Diabetes is a well-recognized independent risk factor for mortality due to coronary artery disease. When diabetic patients need cardiac surgery, either coronary-aortic by-pass (CABP) or valve operations (VO), the presence of diabetes represents an additional risk factor for these major surgical procedures. Because of controversial data on mortality rates and post-operative complications in diabetic patients, probably due to not exactly comparable groups of patients, this retrospective study aimed to compare two homogeneous populations, which were different only for the presence or absence of diabetes. We studied 700 patients undergoing cardiac surgery: 350 with and 350 without diabetes, mean age 62 +/- 9 years (67% males); 441 underwent CABP and 259 VO. Apart from the diabetes, the two groups were strictly matched for age, body mass index, concomitant pathologies and smoking habits, except for previous neurological injuries (more frequent in diabetic patients), and for a slightly lower ejection fraction in the diabetic group. Intra- and post-operative complications or events were evaluated carefully: death, number staying in post-operative intensive care unit (ICU), renal, hepatic and respiratory complications, necessity for reoperation and hemotransfusions. Anesthesia and surgical procedures (including extra-corporeal circulation techniques) remained substantially unchanged over the period of recruitment of patients (1996-1998) and applied equally to both groups of patients. All diabetic patients were treated with insulin by using standard procedures in order to optimize metabolic control. Diabetic patients in our study, did not show higher rates of mortality in comparison with non-diabetic patients, but had more total neurological complications, more renal complications, a higher re-opening rate, more prolonged ICU stay, and they needed more blood transfusions. Diabetes remains an independent risk factor for these events even in a multivariate logistic regression model analysis. In the subgroup of diabetic patients who underwent CABP a higher rate of renal dysfunction, re-opening, need for hemotransfusions and prolonged ICU stay were confirmed. In the subgroup of diabetic patients undergoing VO we found a higher rate of renal dysfunction, reopening, prolonged ICU stay and major lung complications. In conclusion, diabetes does not seem to increase the mortality rates of cardiac surgery, but diabetic patients undergoing CABP have, on the basis of the relative risk evaluation, a 5-fold risk for renal complications, a 3.5-fold risk for neurological dysfunction, a double risk of being hemotransfused, reoperated or being kept 3 or more days in the ICU in comparison with non-diabetic patients. Moreover, diabetic patients undergoing VO have a 5-fold risk of being affected by major lung complications.