OBJECTIVE--To determine the cardiac, renal, and neuroendocrine effects of lisinopril in men with untreated, symptom free left ventricular systolic dysfunction. DESIGN--A randomised, double blind cross over trial with six week treatment periods to compare lisinopril (10 mg/day) and matching placebo. SETTING--Hospital outpatient department. PATIENTS--Patients with pronounced systolic dysfunction on cross sectional echocardiography due to myocardial infarction at least six months previously, without angina and with no or minimal breathlessness. Eighteen men were identified of whom 15 completed the study. INTERVENTIONS--Lisinopril (10 mg) or placebo given once daily by mouth. MAIN OUTCOME MEASURES--Primary: oxygen consumption at peak exercise. Secondary: resting cardiac function as measured by radionuclide ventriculography and echocardiography, renal function estimated radioisotopically, and plasma indices of neuroendocrine activity. RESULTS--Compared with placebo, lisinopril increased (mean (SD)) peak oxygen consumption during exercise (19.8(3.1) ml/kg/min v 21.4(3.2) ml/kg/min; p < 0.003). Lisinopril did not improve indices of cardiac function at rest. It reduced plasma concentrations of angiotensin II (median values 7 pg/ml to 5 pg/ml; p < 0.02), aldosterone (median values 113 pg/ml to 66 pg/ml; p < 0.05) and atrial natriuretic peptide (median values 69 pg/ml to 40 pg/ml; p < 0.04), but noradrenaline and antidiuretic hormone concentrations did not change. Renal blood flow increased and glomerular filtration rate declined. CONCLUSIONS--Even before the onset of heart failure lisinopril improves the cardiopulmonary response to exercise in patients with systolic ventricular dysfunction.