Study Design and Population

Data from 2 population-based cross-sectional studies of Indian adults in Singapore and India were used. The Singapore Indian Eye Study examined adult Indians aged 40-83 years living in Singapore between August 2007 and December 2009 and was approved by the Singhealth Institutional Review Board. The Central India Eye and Medical Study was conducted in central India in 8 villages comprising the rural region of central Maharashtra at a distance of approximately 40 km from Nagpur from April 2006 to August 2008. Both the Medical Ethics Committee of the Medical Faculty Mannheim of the Ruprecht-Karls-University Heidelberg and the Ethical Committee of the Suraj Eye Institute (Nagpur) approved the study. All participants gave informed consent, and both studies were conducted according to the Declaration of Helsinki.

Details of study methodology have been reported. Briefly, for the Singapore cohort, from a list of 12 000 Indians residing in southwest Singapore provided by the Ministry of Home Affairs, an age-stratified random sampling strategy was conducted to select 6350 subjects aged 40 years or older, of which 4497 subjects were eligible. Of these, 3400 (75.6%) participated in the study. A total of 3337 (98.1%) subjects had fundus photographs of adequate quality for AMD grading. For the rural India cohort, 5885 subjects aged above 30 years were invited to participate, of which 4711 (80.1%) participated. For the purpose of the current analysis on AMD, only subjects aged 40 years and above (3591; 76.2%) were included. Of these, 3422 eyes (95.3%) had fundus photographs of adequate quality for AMD grading. Singapore-residing participants were generally older (57.8 years, SD 10.1 vs 53.8 years, SD 11.1; P < .001) and more likely to be male (50.2% vs 47.3%; P = .02) than India-residing participants ( Supplemental Table 1 , available at AJO.com ).

Retinal Photography and Age-Related Macular Degeneration Grading

In both studies, digital retinal photographs were taken after pupil dilation and grading was performed following a modification of the Wisconsin age-related maculopathy grading system. In Singapore, color photographs of Early Treatment Diabetic Retinopathy Study (ETDRS) standard field 1 (centered on the optic disc) and ETDRS standard field 2 (centered on the fovea) of each eye were obtained (Canon CR-DGi with a 10-D SLR back; Canon, Tokyo, Japan). The photographs were graded for AMD signs at the Centre for Vision Research, University of Sydney, Australia. In India, digital photography of the lens, optic disc (20-degree), and the disc and macula (50-degree) was captured (Zeiss FF450 telecentric fundus camera; Zeiss Meditec, Oberkochen, Germany). For the assessment of AMD, the international classification of the Wisconsin age-related maculopathy grading system was used and performed in Nagpur. Each photograph was independently adjudicated by 2 senior graders (V.N. and J.J.). Differences in final classification were discussed and resolved.

Among the AMD features evaluated were drusen size, type, and area; increased retinal pigment; retinal pigment epithelial depigmentation; pure geographic atrophy; and signs of exudative macular degeneration. Drusen were classified as hard or soft; then soft drusen were divided into distinct and indistinct soft drusen. Retinal pigmentary abnormalities were graded as hypopigmentation and hyperpigmentation. Early AMD was defined by either any soft drusen (distinct or indistinct) and pigmentary abnormalities or large soft drusen 125 μm or more in diameter with a large drusen area (>500-μm-diameter circle) or large soft indistinct drusen in the absence of signs of late AMD, following the definitions used in the Blue Mountains Eye Study. Neovascular AMD lesions were defined as the presence of retinal pigment epithelium (RPE) detachment, neurosensory detachment, subretinal or sub-RPE hemorrhages, or intraretinal, subretinal, or sub-RPE scar tissue. Subretinal hemorrhages or hard exudates within the macular area were also considered signs of neovascular AMD if other retinal vascular diseases as alternative causes were ruled out. Geographic atrophy was defined by presence of visible choroidal vessels and a discrete atrophic area with sharp border and an area at ≥175 μm diameter. Late AMD was defined as having either neovascular AMD or geographic atrophy.

Risk Factors

We defined risk factors using similar definitions for the purpose of this analysis. All examinations were performed in the Singapore Eye Research Institute (for Singapore-residing participants) or Suraj Eye Institute (for India-residing participants) by trained research staff. Trained interviewers administered questionnaires at the respective site, to collect information on sociodemographic and lifestyle factors. Demographic variables included age, sex, marital status, education level, occupation, income, and family history of diabetes and hypertension. Age was defined as the age at examination. Lifestyle questions were asked about smoking habits and alcohol consumption. Physical examination included anthropometric measurements, blood pressure (BP), and a detailed ocular examination. In the Singapore-residing participants, blood pressure was measured using a digital automatic blood pressure monitor (Dinamap model Pro100V2; Criticon GmbH, Norderstedt, Germany), following the protocol used in the Multi-Ethnic Study of Atherosclerosis. The average of the 2 systolic and diastolic BP measurements, taken with participant sitting down, was used. In the India-residing participants, BP was measured by a trained technician. The study participants were sitting for at least 5 minutes and had refrained from smoking and drinking of coffee, tea, or alcohol for at least 3 hours. In addition, no exercise was performed for the last 30 minutes prior to the BP measurements. An automated mercury sphygmomanometer (Model IA2 [HEM-7011-C1]; Omron Healthcare Co Ltd, Kyoto, Japan) was used and the cuff size was chosen according to the measured circumference of the upper arm. Laboratory examination included measurement of fasting plasma glucose and lipids.

