In a group of 343 working men, only 34 of whom regularly drank more than 60 ml of ethanol per day, logistic regression was used to determine the combination of biomarkers which best discriminated between those who regularly drank less than 30 ml or 30 ml or more of ethanol daily. The index consisted of apolipoprotein A-II, uric acid, gamma-glutamyl transpeptidase and mean corpuscular volume. Even with the relatively low level of alcohol intake reported in these subjects (assessed using a 7-day retrospective alcohol diary), the groups with higher or lower intake of alcohol were separated with a sensitivity of 68%, a specificity of 74% and a positive predictive value of 71%. Systolic blood pressure, adjusted for age and body mass index, was also linearly related to alcohol intake, and when included in the biomarker index improved the proportion correctly classified from 71% to 75%. Using the biomarkers only, 94% of subjects regularly drinking at least 80 ml of ethanol equivalent per day and 100% of the non-drinkers were correctly classified.