Citation

Reisner A, Chen X, Kumar K, Reifman J. J. Neurotrauma 2014; 31(10): 906-913.

Affiliation

Massachusetts General Hospital, Department of Emergency Medicine, Boston, Massachusetts, United States ; andrewtreisner@yahoo.com.

Copyright

(Copyright © 2014, Mary Ann Liebert Publishers)

DOI

10.1089/neu.2013.3128

PMID

24372334

Abstract

We hypothesized that vital signs could be used to improve the association between a trauma patient's prehospital Glasgow coma scale (GCS) score and his or her clinical condition. Previously, abnormally low and high blood pressures have both been associated with higher mortality for patients with traumatic brain injury (TBI). We undertook a retrospective analysis of 1384 adult prehospital trauma patients. Vital-sign data were electronically archived and analyzed. We examined the relative risk of severe head abbreviated injury scale (AIS) 5-6 as a function of the GCS, systolic blood pressure (SBP), heart rate (HR), and respiratory rate (RR). We created multivariate logistic regression models and, using DeLong's test, compared their area under receiver-operating characteristic curves (ROC AUCs) for three outcomes: head AIS 5-6, all-cause mortality, and either head AIS 5-6 or neurosurgical procedure. We found significant bimodal relationships between head AIS 5-6 versus SBP and HR, but not RR. When the GCS < 15, ROC AUCs were significantly higher for a multivariate regression model (GCS, SBP, and HR) versus GCS alone. In particular, patients with abnormalities in all parameters (GCS, SBP, and HR) were significantly more likely to have high-mortality TBI versus those with abnormalities in GCS alone. This could be useful for mobilizing resources, e.g., neurosurgeons and operating rooms at the receiving hospital, and might enable new prehospital management protocols where therapies are selected based on TBI mortality risk.

Language: en