Data or Money: What’s Holding Back Cerebral Embolic Protection in TAVR?

CHICAGO, IL—A range of new retrospective analyses and single-center series are building the case for cerebral embolic protection in TAVR procedures, hinting at relative reductions in stroke or death approaching 70%.

But as several experts gathered here at the 2018 Structural Heart Disease Summit (SHDS) seemed to conclude today, the single most important thing that will convince operators to use protection devices will be reimbursement.

The conversation is timely. The US Centers for Medicare & Medicaid Services (CMS) has been mulling an expanded “add-on” payment to physicians performing TAVR that would help cover the costs of using cerebral protection. The open comment period for the CMS review ends Monday June 25, 2018.

The only cerebral protection device currently available on the US market is the Sentinel (Claret Medical). As previously reported by TCTMD, Sentinel received FDA clearance in June 2017 despite the fact that the SENTINEL trial failed to meet its primary efficacy endpoint of significant reductions in new brain lesion volume on MRI. At the time of the agency’s review, FDA advisors had mixed thoughts on the device efficacy but were nearly unanimous in agreeing that they themselves would want to have the device in place if they were undergoing TAVR.

Real-world Numbers

Three “snapshots” from single-center experiences presented this morning at SHDS 2018 help fill in some blanks for cerebral embolic protection. While not randomized, and with all the pitfalls of retrospective and propensity-matched analyses, all of them pointed to more and more stroke reduction as operator experience with cerebral protection improves.

Julia Seeger, MD (University of Ulm, Germany), presented 7-day follow-up in a propensity-matched analysis of 560 transfemoral TAVR patients, showing a 69% relative risk reduction (RRR) in mortality or stroke with cerebral embolic protection (2.1% vs 6.8%, P = 0.01) and a 70% RRR in disabling or nondisabling stroke (1.4% vs 4.6%, P = 0.03). The only predictor of stroke in subsequent analyses was lack of protection. In an additional, propensity-matched analysis that combined 363 patients from the randomized SENTINEL trial with 800 patients treated at the University of Ulm, Seeger et al documented a statistically significant 64% reduction in all procedural strokes (1.86% vs 4.95%, P = 0.0079) and a 68% reduction in mortality or stroke (1.86% vs 5.57%, P = 0.0022).

Next up, Nicolas M. Van Mieghem, MD (Thoraxcenter, Erasmus University Medical Center [EMC], Rotterdam, the Netherlands), presented protected (n = 294) versus unprotected (n = 453) TAVR patients from his center, showing significant differences in the rate of any neurological event (1% vs 5%, P < 0.01) and death or neurological events (2% vs 7%, P < 0.01) at 72 hours post-TAVR. Those numbers closely matched an analysis that combined EMC TAVR patients with those from the University of Groningen, also in the Netherlands.

Finally, Tarun Chakravarty, MD (Cedars-Sinai Medical Center, Los Angeles, CA), presented 7-day stroke rates among the 456 patients who have been treated with the Sentinel device at his center as compared with patients who did not receive it. Here the major stroke rates were 1.1% and 4.9%, respectively, yielding a number to treat to prevent one stroke of 26.

These three “snapshots” represent the kind of data that may convince CMS of the need for reimbursement for cerebral protection, said Hemal Gada, MD (UPMC Pinnacle, Harrisburg, PA), who commented on the data for TCTMD after this morning’s session.

“Pretty consistently we are seeing this 70% stroke risk reduction,” he said, adding that the Ulm data “show a palpable stroke benefit that’s really very similar to what I’ve seen at my institution, although I don’t have the numbers. In 108 patients treated with add-on Sentinel use at UPMC Pinnacle, he said, “we’ve only had two clinically significant strokes and I think that speaks for itself.”

If Price Was Not a Concern

Gada was one of the speakers during a February 2018 CMS Town Hall meeting discussing the proposed new technology add-on payment for cerebral embolic protection during TAVR. As noted, the open comment period on the proposed rules will close next week and the final rule will be published mid-summer. October 1, 2018, is the anticipated implementation date.

During this morning’s session at SHDS, Gada gave a presentation addressing cost-effectiveness in TAVR, from the perspective of the patient, the provider, the payer, and society/healthcare system more broadly. He believes the CMS final rule will be critical to wider use of cerebral protection, but as he argued in his talk today, using devices to prevent stroke during TAVR will likely pay off in the long term regardless of whether they are reimbursed.

