Age-related macular degeneration and the incidence of cardiovascular disease: a systematic review and meta-analysis.

IMPORTANCE Research has indicated some shared pathogenic mechanisms between age-related macular degeneration (AMD) and cardiovascular disease (CVD). However, results from prior epidemiologic studies have been inconsistent as to whether AMD is predictive of future CVD risk. OBJECTIVE To systematically review population-based cohort studies of the association between AMD and risk of total CVD and CVD subtypes, coronary heart disease (CHD) and stroke. DATA SOURCES A systematic search of the PubMed and EMBASE databases and reference lists of key retrieved articles up to December 20, 2012 without language restriction. DATA EXTRACTION Two reviewers independently extracted data on baseline AMD status, risk estimates of CVD and methods used to assess AMD and CVD. We pooled relative risks using random or fixed effects models as appropriate. RESULTS Thirteen cohort studies (8 prospective and 5 retrospective studies) with a total of 1,593,390 participants with 155,500 CVD events (92,039 stroke and 62,737 CHD) were included in this meta-analysis. Among all studies, early AMD was associated with a 15% (95% CI, 1.08-1.22) increased risk of total CVD. The relative risk was similar but not significant for late AMD (RR, 1.17; 95% CI, 0.98-1.40). In analyses restricted to the subset of prospective studies, the risk associated with early AMD did not appreciably change; however, there was a marked 66% (95% CI, 1.31-2.10) increased risk of CVD among those with late AMD. CONCLUSION Whereas the results from all cohort studies suggest that both early and late AMD are predictive of a small increase in risk of future CVD, subgroup analyses limited to prospective studies demonstrate a markedly increased risk of CVD among people with late AMD. Retrospective studies using healthcare databases may have inherent methodological limitations that obscure such association. Additional prospective studies are needed to further elucidate the associations between AMD and specific CVD outcomes.