Depot-medroxyprogesterone acetate (DMPA) and cancer: memorandum from a WHO meeting.

Memoranda/Memorandums Depot-medroxyprogesterone acetate (DMPA) and cancer: Memorandum from a WHO meeting* Depot-medroxyprogesterone acetate (DMPA) is a long-acting progestational contraceptive, which is administered by injection. It was developed in the mid- 1960s, when it was seen as a method that would be particularly useful for women seeking reversible contraception who had difficulty taking a pill every day. DMPA has been shown to be a highly effective contraceptive, and it has proved acceptable in a variety of settings. The drug is licensed as a contraceptive in more than 90 countries and has been widely used in a number of them, such as Thailand and New Zealand. On a worldwide basis, the licensing, acceptability, and prevalence of use have been influenced by concern that DMPA may increase the risk of cancer. Cancer of the breast has been a particular concern. This Memorandum reviews comprehensively the results of toxicological tests in animals and epidemiological studies in humans concerning the carcinogenicity of DMPA. Toxicology DMPA has undergone thorough toxicological eva- luation in a number of animal species, and has been tested in short- and long-term toxicity studies in rodents, rabbits and monkeys. It has been examined for its effect on reproduction in mice, rats and rab- bits, and for carcinogenic potential in rats, mice, * This Memorandum was drafted on the occasion of a meeting convened by the World Health Organization in Geneva in May 1993 by the following signatories: P. Boyle, Milan, Italy; C. Chil- vers, Nottingham, England; E. Ferraz, Brasilia, Brazil; B. Hulka, Chapel Hill, NC, USA; R. King, Guildford, England; C. La Vec- chia, Milan, Italy; D. Petitti, San Francisco, CA, USA (Rappor- teur); Pisake Lumbiganon, Khon Kaen, Thailand; D. Skegg, Dunedin, New Zealand (Chairman); D. Thomas, Seattle, WA, USA; 0. Viegas, Singapore. Representative of drug regulatory authorities: A. Jordan, Rockville, MD, USA. Representatives of collaborating agencies: J. Fortney, Research Triangle Park, NC, USA; H. Peterson, Atlanta, GA, USA; P. Ramachandran, New Delhi, India; R. Spirtas, Bethesda, MD, USA. Observer: R. Lande, Baltimore, MD, USA. WHO Secretariat: H. Bathija, S. Mehta, 0. Meirik (Secretary), Special Programme of Research, Development and Research Training in Human Reproduction. Requests for reprints should be addressed to the Director, Spe- cial Programme of Research, Development and Research Train- ing in Human Reproduction, World Health Organization, 1211 Geneva 27, Switzerland. A French translation of this article will appear in a later issue of the Bulletin. Reprint No. 5428 beagle dogs, and rhesus monkeys. Genotoxicity tests have been performed both in vitro and in vivo. Tests of genotoxicity and carcinogenicity are discussed below. Genotoxicity Medroxyprogesterone acetate (MPA) was negative in the Ames test both with and without an in vitro metabolic activation system.a It was also negative in the DNA damage/alkaline elution assay using Chi- nese hamster lung fibroblasts and in the Micronu- cleus test in rats.a In dogs (1) and hamsters (2), MPA produced chromosome aberrations in germ cells when administered in vivo. MPA also increased the induction of sister chromatid exchange in rabbit lym- phocytes (3) and in mouse kidney fibroblasts (4). Carcinogenicity Carcinogenicity studies have been done in monkeys, rats, mice, and dogs. Monkeys. In 1968-1977, a ten-year study of DMPA in rhesus monkeys was done. A total of 52 monkeys were divided into 4 groups, a control group and a Upjohn Corporation, unpublished reports to U.S. Food and Drug Administration (FDA), 1979, 1981. Bulletin of the World Health Organization, 71 (6): 669-676 (1993) © World Health Organization 1993 669 Memorandum groups that received 1, 10, or 50 times the human contraceptive dose of DMPA (5). The low-dose group had 4 animals and the other two groups had 16. There was an increased incidence of total neoplasms in the high-dose group. Endometrial carcinomas were seen in