The relationship between alcohol use and injury severity was investigated in trauma patients admitted to a tertiary referral hospital during a 23-month period. Admission blood alcohol levels (BALs) were obtained on 427 trauma patients, who were stratified into three groups: those with no measurable blood alcohol, those within the legal limit of 100 mg/dL, and those over the legal limit or intoxicated. The no-alcohol group had significantly lower injury severity than the other two groups (p less than 0.001). Even when the BAL was well within the legal limit, injuries suffered by those in the alcohol-positive groups were more severe than those in the no-alcohol group. Confirmatory evidence of the effect of alcohol on injury severity was reflected by a 2.3% mortality in alcohol-negative patients compared with a 13.3% death rate in alcohol-positive patients (p less than 0.0001). To assess the potentially confounding effect of alcohol on injury scoring accuracy, we examined the change in Glasgow Coma Scale (GCS) scores following admission. No significant differences were found when admission GCS values were compared with GCS determinations made 24 hours following admission by separate observers. To correct for any potential bias as a tertiary referral center, repeat analysis with exclusion of transferred patients was done with essentially no change in results. Our data revealed a highly significant relationship between alcohol use, degree of injury, and resource consumption.