Shelley Wood

April 03, 2011

Updated April 3, 2011 (New Orleans, Louisiana) — Long-awaited results from PARTNER cohort A, comparing transcatheter-valve implantation with surgery for severe aortic stenosis, show that the new catheter-based procedure is just as good as surgery in surgery-eligible patients for the primary end point of mortality, with a range of secondary end-point findings that will give surgeons and interventionalists plenty to debate in the days to come.

Principal investigator Dr Craig Smith (Columbia University, New York, NY) presented the one-year results here at the American College of Cardiology 2011 Scientific Sessions/i2 Summit today. Those results show that PARTNER cohort A, conducted in high-risk aortic-stenosis patients deemed eligible for surgery, met its primary end point, demonstrating that TAVI is noninferior to traditional aortic-valve replacement in terms of all-cause mortality.

"These results indicate that [transcatheter aortic-valve replacement] TAVR is an acceptable alternative to AVR in selected high-risk operable patients," Smith concluded. "Future randomized studies should focus on lower-risk patients who are candidates for operation."

The PARTNER cohort A results were met in today's late-breaking clinical trials session with the same level of exuberance that greeted the PARTNER cohort B results at the 2010 TCT meeting. One of the panel members discussing the presentation, Dr Martyn Thomas (London, UK), called it "an absolutely spectacular trial." Dr David Moliterno (University of Kentucky, Lexington), another panelist who was not involved in the study, told members of the press: "You really are witnessing history in the making. This is one of the biggest steps in cardiovascular medicine in our lifetime."

Nuts and Bolts of PARTNER A

At 30 days, deaths were numerically lower in the TAVI group, but not statistically different. By one year, deaths in both groups were nearly identical and met the predefined definition of noninferiority.

All-Cause Mortality

End point TAVI (%) Surgery (%)
30-d mortality 3.4 6.5
1-y mortality 24.2 26.8


Major strokes--the secondary end point that has been provoking the most speculation in the lead-up to this presentation--were higher in the TAVI-treated patients, both at 30 days and one year. In a combined end point of all stroke or transient ischemic attack (TIA), the difference between groups was statistically significant. Additional details on stroke types and etiology will be presented at the upcoming American Thoracic Society meeting, Smith noted.


End point TAVI (%) Surgery (%) p
Major stroke, 30 d 3.8 2.1 0.20
Major stroke, 1 y 5.1 2.4 0.07
All stroke or TIA, 30 d 5.5 2.4 0.04
All stroke or TIA, 1 y 8.3 4.3 0.04


Not unexpectedly, major bleeding was more than twice as common in surgery-treated subjects (19.5% vs 9.3%, p<0.001); new-onset atrial fibrillation was also nearly double in the surgery group (8.6% vs 16%, p<0.001). Vascular complications were nearly four times higher in the TAVI group (11% vs 3.2%, p<0.001). Symptom improvement (NYHA class and six-minute-walk test) was greater in the TAVI group than the surgery group at 30 days, but no different at one year.

In all, 699 elderly patients (median age 84.1) with severe aortic stenosis were randomized to either TAVI or conventional surgery at one of 26 centers in the US, Canada (three centers), or Germany (one center). In the TAVI group, patients received the experimental device via the fully transfemoral route (244 patients) or via a transapical procedure (104 patients), which is typically used when the femoral artery is deemed unsuited to the TAVI catheters. Not surprisingly, patients in the transapical group were slightly higher risk than patients in either the transfemoral TAVI group or the aortic-valve-surgery group. Of note, 28 patients in the surgery arm either refused treatment or withdrew from the study after initial randomization; by contrast only one person in the TAVI arm refused treatment, and none withdrew between the time of trial enrollment and treatment. As such, findings in the intention-to-treat analyses were slightly different from those in the as-treated analyses, although Smith stressed that these differences did not materially affect the main trial findings.

Investigators React

Speaking with heartwire , Dr Michael Mack (Medical City Dallas Hospital, TX), a surgeon and PARTNER investigator, characterized the trial as "a win for everybody. Surgery looked good; transcatheter-valve replacement looked good."

Dr John Webb (St Paul's Hospital, Vancouver, BC) one of the early percutaneous-valve pioneers, also hailed the findings but clearly viewed TAVI as the winner.

"This is a trial of a first-generation transcatheter-valve system, and with operators, many of whom had only done two cases before doing this trial, compared with the very best surgeons, at the very best sites in the United States--a very high bar. And [the patients enrolled] were really just the top 10% of surgical candidates--the worst patients. Despite that, you have a mortality rate that's lower [numerically] with transcatheter-valve implantation. The study was powered as a noninferiority trial, and it met that benchmark."

The Stroke Results

Mack singled out the difference in stroke rates as "the big differentiator" that he expects people will zero in on. Strokes were also a major talking point in the PARTNER cohort B trial (TAVI vs best medical therapy), although in that group, patients ineligible for surgery, a stroke seemed a smaller price to pay for a huge mortality gain. In cohort A-type patients, stroke will be an even more important consideration.

During today's late-breaking clinical-trial session, Dr Robert Bonow (Northwestern University, Chicago, IL) pointed out that stroke is the thing is patients worry about most when considering going forward with valve treatment. Smith, likewise, characterized stroke as something these elderly patients fear more than death itself.

