BackgroundDespite the proven benefit of neurothrombectomy, intracranial hemorrhage (ICH) remains the most serious procedural complication. The aim of this analysis was to identify predictors of different hemorrhage subtypes and evaluate their individual impact on clinical outcome.MethodsPooled individual patient-level data from three large prospective multicenter studies were analyzed for the incidence of different ICH subtypes, including any ICH, hemorrhagic transformation (HT), parenchymal hematoma (PH), subarachnoid hemorrhage (SAH), and symptomatic intracranial hemorrhage (sICH). All patients (n=389) treated with the Solitaire device were included in the analysis. A multivariate stepwise logistic regression model was used to identify predictors of each hemorrhage subtype.ResultsGeneral anesthesia and higher baseline Alberta Stroke Program Early CT score (ASPECTS) were associated with a lower probability of any ICH (OR 0.36, p=0.003), (OR 0.80, p=0.032) and HT (OR 0.54, p=0.023), (OR 0.78, p=0.001), respectively. Longer time from onset to treatment was associated with a higher likelihood of HT (OR 1.08, p=0.001) and PH (OR 1.11, p=0.015). Intravenous tissue plasminogen activator (IV-tPA) was also a strong predictor of PH (OR 7.63, p=0.013). Functional independence at 90 days (modified Rankin Scale (mRS) 0–2) was observed significantly less frequently in all hemorrhage subtypes except SAH. None of the patients who achieved functional independence at 90 days had sICH.ConclusionsGeneral anesthesia and smaller baseline ischemic core are associated with a lower probability of HT whereas IV-tPA and prolonged time to treatment increase the risk of PH after neurothrombectomy.Trial registration numbersSWIFT-NCT01054560; post results, SWIFT PRIME-NCT01657461; post results, STAR-NCT01327989; post results.