Stroke Risk With TAVR: In-Depth Review of PARTNER and CoreValve Data Reveals Opportunities and Unknowns

Two comprehensive analyses of cerebrovascular events in the major randomized controlled trials of transcatheter aortic valve replacement are shedding new light on both the timing and risk factors for stroke in this setting. They also underscore the many unknowns the field is confronting as it moves into younger and lower-risk patients.

The recent papers are just the latest efforts to try to identify ways of mitigating stroke risk in patients undergoing TAVR and in the months following, but also, potentially, to identify patients whose preprocedural stroke risk makes them unsuitable for the procedure or deserving of more attentive follow-up.

The two papers appeared online September 6, 2016, ahead of print in the Journal of the American Heart Association.

In the first study, Neal Kleiman, MD (Houston Methodist DeBakey Heart and Vascular Institute, TX), and colleagues review the rates of stroke, and its predictors, among the 3,687 patients treated as part of the two CoreValve pivotal trials and their continued-access study, and followed for a total of 3,581 person-years. They report that the 1-year stroke rate post-TAVR using the self-expanding CoreValve prosthesis (Medtronic) was 8.4%, with just under half of those events occurring in the first 10 days.

A somewhat similar analysis, led by Samir Kapadia, MD (Cleveland Clinic, OH), reviewed rates of cerebrovascular events among 2,621 participants in the PARTNER trial (testing the balloon-expandable Sapien valve; Edwards Lifesciences) or the continued-access registry who were followed out to 3 years. They report a stroke incidence of 3.3% at 30 days (of which 85% occurred within 1 week). For patients treated via a transfemoral approach, these rates were 3.8% at 30 days, 5.4% at 1 year, and 6.9% at 3 years.   

Of interest, the two papers identified different predictors of stroke. In the CoreValve series, preprocedural predictors of early stroke included National Institutes of Health Stroke Scale (NIHSS) score greater than zero, prior stroke, prior TIA, peripheral vascular disease, angina, low BMI, and a history of falls in the previous 6 months, but no history of previous coronary artery bypass surgery. Predictors of stroke related to the procedure itself included longer procedure times, longer delivery catheter times, use of rapid pacing during valvuloplasty, and need for prosthesis repositioning. Beyond 10 days, the major predictors of stroke were small body surface area, severe aortic calcification, and a history of falls within the previous 6 months.

In the Sapien series, however, different risk factors emerged. Higher pre-TAVR aortic valve peak gradient was the key risk factor for stroke following transfemoral TAVR, whereas more postdilations, pure aortic stenosis without regurgitation, and “possibly” more pacing runs, earlier date of procedure, and lack of dual antiplatelet therapy were identified as risk factors for stroke following transapical procedures. Of note, this last association didn’t hold up in further analyses. For strokes occurring at 30 days or later, higher stroke risk scores, dementia, smaller TAVR valve size, nonwhite race, lower ejection fraction, and atrial fibrillation all emerged as potential risk factors.

Different Stroke Predictors?

Commenting on the two papers for TCTMD, George Dangas, MD, PhD (Icahn School of Medicine, Mount Sinai, New York, NY), admitted he was “a little astonished by how different [the stroke predictors] were.”

Dangas, with Gennaro Giustino, MD (Icahn School of Medicine, Mount Sinai), authored an editorial accompanying the two papers. Dangas said he believes that the differences don’t necessarily reflect the fact that the two devices carry different intrinsic risks or enrolled different patients, but rather that the original studies were not well designed to capture and evaluate stroke risk in a systematic way. Even the definitions of stroke, as well as who was tasked with diagnosing and following up on stroke rates, were different between trials. For example, the editorialists write, use of the NIHSS was applied differently in the PARTNER and CoreValve series and the modified Rankin Scale was used by Kleiman et al, but not Kapadia et al.

Dangas, to TCTMD, said that there is nothing in these papers that would exclude a patient from undergoing TAVR, with the exception of evolving stroke—and these patients were excluded from the pivotal studies.

