Is the EuroSCORE II best suited for reoperative risk estimation in patients with structural deterioration of aortic bioprostheses?

BACKGROUND Operative risk prediction systems (logistic EuroSCORE I, EuroSCORE II and STS Score) are employed together with multidisciplinary discussion to contraindicate conventional surgery in patients with valvular heart disease and propose the employment of alternative transcatheter procedures. The EuroSCORE I has been reported to underperform in these circumstances; we hypothesize that the EuroSCORE II is best suited for the stratification of risk in patients with structural deterioration (SVD) of valvular bioprostheses and potential candidates to the Valve-in-Valve procedure (deployment of a transcatheter valve within a failing valvular bioprosthesis). METHODS AND EVALUATION OF THE HYPOTHESIS A multi-institutional collaboration is required to fully address such hypothesis. Therefore, we performed a preliminary validation study by retrieval of the complete records of 81 patients undergoing reoperative aortic valve replacement for preoperative diagnosis of bioprosthetic SVD at our Institution. Logistic EuroSCORE I, EuroSCORE II and STS Score were calculated by preoperatively available data. Faced to an observed reoperative mortality of 4.9%, average EuroSCORE I was 15.8%±13.4, EuroSCORE II was 7.3%±7.4 and the STS Score was 15%±9.8. The three systems provided sufficient adequacy (Hosmer-Lemeshow p=0.847, p=0.999 and p=0.9948, respectively). Yet, the area under the ROC curve was significantly higher for the EuroSCORE II (0.9903) vs. the EuroSCORE I (0.8994) (p=0.044). The STS Score yielded an intermediate figure (0.9643). The odds ratios (logistic regression) were 1.079 for EuroSCORE I, 1.223 for the STS Score and 1.474 for EuroSCORE II. CONCLUSIONS The three investigated algorithms showed reasonable calibration in the prediction of mortality for reoperative aortic valve replacement, but they evenly overestimated the observed mortality. The hypothesis that the EuroSCORE II is better suited for the selection of candidates to Valve-in-Valve implantation is worth of further multi-institutional investigations on the basis of our preliminary findings and due to the expanding role of transcatheter techniques.