[Skip to Content]
[Skip to Content Landing]
June 27, 2001

Relationship Between Postmenopausal Hormone Replacement Therapy and Ovarian Cancer—Reply

Author Affiliations

Copyright 2001 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2001

JAMA. 2001;285(24):3089-3090. doi:10.1001/jama.285.24.3089

In Reply: The findings of Dr Bosetti and colleagues support the association that we observed between ovarian cancer and HRT use. In both analyses, the increased risk diminished after cessation of use. The critical question for clinical practice continues to be whether estrogen and progestin in combination or only unopposed estrogen use affects ovarian cancer risk.

We agree with Dr Hernandez that the data relating ovarian cancer to ever users of HRT are inconsistent. Case-control studies assessing ovarian cancer risk with 5 or more years of HRT use, however, have consistently reported increased risk, as we mentioned in our article. Nevertheless, we agree with Hernandez that the current evidence is incomplete and does not warrant an immediate change in clinical practice. Hernandez does, however, raise a more general and relevant question. When is the evidence from observational data sufficient to change medical practice? In the case of postmenopausal HRT, appropriate guidelines for individual women ultimately should be based on a full understanding of the balance between risk and benefits. Hernandez states that the association between postmenopausal HRT and breast and ovarian cancer is controversial. While we agree with this statement as it relates to ovarian cancer, we believe that the positive association between breast cancer and HRT use has been clearly established.1,2 Results from the Women's Health Initiative Trial will help clarify the impact of HRT use on risk of cardiovascular disease, osteoporosis, and breast cancer. The effect of postmenopausal HRT use on relatively rare diseases, such as ovarian cancer, cannot be studied with randomized trials and will likely only be evaluated through epidemiologic studies.

First Page Preview View Large
First page PDF preview
First page PDF preview