Early Neurological Change After Ischemic Stroke Is Associated With 90-Day Outcome

Large-scale observational studies of acute ischemic stroke (AIS) promise to reveal mechanisms underlying cerebral ischemia. However, meaningful quantitative phenotypes attainable in large patient populations are needed. We characterize a dynamic metric of AIS instability, defined by change in National Institutes of Health Stroke Scale score (NIHSS) from baseline to 24 hours baseline to 24 hours (NIHSS - NIHSS = ΔNIHSS ), to examine its relevance to AIS mechanisms and long-term outcomes. Patients with NIHSS prospectively recorded within 6 hours after onset and then 24 hours later were enrolled in the GENISIS study (Genetics of Early Neurological Instability After Ischemic Stroke). Stepwise linear regression determined variables that independently influenced ΔNIHSS . In a subcohort of tPA (alteplase)-treated patients with large vessel occlusion, the influence of early sustained recanalization and hemorrhagic transformation on ΔNIHSS was examined. Finally, the association of ΔNIHSS with 90-day favorable outcomes (modified Rankin Scale score 0-2) was assessed. Independent analysis was performed using data from the 2 NINDS-tPA stroke trials (National Institute of Neurological Disorders and Stroke rt-PA). For 2555 patients with AIS, median baseline NIHSS was 9 (interquartile range, 4-16), and median ΔNIHSS was 2 (interquartile range, 0-5). In a multivariable model, baseline NIHSS, tPA-treatment, age, glucose, site, and systolic blood pressure independently predicted ΔNIHSS (R =0.15). In the large vessel occlusion subcohort, early sustained recanalization and hemorrhagic transformation increased the explained variance (R =0.27), but much of the variance remained unexplained. ΔNIHSS had a significant and independent association with 90-day favorable outcome. For the subjects in the 2 NINDS-tPA trials, ΔNIHSS was similarly associated with 90-day outcomes. The dynamic phenotype, ΔNIHSS , captures both explained and unexplained mechanisms involved in AIS and is significantly and independently associated with long-term outcomes. Thus, ΔNIHSS promises to be an easily obtainable and meaningful quantitative phenotype for large-scale genomic studies of AIS.