Embolic Protection in TAVR: If, When, What, and How Debated as Trials Wrap Up

PARIS, France—Recent randomized trials report overall stroke rates after TAVR that range from as low as 1.5% to as much as 6%, but according to experts gathered here at EuroPCR 2016, the true rate of cerebral infarction is likely much higher. 

Moreover, silent cerebral injury “is dramatically more frequent than clinically apparent cerebrovascular events, and has controversial prognostic significance,” Matteo Pagnesi, MD (San Raffaele Scientific Institute, Milan, Italy), said here.

As such, operators and investigators alike are asking questions about stroke: how to track and quantify the risk, how to capture or deflect emboli and debris, and whether this effectively eliminates any risk. 

A number of clinical trials testing use of embolic protection devices specifically developed for transcatheter valve procedures are nearing conclusion. The devices furthest along are the Embrella (Edwards Lifesciences), TriGuard (Keystone Heart), and Sentinel (Claret Medical). A fourth device originally designed for open heart procedures, the EMBOL-X system (Edwards), has also been tested in TAVR.

Pooled Analysis with the TriGuard Device

In the interim, EuroPCR served up a range of studies looking at data from early preapproval registries and from centers in Europe where the devices already hold CE Mark.

Alexandra Lansky, MD (Yale University, New Haven, CT) presented a patient-level pooled analysis of neuroprotection using the TriGuard versus unprotected TAVR, combining data from DEFLECT I, a registry of 23 patients, DEFLECT III, a randomized trial comparing TriGuard in 36 patients with a control group of 44, and NeuroTAVR, a registry of 44 patients.

The TriGuard is a first-generation filter device designed to deflect cerebral emboli while allowing maximal blood flow, positioned across all three cerebral vessels. 

According to imaging tests at 30 days, Lansky said, incidence of central nervous system infarction was significantly higher among patients who underwent TAVR without the TriGuard (92% vs 72%, P = 0.008). So too were the average diffusion-weighted MRI volume and total volume. 

In terms of clinical results, rates of MACCE and National Institutes of Health Stroke Scale (NIHSS) score were significantly lower in the TriGuard group, although cognitive testing was no different between groups. Valve Academic Research Consortium (VARC)-defined stroke was lower in the filter group, with a marginal P value of 0.05. Asked whether the analysis was powered for any of these harder endpoints, Lansky conceded that this was a post-hoc, pooled analysis. “We certainly didn’t power this up front, but it’s essentially [an attempt] to find and establish differences, and we’re actually very pleased that with just 59 patients [in the neuroprotection arm] that we were able to detect a difference.”

The powered studies are being done, she added, with CLARET likely to come out at TCT 2016. “And the REFLECT trial is ongoing,” she said. “Those are the randomized trials that are actually going to answer that question.”

Stroke, she pointed out, is one of the strongest predictors of mortality following TAVR: a three- to ninefold increased risk in mortality has been reported, Lansky said. “What we’ve seen in recent series is that if you send a neurologist in to evaluate the patient, or you do serial [NIHSS] testing and do MRI, the actual stroke rate is higher” than reported in trials, somewhere in the range of 15% to 28%. “So there seems to be an underreporting of this complication.”

Discussing the findings with the press, Lansky underscored a number of problems with the device, and the pooled analysis. The device itself is an early-generation contraption and does not actually seal the aorta, so some emboli still pass through. But these preliminary data suggest that the size of the lesions is reduced, hinting that it is smaller emboli that are making it past the filter, while the larger ones are deflected appropriately.

Nicolo Piazza, MD (McGill University, Montreal, Canada), moderating the press conference, pointed to the baseline characteristics of patients in the pooled analysis, noting that things that are also risk factors for stroke—previous stroke/transient ischemic attack, A-fib on admission, STS score---were all numerically higher in the control group, although none individually was statistically different between groups. 

In an underpowered analysis, this “begs the question” whether the higher stroke rate in the control patients relates to these baseline risk factors and not to the lack of a neuroprotection device, he said.

Registry Results for Sentinel

Elsewhere at the meeting, C.J. Jensen, MD (Contilia Heart and Vascular Center, Essen, Germany), reviewed the prevalence and etiology of thromboembolic debris during TAVR in the SENTINEL-H registry using the Claret Sentinel device. To date, he said, this device has been used in more than 3,000 patients worldwide; there were 217 patients in the registry at the time of the analysis.

