Risk stratification of patients are claimed to be useful before left ventricular assist device (LVAD) implantation and different scoring systems are available. The aim of the study was to validate the Cardiac Health Risk Stratification System (CHRiSS), based on Bayesian network analysis, and the HeartMate II score in our patient population. We retrospectively calculated the CHRiSS using a web-based application and the HeartMate II score of 105 adult patients who underwent consecutive HVAD (HeartWare International, Inc.) implantation as primary LVAD at our institution in a 12-month period (May 2014-April 2015). Survival was 83.8 % (3.6%) at 30 days, 77.8% (4.0%) at 90 days, and 72.8% (4.8%) at 6 months and 1 year. The area under the curve (AUC) of the receiver operating characteristics (ROC) curve for CHRiSS was 0.75 at 30 days, 0.66 at 90 days, and 0.65 at 6 months. The HeartMate II score had an AUC of 0.73 at 90 days. Comparison between the ROC curves of the two models calculated at 90 days showed no statistically significant difference (p = 0.48): CHRiSS presents a high positive predictive value (85 [80-91]), the opposite to the HeartMate II score, which has a high negative predictive value (91 [83.4-96.6]). In our population, application of the CHRiSS was valuable at 30 days, but the overall predictive value of both scores is not satisfactory. The CHRiSS model proved to be a promising tool, suggesting that, with greater sample size and a longer data collection period, it might potentially outperform the HeartMate II scoring system.
- left ventricular assist device
ASJC Scopus subject areas
- Biomedical Engineering