Menopausal Hormone Therapy
For a generation the Women's Health Initiative made menopausal hormone therapy look dangerous for everyone. The 2024–2026 evidence reads the same data as a timing, route, and formulation problem: started within about ten years of menopause, with the right estrogen route and progestogen, MHT is a well-supported treatment for symptoms and bone loss — though not a heart-disease or anti-aging drug.
Roughly two billion women will be postmenopausal by 2030 — meaning a sizeable fraction of the global population will spend a third to half their lifespan in profound estrogen deficiency.[1] Ovarian failure is the earliest programmed organ-level senescence in the human body, and it drags cardiovascular, skeletal, neurological, and metabolic aging trajectories with it. Menopausal hormone therapy (MHT) is the only intervention that directly addresses that hormonal withdrawal. Whether it should be used as more than short-term symptom relief — in whom, with which molecule, by what route, and for how long — is the question the post-WHI literature has spent twenty years answering.
What menopause does, biologically
Menopause marks the terminal depletion of ovarian follicles and a permanent drop in circulating estradiol. Acute manifestations — hot flashes and night sweats (vasomotor symptoms, VMS), the genitourinary syndrome of menopause (GSM), sleep fragmentation, mood disturbance — are the visible top of the iceberg. The systemic consequences are larger: endothelial dysfunction, accelerated atherosclerosis, increased visceral adiposity, insulin resistance, accelerated bone resorption, ~0.6% annual loss of skeletal muscle mass, and faster epigenetic aging. Pooling four cohorts (n > 3,100), earlier menopause and bilateral oophorectomy both track with accelerated DNA-methylation age in blood, and a Mendelian-randomization analysis using menopause-associated variants points the causal arrow from menopause to epigenetic aging rather than the reverse.[2] Age at menopause is itself prognostic: a meta-analysis of 32 studies in 310,329 women found menopause before age 45 carried about 50% higher risk of coronary heart disease (CHD) and roughly 12% higher all-cause mortality versus menopause at 45 or later.[3]
Estrogen receptors (ERα and ERβ) sit on vascular endothelium, hippocampal and prefrontal neurons, osteoblasts, skeletal-muscle satellite cells, and the dermal matrix. Ovarian failure therefore isn't a reproductive event in isolation — it is the withdrawal of a master signaling molecule whose absence accelerates multiple hallmarks of aging simultaneously.
Two pharmacological branches exist:
- Estrogen-only therapy (ET) — for women without a uterus.
- Combined estrogen-progestogen therapy (EPT) — for women with an intact uterus; the progestogen exists solely to prevent endometrial hyperplasia from unopposed estrogen.[4]
The risk-benefit calculus differs sharply between the two and depends on three further decisions: when therapy is started, how it is delivered, and which progestogen is paired with the estrogen.
The timing hypothesis (Moderate)
The single most consequential reframing of MHT since 2002 is the timing hypothesis — the cardiovascular effect of estrogen depends on whether the endothelium is still healthy enough to respond to it.
In healthy vasculature, estrogen upregulates endothelial nitric oxide synthase, inhibits vascular smooth-muscle proliferation, and slows atherosclerotic plaque formation. Initiated in this state, MHT preserves vascular function. In already-diseased vasculature, the same molecules can transiently destabilize advanced atherosclerotic plaques — by activating enzymes that break down the artery-wall matrix — and shift the prothrombotic / pro-inflammatory balance in the wrong direction.[5]
The clinical translation:
- Early initiation (<10 years post-menopause or age <60) — the classic observational cohorts put coronary heart disease roughly 30–40% lower in users than non-users: a meta-analysis of 25 observational studies found about 30% lower risk (RR 0.70, 95% CI 0.65–0.75) and the Nurses' Health Study about 39% lower over 20 years (RR 0.61, 0.52–0.71) [6]. Those numbers have not survived a design built to remove healthy-user bias: a Swedish emulated target trial in 919,614 women aged 50–58 — exactly this window — found oral estrogen-progestin associated with about 21% higher ischaemic heart disease, on an interval that just touches no effect (HR 1.21, 95% CI 1.00–1.46), alongside clearly raised venous thromboembolism (oral continuous HR 1.61, 1.35–1.92), and no excess for any transdermal preparation [7]. A meta-analysis of the randomized trials found MHT had no effect on all-cause mortality overall (relative risk 0.99), but in the subset of trials that started therapy early — mean age under 60 or within 10 years of menopause — mortality was about 30% lower (RR 0.70, 95% confidence interval 0.52–0.95).[8] (A confidence interval is the range where the true effect most plausibly lies; when it stays below 1.0, as here, the benefit is unlikely to be chance.)