Major stroke, Gada noted, is a key contributor to the cost related to TAVR—something that first became clear in the original randomized trials and has played out postapproval. During the index hospitalization, an acute ischemic stroke can add more than $25,000 to the cost of acute care and in patients who leave the hospital with moderate disability, this can be expected to increase annual healthcare costs by up to $60,000. Moreover, he continued, transitioning from mild cognitive impairment to mild dementia could add an annual excess cost of $6,000 per patient.

All of this eats into TAVR’s potential financial margins, which Gada called “small” and “shrinking.” Even a single major cerebrovascular accident could lead to a TAVR program no longer being profitable for hospitals, he said.

Offsetting this, however, is the cost to patients, hospitals, payers, and society if stroke is not mitigated. This will only grow as TAVR moves into lower-risk patients, Gada predicted. As he showed here, lower-risk patients are more likely to be younger and still working, and to have family members who depend on them financially. “A disabling stroke or even mild cognitive impairment in a younger, fully functional, employed patient would be devastating for a patient and for their loved ones,” Gada said.

His hospital, he continued, started using cerebral protection in every patient starting August 2017. Outcomes in these patients support his belief that cerebral protection can be cost-effective. Among the 108 patients treated to date in whom cerebral protection was used, the average modified contribution margin (MCM) with Sentinel is $15,804, “which is very similar to our MCM without Sentinel,” he noted. Average length of stay with Sentinel is 1.61 days, which is actually about 0.5 days less than patients treated without cerebral protection at Gada’s hospital.

“Patients, no matter where they go, deserve and expect excellent care, so we should invest in it,” he concluded, adding that positive patient perceptions have “definitely accelerated our program growth” and have improved patient satisfaction.

Speaking with TCTMD, Gada agreed that the “affordability reality” has been a key reason for the slow adoption of cerebral embolic protection. “Having the new-tech add-on in place is going to help people navigate that, but what I would advise people to do is not just look at the immediate episode of care. Look beyond it. Look at your readmissions and [ask]: are those stroke patients coming back and how frequently are they utilizing healthcare system? That’s a more realistic way of looking at the economic impact of stroke.”

The financial and societal burden of stroke cannot be underestimated, he continued. “We know that there are very few conditions as morbid in a 90-day window of time as stroke. Stroke will deplete people’s ability to function in a variety of different ways, physically, mentally, and emotionally. To have that kind of taxing burden—it’s not just on the patient and their loved ones, you are incorporating that into society as well because society is picking up the tab. That’s our tax dollars at work. I’d rather, for several reasons, have less of a burden associated with [TAVR], and there’s a way to do that.”

Gada said that his hospital managed to cover the cost of cerebral embolic protection by trimming back other procedural costs: moving to an optimized, minimalist procedure that did not require general anesthesia and allowed for shorter hospital stays. But he acknowledged that reimbursement will be key for other TAVR centers throughout the US to offer embolic protection during procedures.

“It’s a big benefit for our patients if we are able to incorporate this, and I don’t want disparities of care to pop up across the country because of this device not being used at some centers,” Gada said.

Indeed, the cost of using cerebral protection has come to dominate the conversation since the Sentinel device was approved, with numerous interventional meetings hosting debates on barriers to use. All have typically concluded with consensus that if cost wasn’t a barrier, both physicians and patients would want a cerebral protection device used.

The Structural Heart Disease Summit was no different. During a debate to wrap up this morning’s session, Axel Linke, MD (Heart Center, Dresden, Germany), argued in favor of the statement that cerebral embolic protection during TAVR should be used in all (or most) patients. John Forrest, MD (Yale University School of Medicine, New Haven, CT), was charged with arguing the “con” position, but in his opening “disclosure” statement he said that if he himself was having aortic valve replacement “I’d want cerebral embolic protection and I’d also want Dr. Linke to do the case.”

Stroke: A Constant Worry

Other presentations in this morning’s session served as a reminder that while stroke rates have declined as TAVR devices and technique have improved, they won’t disappear completely.

