1103reduces the second, and evidently has little effect on the third: thus there is no major penalty to the heart in terms of oxygen uptake for the increased flow that is generated. With the exception of other vasodilators most agents capable of augmenting cardiac output do so at greater metabolic cost.As with most clinical studies, some compromises were made in design to safeguard the overriding interests of the patients. We were anxious to establish the most appropriate treatment for each patient with little delay and were therefore unable to return to control conditions after each infusion rate to distinguish drug effect from possible spontaneous improvement. We thought it justifiable, however, to interrupt treatment once after the 40 ,tg/min dose regimen in eight of our 11 patients to show that the haemodynamic variables returned to the pretreatment values; only mean arterial pressure remained appreciably different from control readings in most patients. We chose to use pulmonary artery end-diastolic pressure as a measure of left ventricular filling pressure to avoid repeatedly wedging the Swan-Ganz catheter, since frequent changes in position or balloon inflation carry small risks. Though the measurement tends to underestimate true left atrial pressure in the presence of heart failure,15 the correlation between the two indices is close and changes are faithfully reflected.Salbutamol has little effect on "backward failure" as shown by the increased pulmonary artery end-diastolic pressure in our patients and therefore cannot be recommended for the treatment of pulmonary oedema in acute myocardial infarction.It is not useful if an increase in blood pressure is believed to be essential. Our results suggest, however, that salbutamol is a useful and safe method of increasing cardiac output in those patients with myocardial infarction in whom poor perfusion is the most important haemodynamic disturbance. Circulation, 1978, 58, 466. 7Walinsky, P, et al, American3Journal of Cardiology, 1974, 33, 37. 8 Franciosa, J A, et al, Lancet, 1972