Performance of the simplified heart failure hydra score in predicting one-year mortality: comparison with hydra score

Abstract Aims We have recently proposed a new prediction model for risk mortality based on four parameters of congestion to predict mortality in patients with acute or chronic heart failure. This risk model (HYDRA Score, HS) is based on the number of parameters of congestion above significant cut–offs (range 0–4): BNP >441 pg/mL, estimated plasma volume status (ePVS) >5.3 dL/gr, bioimpedance vector analysis (BIVA) >73.8%, and BUN/creatinine ratio (BUN/Cr) >25. As bioimpedance vector analysis is a measure of peripheral congestion but it is not frequently used in clinical practice, we evaluated the performance of the risk model by including the presence of peripheral edema in place of BIVA (HYDRA Semplified Score, HSS). Methods and results We analysed data from 436 HF patients. During a one–year follow–up 81 patients died (18.6%). Discrimination for all–cause mortality was compared by Harrell‘s C–statistic. Calibration was assessed by Hosmer–Lemeshow test and global performance by Nagelkerke‘s R2. Correlation between scores was assessed by Spearman rank test. Correlation between the scores was good (rho = 0.91, p<0.001). The two risk models showed similar results in terms of discrimination analyses (HS 0.786 vs HSS 0.760; p=0.09) and global performance (R2 = 0.26 and R2 =0.21). The Hosmer–Lemeshow goodness–of–fit test demonstrated the best calibration of HS risk model (χ2 6.5, P=0.60 for HS model and χ2 7.7, P=0.45 for HSS model). Conclusions Simplified HYDRA score seems a good and friendly model risk to predict one–year mortality in patients with heart failure and could be used as alternative to HYDRA score.