In diabetics with coronary artery disease (CAD), there remains uncertainty as to whether revascularization by percutaneous transluminal coronary angioplasty (PTCA) or coronary artery bypass surgery (CABG) is preferable. To address this, 4-year mortality and level of pre- and postrevascularization angiographic CAD (measured by a series of coronary scores) were compared between both diabetics and nondiabetics and between revascularization modes in the Coronary Angioplasty versus Bypass Revascularization Investigation population as a whole, and then substratified by diabetic status and then by procedure to which they were randomized. The 1,054 randomized subjects contained 125 diabetics (11.9%) who had significantly greater mortality than nondiabetics (RR 2.19, p = 0.001). Among diabetics or nondiabetics, there was no significant mortality difference between those randomized to PTCA versus those to CABG. Diabetics randomized to PTCA and those to CABG had higher mortalities than respective nondiabetics; the association reached significance only in the former (RR 2.41, p = 0.002). All subgroups had similar prerevascularization CAD. Postrevascularization residual CAD was consistently significantly greater in PTCA than in respective CABG subgroups. Most measurements of CAD were greater in diabetic than in nondiabetic subgroups, but none was significant. In the Coronary Angioplasty versus Bypass Revascularization Investigation, diabetics had double the mortality of nondiabetics; this difference was statistically significant both for the entire population and for those randomized to PTCA, but not for those randomized to CABG. Among diabetics or nondiabetics, there was no significant mortality difference between PTCA and CABG. The higher diabetic mortality was more likely related to more rapid disease progression than to greater postrevascularization disease.