Cerebral Embolic Protection During Transcatheter Aortic Valve Implantation: A Meta-Analysis, Meta-Regression, and Trial Sequential Analysis

Stroke is a serious complication of transcatheter aortic valve implantation (TAVI). Cerebral embolic protection devices (CEPDs) were developed to reduce periprocedural embolization; however, their clinical efficacy remains uncertain. We conducted a systematic review and meta-analysis of 8 randomized controlled trials including 11,775 patients undergoing TAVI with or without CEPDs. Primary outcomes included stroke (any, disabling, non-disabling), new ischemic lesions, and cognitive decline assessed by the Montreal Cognitive Assessment and National Institutes of Health Stroke Scale. Secondary outcomes included mortality, bleeding, vascular complications, and acute kidney injury. Data were pooled using a random-effects model, reporting risk ratios (RRs) or mean differences with corresponding 95% confidence intervals (CIs). Subgroup, meta-regression, and trial sequential analyses were performed. CEPDs did not significantly reduce any stroke (RR: 0.92, 95% CI [0.74-1.15], P = 0.48), disabling stroke (RR: 0.80, 95% CI [0.57-1.12], P = 0.19), or non-disabling stroke (RR: 1.08, 95% CI [0.79-1.46], P = 0.64). No significant differences were observed in ischemic lesion count, National Institutes of Health Stroke Scale worsening, or Montreal Cognitive Assessment decline. The Sentinel device showed a marginal reduction in ischemic lesion count without clinical benefit. Sentinel use was associated with increased major vascular complications (RR: 1.42, 95% CI [1.01-2.01], P = 0.046), and sensitivity analysis demonstrated higher acute kidney injury with CEPDs (RR: 1.38, 95% CI [1.01-1.88], P = 0.04). Trial sequential analyses suggested futility for most outcomes. While CEPDs capture embolic debris, they do not significantly reduce stroke or cognitive decline after TAVI, limiting support for routine use. Future efforts should focus on improved device design and identifying high-risk subgroups.