To examine if myopia is a risk factor for age-related cataract.
Systematic review and meta-analysis.
A systematic review of the literature was performed using PubMed and Embase from their inception to March 2013 for population-based studies with data on myopia and age-related cataract, including nuclear, cortical, and posterior subcapsular (PSC) cataract. Separate meta-analyses for case-control/cross-sectional studies and cohort studies were conducted using random-effects models, with results reported as adjusted odds ratios (ORs) and relative risks (RRs), respectively.
A total of 38 007 subjects aged 30-97 years from 12 population-based studies were included in the meta-analysis. Meta-analysis of the 7 cross-sectional studies and 1 case-control study confirmed that myopia was associated with increasingly prevalent nuclear (pooled OR 2.81, 95% CI 1.94-4.06) and PSC cataract (pooled OR 1.93, 95% CI 1.49-2.49) but not with cortical cataract (pooled OR 1.08, 95% CI 0.90-1.30). Meta-analysis from 3 or 4 cohort studies showed nonsignificant associations of myopia with incident nuclear (pooled RR 1.25, 95% CI 0.71-2.21), cortical (pooled RR 1.21, 95% CI 0.67-2.19), and PSC cataract (pooled RR 1.26, 95% CI 0.92-1.74).
The associations of myopia with prevalent nuclear and PSC cataract are confirmed in meta-analysis of 8 study findings. The association of myopia with incidence of age-related cataract could not be confirmed in meta-analysis of 4 study findings.
Age-related cataract, including its subtypes of nuclear, cortical, and posterior subcapsular (PSC) cataract, is a leading cause of visual impairment worldwide. It is well known that increasing nuclear sclerosis of the lens with age leads to a “myopic shift” in refraction. In contrast, whether myopia is a risk factor that contributes to an early onset of different cataract subtypes is unclear. Population-based studies have shown inconsistent and conflicting results. For example, the Barbados Eye Study revealed an associated risk between myopia at baseline and incident nuclear cataract. However, the Beaver Dam Eye Study showed no relationship between baseline refraction and 5-year incident nuclear cataract.
A clearer understanding of whether myopia is a risk factor for cataract provides insights into the pathophysiology of different subtypes of cataract and the changes in refraction with age. It is also of great public health importance as myopia is a common eye disorder, affecting 20%-40% of middle-aged to elderly adults. Considering the conflicting evidence on myopia as a risk factor for age-related cataract, a systematic approach to combine the results of all available studies evaluating the association of myopia with age-related cataract would be informative. To address this gap, we conducted a systematic review and meta-analysis of the literature to examine whether myopia is a risk factor for age-related cataract.
Search Strategy and Inclusion Criteria
We conducted a systematic review and meta-analysis to examine the association of myopia with age-related cataract based on the Meta-analysis of Observational Studies in Epidemiology guidelines. We searched the electronic databases of PubMed and Embase for relevant papers reporting the association of myopia with age-related cataract published up to March 2013 with the following search terms: (“myopia”[MeSH Terms] OR “myopia”[All Fields]) OR (“refractive errors”[MeSH Terms] OR (“refractive”[All Fields] AND “errors”[All Fields]) OR “refractive errors”[All Fields] OR (“refractive”[All Fields] AND “error”[All Fields]) OR “refractive error”[All Fields]) AND (“cataract”[All Fields]). In addition, the reference lists of all identified studies were examined. Titles and abstracts of the studies were independently scanned by 2 authors (C.W.P. and C.Y.C.). The extracted studies were compared, and inconsistencies were resolved by consensus. Duplicate articles from the 2 databases were manually deleted.
We included studies if they were population-based and reported myopia as an independent variable and age-related cataract as the outcome measure. “Population-based” pertains to a general population defined by geopolitical boundaries, which is the sampling frame. Furthermore, we included studies only if the summary estimates, such as the odds ratio (OR) or relative risk (RR) with 95% confidence interval (CI), were reported in the paper or if the studies allowed for the calculation of the summary estimates based on the data presented in the paper. All studies used logistic regression models to adjust for confounders; therefore, we chose OR as the measure of association for case-control/cross-sectional studies. We also summarized the measures of association as RR for all the cohort studies. We excluded studies if they were not population-based or were published in languages other than English. We also included the unpublished data from the Singapore Malay Eye Study (SiMES), which is conducted by our team. The method of SiMES has been described elsewhere.
Data Extraction and Assessment of Study Quality
For each study in the analyses, we extracted the following information: first author, publication year, study name, study design, sample size, age range of the study participants, participation or follow-up rates, definitions of myopia and cataract, summary estimates and corresponding 95% CI, and confounding factors adjusted for. We assessed the study quality using the tool described by Sanderson and associates. The variables examined included the methods for selecting study participants, methods for measuring exposure (myopia) and outcome variable (age-related cataract), design-specific sources of bias (excluding confounding), methods for controlling confounding, statistical methods (excluding control of confounding), and conflict of interest.
Statistical Methods for the Meta-analysis
We performed the meta-analysis using Stata version 12.0 (StataCorp, College Station, Texas, USA). We meta-analyzed the fully adjusted summary estimates using the random-effects model to account for both within- and between-study variability. Different subtypes of cataract, including nuclear, cortical, and PSC cataract, were treated as the outcome measure, while myopia was analyzed as the independent variable. For studies that only reported stratified summary estimates, we pooled summary estimates to obtain an overall estimate for any myopia. We treated emmetropia as the reference category and converted summary estimates accordingly. Statistical heterogeneity among studies was evaluated using I 2 Statistic. Values of 0%-24%, 25%-49%, 50%-74%, and more than 75% denote no, low, moderate, and high heterogeneity, respectively. As the summary estimate was odds ratio for cross-sectional/case-control studies and relative risk for cohort studies, we handled heterogeneity attributable to study design by performing meta-analysis separately for cross-sectional/case-control studies and cohort studies. We evaluated publication bias using the Egger regression asymmetry test and Begg’s test.
We identified 3927 unique titles and abstracts, from which we retrieved 71 full-text articles for review. We included 1 case-control study, 7 cross-sectional studies (including the Singapore Malay Eye Study), and 4 cohort studies in this meta-analysis ( Figure 1 ).
Characteristics of the 1 case-control study, 7 cross-sectional studies, and 4 cohort studies included in the meta-analysis are summarized in Table 1 . These population-based studies were conducted in the United States, Europe, Australia, and Asia on different ethnic groups including whites, blacks, Chinese, Indians, and Malays. All studies graded cataract based on standardized protocols such as the Wisconsin grading system, the Wilmer grading system, or the Lens Opacity Classification System (LOCS). Refractive error was assessed by subjective refraction in all of the studies (1 study did not report the refraction method ). The definition of myopia varied across studies. Some defined myopia as SE <-0.5 diopter (D), while others used -1.0 D or -1.5 D as a cut-off value. All studies excluded participants with previous cataract surgery from the analyses. All cross-sectional or cohort studies adjusted for important confounders such as age, diabetes, and smoking in the multivariate analysis. Most studies reported response rates or lost-to-follow-up rates and described sampling methods, albeit in varying degrees. The follow-up durations varied from 4-5 years in cohort studies. The detailed assessment of study quality of the included study is described in Table 2 .
