Large Databases Hint at Sentinel Benefit During TAVR, but Uncertainty Remains

Information from two large US databases suggest that use of the Sentinel cerebral protection system (Boston Scientific) during TAVR reduces the risk of stroke, although experts continued to call for randomized trials to definitively settle the issue.

In the Society of Thoracic Surgeons/American College of Cardiology (STS/ACC) TVT Registry, the rate of in-hospital stroke was not significantly lower when the Sentinel device was used according to an instrumental-variable analysis (1.39% vs 1.54%; RR 0.90; 95% CI 0.68-1.13), David Cohen, MD (Kansas City, MO), reported during a late-breaking clinical science session at the virtual TCT Connect 2020.

A secondary propensity-weighted analysis of the TVT data, however, indicated that cerebral protection was associated with a reduction in in-hospital stroke (1.30% vs 1.58%; RR 0.82; 95% CI 0.69-0.97).

Consistent with the propensity-weighted analysis, a study involving data from the National Inpatient Sample presented by Michael Megaly, MD (Banner University Medical Center, Phoenix, AZ), during the TCT “Best Abstracts” session yesterday showed that after propensity-matching, Sentinel use was associated with a lower risk of in-hospital ischemic stroke (1.0% vs 3.8%; OR 0.24; 95% CI 0.09-0.62).

Speaking to the TVT Registry data, Cohen said at a press briefing that “both the secondary propensity-adjusted analysis and the confidence interval for the primary analysis are consistent with a possible modest reduction in stroke, with a relative risk reduction of about 20% and a number needed to treat, at least in this study, of about 300 for preventing what I consider to be a major stroke.”

He and others indicated that there’s a need for more-definitive randomized data to determine the impact of cerebral embolic protection devices like Sentinel on patient outcomes. Highlighting that need was the REFLECT II trial presented earlier at the meeting; in that study, another protection device, TriGUARD 3 (Keystone Heart), failed to demonstrate a benefit in terms 30-day stroke or death.

A trial designed to assess any potential clinical effects of cerebral embolic protection—PROTECTED TAVR, which will enroll about 3,000 patients and is estimated to be completed in July 2022—is currently randomizing patients to TAVR with or without the Sentinel device.

The TVT Registry analyses “support clinical equipoise and they provide a strong rationale for the ongoing large-scale randomized trials to test whether embolic protection devices truly provide meaningful clinical benefit—that is, reduced stroke or improved neurocognitive function for patients undergoing TAVR,” Cohen said.

STS/ACC TVT Registry

Despite improvements in TAVR technology and technique, increased operator experience, and refined patient selection, stroke remains a problem, occurring in 2% to 5% of cases. Cerebral protection devices like Sentinel, cleared by the US Food and Drug Administration in June 2017, were developed to lessen that risk. There are, however, no randomized studies definitively showing that the devices—which capture debris bound for the brain in nearly all patients—improve clinical outcomes.

Cohen et al explored the potential clinical impact using data on transfemoral TAVR procedures performed in 2018 and 2019 included in the STS/ACC TVT Registry. They excluded emergent procedures as well as those involving alternative access, done at sites performing fewer than 20 TAVRs each year, and involving concomitant mitral interventions.

Over the 2-year study period, the proportion of hospitals offering embolic protection increased from 7% to 28%, and the proportion of patients receiving protection increased from 5% to 13%. There was substantial variability across sites, however, with two-thirds of hospitals using no embolic protection and about 5% using protection devices in more than half of procedures.

The primary endpoint was site-reported in-hospital stroke assessed in two different ways—instrumental-variable analysis, which uses natural experiments to approximate randomization and control for measured and unmeasured confounders, plus propensity-weighting based on 30 demographic, clinical, and hospital characteristics.

In the instrumental-variable analysis, Sentinel use was not associated with risk of any of the outcomes examined, including stroke, death, and major bleeding in the hospital and stroke and death at 30 days.

