Fifteen-Year Incidence Rate of Primary Angle Closure Disease in the Andhra Pradesh Eye Disease Study





Purpose


To report on the 15-year incidence of primary angle closure disease (PACD) among participants aged ≥40 years in rural southern India


Design


Population-based longitudinal incidence rate study


Methods


Setting: 3 rural study centres. Study population: Phakic participants aged ≥40 years who participated in both examination time points. Observation procedures: All participants at the baseline and at the mean 15-year follow-up visit underwent a detailed interview, anthropometry, blood pressure measurement, and comprehensive eye examination. Automated perimetry was attempted based on predefined criteria. Main outcome measures included development of any form of PACD, as defined by the International Society for Geographical and Epidemiological Ophthalmology (ISGEO), during the follow-up period in phakic participants, who did not have the disease at baseline.


Results


We analyzed data obtained from 1,197 (81.4% out of available 1,470) participants to calculate the incidence of the disease. The mean age (standard deviation) of the study participants at the baseline was 50.2 (8.1) years, with 670 male (45.5%) and 800 female (54.4%) participants. The incidence rate per 100 person-years (95% confidence interval) for primary angle closure suspect, primary angle closure, and primary angle closure glaucoma was 8.8 (8.4, 9.2), 6.2 (5.9, 6.6), and 1.6 (1.4, 1.8), respectively. Thus, the incidence of all forms of PACD was 16.4 (15.9, 17) per 100 person-years. On logistic regression analysis, female gender was a significant risk factor whereas presence of myopia was protective.


Conclusions


This study reports long-term incidence of PACD from rural India. It has implications for eye health care policies, strategies, and planning.


Glaucoma is one of the leading cause of irreversible blindness. Primary angle closure disease (PACD) includes the pre-disease states (primary angle closure suspect [PACS] and primary angle closure [PAC]) and overt disease (primary angle closure glaucoma [PACG]). With an estimated global prevalence of 0.5% (95% confidence interval [CI]: 0.11%, 1.36%), PACG affected more than 20 million people aged 40-80 years in 2013, which is predicted to increase to 32 million by 2040. The prevalence of PACG varies across geographic regions and ethnic groups, and is highest in Asia at 1.09% (95% CI 0.43%, 2.32%). Although PACG is less common than POAG, the prevalence of blindness is higher in people with PACG than in those with POAG. In addition, most forms of the disease are asymptomatic and difficult to diagnose.


In the recent past, several population-based surveys reported the prevalence of glaucoma, especially from Asia. However, data on the incidence rate of PACD are limited. Incidence studies are important as they determine the risk of developing the disease over a period of time. Studies have estimated the incidence of new cases of PACD or have explored the natural history by determining the risk of conversion from one form of the disease to another over time. There is considerably limited published literature on the former than the latter.


The Andhra Pradesh Eye Disease Study (APEDS) is a large, population-based study conducted in Southern India. The study was designed to determine the prevalence of eye diseases and their risk factors, to estimate the magnitude of blindness and low vision and their impact on quality of life, and to describe the barriers to accessing eye care services. The inital study, APEDS I had urban and rural samples. In this publication from APEDS III, we report the incidence of PACD, derived from the mean 15-year follow-up examination, in the 3 rural areas as the urban area could no longer be identified because of rapid urbanization. We also report risk factors associated with the development of the disease.


Methods


The study adhered to the tenets of the Declaration of Helsinki and was approved by the Institutional Review Board of the Hyderabad Eye Research Foundation; L V Prasad Eye Institute (LVPEI), Hyderabad, India; and the London School of Hygiene & Tropical Medicine (LSHTM), London. Written informed consent was obtained from all participants. The first phase of the APEDS (APEDS I) was conducted from 1996 to 2000 and included 10,293 participants of all ages. The sample was selected using a multistage cluster sampling procedure from 1 urban and 3 rural areas of the then undivided Andhra Pradesh state in southern India. The urban area was Hyderabad, and the rural areas were located in the West Godavari District (affluent rural) and in Adilabad and Mahabubnagar districts (poor rural). This was one of the most rigorous population-based surveys conducted in a low-income setting. Findings from this study significantly contributed to the development of eye care policies in India. ,


