Abstract TP229: Measurement of Multiple Blood Markers to Predict Long- term Outcome in Ischemic Stroke

Abstract

Background and Purpose: There is no single blood marker for prediction of prognosis in ischemic stroke. We investigated whether multiple blood markers enhance the predictability of long-term outcome following ischemic stroke. Methods: Blood concentrations of neuronal markers (NSE, VSNL-1, H-FABP, and Ngb), astroglial markers (S100B and GFAP), inflammatory markers (IL-6, TNF-α, WBC count, and CRP), blood-brain barrier marker (MMP-9), and haemostatic markers (D-dimer and PAI-1) were compared with favorable and poor outcome groups regard to 3 month clinical outcome. Multivariate logistic regression analysis, area under receiver operator characteristic (AUROC) curve, and net reclassification improvement (RNI) index were performed to examine whether adding blood markers to baseline clinical model (age, sex, initial NIHSS score, and atrial fibrillation) improves discrimination for poor and good outcome. Results: In multivariate analysis, plasma IL-6 (OR, 1.013, per 10 pg/ml, 95% CI, 1.04 - 1.23, P = 0.003) was independently associated with poor outcome 3 months after onset, and hFABP (OR, 1.05, per 1 ng/ml, 95% CI, 1.00 - 1.09, P = 0.048), WBC (OR, 1.22, per 103/uL, 95% CI, 1.00 - 1.49, P = 0.046), and S100B (OR, 1.05, per 10 pg/ml, 95% CI, 1.01 - 1.09, P = 0.017) showed a marginal association with poor outcome. Addition of single blood marker to baseline clinical model did not improve discrimination for poor outcome from good outcome. Addition of four blood markers including IL-6, hFABP, WBC, and S100B to baseline clinical model showed improvement of discrimination for poor outcome from good outcome (AUROC: 0.95, NRI index, 0.14, p = 0.02). Conclusions: Combination of neuroglial and inflammatory blood markers including IL-6, hFABP, WBC count, and S100B has an additive predictive value for stroke outcome.