Refractory pump failure after cardiopulmonary bypass carries a high mortality. To assess the effectiveness of metabolic support for the heart in patients with refractory heart failure after hypothermic ischemic arrest for aortocoronary bypass surgery, we randomly assigned 22 patients to receive either intravenous glucose (50%), insulin (80 units/I), and potassium (100 meq/l) (GIK) infused at a rate of 1 ml/kg/hr for up to 48 hours or glucose (5%) and NaCl (0.225%) infused at the same rate (control). All patients required inotropic drug support, received intra-aortic balloon pump assistance, and had an initial mean cardiac index (CI) of 2.5 l/min/m2. At 12 and 24 hours, CI had risen significantly in the GIK but not in the control group (3.6 and 3.4 l/min/m2 vs. 2.5 and 2.7 l/min/m2, p less than 0.005). Time on the intra-aortic balloon pump (39 vs. 61 hours) and requirements for inotropic drug support were also significantly less in the GIK compared with the control group. Although the number of patients was relatively small, the GIK group also showed a trend for improved long-term survival: at 60 days after surgery, there were 10 of 11 survivors in the GIK-treated group compared with seven of 11 survivors in the control group. Although the exact mechanism for the beneficial effects of GIK on myocardial contractility remains to be elucidated, we conclude that GIK is safe and effective in the treatment of refractory left ventricular failure after aortocoronary bypass surgery.