This study was undertaken to more clearly define the operative risks and appropriate strategies for the selection of patients who might be candidates for bilateral internal mammary artery grafting. A review of the 674 patients who underwent this procedure was performed over a 10-year period from January 1982 through 1991. These patients represented 5% of the 12,824 patients who underwent bypass grafting during this period. The mean patient age was 54 +/- 9.4 years, and 92% of the patient were male. Diabetes was present in 17.5%, with 24% of this group insulin dependent. The ejection fraction was less than 0.50 in 26.7% and 14% had had previous bypass procedures. Hospital mortality was 1.9% (n = 13). Hospital morbidity included wound infection in 3.6%, reoperation for bleeding in 2.1%, and stroke in 1.6%. Postoperative intraaortic balloon pump support was necessary in 2.8%. Univariate and multivariate analysis revealed that advanced age, an emergent operative status, and the number of diseased vessels (especially left main obstruction) were predictors of hospital death. Except for wound infection (9.3% versus 2.5%) and length of hospital stay (10.8 +/- 11.8 days versus 8.8 +/- 7.6 days), the complications in diabetics were similar to those in patients without diabetes, respectively. At 5 years, freedom from death, death or infarctions, and death, myocardial infarction, or coronary reoperation was 90%, 70%, and 62%, respectively. The only significant multivariate correlates of long-term survival were diabetes and reoperative surgery.