Slideshow | TAVI for Low Risk Operable Population: FDA Perspective

Matthew G. Hillebrenner, M.S.E. Chief, Circulatory Support and Prosthetics Branch Division of Cardiovascular Devices Office of Device Evaluation

Washington, D.C.--Matthew G. Hillebrenner, MSE, chief of the FDA’s circulatory support and prosthetics branch, division of cardiovascular devices at the Office of Device Evaluation, spoke to the complications yet exciting progress with identifying how to trial and treat the low-risk operable patient population with trancatheter aortic valve implantation (TAVI), on Feb. 22 at the Cardiovascular Research Technologies (CRT) conference.

While the reality of assessing risk for TAVI procedures in a low-risk operable population has “not yet arrived, we’re a lot closer to it becoming a reality," Hillebrenner said. "We’ve started to discuss it as tomorrow, as opposed to years down the road when it comes to less sick patients.”

In determining the risk of TAVI in all patient populations, which he defined as high risk, non-risk, elevated risk, unacceptable risk and inoperable, Hillebrenner stressed the importance of collaboration between the interventional cardiology community and the surgical community.

In addition, he emphasized the importance of understanding the patients being enrolled in trials. He acknowledged the “difficulty” in putting a label on the different categories that the patients fall into because it's “usually dependent on the surgeon involved in the case, and his or her past experiences.” As a result, surgeons and interventionalists involved with these cases should have the same level of experience.

In discussing the STS (Society of Thoracic Surgeons) Risk Predictor, Hillebrenner said it helps with selecting patients but is not acceptable as a single-arm performance goal. Specifically, he said the all-comers database is “not perfect,” because:

• A finite number of variables can’t account for rare conditions (mediastinal radiation, scoliosis, previous trauma, etc.;
• Too many known and unknown covariates not captured;
• All comers (versus 20-50 inclusion/exclusions in clinical studies); and
• Does not account for surgeon/hospital experience or quality.

As a result, Hillebrenner said the FDA and clinical practice should continue to rely largely on randomized clinical trials.

However, he cautioned against relying upon an individual surgeon’s perspective in attempting to categorize these patients. Now, the surgeon’s perspective is combined with the STS risk score.

He added that this patient population is very difficult to categorize, as evidenced through Edwards Lifesiences and the FDA’s painstaking process leading up to the PARTNER trial.

“The collaborative process of attempting to identify these patients may only get more difficult as we get down to the lower risk threshold,” Hillebrenner said.

He identified some issues that will likely arise, as the low-risk threshold increases within clinical trials, especially for the control groups:

• Endpoint evaluation time likely longer (more than one year for primary endpoint), adding that mortality might not be the appropriate or the only endpoint for the lower risk group;
• May need co-primary endpoints or powered secondary endpoints with specific statistical plan for secondary endpoints defining “success”;
• Possible endpoints – survival, hemodynamics, echo dimensions, LV mass, function (VO2, 6MWT), aortic insufficiency, QOL;
• Totality of data important; and
• Large amount of missing data may make study uninterpretable.

In his conclusion, he said that the FDA is participating in Valve Academic Research Consortium (VARC) discussions, and it will continue to work with industry to reach the appropriate trial endpoints for this lower risk patient population.

“We are committed to being involved in the process for finding better solutions for these patients, when appropriate, and not allowing disagreements prohibit progress,” Hillebrenner concluded.