Differential Use and Impact of Bleeding Avoidance Strategies on Percutaneous Coronary Intervention-Related Bleeding Stratified by Predicted Risk

Abstract

Background: Procedural anticoagulation with bivalirudin (BIV), trans-radial intervention (TRI), and use of a vascular closure device (VCD) are thought to mitigate percutaneous coronary intervention (PCI)-related bleeding. We compared the impact of these bleeding avoidance strategies (BAS) for PCIs stratified by bleeding risk. Methods: We performed a retrospective cohort analysis of PCIs from 18 facilities within one health care system from 2009Q3 to 2017Q4. Bleeding risk was assessed per the National Cardiovascular Data Registry CathPCI bleeding model, with procedures stratified into 6 categories (first, second, third quartiles, 75th–90th, 90th–97.5th, and top 2.5th percentiles). Regression models were used to assess the impact of BAS on bleeding outcome. Results: Of 74 953 PCIs, 9.4% used no BAS, 12.0% used BIV alone, 20.8% used TRI alone, 26.8% used VCD alone, 5.4% used TRI+BIV, and 25.6% used VCD+BIV. The crude bleeding rate was 4.4% overall. Only 2 comparisons showed significant trends across all risk strata: VCD+BIV versus no BAS, odds ratio (95% CI) range: first quartile, 0.36 (0.18–0.72) to top 2.5th percentile, 0.50 (0.32–0.78); TRI versus no BAS, odds ratio (95% CI) range: first quartile, 0.15 (0.06–0.38) to top 2.5th percentile, 0.49 (0.28–0.86). TRI had lower odds of bleeding compared with BIV for all risk strata except the top 2.5th percentile. Addition of BIV to TRI did not change the odds of bleeding for any risk strata. Factors potentially limiting use of TRI (renal failure, shock, cardiac arrest, and mechanical circulatory support) were present in ≤10% of procedures below the 90th percentile. Conclusions: Among individual BAS, only TRI had consistently lower odds of bleeding across all risk strata. Factors potentially limiting TRI were found infrequently in procedures below the 90th percentile of bleeding risk. For transfemoral PCI, VCD+BIV had lower odds of bleeding compared with no BAS across all risk strata.