BACKGROUND: This study sought to determine whether tight glycemic control with a modified glucose-insulin-potassium (GIK) solution in diabetic coronary artery bypass graft (CABG) patients would improve perioperative outcomes. METHODS AND RESULTS: One hundred forty-one diabetic patients undergoing CABG were prospectively randomized to tight glycemic control (serum glucose, 125 to 200 mg/dL) with GIK or standard therapy (serum glucose <250 mg/dL) using intermittent subcutaneous insulin beginning before anesthesia and continuing for 12 hours after surgery. GIK patients had lower serum glucose levels (138+/-4 versus 260+/-6 mg/dL; P<0.0001), a lower incidence of atrial fibrillation (16.6% versus 42%; P=0.0017), and a shorter postoperative length of stay (6.5+/-0.1 versus 9.2+/-0.3 days; P=0.003). GIK patients also showed a survival advantage over the initial 2 years after surgery (P=0.04) and decreased episodes of recurrent ischemia (5% versus 19%; P=0.01) and developed fewer recurrent wound infections (1% versus 10%, P=0.03). CONCLUSIONS: Tight glycemic control with GIK in diabetic CABG patients improves perioperative outcomes, enhances survival, and decreases the incidence of ischemic events and wound complications.