The internal mammary artery is frequently used as a coronary artery bypass graft conduit because of superior long-term patency. The purpose of this study was to determine if there was also an advantage to the internal mammary artery in terms of operative mortality and morbidity. The Department of Veterans Affairs Cardiac Surgery Database was reviewed for two separate time periods, April 1987 through March 1989 and October 1990 through September 1991. During these periods, 14,172 patients underwent coronary artery bypass grafting. Univariate and multivariate logistic regression analyses were used to determine preoperative variables predictive of operative mortality and morbidity, with the independent variable of use of the internal mammary artery added to previously determined indicators. The total group was analyzed in risk quartiles according to expected mortality. Univariate analysis revealed an operative mortality of 6.8% in the early period and 6.5% in the latter period for the saphenous vein groups compared with 3.7% and 3.2%, respectively, for the internal mammary artery groups (p = 0.000). Multivariate analysis revealed an odds ratio of operative death with use of the internal mammary artery graft versus use of vein grafts of 0.78 during the first period and 0.72 during the second period, reductions of 22% and 28%, respectively. There were 29% and 37% reductions in the odds of operative mortality in the highest-risk quartile group of patients using the internal mammary artery graft. The odds ratio of developing mediastinitis with one internal mammary artery graft was 1.84 (p < 0.01) in the first time period and 1.11 in the second time period (p = not significant). However, with multiple mammary bypass grafts, the odds ratios were 3.70 (p < 0.01) and 2.96 (p < 0.01) in the respective time periods. On the basis of this study, it is concluded that internal mammary artery grafts in addition to providing superior long-term patency also decreased operative mortality after adjustment for patient risk factors. Use of the mammary artery does not consistently increase operative complications with the exception of mediastinitis when both internal mammary arteries are employed.