Ovarian Cancer and Fertility Medications: A Critical Appraisal

S. Kashyap, M.D., F.R.C.S.(C), O.K. Davis, M.D., F.A.C.O.G.


Semin Reprod Med. 2003;21(1) 

In This Article


Ovarian cancer is the fifth most common cancer in the Western world. It accounts for 4% of all malignancies in women. The lifetime risk of ovarian cancer is 1.75%. Despite concern regarding increasing use of infertility therapy, the incidence of ovarian cancer in the Western world has remained stable for several decades.[18,19,20]

Seventy percent of ovarian cancers present as advanced disease (stage III or IV). The 5-year survival for stage IIIa disease, estimated in 1994, is 30 to 40% and for stage IIIb disease is 5%. Most tumors are diagnosed in women between the ages of 55 and 59; the age of peak incidence has been quoted as 62. Ninety percent of ovarian cancers arise from the epithelium: serous (75%), mucinous (20%), endometrioid (2%), clear cell (< 1%), Brenner (< 1%), and undifferentiated tumors (< 1%). Approximately 10% of epithelial tumors are "border-line" or low malignant potential tumors. Borderline tumors are not frankly invasive, occur at a mean age of 40, and have a much better prognosis than their invasive counterparts. Nevertheless, risk factors for borderline tumors are the same as invasive tumors. Two to three percent of ovarian tumors are of germ cell origin and another 2 to 3% are sex-cord stromal tumors.

Several factors modify the risk of this disease. Two independent studies, one case control study, and one cohort study have shown that the risk of ovarian cancer appears to decrease by 19% for each successive live birth.7,21 Other studies support a protective relationship between parity and the incidence of ovarian cancer.[22,23,24] Nulliparity increases the baseline risk twofold. Infertility appears to increase the risk in those patients who do not subsequently conceive.[25,26,27,28] Use of OCP also has been shown, in several studies, to decrease the risk. The most recent study demonstrated a 60% risk reduction after 6 years of use and this benefit appears to last for 10 years after discontinuation of the OCP.[29] For women with a documented high-risk genetic predisposition, the risk may be increased by as much as 4 to 50%.

Weaker data support the following protective factors: tubal ligation, hysterectomy, and lactation. Some authors have suggested that the risk of ovarian cancer is increased with postmenopausal hormone replacement therapy, early age of menarche, late age of menopause, and advanced age of first pregnancy.

Infertility, commonly defined as the inability to conceive after 1 year of unprotected intercourse, affects 10 to 15% of couples. Although the incidence of infertility has remained stable, the use of infertility drugs has increased rapidly. It is estimated that 2.5% of all North American births are the result of artificial reproductive technology.[30] In the United States, clomiphene citrate was registered in 1967. Gonadotropins were registered in 1969. Prior to that time, the hormonal treatment of infertility involved the use of estrogens, progestins, OCP medications, and pituitary radiation.

The article analyzed here is the largest and perhaps most methodologically sound study of its kind to date. It was unable to show a difference in incidence of ovarian cancer between exposed versus unexposed infertile cohorts. Cohort studies are generally more reliable than case control study in assessing adverse events associated with therapy. Cohort studies have inherent biases, as previously discussed. Case control studies generally have even more threatening biases, especially with regard to selection and recall. Nevertheless, one of the advantages of case control studies is that it provides an easier vehicle to study a rare disease. If the frequency of an event is low, the initial preferred study would be a case control study. In such a study, "cases" of disease are identified. Unaffected controls are also identified. The subjects are then investigated for exposure to the agent in question, in this case, ovulation induction. In rare diseases, one can ensure a minimum number of cases by identifying them initially and therefore generally have a greater "power" than cohort studies. However, several biases can be introduced here. Selection bias may influence the conclusion in the direction of the investigator's hypothesis. Patients who have experienced the outcomes (cases) may also experience recall bias because they are more introspective about potential causes of their dis-ease. Patients who have not had the outcome of interest may be less likely to remember such exposure.

The data in the study under review is limited. The follow-up was not sufficient to ensure that the risk is either elevated or reduced. The age of peak incidence of ovarian cancer is 62 years old. One in 70 women will, in their lifetime, develop ovarian cancer. In this study, the observed incidence was higher in untreated infertile patients (6/9044) versus treated infertile patients (7/20,656). The median age at the end of follow-up for exposed and unexposed patients was 39 and 40 years old, respectively. These ages do not approach the average time of diagnosis of this disease (age 55 to 60). Although it is difficult to organize prospective long-term follow-up, further analysis of the cohort at a later date might be very informative.

The authors understandably excluded borderline tumors from their analysis (with the exception of one case that was misclassified). Nevertheless, it is plausible that borderline tumors are increased in treated and/or untreated infertile patients.

Most studies that try to assess the risk of fertility drugs and ovarian cancer employ general population controls. Because the risk of ovarian cancer has been shown to be independently associated with infertility, general population controls are not adequate. Here, the authors have an unexposed infertile cohort that may provide a very useful control group, if the cohort were to be followed over a longer period of time.

Investigation of subsets of patients might also provide further insight into the pathogenesis of ovarian cancer. For example, approximately 40% of couples experience infertility secondary to a male factor. If a male factor is the obvious and exclusive diagnosis, investigation of women who undergo IVF intracytoplasmic sperm injection, who might otherwise be considered potentially fertile, could allow differential analysis of the relationship of ovarian cancer to nulliparity in reproductively intact women versus women with female reproductive disorders versus treatment effects per se. Parity and use of OCPs are important confounders that must also be addressed.

Additionally, ovarian stimulation medications in a cohort of oocyte donors would be an interesting area to investigate. Again, it would be necessary to collect information regarding OCP use as well as parity to make valid conclusions.

When our patient returns we can now tell her the following: To date, the link between ovarian cancer and fertility drugs has not been established. The evidence from this study and a recently published meta-analysis[11] suggests that fertility therapy does not increase the risk of ovarian cancer in infertile patients who already have an increased baseline risk as a result of their infertility. However, current evidence lacks sufficient long-term follow-up, as well as information on confounders, to be conclusive. Results of further follow-up of existing or new cohorts would be helpful.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.