Abstract 2769: A Simple Clinical Index Based On Nihss Score During 1-h Tpa Infusion To Predict Unfavorable Outcome

Abstract

Background: When intravenous tissue-type plasminogen activator therapy (IV-tPA) does not work, additional endovascular thrombectomy is a promising choice. However, there is no clear standard of timing to add thrombectomy after IV-tPA. We sought to create a handy clinical index to judge insufficient effects of IV-tPA immediately after the therapy based on posttreatment NIH Stroke Scale (NIHSS) score. Methods: Consecutive ischemic stroke patients with the baseline NIHSS score ≥8 and occlusion at the ICA or MCA on the initial MRA, who received IV-tPA within 3 h of onset, were retrospectively studied. NIHSS was assessed 30 min and 1 h after the initiation of IV-tPA and change from the baseline score was calculated (ΔNIHSS). The subjects were divided into 2 groups according to the arterial occlusion sites; ICA or proximal M1 with the residual length <5 mm (Group P) and distal M1 or M2 (Group D). Optimal cutoff scores of NIHSS and ΔNIHSS for predicting unfavorable outcome at 3 months, corresponding to the modified Rankin Scale of 3 to 6, were determined using receiver operating characteristic curves. The predictive accuracy of unfavorable outcome by NIHSS score and its combination with ΔNIHSS score was assessed. Results: Forty two patients (16 women, 75±9 years old, 27 ICA, 15 M1) were enrolled as Group P and 78 patients (22 women, 77±12 years old, 50 M1, 28 M2) as Group D. In Group P, 35 patients (83%) had unfavorable outcome; cutoff NIHSS scores predicting unfavorable outcome were ≥14 at both 30 min (AUC 0.855) and 1 h (0.916), and cutoff ΔNIHSS scores were ≤1 at both 30 min (0.829) and 1 h (0.888). In Group D, 42 patients (54%) had unfavorable outcome; cutoff NIHSS scores were ≥11 at 30 min (AUC 0.684) and ≥12 at 1 h (0.723), and cutoff ΔNIHSS scores were ≤6 at 30 min (0.615) and ≤7 at 1 h (0.631). In each time-point of both groups, cutoff NIHSS scores themselves were good predictors of unfavorable outcome, and the combination with cutoff ΔNIHSS scores strengthen the predictive accuracy more. In Group P, ‘NIHSS 30min ≥14 plus ΔNIHSS 30min ≤1’ showed the sensitivity (SEN) of 67%, specificity (SPEC) of 100%, positive predictive value (PPV) of 100%, and negative predictive value (NPV) of 39% in predicting the outcome, and ‘NIHSS 1h ≥14 plus ΔNIHSS 1h ≤1’ showed SEN 66%, SPEC 100%, PPV 100%, and NPV 37%. In Group D, ‘NIHSS 30min ≥11 plus ΔNIHSS 30min ≤6’ showed SEN 71%, SPEC 67%, PPV 71%, and NPV 67%, and ‘NIHSS 1h ≥12 plus ΔNIHSS 1h ≤7’ showed SEN 62%, SPEC 78%, PPV 77%, and NPV 64%. Conclusion: Combination of the NIHSS score and its change from the baseline score during 1-h tPA infusion seems to be a good predictor of unfavorable outcome at 3 months and helpful to decide the timing of adding endovascular thrombectomy. In particular, the combination showed high specificity and PPV of the outcomes in patients with occlusion at the proximal carotid axis.