Comparative effectiveness of coronary artery bypass grafting (CABG) surgery and percutaneous coronary intervention (PCI) in elderly patients with diabetes.

OBJECTIVE: To compare the relative effectiveness of coronary artery bypass grafting (CABG) and percutaneous coronary intervention (PCI) among elderly patients with diabetes regarding acute myocardial infarction (AMI), stroke, repeat revascularization, and all-cause mortality.

METHODS: A retrospective cohort study was conducted using the 2006-2008 5% national sample of Medicare claims data. Elderly (≥65 years) beneficiaries with at least two claims of diabetes separated by ≥30 days and who had at least one inpatient claim for multi-vessel CABG or PCI between 1 July 2006 and 30 June 2008 were identified. The date of beneficiary's first CABG or PCI was defined as the index date. All patients were followed from the index date to 31 December 2008 for outcomes. CABG and PCI patients were 1:1 matched on propensity scores and index dates. Cox proportional hazards models were used to compare postoperative outcomes between patients undergoing CABG versus PCI.

RESULTS: The matched sample consisted of 4430 patients (2215 in each group). The Cox proportional hazards models showed that, compared to patients undergoing PCI, CABG was associated with a lower risk of postoperative AMI (hazard ratio [HR]: 0.494; 95% CI: 0.396-0.616; p < .0001), repeat revascularization (HR: 0.194; 95% CI: 0.149-0.252; p < .0001), the composite outcome (HR: 0.523; 95% CI: 0.460-0.595; p < .0001), and all-cause mortality (HR: 0.775; 95% CI: 0.658-0.914; p = .0024); postoperative risk of stroke was not significantly different between the two groups (HR: 0.965; 95% CI: 0.812-1.148; p = .691).

CONCLUSIONS: CABG appears to be the preferred revascularization strategy for elderly patients with diabetes and coronary heart disease. However, this result should be interpreted considering study limitations, for example, several patient clinical variables and physician-related factors which may affect procedure outcomes are not available in the data. Clinical decisions should be individualized considering all patient- and physician-related factors.