The factors associated with the development of pump failure in patients with acute myocardial infarction (MI) were studied in 475 patients admitted to our coronary care unit (CCU) and a clinical and hemodynamic re-evaluation of vasodilator and catecholamine therapy was carried out in patients with pump failure. The severity of pump failure as represented by Killip's classification was closely related to left ventricular (LV) dysfunction as judged by LV ejection fraction, number of previous MIs and peak creatinine phosphokinase (CPK) value, and the severity of coronary atherosclerotic lesions as assessed by coronary arteriography. Retrospective hemodynamic analysis of vasodilator therapy (intravenous nitroglycerin, nitroprusside and phentolamine) in 94 patients with pump failure revealed that a significant decrease in pulmonary capillary wedge pressure (PCW) with no significant change in cardiac index (CI) resulted in a clinical improvement in pulmonary congestion in Forrester's hemodynamic subset 2. However, in patients in subsets 3 and 4, the slight increase in CI induced by vasodilator therapy was not enough to restore normal systemic perfusion in spite of a significant decrease in PCW and total systemic resistance; catecholamine was far more effective in increasing CI. Thus, the major indication for vasodilator therapy is in patients of subset 2, while combined use with catecholamine is preferable in those with low cardiac output.