Oxymetholone is generally assumed to be a nongenotoxic carcinogen. This assumption is based primarily on the results of an Ames test, existing data in repeat-dose toxicology studies, and the predicted results of a 2-yr National Toxicology Program (NTP) rat carcinogenicity bioassay. To provide a comprehensive assessment of its genotoxicity in a standard battery of mutagenicity assays, oxymetholone was tested in microbial and mammalian cell gene mutation assays, in an in vitro cytogenetics assay (human lymphocytes), and in an in vivo micronucleus assay. Oxymetholone was also tested in an in vitro morphologic transformation model using Syrian hamster embryo (SHE) cells. These studies were initiated and completed prior to the disclosure of the results of the NTP bioassay. Oxymetholone was tested at doses up to 5,000 microg/plate in the bacterial plate incorporation assay using 4 Salmonella strains and the WP2 uvrA (pKM101) strain of Escherichia coil. There was no induction of revertants up to the highest dose levels, which were insoluble as well as toxic. In the L5178Y tk+/- mouse lymphoma assay, doses up to 30 microg/ml reduced relative survival to approximately 30% with no increase in mutants. Male or female human lymphocytes were exposed in vitro to oxymetholone for 24 hr without S9 or 3 hr with S9 and evaluated for the induction of chromosomal aberrations. There was no increase in aberration frequency over control levels and no difference between male and female cells. Peripheral blood from Tg.AC transgenic mice treated dermally for 20 wk with 0, 1.2, 6.0, or 12.0 mg/day of oxymetholone and from p53 transgenic mice treated orally by gavage for 26 wk with 125, 625, or 1,250 mg/kg/day of oxymetholone was evaluated for micronuclei in polychromatic and normochromatic erythrocytes. There was no difference in micronuclei frequency between control and treated animals. These results confirm that oxymetholone is not genotoxic in a comprehensive battery of mutagenicity assays. In the SHE assay, oxymetholone produced a significant increase in morphologically transformed colonies at dose levels of 13-18 microg/ml. The lack of genotoxicity of oxymetholone, the positive response in the in vitro transformation assay, and the results of transgenic mouse carcinogenicity assays will provide an interesting perspective on the results of an on-going NTP rat carcinogenicity bioassay.