When absence of evidence equates to evidence of absence: the case of routine use of cerebral embolic protection devices in transcatheter aortic valve implantation.

Abstract

Objectives: This updated hierarchical Bayesian meta-analysis aims to integrate the latest randomised controlled trials (RCTs) on the use of cerebral embolic protection (CEP) in transcatheter aortic valve implantation (TAVI) into previously available data, providing a definite answer to the clinical effectiveness of CEP in TAVI patients.

Methods: A systematic search was updated on 31 March 2025. RCTs were included when comparing transfemoral TAVI with use of CEP versus transfemoral TAVI without CEP. The primary outcome was all stroke, while the secondary outcome was disabling stroke. A hierarchical Bayesian meta-analysis was performed on the (log) relative risk (RR) scale and transformed to absolute risk differences (ARDs) and numbers needed to treat (NNTs). The threshold for clinical relevance was based on published expert consensus and established on 1.1% ARD (NNT 91).

Results: The study was updated with one new RCT, totalling a number of eight RCTs (n=11 590, CEP n=5921 patients, control n=5669 patients). The prevalence of all stroke and disabling stroke was 2.9% and 1.4% in the control group. The median RR for all stroke was 0.94 (95% credible interval (CrI) 0.72-1.25), translating to a mean of -0.17% ARD (NNT 588), and a posterior probability of a clinically relevant CEP effect of <1%. The median RR for disabling stroke was 0.76 (95% CrI 0.44-1.23), translating to a mean of -0.36% ARD (NNT 278), and a posterior probability of a clinically relevant CEP effect of <1%.

Conclusion: Current-generation CEP devices are ineffective in reducing periprocedural TAVI-stroke risk to a clinically relevant degree, rendering future trials with these devices futile.

Prospero registration number: CRD42023407006.