Ovarian cancer is the second most common gynecologic malignancy in the United States, with an estimated 21,290 new cases expected this year. Ovarian cancer causes 5% of all cancer deaths in women, making it responsible for the highest number of gynecologic cancer deaths.¹ Age, family history, and inherited mutations such as BRCA1 and BRCA2 have been established as significant risk factors for the development of ovarian cancer. However, the role of menopausal hormone therapy remains questionable.

In a study published by The ­Lancet, researchers in the Collaborative Group on Epidemiological Studies of Ovarian Cancer demonstrated that use of menopausal hormone therapy was associated with an increased risk of ovarian cancer, with risk being highest among current users.² The study is summarized in this issue of The ASCO Post. 

The study, funded by the Medical Research Council of Cancer Research UK, consisted of meta-analyses of individual participant data sets from 52 epidemiologic studies, including 17 prospective and 35 retrospective studies. The principal analyses involved prospective studies, with last hormone therapy use extrapolated forward for up to 4 years.

Risk in Current Users

During prospective follow-up in the principal analysis, 12,110 postmenopausal women developed ovarian cancer, including 6,601 (55%) who had used hormone therapy. Compared with never-users, current users of menopausal hormonal therapy had about a 40% increased risk of developing ovarian cancer. This risk was similar for current users regardless of length of use (< 5 years of use: relative risk [RR] = 1.43, 95% confidence interval [CI] = 1.31–1.56; ≥ 5 years of use: RR = 1.41, 95% CI = 1.32–1.50). In an analysis including all studies (including 9,378 cases in 35 retrospective studies), compared with never-users, current users had increased risk with < 5 years of use (RR = 1.27, 95% CI = 1.18–1.37) or ≥ 5 years of use (RR = 1.34, 95% CI = 1.28–1.41).

Risk With Current/Recent Use

Compared with nonuse, current use or recent use (defined as use of any duration stopped < 5 years before diagnosis) was associated with a significantly increased risk (RR = 1.37, 95% CI = 1.27–1.48). This risk was similar in European and American prospective studies and for prospective studies involving estrogen-only and estrogen-progestogen hormone preparations. Analysis in retrospective studies only did not show an increased risk (RR = 1.04, 95% CI = 0.93–1.16).

Ovarian Cancer Subtypes

This increased risk of ovarian cancer was only seen for certain histologic subtypes. Hormone therapy use was associated with an increased risk of serous tumors (RR = 1.53, 95% CI = 1.40–1.66) and endometrioid tumors (RR = 1.42, 95% CI = 1.20–1.67) but not for mucinous tumors (RR = 0.93, 95% CI = 0.77–1.12). Hormone therapy use was associated with a reduced risk of clear cell tumors (RR = 0.75, 95% CI = 0.57–0.98).

Persistence of Risk

Although risk declined with greater duration since last use, women who had used hormone therapy for at least 5 years (median duration = 9 years) and then stopped still had a modestly increased risk of ovarian cancer more than 5 years later, compared with never-users (median time since last use = 10 years; RR = 1.10, 95% CI = 1.01–1.20). There was no increased risk seen in women who reported less than 5 years of use more than 5 years ago (RR = 0.94, 95% CI = 0.88–1.02).

Estimates of Excess Incidence and Death

The investigators interpreted their results as showing that “the increased risk may well be largely or wholly causal.” In this case, for 5 years of hormone therapy use, the absolute excess in ovarian cancer incidence per 1,000 users would be 0.52 (from the expected incidence of 1.2 per 1,000 never-users of hormone therapy) in women aged 50 to 54 years, 0.37 (from the expected 1.6) in those aged 55 to 59 years, and 0.10 (from expected 2.1) in those aged 60 to 64 years. In total, there would be 1 additional ovarian cancer per 1,000 users and 1 additional death per 1,700 users among women of all ages.

Interpreting Results

This study is an important contribution to the literature, with several strengths—particularly the large number of patients who were followed prospectively. As the authors point out, “the robustness of prospective data is demonstrated by the stability of the findings in various sensitivity analyses.” Their search strategy and analyses were comprehensive and included the grey literature in order to decrease the impact of publication bias.

However, their results must be interpreted with the study limitations in mind. First, the exclusion of women less than 55 years of age who underwent hysterectomy results in possibly significant selection bias, given the frequency of prescribing hormonal therapy in this group. Since hysterectomy has been shown to decrease the risk of ovarian cancer,³ including these women would likely have biased the effect measure toward the null. Perhaps a more prudent approach would have been to explore the relationship between hormonal therapy and ovarian cancer in this group through a sensitivity analysis rather than exclude them outright.

Additionally, the authors’ claim that hormonal therapy may be causal is not substantially supported. While the authors base their conclusions on prospective data, which provide evidence of a temporal relationship between the exposure and outcome of interest—thereby supporting their causal claim—there is little discussion about the proposed underlying mechanism by which hormonal therapy causes ovarian cancer.

Hormonal therapy as a cause of ovarian cancer is inconsistent with the well-established protective link between oral contraceptive use and ovarian cancer. Further evidence against biologic plausibility is the differential effect of hormonal therapy based on histologic subtype. Reconciling why hormonal therapy would increase risk of serous and endometrioid but not clear cell or mucinous tumors is problematic, and the authors offer no substantive discussion on this issue.

The authors claim that they demonstrate a dose-response relationship by showing that current users are at greatest risk of ovarian cancer, thereby lending further support for their causal claim. However, this is also problematic. Indeed, while current users are shown to have the highest risk compared to past users, there is no difference in risk for current users based on duration of use.

Furthermore, there are no data on actual dose of hormonal therapy, and prescribing patterns for hormonal therapy have varied substantially over the years. The data as presented are insufficient to provide strong evidence of a dose-response relationship.

Closing Thoughts

All in all, the study provides compelling evidence of a possible association between menopausal hormonal therapy use and serous and endometrioid ovarian cancers. However, the study provides insufficient evidence to claim that hormonal therapy causes ovarian cancer, and many important questions remain.

When counseling patients, it is important to discuss these findings, particularly in terms of absolute risk, within the context of the known benefits of hormonal therapy. Women who are taking or are considering starting hormonal therapy should discuss their personal and family histories and symptoms with their physicians, who can help weigh the risks and benefits and discuss alternative treatments. ■

Disclosure: Dr. Tergas reported no potential conflicts of interest.

References

1. American Cancer Society: Cancer Facts & Figures 2015. Atlanta, American Cancer Society, 2015.

2. Collaborative Group on Epidemiological Studies of Ovarian Cancer: Menopausal hormone use and ovarian cancer risk: Individual participant meta-analyses of 52 epidemiological studies. Lancet 385:1835-1842, 2015.

3. Falconer H, Yin L, Grönberg H, et al: Ovarian cancer risk after salpingectomy: A nationwide population-based study. J Natl Cancer Inst 107:dju410, 2015.

Dr. Tergas is Clinical Instructor of Gynecologic Oncology, at Columbia University Medical Center, New York.