Between 2009 and 2010, a feasibility study called APEDS II was conducted to trace participants examined in APEDS I to estimate migration and mortality rates and to identify participants willing to be re-examined. The 3 rural areas were revisited wherein 5,447 (70.1%) of the 7,771 rural participants examined in APEDS I were traced. Re-examination of this cohort of participants after 15 years (range 13-17 years) between 2012 and 2016 constitutes APEDS III. In this manuscript, we report the incidence of PACD among participants aged ≥40 years at baseline, that is, in APEDS I. Details of the design and methodology for APEDS III have been described previously, and relevant details are summarized here.


A comprehensive eye examination was performed on all participants using similar methods to APEDS I. The study team was trained on the procedures. All 4 clinical investigators underwent interobserver agreement assessments with the principal investigator (PI, a glaucoma specialist) for lens grading, gonioscopy, and optic disc assessment prior to joining the study. There was only one investigator at any given time. Agreement between the PI and other investigators in the binary classification of the anterior chamber angle into occludable or open was high (kappa coefficient range 0.78-0.85). The vertical cup-to-disc ratio (CDR) was assessed subjectively in units of 0.05, with a kappa coefficient ranging between 0.69 and 0.81.


Participants with a presenting distance or near visual acuity >logMAR 0.0 underwent streak retinoscopy followed by subjective refraction and acceptance. Each eye was tested separately and then binocularly. Refraction was performed by a trained optometrist or vision technician. Intraocular pressure (IOP) was measured with Goldmann applanation tonometer (Haag-Streit, Bern, Switzerland). One more reading was taken if the initial reading was >21 mm Hg. Gonioscopy was performed in a dark room with a short and narrow light beam (1-2 mm) to avoid pupillary constriction. An NMR-K 2-mirror lens (Ocular Instruments, Bellevue Washington, USA) and a Sussman 4 mirror lens (Ocular Instruments) was used. The angle was considered occludable if the pigmented posterior trabecular meshwork was not visible for ≥180 degrees of the angle circumferences in the primary position, without manipulation under dim illumination.


All participants underwent pupillary dilatation; participants with occludable angles were dilated after laser iridotomy. The optic disc and peripapillary area were assessed with a 78-diopter (D) lens (Volk Optical, Mentor, OH, USA) at the slit lamp, and the entire fundus was assessed by indirect ophthalmoscopy using a 20-D lens (Volk Optical).


Participants unable to attend the study centre due to frailty or physical morbidity were examined at home using similar methods; that is, they had visual acuity assessment, slit lamp examination, IOP measurement with a Perkins tonometer (Perkins Mk3; Haag-Streit, Bern, Switzerland), gonioscopy with NMR-K 2-mirror as well as Sussman 4 mirror lens and optic disc assessment with a 78-D lens at the slit lamp. Indirect ophthalmoscopy using a 20-D lens was performed to examine the posterior segment. The anterior segments of those who were bedridden were examined with a handheld slit lamp (BA 904; Haag-Streit).


Automated visual field analysis using a Humphrey Visual Field (HVF) analyzer (Humphrey Instruments Inc., San Leandro, California, USA) was attempted for all participants with any of the following optic disc features: asymmetry in CDR of >0.2 between the eyes, a vertical CDR of ≥0.65; neuroretinal rim <0.2 at any clock hour; notch in the disc; disc hemorrhage; and obvious peripapillary nerve fiber layer defect in either eye. Visual fields were also assessed if the IOP was ≥22 mm Hg in either eye, or if there was an IOP difference of ≥6 mm Hg between the 2 eyes, using the threshold central 24-2 strategy (stimulus size III). If the visual field was abnormal or unreliable, the test was repeated. The criteria used to determine glaucomatous visual field defects included a field defect that correlated with optic disc damage and met ≥2 of Anderson’s 3 criteria.


