To evaluate the effect of an assist pump on the metabolic viability of the ischemic myocardium, myocardial pH was continuously monitored using an ion-sensitive field-effect transistor (ISFET) pH sensor in 14 dogs after coronary occlusion. In seven dogs (control group), coronary occlusion (10-20 min) and successive reperfusion (30-60 min) were performed several times. In seven dogs [left ventricular assist device (LVAD) group], the LVAD was implanted between the left atrium and the aorta. Occlusion and reperfusion were performed first with the pump on and then with the pump off. In both groups, myocardial pH fell after occlusion, and increased after reperfusion. In the control group, the fall rate of pH in the later coronary occlusion decreased to 66 +/- 7% (mean +/- SEM) of that in the previous occlusion. Contrarily, in the LVAD group, the fall rate under LVAD-off increased to 174 +/- 32% of that under the preceding LVAD-on. This indicates that progressive cellular damage occurred in the control group, while the myocardium was preserved by the assist pump in the LVAD group. LVAD is effective for preserving the metabolic viability of the ischemic myocardium.