We studied 944 hospitalized heart failure patients 65 years and older to examine the impact of incident atrial fibrillation on mortality. Patients were categorized into four groups based on past medical history and admission electrocardiogram: (1) no (neither past nor current), (2) incident (newly diagnosed), (3) past and (4) chronic (past and current) atrial fibrillation. The primary outcome was 4-year all-cause mortality. Bivariate and multivariable Cox proportion hazards analyses were used to determine risk of all-cause mortality. In the multivariable model, we adjusted for various demographic and clinical covariates. Compared with patients who never had atrial fibrillation, those with incident atrial fibrillation had a 57% higher risk of death (unadjusted hazard ratio, 1.57; 95% confidence interval, 1.22-2.03). After adjustment for other covariates the association remained unchanged (adjusted hazard ratio, 1.41; 95% confidence interval, 1.08-1.83). Past or chronic atrial fibrillation was not associated with increased risk of death. © 2004 European Society of Cardiology. Published by Elsevier B.V. All rights reserved.