- Late initiation (>10 years post-menopause or age >60) — elevated short-term risk of myocardial infarction (MI, heart attack), stroke, and venous thromboembolism (VTE), no offsetting cardioprotective benefit.[9]
The WHI primary cohort had a mean age of 63 and a median time-since-menopause well over 10 years — exactly the population in which the timing hypothesis predicts harm. The original 2002 publications generalized that signal to all postmenopausal women, killing the field for two decades. Re-analyses stratified by age at initiation recover the early-initiation benefit.[10] Two prospective trials — KEEPS (Kronos Early Estrogen Prevention Study) and ELITE (Early versus Late Intervention Trial with Estradiol) — tested the hypothesis directly, with mixed results: ELITE showed the predicted timing-dependent vascular response, while KEEPS was null on its own primary endpoints — neither carotid-artery wall thickness progression nor coronary calcium differed from placebo in its 727 women, all within three years of menopause [11]. But a caution the popular framing tends to drop: both were powered on surrogate endpoints (carotid-artery wall thickness, coronary calcium), not hard cardiovascular events, and neither reduced heart attacks, strokes, or deaths. ELITE supports the mechanism; neither trial carries the coronary-incidence numbers, which come from the observational cohorts above, or the mortality figure, which comes from the early-initiation subgroup of the trial meta-analysis. The 2024 twenty-year follow-up of the WHI reaches the parallel conclusion for prescribing: hormone therapy should not be used to prevent heart disease, stroke, dementia, or other chronic disease at any age, even as its benefit-to-risk profile is more favorable in women under 60 treating symptoms.[12]
The 2025 guideline from the Korean Society of Menopause reaffirms this framework directly: starting standard-dose estrogen within 10 years of menopause or before age 60 is associated with reduced coronary artery disease (CAD) incidence and mortality — the guideline cites a roughly one-third reduction in CAD risk and an all-cause mortality reduction near 30% for early initiators — whereas the same therapy offers no benefit, and possibly harm, in women with established CAD.[13]
| Initiation timing | Endothelial response | Hard outcomes |
|---|---|---|
| <10 years post-menopause / age <60 | Slowed atherogenesis, stable plaque on carotid imaging | ~30–40% lower CHD in observational cohorts, not reproduced in bias-controlled designs; mortality reduction confined to the early-initiation subgroup (RR ~0.70); fewer fractures |
| >10 years post-menopause / age >60 | No slowing of carotid atherosclerosis (ELITE late arm); potential plaque destabilization | Elevated MI, stroke, VTE in the first 1–2 years of initiation |
One line the modern guidelines draw firmly: cardiovascular prevention is not, on its own, an indication for MHT. The timing-hypothesis benefit is a favorable side effect of treating something else, not a licence to prescribe hormones to prevent heart disease. The genuine first-line indications remain vasomotor symptoms, the genitourinary syndrome (best addressed with low-dose vaginal estrogen), and early postmenopausal bone loss — and even the cardiovascular association is graded only Moderate, weighed down by the healthy-user bias that runs through the observational data.