Definitions


Definitions for an occludable angle and PACG were based on the International Society for Geographical and Epidemiological Ophthalmology (ISGEO) classification which uses 97.5th and 99.5th percentiles of IOP and vertical CDR of the normal population. In APEDS I, visual field testing was not performed in the entire sample, so normative data could not be used. Hence, as in our previous publication on prevalence, we used normative data from the Chennai Glaucoma Study (CGS) for the 97.5th and 99.5th percentile cutoffs for the IOP and CDR. The CGS and the APEDS populations were both located in south India and are mainly of similar ethnicity (Dravidians). The 97.5th and the 99.5th percentile cutoffs for IOP were 21 and 24 mm Hg, respectively, whereas those for CDR were 0.7 and 0.8, respectively, for the rural population.


Glaucoma was classified according to 3 levels of evidence. In Level I, the diagnosis was based on structural damage and functional changes, that is, CDR or CDR asymmetry ≥97.5th percentile for the normal population, and a neuroretinal rim width reduced to 0.1 CDR (between 10 and 1 o’clock or 5 and 7 o’clock) with definite visual field defects consistent with glaucoma. Level 2 was based on advanced structural damage with unproven field loss. This comprised participants in whom visual fields could not be determined or were unreliable, with a CDR or CDR asymmetry of ≥99.5th percentile for the normal population. Category 3 included persons with an IOP of ≥99.5th percentile for the normal population, whose optic discs could not be examined because of media opacity. In this category, additional criteria such as visual acuity, clinical evidence of glaucoma filtering surgery, and information in medical records were also taken into consideration.


A PACS was defined as an eye with an occludable angle. PAC was defined as an eye with PACS and peripheral anterior synechiae and/or elevated IOP without glaucomatous optic disc damage. PACG was defined as PAC with evidence of glaucoma as defined by the ISGEO. The entire spectrum of PACD consisted of PACS, PAC, and PACG.


The definitions and relevant denominators for each are shown in Table 1 . For each participant, the form of PACD was defined on the basis of the more affected eye.



Table 1

Definitions and Denominators for Angle Closure Disease
























Form of Angle Closure Disease Population at Risk (Denominator) Incidence
PACS Normal at baseline (X) PACS at follow-up (A)
PAC Normal (X) or PACS (Y) at baseline PAC at follow up (B)
PACG Normal (X) or PACS (Y) or PAC (Z) at baseline PACG at follow-up (C)
PACD Normal at baseline (X) PACS or PAC or PACG at follow up (A+B+C)

PAC = primary angle closure, PACD = primary angle closure disease, PACG = primary angle closure glaucoma, PACS = primary angle closure suspect.


At baseline, hyperopia was defined as a spherical equivalent of ≥0.5 D in phakic eyes, and myopia was defined as a spherical equivalent of –0.5 D or greater in phakic eyes. Nuclear sclerosis was graded using the LOCS III classification system; nuclear opalescence higher than Grade 2 was considered to be nuclear sclerosis. Hypertension was determined by either one or a combination of the following factors: history of high blood pressure diagnosed by a physician, current use of antihypertensive medication, and/or a blood pressure reading of ≥140/90 mm Hg. Diabetes mellitus (DM) was determined by a history of DM and/or diabetic retinopathy on clinical examination.


Two hundred seventy-three of the 1,470 participants (18.5%) were excluded for the following reasons: (1) participants had the following diagnosis at the baseline: PACD (n = 32), POAG (n = 13), and suspicion of glaucoma on the basis of the clinical appearance of the optic disc (n = 1); (2) participants had undergone cataract surgery in the intervening period (n = 180); and (3) no data were available on gonioscopy at baseline (n = 45) or an iridotomy had been performed (n = 2) ( Figure 1 ).




Figure 1


Flow chart showing the number of participants included in analysis.


