Does Hormone Therapy Protect the Heart?

It Is a Stressful Question

Nanette Santoro


J Clin Endocrinol Metab. 2020;105(5) 

The notion that menopausal hormone therapy provides cardioprotection for women is one of the most contentious controversies in menopausal medicine. It is a vexing problem because it forces us to confront an apparent discrepancy between observational cohort studies and randomized, clinical trials. The Nurses' Health Study, one of the most comprehensive and longstanding cohort studies of women, reported an overall, age-adjusted relative risk (RR) of coronary disease of 0.5 for ever users of hormone therapy, and 0.3 for current users.[1] Ten-year follow-up confirmed this reduction in risk (age-adjusted RR 0.72, 95% CI 0.55–0.95, P = .02), but at a lesser magnitude.[2] In the same 1985 issue of the New England Journal of Medicine, the Framingham Heart Study reported an increased RR of heart disease of 1.76 in association with menopausal hormone therapy.[3] The Women's Health Initiative (WHI), which was largely conceptualized to resolve this discrepancy, did not find any cardiovascular benefit of estrogen-alone therapy and found harm with use of estrogen plus progestin in the initial publication,[4] which became nonsignificant with further follow-up.[5] Controversy has raged ever since about the presence of cardioprotection, the possible relationship to age or timing within the menopause transition of the intervention, and the mechanisms by which menopausal hormone therapy might attenuate cardiovascular risk.

To explain these disparate findings, investigators have turned to the "timing hypothesis," which, simply put, posits that intervention with hormones is protective against heart disease only when it is initiated early (ie, within a few years) after the onset of ovarian failure. Subsequent randomized clinical trials reporting intermediate outcomes of cardiovascular disease such as carotid intimal medical thickness (CIMT) and/or coronary calcium have attempted to address the timing hypothesis, and again we are left with conflicting results. The Kronos Early Estrogen Prevention Study (KEEPS) was a multisite trial of 727 women age 45 to 54 years who were within 4 years of their final menstrual period, randomized to transdermal estradiol or conjugated equine estrogen, both with intermittent micronized progesterone, or placebo and did not demonstrate any differences between the 3 groups with respect to CIMT (primary end point) or coronary calcium (secondary end point) accrual over time.[6] However, the ELITE trial, a single-site study that randomly assigned 643 women to oral estradiol or placebo (also with intermittent micronized progesterone), found that hormone therapy initiated within 6 years of menopause resulted in a slower rate of CIMT accrual compared to women treated with placebo, but those who initiated treatment 10 years or more after menopause did not differ from placebo-treated women with respect to CIMT accrual over 5 years.[7] These studies leave open the possibility that hormone therapy exerts an undetermined beneficial effect on a woman's cardiovascular system that is limited to a certain amount of time since menopause, or perhaps to age per se.

In this issue of the Journal, Gordon et al have conducted a trial of 172 women age 45 to 60 years, most of whom were late perimenopausal or recently menopausal, randomly assigned to 12 months of transdermal estradiol and intermittent oral micronized progesterone. Participants were evaluated for dynamic markers of arterial function, with a primary outcome being a composite score for stress reactivity, and secondary measures of endothelial function, baroreceptor reflex sensitivity, and presence of metabolic syndrome. Their findings complement the psychological findings previously reported by the Perimenopausal Estrogen Replacement Therapy (PERT) Study, indicating that perimenopausal and early postmenopausal women treated with this regimen had an almost 50% decrease in depressive symptoms compared to placebo-treated women, as measured by the Center for Epidemiological Studies-Depression Scale.[8]

There are some caveats to the authors' findings that need to be taken into account. As the authors point out, the PERT study was powered for a primary outcome of depressive symptoms, not for these additional outcomes. The arterial outcome measures used, although currently believed to be linked to subsequent cardiovascular disease end points, are not necessarily predictive of disease outcomes. The dose of transdermal estradiol used in this study is twice as high as what has been used in similar studies of menopausal hormone therapy[6] and may have a different risk profile; similarly, the intermittent micronized progesterone regimen awaits much more rigorous testing for endometrial safety before it can be recommended. Moreover, the effect of treatment on all main outcomes did not differ significantly by study visits. Nonetheless, the finding of age-related worsening of measures both of stress reactivity and endothelial function in the placebo group implies that there is more to be explored about the timing hypothesis.

Gordon and colleagues' contribution helps light the path forward for further research, and, as does all good science, raises new questions. It seems that blood vessel function and not merely anatomy is a promising place to look to resolve the discrepancies between cohort studies and randomized clinical trials. Stress, loosely defined, is an overlooked predictor for heart disease in women. The authors bring this concept back to the forefront and cast it in the context of the timing hypothesis. This may well be where it belongs. Among questions that remain to be addressed include the time-limited nature of a hormone intervention, optimal dosing, and the durability of its effect. If early menopausal intervention with hormone therapy is consistently found to preserve endothelial function, does the effect subside with time since menopause and/or age? Do the arterial outcomes demonstrated in this study result in reduced morbidity and mortality? If hormones do, indeed, provide later-life cardioprotection from women, how can we best balance this benefit against their known risks?