Risk of Breast Cancer with Progestins in Combination with Estrogen as Hormone Replacement Therapy

Two recent studies have suggested that progestins substantially increase the relative risk (RR) of breast cancer when added to estrogens as hormone replacement therapy (HRT) (1, 2) (see Figs. 1 and 22). If correct, this information could substantially change clinical practice. Here, we review biological, epidemiological, and clinical data regarding the effects of progestins on the breast. From this analysis, we conclude that no definitive proof exists to establish a causal relationship between progestins and breast cancer risk. However, a wide range of biological and clinical data provides strong supportive evidence of such an effect. Based on this, we believe that it is prudent to inform patients that progestin use may add to the increased risk imparted by estrogens. Patients should understand, however, that this increased risk is small, particularly when associated with the short-term use of HRT. Estimated increase in RR of breast cancer as a function of taking estrogen alone as HRT. The shaded area represents the confidence limits of the percent increase in risk per year. Redrawn from Ref. 1 . Estimated increase in RR of breast cancer as a function of taking estrogen plus a progestin. The shaded area represents the confidence limits of the percent increase in risk per year. Redrawn from Ref. 1 . Before examining the effects of progestins, it is first necessary to question whether estrogen alone increases the risk of breast cancer. Substantial data from animal and human studies provide support for a link between estrogen use and breast cancer risk. Administration of exogenous estrogen to rodents results in a high incidence of breast cancer (3). The use of antiestrogens or blockers of estrogen biosynthesis (aromatase inhibitors) abrogates the development of breast tumors that occur spontaneously or are induced by carcinogens in rats (4, 5). In women, early menarche, late menopause, and increased endogenous circulating estrogen levels increase the RR of developing breast cancer (3, 6, 7). Removal of both ovaries before age 35 lowers the risk of breast cancer by 75% over a 25-yr period of observation (8, 9). Finally, antiestrogens such as tamoxifen and raloxifene reduce the incidence of newly diagnosed breast cancer as demonstrated by randomized, placebo-controlled trials in women (10, 11). More than 50 observational studies in patients have examined whether estrogens cause an increased risk of breast cancer. Individual studies report an increase, decrease, or no change in the risk of breast cancer in menopausal women taking estrogen replacement therapy (ERT) (6). A recent meta-analysis from the Collaborative Group on Hormonal Factors in Breast Cancer (CGHFBC) (6) identified several objective factors that potentially explain differing conclusions among these studies. We draw the following conclusions from this meta-analysis. First, the RR of breast cancer from ERT is small and very large numbers of women must be studied to minimize type I and type II statistical errors.1 Second, the risk of breast cancer seems to increase linearly with duration of use. Consequently, studies comparing “ever users” of estrogen with “never users” have limited validity because they do not consider duration of estrogen use. Third, the increased risk of breast cancer imparted by estrogens seems to dissipate within 4 yr of cessation of therapy. Accordingly, only women using estrogen within 4 yr of study might be found to be at increased risk. Fourth, breast cancer risk seems to diminish over the 4-yr period following the menopause, presumably as a reflection of decreased estrogen levels. As a result, analyses of observational studies need to match users vs. nonusers as to time following menopause. Finally, the increased risk of breast cancer seems to be limited to nonobese women[ i.e. body mass index (BMI), <25 kg/m2]. Inclusion of a large proportion of obese women might obscure the association between estrogen use and breast cancer risk. The CGHFBC meta-analysis (6) was sufficiently large (i.e. 52,705 women with breast cancer and 108,411 without) to take each of these five factors into account. The key finding was a linear 2.3% increase in the RR of breast cancer for each year of HRT use for up to 25 yr. Both the slope of this linear increase in risk and the overall risk of breast cancer among HRT users was found to be highly statistically significant. In the authors’ opinion, the CGHFBC meta-analysis provides substantial evidence that ERT increases the risk of breast cancer. However, the inferences from this study must be considered provisional because they are based on observational data and are subject to various biases. A general theory of carcinogenesis holds that agents that increase the rate of cell proliferation can enhance the development of new genetic mutations (12). Mutations are thought to be necessary for the process of initiation of cancer. Once mutations are present, they need to be propagated by cell replication, a process considered to be responsible for tumor promotion. Estrogens are known to enhance the rate of cell proliferation in glandular tissue of the breast and, thus, could potentially act both in the initiation and promotion of breast cancer. A key issue is whether progestins exert proliferative or antiproliferative effects on the human breast. Progestins oppose the proliferative effects of estrogens on the human endometrium and reduce the risk of endometrial cancer. Gambrell (13) has hypothesized that progestins might abrogate the carcinogenic effects of estrogen on the breast through a similar antiproliferative action. Others argue that progestins exert proliferative and, thus, procarcinogenic effects on the breast (14). This controversy has stimulated a wide range of in vitro and in vivo studies to delineate