Heart failure (HF) is increasing in prevalence in the United States. Little data exists on race and gender differences in HF incidence rates and case fatality. The Atherosclerosis Risk in Communities (ARIC) cohort is a population-based study from 4 United States communities (1987 to 2002). Prevalent HF cases (n = 750) were identified by self-report and were excluded. Incident HF was defined by the International Classification of Diseases codes for HF (428.0 to 428.9, I50) from a hospitalization (n = 1,206) or death certificate (n = 76). There were 1,282 incident HF cases over 198,417 person-years. The age-adjusted incidence rate (per 1,000 person-years) for Caucasian women, 3.4, was significantly less compared with all other groups (Caucasian men, 6.0; African-American women, 8.1; African-American men, 9.1). Age-adjusted HF incidence rates were greater for African-Americans than Caucasians, but adjustment for confounders attenuated the difference. The adjusted African-American-to-Caucasian hazard ratio was 0.86 (95% confidence interval, 0.70 to 1.06) for men, and similarly, 0.93 (95% confidence interval, 0.46 to 1.90) for women during the second half of follow-up. The hazard ratio for women during the first half of follow-up was 1.79 (95% confidence interval, 1.25 to 2.55). Thirty-day, 1-year, and 5-year case fatalities following hospitalization for HF were 10.4%, 22%, and 42.3%, respectively. African-Americans had a greater 5-year case fatality compared with Caucasians (p <0.05). In conclusion, heart failure incidence rates in African-American women were more similar to those of men than of Caucasian women. The greater HF incidence in African-Americans than in Caucasians is largely explained by African-Americans' greater levels of atherosclerotic risk factors.