Prediction of Favorable Outcome by Percent Improvement in Patients with Acute Ischemic Stroke Treated with Endovascular Stent Thrombectomy

Yuezhou Cao,Shuiping Wang,Wen Sun,Qiliang Dai,Wei Li,Jin Cai,Xinying Fan,Wusheng Zhu,Yunyun Xiong,Yunfei Han,Wenjie Zi,Shiquan Yang,Jia Chen,Xinfeng Liu
DOI: https://doi.org/10.1016/j.jocn.2016.12.045
IF: 2.116
2017-01-01
Journal of Clinical Neuroscience
Abstract:Our objective was to investigate a method for assessing early improvement and its predictive value for 3 month functional outcome in patients treated with EST. A total of 97 consecutive AIS patients undergoing EST were prospectively collected and retrospective reviewed. Data on demographics, vascular risk factors, admission National Institutes of Health Stroke Scale (NIHSS) score, 24-h NIHSS score, reperfusion and collateral formation were collected. Percent improvement was defined as ([baseline NIHSS score 24-h NIHSS score]/baseline NIHSS score x 100%), while absolute improvement was calculated by the difference between scores (baseline NIHSS score 24-h NIHSS score). A 3-month functional outcome was assessed using the modified Rankin Scale (mRS). Favorable outcome was defined as a mRS score of 0-2. Areas under the receiver-operating characteristic (ROC) curve (AUC) for percent improvement and absolute improvement in predicting favorable outcome was compared. Finally, we investigated the independent predictors of improvement at 24 h after EST and its relationship with favorable outcome. Pairwise comparison of ROC curves revealed that percent improvement had larger AUC than absolute improvement (p = 0.004). Rapid neurological improvement (RNI), defined as percent improvement >= 30%, was a powerful predictor of favorable outcome (odds ratio [OR] 7.63, confidence interval [CI]: 2.65-21.96; p < 0.001). Good collaterals (OR 2.86; 95% CI: 1.11-7.38; p = 0.030) and short onset-to-reperfusion time (ORT) (OR 3.02, 95% CI: 1.17-7.80; p= 0.022) were independent predictors of RNI. RNI predicted 3-month favorable outcome in AIS patients treated with EST. Good collaterals and short ORT are independent predictors of RNI. (C) 2017 Elsevier Ltd. All rights reserved.
What problem does this paper attempt to address?