Limited Populations, Lost Patient Follow-Up Impact TAVR versus SAVR Data

Three pivotal trials—Evolut, PARTNER 3, and PARTNER 2A—have shown that transcatheter aortic valve replacement (TAVR) was noninferior to surgical AVR (SAVR) in terms of mortality and stroke among low-risk patients with symptomatic severe aortic stenosis. But speakers at Sunday’s session on the evolving impact of TAVR said that many questions about TAVR and SAVR remain unanswered.

Michael J. Reardon, MD, from Houston Methodist Hospital in Texas, emphasized that TAVR may be preferred over surgery for low-risk patients, but only in the population tested in the Evolut trial. The mean ages for Evolut and PARTNER 3 were similar (74 years and 73 years, respectively), while the PARTNER 2A population had a mean age of 82 years.

“We did not test enough young people,” said Dr. Reardon. Physicians must emphasize to younger patients that the results of these trials should not be applied to them, he added.

Another issue is the number of patients who were excluded and those who withdrew from the studies. Dr. Reardon highlighted two key exclusion criteria for Evolut: bicuspid aortic valve (verified by imaging) and coronary artery disease. PARTNER 3 had these same exclusion criteria. Excluding patients with these clinical features created “knowledge gaps,” said Dr. Reardon.

He said that his “biggest regret” about Evolut is not knowing why 260 of 1,728 patients (14.8%) were screened out at the national level. He said that it would have been better to allow these patients to stay in the study and have their data collected in a registry.

Vinod H. Thourani, MD, from Piedmont Heart Institute in Atlanta, reported the 5-year outcomes for the PARTNER 2A trial. He noted that a disproportionate number of SAVR patients withdrew from the study. At the end of 5 years, 91% of patients in the TAVR group and 81% in the surgery arm completed follow-up. “I consider this a limitation of the study,” said Dr. Thourani.

In reporting data from PARTNER 3, Gilbert Tang, MD, from the Mount Sinai Health System in New York, compared the findings of all three studies, saying the data indicate that surgery is better in terms of new left bundle branch block, mild pulmonary vascular resistance, and valve gradients; SAVR and TAVR are similar in terms of the need for new pacemakers, vascular complications, moderate-to-severe pulmonary vascular resistance, and coronary obstruction; and TAVR is better in terms of mortality, stroke, rehospitalization, acute kidney injury, severe bleeding, new-onset atrial fibrillation, 30-day quality of life, and length of stay (hospital and ICU).

Nick Freemantle, PhD, a biostatistician from University College London, discussed flaws in the methodology of the trials. For example, he said that the non-inferiority boundary on the primary endpoint of 6% includes many clinically important values for major morbidity and mortality. He also pointed to biases related to as-treated (per-protocol) analysis and inadequately concealed randomization.

In light of these issues, he stated, “TAVR may or may not be a safe and effective alternative to SAVR.”