OBJECTIVE--To compare the first dose responses to low dose angiotensin converting enzyme inhibitors (captopril, enalapril, and perindopril) in elderly patients with stable chronic heart failure. DESIGN--Double blind, randomised, placebo controlled, parallel, group prospective study of elderly patients with stable chronic heart failure. SETTING--General hospital in-patient admissions for supervised diuretic withdrawal (24-48 hours) and the introduction of angiotensin converting enzyme inhibitor therapy. PATIENTS--48 unselected elderly (58-85 years) patients with symptomatic but stable chronic heart failure (New York Heart Association grades II-IV) confirmed by clinical history, examination, and cardiological investigations. Patients gave their written and informed consent to receive their initial treatment under double blind conditions; blood pressure was monitored and blood samples taken to measure the pharmacokinetic and neurohormonal responses. INTERVENTION--Patients were randomised to receive a daily oral dose of placebo, captopril (6.25 mg), enalapril (2.5 mg), or perindopril (2 mg). MAIN OUTCOME MEASURES--Blood pressure and heart rate responses, drug concentration, and plasma renin and ACE activities. Differences between treatment groups were analysed by analysis of variance. RESULTS--The four randomised groups of patients had similar age, severity of heart failure (NYHA class), pretreatment diuretic dosage, plasma renin activity, and serum electrolyte state. Placebo treatment caused a modest but significant diurnal fall in blood pressure. Captopril produced a significant early (1.5 hours) and brief fall in blood pressure. The blood pressure fall with enalapril was later (4-10 hours), longer lasting, and was associated with significant slowing of supine heart rate. Though perindopril produced a similar plasma ACE inhibition to that produced by enalapril, it only caused changes in blood pressure that were similar to those caused by placebo. CONCLUSIONS--This controlled study is the first to indicate a qualitative difference in the acute response to angiotensin converting enzyme inhibitors with similar structure and metabolism (that is, enalapril and perindopril). Low dose perindopril seems to be less likely to cause hypotension in patients with heart failure. The explanation for the differences is unclear but may reflect differential effects on local tissue angiotensin generation.