March 29th, 2014

TAVR: Are You Impressed Now?

CardioExchange’s Rick Lange and David Hillis ask interventionalist Ajay Kirtane for his perspective on the randomized trial comparing transcatheter aortic valve replacement (with the CoreValve device) and surgical aortic valve replacement. The data were presented at the 2014 ACC conference in Washington, DC, and published in The New England Journal of Medicine. Click here for CardioExchange’s news coverage of the trial.


In the CoreValve trial, researchers randomized 795 patients with severe aortic stenosis and a high estimated mortality risk for surgery to undergo surgical aortic valve replacement (SAVR) or transcatheter valve replacement (TAVR) with the Medtronic CoreValve self-expanding bioprosthesis. TAVR was associated with a significantly lower 1-year mortality rate than SAVR (14.2% vs. 19.1%; P=0.04).

By contrast, results from the Placement of Aortic Transcatheter Valves (PARTNER A) trial of the Edwards SAPIEN valve, involving 699 high-risk patients with severe aortic stenosis, transcatheter and surgical procedures for aortic valve replacement were associated with similar rates of survival at 1 year.


Lange and Hillis: Will the results of the CoreValve study change your threshold for recommending TAVR?  

Kirtane: First, the CoreValve investigators should be commended for completing a second randomized trial of TAVR against SAVR. We know that for patients at the highest end of the risk spectrum, TAVR is the standard of care because it has superior outcomes compared with those of standard medical therapy (as demonstrated in the PARTNER B trial and the nonrandomized CoreValve study just published in JACC). In patients at high risk for SAVR, the PARTNER A trial demonstrated similar overall survival with TAVR and SAVR, with better quality-of-life outcomes for TAVR. Thus, most patients at high risk for SAVR currently undergo evaluation at valve centers for consideration of TAVR; for many, TAVR is the preferred approach.

The new randomized CoreValve study in NEJM confirms and extends the prior findings in a separate population of patients also deemed to be at high surgical risk, with a survival advantage for TAVR at both 30 days and 1 year. Even with results out to only 1 year, this trial further tilts the balance toward TAVR in high-risk patients.

Notably, though, a significant component of the morbidity and mortality in high-risk and “inoperable” patients with severe aortic stenosis is noncardiac. The high event rates in prior TAVR studies have related to comorbidities in many of the patients (in addition to their aortic stenosis). In patients deemed to be at high surgical risk, we have been willing to accept less-certain valve results over the long term (e.g., less in-vivo durability data, greater paravalvular leak) for the advantages of a minimally invasive approach that allows patients to actually survive the procedure. As TAVR gets extended to lower-risk populations, it is important to keep these long-term, valve-specific endpoints in mind, because conventional SAVR is an excellent procedure with proven durability over the long term. Indeed, lower-risk patients are expected to survive long enough for valve-specific endpoints to really matter.

Lange and Hillis: In high-risk patients with severe AS, is the survival benefit of TAVR in the CoreValve study, but the lack of benefit in the PARTNER study, due to a difference in the studies or the valves?

Kirtane: That’s the question everyone will be asking, isn’t it? These are obviously two different types of prostheses from two different companies: Medtronic’s CoreValve is a self-expanding supra-annular prosthesis with a larger longitudinal dimension; the Edwards Sapien valve used in PARTNER is a balloon-expandable valve deployed within the annulus itself. In addition, both devices are iterating with newer, potentially more advantageous designs that may soon be available. In the absence of a head-to-head comparison of the short- and long-term performance of these valves, it is really difficult to speculate about valve-specific differences, except theoretically. Even “apples-to-apples” comparisons across the study programs are challenging, with different clinical-events committees, echocardiographic core laboratories, and study sites. Even the analysis populations differed: PARTNER A primarily reported intent-to-treat analyses, including patients who did not receive SAVR or TAVR, whereas the CoreValve trial primarily reported as-treated analyses. That is one reason that the presentation of the results of the CHOICE trial (also a late-breaker at ACC.14) will be fascinating, representing the first head-to-head randomized trial of these two devices.

Although both the current CoreValve trial and the high-risk cohort of the PARTNER trial randomized patients at high surgical risk, the populations enrolled in these trials were likely very different, precluding fair cross-study comparisons. As an obvious example, PARTNER A included patients for whom the study PIs “concurred that the predicted risk of operative mortality was ≥15% and/or a STS score of ≥10.” In the CoreValve study, the STS score was calculated but not specifically included in the enrollment criteria, which only specified agreement that the “predicted risk of operative mortality is ≥15%.” As a result, the mean STS score (predicted 30-day mortality) in PARTNER A was 11.7%, whereas it was 7.4% in the CoreValve trial. Similarly, the logistic EuroSCORE in PARTNER A was nearly double that in the CoreValve trial, and the mortality rate in the surgical arm of PARTNER A was higher at both 30 days and especially at 1 year versus in the CoreValve trial (26.8% vs. 18.7%). Interestingly, the 30-day mortality rate in the surgical arm of both trials was approximately 40% lower than that correspondingly predicted by the STS algorithm.

Also consider the comparative stroke data across trials. The CoreValve study authors point out that – in contrast to PARTNER – CoreValve conferred no increased risk for stroke compared with SAVR. However, this may have been due to the unexpectedly higher risk for stroke rate in the surgical arm (6.2% at 30 days and 12.6% at 1 year). Given theoretical differences in stroke adjudication across studies, the 1-year rates of stroke in the TAVR arms of both trials are actually quite comparable (6.0% in PARTNER A, 8.8% in CoreValve).

The bottom line is that it will be very difficult to know what led to the observed differences between the two trials and devices. I view this as a “win” for TAVR as a whole.

Lange and Hillis: Given that we lack trials that directly compare TAVR valves, how will you choose which valve to use in an individual patient?

Kirtane: Another great question. A great deal ultimately depends on operator experience and familiarity, as well as specific sizing concerns and limitations. In the United States, the CoreValve is currently available for commercial use in a greater range of annular sizes than is the Sapien valve. For patients who cannot be treated with one device, the availability of another device that can facilitate TAVR is welcome. For patients who could be treated with either device, much of the decision will come down to operator experience and device familiarity. There are sites, like ours in the U.S., that will implant both devices (and even newer devices in clinical trials), but there are minimum volume requirements that are not necessarily transferrable across devices.

CoreValve also has one additional noteworthy advantage: the smaller size of access vessels required for deployment. The CoreValve has a lower-profile design that requires a smaller sheath size for deployment, so a greater number of patients are potentially eligible for transfemoral treatment with this device. In addition, provided that care is taken to exclude patients with arteries that are too small for even the CoreValve, one would expect the rates of vascular complications to be lower than with the commercially available Sapien valve (the newer Sapien XT valve studied in the PARTNER II trial is still not commercially available in the U.S.). In the CoreValve randomized trial, the rate of major vascular complications was only 5.9%, the lowest rate observed to date in major TAVR trials.

Share your reflections on the latest CoreValve study and Dr. Kirtane’s analysis of it.

To view all of our coverage from the ACC meeting, go to our ACC.14 Headquarters page.

One Response to “TAVR: Are You Impressed Now?”

  1. Judith Andersen, AB, MD says:

    These are really helpful comments. It is unfortunate- and already commented upon by Cardioexchange readers — that no direct comparison can be made between CoreValve and Sapien XT results, since the newer Sapien valve, not yet available in the US, appears to represent a major upgrade from its predecessor. As always, operator experience and familiarity are critical to outcomes –but it is gratifying that both devices offer safe access to non-surgical care for fragile patients — and, specifically, options for patients with vascular access limitations and comorbidities.