Strikingly, debris was captured in 99% of cases. In an analysis of debris type, Jensen showed that acute thrombus is far and away the most common, found in 85% of cases. Of note, this was dominated by acute thrombus associated with “foreign material.”

That’s important, session moderator Raj Makkar, MD (Cedars-Sinai Heart Institute, Los Angeles, CA), noted. Makkar has led much of the research studying clot formation and leaflet mobility with new TAVR valves and surgical valves. The nature of the debris analyzed in Jensen et al’s study suggests, “This is not something that forms on the device, and that’s very important.” 

Jensen agreed, noting, “The relatively high prevalence of periprocedural captured foreign bodies is surprising and might indicate the need for improvement of the material involved during TAVI.”

A Silent Threat?

Pagnesi presented results from a meta-analysis combining six prospective studies using the Claret, EMBOL-X, TriGuard, and Embrella devices. In this analysis, the number of new lesions per patient was not statistically greater among TAVR patients treated without embolic protection than among those in whom protection devices were used. However, both the total lesion volume and the single lesion volume was reduced in the pooled neuroprotection patients. 

Importantly, the rate of silent cerebral injury after TAVR in Pagnesi’s study was 77.5%, according to pooled diffusion-weighted MRI—the measurement being used in the ongoing randomized controlled trials.

Session comoderator, Ravinay Bhindi, MBBS, PhD (Royal North Shore and North Shore Private Hospitals, Sydney, Australia), agreed to TCTMD that the magnitude of the stroke problem after TAVR, how to mitigate it, and how to predict which patients are at higher stroke risk, remain key unresolved questions in the field. 

When strokes occur, he said, “we try to blame the device, the anatomy, or the patient characteristics, but I don’t think we’re any closer to knowing what actually predicts stroke. We’re moving to an era where there’s a tension between reducing the risk of stroke versus the risk of not [deploying the device] in the perfect position.”

Even using repositionable devices, he added, it’s not clear that repositioning TAVR devices leads to more stroke. Indeed, one of the presentations in a session Bhindi moderated suggested that stroke was more common among patients treated with balloon-expandable devices, while another found more strokes for a repositionable valve. “The key thing I got from this is that the patients that we’re treating that get stroke are people who would probably get stroke anyway, because they’re high risk.”

Makkar said, “What I take away is that in the vast majority of patients, if you use a filter device, you capture something, and this is something that’s actually coming from the valve, because one study showed that there is tissue attached to the thrombus. The other thing that’s important is that even knowing that this material goes into the brain, as we saw, a lot of the strokes occurred beyond 24 hours, so there is late consequence of this material.”

Whether cerebral protection devices will reduce stroke rates is information that will hopefully come from the ongoing trials. The next question will be how to select the patients in whom they should be used. Will they be warranted in all cases or just in patients at the highest risk for stroke? 

“We don’t know yet,” Makkar said. “Looking at all of this, I’m tempted to use [a protection device] rather than not use one,” he added, noting that while the products are not yet available to US operators, he himself has used certain devices in studies. “If I was practicing in Europe where these are CE marked and available to me, I would very much be tempted to use them for everybody.”  


  • Lansky A. A patient-level pooled analysis of neuroprotection with the TriGuard embolic deflection device compared to unprotected transcatheter aortic valve replacement. Presented at: EuroPCR 2016. May 18, 2016. Paris, France.
  • Jensen CJ. Prevalence and etiopathology of thromboembolic debris during transcatheter interventional aortic valve replacement: results of the SENTINEL-H study. Presented at: EuroPCR 2016. May 19, 2016. Paris, France.
  • Pagnesi M. Silent cerebral injury after TAVI and the preventive role of embolic protection devices: a systematic review and meta-analysis. Presented at: EuroPCR 2016. May 19, 2016. Paris, France.  


  • Lansky reports receiving institutional grant/research support from Keystone Heart.
  • Piazza reports serving as a consultant to and being a stockholder of HighLife and serving as a consultant to Medtronic.
  • Pagnesi and Jensen report no relevant conflicts of interest.

Related Stories:


Shelley Wood is the Editor-in-Chief of TCTMD and the Editorial Director at CRF. She did her undergraduate degree at McGill…

Read Full Bio