By Jeffrey Zimmet, MD, PhD

Associate Professor of Medicine, University of California, San Francisco; Director, Cardiac Catheterization Laboratory, San Francisco VA Medical Center

SYNOPSIS: In this trial of cerebral protection in transcatheter aortic valve replacement, the TriGUARD device was safe vs. historical controls, but failed to meet its primary efficacy endpoint.

SOURCE: Lansky AJ, Makkar R, Nazif T, et al. A randomized evaluation of the TriGuard™ HDH cerebral embolic protection device to Reduce the Impact of Cerebral Embolic LEsions after TransCatheter Aortic Valve ImplanTation: The REFLECT I trial. Eur Heart J 2021;42:2670-2679.

Despite significant advances over the past decade, stroke remains the most common ischemic complication of transcatheter aortic valve replacement (TAVR) and certainly is one of the most feared.1 Imaging studies have demonstrated an overwhelming majority of TAVR patients develop new ischemic lesions in the brain after the procedure.2 Clinically evident stoke occurs more selectively and unpredictably in many cases, affecting somewhere in the range of 2% to 6% of patients.1 These strokes are thought to result from embolization of particulates from the calcified valve itself and from manipulation of atheromatous disease in the ascending aorta and aortic arch. The concept of cerebral embolic protection (CEP) devices that can capture such debris before it reaches the brain has considerable intuitive appeal. To date, a single device, the Sentinel, has received FDA approval.3 This device is deployed from the right radial artery and places filters in the brachiocephalic trunk and left carotid artery. A competing device that has thus far only been available in Europe, TriGUARD, is deployed from the femoral approach and could protect each arch vessel. The REFLECT 1 trial was designed to prospectively evaluate the efficacy of this device in TAVR patients.

To this end, the authors recruited patients to be randomized 2:1 to the TriGUARD device or to standard care without CEP. The trial included a composite safety endpoint (all-cause death and stroke), as well as outcomes including bleeding, acute kidney injury, vascular complications, coronary obstruction, and major valve dysfunction. The primary endpoint was a hierarchical composite of all-cause mortality or any stroke at 30 days, NIH Stroke Scale score worsening from baseline to two to five days after procedure or MoCA cognitive function test worsening at 30 days, and total volume of cerebral ischemic lesions detected by MRI performed two to five days after procedure.

Between June 2016 and July 2017, 258 patients were enrolled from 20 centers in the United States and six in Europe. A total of 54 of these patients were in a predesigned roll-in phase, leaving 141 patients randomized to the CEP device and 63 to control. The balloon-expandable SAPIEN valve was used in approximately 60% of study patients, with the self-expanding CoreValve accounting for most of the remainder. Among patients assigned to the device arm, the device was deployed successfully across all three great vessels for 93.4%, while it remained fully in place throughout the TAVR surgery in just 57.3% of procedures. The operators reported the device interfered significantly with the TAVR system in 8.8% of cases.

The primary safety endpoint occurred in 21.8% of patients in the device group by 30 days. Although this met the safety endpoint by beating the prespecified performance goal of 34.4%, it was significantly higher than the 8.5% event rate observed in the control group. This was driven primarily by a higher rate of major vascular complications in the TriGUARD group (7.1% vs. 0%). The other components of the safety endpoint were not significantly different. The primary efficacy endpoint was not significantly different between groups. This was true for the entire set, as well as for the subset of patients who achieved full cerebral coverage with the device for the duration of the procedure. The rates of stroke at 30 days were numerically higher in the device arm (10%) vs. controls (6.8%), although this difference was not statistically significant. Among patients who underwent brain imaging with diffusion-weighted MRI, the number of ischemic lesions and the total volume of lesions were not different between the groups. The authors reported the TriGUARD device was safe when compared with a performance goal based on historical data, but failed to meet its efficacy endpoint in the primary hierarchical composite of death, stroke, NIH Stroke Scale score worsening, or total ischemic lesion volume.


In a world where positive publication bias is a reality, a negative trial like this one seems rare. Seemingly hidden in the text is the fact the authors paused the study early at the recommendation of the Data Safety and Monitoring Board, although the reasons for this are not included in the publication. The sponsor elected not to resume the trial, instead focusing its attention on its next-generation device, TriGUARD 3. What can we learn from this relatively underpowered trial, with an older-generation device, that ended prematurely?

Even the assertion the device met its safety endpoint is somewhat suspect. Using a prespecified historical event rate of 35%, seemingly high by current standards, as a comparator should rate some notice. The fact the safety composite was significantly lower in the actual control group should raise some eyebrows, especially with a device that requires a relatively large second (in addition to the TAVR sheath) femoral access site.

Stroke rates in REFLECT were higher than what usually is reported in clinical registries. This seems to reflect a general finding that stroke rates are higher when all patients are subjected to formal neurologic assessment. The disconnect between the appearance of ischemic lesions on MRI, seen in 85% of participants, and clinical stroke is well-known and is demonstrated here. Whether the post hoc observation indicating the device reduced larger lesions and carries clinical relevance remains to be seen.

Usually, medical devices continue to improve with further iterations. The TriGUARD device studied here was successful in protecting the great vessels throughout the TAVR procedure in just over half of patients and interacted negatively with the TAVR system itself in nearly 10%. The device is relatively large and requires femoral access. We can expect the devices and outcomes to improve, especially when we can learn from negative trials such as this one.

Much larger trials using the Sentinel device are in the works, including PROTECTED TAVR (3,000 patients) and British Heart Foundation PROTECT-TAVI (7,730 patients).4,5 However, despite a compelling backstory, the routine use of CEP devices in TAVR continues to lack robust data showing unequivocal benefit in reducing the clinically important strokes that matter most. We will continue to watch this space with keen interest as future trials are presented.


  1. Davlouros PA, Mplani VC, Koniari I, et al. Transcatheter aortic valve replacement and stroke: A comprehensive review. J Geriatr Cardiol 2018;15:95-104.
  2. Arnold M, Schulz-Heise S, Achenbach S, et al. Embolic cerebral insults after transapical aortic valve implantation detected by magnetic resonance imaging. JACC Cardiovasc Interv 2010;3:1126-1132.
  3. U.S. Food & Drug Administration. Sentinel cerebral protection system.
  4. Stroke PROTECTion With SEntinel During Transcatheter Aortic Valve Replacement (PROTECTED TAVR).
  5. London School of Hygiene & Tropical Medicine. BHF PROTECT-TAVI.