Abstract 303: The Association Between Change in Hospital-Level Use of Transradial PCI and Periprocedural Outcomes: Insights from the NCDR®

Abstract

Background: Although radial artery access reduces access site bleeding, increasing use of transradial PCI (TRI) may compromise facility-level outcomes due to patient selection, procedural learning curve, and unintended consequences on femoral PCI management. We sought to determine the relationship between increasing facility-level use of TRI and change in rates of access site bleeding. Methods: Within the National Cardiovascular Data Registry CathPCI Registry®, we restricted our analyses to hospitals reporting at least 50 PCI procedures annually and with less than 10% TRI use in the first year of observation. We evaluated 1,438,516 patients undergoing PCI at 818 facilities participating in NCDR CathPCI between 2009 and 2012. We identified categories of hospital change in TRI use from latent class growth analysis with time modeled continuously across quarters. Controlling for bleeding risk calculated from prior CathPCI bleeding risk models, we estimated the association between hospital category of change in TRI use and access site bleeding rates using generalizing estimating equations accounting for clustering within hospitals and time. Results: Latent growth curve analysis identified four classes of hospital-level change in TRI use with widely divergent patterns (Figure 1a). Patients’ predicted bleeding risk was similar across hospital categories of change in TRI use (predicted bleeding from lowest to highest hospital category of TRI increase, 6.0% vs 5.8% vs 6.1% vs 5.7%). Risk-adjusted bleeding decreased over time for all hospitals, regardless of change in TRI use (Figure 1b). The decreasing rate of access site bleeding (determined from the relative risk [RR] of access site bleeding in the last quarter of study compared to the first quarter) was greater for hospitals with moderate or high increases in TRI use (RR 0.45; 95% confidence interval [CI] 0.36-0.56) compared with hospitals with very low or low increases in TRI use (RR 0.65; 95% CI 0.58-0.74; p for comparison=.002). Conclusions: In a national sample of hospitals performing PCI, access site bleeding decreased over time for all hospitals. The trajectory of decline in access site bleeding was greatest at hospitals with the largest increases in TRI use.