In the landmark BARI trial, diabetics assigned to CABG had 57.8% (10-year survival) compared to 45.5% with angioplasty. Angioplasty had substantially higher repeat revascularization rates as well [1]. In contrast, meta-analysis of the six CABG versus PCI RCTs with reported diabetic subgroups shows no survival advantage of CABG over PCI at 10 years [2]. With bare metal stenting (BMS), the ARTSI randomized trial versus CABG for multivessel disease is typical. Although there was no significant difference in mortality, CABG outperformed BMS particularly in diabetics because of the increased frequency of subsequent revascularization procedures with stenting [3]. Restenosis, of course, has been the limiting Achille's heel of balloon angioplasty and bare metal stents in diabetics. When feasible at reasonable risk, therefore, CABG has remained the gold standard for diabetics with significant multi-vessel CAD.Should DES change the revascularization landscape for diabetics? The diabetics in TAXUS IV, SIRIUS, DECODE, and DIABETES RCTS had single digit target lesion revascularization rates (7.8%) with DES versus 24.7% with BMS [4]. The nonrandomized comparison of CABG and DES in ARTSII demonstrated equipoise in diabetic patients [5]. The NHLBI dynamic registry confirms continued improvement in outcomes for diabetics with PCI [6].But can DES be advocated for multivessel disease diabetics instead of CABG? Attempting to adjust for baseline differences, retrospective databases of DES versus CABG for multivessel CAD have conflicting results. The Washington Hospital Center reports lower adjusted risk of major adverse cardiac and cerebrovascular events (MACCE) in diabetics with two and three vessel CAD treated with CABG than PCI [7]. In 6,061 diabetics with multivessel CAD undergoing DES or CABG from October 2003 to December 2004 in New York, adjusted mortality rates at 18 months were not significantly different [8]. However, CABG was associated with a lower risk of combined death or myocardial infarction as well as reduced subsequent revascularization rates.At 12-month follow-up, the SYNTAX RCT demonstrates similar rates of mortality and MI among the 227 diabetics randomized to multivessel Taxus stenting versus 204 assigned to CABG, but 20.3% subsequent revascularization procedures with an initial strategy of multivessel DES versus 6.4% with CABG (p < 0.001). While awaiting longer term SYNTAX follow-up and the results of the larger FREEDOM RCT of multivessel DES versus CABG in diabetics, we must recall another important observation from the landmark BARI trial. In the BARI registry of patients where specific factors mitigated against randomization, the outcomes of CABG and angioplasty were similar [9]. In this issue of CCI, Tarantini et al.'s experience in 220 consecutive diabetic patients with multivessel CAD provides a microcosm for the ''right answer.'' The decision to perform PCI versus CABG revascularization of a DM patient with multivessel CAD is a complex calculus. Important factors including diffuseness of distal disease, ...