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


To calculate a "number needed to treat" to overcome infertility, several factors must be taken into consideration including infertility diagnoses, patient age, and clinic success rates. Therefore, we cannot truly estimate the likelihood of being helped versus being harmed from this large cohort of patients. Indeed, data from this study suggests that fertility drugs might in fact protect against ovarian cancer. Data regarding the success of IVF in this particular cohort would have been helpful.

The cohort is sufficiently large and demonstrates a reasonable sample of infertility diagnoses to accept that our patients are well represented here. However, as mentioned before, the follow-up is not sufficiently long to truly identify whether the incidence of ovarian cancer is or is not increased in infertile patients treated with IVF.

Alternatives to induction of ovulation would include: expectant management, adoption, and oocyte donation. These can be very difficult and occasionally unacceptable alternatives for many patients. As we described in the introduction, Rosen at al assessed the willingness of infertile patients to accept an increased risk of ovarian cancer in exchange for an increased probability of conception.[5]

Unfortunately, this article is not an optimal example to assess harm. However, we can use the extreme end of the crude confidence interval stated in the abstract to demonstrate this point. For this purpose, assume that the relative risk for ovarian cancer in patients exposed to fertility drugs versus unexposed is 1.84, almost twice as frequent in the exposed versus unexposed population. The importance of the relative risk depends on the absolute risk in the unexposed population. If the incidence in the unexposed population is 10%, the incidence in the exposed population will be 20%. However, if the risk is very small (in this case it was 6.64 x 10-4), then the risk in the exposed population is also small (2 x [6.64 10-4]). If 1.84 were the point estimate for a confidence interval 1.70 to 1.90, we could state that the estimate of relative risk is precise. However, if the confidence interval were -5.00 to 5.00, not only is the estimate not precise, but the confidence interval crosses a value of relative risk = 1 and is therefore not informative.


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.