Route matters: oral vs transdermal (Strong)
Estrogen route changes the drug. Oral estrogens pass through the liver before reaching the systemic circulation; the hepatic first-pass triggers synthesis of clotting factors, raises C-reactive protein (CRP, a marker of inflammation), and elevates triglycerides. Transdermal estradiol (patch, gel, spray) bypasses the liver entirely and reaches tissue through systemic microcirculation, mimicking ovarian secretion.
The hard outcome difference is clean, with one condition on it: oral estrogen carries a real signal for venous thromboembolism and ischemic stroke, and transdermal estradiol at standard doses does not raise either above baseline. The European Society of Endocrinology puts it that way for clotting — transdermal administration "does not increase risk above baseline" — and adds that low or standard-dose transdermal estradiol is neutral on stroke while higher doses may not be [14]. The dose caveat is real: current syntheses find higher systemic estrogen doses raise stroke and clotting risk across both routes [15]. The European Society of Endocrinology names transdermal as the preferred route whenever cardiovascular risk factors are present, and specifically in well-controlled diabetes, well-controlled hypertension, obesity, high triglycerides, and migraine with aura [16]. A previous venous thromboembolism is a different case, not a simple route switch: it puts a woman outside the low-risk group and calls for an individual risk-benefit assessment before any hormone therapy. The same guideline advises against testing for inherited clotting mutations before starting.
Transdermal also avoids the oral-estrogen rise in triglycerides and CRP — relevant for any patient with metabolic syndrome.
The progestogen question (Strong)
In the randomised trials, the breast cancer signal that drove the 2002 WHI panic was almost entirely a progestogen signal, not an estrogen signal. Two strands of evidence carry that conclusion.
ET-only post-hysterectomy is breast-neutral, possibly protective. Extended follow-up of the WHI estrogen-alone arm showed lower invasive breast cancer incidence and lower breast cancer mortality on conjugated equine estrogens (CEE) vs placebo.[17] A plausible mechanism: breast cells adapted to a chronic low-estrogen environment can undergo apoptosis on reintroduction.
That is the randomised result, and the observational evidence disagrees with it. Pooling individual data on 108,647 postmenopausal breast cancers, the largest analysis in the field found an excess with every type of hormone therapy except vaginal estrogen — including estrogen alone, at 17% higher risk in the first four years of use and 33% higher in years 5–14. In absolute terms, for women of average weight in developed countries, five years of therapy from age 50 adds about one extra breast cancer per 50 users of estrogen plus daily progestogen, one per 70 with intermittent progestogen, and one per 200 with estrogen alone, counted over ages 50–69 [18]. The same analysis found daily progestogen worse than intermittent (RR 2.30 vs 1.93), so exposure-days matter alongside molecule choice.
Within EPT, the choice of progestogen is the dominant variable. A 2026 systematic review of 34 studies puts numbers on where the signal lives: combined estrogen-progestogen therapy carried about 44% higher breast cancer risk (OR 1.44, 95% CI 1.26–1.64), while estrogen-only therapy showed none at all (OR 1.00, 0.91–1.10, and 22% lower when the analysis was restricted to randomised trials). It could not go further than that — the review states in its own limitations that the published data lacked the detail to separate individual progestogens [19]. The molecule-level evidence is older and observational. Natural micronized progesterone ("body-identical") and the retroprogesterone dydrogesterone show neutral-to-negligible breast risk at standard treatment durations. The cleanest single dataset is the French E3N cohort — 80,377 postmenopausal women, 2,354 invasive breast cancers over a mean 8.1 postmenopausal years — in which, relative to never-use, breast cancer risk was essentially unchanged with estrogen plus micronized progesterone (1.00, 95% CI 0.83–1.22), a non-significant 16% higher with estrogen plus dydrogesterone (1.16, 0.94–1.43 — the interval crosses 1.0, so compatible with no effect), and a clear 69% higher with estrogen combined with other progestogens (1.69, 1.50–1.91). Route of estrogen made no difference; the progestogen did.[20] A later analysis of the same cohort, with longer follow-up, put a limit on that neutrality. Current use of estrogen plus progesterone or dydrogesterone for more than five years carried about 31% higher risk (HR 1.31, 95% CI 1.15–1.48, mean duration 8.7 years), against a non-significant 1.13 for five years or less [21]. "At standard durations" is doing real work in the sentence above.
