Although the benefits of carvedilol in patients with heart failure and depressed ejection fraction (EF) have been elucidated, those in patients with preserved EF are not understood. We enrolled 40 patients with mild or moderate heart failure and EF >/=45%. They were randomly assigned to carvedilol (n = 19) or conventional therapy (n = 21). After 12 months of treatment, carvedilol significantly improved all end points (plasma concentration of B-type natriuretic peptide [BNP] from 175 (35 to 209) to 106 (52 to 160) pg/ml, mean (95% confidence interval) p <0.01; New York Heart Association functional class from 2.37 (2.13 to 2.61) to 1.56 (1.21 to 1.91), p <0.01; exercise capacity estimated with the Specific Activity Scale from 4.75 (4.50 to 5.00) to 5.68 (5.22 to 6.14) METs, p <0.02), whereas conventional therapy did not (plasma BNP concentration from 150 (114 to 186) to 174 (100 to 248) pg/ml; New York Heart Association functional class from 2.29 (2.08 to 2.50) to 2.11 (1.73 to 2.49); exercise capacity from 4.57 (4.34 to 4.80) to 4.72 (4.41 to 5.03) METs). Univariate regression analyses showed that only the use of carvedilol was correlated with the decrease in plasma BNP concentration (p <0.03). Multivariate analyses demonstrated that an ischemic cause of heart failure (p <0.02), high plasma concentration of BNP (p <0.02), left ventricular dilation (p <0.03), and use of carvedilol (p <0.04) at baseline were predictive of a decrease in plasma concentration of BNP. In conclusion, carvedilol potentially decreased neurohumoral activation, decreased symptoms, and increased exercise capacity in patients with heart failure and preserved EF.