Reconstructed meta-analysis of percutaneous coronary intervention versus coronary artery bypass grafting for left main disease

Abstract

Abstract Background Randomized controlled trials (RCTs) comparing percutaneous coronary intervention (PCI) with drug-eluting stents and coronary artery bypass grafting (CABG) for patients with left main coronary artery disease (LMCAD) have reported conflicting results. Objectives We performed a systematic review from inception to 23 May 2021 and one-stage reconstructed individual-patient data meta-analysis (IPDMA) that included 10-year mortality outcomes. Methods The primary outcome was 10-year all-cause mortality. Secondary outcomes included myocardial infarction (MI), stroke and unplanned revascularization at 5 years. We did IPDMA using published Kaplan-Meier curves to provide individual data points in coordinates and numbers at risk were used to increase the calibration accuracy of the reconstructed data. Shared frailty model or, when proportionality assumptions were not met, a restricted mean survival time model were fitted to compare outcomes between treatment groups. Results Of 583 articles retrieved, 5 RCTs were included. A total of 4595 patients from these 5 RCTs were randomly assigned to PCI (N=2297) or CABG (N=2298). The cumulative 10-year all-cause mortality after PCI and CABG was 12.0% versus 10.6% respectively (HR 1.093, 95% CI: 0.925–1.292; p=0.296). PCI conferred similar time-to-MI (RMST ratio 1.006, 95% CI: 0.992–1.021, p=0.391) and stroke (RMST ratio 1.005, 95% CI: 0.998–1.013, p=0.133) at 5 years. Unplanned revascularization was more frequent following PCI compared with CABG (HR 1.807, 95% CI: 1.524–2.144, p<0.001) at 5 years. Conclusion This meta-analysis using reconstructed participant-level time-to-event data showed no statistically significant difference in cumulative 10-year all-cause mortality between PCI versus CABG in the treatment of LMCAD. Funding Acknowledgement Type of funding sources: None.