Choosing TAVR

With transcatheter aortic valve replacement (TAVR), proper patient selection plays a key role in successful outcomes. It may prove to be a linchpin in reimbursement decisions as well.

In November 2011, TAVR hit the ground running in the U.S. with the FDA’s approval of the Sapien aortic valve in inoperable patients with symptomatic aortic stenosis. Less than a year later, the agency gave it another boost by including high-risk patients. The momentum continued into 2013 with encouraging results from the pivotal CoreValve Extreme Risk Iliofemoral study, presented in late October at the Transcatheter Cardiovascular Therapeutics scientific session in San Francisco.

Just as PARTNER (Placement of AoRTic TraNscathetER Valve Trial) data provided the evidence base for approval of Edwards Lifesciences’ Sapien valve, CoreValve Extreme Risk findings may pave the way for commercialization in the U.S. of Medtronic’s CoreValve. The 12-month results—with higher survival and lower stroke rates than in PARTNER—prompted the FDA to announce it would not convene an advisory panel to help it reach a decision on the device’s safety and efficacy. The decision is expected sometime this year.

Patient selection, crucial in clinical trials, is likely to remain a critical factor in the continuing rollout of TAVR, too. 

Recognizing futility

In TAVR, the lines between risk groups can become blurred, particularly at the upper and lower boundaries, according to cardiac surgeons who participated in the two trials. And the grimmest of the gray areas may be inoperable patients with severe aortic stenosis who will die if left untreated but who will not benefit from treatment, either. Michael J. Mack, MD, a PARTNER investigator and medical director of cardiovascular surgery of the Heart Hospital Baylor Plano in Texas, describes these as “the toughest clinical decisions we are making: who is not a candidate for a TAVR because they are too old, too debilitated, too frail and despite the fact that they would get a successful procedure, are not going to have a significant survival benefit.”

These patients may have comorbidities such as severe chronic obstructive pulmonary disease and renal failure in addition to aortic stenosis, says Michael J. Reardon, MD, a cardiothoracic surgeon at Houston Methodist DeBakey Heart & Vascular Center in Texas and a CoreValve investigator. “I fix your valve but you still are so sick from your bad lungs and your bad kidneys that you don’t get better,” he says.

With years of experience, cardiovascular specialists are getting better at recognizing factors that contribute to futility and a poor outcome. Mack lists a Society of Thoracic Surgeons (STS) predicted operative risk score in the high teens to low 20s, significant renal disease, frailty, being wheelchair bound and being on home oxygen as flags that help identify some patients who are not good TAVR candidates. “It is trying to determine where that upper boundary is between patients who are dying with aortic stenosis vs. those who are dying from aortic stenosis,” he says.

David Cohen, MD, a cardiologist and medical director of the Medical Economics and Technology Assessment Group at Mid America Heart Institute in Kansas City, Mo., adds that physicians should consider quality of life as well as survival when choosing TAVR patients, a perspective shared by Mack and Reardon. Patients in their 80s are more likely to be concerned about quality issues such as independence during the period left in their lives than would a younger person, the thinking goes.

“We believe they [survival and quality of life] are both important and different patients may value the different endpoints differently, depending on their circumstances and priorities,” says Cohen, who has conducted analyses in both trials. “That is why we advocated for endpoints that included both survival and a good quality of life as being optimal for patients after TAVR.”

[[{"fid":"18584","view_mode":"media_original","type":"media","attributes":{"height":268,"width":440,"style":"width: 440px; height: 268px;","alt":" - TAVR Reimbursement","class":"media-element file-media-original"}}]]

Using CoreValve Extreme Risk data and quality of life tools, Cohen and his colleagues identified predictors of a good or poor outcome. Preliminary data showed that being wheelchair bound, on home oxygen, having a low albumin, a higher STS mortality risk score, a previous bypass operation and a low aortic valve gradient were associated with a worse outcome at six months. He adds that categories such as frailty and disability were more important than individual factors, the research was preliminary and predictors might change as researchers gather more data.

“We are trying to provide some quantitative ammunition for the clinician and also for the patients, to give them the most realistic expectations,” he says. “They all come in very optimistic. We like to be optimistic as well but we also want to be realistic.”

Understanding the threshold between inoperable and futile also has financial implications. An analysis of TAVR patients at Duke University Medical School in Durham, N.C., presented at the 2013 American Heart Association scientific session in November, pegged the total average cost for TAVR at $84,471 per patient.

“This is expensive technology and if we apply it to people who won’t benefit from it, it will kill off the technology,” Reardon says.

Risk & reimbursement

The distinction between inoperable and high-risk became less of an issue with FDA’s decision in 2012 but physicians still need to carefully assess the line between high risk and intermediate risk. One reason this is challenging is the use of the STS risk score, which has been validated in the surgical aortic valve replacement patient population but not TAVR. “However, there are things that make a [TAVR] patient high risk that are not reflected in the STS scores,” Mack says.

To be eligible for Medicare reimbursement for TAVR, a facility must submit data to the STS/American College of Cardiology Transcatheter Valve Therapy registry. Mack and colleagues used that database to peek into practice patterns after Sapien’s approval through May 2013 (JAMA 2013;310[19]:2069-2077). They found that success rates and complications at 30 days mirrored results from clinical trials, alleviating possible concerns about safety. But they also noted that the STS risk score was lower than in the trials, with a median score of 7 vs. 11. 

“It could be risk creep going on,” Mack says, referring to the possibility that patients with lower surgical risk had been treated with TAVR, which would contribute to better outcomes. Other explanations may include patient characteristics not adequately captured in the STS score, underpredicting by centers unfamiliar with the STS predictor or undercoding what the true risk is. “There is no way of knowing that at this stage.”

After Sapien’s approval, the Centers for Medicare & Medicaid Services (CMS) issued a National Coverage Determination (NCD) that spelled out requirements for reimbursement. It warned that treatment in patients other than those indicated—inoperable and high risk only—would be denied, with the exception of patients enrolled in clinical trials. Mack says CMS is aware of the registry data and is watching developments.

“It is a big unknown as to what the CMS enforcement will be and how large of a concern that is,” he says. “Certainly, the concern is out there that RAC [Recovery Auditing Contract] auditors would look at the practice of TAVR and see how closely or loosely the NCD is being followed. It is total speculation at this stage.”