two high-dose monkeys. Because DMPA caused uterine atrophy in the monkeys, there was debate about whether the endometrial tumours were uterine in origin or originated from the endo- cervix, or perhaps from endometrial plaques. This debate has still not been resolved. Mice and rats. In a study done to support an applica- tion to the U.S. Food and Drug Administration (FDA) for new drug approval, rats and mice (50 ani- mals/sex/group) were injected intramuscularly with DMPA monthly at doses of 0, 2, 100 or 200 mg/kg (5). The control and high-dose mice were examined his- tologically after 18 months and the rats after 24 months. There was some increase in the incidence of malignant lymphoma, hepatoma, and haemangioma in the high-dose mice, but these findings were not considered significant by the reviewing FDA phar- macologist. There was no increase in any tumour type in the rats. In a more recent study, BALB/c mice were treat- ed with 40 mg MPA injected subcutaneously every 2 months for 1 year (6). Mammary adenocarcinomas were found in 16 of 40 treated mice, compared with none in the untreated control group. MPA treatment has also been reported to increase the incidence of mammary tumours in C3H (7) and SHN (8) mice. There are a number of published studies on the promoter effect of DMPA which show that it may inhibit (9, 10) or stimulate (11) growth of dimethyl- benz[a]-anthracene-induced mammary tumours in rats. The inconsistency of the results may reflect both temporal and dosage differences in experimen- tal design (5). Recently, DMPA was shown to increase the incidence of tumours of the uterine cer- vix in mice treated with methylcholanthrene and a beeswax-impregnated thread, when placed inside the canal of the uterine cervix (12). DMPA treatment decreased the latency period and increased the inci- dence of mammary tumours in N-methyl-N-nitroso- urea-treated BALB/c mice (13). In the human breast cancer cell line MCF-7, MPA initially inhibited cell growth, but following continued treatment, the effect was lost and further exposure to the drug enhanced cell proliferation (14). Dogs. In beagle dogs, DMPA increases the incidence of both malignant and benign mammary tumours when given at doses that are at and above the human contraceptive dose. Table 1 shows data from a study which is representative of other studies of the effects of DMPA in beagles (15). In this study, six-month- old female beagle dogs were hysterectomized prior to the start of the study. Forty dogs served as con- trols and 20 per group were treated with DMPA at 1, 10, or 25 times the human contraceptive dose (3, 30, or 75 mg/kg), which was given by intramuscular injection every 3 months. Survival in this study, as in other studies of DMPA in dogs, was poor. Although 30 of the 40 control dogs survived 7 years, survival in the treated groups was 14, 0, and 1 out of 20 dogs in each of the low-, mid-, and high-dose groups, respectively. None of the control dogs developed malignant mammary tumours during the study. In contrast, malignant mammary neoplasia (some with metastases) devel- oped in 5 dogs in the low-dose group, in 8 dogs in the mid-dose group, and in 8 dogs in the high-dose group. In beagle dogs, progestogens stimulate pituitary secretion of growth hormone (16), which has a marked mammotrophic effect in this species (17). Different progestogens have different potencies for stimulation of the mammary gland in beagles (18). In one study (19), induction of mammary tumours by high doses of MPA was reduced by prior hypophy- sectomy, suggesting that enhanced pituitary secretion of growth hormone is a mechanism for MPA- induced tumours. In humans, contraceptive doses of DMPA do not alter the secretion of growth hormone (20). Thus, carcinogenicity testing of DMPA in beagle dogs results in hormonal changes not seen in humans under conditions of normal use. The toxicological profile of DMPA is often compared to that of the orally administered steroidal contraceptives that are in use worldwide. As noted recently (21), oral doses of 19-nortestosterone deriv- atives (83 to 200 times