P.124 Assessing the fragility index of randomized controlled trials on carotid artery stenosis: systematic review

Abstract

Background: The fragility index (FI) is the minimum number of patients whose status would have to change from a nonevent to an event to turn a statistically significant result to a non-significant result. We used this to measure the robustness of trials comparing carotid endarterectomy (CEA) to carotid artery stenting (CAS). Methods: A search was conducted in MEDLINE, Embase, and PubMed on RCTs comparing CEA to CAS. The trials need to have statistically significant results and dichotomous primary endpoints to be included. Results: Our literature search identified 10 RCTs which included 9382 patients (4734 CEA, 4648 CAS). The primary end points of all included trials favoured CEA over CAS. The median FI was 9.5 (interquartile range 2.25 - 21.25). All of the studies that reported lost-to-follow-up (LTFU) had LTFU greater than its fragility index, which raises concern that the missing data could change the results of the trial from statistically significant to statistically insignificant. Conclusions: A small number of events (FI, median 9.5) were required to render the results of carotid artery stenosis RCTs comparing CEA to CAS statistically insignificant. All of the studies that reported LTFU had LTFU greater than its fragility index.