Late Cerebrovascular Events ‘Devastating’ in TAVR-Treated Patients

(UPDATED) Roughly one in 20 patients who undergo transcatheter aortic valve replacement for severe aortic stenosis experience a late stroke or transient ischemic attack, and these cerebrovascular events tend to have devastating consequences, according to the results of a new analysis.

The majority of late strokes that occurred were disabling, and nearly one-third of patients who had them died in the hospital.

While it’s uncertain if anything can be done to mitigate the risk, Guillem Muntané-Carol, MD (Quebec Heart and Lung Institute/Laval University, Quebec City, Canada), and colleagues say the findings should at least provide some assurances. An analysis of echocardiographic data available in 58% of patients at the time of the cerebrovascular event suggested the late events were not related to valve thrombosis or structural valve degeneration, but rather other potential reasons, such as the patient’s higher-risk profile.   

“There has been a lot of discussion and controversy about valve thrombosis, that TAVR may be potentially thrombogenic,” said senior investigator Josep Rodés-Cabau, MD (Quebec Heart and Lung Institute/Laval University). “I think this study provides some reassuring data regarding this issue. In general, the vast majority of strokes occurred either at 1 year or later, so not in the initial days and weeks following the procedure, which is the time period where valve thrombosis seems to peak. We had some incidental echo data—not in all patients and not core-lab adjudicated—and in the vast majority we didn’t see any signs of valve thrombosis or degeneration.”

The study, published online ahead of print March 11, 2020, in JACC: Cardiovascular Interventions, did not include CT evaluations of patients, which Rodés-Cabau admits is a limitation. “But when you look at gradients in the echo reports, we at least didn’t see any clinically relevant signs” suggesting valve thrombosis or degeneration, he said.

TAVR Population Is Unique

To TCTMD, Rodés-Cabau said there are limited data on the occurrence and factors associated with late cerebrovascular events in the TAVR population, which is what the researchers hoped to address with their analysis. The observational study was comprised of patients treated at seven centers in Canada, France, and Spain (mean age 80 years; 50.5% women). There was roughly an equal split between balloon-expandable and self-expanding transcatheter valves (46% and 54%, respectively), and in-hospital stroke occurred in 2% of patients. Slightly more than 34% of patients were discharged home on anticoagulation therapy, and for those not on anticoagulation, dual antiplatelet therapy was prescribed in 75.2% of cases.

Of 3,750 consecutive patients who underwent TAVR and survived beyond 30 days, 192 patients (5.1%) had a late cerebrovascular event, which was defined using the VARC-2 criteria, at a median of 16 months after TAVR. There were 38 TIAs, 124 ischemic strokes, and 29 hemorrhagic strokes (and one undetermined stroke). The annual event rates ranged from 1.5% to 2.1%. Strokes of ischemic, hemorrhagic, and unknown origin occurred in 80.5%, 18.8%, and 0.7% of patients, respectively. Overall, late stroke was disabling in 69% of patients and associated with a high rate of in-hospital mortality (29%).  

Rodés-Cabau said the rate of late cerebrovascular events in these TAVR-treated patients is slightly higher that what has been reported in a general population of similarly-aged patients. However, such comparisons are challenging because the TAVR population in general is unique, as was their patient cohort. “It wasn’t a low-risk group,” he said. “They were patients either at high or prohibitive risk [for surgery], and at the very least moderate risk. They all had comorbidities, frailty, among other conditions.”

Neal Kleiman, MD (Houston Methodist DeBakey Heart & Vascular Center, TX), who was not involved in the study, also said the rate of late cerebrovascular events is considerably higher than would be expected in similar patients not treated with TAVR, but he agreed the finding is explicable given the patient profile. With respect to the echocardiographic finding showing no evidence of valve thrombosis or structural deterioration, it is a reminder that “these valves don’t disintegrate early,” but without available CT data it is difficult to make too much of the observation, he said. 

To TCTMD, Vivian Ng, MD (Columbia University Irving Medical Center, New York, NY), said the present analysis provides some preliminary data suggesting that clinically-evident valve degeneration by transthoracic echocardiography (TTE) is not related to late cerebrovascular events. However, she pointed out that only about half of the patients with a stroke had a TTE at the time of their event. That, and the fact that valve thrombosis leading to valve dysfunction is relatively uncommon, makes it “difficult to make a strong argument that there is no relationship between stroke and valve thrombosis,” she wrote in an email.    

