Abstract Number: Esoc2026a939 Early Adjuvant Tirofiban After Endovascular Thrombectomy: A Real-World, Propensity-Matched Analysis From the Taiwan Treat-AIs Registry

Abstract Background and aims The efficacy and safety of tirofiban administered within 24 hours after endovascular thrombectomy (EVT) remain debated. We evaluated the impact of adjuvant tirofiban on functional outcomes and hemorrhage risks using the Taiwan TREAT-AIS Nationwide Registry. Methods We conducted a retrospective multicenter study (Jan 2019–Feb 2025) of acute ischemic stroke patients receiving EVT. Tirofiban users were matched 1:4 with non-users using propensity scores based on age, gender, initial NIHSS, baseline blood pressure, and large-artery atherosclerosis. Primary outcomes were 90-day functional independence (modified Rankin Scale, mRS 0-2) and ordinal mRS shift. Safety outcomes included symptomatic intracranial hemorrhage (sICH), any radiological ICH, and mortality. Analyses employed multivariable logistic and partial proportional-odds models adjusting for confounders. Results Of 3,021 eligible patients, 245 received tirofiban. After matching, 245 users were compared with 980 non-users. Baseline characteristics were well-balanced. Tirofiban use was associated with significantly higher 90-day functional independence (adjusted OR aOR 2.06, 95% CI 1.48–2.87, p0.001) and a favorable shift in mRS distribution (aOR 0.55, 95% CI 0.42–0.71, p0.001). Moreover, 90-day mortality was significantly lower in the tirofiban group (aOR 0.49, 95% CI 0.30–0.79, p=0.004). No significant differences were observed for sICH (aOR 0.97, 95% CI 0.49–1.90, p=0.919) or any radiological ICH (aOR 0.77, 95% CI 0.55–1.08, p=0.129). Conclusions In this nationwide registry, tirofiban use within 24 hours post-EVT was associated with improved functional outcomes and reduced mortality without increasing intracranial hemorrhage risks. These findings support the therapeutic benefit and safety of early adjuvant tirofiban in selected patients. Conflict of interest All authors, Kuo-Cheng Chang, Nien-Chen Liao, Po-Lin Chen, nothing to disclose. Table 1 - belongs to Methods Table 2 - belongs to Results