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


Overall, the authors suggest that there is no increased risk of ovarian cancer in the cohort compared with what would be expected in the general population. As mentioned above, we are concerned with the risk of ovarian cancer in treated infertile patients versus untreated infertile patients. Therefore, we recalculated the relative risk based on the crude numbers supplied by the authors. Seven cases occurred in 20,656 infertile patients who underwent IVF therapy. Six cases occurred in 9044 infertile patients who did not undergo IVF therapy. Therefore, the relative risk appears lower for infertile patients who do receive treatment than for those who did not undergo IVF therapy (RR 0.51, [confidence interval 0.17, 1.52]). However, the confidence interval does cross 1. In any case, the risk for treated patients is not elevated in comparison with the risk for untreated patients.

The risk associated with treatment is 7/20656 = 3.39 x 10-4 . The risk without treatment is 6/9044 = 6.64 x 10-4 . The absolute risk difference is -3.35 x 10-4 or 1 case per 100,000 women years or 1 per 100,000 woman-years. All we can conclude is that in this cohort, for this period of follow-up, fertility drugs do not appear to increase the risk of ovarian cancer.

When the incidence of disease is low (i.e., the disease is very rare), the relative risk is virtually mean-ingless. The lifetime incidence of ovarian cancer is 1 in 70. However, the peak age of incidence is age 62. The incidence at the ages provided in this study would be very low, even for infertile patients, and therefore it would be difficult to assess a true difference between exposed versus unexposed cohorts.

Although cohort studies are often the most reliable available data for assessment of adverse events related to therapy, case control data usually has greater power. However, case control data also has more inherent biases that threaten the validity. In a case control study, the "cases" of disease are identified initially, thereby ensuring a certain number of events. Controls are also identified by some means of selection. The subjects are then investigated for the exposure of interest. This method can potentially introduce several biases. Selection bias might skew the conclusions toward the authors' hypothesis. Patients who have the outcome of interest might experience recall bias because they are more introspective about the cause of their disease versus unaffected controls. Also, the investigators might search more or less vigorously depending on their convictions about the hypothesis. Nevertheless, for rare disease, case control studies are often the studies of choice because of their ability to ascertain a predetermined number of cases.

A cohort study is an appropriate design here. The incidence of ovarian cancer is not rare when considered over a lifetime. However, as mentioned above, the follow-up needs to be extended.


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.