greater than the human dose) produced serum drug levels in rats that were actually lower than the drug levels in humans taking the con- Table 1: Malignant and benign mammary tumours in beagle dogs treated with 1, 10, and 25 times the human contraceptive dose of DMPA as an intramuscular injec- tion every three months, compared with untreated dogs Control DMPA-treated: group 1 x dose 10 x dose 25 x dose No. of dogs No. alive after 7 years No. of malignant mammary tumours Benign mammary tumours (%) 40 30 0 20 20 20 14 0 1 8 10 85 95 85 WHO Bulletin OMS. Vol 71 1993670 Depot-medroxyprogesterone acetate and cancer traceptive dose. In contrast, DMPA administration resulted in drug exposure far in excess of human drug exposure. Levonorgestrel was administered orally to dogs at doses up to 25 times the human dose on a mg/kg basis; no increase in tumours was noted and the drug was subsequently approved for contraception. However, when given at much higher doses, the steroid was significantly tumorigenic despite the fact that even at this high dose the animal exposure was only 7 times higher than the exposure in humans (5). Conclusions. Because of significant differences in bioavailability, animals injected with DMPA were exposed to far greater concentrations of the progesto- gen than were animals tested with orally adminis- tered 1 9-nortestosterone derivatives. Moreover, the beagle seems to respond uniquely to DMPA and other progestogens, with an increase in serum growth hormone which is probably responsible for the tumorigenic effect of DMPA on the mammary gland. Women taking contraceptive doses of DMPA have no significant elevation of serum growth hormone. Considering all the available toxicological data there seems to be no reason to believe that DMPA is different from other progestogens in its tumorigenic potential. Breast cancer Although now considered an inappropriate model, the effect of DMPA on mammary tumours in beagle dogs has made its effect on breast cancer in humans of particular concern. The lack of widespread use of DMPA has hampered the epidemiological study of breast cancer in relation to DMPA use. Only recently have high quality epidemiological data on this topic become available. Epidemiological studies. The results of six epidemi- ological studies of DMPA and breast cancer have been published. Four early studies (22-25) all suffer from severe methodological problems and do not contribute to an understanding of the relationship between DMPA and breast cancer. Table 2 summarizes the two case-control studies (26, 27) that provide meaningful information on the risk of breast cancer in users of DMPA. The first study (26) was a population-based case-control study that involved the entire country of New Zea- land. The second study (27) was a WHO hospital- based case-control study, in which data were gath- ered in five centres in three countries. Both studies enrolled a large number of cases of breast cancer, and the prevalence of use of DMPA in the control groups of both studies was greater than 10%. The estimated relative risk of breast cancer in women who have ever used DMPA was 1.0 in the New Zea- land study and 1.21 in the WHO study. In both studies, there was no trend in risk with the duration of DMPA use. The estimated relative risk of breast cancer for women of all ages who had used DMPA for 3 or more years was 0.93 (95% CI, 0.62 - 1.41) in the WHO study. Estimated risk was 1.2 (95% CI, 0.59-2.2) for women who used DMPA for 6 or more years in the New Zealand study. Table 3 shows the results of analyses of data examining the risk of breast cancer in women who have ever used DMPA, by age at diagnosis and by age at first use. Both studies show evidence of an association of ever having used DMPA with increased risk of breast cancer diagnosed before 35 years of age. There was no association of DMPA use with increased risk of breast cancer diagnosed after age 35 years in either study. Table 4 shows results of analyses from the New Zealand and WHO studies which examined breast can- cer risk in relation to time since first use and