Abstract 16927: Validation of CathPCI Bleeding Model in Transradial vs. Transfemoral PCI Patients

Abstract

Background: Estimation of bleeding risk in patients undergoing PCI informs the use of bleeding avoidance strategies, allows for personalized consent, and supports quality improvement efforts. The CathPCI Bleeding Model (CPBM) was developed in patients undergoing PCI between 2008 and 2011, when PCI was performed predominantly via transfemoral (TF) access. Whether this model functions equally well in transradial (TR) PCI is unclear. Objectives: This study sought to determine if the CPBM behaves differently for PCI performed via TF vs TR access. Methods: Patients who underwent PCI between Q4 2010 and Q4 2016 were identified from an institutional CathPCI Registry. The primary outcome was post-PCI bleeding within 72 hours or before hospital discharge. A logistic regression model was created predicting CPBM on the 10-variable reduced model. Interactions between each variable and access site were tested. Using the CPBM integer scoring system, the interaction between low, moderate, and high risk groups with the access site was tested. Results: TR access increased from 14% in 2010 to 60% in 2016. Major bleeding was lower for TR than TF PCI (2.5% vs 7.3%, p < 0.001). Prior PCI was protective of bleeding in the TF, but predictive of bleeding in the TR group. There was no significant interaction between access site and all remaining variables (Fig. 1). There was no difference in the discriminative capacity of the CPBM between TR and TF access (Fig. 1). The CPBM integer score behaved similarly between groups, without significant interaction, p = 0.53 (Fig. 2). Conclusions: The CPBM may be used to estimate patients’ bleeding risk regardless of PCI access site.