Late effects of radial artery vs saphenous vein grafting for multivessel coronary bypass surgery in diabetics: a propensity-matched analysis.

OBJECTIVES To determine whether the use of the radial artery (RA) vs the saphenous vein (SV) as the second grafting conduit with the internal thoracic artery (ITA) confers a late-survival advantage in diabetes mellitus (DM). METHODS We reviewed our 1996-2007 DM coronary artery bypass grafting (CABG) experience. Study patients (N = 2281) included all primary, non-salvage multigraft CABG discharged alive and receiving ≥1 ITA graft. Bilateral ITA, ITA-only grafts or concomitant valve/aortic surgery patients were excluded. A non-parsimonious, RA use propensity model (42 variables) was derived excluding five factors [gender, vessel disease, insulin, renal failure and left ventricular (LV) dysfunction] that were always strictly matched for all pairs. Greedy matching resulted in well-matched ITA/RA and ITA/SV cohorts (N = 578 each). The late follow-up was truncated at 16 years, and survival comparisons were done by Kaplan-Meier analysis. RESULTS RA grafting was used in 933 (41%) DM patients and was more frequent for non-insulin (513/1348; 49.1%) compared with insulin (271/784; 34.6%) dependent patients. Relatively fewer insulin ITA/RA (169; 62%) could be matched vs non-insulin (469; 71%). Late survival was significantly better for ITA/RA overall [hazard ratio, HR (95% confidence interval) = 0.78 (0.65-0.95); P = 0.012], but this was primarily due to the non-insulin sub-cohort [HR = 0.72 (0.56-0.91); P = 0.007] as opposed to no effect for insulin [HR = 0.92 (0.68-1.26); P = 0.61]. Sub-cohort analysis revealed a significant ITA/RA survival advantage in males, preserved LV function and three-vessel disease. No sub-cohorts were associated with superior ITA/SV survival. CONCLUSION Our analysis indicated that RA grafting confers a significant late-survival advantage and, thus, supports its liberal use in DM patients undergoing multivessel CABG.