the effects of progestins on breast tissue. The resulting reports highlight various complexities underlying the effects of progestins on breast tissue. It is important to understand that not all progestins are alike in structure and function. Progestins can be classified into two major subtypes, the 17α-acetoxyprogesterone and the nortestosterone derivatives (15). The 17α-acetoxyprogesterone derivatives, such as medroxyprogesterone acetate (MPA), possess glucocorticoid-like as well as progestational activity. The group of 19-nortestosterone compounds includes two subclasses, the estranes and gonanes. The estranes, such as norethindrone acetate and ethynodiol diacetate, are more androgenic, and the gestanes, such as gestodene and desogestrel, are more progestational. Depending on their structure and the tissues in which they are studied, the various progestins can exert either androgenic, synandrogenic, antiandrogenic, estrogenic, glucocorticoid-like, or progestational effects (16–28). These disparate actions of progestins on human breast cells in culture have confounded interpretation regarding effects on proliferation. For example, various studies report that human breast cancer cell lines such as MCF-7, T47-D, and ZR-75–1 can be either stimulated or inhibited by progestins through their androgenic, estrogenic, or progestational effects (17–28). Normal human breast cells obtained at reduction mammoplasty and grown in primary culture also respond to various progestins with either proliferative or antiproliferative responses. A recent series of studies by Horwitz and colleagues (29–31) highlight the complexity of mechanisms whereby progestins regulate the proliferative process. They demonstrated that progestins act to up-regulate growth factor and cytokine receptors and interact with key downstream cell cycle mediators such as cyclin D. Substantial cross-talk between progesterone receptors and growth factor-related pathways occurs. Progestins increase the level of epidermal growth factor receptors, activate the transcription factor stat 5, and result in stimulation of several factors involved in regulating the proliferative process such as mitogen-activated protein kinase, p38 kinase, and c-jun-NH2-kinase. Whereas much is now known about the in vitro effects of progestins from these studies, critical evaluation of these data do not establish whether the predominant effect of progestins is to stimulate or inhibit breast cell proliferation. A clearer understanding that the predominant effects of progestins on breast are to induce proliferation has emerged from patient studies. Anderson and colleagues (14, 32) examined breast biopsies taken from women during the follicular phase when estradiol is the predominant circulating hormone and again during the luteal phase when progesterone increases. They found a substantial increase in tritiated thymidine uptake in association with luteal phase progesterone increments and with use of progestin containing oral contraceptives. These observations were confirmed by an additional study using fine-needle aspiration and markers of cell proliferation (Ki67 or MIB 1) to assess differences between follicular and luteal phase proliferation (33). Some doubt persisted, however, as a result of findings from topical administration of progestins that reduced breast epithelial cell proliferation. However, the amounts of topical progestin used were sufficient to increase tissue levels to pharmacological levels and, thus, may not reflect normal physiology (34). More compelling data regarding the proliferative effects of progestins resulted from histological studies of breast tissue in postmenopausal women receiving either estrogen alone, estrogen plus a progestin, or no HRT for varying periods of time up to 10 yr. Hofseth et al. (35) examined breast tissue from women undergoing excisional biopsy for mammographic lesions. Tissue for assessment was taken from areas distant from the focal lesion. These investigators assessed proliferation by proliferating cell nuclear antigen (PCNA) and Ki67 measurements and quantitated the percent area of breast occupied by glandular tissue with computer-assisted morphometry. The results demonstrated that long-term estrogen use increased the rate of cell proliferation, the number of cells present in terminal ductal lobular units, and the percentage of breast tissue made up of glandular tissue as opposed to adipose and stromal tissue (Fig. 3). Notably, the addition of a progestin to estrogen replacement enhanced the rate of cell proliferation, terminal duct lobular units (TDLUs), and glandular mass. These effects of progestins appeared to increase linearly with time. Top, The percentage of cells that stain positively for the proliferation marker PCNA in ductal tissue and in TDLUs. The number under each bar represents the number of individuals from whom ducts or TDLUs could be analyzed. *, P< 0.002 that the percentages of PCNA-positive cells in the TDLUs of the E+P group were significantly greater than in the TDLUs or ducts of the no HRT group or E alone group; +, P < 0.007 that the percentages of PCNA-positive cells in the TDLUs or ducts of the E group or ducts of the E+P group were significantly greater than in the TDLUs or ducts of the no HRT group; ±, P < 0.005 that the percentages of PCNA-positive cells in the TDLU of the luteal phase group were significantly greater than in the TDLUs of the follicular phase group; §, P < 0.05 that the percentages of PCNA-positive cells were greater in TDLUs than in the ducts of the same group. E, Estrogen; P, progestin; L, luteal phase; F, follicular phase. Middle, The percentage of cells that stain positively for the proliferation marker Ki67 in ductal tissue and in TDLUs. The number under each bar represents the number of individuals from whom ducts or TDLUs could be analyzed. *, P < 0.002 that the percentages of Ki67-positive