The present study enrolled 214 patients, aged 26 to 83 years, with symptomatic New York Heart Association class II through IV congestive heart failure. Patients were continued on their previous therapeutic regimens, which included an angiotensin-converting enzyme (ACE) inhibitor and a loop diuretic with or without digitalis. Patients were randomized to 1 of 5 parallel treatment groups: placebo or spironolactone at a single daily dose of 12.5, 25, 50, or 75 mg for 12 weeks. Serum levels of creatinine, urea nitrogen, potassium, plasma renin activity, and N-terminal proatrial natriuretic factor (pro-ANF), as well as urinary aldosterone levels, were measured periodically. Measurements at 12 weeks versus baseline values indicated significant increases in plasma renin activity and aldosterone excretion and significant decreases in systolic and diastolic blood pressure and pro-ANF. Hypokalemia (serum potassium < 3.4 mmol/L) occurred in 10% of placebo-treated patients and in 0.5% of the spironolactone group. The incidence of hyperkalemia (serum potassium > or = 5.5 mmol/L) was 5% for the placebo group, whereas it was 5%, 13%, 20%, and 24% for the 12.5-, 25-, 50- and 75-mg spironolactone treatment groups, respectively. Predictors of hyperkalemia included the use of ACE inhibitors other than captopril, ACE inhibitor dose, and baseline elevation of serum creatinine or potassium levels. Thus, daily doses of 12.5 to 25 mg of spironolactone coadministered with conventional therapy of ACE inhibitors, loop diuretics, and digitalis are relatively safe (provided that serum potassium levels are monitored) and effective in blocking the effects of aldosterone, while reducing the potential for hypokalemia in patients with heart failure.