Predictors of Hemorrhagic Transformation Following Intra-Arterial Thrombolysis
Background: Hemorrhagic transformation is a major complication of intra-arterial (IA) thrombolytic therapy. Delineation of significant predictors of intracerebral hemorrhage (ICH) following thrombolysis would be useful to guide patient selection for IA treatment. Methods: Data were collected on consecutive patients with acute cerebral ischemia treated with IA thrombolysis over an 8 year period at UCLA Medical Center. Patients received either pure IA thrombolytics (tPA or urokinase) or combined intravenous/intra-arterial (IV/IA) tPA. Using univariate and multivariate analyses, 58 variables were examined as potential predictors of ICH. Results: A total of 89 patients were treated, 77 anterior circulation, 12 posterior circulation. Median baseline NIHSS score was 16. 26 patients received IA tPA only, 22 IV/IA tPA, and 41 IA urokinase only. Asymptomatic ICH occurred in 29 patients (33%), minor symptomatic ICH in 10 (11%), and major symptomatic ICH in 6 (7%). Of the 45 patients with M1 or M2 MCA occlusions treated with pure IA thrombolysis, 4 (9%) had major symptomatic ICH. There was no difference in the rate of any ICH with pure IA thrombolysis (39%) vs. combined IV/IA thrombolysis (41%). In pure IA cases, there was no difference in ICH rate between tPA and urokinase. 86% of patients with any ICH were dead or disabled (MRS > 2) at day 7 vs. 39% without ICH (p<0.0001). On univariate analysis, significant predictors of ICH were NIHSS score, time to recanalization, serum glucose level, platelet count, hematocrit, active treatment with warfarin, and degree of recanalization. In both multivariate stepwise linear discriminant and CART analyses, independent predictors of ICH were NIHSS score, time to recanalization, and platelet count. A model using these variables correctly predicted hemorrhagic transformation with PPV 80% and overall accuracy 75%. Conclusions: In this large series of patients treated with intra-arterial thrombolysis, higher NIHSS score, longer time to recanalization, and lower platelet count were independent predictors of hemorrhagic transformation. A model incorporating these variables identified patients who would develop ICH with 75% accuracy.