Closer Look at SAVR Patients in PARTNER 2A Reveals a Quarter Also Had Additional Surgeries

Patients in the PARTNER 2A trial who were randomized to surgery rather than transcatheter aortic valve replacement ultimately faced an operative mortality rate that was lower than what their STS score predicted going in, a new analysis suggests. Investigators for the study say their detailed look at the operative outcomes and comorbidities among the 937 intermediate-risk individuals in the surgical aortic valve replacement cohort should be reassuring for patients who ultimately receive surgery rather than a transcatheter treatment for their aortic stenosis.

Mortality rates, according to Vinod Thourani, MD (Emory University School of Medicine, Atlanta, GA), who presented the findings earlier this week at the Society of Thoracic Surgeons (STS) 2017 meeting in Houston, TX, were “commensurate with national benchmarks.”

But in a finding that is likely to spur discussion in the months to come, Thourani et al’s analysis also showed that a quarter of the patients randomized to receive SAVR in PARTNER 2A also concomitantly underwent some other form of cardiac surgery. In all, 198 patients in the surgical arm of the trial also received these additional surgeries, mostly CABG but also root replacement, root enlargement, and mitral valve procedures. The observation begs the question of how outcomes among the isolated SAVR patients compared with those of the transfemoral TAVR patients. Here, however, Thourani told TCTMD the numbers are too small to permit a “statistically sound” comparison.

“Those patients who had other procedures done, not the isolated AVRs, had a higher stroke rate and higher mortality,” as well as longer procedure times, Thourani said. And while differences between the subsets were not statistically significant, he told TCTMD, “I think that is potentially important and things for us to take into consideration.”

Scrutinizing SAVR

As previously reported by TCTMD, PARTNER 2A randomized 2,032 intermediate-risk patients with severe symptomatic aortic stenosis at one of 57 centers to either open SAVR or TAVR with the Sapien XT device (Edwards Lifesciences). Death or disabling stroke at 2 years—the trial’s primary endpoint—was not significantly different between groups in both an intention-to-treat and as-treated analysis. Of note, a prespecified subanalysis looking only at the 76% of TAVR patients treated by a transfemoral approach found TAVR was superior to SAVR, although this finding, as co-principal investigator Craig Smith pointed out at the time, was of “borderline significance” and underpowered. Subsequently, the SAPIEN 3 observational study showed the next-generation, balloon-expandable Sapien 3 transcatheter aortic valve (Edwards Lifesciences) to be superior to surgery among intermediate-risk patients in a propensity-score analysis.

In the detailed review Thourani presented at the STS meeting, procedural-related mortality among SAVR treated patients was “better than expected” at 4.1%. One third of patients were found to have patient-prosthesis mismatch (PPM), although at least out to 2 years, PPM did not appear to have an impact on mortality. To TCTMD, Thourani predicted that this may arise as a bigger problem when more time has passed. “We have to keep watching these patients,” he said. Of note, a separate study released at STS 2017 showed that moderate and severe PPM after surgical AVR were associated with lower survival through 10 years of follow-up.

In-hospital stroke in Thourani et al’s analysis was also higher than expected at 5.4% overall, slightly above what was seen in the patients who underwent isolated AVR (5.3%) and twice what trialists had expected when the study was designed, although strokes overall were numerically more common in the surgery-treated patients than in TAVR-treated patients. During his presentation, Thourani noted that all patients had neurological assessment postoperatively, so the higher than expected numbers may just speak to the rigor of the follow-up testing.

Finally, there were no differences in mortality or stroke between patients treated with minimally invasive AVR versus full sternotomy for patients with isolated AVR.

An Intriguing Piece of the Puzzle

One piece of the puzzle left unanswered is how isolated SAVR might compare to transfemoral TAVR. On this, Thourani was explicit: “I specifically don’t make any comparisons to transcatheter valves.”

He continued: “What we can say is that the results of isolated SAVR in intermediate-risk patients is excellent at 2 years, and when patients require concomitant procedures to aortic valve replacement . . . clearly they [face] a higher risk [than that] associated with isolated AVR.”

Commenting on the abstract for TCTMD, Timothy Gardner, MD (Christiana Hospital, Newark, Delaware), who clarified that he did not see Thourani’s presentation in person at STS 2017, said, “I do agree that with almost a quarter of the SAVR patients having more extensive procedures than AVR only, any implied comparison of outcomes with TAVR patients is not valid and reliable. It also is important that the actual randomization to TAVR and SAVR in this intermediate-risk PARTNER 2A cohort was really free of case selection and, therefore, potential bias.”

But reiterating a point made by Thourani, he added that the aim of this analysis was to take a more detailed look at the SAVR arm, not to compare the TAVR and SAVR groups.

Asked whether an apples-to-apples study might be warranted—for example, one comparing isolated AVR versus transfemoral TAVR to clarify whether the two procedures deliver equivalent results—Gardner said, “Such a ‘clear’ comparison would be ideal, but looking at the European trends with TAVR and the growing enthusiasm for TAVR in North America, it is unlikely that we will see such a comparative study done.”