| Progestogen | Classification | Breast cancer signal | Current standing |
|---|---|---|---|
| Medroxyprogesterone acetate | Synthetic | Elevated; the driver of the WHI breast signal | Largely deprecated in modern practice |
| Norethisterone | Synthetic | Elevated | Used selectively; avoided when breast risk is prioritized |
| Micronized progesterone | Body-identical | Neutral / negligible at standard durations | First-line for endometrial protection; also improves sleep |
| Dydrogesterone | Retroprogesterone | Neutral / negligible | Comparable safety to micronized progesterone where available |
The modern gold standard for a woman with an intact uterus is transdermal estradiol plus oral micronized progesterone — minimizing thrombotic and breast risks simultaneously.[22]
Cardiometabolic effects (Weak / preliminary)
The menopausal transition redistributes adipose tissue centrally, reduces lean mass, and raises the risk of metabolic syndrome — and the earlier the transition, the worse the downstream cardiometabolic trajectory.[23] Guidelines describe MHT initiated in the early window as blunting visceral fat accumulation and improving glucose homeostasis, but no trial effect size for those outcomes is presented here, so the claim is graded Weak; the proposed pathways overlap with the metabolic-rescue mechanisms identified by other geroprotector classes (GLP-1, geroprotectors).
The Wilder pattern repeats — the rule that a treatment helps roughly in proportion to how far the patient has already drifted from a healthy baseline. Any cardiometabolic benefit would be expected to concentrate in women whose risk has begun to move in the wrong direction. In a metabolically healthy 50-year-old, MHT's net cardiometabolic effect is modest. In one with rising visceral adiposity, deteriorating insulin sensitivity, or worsening lipids, the effect may be larger — a plausible pattern, not a measured one.
Bones — and the post-cessation rebound (Strong)
MHT is among the most effective osteoporosis-prevention interventions available. The largest analysis to date — a UK primary-care nested case-control study matching 648,747 women with a first fracture to 2,357,125 fracture-free controls — found roughly a 25% reduction in fracture odds during current use (estrogen-only OR 0.76, 95% CI 0.74–0.78; estrogen-progestogen OR 0.75, 0.73–0.76).[24] The randomised evidence agrees, which is what earns this section its rating: pooling 24 trials in 45,660 participants, combined therapy cut all clinical fractures by about 22% (RR 0.78, 95% CI 0.71–0.86) and estrogen alone by about 27% (RR 0.73, 0.65–0.80), both at moderate certainty [25]. The caveat is population: roughly 70% of that data comes from HERS and the WHI, meaning older American women on oral conjugated estrogens — not the early-window transdermal regimen this page recommends.
The age question cuts against the timing narrative for bone. A 2001 meta-analysis of 22 small trials found about a quarter fewer non-vertebral fractures overall (RR 0.73), with a larger effect in trials whose participants averaged under 60 (RR 0.67).[26] But the largest randomised dataset since then found the benefit the same at every age: 8.6% of women on estrogen-plus-progestin fractured over 5.6 years against 11.1% on placebo, about 2.5 fewer fractures per hundred women, and "the effect did not differ in women stratified by age." Across all outcomes in that trial, however, there was no net benefit even in women at high fracture risk, so MHT is not a bone-only prescription.[27] The timing hypothesis has somewhat better (still contested) support for cardiovascular outcomes; for fractures it rests on subgroups of underpowered trials. The 2022 position statement of The North American Menopause Society (renamed The Menopause Society in 2024) affirms hormone therapy for prevention of bone loss and fracture in appropriate candidates.