“Is there something that's going to tell me, don't even try TAVR because a stroke is imminent?” he asked.  “I don't think that decision is helped by this data, and I don't know that there is a condition that one can say, with this, TAVR should not be done because of stroke risk . . . . In the regular population, the benefit of TAVR is so tremendous that it's difficult to say I’m not going to do this because of a stroke risk.”

Modifying Risk Postprocedure

That leaves the possibility of modifying the stroke risk during or postprocedure and here, says Dangas, there are three avenues to consider: the valve and delivery systems themselves, add-on devices for cerebral protection, and concomitant pharmacotherapy.

The devices themselves, said Dangas, have already changed substantially since the original trials were conducted. “Devices are smaller and more flexible, with smoother materials—so many of the improvements from the device point of view have already happened and hopefully will continue to happen.

As for cerebral protection using filters, baskets, or deflectors, these trials are still ongoing, said Dangas. He added that he believes the carotid stenting space, where protection devices led to improved outcomes, offer a clue to what the future may hold for TAVR, although this remains to be established. “There is a lot of theory and hope, but not a lot of proof so far.”

Results of the SENTINEL randomized trial evaluating a filter-based cerebral protection device (Claret Medical) are slated for presentation at the upcoming TCT 2016 meeting.

“The third thing is, is there something we can do with the pharmacology that could prevent stroke?” Dangas said. This question has risen to the forefront of TAVR discussions in the wake of observations of leaflet mobility and thrombosis concerns. In the pivotal trials to date, both periprocedural and postprocedure pharmacotherapy were not standardized, patients were not randomized to different strategies, and the use of antiplatelet and anticoagulant drugs was typically left to the treating physician, with no requirement to document strategies in a rigorous fashion, Dangas and Giustino note.

A number of ongoing studies may help clarify a role for different pharmacological strategies, including the GALILEO and ATLANTIS trials, both of which are testing non-vitamin K antagonist oral anticoagulants in the setting of TAVR, but many of these remain a ways off from delivering results. GALILEO, said Dangas, who is co-PI on the trial, has enrolled 250 patients so far out of an estimated 1,500 patients, in 15 countries. The phase III study is looking at whether a low-dose rivaroxaban-based strategy long term is better than an antiplatelet drug strategy in preventing death or thrombotic events in patients who have undergone successful TAVR. ATLANTIS is “taking a different approach,” he said, looking at full-dose apixaban versus different control groups based on upfront anticoagulation needs related to non-TAVR indications.

What to Do in the Meantime

In the meantime, what can physicians performing TAVR, or following up with patients who’ve undergone successful transcatheter valve replacement do to prevent stroke? Here, says Dangas, the papers offer a couple insights that he thinks people can consider while waiting for ongoing trials to wrap up.

Using some of the observations gleaned in Kapadia and Kleiman’s analyses, Dangas and Giustino constructed a flow-chart outlining the risk factors and preventive strategies for stroke and how these could be applied over the days, weeks, and months following TAVR. These include procedural and pharmacologic strategies that could be adopted during the procedure and in the days immediately following TAVR. They also highlight the role of the CHA₂DS₂-VASc score in predicting stroke not only in the first month, but at 3 months and 1 year.

To TCTMD, Dangas highlighted the fact that physicians should be screening for certain stroke risk factors beyond the 1-month mark. “The fact that the acute and subacute stroke risk is out to 3 months—I’m not sure everyone has understood that. They might see their patient at 1 month and not again until 1 year so I think they don't quite click that the 3-month time point there may be a need to reevaluate the situation. So that might be one of the messages to apply in their practice.”

Dangas added that the door remains open to using clinical judgement with regards to what kinds of antiplatelet and anticoagulation therapy could be prescribed long-term, advocating an approach that isn’t one-size-fits-all.

“In clinical practice, most medical centers prescribe one thing to everybody with TAVR and just hope for the best, and I think in this respect, we are trying to make them start changing that a little bit and use the knowledge and the risk scores in a little bit better way. Do I have firm evidence to show that this will improve patient outcomes? No, but I have a firm understanding that this makes pathophysiological sense and that it should make a difference with careful patient evaluation.”