Author + information
- Matthew R. Reynolds, MD, MSc⁎ ( and )
- David J. Cohen, MD, MSc
- ↵⁎Harvard Clinical Research Institute, 930-W Commonwealth Avenue, Boston, Massachusetts 02215
As Dr. Cao and colleagues point out, our recently published cost-effectiveness analysis, based on patient-level data from the PARTNER (Placement of Aortic Transcatheter Valves Trial) (1) has some important limitations. Our analysis used a 12-month timeframe. From a health economic perspective, this analytic timeframe is similar to an analysis that assumes that long-term outcomes between study groups will be comparable. Although there is presently no empirical evidence suggesting otherwise, this remains an unproven assumption, which we explicitly noted in the manuscript.
Also correctly noted was the fact that hospital length of stay (LOS) was shorter after transcatheter aortic valve replacement (TAVR) than after surgical aortic valve replacement (AVR), and the savings associated with reduced LOS play a key role in offsetting the higher acquisition cost of a transcatheter valve. The LOS following surgical AVR in the PARTNER trial was indeed long—about 6 days longer than average LOS for AVR for all U.S. patients (2). We believe the long LOS after surgical AVR in the PARTNER trial was primarily related to the unique risk profile of the patient population, who represent the highest 5% to 8% risk of patients operated on for aortic stenosis in the United States. Based on recent publications, it is clear that “high-risk” patients who currently undergo TAVR in other countries tend to be lower risk than those who were studied in the PARTNER A trial. For example, the SOURCE-XT registry of 2,700 consecutive patients undergoing TAVR in 17 countries enrolled patients with a mean Society of Thoracic Surgeons (STS) mortality risk of 8.5% (4) whereas the PARTNER trial population had a mean STS mortality risk of 11.8% (3). Moreover, as we noted in our discussion, even for similar patients, clinical results and resource utilization patterns may differ across healthcare systems. We agree that country-specific analyses are needed, but we are concerned about conclusions reached by comparing nonrandomized groups (5). In the absence of randomization, careful risk adjustment is required to reach balanced conclusions.
Finally, it is not correct to state that our analysis ignored or neglected complications associated with TAVR. For the TAVR and AVR groups in the trial, we measured costs for index and follow-up hospitalizations primarily by collecting and analyzing hospital bills from the enrolling centers. Thus, while we have not yet reported the excess costs associated with specific complications, those costs were incorporated into analysis through the 12-month timeframe of the study.
- American College of Cardiology Foundation
- Reynolds M.,
- Magnuson E.,
- Lei Y.,
- et al.
- Agency for Healthcare Research and Quality
- ↵Vahanian A. Thirty-day outcomes in a 2700 patient international clinical registry with a second generation balloon expandable transcatheter heart valve using multiple access techniques. Presented at: EuroPCR Conference; May 2012.
- Neyt M.,
- Van Brabandt H.,
- Devriese S.,
- Van De Sande S.