The prognostic reliability of the Glasgow coma score in traumatic brain injuries: evaluation of MRI data

To clarify the predictive power of the Glasgow coma score (GCS) after traumatic brain injury (TBI) and in the context of brain stem lesions.

In 143 patients who had suffered severe TBI, the GCS was correlated to brain damage as visualized by cranial magnetic resonance imaging (MRI). This technique evaluates the damage to the brain stem in particular. The Brussels coma score (BCS) was also used.

The GCS was not significantly correlated to brain stem lesions when it was only scored at the time of admission. When MRI was not used later on, the GCS showed a poor ability to predict the outcome. After 24 h, and on the day of MRI screening, the GCS was significantly correlated with two parameters: outcome (the higher the GCS, the better the outcome) and the frequency of patients without injuries to the brainstem in MRI (the higher the GCS, the higher this frequency). These correlations were much more evident when the BCS was used. The prognostic power of the GCS was found to vary over time; for example: a GCS of 3 at admission was associated with a favorable prognosis; a GCS of 4 signified a poor prognosis, irrespective of the time point at which the GCS was scored; and the prognostic power of a GCS of 5 deteriorated from the day of the MRI onwards, whereas the prognostic power of a GCS of 6 or 7 varied little over time.

We only recommend the use of the GCS for prognostic evaluation in a multidimensional model. Study protocols should contain additional brain stem function parameters (BCS, pupil condition, MRI).