1.00
Present	1710	0.68	0.40–1.17
Hypertension
Absent	2651	1.00
Present	1273	1.02	0.60–1.74

 Myopia: spherical equivalent less than −0.50 D in a phakic eye. Hypertension: the current use of systemic antihypertensive medications or a measured systolic blood pressure ≥140 mm Hg and/or a diastolic blood pressure ≥90 mm Hg.

Table 5.

 

Age Specific POAG Prevalence Rates in All Population-Based Studies from India

	VES	APEDS	ACES	CGS (Current Study)
Diagnostic Criteria	Elevated IOP and/or disc changes with typical glaucoma fields defects on automated perimetry	Disc changes with field changes on automated perimetry	Disc changes with field changes on automated perimetry	ISGEO recommendations
Age (y)	30–60	≥40	≥40	≥40
Total Subjects (% response)	972 (63.9)	934 (85.4)*	5150 (93.0)	3934 (81.95)
Age groups, †
40–49	Data not provided	5 (1.26%)	7 (0.3%)	10 (0.63%)
50–59		6 (2.31%)	23 (1.6%)	16 (1.62%)
60–69		9 (4.89%)	22 (1.8%)	23 (2.64%)
>70		6 (6.32%)	12 (2.9%)	15 (3.25%)
Total	(0.41%)	26 (2.56%)	64 (1.2%)	64 (1.62%)

*  Response rate for the entire study population of 2522.

†  Data are the number of subjects, with the prevalence in parentheses.