Cigarette smoking was categorized into current smokers, former smokers, or nonsmokers and alcohol consumption into drinkers (irrespective of quantity) and nondrinkers. Cardiovascular disease (CVD) was defined as self-reported myocardial infarction or angina or stroke. Hypertension was defined as systolic BP ≥140 mm Hg or diastolic BP ≥90 mm Hg or self-reported physician-diagnosed hypertension. For the Singapore-residing participants, diabetes mellitus was defined as random glucose of 11.1 mmol/L or more, a self-reported physician diagnosis of diabetes, or use of glucose-lowering medication. For the India-residing participants, diabetes mellitus was defined as hemoglobin A1C (HbA1C) ≥6.0% (performed in only 2423 participants) or random blood glucose ≥200 mg/dL or self-reported physician diagnosis. Dyslipidemia was defined as total cholesterol ≥6.2 mmol/L, low-density lipoprotein (LDL) cholesterol ≥4.1 mmol/L, high-density lipoprotein (HDL) cholesterol <1.0 mmol/L, or self-reported physician-diagnosed dyslipidemia.

Ocular Parameters

Refractive error in both eyes of each subject was determined by a trained technician using an autorefractor (Canon RK-5 Auto Ref-Keratometer; Canon Inc, Ltd, Tokyo, Japan) (both Singapore-residing and India-residing participants), after which subjective refraction was performed by an optometrist to achieve best-corrected visual acuity. The final subjective refraction result was used in the analysis. Spherical equivalent (SE) was defined as sphere plus half negative cylinder. Refractive errors were analyzed as a continuous variable.

Axial length (AL) and anterior chamber depth (ACD) were measured by trained technicians using noncontact partial coherence laser interferometry (IOLMaster version 3.01; Carl Zeiss Meditec AG, Jena, Germany) (for Singapore-residing participants). Corneal pachymetry was measured with US-1800 Echoscan (Nidek, Fremont, California, USA) in the Singapore cohort. Corneal pachymetry and ocular biometry were obtained by trained technicians with Pacscan (Sonomed, Lake Success, New York, USA) in the India cohort.

Statistical Analyses

All statistical analyses were performed using R version 2.13.2 (R Development Core Team, 2011; R Foundation for Statistical Computing, Vienna, Austria). Characteristics of the study population were described using proportions or means and standard deviation (SD). Age-standardized prevalence estimates were calculated using the direct method. The 2010 Singapore Indian population was used as the standard population for comparison of the Singapore and India cohort. Age strata–specific prevalence rates were weighted by the proportions of the corresponding strata in the population census, and then summed to give a summary prevalence rate. The relationship between early AMD and risk factors was analyzed using logistic regression adjusting for age and sex. Mixed-effects logistic regression was used to model combined data from Singapore site and India site with study effect as the random effect to control for heterogeneity between 2 studies. Age- and sex-adjusted mixed-effects logistic regression was performed to select potential confounders ( P < .05) for multivariate analysis, with final covariates selected including age, sex, body mass index (BMI), cholesterol, arterial hypertension, refractive error, myocardial infarction, intraocular pressure (IOP), central corneal thickness, and study effect.

Finally, we performed a comparative analysis of the age-specific prevalence of AMD from Indian subjects aged 40 years and above. We performed a PubMed search for relevant papers of population-based studies reporting AMD prevalence in Indians published from January 1, 1990 through September 18, 2012, using the terms “age-related macular degeneration” AND “prevalence” AND “population study” AND (“Indian” OR “Indians” OR “India”). Twenty-six articles were identified initially, of which 17 were assessed as relevant (C.M.G.C.) for the purpose of reporting prevalence of AMD among Indian persons within a population-based study sample. All 17 articles were from one of the 6 population studies, namely the Aravind Comprehensive Eye Study, the Andhra Pradesh Eye Disease Study, the Central India Eye and Medical Study, the INDEYE feasibility study, the INDEYE study, and the Singapore Prospective Study Program. These studies were included in the meta-analysis.

We used a random-effects model because of the heterogeneity among these studies. These studies were selected after publication bias was assessed in a funnel plot, using Egger’s test. For ocular parameter analysis, data of the eyes with AMD was selected. If both eyes or neither eye had AMD, then data of the right eye were used for analysis.