This article is more than 5 years old. Information may no longer be current.
No volume-outcomes relationship apparent after learning curve in TAVR
Click Here to Manage Email Alerts
After overcoming a learning curve, procedural volume is no longer associated with outcomes for transcatheter aortic valve replacement with balloon-expandable valves, according to a recent study.
Using the Transcatheter Valve Therapy registry, Mark J. Russo, MD, MSc, from the department of cardiothoracic surgery at Rutgers Health in New Brunswick, New Jersey, and colleagues collected data on balloon-expandable valve implants from 2011 to 2017 (n = 61,949). They then evaluated 30-day all-cause mortality, stroke and major vascular complications to identify the length of the learning curve for TAVR with balloon-expandable valves and determine whether a relationship existed between a center’s monthly procedural volume and outcomes.
“I have a particular interest in the adoption and diffusion of new technology. I have studied volume-outcomes relationships in other areas, including aortic dissection, heart transplantation, lung transplantation and pediatric cardiac surgery,” Russo wrote in an email to Cardiology Today’s Intervention. “Given the dynamic state of TAVR with the rapid introduction of new device technology and changing practices, it seemed that the study of learning curves and volume outcome in TAVR might provide unique insights into these issues.”
Also, in a previous study by Carroll et al, the researchers used Transcatheter Valve Therapy data to explore a similar question, Russo noted.
“Their study in contrast to our more recent analysis concluded that a positive volume-outcomes relationship persisted or sites even after TAVR proficiency may have been achieved. This implied that a continuous learning curve was inseparable from the volume-outcomes relationship. However, their analysis examined TAVR experience only until November 2015, and therefore contains very little experience with the current-generation balloon-expandable device. It also included lower rates of cases done using transfemoral access and conscious sedation than is the current practice. This is important given that their subanalysis of transfemoral-only patients, Carroll et al found that after adjustment for patient and procedural characteristics, the association between volume and mortality and stroke was no longer statistically significant—though the relationship between volume and vascular complications persisted. This is consistent with our hypothesis and findings that the learning and volume-outcome relationship would be less pronounced, or even nonexistent, with present-day technologies and current practices,” he said.
Outcomes in relation to procedural volume
The researchers gauged experience by case series number and grouped them into quartiles. For all commercially available balloon-expandable valve types, there was an association between case sequence number quartile and 30-day mortality (P < .0001) and 30-day stroke (P = .009) after adjustment for Society of Thoracic Surgeons Predicted Risk of Mortality (STS PROM) score and site random effects.
Case sequence analysis showed that the learning curve for TAVR ended around the 200th case, the point at which there were no longer differences in 30-day mortality (P = .39), 30-day stroke (P = .27) or 30-day major vascular complications (P = .39), even after adjustment for STS PROM score and site random effects.
After the learning curve period ended, there was no longer an association between increasing procedural volume and 30-day mortality (P = .83) or stroke (P = .45). However, there was a significant association between TAVR frequency and major vascular complications (P = .01).
In analyses restricted to TAVR with the Sapien 3 valve (Edwards Lifesciences), the researchers did not identify a learning curve, with results demonstrating no significant associations between case sequence number and 30-day mortality, stroke or major vascular complications. Additionally, when comparing low-volume with high-volume centers, there was no association between TAVR frequency and 30-day mortality (P = .63), stroke (P = .64) or major vascular complications (P = .99).
Of those included in the study, 54 centers that had no documented TAVR cases with the Sapien or Sapien XT valves (Edwards Lifesciences) performed 1,209 procedures with the Sapien 3 valve. The researchers again found no associations between frequency of TAVR cases with the Sapien 3 valve and 30-day mortality (P = .69) or 30-day stroke (P = .9) after adjusting for STS PROM score and site random effect. However, results demonstrated a trend toward increased major vascular complications at centers performing fewer than two procedures per month (P = .052).
With increasing experience, there were decreases in the patients’ age and STS PROM scores and a trend toward increased transfemoral access, use of conscious sedation and less cardiopulmonary bypass.
“The findings themselves were not a total surprise. After all, we hypothesized that the learning curve and volume-outcome relationship would be less pronounced — or even nonexistent — with each successive device iteration,” Russo said.
“It is, however, a little surprising that people have continued to interpret the findings of the Carroll study and this current analysis to be in conflict. They, in fact, have very similar conclusions. Early in the US TAVR experience, there was a clear learning curve and volume-outcomes relationship. After adjustment for patient and procedural characteristics, there was no demonstrable volume relationship with mortality or stroke in transfemoral-only cases.”
Quantity vs. quality
This study shows that elements aside from procedural volume may have a hand in producing better outcomes after TAVR, according to the researchers.
“First, after the maturation of the TAVR procedure and introduction of enhanced new technology, there was no meaningful relationship between volume and mortality (or stroke). That is not to say experience doesn't matter; it obviously does. However, in current practice, excellent outcomes should be achievable even early in a centers experience.
Second, the roll-out and adoption of TAVR represent a new paradigm for the introduction of device therapies in the modern era. Intensive training, the heart team approach, group learning and crowd wisdom can mitigate learning curves and volume-outcome relationships,” Russo said.
Because previous data have demonstrated a link between TAVR volume and outcomes, this study raises some important questions as to whether procedural volume is the best way to assess quality, especially with the availability of newer, more advanced devices, according to Saif Anwaruddin, MD, and Matthew D. Saybolt, MD, both from the Hospital of the University of Pennsylvania and Perelman School of Medicine at the University of Pennsylvania.
“However, like most things in our world, the truth is somewhere in the middle, and it would seem unlikely that a [volume-outcomes relationship] is completely absent even for latest-generation TAVR,” they wrote in an accompanying editorial. “What remains apparent is that work needs to be done to better assess and to improve upon quality and outcomes in TAVR across all sites in the best interest of our patients.”
Russo agreed, noting that just as physicians refine technologies and practices, so must they continue to refine quality metrics to measure related outcomes.
“As death and stroke after TAVR become increasingly uncommon, other outcomes including pacemaker, paravalvular leak, vascular complications, functional status and quality of life must be assessed to measure quality. Moreover, in this setting, the volume is not a surrogate for quality and should not be used alone to restrict TAVR availability in lower-volume high-quality centers as well as underserved populations,” he told Cardiology Today’s Intervention. – by Melissa Foster
Disclosures: The statistical analyses for this manuscript were performed by Edwards Lifesciences. Russo reports he has served as a study investigator, a consultant and a proctor for Abbott, Boston Scientic and Edwards Lifesciences. Please see the study for all other authors’ relevant financial disclosures. Anwaruddin reports he has served as a consultant/speaker for Edwards Lifesciences and Medtronic. Saybolt has received research support from Infraredx.
Click Here to Manage Email Alerts