Refractive Error and Risk of Early or Late Age-Related Macular Degeneration: A Systematic Review and Meta-Analysis

OBJECTIVE To summarize relevant evidence investigating the associations between refractive error and age-related macular degeneration (AMD). DESIGN Systematic review and meta-analysis. METHODS We searched Medline, Web of Science, and Cochrane databases as well as the reference lists of retrieved articles to identify studies that met the inclusion criteria. Extracted data were combined using a random-effects meta-analysis. Studies that were pertinent to our topic but did not meet the criteria for quantitative analysis were reported in a systematic review instead. MAIN OUTCOME MEASURES Pooled odds ratios (ORs) and 95% confidence intervals (CIs) for the associations between refractive error (hyperopia, myopia, per-diopter increase in spherical equivalent [SE] toward hyperopia, per-millimeter increase in axial length [AL]) and AMD (early and late, prevalent and incident). RESULTS Fourteen studies comprising over 5800 patients were eligible. Significant associations were found between hyperopia, myopia, per-diopter increase in SE, per-millimeter increase in AL, and prevalent early AMD. The pooled ORs and 95% CIs were 1.13 (1.06-1.20), 0.75 (0.56-0.94), 1.10 (1.07-1.14), and 0.79 (0.73-0.85), respectively. The per-diopter increase in SE was also significantly associated with early AMD incidence (OR, 1.06; 95% CI, 1.02-1.10). However, no significant association was found between hyperopia or myopia and early AMD incidence. Furthermore, neither prevalent nor incident late AMD was associated with refractive error. Considerable heterogeneity was found among studies investigating the association between myopia and prevalent early AMD (P = 0.001, I2 = 72.2%). Geographic location might play a role; the heterogeneity became non-significant after stratifying these studies into Asian and non-Asian subgroups. CONCLUSION Refractive error is associated with early AMD but not with late AMD. More large-scale longitudinal studies are needed to further investigate such associations.