Quantitative lobar cerebral blood flow for outcome prediction after traumatic brain injury

The aim of this study was to examine cortical cerebral blood flow (CBF) in patients with traumatic brain injury (TBI) and determine whether lobar cortical CBF is a better predictor of long-term neurological outcome assessed by the Glasgow Outcome Scale (GOS) than global cortical CBF. Ninety-eight patients with TBI had a stable xenon computed tomography scan (Xe/CT-CBF study) performed at various time points after their initial injury. Spearman's correlation coefficients and Kruskall-Wallis' test were used to examine the relationship between patient age, emergency room Glasgow Coma Scale (GCS), Injury Severity Score, prehospital hypotension, prehospital hypoxia, mechanism of injury, type of injury, side of injury, global average CBF, lobar CBF, number of lobes with CBF below normal, and GOS (discharge, 3 and 6 months). Univariate ordinal regression was performed using these same variables and in combination with principle component analysis (PCA) to determine independent variables for multi-variate ordinal regression. Significant correlation between age, GCS, prehospital hypotension, type of injury, global average CBF, lobar CBF, number of lobes below normal CBF, and GOS was found. Individual lobar CBF was highly correlated with global CBF and the number of lobes below normal CBF. PCA found one principle component among these three CBF variables; therefore, average global CBF and number of lobes with CBF below normal were each chosen as independent variables for multiple ordinal regression, which found age, GCS, and prehospital hypotension, global average CBF, and number of lobes below normal CBF significantly associated with GOS. This study found global average CBF and lobar CBF significantly correlated with GOS at follow-up. There was, however, no individual cerebral lobe that was more predictive than any other, which puts into question the value of calculating lobar CBF versus global CBF in predicting GOS.