Prevalence and predictors of myopic macular degeneration among Asian adults: pooled analysis from the Asian Eye Epidemiology Consortium

Authors: Wong YL, Zhu X, Tham YC.

Geographical coverage: Included studies were all conducted in Asia. Specific calculations generated for India, Beijing, Russia and Singapore.

Sector: Epidemiology

Sub-sector: Burden of disease

Equity focus: Asian adults.

Study population: Asian adults with myopia.

Review type: Other review

Quantitative synthesis method: Multivariate regression analysis

Qualitative synthesis method: Not applicable

Background: The prevalence of Myopic Macular Degeneration (MMD) in adults varies from 0.2% to 3.8% globally and is higher in individuals with high myopia (13.3% to 72.7%). However, the lack of a standardized definition of MMD in past studies has hindered direct comparisons. Recent studies have adopted the International Photographic Classification and Grading System for Myopic Maculopathy, and some clinic-based studies have used the META-PM classification. A standardized META-PM classification would enable better comparison of MMD prevalence across studies. Established risk factors for MMD include older age, greater myopic refractive error, and longer axial length. Other demographic and systemic factors need further investigation due to limited MMD cases and statistical power in previous studies. Meta-analysis attempts to identify new associations are limited by the scarcity of published reports on MMD risk factors.

Objectives: To determine the prevalence and predictors of myopic macular degeneration (MMD) in a consortium of Asian studies.

Main findings:

Overall, authors found that older age, female, lower education, greater myopia severity and longer axial length (AL) were risk factors of MMD, and myopic SE was the strongest single predictor of MMD.

A total of 19,885 participants from the population-based studies and 1,379 highly myopic individuals from the clinic and school-based studies were included in the study.

The prevalence of Myopic Macular Degeneration (MMD) was found to be 0.4% in rural India, 0.5% in Beijing, 1.5% in Russia, and 5.2% in Singapore among Asian populations. In population-based studies, factors such as older age (OR = 1.13 per year), being female (OR = 2.0), having a higher spherical equivalent (SE; OR = 1.7 per negative dioptre), longer axial length (AL; OR = 3.1 per mm), and lower education level (OR = 1.9) were all associated with MMD, with all p-values less than 0.001. These associations were also observed in clinic-based and school-based studies, with older age (OR = 1.07; p < 0.001), being female (OR = 2.1; p < 0.001), longer AL (OR = 2.1; p < 0.001), and lower education level (OR = 1.7; p = 0.005) linked to MMD. SE had the highest Area Under the Curve (AUC) of 0.92, followed by AL with an AUC of 0.87. The combination of SE, age, education level, and gender resulted in a slightly higher AUC of 0.94.

Methodology:

Review authors stated including pre-determined studies based on pre-defined inclusion criteria (not reported within the review).

This study involved seven cross-sectional studies from the Asian Eye Epidemiology Consortium (AEEC), including four population-based and three high myopia studies. These studies provided information on Myopic Macular Degeneration (MMD) using the META-PM classification, along with various risk factors such as demographic (age and gender), socioeconomic (education level), and ocular factors in individuals of Asian ancestry.

This combined consortium analysis, studying the pattern of Myopic Macular Degeneration (MMD) lesions and risk factors, is divided into two groups: (1) population-based studies, and (2) studies involving participants with high myopia (axial length >26.0 mm in at least one eye). The population-based studies include the Singapore Epidemiology of Eye Diseases (SEED) studies with 8,901 participants aged 40-80 years, the Beijing Eye Study (BES) with 3,145 Chinese participants aged 50-90 years, the Ural Eye and Medical Study (UEMS) in Russia with 3,376 participants of Asian ancestry aged 40-90 years, and the Central India Eye and Medical Study (CIEMS) with 4,463 participants aged 30-100 years. The high myopia studies include the Shanghai High Myopia Study with 818 highly myopic Chinese participants aged 30-90 years, the Kangbuk Samsung Hospital in South Korea with 252 highly myopic participants aged 15-90 years, and the Hong Kong High Myopia Study with 309 highly myopic Chinese participants aged 30-70 years.