In the secondary propensity-weighting analysis, however, use of cerebral embolic protection was associated with lower rates of in-hospital stroke and death or stroke (2.1% vs 2.5%; RR 0.84; 95% CI 0.73-0.98), as well as 30-day stroke (1.9% vs 2.2%;  RR 0.85; 95% CI 0.73-0.99) and death (1.7% vs 2.2%; RR 0.78; 95% CI 0.64-0.95).

Though the findings appear to conflict, “in some ways, these two analyses are more similar than different,” Cohen said, noting that the confidence intervals are wider in the instrumental-variable analysis. “These results are actually not that inconsistent with one another.”

National Inpatient Sample

Megaly et al explored the impact of the Sentinel device using data from the National Inpatient Sample. Their analysis included 36,220 patients who underwent TAVR in the last three quarters of 2017; Sentinel was used in 525 of them (1.4%). Cerebral protection was more frequently used at large teaching hospitals. Patients in whom Sentinel was used were less likely to have PAD and a history of coronary revascularization, and none had carotid disease, which is a contraindication to use.

Sentinel patients were propensity-matched to 1,050 patients who underwent unprotected TAVR. Those who were protected had a lower risk of in-hospital ischemic stroke, as well as a lower risk of in-hospital death (0 vs 1%; P = 0.036). The mortality finding should be interpreted with caution because of the study’s retrospective nature, Megaly said.

There were no differences between groups for vascular complications, blood transfusions, renal complications, length of stay, and the likelihood of discharge to a nursing facility, although index hospitalization costs were higher in the Sentinel group ($47,783 vs $44,578; P = 0.002).

On multivariate analysis, several factors independently associated with a higher risk of post-TAVR ischemic stroke emerged: history of carotid disease, PAD, atrial fibrillation or flutter, older age, bicuspid aortic valve, and female sex. All of these factors, Megaly said, would suggest a need to use cerebral embolic protection.

Waiting on a Randomized Trial

In a panel discussion during the press briefing, Susheel Kodali, MD (NewYork-Presbyterian/Columbia University Irving Medical Center, New York, NY), who co-led the SENTINEL trial, questioned Cohen about how his analyses could account for all confounders. For example, he said, “the sites that use Sentinel may be more concerned about stroke and maybe they’re more focused on the neurologic analysis and do neurologic checks.”

Cohen said the data suggested that might be an issue, so they adjusted for the rate of stroke at the sites in 2017, before anybody started using embolic protection. He noted that it’s possible the propensity-weighted analysis was less affected by that potential confounding.

What Cohen said he would take away from his study is that equipoise exists to support randomization in ongoing trials. In addition, “what is sobering to me is that even if there is benefit here, the benefit may be less than we think,” he said, adding that the 20% to 25% relative risk reduction suggested by the analysis is “a little bit sobering for a device that’s supposed to protect three out of the four great vessels.”

Commenting for TCTMD, Dharam Kumbhani, MD (UT Southwestern Medical Center, Dallas, TX), said Cohen’s study is “helpful to kind of see the overall landscape for the field. And I agree with his conclusion that it’s probably appropriate to have an adequately powered trial that assesses clinical endpoints with this device.”

Asked whether the totality of the evidence on cerebral protection devices justifies use in 13% of patients undergoing TAVR, Kumbhani responded, “The use is a little bit higher than what I think the evidence suggests. The evidence does not support widespread use of this device.”

He said that at his center, Sentinel is used primarily in patients with prior stroke and in those with heavy calcification in the ascending aorta and aortic arch.

Because of the inherent limitations of observational analyses, the TVT Registry data are not likely to push operators one way or another in terms of how they’re currently using the Sentinel device, Kumbhani said. “It is FDA-approved and I think operators will probably want to use it in the right patient population.”

At the press briefing, Chad Rammohan, MD (Sutter Health, Mountain View, CA), said about the impact of the Sentinel observed in the TVT Registry, “The fact that the absolute risk reduction is so low and the number needed to treat is so high, it allows us to realize that they’re pretty much equal based on this data and we have to wait for the big trial. And if you’re not using it, it’s probably okay to not use it, and if you’re using it, it’s probably okay to use it until the big trial comes out.”