Key Predictors of 90-Day Functional Outcome for Patients with Large Infarct Core Undergoing Endovascular Therapy in Real-world Practice
Harshita Cherukuri1, Demi Waworuntu1, Sergio Marioni1, Monica Mallavarapu1, Ananya Iyyangar1, Sunil Sheth1, Jerome Jeevarajan1
1The University of Texas Health Houston, McGovern Medical School, Houston, TX
Objective:
To evaluate the clinical and imaging variables that have the largest impact on 90-day modified Rankin Scale (mRS) outcome for patients with large infarct core (LIC) undergoing endovascular therapy (EVT).
Background:
Recent randomized trials have shown that patients with LIC can benefit from EVT, however the key factors to predict long-term functional outcome remain unclear.
Design/Methods:
From a prospectively collected multicenter registry, consecutive patients undergoing EVT were identified with pre-treatment ASPECTS 0 to 5. The cohort was divided a priori into 80:20 training and testing sets for the final analysis. Multivariable logistic regression was used to evaluate the association between variables of interest and the primary outcome, defined as 90-day mRS 0-3. The final model performance was measured by area under the receiver operating characteristic curve (AUROC). Secondarily, we evaluated predictors of symptomatic intracranial hemorrhage (sICH) by ECASS-II definition.
Results:
Among 114 patients with LIC, median age was 65, 50% were female, median NIHSS was 19 [IQR 16-22], 35.1% received IV thrombolysis, and 8.8% experienced sICH. In univariable analysis, 3 or more EVT passes was associated with poor 90-day outcome (OR 2.80, p = 0.041); however, this association became nonsignificant after adjusting for covariables including age, NIHSS, and TICI score. In multivariable analysis, only age and NIHSS remained significantly associated with poor 90-day mRS. Age ≥75 led to about 5 times greater likelihood of poor outcome, and each 5-point increase of NIHSS increased the likelihood by approximately 10%. No significant variables predicting sICH were identified. Final model performance was reasonable (AUROC 0.771) for predicting 90-day mRS.
Conclusions:
In this real-world cohort of patients with LIC undergoing EVT, ASPECTS and number of EVT passes were not associated, whereas age and NIHSS were the only significant predictors of poor 90-day outcome. More nuanced tools will be necessary to triage patients with LIC.
Disclaimer: Abstracts were not reviewed by Neurology® and do not reflect the views of Neurology® editors or staff.