Glasgow Coma Scale Motor Score and Pupillary Reaction To Predict Six-Month Mortality in Patients with Traumatic Brain Injury: Comparison of Field and Admission Assessment

M Majdan, Ewout Steyerberg, Daan Nieboer, W Mauritz, M Rusnak, Hester Lingsma

Research output: Contribution to journalArticleAcademicpeer-review

46 Citations (Scopus)

Abstract

The Glasgow Coma Scale (GCS) and pupillary reactivity are well-known prognostic factors in traumatic brain injury (TBI). The aim of this study was to compare the GCS motor score and pupillary reactivity assessed in the field and at hospital admission and assess their prognostic value for 6-month mortality in patients with moderate or severe TBI. We studied 445 patients with moderate or severe TBI from Austria enrolled to hospital in 2009-2012. The area under the curve (AUC) and Nagelkerke's R-2 were used to evaluate the predictive ability of GCS motor score and pupillary reactivity assessed in the field and at admission. Uni- and multi-variable analyses-adjusting for age, other clinical, and computed tomography findings-were performed using combinations of field and admission GCS motor score and pupillary reactivity. Motor scores generally deteriorated from the field to admission, whereas pupillary reactivity was similar. GCS motor score assessed in field (AUC=0.754; R-2=0.273) and pupillary assessment at admission (AUC=0.662; R-2=0.214) performed best as predictors of 6-month mortality in the univariate analysis. This combination also showed best performance in the adjusted analyses (AUC=0.876; R-2=0.508), but the performance of both predictors assessed at admission was not much worse (AUC=0.857; R-2=0.460). Field GCS motor score and pupillary reactivity at hospital admission, compared to other combinations of these parameters, possess the best prognostic value to predict 6-month mortality in patients with moderate-to-severe TBI. Given that differences in prognostic performance are only small, both the field and admission values of GCS motor score and pupillary reaction may be reasonable to use in multi-variable prediction models to predict 6-month outcome.
Original languageUndefined/Unknown
Pages (from-to)101-108
Number of pages8
JournalJournal of Neurotrauma
Volume32
Issue number2
DOIs
Publication statusPublished - 2015

Cite this