We thank Dr. Noyez for his comments regarding our study (Ben-Dor I, Gaglia MA, Barbash IM, et al. Comparison between Society of Thoracic Surgeons score and logistic EuroSCORE for predicting mortality in patients referred for transcatheter aortic valve implantation. Cardiovasc Revasc Med . 2011;12:345–9).

We agree with Dr. Noyez in that both the STS score and the EuroSCORE were developed from clinical databases in order to predict expected patient mortality from surgical cardiac procedures. Currently, however, both risk scores are used as key criteria for inclusion in transcatheter aortic valve replacement (TAVR) trials. The EuroSCORE is used mostly outside of the USA, while the STS score is used mostly in the USA. Both scores provide an objective benchmark for standardized patient enrollment, and both have inherent limitations.

Understanding that the logistic EuroSCORE overpredicts the risk of mortality and the STS score is superior for patients undergoing surgical AVR, we chose to use both risk models in our study due to the lack of a specific risk score for TAVR and balloon aortic valvuloplasty (BAV). The data sets for both risk models did not include patients undergoing TAVR, yet this does not mean that we cannot use those risk models for predicting mortality after TAVR and medical/BAV. These risk scores include variables, such as renal failure functional class ejection fraction, known to increase mortality risk in patients with severe aortic stenosis regardless of the treatment choice.

The STS score and the EuroSCORE were standard when our data were collected, and both lack the ability to predict 30-day mortality; therefore, we agree with Dr. Noyez and support the need for a new risk model. With the increased number of TAVRs performed worldwide and a new accumulating data set, of course, a new risk score with variables that are more relevant for TAVR and/or BAV will be more accurate than the STS score and the EuroSCORE.

It remains open-ended whether or not one risk model for all aortic valve treatment or for individual models based on the treatment data sets is needed. EuroSCORE II may be a good starting point; however, the introduction of TAVR to the field of aortic stenosis treatment warrants the determination of whether or not such surgical scores can be and/or should be applied to nonsurgical modalities.

Sincerely,