time since last use. In both studies, estimates of relative risk were highest in relation to recent use. Younger women tend to be more likely to be recent users and this may explain the apparent effect of age at diagnosis. Both the New Zealand and WHO studies consid- ered use of DMPA relative to first full-term pregnan- cy. The numbers of exposed women were very small, and firm conclusions about risk could not be drawn for this reason. Conclusions. Epidemiological studies provide reas- surance that use of DMPA does not increase breast Table 2: Summary of the New Zealand (26) and WHO (27) case-control studies (estimated relative risk (RR) and 95% confidence intervals (Cl)) Age range No. of subjects Ever used DMPA RR (95% Cl) Description (years) Cases Controls among controls (%) in ever-users New Zealand: population-based 25-54 891 1864 13.5 1.0 (0.8-1.3)a WHO-coordinated: hospital-based <63 869 11 890 12.2 1.21 (0.96-1.52)b a Adjusted for age, parity, ethnic group, and year of interview. b Adjusted for age, centre, and age at first live birth. WHO Bulletin OMS. Vol 71 1993 Memorandum Table 3: Estimated relative risk (RR) with 95% confidence intervals (Cl) for breast cancer in ever-users of DMPA, by age at diagnosis and by age at first use, in two studies on DMPA and breast cancer Study and reference New Zealand (26) WHO-coordinated (27) By age at diagnosis Years RR (95% Cl) 25-34 35-44 45-54 2.0 (1.0-3 8)a.b 0.94 (0.65-1.4) 0.95 (0.63-1.4) By age at first use Years RR (95% Cl) <25 25-29 >30 <25 >25 <35 1.40 (0.88-2.22)c 35-44 1.08 (0.75-1.55) >45 1.01 (0.68-1.51) 1.5 (0.85-2.6)a 0.79 (0.50-1.3) 1.1 (0.75-1.5) 1.32 (0.81-2.17)c 1.21 (0.96-1.53) a Adjusted for age, parity, ethnic group and year of interview. b Test for trend with duration of use: P <0.001. c Adjusted for age, centre, and age at first live birth. cancer risk overall. Elevated risks in the same or similar subgroups were observed in the WHO and New Zealand studies. These patterns of elevation in risk are difficult to interpret. They would be consis- tent with an acceleration in detection of pre-existing cancer, but other interpretations are possible. The data are not compatible with an effect of DMPA as an initiating agent. Our limited understanding of how progestogens act on the human breast does not allow us to interpret the data further. Recommendations. Further epidemiological studies of DMPA and breast cancer should be designed to clarify issues raised in existing studies. In particular, they should seek to distinguish between any effect of recent use and any effect of breast cancer occurring at a young age. Further studies should, however, include women in all age groups that have been exposed. Research is needed on the biological mecha- nisms of action of progestogens on the human breast. Cervical cancer Data from epidemiological studies of cervical neo- plasia in users of combined oral contraceptives sug- gest that such use may increase the risk of intra- epithelial and invasive cervical cancer, although concern about residual confounding makes a defini- tive conclusion impossible at present (28). It is un- certain whether the possible effect of combined oral contraceptives on risk of cervical neoplasia is due to the estrogenic or the progestogenic component, or to both. Therefore, questions have been raised about the possible role of DMPA, a progestogen, in cervical carcinogenesis. Epidemiological studies. There are only two epi- demiological studies that have assessed the relation- ship between DMPA and invasive cervical cancer (29, 30). Both of these were case-control studies and the relevant findings are summarized in Table 5. The relative risk of invasive cervical cancer was above Table 4: Estimated relative risk (RR) with 95% confidence intervals (Cl) for breast cancer in ever-users of DMPA, by time since first and last use, in two studies on DMPA and breast cancer. Study and reference New Zealand (26) WHO-coordinated (27) By time since first use Months <60 60-119 120-179 >180 OR (95% Cl) 1.7 (0.88-3.4)a 1.2 (0.76-1.9) 0.92 (0.64-1.3) 0.73 (0.39-1.4) <48 2.02 (1.35-3.01)b By time since last use Months RR (95% Cl) <60 60-119 >120 1.6 (1.0-2.5)a 0.99 (0.65-1.5) 0.78 (0.53-1.2) Current 1.65 (1.08-2.52)b 49-96 97-156 >157 1.24 (0.81-1.90) 0.90 (0.60-1.36) 1.04 (0.68-1.58) <12 13-36 37-84 1.17 (0.53-2.56) 1.28 (0.74-2.20) 1.09 (0.68-1.75) a Adjusted for age, parity, ethnic group, and year of interview. b Adjusted for age, centre, and age at first live birth. WHO Bulletin OMS. Vol 71 1993672 Depot-medroxyprogesterone acetate and cancer Table 5: Summary of results of two studies of DMPA and invasive cervical cancer (estimated relative risk (RR) with 95% confidence intervals (Cl)) Controls Population-based Hospital-based Number of subjects Cases Controls RR (95% Cl) Ever-used Long-term DMPA DMPA use 133 646 1.4 (0.6-3.1)a 1.2 (0.34.5)aab 2009 9583 1.1 (0.96-1.29)c 1.2 (0.84-1.72)Cd a Adjusted for major identified confounding factors. b Use for one or more years. I Adjusted for age, number of pregnancies, number of prior Pap smears, use of oral contraceptives. d Use for more than 8 years. 1.0 in both studies, but the 95% confidence intervals did not exclude 1.0 in either of them. In neither the Costa Rica study (29) nor the WHO study (30) was there a consistent pattern of risk of invasive cervical cancer with the duration of use or with time since first or last use. A third study (31) assessed the use of injectable contraceptives and the risk of invasive cervical can- cer. Of the injectable users 55% used monthly inject- able(s) and 45% reported use of a 3-monthly inject- able (presumed to be DMPA). The overall estimate of relative risk for women who ever used injectable contraceptives was 0.8. The findings are difficult to interpret because separate analyses for different types of injectable preparations were not presented. The relationship between DMPA and carcinoma in situ of the cervix was studied in Costa Rica (29) and in two small cohort studies (32, 33). In the Costa Rica study (29), there were 369 cases of carcinoma in situ. The relative risk in women who had ever used DMPA was estimated to be 1.1 (95% CI, 0.6-3.1) after adjustment for the main recognized risk factors for cervical neoplasia. In long-term (.2 years) users, the adjusted relative risk estimate was 1.0 (95% CI, 0.3-3.2). There was no consistent pat- tern of risk of carcinoma in situ with duration of use, with time since first or last use, or with age at first use in this study. Because of the small size and other methodological limitations, the two cohort studies are not considered here. Conclusions. Findings from studies of DMPA and invasive cervical cancer are generally reassuring.The results of the studies show no overall increase in risk of cervical cancer. In none of the studies to date was there a consistent pattern of risk with the duration of use or with other time-related factors. These findings do not support an association between DMPA and cervical cancer. Recommendations. Additional studies of invasive cervical cancer and DMPA should be conducted. These studies should include collection of biological specimens to allow the assessment of risk separately for women with and without evidence of human papilloma virus infection. Endometrial cancer Factors that increase exposure of the uterus to unop- posed estrogen, endogenous or exogenous, produce a progression of endometrial changes from benign pro- liferation to atypical hyperplasia to adenocarcinoma (34). In post-menopausal women, the administration of a progestogen with exogenous estrogen mitigates the carcinogenic effect of estrogen on the endo- metrium (35). There is a highly significant trend of decreasing risk of endometrial cancer with increasing duration of use of combined oral contraceptives (36). These observations suggest, on theoretical grounds, that DMPA might decrease the risk of endometrial cancer. Epidemiological studies. Two epidemiological stud- ies with information on DMPA and endometrial can- cer have been carried out (24, 37). The first study (24), a record linkage study of women attending a large family planning clinic in the USA, observed only one case