ABSTRACT NUMBER: ESOC2026A672 PROXIMAL FLOW ARREST USING BALLOON GUIDE CATHETER IN ENDOVASCULAR THROMBECTOMY (PROACT): INDIVIDUAL PARTICIPANT DATA META-ANALYSIS OF TWO RANDOMIZED TRIALS

Abstract Background and aims Proximal flow arrest using a balloon-guide catheter (BGC) is recommended during endovascular thrombectomy for acute ischemic stroke due to large-vessel occlusion, but recent randomized trials have yielded conflicting results. This study aims to evaluate the effectiveness of proximal flow arrest with BGC vs. non–flow-arrest in patients undergoing endovascular thrombectomy. Methods Systematic search of PubMed and Medline from inception to August 1, 2025 was conducted to include randomized clinical trials comparing endovascular thrombectomy with vs. without proximal flow arrest using a BGC.Both eligible trials provided complete Individual participant data, which were harmonized and analyzed using one-stage mixed-effects models. The primary outcome was functional disability at 90 days, assessed by the modified Rankin Scale (mRS) using a mixed-effects ordinal logistic regression model. Results Two trials (PROTECT-MT and ProFATE-MT) with 463 patients were included (232 assigned to BGC with flow arrest, 231 to non–flow-arrest). Variability between studies was small. Median 90-day mRS was 4 (IQR 2–5) with flow arrest and 3 (IQR 2–5) with non–flow arrest (adjusted common OR 0.83, 95% CI 0.60–1.14). Symptomatic intracranial haemorrhage and mortality did not differ significantly. Subgroup analysis suggested onset-to-randomization time modified the effect (p for interaction=0.032): BGC tended toward worse outcomes when randomized >6 h from symptom onset, whereas outcomes were comparable otherwise. Conclusions In this IPD meta-analysis of randomized trials, proximal flow arrest using a BGC provided no overall benefit compared with non-flow arrest. Future studies were needed to assess whether newer-generation devices or more selective indications might provide clinical benefit. Conflict of interest All authors: nothing to disclose