Methods:

This is a retrospective review of linked data from multiple clinical and administrative databases in Ontario, Canada. The study cohort consisted of patients with diabetes and multivessel coronary artery disease, defined as: ≥70% stenosis of two vessels including the proximal left anterior descending artery (LAD) and either the circumflex (Cx) or right coronary artery (RCA); or of three vessels involving the LAD, Cx, and RCA. Patients underwent revascularization with CABG or PCI within 90 days of index angiography. Exclusion criteria included significant left main disease (although these patients were included in a secondary analysis), patients who had concomitant valve or aortic procedures at the time of CABG, patients who underwent transcatheter aortic valve replacement (TAVR) after PCI, history of prior cardiac surgery, PCI within 6 months prior to angiography, and patients with ST-segment elevation myocardial infarction (STEMI) or cardiogenic shock. Propensity-score matching accounted for baseline differences of 23 variables using multivariable logistic regression with standardized mean difference cut-off of 0.10 for balance. Time-to-event Kaplan-Meier survival analysis and Cox proportional hazards modeling were performed. Primary outcomes were 30-day mortality and long-term all-cause mortality. Secondary outcomes were long-term major adverse cardiac and cerebrovascular events (MACCEs).