Abstract OBJECTIVES The very long-term mortality of off-pump and on-pump coronary artery bypass grafting (CABG) versus percutaneous coronary intervention (PCI) in a randomized complex coronary artery disease population is unknown. This study aims to investigate the impact of on-pump and off-pump CABG versus PCI on 10-year all-cause mortality. METHODS The SYNTAX trial randomized 1800 patients with three-vessel and/or left main coronary artery disease to PCI or CABG and assessed their survival at 10 years. In this sub-study, the hazard of mortality over 10 years was compared according to the technique of revascularization: on-pump CABG (n = 725), off-pump CABG (n = 128) and PCI (n = 903). RESULTS There was substantial inter-site variation in the use of off-pump CABG despite baseline characteristics being largely homogeneous among the 3 groups. The crude rate of mortality was significantly lower following on-pump CABG versus PCI [25.6% vs 28.4%, hazard ratio (HR) 0.79, 95% confidence interval (CI) 0.65–0.96], while it was comparable between off-pump CABG and PCI (28.5% vs 28.4%, HR 0.98, 95% CI 0.69–1.40). After adjusting for the 9 variables included in the SYNTAX score II 2020, 10-year mortality remained significantly lower with on-pump CABG than PCI (HR 0.75 against PCI, P = 0.009). CONCLUSIONS In the SYNTAXES trial, 10-year mortality adjusted for major confounders was significantly lower following on-pump CABG compared to PCI. There was no evidence for unadjusted difference between off-pump CABG and PCI, although the unadjusted estimated HR had a wide CI. Site heterogeneity in the technique used in bypass surgery has had measurable effects on treatment performance.