The discontinuation trajectory is the part that matters clinically. Protection does not persist smoothly: in the 1–10 years after stopping, fracture odds climb back to — and for estrogen-progestogen slightly above — never-user levels (OR 1.06, 1.05–1.08), an excess concentrated in women who used MHT for under 5 years and equivalent to roughly 14 extra fractures per 10,000 woman-years. Beyond 10 years post-cessation the odds fall below never-use again (OR 0.93–0.95). The microarchitectural gains banked during treatment do translate into a durable long-run benefit — but the cessation window itself is a real vulnerability requiring either a slow taper or bridging anti-resorptive therapy. See bone density for the LIFTMOR-style resistance protocol that complements MHT.
Muscle and physical vitality (Weak / preliminary)
Estrogen receptors on skeletal muscle directly modulate protein synthesis, satellite-cell activation, and oxidative-stress buffering. After menopause, muscle mass declines by roughly 0.6% a year [28]. How much of that is estrogen withdrawal rather than ordinary aging is unresolved — the systematic reviews say so explicitly rather than apportioning it.
The honest position is that the muscle case for MHT is mechanistic, not quantitative. A systematic review of 32 studies found that postmenopausal women do have less muscle than premenopausal women, but could not separate estrogen withdrawal from ordinary chronological aging, and rated the evidence on whether estrogen therapy preserves muscle size or strength "largely conflicting and inconclusive" [29]. No pooled effect size for MHT on muscle mass, grip strength, or gait speed currently exists.
What does have numbers is the thing that works regardless of hormones. Pooling 15 randomised trials in 2,491 older adults with diagnosed sarcopenia, structured exercise plus nutritional supplementation — against a control group carrying on with usual daily activity — added about 0.22 kg/m² of skeletal muscle mass index, roughly 1.8 kg of grip strength, and 0.09 m/s of gait speed [30]. That trial set excluded drug interventions altogether, so it says nothing about hormone therapy — but it is the closest thing to a quantified answer anyone has, and it points at resistance training and protein rather than at a prescription.
A separate 2025 cross-sectional analysis showed physical inactivity during the transition is the single strongest predictor of functional decline — nearly four times the odds of slow gait speed in inactive vs active women (odds ratio 3.93), with type 2 diabetes and obesity multiplying the sarcopenia risk independently.[31] MHT is a scaffold, not a substitute for training and protein.
Brain: cognition, mood, sleep (Moderate)
The neurology version of the timing hypothesis is the cleanest. The WHI Memory Study (WHIMS), reported in 2003, famously doubled dementia risk when oral CEE plus MPA was started in women aged 65 and older — 40 cases against 21 on placebo, or 45 versus 22 per 10,000 women a year (HR 2.05, 95% CI 1.21–3.48) — a finding that defined a generation of geriatric prescribing [32]. Estrogen alone, in the same programme, did not significantly raise it (28 cases vs 19; HR 1.49, 0.83–2.66).
Two prospective trials in early initiators tell a very different story:
- KEEPS — 4 years of transdermal or oral estrogen, both with cyclic micronized progesterone, started within 3 years of menopause. Ten years on, cognition was no different from placebo in either direction — no lasting harm, and no protective benefit [33]. The imaging arm, reported separately in 266 women, found the same: amyloid on PET and hippocampal and prefrontal measures on MRI were unchanged versus placebo, and carrying the APOE ε4 risk variant made no difference [34].
- ELITE — confirmed the timing-stratified vascular response, with neutral cognitive findings in early initiators.[35]
The honest summary: MHT initiated early does not harm long-term cognition; it does not, on current evidence, independently preserve it either. The exception is premature menopause — surgical or natural ovarian failure before age 45, where prolonged hypoestrogenism is itself a dementia risk factor and immediate hormonal replacement is indicated.[36]
One observational study cuts against all of this and is worth knowing about, because it was reported widely. Danish national registries covering 5,589 dementia cases and 55,890 matched controls found estrogen-progestogen use associated with about 24% higher dementia rates (HR 1.24, 95% CI 1.17–1.33), rising with duration and present even in women who started at 55 or younger [37]. The objection is confounding by indication — women with early cognitive or vasomotor symptoms are more likely to be prescribed hormones — and the randomised biomarker data above point the other way. It is a signal to hold in mind, not a finding to act on.
