INTRODUCTION A common presentation to the emergency department (ED) is the trauma patient with altered sensorium who is presumed to be alcohol intoxicated by physicians based on their olfactory sense. ED physicians may often leave patients suspected of alcohol intoxication aside until the effects wear off, potentially missing trauma as the source of confusion. This often results in delays in diagnosing acute potentially life-threatening injuries in patients with presumed alcohol intoxication. OBJECTIVE This study aimed to determine the accuracy of a physician's olfactory sense for diagnosing alcohol intoxication. METHODS Patients suspected of major trauma in the ED underwent an evaluation by the examining physician for alcohol odour and a blood alcohol level. Alcohol intoxication was defined as a serum ethanol level ≥80 mg/100 ml. Data were reported as means with 95% CI or proportions with IQR 25-75%. RESULTS 151 patients (70% men) were enrolled, with a median age of 45 years (IQR 33-56). The prevalence of alcohol intoxication was 43% (95% CI 35% to 51%). OPERATING CHARACTERISTICS Physician assessment of alcohol intoxication: sensitivity 84% (95% CI 73% to 92%), specificity 87% (95% CI 78% to 93%), positive likelihood ratio 6.6 (95% CI 3.8 to 11.6), negative likelihood ratio 0.18 (95% CI 0.1 to 0.3) and accuracy 86% (95% CI 80% to 91%). 7.3% (95% CI 4% to 13%) of patients were falsely suspected of being intoxicated. CONCLUSIONS Although the physicians had a high degree of accuracy in identifying patients with alcohol intoxication based on their olfactory sense, they still falsely overestimated intoxication in significant numbers of non-intoxicated trauma patients.