Propensity analysis of long-term survival after surgical or percutaneous revascularization in patients with multivessel coronary artery disease and high-risk features

Background - Although most randomized clinical trials have suggested that long-term survival rates after percutaneous coronary intervention (PCI) or surgical multivessel coronary revascularization (CABG) are equivalent, some post hoc analyses in high-risk groups and adjustment for severity of coronary disease have suggested higher mortality after PCI. Methods and Results - We studied 6033 consecutive patients who underwent revascularization in the late 1990s. PCI was performed in 872 patients; 5161 underwent CABG. Half the patients had significant left ventricular dysfunction or diabetes. Propensity analysis to predict the probability of undergoing PCI according to 22 variables and their interactions was used. The C-statistic for this model was 0.90, indicating excellent discrimination between treatments. There were 931 deaths during 5 years of follow-up. The 1- and 5-year unadjusted mortality rates were 5% and 16% for PCI and 4% and 14% for CABG (unadjusted hazard ratio, 1.13; 95% CI, 1.0 to 1.4; P = 0.07). PCI was associated with an increased risk of death (propensity-adjusted hazard ratio, 2.3; 95% CI, 1.9 to 2.9; P < 0.0001). This difference was observed across all categories of propensity for PCI and in patients with diabetes or left ventricular dysfunction. Other independent predictors of mortality (P <= 0.01 for all) were renal dysfunction, age, diabetes mellitus, chronic lung disease, peripheral vascular disease, left main trunk stenosis, and extent of coronary disease (Duke angiographic score). Conclusions - In patients with multivessel coronary artery disease and many high-risk characteristics, CABG was associated with better survival than PCI after adjustment for risk profile.