Use of fertility medications and cancer risk: a review and update.

With approximately 9% of the population experiencing infertility, an increasing number of couples are seeking fertility treatment. Use of in vitro fertilization (IVF) has steadily risen, contributing to 1.5% of live births in the United States, and adding to those pregnancies resulting from ovulation induction and superovulation [ 1 ]. Consequently, concerns exist regarding long-term risks of fertility medications causing cancer, since despite their multiple modes of action, all can induce multiple ovulation and alter steroidogenesis [ 2 ]. Moreover, some cancers, including female reproductive and breast cancers, are hormone-dependent, offering a physiological mechanism behind an association between fertility medication use and cancer risk [ 3 ]. Nevertheless, although some epidemiologic studies have also suggested such a link, most others offer conflicting results often difficult to interpret. Studying cancer risk in infertile couples poses several challenges. Several conditions, including anovulation, endometriosis and nulliparity accompany infertility and themselves increase the risk of cancer. Additionally, most of these cancers are rare and tend to occur several years after the time of fertility treatment, making proof of causal relationship difficult. Most studies are case-control and cohort studies due to rare disease outcomes, but introduce selection and recall biases, respectively [ 4 ]. In particular older studies are limited by short-term follow-up, confounding factors, improper controls and multiple causes and treatments of infertility. Although limitations exist, recent studies have improved knowledge through better study design, including using long-term follow-up, identifying type of infertility and method of treatment, selecting appropriate controls and assessing specific individuals at high risk for disease. In this review, literature on fertility medication use and cancer risk is discussed with an emphasis on recent publications.

Thyroid disorders including cancer are more common in women than men, with a peak incidence during the reproductive years [ 41 ]. Furthermore, the presence of estrogen receptors in thyroid cancers [ 42 ] implicates sex steroid action with thyroid cancer. With the rising incidence of thyroid cancer in the United States, perhaps from improved surveillance and diagnostic testing [ 43 ], several studies have investigated the relationship between fertility medication use and thyroid cancer. Most studies have found no such association [ 44 ] [ 45 ]. In a retrospective cohort of 8,422 women evaluated for infertility, neither clomiphene citrate (RR: 1.42, 95% CI 0.5–3.7) nor gonadotropin (RR: 1.1; 95% CI 0.2–4.9) use demonstrated an increased risk of thyroid cancer (18 cases) (RR: 1.42; 95% CI 0.5–3.7) [ 46 ]. It is possible that clomiphene citrate use may have had a stronger effect on thyroid cancer risk among women who remain nulliparous (RR: 4.23; 95% CI 1.0–17.1), although 6 of 18 thyroid cancer cases in this study were missing parity data. Another recent retrospective cohort study [ 47 ] showed a non-significant increase in thyroid cancer risk with the use of clomiphene citrate (HR: 1.57; 95% CI 0.89–2.75), based upon 55 cases of thyroid cancer in a cohort of 9,892 women. Thyroid cancer risk was greatest among those women (HR: 1.96; 95% CI 0.92–4.17) who received >2,250 mg of clomiphene citrate, although gonadotropin administration was unrelated to increased risk of any of the assessed cancers (HR: 1.16; 95% CI 0.52–2.58). A Danish cohort of 54,362 women with infertility, however, did show a significant association between fertility medication use and thyroid cancer [ 48 ], in which clomiphene citrate use increased the risk of thyroid cancer, based upon 29 cases (RR: 2.29; 95% CI 1.08–4.82). This increased risk was primarily associated with clomiphene citrate use in parous women. An increased risk of thyroid cancer was not found after gonadotrophin use (RR: 1.43; 95% CI 0.54–3.83), but was observed following progesterone use (RR: 10.14; 95% CI 1.93–53.34), although based upon only two patients. In a large cohort study [ 46 ], clomiphene citrate use did not significantly increase the risk of colon cancer (RR: 0.83; 95% CI 0.4–1.9). Furthermore in retrospective cohort study [ 47 ] of 9,892 women followed for a median of 30 years to detect 91 colorectal cancers, clomiphene citrate use was unrelated to colorectal cancer risk (HR: 0.82; 95% CI, 0.52–1.30). Another recent cohort study from the Netherlands [ 49 ] evaluated the risk of colorectal cancer in 19,158 women undergoing ovarian stimulation for IVF compared to both women receiving other fertility treatments or the general population. A total of 109 colorectal cancers were observed after a median follow-up of 21 years. The IVF group did not have an increased colon cancer risk compared to the general population (SIR: 1.00; 95% CI, 0.80–1.23) but did have an increased such risk compared to women receiving other fertility treatments (HR: 1.80; 95% CI, 1.10–2.94). Given several studies examining occurrence of malignant melanoma following use of fertility treatments, most have not shown an increased risk of this disease [ 46 ] [ 50 ]. One retrospective cohort study from Australia [ 51 ] found that women giving birth following IVF had an increased rate of invasive melanoma versus women who failed to conceive after IVF, based on 139 invasive melanoma cases (HR: 3.61; 95% CI 1.79–7.26). There was, however, no increased risk in women receiving nonIVF-related fertility treatment (HR: 1.39; 95% CI 0.88–2.20). In another study, fertility treatment was not associated with malignant melanoma, except for an increased risk following use of gonadotrophins or gonadotropin-releasing hormone in parous women [ 52 ]. Finally, use of clomiphene citrate has been associated with increased risk of melanoma in some [ 53 ] [ 47 ] but not all [ 46 ] [ 52 ] studies.

