Ninety-seven chronic alcoholics, both impaired and unimpaired, without clinically evident liver disease, showed significantly more cerebral atrophy on computed tomography than age-matched neurologic controls. Age was the variable most highly correlated with cerebral atrophy measurements, and it accounted for most of the correlations between atrophy and functional impairment, except in the Wernicke-amnesic group. Analysis of the slopes of atrophy scores versus age showed a more rapid "rate" of development of cerebral atrophy in alcoholics compared with controls. There were no correlations between liver biopsy scores (51 cases), drinking history (47 cases), or dietary intake (39 cases) and cerebral atrophy measurements.