Abstract
BACKGROUND: The association of alcohol use with in-hospital trauma deaths remains unclear. This study identifies the association of blood alcohol content (BAC) with in-hospital death accounting for injury severity and mechanism. METHODS: This study involves a historical cohort of 46,222 admissions to a statewide trauma center between January 1, 2002, and October 31, 2011. Blood alcohol was evaluated as an ordinal variable: 1 mg/dL to 100 mg/dL as moderate blood alcohol, 101 mg/dL to 230 mg/dL as high blood alcohol, and greater than 230 mg/dL as very high blood alcohol. RESULTS: Blood alcohol was recorded in 44,502 patients (96.3%). Moderate blood alcohol was associated with an increased odds for both penetrating mechanism (odds ratio [OR], 2.22; 95% confidence interval [CI], 2.04-2.42) and severe injury (OR, 1.25; 95% CI, 1.16-1.35). Very high blood alcohol had a decreased odds for penetrating mechanism (OR, 0.75; 95% CI, 0.67-0.85) compared with the undetectable blood alcohol group. An inverse U-shaped association was shown for severe injury and penetrating mechanism by alcohol group (p < 0.001). Moderate blood alcohol had an increased odds for in-hospital death (OR, 1.50; 95% CI, 1.25-1.79), and the odds decreased for very high blood alcohol (OR, 0.69; 95% CI, 0.54-0.87). An inverse U-shaped association was also shown for in-hospital death by alcohol group (p < 0.001). Model discrimination for in-hospital death had an area under the receiver operating characteristic curve of 0.64 (95% CI, 0.63-0.65). CONCLUSION: Injury severity and mechanism are strong intermediate outcomes between alcohol and death. Severe injury itself carried the greatest odds for death, and with the moderate BAC group at greatest odds for severe injury and the very high BAC group at the lowest odds for severe injury. The result was a similar inverse-U shaped curve for odds for in-hospital death. Clear associations between blood alcohol and in-hospital death cannot be analyzed without consideration for the different injuries by blood alcohol groups. LEVEL OF EVIDENCE: Epidemiologic study, level III.