Although hydralazine improves cardiac performance in patients with chronic left ventricular failure, its long-term clinical efficacy has not been established in controlled trials. We carried out a double-blind randomized trial of hydralazine (200 mg daily in 16 patients) versus placebo (16 patients) in patients with class III and IV symptoms while they were taking digitalis and diuretics. Maximal treadmill exercise time was determined prior to and at 4, 10, 18, and 26 weeks of hydralazine or placebo treatment; average follow-up was 20 weeks. We found no change in body weight, clinical class, resting heart rate and blood pressure, or heart size (by chest x-ray examination and echocardiogram) during treatment in either group. The total number of complicating clinical events was insignificantly fewer in the hydralazine treated group (8 vs 13). Control exercise duration in the hydralazine group averaged 259 ± 21 seconds (SEM), and increased to 347 ± 35 seconds at 4 weeks (p < 0.01) and 421 ± 38 seconds at 26 weeks (p < 0.001). Exercise duration also increased significantly in the placebo group, from 271 ± 30 seconds at control to 340 ± 44 seconds at 4 weeks (p < 0.02) and 339 ± 46 seconds at 26 weeks (p < 0.02). No differences between groups were significant. Left ventricular ejection fraction remained depressed and unchanged in both groups. Thus long-term vasodilator treatment with hydralazine alone is not significantly more effective than placebo in chronic heart failure.