https://orcid.org/0000-0001-5028-8186
-
PDF
- Split View
-
Views
-
Cite
Cite
Fausto Biancari, Marco Barbanti, Paola D’Errigo, Stefano Rosato, Reply to Nezic, European Journal of Cardio-Thoracic Surgery, Volume 60, Issue 4, October 2021, Pages 1002–1003, https://doi.org/10.1093/ejcts/ezab204
Close - Share Icon Share
We are grateful to Dr Nezic for his comments on our recent study [1, 2], the goal of which was to develop a new scoring method to stratify the operative risk of patients undergoing surgical aortic valve replacement. Dr Nezic criticized our proposal also to externally validate this risk-scoring method in patients undergoing transcatheter aortic valve replacement (TAVR). External validation is an important step when evaluating the predictive ability of a risk stratification tool that is usually performed in patient populations exposed to the same treatment. In patients with severe aortic stenosis, a risk score would allow the identification of patients with high or prohibitive operative risk for surgical aortic valve replacement and who may benefit from a less invasive treatment such as TAVR or in whom conservative treatment is indicated. Such a risk-scoring method may still be valuable in stratifying the risk of patients undergoing TAVR, in a way similar to the way the EuroSCORE II and the Society of Thoracic Surgeons score are currently used. Therefore, the OBSERVANT aortic valve replacement (OBS AVR) score is expected to help clinicians make decisions about the care of patients with severe aortic stenosis requiring either surgical aortic valve replacement or TAVR.
We do agree with Dr Nezic about the expected better performance of the OBS AVR score compared with the EuroSCORE II in our data. It is our goal to assess its predictive ability in an external validation data set. Nevertheless, in the validation data set, the area under the receiver operating characteristic curve was 0.76 (95% confidence interval 0.67–0.85) for the OBS AVR score and 0.68 (95% confidence interval 0.57–0.79) for the EuroSCORE II. This difference was statistically significant (P < 0.0001).
Regarding the performance of the EuroSCORE II in predicting 30-day instead of hospital mortality, Nashef et al. [3] estimated that hospital mortality (4.015%) and 30-day mortality (4.048%) were virtually identical, although some overlap between the 2 outcomes occurred. Therefore, we may expect that the performance of the EuroSCORE II might not suffer significantly by investigating 30-day mortality instead of hospital mortality as we did in our study.
