To the Editor We read with great interest the meta-analysis by Sipahi et al1 in which 6 randomized clinical trials (RCTs) of percutaneous coronary intervention (PCI) vs coronary artery bypass grafting (CABG) in multivessel disease (MVD) were included. There was a significant reduction in 1-year-or-greater all-cause mortality with CABG compared with PCI (risk ratio [RR], 0.73; 95% CI, 0.62-0.86; P < .001). The design of included RCTs, however, is heterogeneous: ie, using exclusively bare-metal stents (BMS) in 3 trials, drug-eluting stents (DES) in 2 trials (SYNTAX [Synergy Between Percutaneous Coronary Intervention With Taxus and Cardiac Surgery] multivessel and FREEDOM [Future Revascularization Evaluation in Patients With Diabetes Mellitus: Optimal Management of Multivessel Disease] [references 12 and 16, respectively, in the original article1]), and both BMS and DES in 1 trial (CARDia [Coronary Artery Revascularization in Diabetes] [reference 7 in the original article1]). The authors stated that their aim was to “overcome the power limitation” of the existing data sets.1(p224) Pooling heterogeneous-design RCTs just merely to overcome the power limitation (thereby, probably expecting statistically significant results), however, would be inappropriate.
Takagi H, Umemoto T, for the ALICE Group. Coronary Artery Bypass Grafting vs Percutaneous Coronary Intervention in Multivessel Disease. JAMA Intern Med. 2014;174(6):1006-1007. doi:10.1001/jamainternmed.2014.773