We examined existing evidence concerning the relative efficacy of various vasodilator agents in chronic congestive heart failure. Only randomized placebo-controlled trials with clinical end points and treatment durations of four weeks or more were selected from an exhaustive search of the English-language medical literature. Twenty-eight trials involving 1976 patients were found. Treatment durations of the trials varied from one month to two years. Patients with symptomatic heart failure despite digitalis and diuretic therapy were studied; most were middle-aged men and approximately half had coronary artery disease. Results of the trials were appraised by three independent observers, and mortality and functional status outcomes were pooled in a meta-analysis. All vasodilator agents except hydralazine hydrochloride were associated with improvements in functional status. Angiotensin converting-enzyme inhibitors were the only agents associated with both decreased mortality (odds ratio, 0.51; 95% confidence interval, 0.34 to 0.75) and improved functional status (odds ratio, 4.53; 95% confidence interval, 3.46 to 5.92). The optimal timing for initiation of these agents was not established.