The aim of this prospective study was to identify alcoholic liver disease or covert alcohol abuse in unselected consecutive patients referred to a gastroenterologic out-patient's clinic for elevated liver enzymes. One-hundred-and-thirteen patients were questioned about alcohol consumption, trauma history and loss of driver's licence. Laboratory tests claimed to reflect alcohol consumption and liver biopsy were taken. Using data from patients with the highest (26 patients) and lowest (29 patients) stated consumptions, logistic regression analysis identified violence score (trauma score + driver's licence score) and a laboratory index combining MCV, ASAT/ALAT ratio and IgA values as significant independent predictors of alcohol abuse (index = 2.3 x violence score +1.08 x laboratory index-3.19). Twenty-six patients openly admitted alcohol abuse (> 300 g week-1) and 85% of these had alcoholic liver damage. The index identified a further 12 patients as abusers who stated an intermediate alcohol consumption (25-300 g week-1). Half of these had alcoholic liver damage. Thus, a total of 34% of the patients were identified as abusers and in one-third of these the abuse was covert.