OBJECTIVE: To examine the association of age-related macular degeneration (AMD) with incident coronary heart disease (CHD) and all-cause mortality. DESIGN: Population-based prospective cohort study. PARTICIPANTS: From the Atherosclerosis Risk in Communities Study (n = 12 536; age range, 49-73 years). METHODS: Participants had retinal photographs of one eye taken between 1993 and 1995. Photographs were evaluated for the presence of early and late AMD signs according to the Wisconsin grading system. Incident CHD events (acute myocardial infarction, silent myocardial infarction, fatal CHD, and cardiac revascularization procedures) and all-cause mortality were identified prospectively using standardized methods. MAIN OUTCOME MEASURES: Incident CHD events and all-cause mortality. RESULTS: Of 11,414 persons at risk of CHD, there were 555 (4.9%) with AMD at baseline, of whom 540 were early AMD and 15 were late AMD cases. Over a 10-year follow-up, 922 persons developed an incident CHD event. After controlling for age, gender, race, systolic and diastolic blood pressure, pack-years of cigarette smoking, and other variables, early AMD was not associated with incident CHD (relative risk, 1.08; 95% confidence interval, 0.82-1.42). However, individuals with late AMD were significantly more likely to have an incident CHD event, with 4 CHD events among the 15 participants with late AMD at baseline (10-year cumulative incidence, 30.9%) as compared with 918 CHD events among the 11 399 participants without late AMD (incidence of 10.0%; P = 0.049, Fisher exact test). In the full cohort (n = 12 536), early AMD was not significantly associated with all-cause mortality. However, individuals with late AMD were more likely to die (10-year cumulative mortality rate, 23.5%) than those without late AMD (mortality rate, 8.9%; P = 0.088, Fisher exact test). CONCLUSIONS: These data provide no evidence of an association between early AMD signs with incident CHD and all-cause mortality in middle-aged persons. Individuals with late AMD appear to have a higher rate of CHD events than those without late AMD, but due to a small number of late AMD cases, this finding should be interpreted cautiously.