Author + information
- Pierre Lantelme, MD, PhD∗ (, )
- Pierre-Yves Courand, MD, PhD and
- Brahim Harbaoui, MD, PhD
- ↵∗Hôpital Croix-Rousse and Hôpital Lyon Sud, Cardiology Department, 103 Grande Rue de la Croix-Rousse, F-69004 Lyon, France
Patient selection is a critical concern for candidates for transcatheter aortic valve replacement (TAVR) procedures. In this elderly population, frailty is intuitively important to consider during the pre-operative workup, but its prognostic value has not been unequivocally assessed in this setting; one explanation is the lack of an objective definition of frailty. In their recently published paper, Shoenenberger et al. (1) addressed the value of combining several frailty indexes with conventional surgical scores, namely, the European System for Cardiac Operative Risk Evaluation score and the Society of Thoracic Surgeons score for predicting 1-year mortality; they showed an improvement of the predictive value of these scores by adding frailty indexes. Yet the performance of the European System for Cardiac Operative Risk Evaluation score appeared surprisingly high, with C statistics close to 0.7, contrasting with many previous reports showing much lower performance for predicting 1-year outcomes. This raises a statistical issue related to the small size of the cohort. Predictive models are indeed generated to provide the best fit for the available data, but they may be overfitted in small datasets and hence provide an optimistic assessment of predictive ability, exemplified by unexpectedly high C statistics. Techniques of cross-validation or external validation are useful to provide unbiased estimates of predictive value (2). Unfortunately, they were not used in this study. Another limitation of the study, in view of the evolving indications for TAVR, is the period of inclusion (2009 to 2013), which makes this cohort rather different and perhaps not representative of current practice. Taken together, these limitations call into question the applicability of the conclusions to other cohorts and to current practice.
That said, the investigators should be commended for their attempts to improve the identification of potentially futile TAVR procedures. Considering the difficulty of assessing frailty, our group has proposed another meaningful predictor: aortic calcification burden assessed during systematic pre-operative computed tomography; this calcification burden was indeed highly predictive of morbidity and mortality outcomes after TAVR (3). Aortic calcification represents an objective indicator of cardiovascular frailty but also of overall frailty, as it is correlated with bone and muscle loss. We have combined this calcification burden with other conventional risk factors to build a score, the CAPRI score, to predict 1-year mortality after TAVR. This score was constructed in a large multicenter cohort and validated in an independent cohort with excellent calibration. So far it represents the first dedicated score to predict futility after TAVR with statistical methods used to minimize optimism (4). According to this score, more than 5% of TAVR procedures currently indicated could be considered futile.
It is necessary to test whether conventional frailty indexes would further improve this score. The study by Shoenenberger et al. (1) and ours underscore the need to develop tools for improving patient selection. Avoiding futile intervention is mandatory both for ethical reasons (i.e., to prevent patients from being exposed to periprocedural discomfort and complications) and for economic ones, to optimize the allocation of financial resources. So far, we believe that the CAPRI score represents the most comprehensive score for predicting 1-year mortality; heart teams should be encouraged to use it.
Please note: The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- 2018 American College of Cardiology Foundation
- Schoenenberger A.W.,
- Moser A.,
- Bertschi D.,
- et al.
- Harbaoui B.,
- Courand P.Y.,
- Charles P.,
- et al.
- Lantelme P.,
- Eltchaninoff H.,
- Souteyrand G.,
- et al.