Abstract & Commentary
Synopsis: There appears to be an increase in ovarian cancer with long-term estrogen-only hormone replacement therapy.
Source: Lacey J, et al. JAMA. 2002;288:334-341.
Between 1973 and 1980, the American Cancer Society and the National Cancer Institute conducted the Breast Cancer Detection Demonstration Project (BCDDP). More than one quarter of a million women were involved in this study. Follow-up of these participants began in 1979, and 4 phases have been completed. Questions concerning hormone therapy were asked. Initially no distinction was made between estrogen-only replacement therapy (ERT) and estrogen-progestin replacement therapy (EPRT). A distinction was made beginning in 1987.
Using appropriate exclusions, 44,247 women were available for inclusion in this study. Diagnoses of ovarian cancer were verified through medical record review, cancer registry information and the National Death Index. Three hundred twenty-nine cases of ovarian cancer were diagnosed, virtually all of the epithelial type. The mean length of follow-up of the women in the study was 13½ years. The mean age at the start of follow-up was slightly less than 57 years. Parity, use of oral contraceptives, and hysterectomy were inversely associated with ovarian cancer. Estrogen-only hormone replacement therapy was significantly associated with ovarian cancer. Longer use of estrogen resulted in higher risk ratios. For those women who used ERT for 20 years or more, there was a greater than 3-fold increase in ovarian cancer. The increase was approximately 7% per year of ERT usage. There was no significant increase in risk for those women who took only EPRT.
Comment by Kenneth L. Noller, MD
This important study was greatly overlooked because of the tremendous publicity surrounded by the decision of the Women’s Health Initiative not to continue the estrogen plus progestin arm of their study due to various increased risks. That is unfortunate as this well-done study suggests that there also are risks associated with estrogen-only therapy.
Over the past several years, I have frequently commented on study design, P-values, and confidence intervals. Certainly the most powerful type of study uses a prospective, randomized study (experimental) design. Cohort studies, such as this report, are less powerful. Nonetheless, if performed properly and if the risk ratio is significantly elevated, they are quite believable. While an increase of risk of 20% or 30% is usually believable in a prospective randomized trial, a similar increase in a cohort study is questionable.
In this study by Lacey et al, there was a 3-fold (300%) increase in ovarian cancer among estrogen-only hormone replacement users. There was a smaller increase in women who took the medication for only 10 years, but there was a definite dose response association. That is, the risk of ovarian cancer increased with increasing length of usage.
This is quite a good study and it could stand alone. However, other articles have found a similar association, and thus I think it is reasonable for us to believe that the use of estrogen replacement therapy may actually increase a woman’s risk of developing epithelial ovarian cancer. What is harder to understand is why this association would be true. So far I am not aware that anyone has developed a believable pathophysiological explanation.
Dr. Noller is Professor and Chairman Department of OB/GYN, Tufts University School of Medicine, Boston, Massachusetts.