OBJECTIVE--To determine whether echocardiography and radionuclide angiography give comparable results when the left ventricular ejection fraction is measured early after myocardial infarction and thus whether, irrespective of the method used, a single value for the ejection fraction could be used as a guide for starting treatment with an angiotensin converting enzyme inhibitor. DESIGN--Prospective comparison of measurement of left ventricular ejection fraction by echocardiography and radionuclide angiography. SETTING--Coronary care units of two university teaching hospitals in Glasgow. PATIENTS--99 patients studied within 36 hours of acute myocardial infarction. OUTCOME MEASURES--Left ventricular ejection fraction assessed by echocardiography and radionuclide angiography. RESULTS--70 (77%) of the 99 patients had ejection fraction measured by both echocardiographic and radionuclide techniques, 30 in centre 1 and 40 in centre 2. In centre 1 the mean difference (SD) in ejection fraction (radionuclide angiography--echocardiography) was -8 (10%); 95% CI -12 to -4%. In centre 2 the mean difference was -14 (11%); 95% CI -17 to -11%. If patients had been treated with an ACE inhibitor on the basis of a radionuclide ejection fraction of < 40% then 93% in centre 1 (28 of 30) and 98% in centre 2 (39 of 40) would have been treated. This compares with 63% (19 of 30) and 50% (20 of 40), respectively if echocardiography had been used as a guide. CONCLUSION--Measurement of ejection fraction is highly dependent on the method used and it is therefore impossible to quote a universally applicable figure for left ventricular ejection fraction below which an ACE inhibitor should be used after myocardial infarction.