"You may pay a higher risk of stroke with transcatheter aortic-valve replacement, but you have a quicker recovery with TAVI, and you end up at the same point at one year," Mack says. That said, he thinks people were too fixated on the stroke rates in PARTNER cohort B. "It's almost like the primary focus got lost because of the stroke issue."

Webb, for his part, is less concerned about the stroke rate, which he says will fall considerably as lower-risk patients are treated, as operators gain more experience, and with the newer devices now available. The first-generation Sapien (Edwards Lifesciences) device and delivery system were used in PARTNER; his center is now using a fourth-generation device that is much less bulky.

A number of experts also pointed out that the stroke rate in cohort A is lower than that of cohort B, which involved the same enrolling centers but wrapped up earlier, suggesting that the stroke rates were already declining as operators gained more experience.

Not only will stroke rates decline, Webb predicts, but vascular adverse events will drop dramatically. Compared with the 11% rate in this study, major vascular complications at Webb's hospital are now down to 2%.

Speculating on Subsets

Webb also pointed to what he thought were important subset findings--the rate of the primary end point in the patients treated with the transfemoral procedure. In an analysis looking at patients "as-treated," rates of the primary end point at 30 days in the transfemoral subset were just 3.7%, vs 8.2% (p=0.05) in those treated with surgery.

"It's hard to make a noninferiority study sound exciting, but the point is that, overall, TAVI was better than equivalent, it was noninferior, it had a lower mortality [numerically], and the percutaneous group had a dramatically lower mortality," Webb said.

Smith, asked about this finding after the press conference this morning, agreed that it was an "exciting" finding.

An open question remains valve durability and rate of paravalvular leaks with transcatheter valves. "We do have an idea of what valve durability is with surgery, and that's 15 to 20 years," Mack reminded heartwire , "and we have virtually no paravalvular leaks. That may or may not be significant in the long term for transcatheter valves; right now, so far, it doesn't seem to be a major issue."

What the Future Holds

Industry analysts have predicted the FDA may approve the Sapien device as early as late fall 2011, at least for surgery-ineligible patients. The device is already CE Mark–approved in Europe and on the market in other countries worldwide. A US randomized clinical trial of the CoreValve (Medtronic) device--the other major transcatheter valve in clinical testing in the US and approved in Europe--now has 35 sites "activated." As excitement mounts in the US, Mack has one more reminder for physicians.

"What we have learned through this trial is that there are clearly patients who are not candidates for surgery now who can be helped by this, but we've also learned that there is a ceiling to this, and not everybody is a candidate for transcatheter-valve therapy. I think we have found through both trials, A and B, that there were patients that we enrolled early on that we would not enroll today, because they were too old, too frail, or too debilitated, and they can't even withstand a transcatheter procedure."

He continued: "I also think that, as great as the results for cohort B were, and as great as these results are from cohort A, you still have a therapy in which between 25% to 35% of patients are dead at one year. So you have to ask, is that an appropriate use of expensive technology?"

Going forward, says Mack, surgeons and interventionalists, in collaboration, need to do a better job of recognizing when comorbid diseases are too advanced.

"We really need to be more selective about who we think has the greatest chance of surviving one, two, three, and five years out. I would be disappointed if, in the next series, the one-year survival doesn't climb up to 80%, 85%, and 90%."

It should do that, he notes, as lower-risk populations are treated. "If you look at the one-year survival of all patients undergoing valve-replacement surgery in the US today, it's 93%. So I think that's the target to aim for."

During the morning press conference, several of the experts assembled, including Smith, stressed the need for the multidisciplinary "heart-team" approach in deciding on what procedure was best for each patient. Dr Murat Tuzcu (Cleveland Clinic, OH) warned that, assuming the device is approved, "if we don't take note of how we collaborated in this trial, we won't get as good results."

Others noted that patients "never want surgery," underscoring the need for physicians to thoroughly and "honestly" convey the true known risks and benefits of each procedure to patients--many of whom will want a less invasive procedure.

Cohort B and a Publication Scuffle

In the PARTNER cohort B comparison, the relative risk reduction in all-cause mortality was 46% compared with medical therapy (a 20% absolute difference), and cardiovascular mortality risk was reduced by 61% compared with patients receiving the best medical care, including balloon valvuloplasty. Risk of a composite end point of dying or needing repeat hospitalization was also lower in TAVI-treated patients. In that arm of the trial, the risk of stroke and of vascular complications was higher in the TAVI group.

In a controversy that still rankles, the PARTNER cohort B investigators had agreed to simultaneously publication of their results in the New England Journal of Medicine when the presentation was made at TCT on a Thursday morning. Inexplicably, the journal released them early online on Wednesday evening.

Speaking with heartwire , co–principal investigator for PARTNER, Dr Martin Leon (Columbia University), said that the cohort A results were not simultaneously published publication "by intention."

"One of the reasons we did that was that when we worked so hard to get [cohort B] published simultaneously, it was prereleased, and it took a lot of the shine out of the presentation and was frustrating for us, and we could not convince the journal to hold off on that. So it was such stress for everybody, having data that were potentially being exposed, that I think we wanted to reduce the stress level. So it has been submitted to the NEJM but will not be presented and published simultaneously."


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.