February 21st, 2014

Unpacking a Systematic Review of TAVR Research


The Annals of Internal Medicine recently published an article by Caroline A. Kim and colleagues, titled “Functional Status and Quality of Life After Transcatheter Aortic Valve Replacement: A Systematic Review.” We believe that this paper warrants careful analysis.

Kim et al. reviewed 62 studies, involving more than 11,000 patients who underwent TAVR; all the studies included some measure of functional status or quality of life. Despite heterogeneity in design, quality, and results among the studies, a clear message emerged: Overall, TAVR is associated with statistically significant, though at times clinically modest, improvements in symptoms, functional status, and quality of life 6 to 23 months after the procedure. Two aspects of the review by Kim and colleagues merit further examination.

First, the authors state that function and symptoms do not improve for a substantial minority of TAVR recipients. This conclusion must be interpreted carefully. Although therapeutic failures do occur after TAVR, many patients who report no symptom improvement are actually doing quite well. Often patients with the worst baseline functional status have the highest mortality rate after TAVR, but they also often experience the greatest improvements in function and quality of life. The opposite is frequently true for patients whose symptoms were already mild before TAVR and who, therefore, do not feel much better after the procedure. For patients with mild preprocedural symptoms, we must ensure that their aortic stenosis is truly severe and that they do not have other life-limiting conditions, to maximize the likelihood that TAVR will confer a meaningful survival benefit. To that end, researchers must better characterize the heterogeneity in symptom response so that we can better predict improvements in function and symptoms for individual patients. This goal may be more important than understanding the overall expected benefit across broad study cohorts.

Second, as the authors note, survival bias may have influenced the study findings. Specifically, patients with the worst functional status and symptom scores might have been the most likely to die and, therefore, the least likely to be represented in analyses of outcomes 6, 12, or 24 months after the procedure. We therefore need to integrate survival, functional status, and symptoms into a single, validated outcome measure. Suzanne Arnold and colleagues have proposed combining both mortality and quality-of-life measures into a composite endpoint on the basis of empirical data in the PARTNER trial. Arnold argues that the most appropriate definition of a poor outcome 6 months after TAVR involves consideration of (1) death, (2) a Kansas City Cardiomyopathy Questionnaire summary score <45 (corresponding to class IV heart failure), and (3) a KCCQ summary score decrease of ≥10 points. In essence, patients who have died, have class IV heart failure, or have significantly worse symptoms post-TAVR have failed to achieve a therapeutic endpoint. The exact parameters of such a definition can be debated, of course, but we nonetheless see a need for an integrated endpoint that includes survival and quality of life, to help us better identify which patients benefit most from TAVR.


Do you agree with the analysis of Dr. Green and Dr. Dharmarajan? 

Comments are closed.