Five-Year Outcomes of PCI and CABG for Multivessel Disease: A National Population-Based Study of Regional Practice

Objectives To study how the regional preferences for less invasive multivessel coronary revascularisation would adversely affect long-term clinical outcomes. Design A national retrospective cohort study utilising instrumental variable analysis to estimate the causal effect of revascularisation strategy. Settings England, using national linked Hospital Episode Statistics (HES) with office of national statistic mortality data from 2007–2020. Participants The analysis included 173,771 individuals with complete 5-year follow-up who underwent multivessel revascularisation for coronary artery disease. Of this cohort, 63,189 (36.4%) received PCI and 110,582 (63.6%) received CABG. In total, 37,894 (21.8%) participants were female, and 153,048 (88.2%) were of White ethnicity. Interventions The exposure was the preference between multivessel percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG). The regional ratio of CABG to PCI procedures was used as the instrumental variable. Main outcome measures The primary outcome was all-cause mortality, assessed in-hospital and up to 5 years post-procedure. Results All-cause mortality was 2.1% (n=3,587) in-hospital and 16.4% (n=28,474) at 5 years. The proportion of patients undergoing CABG varied significantly across regions (25.4%–82.3%), demonstrating validity as an instrumental variable. In the primary analysis, CABG was associated with higher in-hospital all-cause mortality versus PCI (ATE, 1.1%; 95% CI, 0.6% to 1.6%) but lower 5-year all-cause mortality (ATE, -5.4%; 95% CI, -7.0% to -3.7%). Adjusted hazard ratios stratified by quartiles of regional CABG: PCI ratios showed an increase in in-hospital mortality but a decrease in 5-year mortality as the proportion of CABG increased. Conclusion Regional preferences for revascularisation with multivessel PCI result in lower in-hospital all-cause mortality, a key quality metric, but worse long-term outcomes for individuals with multivessel coronary artery disease.