|Author (Year)||Study||Region||Ethnicity||Sample Size||Age (y)||Response or Follow-up Rate (%) a||Myopia Cut-off||Cataract Assessment||Summary Estimates (OR or RR, 95% CI)|
|Nuclear Cataract||Cortical Cataract||PSC Cataract|
|Giuffré et al (2005)||CES||Italy||White||1068||≥40||67.3||≤−1.5D||LOCS II||8.05 (2.91, 22.3)||1.84 (0.79, 4.13)||3.22 (1.29, 8.03)|
|Lim et al (1999)||BMES||Australia||White||3654||≥49||82.4||<−1.0D||Wisconsin||1.3 (1.0, 1.6)||1.2 (0.8, 1.6)||2.5 (1.6, 4.1)|
|Wong et al (2001)||BDES||US||White||4926||≥43||83.2||≤−1.0D||Wisconsin||1.74 (1.28, 2.37)||0.86 (0.64, 1.16)||1.23 (0.75, 2.03)|
|Duan et al (2013)||HES||China||Chinese||6830||≥30||90.4||<−0.5D||LOCS III||6.47 (3.14, 13.32)||2.17 (1.60, 2.95)||6.67 (2.56, 17.36)|
|Chang et al (2005)||SEEP||US||White and black||2520||≥65||NA||≤−0.5D||Wilmer||3.18 (2.22, 4.56)||0.83 (0.61, 1.11)||4.07 (1.71, 9.66)|
|Wong et al (2003)||TPS||Singapore||Chinese||1090||≥40||63.5||<−0.5D||LOCS III||2.89 (2.49, 3.24)||0.99 (0.98, 1.09)||1.61 (1.51, 1.82)|
|Pan et al (2013)||SINDI||Singapore||Indian||3337||≥40||75.6||<−0.5D||LOCS III||1.57 (1.13, 2.20)||1.06 (0.84, 1.33)||1.73 (1.10, 2.72)|
|Unpublished data||SiMES||Singapore||Malay||3280||≥40||75.6||<−0.5D||Wisconsin||4.12 (3.89, 4.36)||0.85 (0.81, 0.91)||1.35 (1.31, 1.40)|
|Wong et al (2001)||BDES||US||White||3684||5-year follow-up||81.1||≤−1.0D||Wisconsin||0.86 (0.63, 1.16)||1.08 (0.77, 1.51)||1.18 (0.73, 1.92)|
|Younan et al (2002)||BMES||Australia||White||2334||5-year follow-up||63.9||≤−1.0D||Wisconsin||1.0 (0.7, 1.5)||0.7 (0.4, 1.3)||2.1 (1.0, 4.8)|
|Mukesh et al (2006)||VIP||Australia||White||2594||5-year follow-up||85.0||≤−1.0D||Wilmer||1.0 (0.62, 1.60)||2.2 (1.4, 3.4)||1.1 (0.69, 1.9)|
|Leske et al (2002)||BISED||US||Black||2609||4-year follow-up||85.0||<−0.5D||LOCS II||2.8 (2.0, 4.0)||–||–|
|Study||Methods for Selecting Study Participants||Methods for Measuring Exposure (Myopia)||Methods for Measuring Outcome (Cataract)||Design-Specific Sources of Bias||Methods for Controlling Confounding, and Statistical Methods||Conflict of Interest|
|Giuffré et al (2005)||Using random numbers, a stratified random sampling was carried out in order to enroll half of the eligible population in each age group.||Not reported||Lens Opacities Classification System II||Selection bias; misclassification bias; chance finding; residual confounding; small sample size||Multivariate analysis adjusted for confounders: Nuclear cataract: iris atrophy; Cortical cataract: none; PSC cataract: iris atrophy, family history||None reported|
|Lim et al (1999)||Total of 3654 (82.4%) were selected from 4433 eligible residents in 2 postcode areas in the Blue Mountains area, west of Sydney.||Subjective refraction||Wisconsin Grading System||Selection bias; misclassification bias; chance finding; residual confounding||Multivariate analysis adjusted for age, sex, smoking, hypertension, diabetes, use of oral or inhaled steroids, and sun-related skin damage||None reported|
|Wong et al (2001)||All 5924 people who were 43-84 years of age were invited for baseline examinations. Of eligible persons, 4926 participated in the baseline examination.||Subjective refraction||Wisconsin Grading System||Selection bias; misclassification bias; chance finding; residual confounding||Multivariate analysis adjusted for age, sex, diabetes, smoking, and education||None reported|
|Duan et al (2013)||Residents of Yongnian County in China aged 30+ years were randomly selected using a clustered sampling technique with probabilities proportionate to the size of population in each cluster.||Subjective refraction||Lens Opacities Classification System III||Selection bias; misclassification bias; chance finding; residual confounding||Multivariate analysis adjusted for age, sex, fasting glucose, serum cholesterol, diabetes, body mass index, smoking, drinking, hypertension, time spent outdoors, level of income and education||None reported|
|Chang et al (2005)||Total of 2520 individuals aged 65 to 84 years, identified through Medicare rolls, who resided in the Salisbury area of the eastern shore of Maryland, were enrolled.||Subjective refraction||Wilmer grading system||Selection bias; misclassification bias; chance finding; residual confounding||Multivariate analysis adjusted for age, race, and sex. Additionally adjusted: nuclear cataract: smoking, education; cortical cataract: diabetes, ultraviolet-B exposure; PSC cataract: diabetes||None reported|
