Tamoxifen and tumorigenicity: a predictable concern.

Abstract

In 1988, we (1) demonstrated that tamoxifen encouraged the growth of a hormone-responsive endometrial carcinoma but blocked the estrogen-stimulated growth of a breast carcinoma cotransplanted in the same athymic mouse. Tamoxifen produced target site-specific actions, so that the growth of an occult endometrial carcinoma might continue during long-term tamoxifen therapy despite the control of breast cancer growth. We (1) suggested that . . . a [large] cohort of patients under long term tamoxifen therapy (>5 years) needs to be monitored for the occurrence of the tamoxifen-stimulated In 1989, Fornander et al. (2) provided the first clinical evidence that tamoxifen reduced the incidence of second primary breast cancers but increased the incidence of endometrial carcinomas. The concept of target site specificity was apparently also true for humans. There are now more than 200 cases of endometrial carcinoma that have been described in patients who are taking or have completed taking tamoxifen therapy (3). These data, however, should be placed in perspective, because the total number of women with endometrial carcinomas is very small compared with the millions of women who have been exposed to tamoxifen and have benefited from the drug. Nevertheless, the report by Magriples et al. (4) stated that . . . women receiving tamoxifen as treatment for breast cancer, who subsequently develop uterine cancer, are at risk for high grade endometrial cancers that have poor prognosis; their findings were understandably of major concern. At about the same time, high-dose tamoxifen was shown to produce liver cancer in rats (5,6). One could, therefore, predict that the clinical and laboratory observations would eventually be linked. In this issue of the Journal, Rutqvist et al. (7), reporting for the Stockholm Breast Cancer Study Group, extend their initial observations (2) on the incidence of second primary tumors after adjuvant tamoxifen therapy by pooling their results with selected data from the Danish Breast Cancer Group Trial (8) and the South-Swedish Trial (9). The good news is that the Stockholm group still confirmed the ability of tamoxifen to reduce the incidence of second primary breast cancers, despite the fact that the Danish Trial (8) originally found no effect because the treatment time (48 weeks) was too short. Most importantly, they found no evidence that tamoxifen increases liver cancer in humans. (The liver is the primary site of carcinogenesis during high-dose tamoxifen treatment in the rat.) However, their conclusion (7) that . . . the endometrium and gastrointestinal organs may be target sites for tamoxifen-induced carcinogenesis in demands a thorough examination. Tamoxifen does act as a carcinogen and does produce liver cancer in rats (5,6); however, it is extremely important to appreciate that the phenomenon is dependent on dose, duration of treatment, and species (10). There appears to be a threshold dose (i.e., >3 mg/kg) at which daily administration of tamoxifen from 6 weeks of age for the rat's life produces both DNA adducts and liver tumors in the surviving animals over the next 2-year period (5). The doses used, however, are not comparable to those used clinically in the United States, i.e., 20 mg daily or approximately 250 |ig/kg for a 70-kg woman. Moreover, the treatment durations in patients are not comparable to those in the rat model (10). Patients are treated for 5 years, which is usually 6% of a postmenopausal woman's life; in contrast, animals have been treated beginning at puberty and continuing for the rest of their lives to produce tumors. It is important to appreciate that the therapeutic studies to determine the value of tamoxifen as an antitumor agent were completed using doses in rats comparable to those used in clinical trials (11). A realistic model to evaluate concerns about carcinogenesis with tamoxifen would be to use the therapeutic dose and equivalent times and durations to those used clinically. This model has been suggested (10,12), but no studies have been undertaken by the toxicologists who have raised concerns about tamoxifen. Current research is focused on the metabolic activation of tamoxifen in the rat liver to determine its relevance (if any) to humans (70). If DNA adduct formation is critical to carcinogenesis in humans, it will be important to detect comparable levels of DNA adducts in human tissues. At present, the levels of DNA adducts in the livers of patients treated with tamoxifen have not been reported, but unpublished observations (Medical Research Council Toxicology Meeting, Leicester, England, 1994) suggest that these levels are not elevated above control levels. It is, therefore, reassuring that