Routine Cerebral Embolic Protection during Transcatheter Aortic-Valve Implantation (BHF PROTECT-TAVI)

Authors, Journal, Affiliations, Type, DOI

• Authors: Rajesh K. Kharbanda et al. for the BHF PROTECT-TAVI Investigators (44 named investigators; 33 UK centres)
• Journal: New England Journal of Medicine 2025; 392:2403–12
• Affiliations: University of Oxford (lead); London School of Hygiene and Tropical Medicine Clinical Trials Unit (coordination); 32 NHS centres + 1 private UK TAVI centre
• Type: Prospective, open-label, multicenter, randomized controlled trial with blinded endpoint adjudication
• Funding: British Heart Foundation (BHF Clinical Study CS/20/1/34732); Boston Scientific investigator-sponsored research grant (CEP devices only — not involved in design, conduct, or reporting)
• DOI: 10.1056/NEJMoa2415120

Overview

BHF PROTECT-TAVI is the largest RCT (n=7,635) evaluating routine use of the Sentinel cerebral embolic protection (CEP) device during TAVI. Patients with aortic stenosis were randomized 1:1 to TAVI with or without CEP across 33 UK centres. The primary endpoint — clinical stroke within 72 hours — occurred in 2.1% with CEP vs 2.2% without (p=0.94). The trial was stopped early for futility after ruling out a ≥40% relative risk reduction. There were no differences in disabling stroke, severe stroke, or death; CEP was associated with a slight increase in serious adverse events (0.6% vs 0.3%).

Keywords

Transcatheter aortic-valve implantation, cerebral embolic protection, Sentinel device, stroke, randomized controlled trial, BHF PROTECT-TAVI, PROTECTED TAVR, aortic stenosis, futility

Key Takeaways

Background and Rationale

• TAVI is associated with procedure-related stroke from embolism, hemorrhage, or cerebral hypoperfusion
• CEP devices (specifically the Sentinel device, Boston Scientific) are designed to filter debris in the left common carotid artery and right innominate artery, preventing embolic stroke
• The Sentinel device is the only CEP device approved in the US and Europe
• Prior trial: PROTECTED TAVR (n=3,000; 51 international sites) stopped early by prespecified rules; stroke within 72h NS; disabling stroke numerically lower with CEP — warranted further evaluation

Study Design

• Design: Prospective, open-label, multicenter RCT; blinded independent clinical events committee adjudicated primary endpoint
• Sites: 33 UK centres (32 NHS + 1 private); 29.6% of all TAVI procedures at NHS participating sites enrolled
• Enrollment: October 2020 – October 2024; stopped for futility October 9, 2024
• Population: Aortic stenosis patients scheduled for TAVI, judged anatomically/clinically suitable for Sentinel device; age ≥18 years
• Randomization: 1:1, stratified by site with random permuted blocks
• Device: Sentinel CEP device delivered percutaneously from right radial artery; distal filter in left common carotid artery; proximal filter in right innominate artery
• Stroke assessment: Questionnaire for Verifying Stroke-Free Status (QVSFS) administered daily for 72 hours or until discharge — validated tool; positive response prompted formal stroke assessment

Participants

• n=7,635 randomized: 3,815 CEP vs 3,820 control
• Modified ITT population: 3,798 CEP vs 3,803 control (17 withdrawn per group)
• Mean age 81.2 ± 6.5 years; 38.7% women
• Trial cohort representative of UK TAVI population (BCIS/NICOR registry comparison)

Device Deployment

• Both filters fully and correctly deployed for duration of procedure: 3,058/3,768 (81.2%) in CEP group
• At least one filter deployed: 87.5% of CEP group
• No difference in device deployment success between early vs later cases at each site — no learning effect detected

Primary Outcome

• Stroke within 72 hours after TAVI (or before discharge if earlier): 2.1% CEP (81/3,795) vs 2.2% control (82/3,799)
• Difference: −0.02 percentage points; 95% CI −0.68 to 0.63; P=0.94
• Most strokes occurred within 24 hours of TAVI

