Ninety-two patients with heart failure refractory to digitalis and diuretic therapy had captopril (n = 50) or placebo (n = 42) added to their therapeutic regimen in a randomized, double-blind trial. During a 2 week dosage titration period, one captopril-treated patient died of an intracerebral hemorrhage. Over the remaining 10 week evaluation period, 1 captopril-treated patient (2%) was excluded from the study because of treatment failure as compared with 12 discontinuations (4 deaths and 8 failures [29%]) among the placebo group (p less than 0.001). Eighty percent of patients in the captopril group exhibited some degree of clinical improvement, whereas only 27% in the placebo group did so (p less than 0.001). The therapeutic advantage of captopril over placebo was evidenced by a mean improvement of 0.52 (2.8 +/- 0.1 to 2.3 +/- 0.1) in the New York Heart Association functional class value as compared with 0.03 (2.9 +/- 0.1 to 2.8 +/- 0.1) with placebo (p less than 0.0001). There was a 24% mean increase in exercise tolerance with captopril (495 +/- 22 to 614 +/- 27 seconds) as compared with 0.4% with placebo (480 +/- 28 to 483 +/- 43 seconds) (p less than 0.01); the captopril group had an increase in the ejection fraction from a mean baseline value of 0.19 +/- 0.02 to 0.22 +2- 0.02 as compared with 0.19 +/- 0.02 to 0.18 +/- 0.002 (p less than 0.05) in the placebo group. A cohort analysis revealed that improvement in exercise tolerance with captopril was gradual and progressive throughout the 12 weeks. Improvement in specific symptoms of heart failure, that is, dyspnea, fatigue and orthopnea, and the reduction of edema also were greater in the captopril-treated patients (p less than 0.05 to p less than 0.001). Captopril therapy was well tolerated, although symptomatic hypotension after the first dose caused withdrawal of three patients (3%) from the study. It was concluded that captopril is an effective adjunctive treatment to digitalis and diuretic drugs for patients with refractory heart failure.