Ovarian cancer is a rare but often fatal disease, accounting for 3% of all annual cancer diagnoses in women. This complex, multifactorial disease is often diagnosed at a late stage, since effective screening techniques for early disease detection do not yet exist. Risk factors for invasive ovarian cancer include nulliparity, infertility, late-onset menopause, family history and genetic susceptibility; while factors protective against disease are multiparity, breastfeeding and oral contraceptive use [ 5 ] [ 6 ]. Each of these factors must be considered when examining the mechanistic relationship between ovarian cancer and use of fertility medications. Uninterrupted ovulation causing repeated ovarian epithelial surface damage underlies the ‘incessant ovulation theory’ [ 7 ] [ 2 ], based upon the theory that factors decreasing lifetime ovulation rates, such as multiparity, concomitantly reduce ovarian cancer risk. Conversely, use of fertility medications could increase ovarian cancer risk by promoting polyfollicular ovulation [ 7 ]. This theory has been questioned by recent data that serous ovarian tumors can originate from the fallopian tube rather than the ovary itself [ 8 ]. An alternate mechanism underlying the relationship between fertility treatment and ovarian cancer is the ‘elevated gonadotropin level’ theory, which postulates that elevated gonadotropin levels resulting from fertility medications stimulate the ovarian epithelium and induce malignant changes [ 9 ]. Although studies examining the relationship between fertility medication use and ovarian cancer risk show mixed results, most studies do not demonstrate a significant increase in risk [ 10 ] [ 11 ] [ 12 ] [ 13 ] [ 14 ] [ 15 ] [ 16 ] [ 17 ]. In the 1990s, two studies found an association between fertility medication use and ovarian cancer [ 18 ] [ 19 ] but were limited by specific disease characteristics, such as low prevalence and late onset (i.e., 7th decade of life), study design (i.e., inappropriate use of fertile controls, limited length of follow-up, lack of control for parity and other diseases) and lack of stratification by type and duration of fertility medication use. Most recent studies and systematic reviews show no overall association between fertility medication use and invasive ovarian cancer when using a subfertile control group or the general population [ 10 ] [ 11 ] [ 12 ] [ 13 ] [ 14 ] [ 15 ] [ 16 ] [ 17 ], although some studies finding such an association used the general population rather than a subfertile group as controls. Asante et.al. and Briton et.al did not find an association between ovarian cancer and fertility treatment when using a subfertile control group [ 11 ] [ 20 ], even when women had undergone ≥4 cycles of IVF [ 20 ]. A meta-analysis of 9 cohort studies including 109,969 patients compared the ovarian cancer risk in women receiving fertility treatment with that in an infertile reference group and the general population. The ovarian cancer risk in women receiving fertility treatment was increased compared to the general population (relative risk [RR]: 1.50, 95% CI 1.17 – 1.92) [ 17 ], but similar to that of the infertile reference group (RR: 1.26, 95% CI 0.62 – 2.55). That subfertile women have an inherently increased risk for ovarian cancer compared to that of the general population likely explains why a cohort study of parous women found fertility treatment as a significant risk factor for ovarian cancer over a 25-year follow-up (hazard ratio [HR]: 3.9, 95% CI 1.2–12.6), based on only 3 women with invasive ovarian cancer following IVF [ 21 ]. A number of more recent studies have attempted to separately stratify for type of fertility treatment, use of IVF and administration of gonadotropins and/or clomiphene citrate. Those specifically studying women undergoing IVF have generally not found an increased risk in ovarian cancer [ 15 ] [ 12 ] [ 10 ]. One cohort study of 53,859 IVF patients found no association between invasive ovarian cancer and IVF compared to that in the population (standardized incidence ratio [SIR]: 0.96–1.18, NS). Paradoxically, IVF patients had a decreased risk of all cancers (SIR: 0.78, 95% CI 0.73–0.83), although a study limitation was a short average follow-up time of only 5 years [ 12 ]. Another population-based cohort study of parous women in Norway found no significant association between IVF and ovarian cancer risk (HR: 1.56, 95% CI 0.94–2.60). An increased risk of ovarian cancer occurred in IVF patients who had only had one child but no longer existed after adjusting for multiple comparisons [ 13 ]. Despite these reassuring findings, two cohort studies show that