Between 35% to 50% of traumatic brain injury (TBI) patients are under the influence of alcohol. Alcohol intoxication may limit the ability of the Glasgow Coma Scale (GCS) to accurately assess severity of TBI. We hypothesized that alcohol intoxication significantly depresses GCS scores of TBI patients. A 10-year, retrospective analysis of a Level I trauma center registry was undertaken. The study population consisted of all blunt injured TBI patients tested for blood alcohol concentration (BAC, n = 1,075). Patients were divided into two groups; intoxicated (mean BAC 202 +/- 77 mg/dL, n = 504) and nonintoxicated (BAC = 0, n = 571). TBI was classified using ICD-9 codes as concussion alone (ICD-9 850, n = 90) and intracranial injury (ICI, ICD-9 851-854, n = 985). Severity was further classified using the Abbreviated Injury Score (AIS). Mean GCS score was compared between the two groups. Patients who were either intubated or hypotensive upon arrival were analyzed separately to rule out confounding effects on GCS score. Severely intoxicated patients (BAC >250 mg/dL, [mean +/- SD] 309 +/- 54 SD, n = 118) were similarly compared. Finally, multivariate linear regression analysis was undertaken to determine whether BAC level was an independent predictor of GCS score while controlling for confounding factors. Intoxicated and nonintoxicated TBI patients were clinically similar. Alcohol intoxication had little effect on GCS score, with less than a single point difference in all types of TBI, except the most severely injured (AIS 5 injuries, GCS score difference 1.4 points). These results were not altered by endotracheal intubation, systemic hypotension, or severe intoxication. Similarly, BAC was not a significant independent predictor of GCS score in a multivariate model. Alcohol intoxication does not result in clinically significant changes in GCS score for patients with blunt TBI. Hence, alterations in GCS score after TBI should not be attributed to alcohol intoxication, as doing so might result in inappropriate delays in monitoring and therapeutic interventions.
Read full abstract