Ultra long-term outcomes after percutaneous coronary intervention versus coronary artery bypass grafting in complex coronary artery disease: a reconstructed individual patient data meta-analysis

Background The optimal revascularisation strategy for complex coronary artery disease (CAD) remains debated, particularly in the context of rapidly evolving stent technology, intravascular imaging, physiologic assessment and pharmacotherapy. Randomised trials comparing percutaneous coronary intervention (PCI) to coronary artery bypass grafting (CABG) span multiple technological eras, limiting direct interpretation of long-term outcomes across generations of stents. Purpose To evaluate long-term survival after PCI versus CABG across stent generations using a large-scale reconstructed individual patient data (IPD) approach derived from eight published randomised trials. Methods Kaplan-Meier survival curves from landmark randomised trials (FREEDOM, SYNTAX, VA-CARDS, PRECOMBAT, BEST, EXCEL, NOBLE and FAME3) comparing PCI and CABG were digitised directly from published figures using image-based analytical reconstruction. Time-to-event data and numbers at risk were extracted to generate pseudo-IPD, enabling pooled survival analyses with long-term follow-up (Figure). Cox proportional hazards models stratified by trial were used to estimate hazard ratios (HR) for PCI versus CABG in the total cohort as well as according to stent generation (first-generation versus second-generation or later stents) and degree of CAD. Results The reconstructed database represents the largest aggregated long-term survival dataset comparing PCI and CABG derived from randomised evidence, including 9,967 patients with complex CAD with follow-up extending to 10 years. In the total cohort, PCI was associated with higher mortality compared with CABG (HR 1.22, 95% confidence interval [CI] 1.09-1.36; p<0.001). When stratified by technological era, first-generation stents demonstrated worse outcomes compared to CABG (HR 1.25, 95% CI 1.09-1.44; p=0.002). In contrast, outcomes with second- and later-generation stents showed equipoise to CABG, with no significant difference in long-term mortality (HR 1.11, 95% CI 0.93-1.33; p=0.26) (Figure). When stratified by coronary anatomy, PCI was associated with higher long-term mortality compared with CABG in patients with three-vessel disease (HR 1.23, 95% CI 1.09-1.38; p=0.001), whereas no significant difference was observed in left main disease (HR 1.18, 95% CI 0.91-1.53; p=0.21). Conclusion This study provides the largest long-term comparative randomised assessment of PCI and CABG across technological eras. While early-generation stents were associated with inferior survival compared with CABG, contemporary PCI demonstrates neutral long-term outcomes, suggesting technological evolution has fundamentally narrowed the historical survival gap. RCTs are urgently required to compare state-of-the-art PCI including imaging and physiologic assessment to CABG to potentially redefine guidelines.