Meta-analysis on the performance of the EuroSCORE II and the Society of Thoracic Surgeons Scores in patients undergoing aortic valve replacement

Objective: To evaluate the performance of the EuroSCORE II (ESII) and the Society of Thoracic Surgeons (STS) scores in surgical (SAVR) or transcatheter aortic valve replacement (TAVR). Design: Systematic review of the literature and meta-analysis. Setting: University hospitals. Participants: Studies reporting data on the performance of ESII and STS scores in patients undergoing SAVR or TAVR. Interventions: SAVR or TAVR. Measurements and Main Results: Ten studies validated these scores in 13,856 patients who underwent either TAVR or SAVR. Operative mortality was 5.9% (SAVR 3.1%; TAVR 9.6%). ESII-expected mortality was 5.1% (O/E ratio: 1.15, SAVR, O/E ratio 0.94; TAVR, O/E ratio 1.23) and STS-expected mortality was 6.3% (O/E ratio: 0.94, SAVR, O/E ratio 0.84; TAVR, O/E ratio 1.13). The area under the ROC curve for ESII was 0.70 and for STS was 0.70 (SAVR patients: 0.73 for ESII and 0.75 for STS; TAVR patients; 0.66 for ESII and 0.63 for STS). The difference between observed/expected mortality was not significant for ESII (Peto's OR 0.99, p = 0.88) and was significant for STS (Peto's OR 0.86, p = 0.008). ESII (Peto's OR 1.35, p < 0.00001) and STS (Peto's OR 1.23, p < 0.00001) significantly underestimated the mortality risk in TAVR patients. The STS (Peto's OR 0.74, p < 0.0001) and, to a lesser extent, the ESII (Peto's OR 0.86, p = 0.0.04) overestimated the mortality risk in SAVR patients. Conclusions: The ESII and STS scores have good O/E ratios for either TAVR or SAVR patients, but both scores significantly underpredicted the risk of TAVR patients. ESII seemed to be accurate in predicting the risk of SAVR patients.