Relevance of rat liver tumors to human hepatic and endometrial cancer.

Concern for a risk of cancer in humans has been prompted by the appearance of hepatocellular carcinomas in tamoxifen-treated rats. However, there is no evidence of excess hepatocellular carcinoma among tamoxifen-treated women. Moreover, liver tumors are not induced in tamoxifen-treated mice or hamsters. An explanation for the species selectivity of this toxic effect may relate to the greater rate of formation of reactive intermediates in rats than either mice or humans. This results in persistent liver DNA adducts in tamoxifen-treated rat liver exceeding those produced in treated mice or in background levels measured in women taking tamoxifen. Another observation that reduces concern for human carcinogenesis is that bioactivation of tamoxifen may be inducible at dosages used in rodent cancer bioassays but not in those used clinically. Suggestions that endometrial cancer may be tamoxifen related are not supported by rodent data for endometrial cancer. Indeed, endometrial tissue lacks the necessary tamoxifen bioactivating capacity of liver consistent with the absence of DNA adducts in the endometrium of women taking tamoxifen. Finally, it seems doubtful that the colon is a target for human carcinogenesis of tamoxifen given the negative epidemiologic studies and high-dose rodent studies. In summary, there is at present no sound scientific basis for extrapolation of rat liver cancer findings to risks of liver, endometrial, and colon cancer in women receiving tamoxifen.