Considerable controversy has surrounded the effects of fertility drug use on cancer risk, primarily as related to ovarian cancer. In this issue of the Journal, Fei et al. (1) presented unique data from a case–control study of sister pairs (one of whom had breast cancer) that allowed evaluation of associations between fertility drug use and young-onset (<50 years at diagnosis) breast cancer. Although a reduced risk associated with drug usage was noted overall, this association was not observed among women who conceived as a result of treatment and who subsequently carried their pregnancies for at least 10 weeks.
These results (1) are difficult to interpret in the context of previous studies, given that findings have ranged from reduced risks to increased risks to the absence of any relationship (2). Several biological explanations are possible for these associations. On the one hand, reduced risks are possible given that clomiphene is a selective estrogen receptor modulator (SERM) with structural similarities to tamoxifen, a well-established chemopreventive agent for breast cancer. On the other hand, increased risks could reflect that fertility drugs induce ovulation and raise estrogen levels (3). For ovarian cancer, some studies have shown elevated risks primarily among patients who are unable to conceive and remain nulliparous (4–6). This has been interpreted as indicative of these patients having either resistant types of infertility linked with higher cancer risks, genetic predispositions to both infertility and cancer, or higher drug exposures.
In contrast, in the current study (1), stratified analyses focused on pregnancies that were conceived after ovulation induction and continued at least 10 weeks. The risk among these patients was compared to a referent group of women receiving ovulation induction and who did not conceive or had short-term pregnancy losses, possibly presenting interpretative complexities. It was suggested that the higher risk of breast cancer among women with successful conceptions was indicative of ovulation-induction causing abnormally high exposures to ovarian hormones during early pregnancy that could affect subsequent breast remodeling. However, there could be several alternative explanations, including patients who successfully conceived being those who received higher drug exposures. In addition, the increased risk among women with pregnancies after ovarian induction could reflect the well-recognized dual effect of pregnancy on breast cancer risk, namely a short-term transient increase that dissipates with time and eventually leads to a long-term risk reduction (7). Further insights might have derived from analyses related to the intervals since these pregnancies, as well as on whether the pregnancies lasted full-term, because the impact of pregnancies on breast cancer risk appears dependent on their being carried to term (8).
Additional difficulties in interpreting fertility drug effects relate to the complex manner in which these drugs are prescribed. Accurately accounting for previous drug usage is complicated when patients are asked to recall such information—as recognized with respect to early findings related to ovarian cancer (9). The fact that breast cancer patients in the current investigation (1) were not interviewed until at least a year after being diagnosed may have led to recall problems regarding the drugs prescribed, especially among unsuccessfully treated women who tend to be averse about recalling their experiences (10).
The authors evaluated risks related to two main types of fertility drugs—clomiphene (a SERM) and gonadotropins (a wide class of glycoprotein hormones, including follicle-stimulating hormone [FSH]). Although gonadotropins are most often used in conjunction with in vitro fertilization (IVF), many women undergoing ovulation induction also are given clomiphene—either as a less expensive first-line treatment approach or in combination with gonadotropins to lower their effective dose. Thus, it was not surprising in the study that few women had been exposed only to FSH. Because of small numbers of women receiving FSH, the approach of examining relationships according to use of any fertility drug may have obscured unique effects of these two different types of drugs. Women who receive IVF are also given gonadotropin-releasing hormone agonists or antagonists as well as progestogens (a class of steroid hormones including progesterone), which could also impact cancer risks. It would, therefore, have been of interest to evaluate risk according to IVF treatment rather than merely assessing risks related to the two evaluated drugs.
As discussed by the investigators (1), a major limitation that has affected most studies evaluating the associations of fertility drugs with cancer risk is the inability to accurately control for the precise causes of infertility—which can be associated with cancer risks independent of fertility drug use (2). Such information is extremely difficult to obtain, especially from patients who often do not understand the distinctions between different causes of infertility. Further complicating interpretations of associations are that distinctive causes can lead to the types of drugs prescribed (eg, women with anovulatory problems being preferentially prescribed clomiphene), as aptly described in the article.
Although this study will undoubtedly stimulate additional controversy regarding the associations of fertility drugs with breast cancer risk, it must be emphasized that the investigation focused on women who developed breast cancer before the age of 50 years. Such cancers often show distinctive risk predictors, including being more strongly associated with genetic factors than older-onset breast cancers. However, because the study design involved sisters, it was not possible to assess whether there could be unique gene–environment interactions. Several previous studies have hypothesized that fertility drugs may have unusual relationships with cancer among women with either a family history of cancer (11) or BRCA mutations (12), but further studies are needed to fully understand potential interactions. Breast cancers occurring among young women (age <50 years) also often have unique tumor markers (13) beyond the estrogen or progesterone markers examined in the current study (1).
Because of the complexities in assessing the relationships of fertility drugs to cancer risk, it is clear that additional research will be needed to fully understand the associations. Case–control studies can generate hypotheses, but further resolution of relationships is ultimately dependent on cohort studies that obtain detailed information on drug exposures and clinical indications for usage before the disease onset. These studies are complicated by diverse and changing treatments, which could have unique effects. Because of such complexities, results from individual investigations must be cautiously interpreted and weighed against the considerable benefits associated with fertility drug usage, including a high probability of carrying pregnancies to term, which can lead to substantial long-term reductions in breast cancer risks.
Supported by the Intramural Research Program of the National Cancer Institute, National Institutes of Health.
The author declares no conflicts of interest.