In an editorial, Paul Sorajja, MD, and Santiago Garcia, MD (Minneapolis Heart Institute Foundation/Abbott Northwestern Hospital, MN), point out that the clinical consequences seen here were “devastating,” noting that the occurrence of a late cerebrovascular event was an independent predictor of long-term mortality. With respect to etiology, though, more comprehensive imaging is necessary to definitively address the possible association between valve degeneration and late cerebrovascular events. The imaging substudies with cardiac CT that were part of the low-risk TAVR trials will hopefully provide more information, they write.  

Predictors of Late Cerebrovascular Events

In multivariate analysis, independent predictors of late ischemic events (ischemic stroke or TIA) were older age (HR 1.04 for each year increase; 95% CI 1.02-1.06), history of cerebrovascular disease (HR 1.87; 95% CI 1.57-2.21), higher baseline mean aortic gradient (HR 1.05 for each 10-mm Hg increase; 95% CI 1.01-1.09), periprocedural stroke at the time of TAVR (HR 3.21; 95% CI 1.46-7.07), and lack of anticoagulation at the time of discharge (HR 1.41; 95% CI 1.20-1.64).

The strong association between periprocedural stroke and later ischemic events is an intriguing finding in that it may identify a potentially more at-risk patient group, say researchers.

“Having a stroke at any time in your life puts you automatically at a higher risk for having a second stroke,” said Rodés-Cabau. “It’s not surprising. Older age, not surprising. However, having a stroke during the procedure also increases your risk for having a late stroke. That’s interesting. It means that patients having a stroke during TAVR, or in the periprocedural period, maybe they’re susceptible to stroke? It might be a risk marker. If you have had the bad luck to have a stroke, it might mean you’re more vulnerable. Embolization occurs in all patients, but not all patients have a stroke in the end.”

With respect to the lack of anticoagulation being linked with more ischemic strokes, the echo data don’t support the idea that these late events were the result of valve thrombosis, he added. TAVR patients, said Rodés-Cabau, have a high burden of atrial fibrillation, not all of which is clinically documented, and this may explain why lack of anticoagulation is associated with a higher risk of ischemic stroke.

“The problem is that most late strokes were ischemic, but close to 20% were hemorrhagic,” he said. “This is the difficult balance. Maybe, and I stress maybe, if we systematically anticoagulate all patients we will reduce ischemic strokes. That’s what the data seems to say. However, there is no doubt that we will increase hemorrhagic strokes because this is a vulnerable population. These aren’t 50-year-old patients. . . . In this population, when you do some very aggressive measures, usually there is a price to pay.”

The consequences of late hemorrhagic stroke are particularly damning, too. In their series, 79% of these hemorrhagic strokes were disabling and 66% of patients died. Recently, the GALILEO investigators showed that in patients without an indication for oral anticoagulation after TAVR, a rivaroxaban (Xarelto; Bayer/Janssen)-based antithrombotic strategy was associated with higher risk of death or thromboembolic events and bleeding compared to an antiplatelet-based strategy.

To TCTMD, Kleiman said their protocol is to discharge TAVR-treated patients home on dual antiplatelet therapy, with the plan typically being to stop one of the agents after 3 months. Oral anticoagulation is not prescribed unless the patient has an indication for treatment, and in that situation they are not prescribed anything else. For patients who had a periprocedural stroke, Kleiman said they have a conversation with the patient to tell them there is a “pretty substantial chance it will happen again” outside the hospital, but the concern doesn’t rise to level of prescribing oral anticoagulation.

“If you knew that anticoagulation would prevent the stroke, then you would, but I think the mechanism is probably that there just isn’t much reserve,” he said. “They’ve had a stroke and lost some brain tissue. It means that if they lose another blood vessel they’ll probably have a noticeable event. At least that’s my interpretation. I certainly would not swear to that.”

At their center, Ng said, they also discharge patients home on aspirin and clopidogrel and do not routinely start oral anticoagulation in TAVR patients given the lack of evidence to support it. “Furthermore, the patients that we have been treating, up until the recent approval of low-risk patients, are not only at high risk of ischemic stroke, but also frail and tend to be at higher risk of bleeding,” she said. 

In their editorial, Sorajja and Garcia state that the risk of periprocedural stroke might be mitigated by greater use of cerebral embolic protection. They also note than nonpharmacological alternatives to oral anticoagulation might be an option in some TAVR patients with atrial fibrillation at increased risk for bleeding. The WATCH-TAVR trial is currently evaluating left atrial appendage closure for the prevention of stroke and bleeding in patients with atrial fibrillation undergoing TAVR.