OBJECTIVE The goal of this study was to examine the metabolic and hemodynamic effects of a glucose-insulin-potassium infusion in elective coronary surgery, when blood cardioplegia was used for cardiac protection. METHODS A prospective, randomized, open, clinical comparison was performed between 2 perioperative infusion regimens in 40 elective nondiabetic coronary artery bypass graft (CABG) surgery patients. METHODS 20 patients (glucose-insulin-potassium-GIK-group) received glucose 0.2 g/kg/h, insulin 0.12 U/kg/h, potassium 0.15, magnesium 0.032 and phosphate 0.024 mmol/kg/h from anesthesia induction to the start of bypass, when infusion rate was reduced to 30%, and after bypass increased to 50% of the initial rate. The infusion was continued until the first postoperative morning. Another 20 patients (control-R-group) received glucose 0.05 g/kg/h, potassium 0.075, magnesium 0.016 and phosphate 0.008 mmol/kg/h from the end of bypass to the next morning. Pump prime was glucose-free and a blood cardioplegia technique was used for cardiac protection. RESULTS The GIK patients needed less dopamine support in the intensive care unit (ICU) (p < 0.05). No difference was found between the groups with regard to myocardial injury, the MB-fractions of serum creatine kinase (CK-MB) being elevated to a similar degree in both groups. Likewise there were no significant differences in hemodynamic changes or duration of ICU stay. Although the glucose infusion was continued during bypass in the GIK patients, there was a considerable risk of hypoglycemia (due to insulin and hemodilution) after the onset of bypass: in 5 GIK patients (25%; 95% confidence interval 8.7 to 49.1%) blood glucose was less than 2 mmol/l. However, the hypoglycemia was of short duration and no detrimental effects were seen. CONCLUSIONS Perioperative GIK infusion entailed a slight decrease in the postoperative need for dopamine support, but was connected with a considerable risk of hypoglycemia.