Abstract P819: Individualized Patients’ Risk of Post Rt-pa ICH in Acute Ischemic Stroke by Using a Simple Predictive Model

Abstract

Background and Purpose: A simple and reliable prediction of post intravenous recombination tissue plasminogen activator (rt-PA) intracerebral hemorrhage (ICH) is useful for stroke team to inform individual risk and improve hospitalization care. We aim to develop and validate a simple predictive score model to assess the 24-hour risk of ICH post rt-PA. Methods: This retrospective study included 739 acute ischemic stroke patients who received intravenous rt-PA between October 2005 and June 2020 at Siriraj Hospital, Bangkok, Thailand. Multiple logistic regression was used to evaluate the value of independent variables associated with any ICH and symptomatic intracerebral hemorrhage (SICH) measure by the European Cooperative Acute Stroke Study II (ECASS II) definition. Adjusted ORs of independent variables were converted to point score, summated of point score provide a probability of ICH and SICH. The predictive scores were internal validated in 95 patients and performance tested with an area under a receiver operating characteristic curve (AUC-ROC). Results: From a total of 739 patients, 19.2% had an ICH in which 5.9% was SICH. The predictive model of ICH included hyperdense middle cerebral artery sign (HDMCA) (3 points), initial blood glucose > 180mg/dl (3 points), coronary artery disease (CAD) (2 points), international normalized ratio (INR) > 1.0 (2 points) and 10 year increase of age ( 1 point per 10 year increase; age ≤ 20 year = 1). The performance of the predictive model showed AUC-ROC was 0.72 (95% confidence interval, 0.60 - 0.82) in the derivative cohort and AUC-ROC was 0.74 (95% confidence interval, 0.66 - 0.85) in the validation cohort. Conclusions: This simple predictive model provides a better prediction of a 24-hour risk of ICH post rt-PA in AIS among Thai patients. An individualized ICH risk post rt-PA can assist frontline physicians and relatives in rt-PA decision making process. Further external validation and prospectively confirmed in a larger cohort is recommend.