cells in the TDLUs of the E+P group were significantly greater than in the TDLUs or ducts of the no HRT group or E alone group; +, P < 0.007 that the percentages of Ki67-positive cells in the TDLUs or ducts of the E group or ducts of the E+P group were significantly greater than in the TDLUs or ducts of the no HRT group; ±, P < 0.05 that the percentages of Ki67-positive cells in the TDLUs of the luteal phase group were significantly greater than in the TDLUs of the follicular phase group; §, P < 0.05 that the percentages of PCNA-positive cells were greater in TDLUs than in the ducts of the same group. Bottom, Effects of HRT on breast epithelial density in postmenopausal women. The number under each bar represents the number of individuals for whom epithelial density was determined. *, P < 0.001 that the percentages of epithelial area in the E or E+P groups were significantly greater than that of the no HRT group; + P < 0.02 that the percentage of epithelial area in the E+P group was significantly greater than that of the E alone group. Further evidence of the proliferative effect of progestins derives from quantitative studies of mammographic density in women receiving HRT (36–40). Glandular tissue enhances the density of mammograms, and adipose tissue reduces it. Thus, breast density can serve as a surrogate marker for long-term glandular cell proliferation. The Postmenopausal Estrogen/Progestin Interventions trial analyzed mammographic density (36) in 307 eligible candidates out of a total of 875 women in the a a at or for with the and no use of estrogen for yr before the the percentage of women with density increases was confidence in the group; in the estrogens alone group; in the estrogens plus group; in the estrogens plus group; and with the estrogens plus progesterone group. density may be a marker for increased risk for breast cancer. If the when a progestin is added to estrogen therapy may be These results are with studies and provide compelling evidence of the as opposed to the effects of the progestins on human breast tissue in The two recent studies, within 1 of each in the of (1, that progestins add to the risk of breast cancer to these studies are observational with a for they are and by not studies. the validity of the conclusions from these studies, we have to review data from studies very We believe that of the of findings among these key studies provides the of conclusions from We do not believe that a of in each study serve to these In opinion, the from the CGHFBC meta-analysis of for study validity (6). First, the study must be large to among various Second, women must have used HRT within at 4 yr before assessment of breast cancer risk. Third, the duration of HRT use must be taken into account. For these we have to on studies the following 1) of at women with breast 2) of data on women receiving HRT within 4 yr of breast cancer risk of risk duration (i.e. at 4 of HRT and of estrogen alone with the of estrogen plus a progestin. studies these of data from five key studies of estrogens alone vs. E plus a progestin and only for of data from five key studies of estrogens alone vs. E plus a progestin and only for The study in the of the in by et al. represents a study of postmenopausal women in a mammographic The investigators report that the RR of breast cancer from estrogen alone was and that the RR from estrogen plus a progestin was the RR increased by per year of estrogen use alone and per year of use of estrogen plus a progestin 1 and 22). increase in risk was in women with a greater than However, in women with a to or than the increase in risk was with estrogen alone and with estrogen plus a progestin. The study by et al. used the and HRT use in a group of postmenopausal women with diagnosed breast cancer and The risk of breast cancer with ERT alone was only increased for women taking this for yr or more no estrogen plus a progestin, the 10 yr was as a increase in ERT was associated with a per year estrogen in with a progestin was associated with a per year These were as per yr with confidence limits alone with with The risk of progestin use to be was not significantly than that associated with progestin use. 10 the for the was vs. for the The only in involved and yr of In women breast cancer and information regarding use of breast cancer risk increased with estrogen use alone by and with estrogen plus a progestin by per year A study on postmenopausal women with breast cancer. 10 yr of the for breast cancer risk associated with estrogen alone was that associated with estrogen plus a progestin was The risk was with estrogens alone and for estrogen plus a progestin. This study also with obese women. The of HRT taken for more than 10 yr estrogen alone or estrogen plus a were not increased for with a more than as opposed to with of and of than The study the results of which with of the that the RR associated with estrogen use alone was that associated with estrogen plus a progestin was studies not the are with an effect of et al. two studies from within the yr. a RR of for an for 10 yr vs. a RR of for estrogen alone The a RR of for the vs. a RR of for estrogen alone study a RR of for estrogen plus a progestin vs. for estrogen alone In additional studies large numbers of women with breast cancer and women no increased risk of progestins to estrogens. studies women with breast cancer to (i.e. breast cancer the CGHFBC meta-analysis information regarding estrogen plus progestin and no conclusions were regarding the added effects of progestins on breast cancer risk (6). data that progestins may exert various actions in addition to their progestational Accordingly, the various progestins used by patients could exert actions on their data be to whether are differences in breast cancer risk associated with the use of these The study of et al. that this might be the They a greater risk of breast