Warm heart surgery has documented myocardial protection benefit, but with an added neurologic threat. It is hypothesized that moderately hypothermic aerobic heart surgery will maintain the myocardial protection and reduce neurologic risk. This study compared 493 patients undergoing coronary artery bypass graft operations with normothermic (35 degrees to 37% degrees C) continuous blood cardioplegia and normothermic perfusion to 379 coronary artery bypass grafting patients with hypothermic (33 degrees to 29 degrees C) continuous blood cardioplegia and hypothermic perfusion to test this hypothesis. There was no difference in age, sex, prior myocardial infarction, hypertension, prior neurologic event, congestive failure, or diabetes. The hypothermic group had more reoperations (24% versus 14%; p = 0.0002), class III/IV angina (83% versus 71%; p = 0.002), a trend to more triple-vessel (54% versus 47%; p = 0.10) and left main disease (18% versus 14%; p = 0.10), lower ejection fractions (0.52 +/- 0.15 versus 0.55 +/- 0.13), more grafts placed (3.6 +/- 1.1 versus 3.4 +/- 1.1; p = 0.04), but fewer internal mammary arteries (62% versus 78%; p < 0.0001). Postoperative myocardial infarction rate was 1.2% in the hypothermic group and 1.3% in the normothermic group (p = not significant). Intraaortic balloon pump requirement was 3.4% with hypothermic and 1.4% with normothermic groups (p = 0.05). The incidence of postoperative neurologic events was significantly higher in the normothermic group (4.7% versus 1.8%; p = 0.038). The multivariate correlates of stroke were older age and normothermic cardioplegia, whereas the only multivariate correlate of death was older age.(ABSTRACT TRUNCATED AT 250 WORDS)