Groups of 50 male and 50 female Sprague-Dawley rats were given food containing sufficient doxefazepam, a benzodiazepine derivative, to ensure intakes of 0, 3, 10, or 30 mg/kg/day. These dosages respectively correspond to 2, 20, and 60 times the mean daily hypnotic dose level of an adult man. Rats were treated for 104 weeks and then euthanized. An extensive autopsy was performed on those animals that died intercurrently and on euthanized animals. The chronic administration of doxefazepam did not influence the survival of the rats. No treatment-related changes in clinical signs and body weight gains occurred and malignant tumor rates were similar in controls and treated animals. A significant linear trend in the incidence of hepatocellular neoplasms, primarily benign, was observed in the female treated groups. This higher incidence was not associated to a higher occurrence of focal hyperplasia or other preneoplastic lesions in treated rats. The brain, a target organ for the pharmacological activity of doxefazepam, was carefully examined to search for microscopic foci of proliferative cells. A total of 12 and 6 malignant gliomas were observed in male and female rats, respectively; only two were noticed at autopsy. These tumors were mainly of the oligodendroglioma type commonly found in aged rats. Their incidence was slightly higher in treated rats, but results were not of statistical significance. The overall evaluation of the present study indicates that doxefazepam is noncarcinogenic in rats. However, the increase in liver adenomas found here as well as in previous bioassays with similar drugs and the lack of reliable historical data on the incidence of brain tumors in benzodiazepine-treated rodents suggest that additional experimental and epidemiological studies should be undertaken to exhaustively assess the toxic potential of this widely used class of drugs.