of leiomyosarcoma among 5000 women who received at least one DMPA injection. No endometrial cancers occurred whereas the expected number was 0.83. Because of the small number of cases expected and the possibility that cancers were not ascertained completely in members of the cohort, no meaningful conclusion can be drawn from this study. In a multicentre WHO study (37), only the centres in Thailand provided enough women who used DMPA to contribute to the analysis. There were 122 women with pathologically confirmed endometrial cancer and 939 controls. There were 3 cases and 84 controls who had ever used DMPA. The estimated relative risk of endometrial cancer in ever-users of DMPA was 0.21 (95% CI, 0.06-0.79). All 3 women with endometrial cancer who had used DMPA had also received estrogen premenopausally. The protective WHO Bulletin OMS. Vol 71 1993 Study location and reference Costa Rica (29) Thailand, Kenya, Mexico (30) 673 Memorandum effect of DMPA for endometrial cancer appeared to be at least as strong as for use of combined oral con- traceptives. Conclusions. Taken together with other information about the effects of progestogens on the endometri- um, data from the WHO study provide evidence that DMPA protects against endometrial cancer. Recommendations. Studies should be undertaken to determine if the protective effect of DMPA, observed in Thailand (37), occurs in other popula- tions. Studies should be done to determine whether past use of DMPA continues to exert a protective effect on endometrial cancer risk in older women, taking into account the influence of menopausal hor- mone therapy. If estrogen is used with DMPA, the effect of this practice on risk of endometrial cancer should be evaluated. Ovarian cancer A number of lines of evidence implicate risk of ova- rian cancer with ovulatory activity. In general, fac- tors that are associated wih reduced ovulation are associated with a lower risk of epithelial ovarian cancer (38). Taking oral contraceptives substantially reduces the risk of epithelial ovarian cancer. Five or more years of use is associated with a 50% reduction in risk (39). Since DMPA inhibits ovulation, it might also be expected to reduce the risk of ovarian cancer. Epidemiological studies. Two epidemiological stud- ies have assessed the risk of ovarian cancer in DMPA users. The first study of women attending a family planning clinic described earlier (24) observed 1 case of ovarian cancer in women who had at least one DMPA injection, whereas 1.16 cases were expected. The small number of cases of ovarian cancer and concern that cases of cancer may not have been completely ascertained limit the conclu- sions based on this study. The question of DMPA and ovarian cancer was assessed in a WHO study (40). A total of 224 women with histologically confirmed epithelial ovar- ian cancer were compared with 1781 hospital con- trols matched by age, hospital, and year of interview. The estimated relative risk of ovarian cancer in women who had ever used DMPA, adjusted for the number of live births and oral contraceptive use, was 1.07 (95% CI, 0.6-1.8). Conclusions. Data from the WHO study found that DMPA use was not associated with either an increased or decreased risk of ovarian cancer. The low power of the study to detect changes in risk in long-term DMPA users is a limitation. Given the public health importance of ovarian cancer in many countries, more data are required in order to clarify this issue. Recommendations. Studies are needed to determine whether the lack of association between DMPA use and epithelial ovarian cancer risk seen in Thailand is confirmed. Studies are needed to determine the bio- logical mechanisms of epithelial ovarian carcinogen- esis, with particular attention to the ovulation inhibi- tion hypothesis. Overall conclusions The Group concluded that there was no evidence for an overall increase in risk of cancer in users of DMPA at any of the four sites reviewed (breast, cer- vix, endometrium and ovary). Given this evidence, the Group does not recommend restriction of DMPA use as a contraceptive on the