Frequency and impact of bleeding in elective coronary stent clinical trials--utility of three commonly used definitions.

ABSTRACT

BACKGROUND: Bleeding events are common after percutaneous coronary intervention (PCI) and have been shown to increase mortality in studies of acute coronary syndrome (ACS) and anti-thrombotic therapy. Despite this evidence, bleeding has not been included as a traditional major endpoint in clinical trials of low-risk populations enrolled in PCI clinical trials. Thus, the impact of specific bleeding definitions has not been evaluated fully among these patients. METHODS AND RESULTS: Using patient-level pooled data from sirolimus and zotarolimus drug-eluting stent clinical trials, we identified bleeding events using three common definitions of bleeding, ACUITY, TIMI, and GUSTO, and assessed the impact on mortality and MI at 12 months after PCI. The GUSTO, ACUITY, and TIMI classifications identified bleeding rates of 2.3%, 1.9%, and 2.1%, respectively. The GUSTO criteria classified all 118 suspected bleeding events. There were 22 (18.6%) and 8 (6.8%) suspected bleeding events that did not meet ACUITY and TIMI criteria, respectively. The combined endpoint of all-cause death or myocardial infarction (MI) at 12 months was significantly higher for patients with a bleeding event compared with those who did not bleed [hazard ratio 1.95 (95% CI 1.06-3.60)]. CONCLUSION: There is a substantial variability in the utility and inclusiveness of three widely used bleeding definitions in identifying clinically significant bleeding events in clinical trials of low risk patients undergoing PCI with DES. Patients with bleeding after elective PCI have an increased one-year risk of death or MI compared to those patients who do not bleed.