PARTNER A: TAVR Noninferior to Surgery, But Stroke May Be A Problem

Updated with commentary at end of article– Here are the main results of the much-anticipated PARTNER A trial comparing transcatheter aortic valve replacement (TAVR) versus surgery for aortic valve replacement (AVR).

699 high-risk older patients with severe aortic stenosis were randomized to either TAVR or AVR.  The primary endpoint, all cause mortality at 1 year, was 24.2% in the TAVR group versus 26.8% in the AVR group (HR 0.93, CI 0.71-1.22, p+0.62), thereby meeting the prespecified margin for noninferiority. 30 day mortality was 3.4% versus 6.5% (p = 0.07).

The rate of major stroke rate at one year was 5.1% for TAVR versus 2.4% for AVR (p=0.07). For all strokes the difference achieved statistical significance: 8.3% versus 4.3% (p=0.04). There were no significant differences at either 30 days or 1 year in cardiac mortality, rehospitalization, MI, or acute renal injury requiring renal replacement therapy.

Major vascular complications occurred more frequently in the TAVR group, both at 30 days (11% versus 3.2%, P <.01) and at 1 year (18% versus 4.8%, p<0.01). Major bleeding, on the other hand, occurred more often in the AVR group, both at 30 days (9.3% versus 19.5%, p<0.01) and at 1 year (14.7% versus 25.7%, p <0.01)

New AF occurred more frequently in the surgery group: 12.1% versus 17.1% at 1 year (p=0.07).

The rate of all-cause mortality or stroke was 26.5% for TAVR versus 28% for AVR (p=0.70).

The authors concluded:

“Both TAVR and AVR were associated with important but different peri-procedural hazards: Major strokes at 30 days and one year and major vascular complications were more frequent with TAVR. Major bleeding and new onset atrial fibrillation were more frequent with AVR. TAVR and AVR are both acceptable therapies in these high-risk patients; differing peri-procedural hazards may impact case-based decision-making.”

Over on CardioExchange, Rick Lange and David Hillis question the PARTNER A press release which cites TAVR as an “excellent alternative” to AVR because it was associated with less atrial fibrillation (8.6% vs 16%) and bleeding (9.3% vs 19.5%). “This is an interesting conclusion,” they write, “since most physicians and patients are more concerned about periprocedural stroke and vascular complications than atrial fibrillation or transfusions.”

Comment: It was hard not to be stirred by the high praise showered on Craig Smith and the PARTNER investigators at the big session today– and I’m only being slightly ironical here. The fact is, TAVI (despite the  use of “TAVR” today everyone seems to have reverted to using “TAVI,” and it’s a lot easier to say) represents a major advance, and though the comparisons to CABG and angioplasty were perhaps a bit overdone, PARTNER does represent one of those rare occasions where the word “breakthrough” is actually justified.

However, sometimes people’s minds get a little fuzzy when they encounter this kind of terminology, so I want to offer an important distinction that may have been missed in some of the discussion. Although TAVI may be a legitimate breakthrough, that doesn’t mean that it will immediately revolutionize (another word that should be used carefully) clinical practice. From my conversations with a number of thoughtful cardiologists today, it appears that in the immediate future the clinical application of TAVI is likely to be somewhat limited. For instance, a significant number of patients who now get AVR are at much lower risk than patients in PARTNER, and there is no evidence yet to support TAVI use in this population. And many patients get AVR in conjunction with CABG, so these patients will not be eligible for TAVI either. Further, as was seen in the taped case presentation, TAVI is an extremely high tech procedure, requiring the facilities and personnel of advanced surgical, catheterization, and imaging laboratories.

It’s likely that following the excitement and hype surrounding PARTNER, some may rush in to offer TAVI before they are ready. The results might not be so pretty.


  1. I’m not a science expert, but saying that there were more strokes in the TAVR arm of the study is meaningless without a P value.

    Was the risk STATISTICALLY significant or not?

    Unless there’s a statistically increased risk of stroke in the TAVR group, I’m inclined to say that TAVR is a better option than AVR considering that TAVR is much less invasive with a faster recovery time.

  2. Quick answer: the overall stroke and TIA risk was statistically significant. The major stroke risk was borderline (p=0.07), as reported above.

  3. Angie Long says

    And is there anyway to determine how any of this compares to “do nothing?”

  4. Elaine Duryea says

    How common are renal problems after this surgery? My uncle just had the procedure done and he went into renal failure in the first 24 hrs –the doctors said it was from the dye used in the procedure??

Speak Your Mind