Rationale The increasing use of thrombectomy for acute ischemic stroke highlights the need to better understand its complications. Vessel perforation leading to active intracranial bleeding is a severe complication associated with a poor outcome and a mortality rate of approximately 50%. The etiology and risk factors for this severe complication remain largely unknown, and there is limited evidence to guide decision-making regarding endovascular hemostatic strategies and the continuation of thrombectomy after perforation.Design The Perforation EVents during ENdovascular Therapy for acute ischemic stroke (PREVENT) Registry is an international, multicenter, prospective, and retrospective study, collecting data on vessel perforations during thrombectomy. The registry aims to include 500 cases of vessel perforation and 500 matched controls without perforation. The collected data will consist of both tabulated data and periprocedural imaging, which is to be analyzed centrally by an imaging core laboratory. Data will be analyzed using a case-control method, employing both univariate and multivariate statistical analyses.Objectives The primary objectives of this study are to (1) identify risk factors for vessel perforation, (2) explore the underlying pathophysiology, (3) develop a classification system for vessel perforations, (4) compare different hemostatic treatment strategies, (5) evaluate the impact of continuing versus aborting thrombectomy after perforation, and (6) propose a safety-optimized thrombectomy technique.Discussion Given the frequency and hyperacute nature of vessel perforations, a prospective randomized study is not feasible. This large, international registry provides a robust approach to collecting real-world evidence, reducing bias through multicenter data collection and case matching. The findings may help improve clinical decision-making and enhance patient safety during thrombectomy procedures.Clinical Trial Registration ClinicalTrials.gov, identifier NCT06394180.
Schulze-Zachau et al. (Fri,) studied this question.