Mood is where MHT delivers fast, consistently measurable benefit. The European Society of Endocrinology (ESE) explicitly recognizes perimenopause as an independent risk factor for depressive symptoms and notes MHT effective for the transition-associated mood disturbance specifically.[38] The sleep evidence is thinner than the mood evidence. In a single-centre series of 342 women followed on hormone therapy, Pittsburgh Sleep Quality Index scores fell by about a quarter from their own starting point and were still down three years later — but there was no untreated comparison group, and only 79 women reached the three-year mark, so this cannot separate the drug from the passage of time [39]. The plausible mechanism is simply that treating night sweats stops them waking you — see insomnia treatment, and screen for sleep-disordered breathing before attributing all sleep symptoms to menopause.
Other cancers (Moderate)
Ovarian cancer. Current or recent MHT use carries a small absolute increase in ovarian cancer risk, and the increase appears even with under five years of use — about 40% higher relative risk in the pooled epidemiology, concentrated in the two commonest tumour types, serous and endometrioid. In absolute terms, five years of use starting around age 50 works out at roughly one extra ovarian cancer per 1,000 users and one extra ovarian cancer death per 1,700 users, modelled on English incidence and death rates; ten years of use works out at about one extra ovarian cancer per 600 users and one extra death per 800 [40]. Risk falls the longer ago use stopped, though about ten years after stopping long-duration therapy there was still an excess of those two tumour types (RR 1.25, 95% CI 1.07–1.46). In women already diagnosed with non-serous epithelial ovarian cancer, the limited evidence suggests MHT for symptom control does not shorten overall or disease-free survival. The restriction matters: serous and granulosa-cell tumours are hormone-dependent, and non-hormonal options are recommended first there [41].
Endometrial cancer. Preventing this is the entire reason EPT exists, and continuous combined progestogen does appear to bring the risk of endometrial hyperplasia back down to roughly placebo levels — about 0–16 cases per 1,000 women a year, against 5 per 1,000 on placebo and 22–43 per 1,000 on unopposed estrogen — though Cochrane rates that comparison low-certainty, with an interval (OR 0.51, 95% CI 0.08–3.38) still compatible with a threefold increase. Two qualifications matter clinically. Sequential (cyclic) regimens do not achieve this: at one year they carried roughly 5.5 times the odds of hyperplasia versus placebo (OR 5.53, 95% CI 2.60–11.76). And for endometrial carcinoma the trials have never accumulated enough cases to answer the question at all — the current Cochrane review of 72 trials in 40,652 women says outright that events were too few to draw conclusions [42]. Protection is inferred from the hyperplasia surrogate and from observational data, not demonstrated against cancer.
Colorectal cancer. Observational signals are modestly inverse (suggesting protection), but the randomized evidence is mixed; the effect, if real, is small.
Epigenetic biomarkers of aging (Weak / preliminary)
The menopausal transition accelerates DNA methylation–based aging, as the cohort pooling above shows [43]. The clocks used to measure it are not equal. Second-generation clocks were trained on clinical biomarkers and mortality data rather than on chronological age, and in a head-to-head test across nine endpoints including all-cause mortality, in 490 adults, GrimAge was the one that held up — PhenoAge was associated with none of the nine once social and lifestyle factors were adjusted for, and the first-generation clocks (Horvath, Hannum) predicted health outcomes not at all [44]. The mechanistic argument that MHT decelerates this acceleration is plausible — estrogen preserves telomerase activity, dampens oxidative stress, and maintains mitochondrial DNA copy number — but the direct RCT evidence that MHT slows specific epigenetic clocks in healthy adult women is still preliminary; large longitudinal cohorts are running.