grounds of risk of neo- plasia. Although an increased risk of breast cancer has been observed in certain subgroups of women using DMPA, these findings are difficult to interpret. It is unlikely that these observations represent new tumours caused by DMPA. There is good evidence for protection against endometrial cancer. Much of the epidemiological information regarding DMPA and cancer comes from a single, well-conducted study, and conclusions drawn from epidemiological observations are more secure when based on a more extensive body of evidence. Addi- tional research is needed to confirm or refute exist- ing findings, as well as to address additional issues. In particular, the Group noted that more data were needed on long-term use by young women and that the effect of estrogen use in combination with DMPA should be evaluated. In assessing the benefits and risks of DMPA it would be a failure of perspective to consider cancer in isolation. In deciding whether to choose this meth- od of contraception, women and family-planning providers need to consider all the benefits and risks. References 1. Williams, D.L. et al. Chromosome alterations pro- duced in germ cells of dogs by progesterone. Jour- nal of laboratory and clinical medicine, 77: 417- 429 (1971). 2. Williams, D.L. et al. Progesterone-induced altera- tions of oogenesis in the Chinese hamster. Journal of laboratory and clinical medicine, 79: 972-977 (1972). 3. Garcia Heras, J. & Coco, R. Induction of sister chromatid exchanges in rabbit lymphocytes by me- droxyprogesterone. Medicina, 42: 250-254 (1982). 4. Garcia Heras, J. et al. Induction of sister chromatid 674 WHO Bulletin OMS. Vol 71 1993 Depot-medroxyprogesterone acetate and cancer exchanges in mouse kidney fibroblasts by medroxy- progesterone acetate. Medicina, 42: 255-258 (1982). 5. Jordan, A. Toxicology of depot-medroxyprogester- one acetate. Contraception (submitted) (1993). 6. Lanari, C. et al. Induction of mammary adenocar- cinomas by medroxyprogesterone acetate in BALB/c female mice. Cancer letter, 33: 215-223 (1986). 7. Lanari, C. et al. [Induction of mammary adenocarci- nomas by medroxyprogesterone in females C3h rats]. Medicina, 47: 652-653 (abs. no. 339) (1987) (in Spanish). 8. Nagasawa, H. et al. Medroxyprogesterone acetate enhances spontaneous mammary tumorigenesis and uterine adenomyosis in mice. Breast cancer research and treatment, 12: 59-66 (1988). 9. Zaccheo, T. et al. Combined and sequential treat- ment using FCE 21336, a new prolactin-lowering drug, and medroxyprogesterone acetate (mpa) in DMBA-induced tumors in rats. European journal of cancer and clinical oncology, 20: 1193-1197 (1984). 10. Dauvois, S. et al. Additive inhibitory effects of bromocryptine (CB-154) and medroxyprogesterone acetate (mpa) on dimethylbenz[a]anthracene (DMBA)-induced mammary tumours in the rat. Euro- pean journal of cancer and clinical oncology, 25: 891-897 (1989). 11. Russo, I.H. et al. Effect of medroxyprogesterone acetate on the response of the rat mammary gland to carcinogenesis. British journal of cancer, 59: 210-216 (1989). 12. Hussain, S.P. & Rao, A.R. Modulatory influence of injectable contraceptive steroid medroxyprogester- one acetate on methylcholanthrene-induced carcino- genesis in the uterine cervix of the mouse. Cancer letter, 61: 187-193 (1991). 13. Pazos, P. et al. Mammary carcinogenesis induced by methyl-N-nitrosourea (MNU) and medroxyproges- terone acetate (MPA) in BALB/c mice. Breast can- cer research and treatment, 20: 133-138 (1991). 14. Braunsberg, H. et al. Hormonal therapies for breast cancer: can progestogens stimulate growth? Cancer letter, 30: 213-218 (1986). 15. Larsson, K.S. & Machin, D. Predictability of the safety of hormonal contraceptives from canine toxi- cological studies. In: Michal, F., ed. Safety require- ments for contraceptive steroids. Cambridge, Cam- bridge University Press, 1989, pp. 230-269. 16. Coleman, M.E. et al. Acromegalic changes in bea- gles after long-term progestogen treatment. Toxicol- ogy and applied pharmacology, 37: 181-182 (1976). 