Abstract 6: A Validated Contemporary Risk Model of Acute Kidney Injury in Patients Undergoing Percutaneous Coronary Interventions: Insights From the NCDR Cath-PCI Registry

Abstract

Background: Acute Kidney Injury (AKI) is a serious complication of percutaneous coronary intervention (PCI) and is associated with adverse outcomes. Accurately estimating this risk may help clinicians better inform patients considering PCI of their potential risk of this complication. Methods: Data were analyzed from 985,737 consecutive PCI patients performed at 1253 sites participating in the NCDR-Cath-PCI registry. AKI was defined as an absolute increase of ≥ 0.3 mg/dL or a relative increase of 50% in serum creatinine or a new requirement for dialysis following PCI (AKIN Stage 1 or greater). A parsimonious multivariable model in a 70% training cohort identified factors associated with AKI and assigned weighted integer scores. This score was validated in the remaining 30% of the population. Results: Overall, 69,658 (7.1%) patients developed AKI with an associated in-hospital mortality of 9.7% versus 0.5% in those without AKI. The model identified 16 variables predictive of AKI (Figure 1). The parsimonious integer showed good discrimination in validation (c-statistic = 0.72). Patients with increasing risk score had increasing risk of predicted and observed rates of AKI across low, medium and high risk groups (Figure 2) Conclusion: The NCDR AKI prediction model is a simple, contemporary and robust tool for predicting AKI in patients undergoing PCI. In certain clinical scenarios, the use of this tool may aid clinicians in counseling patient regarding the risk of PCI and identify patients for targeted interventions.