1. Liver biopsies were performed in healthy control subjects and in subjects with alcoholic and non-alcoholic liver disease in order to examine alcohol dehydrogenase (ADH; EC 1.1.1.1) and aldehyde dehydrogenase [ALDH; aldehyde dehydrogenase (NAD+); EC 1. 2. 1. 3] activities. Erythrocyte ALDH and ethanol metabolism were also investigated in the same subjects. 2. Fifteen per cent of the subjects studied (seven of 48 subjects tested) presented atypical ADH activity, characterized by elevated activity at pH 7.4 or 8.8 compared with that found in subjects with the usual ADH form. However, the ethanol elimination curves obtained in two subjects with atypical ADH were indistinguishable from the kinetics of the group with normal ADH. Subjects displaying atypical ADH activity showed normal liver and erythrocyte ALDH activities. 3. Considering only the subjects with the normal ADH form, hepatic ADH activity was unaltered in subjects with non-alcoholic liver disease (chronic hepatitis or cirrhosis) and in those with alcoholic steatosis. Subjects with alcoholic hepatitis or alcoholic cirrhosis showed a lower ADH activity compared with the healthy control group. 4. In spite of the changes detected in subjects with alcoholic liver disease, curves of blood ethanol concentration after oral administration of 0.4 g of ethanol/kg were indistinguishable between the alcoholic hepatitis group and the control group. 5. Hepatic ALDH activity, assayed at 300 μmol/l acetaldehyde, was found to be diminished in all liver pathologies investigated, regardless of their aetiology. Nevertheless, erythrocyte ALDH activity was not modified in subjects with non-alcoholic or alcoholic liver disease. As a result of these findings, no relationship was found between hepatic and erythrocyte ALDH. 6. In summary, our data demonstrate that (a) marked modifications in ADH activity, as found in patients with atypical ADH or in subjects with alcoholic liver disease, are not accompanied by parallel alterations in the kinetics of ethanol disappearance, suggesting that ADH activity per se does not limit ethanol metabolism in vivo, (b) hepatic high-Km ALDH activity is decreased in patients with liver disease independent of alcoholism, and therefore decreased ALDH activity cannot be considered as a primary defect in alcoholism but as a consequence of liver damage, and (c) erythrocyte ALDH does not reflect hepatic high-Km ALDH.