Author + information
- Jonathan Afilalo, MD, MSc∗ ( and )
- Dae Hyun Kim, MD, MPH, ScD
- ↵∗Jewish General Hospital, McGill University, 3755 Cote Ste. Catherine Road, E-222, Montreal, QC H3T 1E2, Canada
In a recent issue, Schoenenberger et al. (1) eloquently presented a post hoc analysis from their single-center cohort of 330 older adults undergoing transcatheter aortic valve replacement between 2009 and 2013, in which they assessed the predictive value of frailty alongside conventional risk scores. Given the modest number of events (52 deaths over 1 year), to avoid overfitting, the investigators selected 2 independent variables: a composite clinical risk score (logistic European System for Cardiac Operative Risk Evaluation score or Society of Thoracic Surgeons score) and a multidimensional geriatric assessment (MGA) score reflecting mobility, cognition, nutrition, and disability. By combining a clinical risk score with a frailty scale, they captured the use case of a practicing clinician who wishes to enrich his or her assessment of a complex older patient referred for transcatheter aortic valve replacement.
Frailty metrics provide unique information that is highly predictive and yet does not overlap with the realm of traditional risk factors that are encrusted in our cardiocentric vernacular; this may simultaneously be a strength of frailty scores and a limitation of current clinical risk scores. A small but salient caveat is that although the investigators used the older European System for Cardiac Operative Risk Evaluation I, the newer version (2) has added a question for “poor mobility” (broadly defined as severe impairment of mobility secondary to musculoskeletal or neurological dysfunction), borrowing from geriatricians’ vernacular and initiating a paradigm shift from disease- or surgery-specific risk scores to much needed patient-centered risk scores.
The investigators of this study astutely used a bevy of model performance statistics to demonstrate that the MGA added incremental predictive value above the clinical risk scores. This finding is consistent with the 14-center FRAILTY-AVR (Frailty in Aortic Valve Replacement) cohort including >700 older adults undergoing transcatheter aortic valve replacement (119 deaths over 1 year), in which we had previously shown that the MGA added incremental value above the Society of Thoracic Surgeons score, achieving a C statistic of 0.70 (comparable with their reported C statistic of 0.68) and an integrated discrimination improvement statistic of 0.030.
The FRAILTY-AVR study (3) prospectively compared the incremental value of 7 different frailty scales, including the MGA, and ultimately found that the Essential Frailty Toolset was more predictive of 1-year mortality and disability and was shorter to administer. This leads us to respectfully question the omission of FRAILTY-AVR in the background and discussion of Schoenenberger et al.’s (1) study. Too often, readers are presented with the merits of frailty scale X without being presented data on frailty scales Y and Z. Alternatively, they are presented modified versions of frailty scales Y and Z constructed with one-off variables available in retrospective databases, leading them to believe that these scales perform poorly. This speaks to the clinical community’s perplexity with the plethora of frailty scales being promulgated and the lack of consensus that has impeded frailty from going mainstream.
Although we cannot humbly assert a simple solution to this conundrum, pending external validation of the Essential Frailty Toolset, we would like to call on frailty researchers to take an inclusive view of frailty that acknowledges and compares the effectiveness of different scales to predict outcomes and affect care. Because the choice of frailty scale also depends on feasibility, expertise, and resources, relevant measures such as administration time and assessor training should be considered. Once we as frailty researchers show that we can advance a cohesive agenda, then clinicians will follow suit and carry our research forward to the ultimate stage of knowledge translation.
Please note: Both 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.
- Afilalo J.,
- Lauck S.,
- Kim D.H.,
- et al.