Performance of surgical risk scores to predict mortality after transcatheter aortic valve implantation

• Published in: Arquivos brasileiros de cardiologia 2015-09, Vol.105 (3), p.241-247
• Main Authors: Silva, Leonardo Sinnott, Caramori, Paulo Ricardo Avancini, Nunes Filho, Antonio Carlos Bacelar, Katz, Marcelo, Guaragna, João Carlos Vieira da Costa, Lemos, Pedro, Lima, Valter, Abizaid, Alexandre, Tarasoutchi, Flavio, Brito, Jr, Fabio S de
• Format: Article
• Language: English
• Subjects: Aged; Aged, 80 and over; Aortic Valve Stenosis - mortality; Aortic Valve Stenosis - surgery; Brazil; Calibration; CARDIAC & CARDIOVASCULAR SYSTEMS; Epidemiologic Methods; Estenose da Valva Aórtica / cirurgia; Fatores de Risco; Female; Humans; Male; Original; Probabilidade; Reference Values; Reproducibility of Results; Risk Assessment - methods; Substituição da Valva Aórtica Transcateter; Time Factors; Transcatheter Aortic Valve Replacement - mortality
• ISSN: 0066-782X; 1678-4170; 1678-4170
• DOI: 10.5935/abc.20150084

Predicting mortality in patients undergoing transcatheter aortic valve implantation (TAVI) remains a challenge. To evaluate the performance of 5 risk scores for cardiac surgery in predicting the 30-day mortality among patients of the Brazilian Registry of TAVI. The Brazilian Multicenter Registry prospectively enrolled 418 patients undergoing TAVI in 18 centers between 2008 and 2013. The 30-day mortality risk was calculated using the following surgical scores: the logistic EuroSCORE I (ESI), EuroSCORE II (ESII), Society of Thoracic Surgeons (STS) score, Ambler score (AS) and Guaragna score (GS). The performance of the risk scores was evaluated in terms of their calibration (Hosmer-Lemeshow test) and discrimination [area under the receiver-operating characteristic curve (AUC)]. The mean age was 81.5 ± 7.7 years. The CoreValve (Medtronic) was used in 86.1% of the cohort, and the transfemoral approach was used in 96.2%. The observed 30-day mortality was 9.1%. The 30-day mortality predicted by the scores was as follows: ESI, 20.2 ± 13.8%; ESII, 6.5 ± 13.8%; STS score, 14.7 ± 4.4%; AS, 7.0 ± 3.8%; GS, 17.3 ± 10.8%. Using AUC, none of the tested scores could accurately predict the 30-day mortality. AUC for the scores was as follows: 0.58 [95% confidence interval (CI): 0.49 to 0.68, p = 0.09] for ESI; 0.54 (95% CI: 0.44 to 0.64, p = 0.42) for ESII; 0.57 (95% CI: 0.47 to 0.67, p = 0.16) for AS; 0.48 (95% IC: 0.38 to 0.57, p = 0.68) for STS score; and 0.52 (95% CI: 0.42 to 0.62, p = 0.64) for GS. The Hosmer-Lemeshow test indicated acceptable calibration for all scores (p > 0.05). In this real world Brazilian registry, the surgical risk scores were inaccurate in predicting mortality after TAVI. Risk models specifically developed for TAVI are required.