Related Experiment Video
Updated: Jan 3, 2026
Optimized Management of Endovascular Treatment for Acute Ischemic Stroke
Published on: January 18, 2018
Frequency of early rapid improvement in stroke severity during interfacility transfer
Kori S Zachrison1, Thabele M Leslie-Mazwi1, Gregoire Boulouis1
1Department of Emergency Medicine (KSZ, JNG), Massachusetts General Hospital, Boston, MA; Department of Neurology (TML-M, RWR, AV, AC, NR, LHS), Massachusetts General Hospital, Boston, MA; Department of Neuroradiology (GB), Université Paris-Descartes, INSERM U894, Centre Hospitalier Sainte-Anne, Paris, France; Department of Neuroradiology (AL), Goethe University, Frankfurt, Germany; and Department of Neurology (KAS), Baylor College of Medicine, Houston, TX.
Background:
As interfacility transfer of patients with stroke becomes increasingly common, understanding fluctuations in deficits during transfer may help predict resource needs. We sought to characterize changes in NIH Stroke Scale (NIHSS) scores during transfer and identify factors associated with early rapid improvement (ERI).
Methods:
We used prospectively collected data from our Comprehensive Stroke Center's (CSCs) stroke and telestroke network databases. We calculated changes in NIHSS scores for all patients transferred to our CSC after an initial telestroke evaluation from January 2010 to December 2015. Logistic regression identified factors associated with ERI, controlling for patient characteristics available on arrival.
Results:
Among the 505 patients included, the median initial NIHSS score was 11 (interquartile range [IQR] 5-18), and on CSC arrival, it was 9 (IQR 3-17), with a median change of 0 (-3 to -0). Of note, 74.5% of scores changed by fewer than 4 points (7% increased ≥4 points, and 19% decreased ≥4). In 85% of cases, the NIHSS score change did not cross a threshold to alter eligibility for thrombectomy. In multivariable modeling, ERI was associated with ability to ambulate before the index stroke (odds ratio [OR] 5.79, p = 0.02) and higher initial NIHSS (OR 1.06 per point, p = 0.001).
Conclusions:
These findings may be valuable for resource planning and for inclusion in thrombectomy alert activation processes at the receiving hospital.
