Abstract 18674: Predicting Long Term Outcome in Patients Treated With Continuous Flow Left Ventricular Assist Device - The Penn - Columbia Risk Score

Abstract

Mechanical assist devices have emerged as an established therapeutic option for patients with end-stage heart failure. Predicting which patients are unlikely to benefit from continuous flow left ventricular assist device (cf-LVAD) treatment is crucial for the identification of appropriate patients while excluding those considered futile. Previously developed scoring systems are limited to prior eras of device and medical support or restricted to specific devices. Methods: We performed a retrospective analysis of all patients implanted with a cf-LVAD (Heartmate II or HVAD) at our institution between 2005 and 2014. From all plausible pre-operative covariates, we performed univariate Cox regression analysis for covariates affecting the odds of 1yr post implantation survival (p<0.4). These variables were included in a multivariable model and dropped if significance rose above p=0.4. From this base model, we performed stepwise forward and backward selection for other covariates that improved power by minimizing Akaike Information Criteria while maximizing the Harrell Concordance Index. We then used Kaplan-Meier curves, the log rank test, and Cox proportional hazard models to assess internal validity of the scoring system and its ability to stratify survival. The risk score was validated in another VAD center. Results: A total of 204 patients (male 78%; mean age 57+15, mean follow up 465+486 days). After stepwise selection for variables that increased predictive power, a final optimized model was identified (Figure). One yr survival was significantly lower in patients with higher risk score (HR 2.48, p=<0.001, figure). The validation cohort of 260 patients confirmed the prognostic utility of this risk score (p=0.0237). Conclusion: We present a novel risk score and its validation for prediction of long-term survival in patients on cf-LVAD support. This score may prove useful for patient selection and identification of patients requiring adjuvant therapies