Long-term impact of drug-eluting stents versus bare-metal stents on all-cause mortality.

Abstract

Our purpose was to examine the incidence of all-cause mortality among drug-eluting stents (DES) and bare-metal stents (BMS) while adjusting for many confounding factors generally not considered in prior studies. DES use in the U.S. declined by up to 50% in recent years, primarily due to concerns about late stent thrombosis and possibly increased mortality. However, recent data suggest that DES are as safe as BMS and may actually be associated with a lower incidence of myocardial infarction and mortality. All patients undergoing percutaneous coronary intervention with a DES or BMS alone from March 1, 2003, to June 30, 2007, at a tertiary care center were assessed. Multivariable Cox proportional hazards modeling was performed for overall and propensity-matched patients. Socioeconomic status was calculated using U.S. Census 2000 data. The primary end point was all-cause mortality. There were a total of 832 deaths over a 4.5-year interval among 8,032 patients. Of these, 6,053 received a DES and 1,983 patients had a BMS. All-cause mortality was significantly lower in unadjusted and adjusted Cox proportional models with DES (hazard ratio: 0.62, 95% confidence interval: 0.53 to 0.73; p < 0.001). Similarly, in the propensity-matched group, DES remained associated with lower mortality compared with BMS (adjusted hazard ratio: 0.54, 95% confidence interval: 0.45 to 0.66; p < 0.001). DES were associated with lower mortality in this "real-world" setting. However, despite multiple adjustments, potential confounding may still play a role.