A Food and Drug Administration advisory panel struggled to muster support for marketing clearance of the TriGuard 3 (Keystone Heart) device for use during transcatheter aortic valve replacement (TAVR).
The Circulatory Systems Devices Panel of the Medical Devices Advisory Committee took no vote when it met Aug. 3, but weighed evidence for a proposed indication for the device “to minimize the risk of cerebral damage by deflecting embolic debris away from the cerebral circulation” during TAVR.
“While this device may deflect some debris, the data would suggest it may also create issues,” said Keith B. Allen, MD, director of surgical research at the Mid America Heart & Lung Surgeons, Kansas City, Mo. “I am really concerned that our desire and the emotion that surrounds preventing stroke are not being supported by the data.”
TriGuard 3 received CE Mark in Europe in March 2020. It was submitted for 510(k) clearance and seeks to prove substantial equivalence to the predicate Sentinel device (Claret Medical), currently the only approved embolic protection device in the United States.
The device is designed to cover all three major aortic vessels (innominate, left carotid, and left subclavian arteries) and is delivered transfemorally through an 8F sheath, whereas the Sentinel is positioned within the branch vessels, doesn’t cover the left subclavian artery, and is introduced through the radial or brachial artery via a 6F sheath.
TriGuard 3 faced an uphill battle, however, after failing to meet the primary composite efficacy endpoint in the REFLECT phase 2 trial (P = .857), with numeric trends showing higher all-cause mortality or any stroke at 30 days (9.8% vs. 6.7%) than pooled control subjects without embolic protection.
Rates for other components of the endpoint also trended higher with the device: National Institutes of Stroke Stroke Scale score worsening 2-5 days after the procedure, cerebral ischemic lesions on MRI 2-5 days after the procedure, and total cerebral ischemic lesion volume.
The Sentinel device was approved in 2017 after it failed to meet its primary efficacy endpoint of new brain lesion volume on MRI, but death and stroke rates favored the device over control, the panel pointed out.
The sponsor provided additional analyses in the per treatment (PT) population, defined as those with complete three-vessel coverage in at least two of three procedural time points. Compared with pooled control subjects, most of the imaging endpoints favored the TriGuard 3 device, but clinical neurologic event rates continued to favor the control group.
“The data used to demonstrate efficacy are all based on the PT subpopulation of the whole population, and those have to be considered promissory data,” said John Hirshfeld, MD, emeritus professor, University of Pennsylvania, Philadelphia. “This is the group where everything went well and for us to decide that’s achievable in the general population is speculative.”
Safety data
The REFLECT trial did meet its primary safety endpoint, with a 30-day major adverse cardiovascular event rate of 15.9%, compared with a performance goal of 34.4% (P < .0001).
Although prespecified, panel members pushed back, saying that the performance goal was unacceptably high, with several members remarking they’d never heard of a trial adding 9% as a “fudge factor” to a 25% historic control rate to get to the 34% performance target.
Keystone health officials noted that REFLECT was not designed to demonstrate a significant difference in the rate of primary safety events, compared with control. Instead, its purpose was to demonstrate that TriGuard 3 did not increase the risk associated with a TAVR procedure.
The TriGuard 3 device was successfully placed and retrieved in 100% of patients, but complete coverage was not uniform, with 72% of 157 as-treated patients having complete three-vessel coverage post TAVR but 15% having no coverage.
Panel members also expressed concern over device interference during TAVR, which was reported in nearly 10% of all TriGuard patients.
The TriGuard 3 group had 11 major vascular complications, 2 directly related to the device, and 3 stage 3 acute kidney injuries, whereas neither complication occurred in the control group.
Throughout the 9-hour hearing, the panel wrestled with what was described as a highly select patient group and small patient numbers that made it difficult to interpret observed differences. The trial involved 157 TriGuard 3 patients (including 41 from the roll-in phase) and 119 control subjects pooled from phase 2 of the trial (n = 57) and from phase 1 using the early-stage TriGuard HDH device (n = 57).
Pieter Stella, MD, PhD, Utrecht (the Netherlands) Medical Center, also presented “real-world” evidence from 75 patients in the Netherlands using the latest iteration of the device available in Europe with updates to the crimper and additional training materials to prevent the device from torquing during delivery. No strokes were reported, one patient had a transient ischemic attack (TIA), and two patients had a dissection, which resolved without sequelae.
Ralph Brindis, MD, MPH, professor of medicine, University of California, San Francisco, countered that there were only three experienced operators from a single center and that the stroke incidence was physician reported, “not data we can really embrace.”
There was much debate over why enrollment in phase 2 of the RHYTHM trial was temporarily paused in February 2019, briefly restarted, and then prematurely stopped in April 2019.
FDA officials said the study was paused at the recommendation of the data monitoring committee (DMC) because rates of safety events were different between patients and control subjects and operational errors called into question the accuracy of the data being reviewed. Ultimately, both the DMC and FDA recommended study suspension.
During the public hearing, TAVR pioneer Alain Cribier, MD, University of Rouen’s Charles Nicolle Hospital, Mont-Saint-Aignan, France, said the TriGuard 3 is of interest because it can be used with minimal need for manipulation and complete coverage of the cerebral vessels that is achieved by diverting rather than capturing debris. “The rapid and exponential growth of TAVR procedures demands safe TAVR interventions and the use of cerebral protection devices is a step in this direction.”
Dr. Rita F. Redberg
Others took a dim view. “Given that the Sentinel device has not demonstrated benefit on clinical outcomes, there is significant concern about similar devices, such as the TriGuard 3, providing clinical benefit,” Rita Redberg, MD, Sanket Dhruva, MD, and Robin Ji, University of California, San Francisco, wrote in a letter submitted to the panel.
Commenting further, they added: “With the results from the REFLECT II trial demonstrating no evidence for clinical outcome benefit in TAVR patients, and numerically higher rates for stroke risk, mortality, bleeding risk, and other dangerous adverse complications among those treated, it is concerning and dangerous for patient safety that the TriGUARD 3 cerebral embolic protection device is being considered for FDA 510(k) clearance.”
The FDA panel members reported no financial relationships.
A version of this article first appeared on Medscape.com.