A prospective randomised trial compared the immediate haemodynamic effects of intravenous diuretic (frusemide), venodilator (isosorbide dinitrate), arteriolar dilator (hydralazine), and positive inotropic stimulation (prenalterol) as first-line therapy for acute left ventricular (LV) failure following myocardial infarction. Forty-eight patients with transmural myocardial infarction and a pulmonary artery occluded pressure (PAOP) of greater than 20 mm Hg were studied within 18 h of admission to a coronary care unit. Both frusemide (-4 mm Hg; p less than 0.01) and isosorbide dinitrate (-6 mm Hg; p less than 0.01) reduced LV filling pressure without change in cardiac index and heart rate. Although both hydralazine and prenalterol increased cardiac index (p less than 0.01), the reduction in LV filling pressure (-2 mm Hg; p less than 0.05) was less than with frusemide and isosorbide dinitrate, and was associated with an increased heart rate (+8 and +13 beats min-1; p less than 0.01). These data suggest that in acute heart failure following myocardial infarction the four treatment modalities could be ranked in descending order of potential benefit as follows: venodilatation (isosorbide dinitrate)--decrease of LV pressure/work; diuretic therapy (frusemide)--decrease of LV pressure/work offset by a transient pressor effect; arteriolar dilatation (hydralazine)--decrease of LV pressure/work and of PAOP, but offset by tachycardia; and positive inotropic therapy (beta 1-agonist prenalterol)--tachycardia and augmented LV afterload. Combination of the former and latter agents, because of their differing modes of action, should offer haemodynamic advantages over monotherapy and deserves further evaluation.