Posts Tagged Glasgow Outcome Score

[ARTICLE] Traumatic Brain Injury Severity, Neuropathophysiology, and Clinical Outcome: Insights from Multimodal Neuroimaging – Full Text

Background: The relationship between the acute clinical presentation of patients with traumatic brain injury (TBI), long-term changes in brain structure prompted by injury and chronic functional outcome is insufficiently understood. In this preliminary study, we investigate how acute Glasgow coma score (GCS) and epileptic seizure occurrence after TBIs are statistically related to functional outcome (as quantified using the Glasgow Outcome Score) and to the extent of cortical thinning observed 6 months after the traumatic event.

Methods: Using multivariate linear regression, the extent to which the acute GCS and epileptic seizure occurrence (predictor variables) correlate with structural brain changes (relative cortical atrophy) was examined in a group of 33 TBI patients. The statistical significance of the correlation between relative cortical atrophy and the Glasgow Outcome Score was also investigated.

Results: A statistically significant correlative relationship between cortical thinning and the predictor variables (acute GCS and seizure occurrence) was identified in the study sample. Regions where the statistical model was found to have highest statistical reliability in predicting both gray matter atrophy and neurological outcome include the frontopolar, middle frontal, postcentral, paracentral, middle temporal, angular, and lingual gyri. In addition, relative atrophy and GOS were also found to be significantly correlated over large portions of the cortex.

Conclusion: This study contributes to our understanding of the relationship between clinical descriptors of acute TBI, the extent of injury-related chronic brain changes and neurological outcome. This is partly because the brain areas where cortical thinning was found to be correlated with GCS and with seizure occurrence are implicated in executive control, sensory function, motor acuity, memory, and language, all of which may be affected by TBI. Thus, our quantification suggests the existence of a statistical relationship between acute clinical presentation, on the one hand, and structural/functional brain features which are particularly susceptible to post-injury degradation, on the other hand.


Long-term clinical outcome after traumatic brain injury (TBI) is predicated upon a large variety of often poorly understood factors which substantially complicate the task of identifying the relationship between acute clinical variables and chronic functional deficits. Nevertheless, understanding how post-TBI cortical atrophy patterns reflect acute-stage patient presentation may help to identify cortical areas that are likely to undergo substantial atrophy, and implicitly to isolate aspects of cognitive, affective and neural function which are at highest risk for long-term degradation.

Attempts to relate TBI-related changes in brain structure to clinical variables often involve structural brain variables provided by neuroimaging methodologies, such as magnetic resonance imaging (MRI) and diffusion tensor imaging (DTI) (13). In previous studies, quantitative metrics provided by acute neuroimaging of TBI patients have been used to describe the relationship between acute injury profiles and chronic dysfunction (47). By contrast, hardly any non-neuroimaging clinical variables have been identified which can be used to elucidate the pattern of structural brain changes after TBI. Nevertheless, the ability to incorporate such non-neuroimaging clinical descriptors into outcome forecasting models is important because many such descriptors—including the Glasgow Coma Score (GCS)—are recorded routinely by clinicians and relied upon during the treatment decision-making process.

In this study, we illustrate how two important TBI severity indicators that are routinely assessed by clinicians in the acute care setting and without the use of neuroimaging can be used to relate patient presentation in the acute stage of TBI to the pattern and extent of post-TBI cortical atrophy as well as to neurological outcome. These two indicators—the GCS and the occurrence of epileptic seizures during the acute stage of TBI—can likely assist in predicting cortical atrophy patterns and in evaluating the risk for poor neurological outcome. This study additionally identifies cortical regions whose susceptibility to post-traumatic atrophy is correlated significantly and reliably—in a statistical sense—with functional outcome and with clinical descriptors of TBI severity. […]

Continue —>  Frontiers | Traumatic Brain Injury Severity, Neuropathophysiology, and Clinical Outcome: Insights from Multimodal Neuroimaging | Neurology

Figure 1. (A) Quantification of the linear model’s ability to predict cortical atrophy extent at 6 months after injury. For each gyrus and sulcus, the null hypothesis that there is no statistically significant correlation between the predictor variables and the response variable (cortical thinning, in millimeters) was tested. Values of the F2,30 statistic for each statistical test are encoded on the cortical surface, subject to the false discovery rate correction for multiple comparisons. Darker red hues indicate higher significance of the statistical test and, consequently, stronger ability to predict cortical thinning for the areas in question. Regions where the null hypothesis was not tested because less than 90% of cortical thickness data were available (see text) are drawn in black. Regions where the test statistic was lower than the threshold F statistic of the reliability analysis permutation test are drawn in white. (B) Statistical significance of the correlation between relative cortical atrophy and the GOS-E. Values of the t31 statistic for each statistical test are encoded on the cortical surface, as in panel (A). Note that all values of this statistic are negative, which confirms that greater regional atrophy is associated with lower GOS-E values (i.e., poorer functional outcome), as expected. The values of F and t statistics in (A) and (B), respectively, are associated with different statistical tests and different degrees of freedom and, therefore, they should not be compared to one another.

, , , , , , , , ,

Leave a comment

%d bloggers like this: