The trusted source for
healthcare information and
By Jeffrey T. Jensen, MD, MPH
Leon Speroff Professor and Vice Chair for Research, Department of Obstetrics and Gynecology, Oregon Health & Science University, Portland
Jensen, MD, MPH, reports that he is a consultant for and receives grant/research support from Bayer, Merck, ContraMed, and FHI360; he receives grant/research support from Abbvie, HRA Pharma, Medicines 360, and Conrad; and he is a consultant for the Population Council.
SYNOPSIS: An evaluation of outcomes of users and nonusers of postmenopausal hormonal therapy followed longitudinally in the Danish database showed no overall difference in mortality.
SOURCE: Holm M, Olsen A, Au Yeung SL, et al. Pattern of mortality after menopausal hormone therapy: Long-term follow up in a population-based cohort. BJOG 2018; Aug. 14. doi: 10.1111/1471-0528.15433. [Epub ahead of print].
The use of postmenopausal hormone therapy (HT) remains controversial despite increasing evidence supporting at least short-term benefits for women experiencing menopausal symptoms and intriguing suggestions of other long-term benefits for healthy menopausal women. The Women’s Health Initiative (WHI) randomized studies ended prematurely because of a suggestion of an overall increased risk of morbidity and mortality. However, most recently, a new analysis of long-term outcomes of women enrolled in WHI revealed no evidence of a differential effect on mortality after up to 18 years of follow-up.1 Although informative, data from the WHI relate to the use of conjugated estrogens with and without medroxyprogesterone acetate, and are not generalizable to other treatment regimens. To provide more context, Holm et al used a population-based cohort to study the association between HT and mortality. The Diet, Cancer and Health Cohort was a large, Danish, population-based study. The cohort, established between 1993 and 1997, included 29,875 women (age 50-64 years and without a previous cancer diagnosis) who responded to an invitation to participate in a longitudinal study. This corresponds to 7% of the Danish female population in the age range. A strength of this database was the collection of baseline data that included two self-administered questionnaires, anthropometric measurements, and blood samples obtained at a clinic visit. The questionnaires asked women about use of HT (never, previous, and current use), route of administration (oral, systemic nonoral, and local), and brand names used. Based on this information, investigators categorized women as users of estrogen alone, combination therapy (estrogen and progestogen), and sequential or continuous regimens.
To assess outcomes, investigators linked cohort members to the Danish Causes of Death registry using their unique national personal identification numbers. In the analysis, researchers considered alcohol intake, smoking, physical activity, and body mass index as possible confounders based on a priori hypotheses of associations with the outcome. They calculated associations between HT and mortality using Cox’s proportional hazards models. At baseline, 54.4% of the women had never used hormones, 15.5% were previous users, and 30% were current users. Among those women using hormones at baseline, 30% used estrogen only, 21% used continuous combined, and 32% used combined sequential. The type of HT was not defined in 17% of women. Most participants reported use of oral HT. During a median follow-up interval of 17.6 years, 4,098 participants died. Of these, 2,155 died from cancer, 671 died from cardiovascular disease, and 1,084 died from other causes. The cause of death was not determined in 188 cases. After adjustment for relevant lifestyle risk factors, hormone use did not affect all-cause mortality, regardless of type or route.
Looking more closely at specific causes of death, investigators found no differences in cancer-specific mortality or mortality from other causes between women on HT and never users. They further subdivided death due to specific causes of cancer and specific causes of cardiovascular diseases. They found colorectal cancer mortality markedly lower among both current (hazard ratio [HR], 0.64; 95% confidence interval [CI], 0.46-0.89) and previous users (HR, 0.67; 95% CI, 0.46-0.99) compared with never users. They found the opposite association with breast cancer mortality; current users had significantly higher mortality compared with never users (HR, 1.34; 95% CI, 1.05-1.72). The risk was highest and statistically significant only for combined continuous HT (HR, 1.56; 95% CI, 1.05-2.31) and not statistically significant (HR, 1.37; 95% CI, 0.95-1.98) for women reporting use of estrogen only. Although no overall associations were observed between HT and cardiovascular disease, an interesting increase was seen in the risk of ischemic heart disease associated with use of local estrogen only (relative risk, 2.74; 95% CI, 1.38-5.44). Aside from this, the authors found no other associations between current or past HT use or type of HT and risk of death.
This paper adds to our understanding of the long-term risks of postmenopausal hormone therapy. The research offers few surprises with respect to cancer. A reduction in the risk of colon cancer death and a slight increase in breast cancer mortality were seen with users of combined HT, findings also reported in the WHI.2 The lack of effect of HT on overall cancer mortality is highly reassuring. Even more reassuring is the lack of an association between systemic HT and death from ischemic heart disease or stroke. Interestingly, the only cardiovascular association with a moderate effect size (HR > 2), death from ischemic heart disease, occurred in users of local (e.g., vaginal) estrogen, a route of administration not associated with systemic risks or benefits. This effect may be explained by the very low numbers of women who reported use of local therapy only. Adverse prescription bias also may have influenced this outcome if clinicians were less likely to prescribe systemic HT for women with underlying cardiovascular disease risk. However, it is important to note that these results reflect HT use initiated prior to the publication of WHI, a time in which clinicians generally agreed that use of HT was cardioprotective.
Similar to many of the database studies reported by this group, there were strengths and weaknesses to the Holm et al work. A major strength was the establishment of a cohort that included accurate collection of baseline data on body mass index (BMI) and hormone use. Unfortunately, we do not know much about continuation of use or how many women discontinued or initiated HT after the baseline assessment. We also rely on self-report to categorize the type of HT, which is surprising given the availability of accurate pharmacy records linked to unique national identity numbers. This group has used these pharmacy records to track exposure to hormonal contraception, so I am surprised at the lack of attempt to better follow HT exposure in this cohort. A major limitation of the other Danish registries studies was the lack of ability to adjust for baseline confounders, such as BMI, a notable strength of this study. In a companion paper using the same cohort, the same author group reported that a healthy lifestyle (e.g., regular exercise, nonsmoking) contributes to benefits seen with HT.3
We all will die someday, but we want to avoid premature mortality. This study supports a lack of association between HT and overall mortality. While I believe there may be additional benefits to HT, this study shows that at a minimum, women using HT should experience no overall risk. This allows us to focus on counseling about important quality-of-life issues such as reduction in hot flashes,4 protection against fracture risk,5 and improved sexual function.6
Financial Disclosure: Internal Medicine Alert’s Physician Editor Stephen Brunton, MD, is a retained consultant for Abbott, Acadia, Allergan, AstraZeneca, Avadel, Boehringer Ingelheim, GlaxoSmithKline, Janssen, Mylan, and Salix; he serves on the speakers bureau of AstraZeneca, Boehringer Ingelheim, Janssen, Lilly, and Novo Nordisk. Peer Reviewer Gerald Roberts, MD; Editor Jonathan Springston; Executive Editor Leslie Coplin; and Editorial Group Manager Terrey L. Hatcher report no financial relationships relevant to this field of study.