women failing to conceive after clomiphene citrate therapy have an increased risk for ovarian cancer [ 15 ] [ 14 ], warranting further investigation of this subgroup of women. Two recent studies have assessed risk of ovarian cancer in BRCA mutation carriers undergoing fertility treatment, since this high-risk group may also be more likely to undergo fertility treatment for fertility preservation or diminished ovarian reserve [ 22 ]. Both studies are small in size but reassuring [ 23 ] [ 24 ]. One cohort study of 1,073 BRCA mutation carriers, of which 164 (15%) patients received fertility treatment, showed no associated between fertility treatment and ovarian cancer, regardless of type of fertility treatment [ 24 ]. Another matched case-control study of 941 pairs of BRCA mutation carriers with and without a cancer diagnosis found no significant relationship between fertility medication use and subsequent risk of ovarian cancer [ 23 ]. Borderline ovarian tumors (i.e., tumors of low malignant potential) are non-invasive, indolent tumors that differ from invasive ovarian cancer in having an excellent prognosis with a five-year survival rate of 95% and occurring more commonly in reproductive-aged women. These tumors also are rare, with an incidence of 1.8 – 4.8 per 100,000 women years, making the absolute versus relative risk of this disease an important consideration [ 25 ]. Unlike invasive ovarian cancer, studies have found a more consistent association between borderline ovarian cancer and use of fertility medication. Early studies, including a cohort study demonstrated a significant association between borderline ovarian tumors and women receiving fertility treatment versus those not receiving treatment (SIR: 3.3, 95% CI 1.1–7.8), based on 5 cases of borderline ovarian cancer [ 18 ]. Larger recent studies using infertile controls have found similar results [ 26 ] [ 10 ]. A Dutch cohort study of 19,861 women undergoing IVF versus 6,604 subfertile women with an average follow-up of 14.7 years found an increased risk of borderline ovarian tumors in the former women, adjusting for age, parity and fertility diagnosis (SIR: 1.79, 95% CI 1.16–2.56) [ 10 ]. Another cohort study found a similarly increased rate of borderline ovarian cancer in women who received IVF, controlling for confounders (HR: 2.66, CI 1.2–5.04). Unlike its invasive counterpart, borderline ovarian cancer has not been associated with parity, endometriosis or prior surgery [ 26 ]. Conversely, not all evidence demonstrates an association between fertility treatment and borderline ovarian tumors [ 27 ] [ 11 ] [ 28 ]. A recent case-cohort study of 96,545 Dutch subfertile women, including 142 cases of borderline ovarian tumors, found no association between all subtypes of ovarian borderline cancer and fertility medication use (RR: 1.00, 95% CI 0.67–1.51), which was consistent across fertility treatment cycles and follow-up duration [ 27 ]. An increased association between serous borderline ovarian tumors and progesterone use currently lacks biologic explanation. Although a large systemic review shows heterogeneity of data [ 16 ], a possible link between borderline ovarian tumors and fertility medications cannot be ignored. Despite possible surveillance bias, most borderline ovarian tumors have been diagnosed up to 7 to 9 years after fertility treatment [ 26 ], raising concern about such an association, although the absolute increased disease risk is extremely small, given the low incidence of disease.

Data on the association between fertility treatments and cancer risk are limited, being mostly derived from observational cohort or case-control studies. Many such studies have methodological problems including small sample size, heterogeneous populations and treatments, multiple co-variables and short-term follow-up. Additionally, infertility per se is a risk factor for certain cancers including those of the breast, ovary and uterus. Recent studies have attempted to address these limitations by investigating more homogeneous populations, employing longer follow-up intervals, assessing cancer risk in low- versus high-risk individuals and using infertile women as appropriate controls. Overall most studies show that fertility treatments do not increase the risks of invasive ovarian cancer, malignant melanoma or cancers of the endometrium, cervix, breast, thyroid or colon. There is, however, limited evidence that use of fertility treatments may modestly increase the risk of borderline ovarian cancer, although this risk is small in absolute terms and such an ovarian cancer with low-malignant potential, if found, has a favorable prognosis.