International Agency for Research on Cancer (IARC) - Summaries & Evaluations


VOL.: 21 (p. 131)

Steroid hormones are essential for the growth, differentiation and function of many tissues in both animals and humans. It has been established by animal experimentation that modification of the hormonal environment by surgical removal of endocrine glands, by pregnancy or by exogenous administration of steroids can increase or decrease the spontaneous occurrence of tumours or the induction of tumours by applied carcinogenic agents. In humans, endogenous hormones are important in the initiation and progression of tumours. The incidence of tumours in humans could be altered by exposure to various exogenous hormones, singly or in combination.

For an administered oestrogen seriously to influence the human hormonal environment, its intake must be equal to, or greater than, the amounts of oestrogens produced endogenously. For contraceptive medication to be effective, the intake of steroids must be sufficient to alter the hormonal environment. The doses at which steroids are administered for other therapeutic purposes are considerably higher; the possibility that a carcinogenic risk may be involved in such medication must therefore be considered. For example, the minimum effective dose of diethylstilboestrol that produces mammary carcinogenesis in mice is of the same order as the doses used in women for therapy of the climacteric or following ovariectomy.

Animal data

Administration of the naturally occurring oestrogens, oestradiol-17b and oestrone, increases the incidence of tumours in a number of organs in a variety of animal species. There is limited evidence for the carcinogenicity of oestriol, whereas insufficient evidence was available concerning the conjugated oestrogens. Diethylstilboestrol has a carcinogenic potential comparable with those of oestradiol-17b and of oestrone, and there is no evidence to support the suggestion that its carcinogenic properties are due to some special biological function other than its oestrogenic activity. Ethinyloestradiol and mestranol have been shown to be carcinogenic in animals, but there is no evidence that they are more or less carcinogenic than are other oestrogens with comparable levels of oestrogenic activity. Insufficient data were available to evaluate the carcinogenicity of dienoestrol and chlorotrianisene, which are structurally similar to diethylstilboestrol. Because of lack of experimental data, no attempt has been made to show a relationship between carcinogenic potential and oestrogenic activity for any of the compounds considered.

In the majority of experiments in which animals were treated with oestrogens and which resulted in carcinogenesis, high dose levels were used. The information currently available is, however, inadequate to indicate the minimum doses required to produce carcinogenesis, and these could be much lower than those commonly employed in animal studies.

There is limited evidence that natural progestins have a carcinogenic potential. There is, however, evidence that low doses of progesterone administered over long periods act in combination with carcinogenic agents, such as some viruses or chemicals, to enhance tumour development. In part, therefore, long-term administration of synthetic progestins may produce a comparable hazard by increasing the incidence of tumours due to other agents. This is dependent on the degree of progestational activity possessed by the compound in question relative to its other hormonal characteristics.

Animal data concerning the synthetic progestins provide either insufficient or limited evidence of their carcinogenic potential. Combination with oestrogens seems to enhance their carcinogenic potential. There is limited evidence for the carcinogenicity of some progestins (chlormadinone acetate, medroxyprogesterone acetate, megestrol acetate) in dogs, which is associated with the production of mammary tumours.

There is sufficient evidence for the carcinogenicity of testosterone in mice and rats.

The mechanism(s) by which hormones act in the induction of cancer is not understood. Although many carcinogens show a mutagenic action, no sex hormones, including diethylstilboestrol, nor any of their metabolic products, have so far convincingly been shown to be mutagenic; however, covalent binding of their metabolites to DNA and other results from short-term tests that indicate interaction with DNA have been reported.

The evidence suggests that steroid hormones may stimulate carcinogenesis in several ways and, in addition, provide a background for subsequent tumorigenesis by chemical, physical or viral agents and promote the growth and metastasis of tumours once they have been initiated.

The administration of sex steroids and diethylstilboestrol prenatally or neonatally, especially to mice, results in teratogenic changes and altered endocrine function. Carcinogenesis may occur consequent to these alterations. Both in neonates and adults, sex steroid-stimulated prolactin secretion may lead to increased mammary carcinogenesis.

Human data

Sex hormones, such as those considered in this monograph, have been and are used extensively in human therapy. When they are used for the treatment of disseminated cancer, such as that of the breast, prostate and endometrium, their effect on tumour growth and the severity of their side-effects are the major considerations. The use of sex hormones in therapy for other conditions (for example, menstrual disorders, climacteric syndrome, pregnancy maintenance, osteoporosis, abnormal protein metabolism, gonadal deficiency) makes the question of carcinogenic hazard more pertinent. With the continuing development of steroid use for the control of conception, the question of possible carcinogenic hazards has become of major importance.

Epidemiological studies to evaluate possible carcinogenic effects of administered oestrogens and progestins in humans suffer from two major difficulties. Firstly, the interval between the commencement of administration and the possible appearance of cancer may be long. Secondly, to detect a small or moderate change in risk, observations on large numbers of subjects are required.

With these reservations in mind, the following conclusions can be made:

1. Intrauterine exposure to diethylstilboestrol - The administration of diethylstilboestrol to women during pregnancy is causally associated with an increased risk of vaginal and cervical clear-cell adenocarcinoma in exposed daughters. Non-neoplastic changes of the female genital tract, including transverse fibrous septae, vaginal adenosis and cervical ectropion, have been observed frequently; non-malignant structural changes in the male reproductive tract have also been reported.

2. Oestrogens administered to adults - The administration of oestrogens (including diethylstilboestrol) to adult women is causally associated with an increased incidence of endometrial cancer. There is also a possibility that the risk of breast cancer is increased by such therapy, but the evidence is not conclusive. There is good evidence that factors which increase or prolong exposure of the uterus to endogenous oestrogens result in an increased risk of developing endometrial cancer. At present, the specific role of endogenous hormones in the development of breast cancer is unclear.

3. Progestins - There are no adequate data in humans to assess whether progestins used as contraceptives (either as pills or as injections) alter the risk of developing cancer.

4. Androgens - Prolonged androgen therapy may be associated with an increased risk of hepatocellular tumours, but the evidence is not conclusive.

5. Oral contraceptives - Oral contraceptive use decreases the risk of benign breast disease. There is no clear evidence that their use alters the risk of breast cancer, although limited data suggest that the preparations may interact with other risk factors (e.g., late age at first pregnancy, presence of benign breast lesions) in the development of the disease. Oral contraceptives may increase the risk of cervical dysplasia and carcinoma in situ after long-term use. Few data are available concerning invasive cancer of the cervix. Sequential oral contraceptives may increase the risk of endometrial cancer, but the evidence is not conclusive.

Long-term use of oral contraceptives markedly increases the relative risk of hepatocellular adenoma. The effect is greater in older women and in those using preparations containing high doses of steroids.

Oral contraceptives may increase the risk of malignant melanoma and of pituitary adenoma and decrease the risk of ovarian cancer, but the available data are inadequate for a proper assessment.

    See Also:
       Toxicological Abbreviations
       TESTOSTERONE (JECFA Evaluation)
       Testosterone (PIM 519)
       Testosterone (IARC Summary & Evaluation, Volume 6, 1974)