17. El Etreby, M.F. et al. Evaluation of effects of sexual steroids on the hypothalamic-pituitary system of ani- mals and man. Archives of toxicology, Suppl. 2: 11-39 (1979). 18. El Etreby, M.F. et al. Comparative endocrine phar- macodynamics of contraceptive steroids. In: Michal, F., ed. Safety requirements for contraceptive ster- oids, Cambridge, Cambridge University Press, 1989, pp. 179-192. 19. Concannon, P.W. et al. Gross and histopathologic effects of medroxyprogesterone acetate and proges- terone on the mammary glands of adult beagle bitches. Fertility and sterility, 36: 373-387 (1981). 20. Spellacy, W.N. et al. The effects of medroxypro- gesterone acetate on carbohydrate metabolism: measurements of glucose, insulin, and growth hor- mone after twelve months' use. Fertility and sterility, 23: 239-244 (1972). 21. Jordan, A. FDA requirements for non-clinical testing of contraceptive steroids. Contraception, 46: 499- 509 (1992). 22. Greenspan, A.R. et al. The association of depot- medroxyprogesterone acetate and breast cancer. Contraception, 21: 563-569 (1980). 23. Danielson, D.A. et al. Nonestrogenic drugs and breast cancer. American journal of epidemiology, 115: 329-332 (1982). 24. Liang, A.P. et al. Risk of breast, uterine corpus and ovarian cancer in women receiving medroxyproges- terone injections. Journal of the American Medical Association, 249: 2909-2912 (1983). 25. Lee, N.C. et al. A case-control study of breast can- cer and hormonal contraception in Costa Rica. Jour- nal of the National Cancer Institute, 79: 1247-1254 (1987). 26. Paul, C. et al. Depot medroxyprogesterone (Depo- Provera) and risk of breast cancer. British medical journal, 299: 759-762 (1989). 27. WHO Collaborative Study of Neoplasia and Ster- oid Contraceptives. Breast cancer and depot- medroxyprogesterone acetate: a multinational study. Lancet, 338: 833-838 (1991). 28. La Vecchia, C. Depot-medroxyprogesterone acet- ate, other injectable contraceptives and cervical neoplasia. Contraception (submitted) (1993). 29. Oberle, M.W. et al. Cervical cancer risk and use of depot-medroxyprogesterone acetate in Costa Rica. International journal of epidemiology, 17: 718-723 (1988). 30. WHO Collaborative Study of Neoplasia and Ster- oid Contraceptives. Depot-medroxyprogesterone acetate (DMPA) and risk of invasive squamous cell cervical cancer. Contraception, 45: 299-312 (1992). 31. Herrero, R. et al. Injectable contraceptives and risk of invasive cervical cancer: evidence of an associa- tion. International journal of cancer, 46: 5-7 (1990). 32. Powell, L.C. Jr. & Seymour, R.G. Effect of depot- medroxyprogesterone acetate as a contraceptive agent. American journal of obstetrics and gynecolo- gy, 110: 36-41 (1971). 33. Dabancens, A. et al. Intraepithelial cervical neopla- sia in women using intrauterine devices and long- acting injectable progestogens as contraceptives. American journal of obstetrics and gynecology, 119: 1052-1056 (1974). 34. Gusberg, S.B. Hormonal dependence of endometri- al carcinoma. Obstetrics and gynecology, 30: 287-292 (1967). 35. Persson, I. et al. Risk of endometrial cancer after treatment with estrogens alone or in conjunction with progestogens: results of a prospective study. British medical journal, 298: 147-151 (1989). 36. Schlesselman, J.J. Oral contraceptives and neo- plasia of the uterine corpus. Contraception, 43: 557-579 (1991). 37. WHO Collaborative Study of Neoplasia and Ster- oid Contraceptives. Depot-medroxyprogesterone WHO Bulletin OMS. Vol 71 1993 675 Memorandum acetate (DMPA) and risk of endometrial cancer. International journal of cancer, 49: 186-190 (1991). 38. Brinton, L.A. & Hoover, R.N. Epidemiology of gynecologic cancers. In: Hoskins, W.J. et al., ed. Principles and practice of gynecologic oncology. Philadelphia, Lippincott, 1982, pp. 1-12. 39. Stanford, J.L. Oral contraceptives and neoplasia of the ovary. Contraception, 43: 543-555 (1991). 40. WHO Collaborative Study of Neoplasia and Ster- oid Contraceptives. Depot-medroxyprogesterone acetate (DMPA) and risk of epithelial ovarian cancer. International journal of cancer, 49: 191-195 (1991). 676 WHO Bulletin OMS. Vol 71 1993