Statistical Analysis


Data were analyzed for participants aged ≥40 years at baseline and re-examined in APEDS III. The Shapiro-Wilk test was used to check the normality of distribution. Data are presented as means (standard deviations [SDs]) and medians (first, third quartile), as appropriate. The incidence estimates were adjusted for the age and sex distribution of the population. Participants were classified into 3 groups on the basis of their age at baseline, that is, APEDS 1, as 40-49 years, 50-59 years, and 60 years and older. For categorical variables in univariable analysis, χ 2 or Fisher exact tests were used. And to compare continuous variables, t tests and 1-way analysis of variance were used. Age was used as a continuous variable; the age interval was per 1-year increase. The association of PACD with age, sex, hyperopia, myopia, nuclear sclerosis, hypertension, DM, and body mass index were evaluated first with univariable analysis followed by multivariable analysis using logistic regression. Multivariable regression model included variables that achieved definite ( P < .05) or borderline significance ( P < .1) in the univariable model. We also used the AIC (Akaike information criterion) while selecting the regression model. Multicollinearity was checked by calculating the variance inflation factor (VIF), and the goodness of fit for logistic regression models was checked using the Hosmer-Lemeshow test. Statistical analyses were undertaken using Stata 12.1 (StataCorp, College Station, TX). A 2-sided P value <.05 was considered statistically significant.


Results


A total of 2,790 participants aged ≥40 years were examined in APEDS I. After a mean 15 years, 1,470 (52.6%) were re-examined in APEDS III. The mean (SD) age of these participants was 50.2 (SD 8.1) years; median (first, third quartiles) age was 48 (44, 55) years and ranged between 40 and 82 years at baseline, that is, APEDS I. The distribution of participants by age group was as follows: 774 (52.6%) 40-49 years, 454 (30.8%) 50-59 years, and 242 (16.4%) 60 years and older. There were 670 (45.5%) men. Perimetry was performed in 380 participants, 256 (67.3%) of whom underwent repeat tests as per the study protocol.


We compared baseline demographic characteristics of participants and (1) all non-participants (ie, those who had died since APEDS I and those who did not respond in APEDS III) and (2) those who did not respond in APEDS III (“non-responders,” ie, participants who migrated, could not be traced, or refused to participate) ( Table 2 ). Comparing participants with all non-participants, the latter were older, were more likely to be male, to have nuclear sclerosis and myopia but not hyperopia, to have hypertension and DM, and a leaner body mass index. There was also no difference in baseline PACD between participants and non-participants. Comparing participants with non-responders, participants were more likely to be younger, to be male, nonmyopic, and not to have nuclear sclerosis.



Table 2

Comparison of Baseline Demographic Characteristics Between Participants and Non-participants in Andhra Pradesh Eye Disease Study 3





















































































































Variable Participants (n=1,470; 52.6%) Non-participants (n=1,320; 47.4%) Subdivision of Non-participants P Value b P Value c
Died (n=1,106; 39.6%) No response a (n=214; 7.6%)
Baseline age, y, mean (SD) 50.2 (8.1) 59.6 (10.4) 61 (9.9) 52.4 (9.5) <.01 <.01
Female 800 (54.4) 668 (50.6) 535 (48.3) 133 (62.1) <.01 .03
Hyperopia 259 (17.6) 211 (15.9) 162 (14.6) 49 (22.9) <.01 .06
Myopia 397 (27.0) 623 (47.2) 551 (49.8) 72 (33.6) <.01 .04
Nuclear sclerosis 181 (12.5) 520 (41.6) 473 (45.2) 47 (22.9) <.01 <.01
PACD 32 (2.1) 34 (2.5) 31 (2.8) 3 (1.4) .48 .3
Hypertension 545 (37.7) 654 (50.4) 560 (51.6) 94 (44.1) <.01 .07
Diabetes mellitus 20 (1.3) 51 (3.8) 48 (4.3) 3 (1.4) <.01 .9
Body mass index
18.5-24.99 705 (49.2) 547 (44.3) 437 (42.6) 110 (52.8) <.01 .71
<18.5 583 (40.7) 575 (46.6) 497 (48.4) 78 (37.5)
25-29.99 116 (8.1) 90 (7.3) 75 (7.3) 15 (7.2)
≥30 27 (1.8) 21 (1.7) 16 (1.5) 5 (2.3)

PACD = primary angle closure disease.

Unless otherwise noted, values are n (%).

a Includes participants who migrated, could not be traced, or refused to participate.


b Comparison between participants and non-participants.


c Comparison between participants and non-respondents.

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Dec 24, 2021 | Posted by in OPHTHALMOLOGY | Comments Off on Fifteen-Year Incidence Rate of Primary Angle Closure Disease in the Andhra Pradesh Eye Disease Study

Full access? Get Clinical Tree

Get Clinical Tree app for offline access