To The Editor In a meta-analysis from 6 randomized clinical trials (RCTs), Sipahi et al1 reported 4.1 years’ outcome of 6055 patients randomized to percutaneous coronary interventions (PCIs) either with bare-metal stents (BMS) or drug-eluting stents (DES) vs coronary artery bypass graft (CABG) in multiple vessel disease. The authors concluded that CABG leads to a significant reduction in long-term mortality, myocardial infarction, and repeat revascularization procedures independently of diabetic status without differences in nonfatal stroke. Trial selection was made according to the number of stents and arterial conduits used in PCI/CABG arms. However, in the Estudio Randomizado Argentino Angioplastia vs Cirugia (ERACI) II trial, which was excluded,2 BMS were used in all patients and arterial conduits in 88.5% of patients, which is not significantly different from other BMS and CABG trials.1 In ERACI II, a lower incidence of death and myocardial infarction with PCI was observed in the first year and an advantage remained at 5 years. A meta-analysis of the 4 RCT comparing BMS vs CABG3 found almost identical survival (8.5% and 8.2%, respectively; P = .74); furthermore, in patients with diabetes, the survival for a composite of death, MI, and stroke were similar (21.4% [BMS] and 20.9% [CABG]; P = .79). There have been 8 RCTs that compared PCI either with BMS or DES vs CABG—2 of them were not included2,4 and in 1 trial the authors reported incomplete and underpowered data.1 If we pool the results from the 8 RCTs (7468 patients), mortality between PCI and CABG in a nondiabetic population remains similar (risk ratio [RR], 1.03; 95% CI, 0.77-1.37) and the only survival advantage with CABG is in a diabetic population (RR, 0.67; 95% CI, 0.52-0.86), despite an increased risk of nonfatal stroke (RR, 1.72; 95% CI, 1.18-2.53).