Secondary Outcomes

• Severe stroke (NIHSS ≥10): 0.5% vs 0.5% (diff 0.0 pp; 95% CI −0.3 to 0.3)
• Disabling stroke at 6–8 weeks (mRS ≥2, increase ≥1 from baseline): 1.2% vs 1.4% (diff −0.2 pp; 95% CI −0.7 to 0.4)
• Death within 72h: 0.8% vs 0.7% (diff 0.1 pp; 95% CI −0.3 to 0.5) — NS
• Death within 8 weeks: 2.1% vs 1.9% — NS
• Access-site complications (VARC-2) before discharge: 8.1% vs 7.7% (diff 0.4 pp; NS)
• Access-site complications at 6–8 weeks (at aortogram access site): 0.8% vs 0.4% (diff 0.4 pp; 95% CI 0.1 to 0.8) — statistically higher with CEP; 25 minor vs 12 minor complications

Adverse Events

• Serious adverse events: 24 events in 22/3,798 participants (0.6%) CEP vs 13 events in 13/3,803 participants (0.3%) control — numerically doubled with CEP

Statistical Analysis and Stopping

• Original sample size target: 7,730 (80% power for 3% vs 2% stroke reduction)
• After second interim analysis (5,411 enrolled): incidence 2.0% → revised target 9,712 for 2.4% vs 1.6%
• Stopped for futility: at 134 stroke events, lower limit of 99% CI excluded 40% relative risk reduction — pre-specified futility criterion met
• CACE (complier average causal effect) analysis confirmed no treatment effect adjusting for non-adherence: difference −0.2 pp (95% CI −1.0 to 0.6) for stroke; −0.2 pp (95% CI −0.9 to 0.4) for disabling stroke
• Age- and sex-adjusted analyses concordant

Comparison with PROTECTED TAVR

• PROTECTED TAVR: n=3,000; 51 international sites; stopped early; stroke within 72h NS; disabling stroke numerically lower (CEP); overall stroke incidence 2.6% (higher than PROTECT-TAVI 2.2%)
• BHF PROTECT-TAVI used clinical stroke definition (symptoms >24h); PROTECTED TAVR used imaging-inclusive shorter duration — explains PROTECT-TAVI having 31 TIAs vs 3 in PROTECTED TAVR
• BHF PROTECT-TAVI required both filters for adherence; less restrictive eligibility criteria → enrolled more complex anatomies
• Both trials consistently null for clinical stroke prevention with routine CEP

Limitations of the Document

• Open-label design (blinded adjudication only for primary endpoint — unblinded care staff could influence secondary decisions)
• Predominantly White population — minority ethnic groups underrepresented; generalisability uncertain
• COVID-19 pandemic affected enrollment cadence and clinical context
• Less restrictive eligibility criteria may have diluted CEP efficacy in complex aortic arch anatomy patients (lower device deployment success)
• Clinical stroke definition (>24h symptoms) may miss subclinical embolic events detectable by DWI-MRI — CEP may still reduce subclinical brain lesions not captured by this trial
• Stopped before reaching revised target (9,712) — powered for 40% RRR; smaller benefit not excluded (but would be clinically marginal)
• Aortogram-access site complications slightly higher with CEP at 6–8 weeks (0.8% vs 0.4%; p<0.05) — primarily minor; clinical significance uncertain

Key Concepts Mentioned

• concepts/Cerebral-Embolic-Protection-TAVI — primary subject of this trial
• concepts/TAVI — the procedure in which CEP was tested
• concepts/Aortic-Stenosis — primary indication
• concepts/Stroke-Prevention — clinical goal of CEP

Key Entities Mentioned

• entities/Boston-Scientific — manufacturer of Sentinel CEP device (provided devices; no design/conduct role)
• entities/British-Heart-Foundation — primary funder
• entities/Sentinel-CEP-Device — the specific device tested

Wiki Pages Updated

• Created: wiki/sources/cep-tavi-bhfprotecttavi-nejm-2025.md
• Created: wiki/concepts/Cerebral-Embolic-Protection-TAVI.md
• Updated: wiki/concepts/TAVI.md — added CEP section and new contradiction
• Updated: wiki/sourceindex.md
• Updated: wiki/wikiindex.md