Objectives: To characterize prescribing of carvedilol and controlled-release (CR) metoprolol for veterans with heart failure who were predominantly elderly (aged > or = 65) and to compare the time to first hospitalization or death.
Design: Retrospective cohort.
Setting: Veterans Health Administration (VHA) nationwide healthcare system.
Subjects: Veterans enrolled in fee-for-service Medicare that had a VHA diagnosis of heart failure and a first prescription for carvedilol (n=17,429) or CR metoprolol (n=8,683) between October 1999 and September 2003.
Measurements: Prescribed daily doses and CR metoprolol-to-carvedilol hazard ratios (HRs) estimated using Cox regression within strata defined according to propensity to prescribe carvedilol.
Results: After several months, the majority of prescribed doses in both cohorts were less than 50% of target doses. The risk-adjusted HR was 0.99 (95% confidence interval=0.96-1.03) for hospitalization or death and 0.91 (0.85-0.96) for death alone.
Conclusion: Doses of beta-blockers being prescribed to predominantly elderly veterans with heart failure were much lower than proven doses. Whether efforts to increase doses can improve the effectiveness of beta-blocker therapy warrants further study. At the prescribed doses, CR metoprolol was associated with better survival than carvedilol, although unmeasured confounding variables might explain the observed difference in survival.