Non-selective beta-blockers decrease thrombotic events in patients with heart failure

Background: Beta-blockers are often prescribed to patients with heart failure (HF) without distinctions between types of beta-blockers. The 2002 COMET study showed superiority of carvedilol (a non-selective beta-blocker) over metoprolol (selective beta-blocker) on mortality and cardiovascular events in patients with HF. However, this study was criticised for several reasons. Laboratory findings suggest a reduced prothrombotic response upon sympathetic activation by non-selective beta-blockers. We therefore hypothesised that non-selective beta-blockers reduce vascular events compared to selective beta-blockers in patients with HF. Methods: Data were obtained from the PHARMO Record Linkage System, a population-based registry of pharmacy records linked with hospital discharge records in The Netherlands. We identified a cohort of 20,870 patients with documented HF in the period 1998 to 2007, based on hospital discharge diagnosis. This method has been validated for selecting HF patients. We used Cox regression analysis, with time varying beta-blocker covariate to assess the difference in the incidence of thromboembolic events (acute coronary syndrome (ACS), stroke, or pulmonary embolism) between patients using selective and non-selective beta-blockers. Results: Median follow-up was 2.0 years (interquartile range (IQR): 0.7-4.1). Directly after discharge, 6,980 patients were prescribed a selective beta-blocker and 2,504 patients a non-selective beta-blocker. Total follow-up was 56,667 person-years, of which 18,245 person-years for selective beta-blockers and 6,455 for non-selective beta-blockers. The hazard ratio (HR) for any thrombotic event for non-selective beta-blockers compared to selective beta-blockers was 0.76 (95% confidence interval (CI): 0.64-0.89). After adjustment for potential confounders the difference remained significant (HR 0.84, 95%CI: 0.72-0.99). Conclusion: Non-selective beta-blockers are associated with a lower risk of thromboembolic events compared to selective beta-blockers in patients with HF. The hypothesis that non-selective beta-blockers reduce the prothrombotic state in these patients should be further explored.