BACKGROUND: In addition to the risk of stroke, patients undergoing transcatheter aortic valve implantation (TAVI) are susceptible to a decline of neurocognitive function. This may occur because of embolization of material (eg, valve or calcium) to the brain. Cerebral embolic protection (CEP) devices are engineered to capture this debris, potentially mitigating its incidence. METHODS: This is a secondary analysis of the BHF PROTECT-TAVI trial (British Heart Foundation Randomized Trial of Routine Cerebral Embolic Protection in Transcatheter Aortic Valve Implantation), in which participants with aortic stenosis from across 33 centers in the United Kingdom were randomly assigned at a 1:1 ratio to undergo TAVI with a CEP device (SENTINEL, Boston Scientific; SENTINEL CEP group) or TAVI without a CEP device (control group). This analysis is restricted to those who underwent cognitive assessment. The primary outcome was the mean change in the telephone version of the Montreal Cognitive Assessment (t-MoCA) between baseline and 6 to 8 weeks after TAVI. The secondary outcome was a ≥3-point drop in total t-MoCA score between baseline and 6 to 8 weeks after TAVI. RESULTS: A total of 3535 participants, 1763 in the SENTINEL CEP group and 1772 in the control group (mean age 81.0 years, 37.7% women) randomized in BHF PROTECT-TAVI were included in the modified intention-to-treat population for this analysis. The median t-MoCA at presentation was 18 (interquartile range, 16-20). The median t-MoCA at 6 to 8 weeks was 20 (interquartile range, 17-21). The mean change in total t-MoCA score between baseline and 6 to 8 weeks adjusted for the baseline score was 0.83 (95% CI, 0.70-0.96) in the SENTINEL CEP group and 0.91 (95% CI, 0.79-1.04) in the control group. There was no difference in means between the treatment groups (-0.07 [95% CI, -0.22 to 0.09], P=0.42). The incidence of a ≥3-point drop in the total t-MoCA score was 154 of 1763 (8.7%) in the SENTINEL CEP group and 142 of 1772 (8.0%) in the control group. The corresponding risk difference was 0.72% (95% CI, -1.10 to 2.55; P=0.44). These findings were robust to sensitivity analyses. There was no evidence of an interaction between treatment assignment and any of the subgroups assessed. CONCLUSIONS: In the BHF PROTECT-TAVI trial, the use of CEP did not impact cognition after TAVI. REGISTRATION: URL: https://www.isrctn.com; Unique identifier: ISRCTN16665769.