Diabetes mellitus is a well established risk factor for coronary artery disease, but its effect on the conduct and results of bypass surgery (CABG) are not well documented. We identified 261 (11.9%) diabetics among 2192 patients operated August, 1972, through December, 1977. All patients were surveyed, in identical manner, during January, 1981. There were 51 (19.5%) dead, seven (3.3%) lost, and seven (3.3%) residing in foreign countries whom we did not attempt to follow. We divided the 261 patients into three groups by severity of diabetes: drug therapy 106 (40.6%), diet therapy 60 (23.0%), and borderline 95 (36.4%). Severity of diabetes had no effect on any of the factors we investigated. Compared with nondiabetics, diabetics had the same average age, a higher proportion of women, the same number of grafts per patient, slightly more bad ventricles, and a higher surgical and late mortality. Frequency of preoperative angina was slightly higher in diabetics. Relief of angina was essentially the same in both groups. There were 54 serious complications which occurred in 45 (17.2%) patients; of these, 20 (7.7%) were hospital mortalities. The only factors reaching statistical significance were the proportion of women and the surgical mortality in good ventricles. We conclude that the presence of diabetes does increase the morbidity and mortality associated with bypass surgery, but only to a relatively small degree and that controlled diabetes is not sufficient reason to avoid bypass surgery.