Alexandra Lansky, MD (Yale University, New Haven, CT), who opened the session this morning with an overview of stroke incidence in TAVR in the “modern era,” noted that the weighted average rate of major strokes has fallen from 4.2% with the first-generation devices studied to 3.1% with current-generation devices. Moreover, while the original PARTNER trial first raised concerns of higher stroke rates with TAVR, subsequent studies have allayed those fears. “In current randomized controlled trials,” said Lansky, “particularly in the intermediate-risk population, stroke rates are actually lower as compared with surgical AVR, and this represents truly an advantage of minimally invasive TAVR as compared to SAVR.”

That said, other data have made it clear that stroke rates do not appear to go down as operator experience goes up. “Experience is associated with a reduction in mortality and vascular complications but not stroke,” Lansky said. “So, stroke is probably an inherent part of this procedure.”

If some residual stroke risk during TAVR is inevitable, the best solution is better protection, she concluded.

“Stroke rates are devastating, unpredictable, and have remained constant despite lower-risk patients, improved TAVR devices, and operator experience,” she concluded. Moreover, “cerebral embolization and injury occurs in all patients,” as evidenced by MRI or in the capture device itself.

Another analysis of the PARTNER trials, this one presented by Chetan P. Huded, MD (Cleveland Clinic, OH), looked specifically at surgical aortic valve replacement (SAVR) compared with transfemoral (TF) TAVR, which has become the standard of care in 90% of patients, he noted. The analysis pooled 1,204 TF-TAVR patients from PARTNER 1A and B, PARTNER 2A and B, and PARTNER 2 S3—yielding a mix of Sapien (41%), Sapien XT (43%), and Sapien 3—who were propensity matched to 1,204 SAVR-treated patients.

What they found is heartening. Major stroke among TF-TAVR patients was 2.2% as compared to 3.9% for SAVR-treated patients (P = 0.018). Minor stroke, TIA, any stroke, and combined stroke/TIA were all not significantly different between groups.

Importantly, for both TF-TAVR and SAVR, the risk of stroke peaked within the first 24 hours, and both procedures were associated with similar, low, and constant late neurologic risk out to 48 months.

“I would say that our data supports the hypothesis that the most bang for your buck in preventing stroke in TAVR patients is at the time of the procedure.,” Huded told TCTMD. Studies of cerebral protection in surgical AVR have yielded “less promising” results.

Early trials as well as other retrospective registries have hinted that the stroke rate may be higher with TAVR. “That’s still a little bit controversial,” Huded said. “That’s one of the reasons we wanted to look into this a little deeper, with more contemporary devices and using a very matched SAVR-TAVR population. Our data support the idea that TAVR has a lower risk of stroke, and I think that’s very encouraging.”

What Comes Next

In panel discussions during today’s session, Susheel Kodali, MD (NewYork-Presbyterian/Columbia University Irving Medical Center, New York, NY), cautioned that it’s difficult to draw too much from registry experiences and propensity matching when patients were not randomized. Many of the same reasons for not using cerebral embolic protection, such as diffuse calcium, may also predispose patients to a higher stroke risk, he noted.

But he also raised the question of whether cardiologists are comfortable enough with observational data to adopt a new, expensive technology—reimbursed or not.

“The reality is cost is a factor, the majority of centers both in the US and Europe are not using embolic protection, and unless we do another large randomized trial we’re not going to get a class I indication,” Kodali commented. “So how do we move forward from here?”

Lansky responded that it would be “fantastic to have a 2,000-patient clinical trial showing a significant percent reduction in stroke, but we cannot expect a Claret or anyone of these companies to actually foot the bill.”

And would physicians even want to randomize patients, given the growing body of evidence in favor of protection? Lansky thinks they would. “I would presume that they would because to your point, many of these centers are not using cerebral embolic protection, so if they are not, the question is why not? And the answer we get is that there is not enough evidence out there,” she said.

Propensity matching leaves observers plenty to “quibble” over, Lansky continued, but as the three “snapshot” presentations make clear, “we’re seeing potentially very significant reductions in stroke and we’re seeing benefit from cerebral embolic protection. So at least currently, in the absence of a randomized trial, I think the evidence is very strong.”

One way to find out whether it’s data or money that’s holding back this field will be to see what happens, assuming the reimbursement falls into place, Kodali said. “If the new add-on technology payment goes through and the number of centers goes from 90 to 400 in the US, we’ll find out what people really thought: that it was not an evidence issue, it was a cost issue.”