Few studies have investigated whether ethnic groups differ in the prevalence of peripheral arterial disease (PAD). We compared the distribution of the ankle-brachial index (ABI), a measure of PAD, between African Americans and non-Hispanic white individuals. Subjects (n = 931) belonged to the Genetic Epidemiology Network of Arteriopathy (GENOA) study, a community-based study of hypertensive sibships, and included 453 African Americans from Jackson, Mississipi (mean age 72 +/- 6 years, 69% women) and 478 non-Hispanic white individuals from Rochester, Minnesota (mean age 58 +/- 7 years, 64% women). ABI was determined at two sites in each lower extremity and the lowest of four indices was used in the analyses. PAD was defined as an ABI of < or = 0.95. Information about conventional risk factors was derived from interviews and from blood samples drawn at the study visit. The prevalence of diabetes and hypertension was significantly higher in African Americans than in non-Hispanic white individuals. After adjusting for age, African American subjects had a lower mean ABI (women 0.97 vs 1.04, p < 0.001; men 0.96 vs 1.12, p < 0.001) and a greater prevalence of PAD (women 34% vs 22%, p = 0.010; men 33% vs 11%, p < 0.001) than their non-Hispanic white counterparts. In multiple regression analyses, African American ethnicity was a predictor of a lower ABI and the presence of PAD in each sex after adjusting for age and other conventional risk factors. In conclusion, the lower ABI and greater prevalence of PAD in African Americans than in non-Hispanic white individuals is not explained by differences in conventional risk factors. Identifying additional 'novel' risk factors that account for the ethnic differences in PAD is an important next step towards understanding why such differences exist and developing more effective strategies to reduce the burden of PAD.