cancer in association with the 19-nortestosterone derivatives as opposed to the The of progestin administration may also breast cancer risk. clinical data to use of progestins, clinical now use of A in the study by et al. that the 10 yr of a RR of the RR for the was Finally, the reduction of endometrial cancer with progestins may be by the increased risk of breast cancer. analyses need to the regarding long-term progestin use to endometrial cancer. Notably, should progestins be in HRT The increased risk of breast cancer from progestins may effects to endometrial cancer. data examining the risk of breast cancer from HRT report to statistical (6). This provides substantial statistical to the effects of these agents that might be small in As in a the and the relative with understanding of the of these as is important to the of RR represents the of the risk of breast cancer in women taking HRT to not taking HRT. The not take into the of breast cancer in the group risk is by the rate of breast cancer in the group considered by the For example, women have an risk of developing breast cancer of per women over a A increase in RR from estrogens alone increase the of a breast cancer over a period to per women. risk to the number of women a breast cancer that not have use of estrogen the the between breast cancer risk of per and per represents the increased risk to or per women. in 1 in women taking ERT over 10 yr a breast cancer that not have ERT were not used and from HRT and from HRT of risk data regarding the incidence of breast cancer in the under and the increase in RR with duration of hormone use. can use the data of et al. 1 and to risk. In women, use of estrogen for only yr increases the RR of breast cancer by per year over to and in women have a new breast cancer diagnosed over a a increase in the patient taking ERT have a in of a breast cancer. The risk to HRT is per women. in 1 in women a breast cancer as a result of taking for an an risk of 1 in women. This small increase in risk the RR is increased by per year over or in The use of HRT over a period increases the risk The incidence of breast cancer increases over this time and the risk increases linearly per year. For estrogen alone, the risk is increased by per year or over 10 yr. The rate of breast cancer in a is 10 per women at 10 yr. A increase the rate of breast cancer per over 10 yr. The risk to estrogen thus, 1 in the of estrogen plus a progestin, the RR increases by per year or The risk for this group is now 1 in use of for menopausal of ERT or HRT for than yr only a increase in risk of breast cancer in a in risk from estrogen alone and 1 in for estrogen plus a Consequently, a could be to take short-term ERT or HRT for menopausal a of regarding risk of breast cancer. use of to or The a the greater is risk of developing breast cancer to this The risk associated with estrogen alone, when taken by the for 10 is a 1 in increase in the of a breast cancer. For women, the risk is 1 in If we the data of et al. as the RR of breast cancer increases by per year with use of an estrogen plus a progestin. these we may for the taking HRT for a the breast cancer risk to therapy be 1 in For the the risk increases to 1 in In the this risk of developing breast cancer the of a If the data HRT a new in only 1 of women taking this for 10 yr. In the women taking HRT for 10 1 in have a However, this must be in of recent information from the and Estrogen/Progestin the and and trials that whether estrogen results in primary the and of women therapy because the risk of breast cancer is small in For example, based on the analysis, a women taking an as HRT for 10 yr has only a of breast cancer. risk be These more as the number of women of breast cancer. For example, women taking HRT for 10 yr have a of of breast cancer vs. of not taking HRT. The two recent studies of HRT and breast cancer risk that only patients (i.e. of in study in the an increased risk of breast cancer from either estrogen alone or the of estrogen and a progestin (1, to women, only taking HRT (i.e. greater than a statistically increase in breast cancer risk. women not have an increased risk of breast cancer to HRT. A of breast early age of menarche, late age of a increased breast density on mammograms, and breast increase the underlying risk of developing a breast cancer. Finally, several studies that the breast that in women receiving HRT are in type and are associated with a A number of that can be used in of estrogen to to estrogen in women about breast cancer risk estrogen can be used to the of estrogen levels to a The of can of data from a randomized, trial that the also cause 75% of trials demonstrated that is more than in For of density and of and the raloxifene estrogen and can be In postmenopausal patients at high risk for breast tamoxifen may be used both for of breast cancer as well as of A series of recent report the finding that the increase and reduce risk studies now be to these observational studies. For of the are to be These can be in of estrogens alone or in patients at high risk of breast cancer or of taking HRT. and E are studied as for the of or for which is not definitive evidence that HRT may the of epidemiological, and clinical data support the that progestins enhance cell proliferation of breast tissue inhibit cell in the Based on this it is to that progestins may increase the risk of breast cancer over and that resulting from estrogens The risk from estrogen plus a progestin is for short-term use may be substantial in the of long-term data that the of breast cancer associated with HRT to not obese women. We that is much to and the is definitive data from randomized, trials are it is prudent to present the to patients and inform of their level of risk from estrogens with or a progestin. Based on this short-term use of HRT is associated with risk the and of long-term use more and and of and