BACKGROUND: High-risk percutaneous coronary intervention (HRPCI) procedures supported by percutaneous left ventricular assist devices (pLVAD) are increasingly common, but existing PCI risk scores were developed in patients across the risk spectrum, including few pLVAD-assisted patients. OBJECTIVES: Assess the performance of existing PCI risk scores in patients receiving pLVAD-assisted HRPCI and create a novel risk score specific to this group. METHODS: Patients in the PROTECT III multicenter, observational (46 US centers) study undergoing pLVAD-assisted HRPCI were assessed. The National Cardiovascular Data Registry (NCDR) bedside risk score and the Complex High-Risk Indicated PCI (CHIP-PCI) risk score were calculated for each patient, and their accuracy in predicting in-hospital events was assessed. A novel risk score for in-hospital mortality was created using pre-procedural variables which were significant in univariable and multivariable regressions. RESULTS: Among 1237 patients, the NCDR bedside risk score showed modest discrimination (C-index 0.71) but poor goodness of fit (R2 = 0.30). The CHIP-PCI score had poor discrimination (C-index 0.61) and reasonable goodness of fit (R2 = 0.62). Five independent predictors of in-hospital mortality were identified: age >80 years, eGFR <30, left main disease, acute myocardial infarction, and left ventricular ejection fraction <30 %. These formed the "HRPCI" risk score (C-index 0.75), which correlated with 30-day mortality (5.4 % vs. 17.0 %, p<0.0001). CONCLUSIONS: Existing PCI risk scores perform poorly in patients undergoing pLVAD-assisted HRPCI. A novel easily, calculable HRPCI risk score can assist in clinical decision making once validated. CLINICAL TRIAL INFORMATION: Trial Name: The Global cVAD Study (cVAD). URL: https://clinicaltrials.gov/ct2/show/NCT04136392?term=cvad&draw=2&rank=2 ClinicalTrial.gov Identifier: NCT04136392.