Authors collected data on colour fundus photographs-captured and fundus photograph grading performed for all study participants based on the International META-PM classification and categorised into five categories: no myopic macular lesions; tessellated fundus only; diffuse choroidal atrophy; patchy chorioretinal atrophy; and macular atrophy. ‘Plus’ lesions, which supplemented the META-PM categories, comprised lacquer cracks, choroidal neovascularisation (CNV) and Fuchs’ spot. An eye was considered to have MMD if META-PM category 2, 3, 4 or any ‘plus’ lesion, observed.

Demographic (age, gender and ethnicity), socioeconomic characteristics (education level and urban/rural area) and general medical history were collected from participants in all studies. Education level was dichotomised for all participants in each study. The study sites were differentiated between urban and rural areas. Arterial hypertension was defined as systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg, physician-diagnosed hypertension or self-reported history of hypertension. Diabetes mellitus was defined as random blood glucose concentration of ≥11.1 mmol/L, fasting blood glucose concentration ≥7.0 mmol/L or ≥126 mg/dL, an HbA1c value ≥6%, diabetic medication use or a physician-diagnosed history of diabetes.

MMD grade in the worse eye (the eye with longer AL) was the outcome of this study. For the population-based studies, analyses of associations were performed using two multivariable logistic regression models; model 1 adjusted for demographic (age and gender), socioeconomic (education and urban/rural area), AL and general health comorbidities (hypertension and diabetes) variables; and model 2 adjusted for demographic (age and gender), socioeconomic (education and urban/rural area), SE and general health (hypertension and diabetes) variables. SE was only evaluated in model 2, as SE data was available in the population-based studies but not the high myopia studies. For the high myopia studies, risk factors of MMD, including demographic (age and gender), socioeconomic (education), AL, study site and general health (hypertension and diabetes) variables were assessed. ORs were estimated using backward stepwise multivariable logistic regression models. To examine the predictive value of risk factors of MMD (for instance, SE, AL and age), receiver operating characteristic (ROC) curves were constructed separately using univariable logistic regression models and in combination using multivariable logistic regression models for participants in the population-based studies. The area under the curve (AUC) was compared between the different models. Statistical software (Stata) was used.

Applicability/external validity:

Authors noted significant association between MMD and low level of education (at odds with previous studies) and female gender (endorsing previous studies).

Authors noted a number of limitations. Selection bias is likely to be present for the highly myopic subjects from the school-based and clinic-based samples, especially when patients were selected from only one single hospital site which may not be representative of other eye clinics and hospitals in China and South Korea. Therefore, this may have resulted in overestimated rates of MMD in the highly myopic patients. There is heterogeneity between the studies, in terms of differences in study methodologies, which may limit direct comparability of prevalence rates and associated risk factors, in particular, the high myopia studies. Other potential risk factors of MMD, such as the presence of other ocular comorbidities (including cataract, glaucoma and diabetic retinopathy) and choroidal thickness, were not available. As spectral-domain optical coherence tomography images were not available for most of the population-based surveys in this study, the exact cause of CNV (due to pathologic myopia or AMD) could not be definitively determined for all studies. Nevertheless, misclassification of CNV due to AMD was unlikely, as these four studies also applied the standardised Wisconsin Age-Related Maculopathy Grading System, which would include the presence of drusen and retinal pigment epithelial changes as part of the classification of CNV due to AMD. Furthermore, the majority of CNV in our study had high myopia (73%) and MMD category 2 or worse (87%). As the design of the study is cross-sectional, the temporality of associations cannot be established and thus inference of causal relationships of MMD is limited.

Geographic focus: Included studies were all conducted in Asia. Specific calculations generated for India, Beijing, Russia and Singapore (meaning that LMICs were represented).

Summary of quality assessment:

This review is somewhat unusual, in that the studies to be included were pre-selected (as members of an existing consortium), rather than an independent search being undertaken. For this reason, many of the criteria conventionally used to assess approaches to identifying, including and appraising studies are not relevant. However, it is worth noting that no attempt was made to assess the risk of bias of included studies. When it comes to the analysis, approaches were generally robust; however, it is unclear whether data was extracted by more than one author. In addition, risk of bias of individual studies was not considered in the analysis. For these reasons, we have “low confidence” in the findings of this review.