|Wong et al (2003)||The electoral register listed 15 082 names of Chinese aged between 40 and 79 years residing in Tanjong Pagar district, Singapore. Two thousand names (13.3%) were initially selected by a stratified, clustered, random sampling method, with more weights given to the older age groups.||Subjective refraction||Lens Opacities Classification System III||Selection bias; misclassification bias; chance finding; residual confounding; small sample size||Multivariate analysis adjusted for age, sex, diabetes, smoking, and education||None reported|
|Pan et al (2013)||An age-stratified random sample of the Indian population aged over 40 years was drawn from a computer-generated random list of 12 000 Indians living in southwest Singapore. In total, 3400 individuals (75.6%) participated in the study.||Subjective refraction||Lens Opacities Classification System III||Selection bias; misclassification bias; chance finding; residual confounding||Multivariate analysis adjusted for age, sex, diabetes, smoking, and education||None reported|
|The Singapore Malay Eye Study (unpublished)||An age-stratified random sample of the Malay population aged over 40 years was drawn from a computer-generated random list of 16 069 Malays living in southwest Singapore; 3280 individuals (78.7%) participated in the study.||Subjective refraction||Wisconsin Grading System||Selection bias; misclassification bias; chance finding; residual confounding||Multivariate analysis adjusted for age, sex, body mass index, systolic blood pressure, HbA1c, smoking, and education||None reported|
|Wong et al (2001)||All subjects identified during the initial census were invited for a second examination 5 years after the first. Of the 4541 participants from the baseline examination surviving, 3684 (81.1%) returned for the follow-up examination.||Subjective refraction||Wisconsin Grading System||Lost-to-follow-up bias; survival bias; chance finding; residual confounding||Multivariate analysis adjusted for age, sex, diabetes, smoking, and education||None reported|
|Younan et al (2002)||Five-year follow-up examinations were conducted, 2334 of the survivors (75.1%) were reexamined. Of those not seen, 383 (12.3%) had moved from the area and 394 (12.7%) refused the examination.||Subjective refraction||Wisconsin Grading System||Lost-to-follow-up bias; survival bias; chance finding; residual confounding||Multivariate analysis adjusted for age, sex, diabetes, smoking, and education||None reported|
|Mukesh et al (2006)||Baseline assessment was conducted on 3271 residents, recruited by door-to-door survey from 9 randomly selected adjacent pairs of census collector districts within urban Victoria. Five-year follow-up assessment was conducted on all available participants.||Subjective refraction||Wilmer grading system||Lost-to-follow-up bias; survival bias; chance finding; residual confounding||Multivariate analysis adjusted for age, sex, country of birth, occupation, smoking, arthritis, diabetes, vitamin C and calcium channel blockers intake||None reported|
|Leske et al (2002)||The studies were based on a simple random sample of Barbados African Americans, 40-84 years old (84% participation); 4631 persons completed baseline examinations at the study site. Surviving members of the cohort were invited to return for a 4-year follow-up visit.||Subjective refraction||Lens Opacities Classification System II||Lost-to-follow-up bias; survival bias; chance finding; residual confounding||Multivariate analysis adjusted for age, sex, body mass index, iris color, diabetes, IOP, and IOP-lowering treatment||None reported|
In the meta-analysis of 7 cross-sectional studies and 1 case-control study, myopia was associated with increasingly prevalent nuclear cataract (pooled OR 2.81, 95% CI 1.94-4.06; I 2 = 91.5%) and PSC cataract (pooled OR 1.93, 95% CI 1.49-2.49; I 2 = 66.7%) but not cortical cataract (pooled OR 1.08, 95% CI 0.90-1.30; I 2 = 78.4%) ( Figures 2-4 ). In the meta-analysis of 4 or 3 cohort studies, there were no statistically significant associations between baseline myopia and incident nuclear cataract (pooled RR 1.25, 95% CI 0.71-2.21; I 2 = 89.6%), cortical cataract (pooled RR 1.21, 95% CI 0.67-2.20; I 2 = 81.3%), or PSC cataract (pooled RR 1.26, 95% CI 0.92 1.74; I 2 = 0%) ( Figures 5-7 ). There was no evidence of publication bias as indicated by a nonsignificant Egger test (all P > .05) and Begg’s test (all P > .05) in all analyses.