For now, treat this as a mechanism-rich, hard-outcome-pending claim — similar to where rapamycin and metformin sit on the geroprotector ladder.
Estetrol — a next-generation estrogen (Moderate in the EU; Emerging elsewhere)
Estetrol (E4) is a naturally occurring fetal estrogen produced exclusively by the human fetal liver during pregnancy. Its developers propose it as a distinct class — a native estrogen with selective action in tissues. What the evidence currently supports is narrower: estrogenic activity in vaginal tissue, bone and the central nervous system with comparatively limited effect on the liver, including on clotting factors, and early breast samples that look reassuring rather than conclusive [45]. The favourable vascular actions are so far preclinical; no cardiovascular outcome data in postmenopausal women exist yet. The European Commission approved estetrol (as FYLREVY, 14.2 mg and 18.9 mg, with a progestogen prescribed separately for women with a uterus) for menopausal symptoms in March 2026; it is not approved in the US [46].
The phase 3 E4COMFORT I trial randomised 640 postmenopausal women aged 40–65 to estetrol 15 mg, estetrol 20 mg, or placebo for 12 weeks:[47]
- Both doses significantly reduced the weekly frequency and the severity of moderate-to-severe vasomotor symptoms at weeks 4 and 12 versus placebo; 15 mg was the lower of the two effective doses.
- Drug-related adverse events occurred in 51.6% (15 mg), 57.3% (20 mg), and 20.6% (placebo), with discontinuation for adverse events in 6.6%, 7.5%, and 2.3% respectively — few women stopped the drug, but it is not side-effect free.
One phase 2 trial, E4Relief, carries the metabolic and bone picture: in 180 women, estetrol produced minimal change in triglycerides, a rise in HDL cholesterol, no change in fasting glucose, significant falls in insulin resistance and HbA1c at the middle and upper doses, and a fall in the bone-resorption marker CTX-1 [48], [49]. Both were authored by the developer. The separate claim of no clinically meaningful change in blood pressure, including in women with pre-existing cardiovascular risk factors, was presented at a 2025 conference and is not yet peer-reviewed. Multi-decade thrombotic safety data in postmenopausal women are still pending, but the proposed tissue-selective mechanism, if it holds up, would be the most promising effort in years to deliver oral estrogen with a transdermal-like safety profile.
Non-hormonal alternatives (Moderate)
For women with hard contraindications to estrogen — hormone-receptor-positive breast cancer history, active thromboembolic disease, unexplained vaginal bleeding, severe uncontrolled cardiovascular disease — the modern non-hormonal options have moved well beyond selective serotonin reuptake inhibitors (SSRIs).
Neurokinin receptor antagonists — fezolinetant (NK3) and elinzanetant (dual NK1/NK3) — act on the hypothalamic temperature-control neurons that trigger hot flashes, with no hormonal exposure. Both cut hot-flash frequency by roughly 60–65% over 12 weeks, but the placebo arms of the same trials fell 37–43%, leaving the drugs' own contribution at roughly 25 percentage points [50], [51].
They are not interchangeable on safety. Fezolinetant has carried a US boxed warning for serious liver injury since December 2024, requiring liver tests before starting, monthly for the first three months, and again at months 6 and 9; it must not be started in anyone whose liver enzymes or bilirubin are already at twice the upper limit of normal [52]. The event is described as rare and no incidence rate has been published. Elinzanetant, approved in the US in October 2025 and in the EU a month later, carries no boxed warning — but its label also requires liver tests before starting and one recheck at three months, and bars initiation at the same twice-the-upper-limit threshold: one follow-up test against fezolinetant's five. This is a drug class for someone who needs it, not a casual choice.
SSRIs and SNRIs (serotonin-norepinephrine reuptake inhibitors) — paroxetine, escitalopram, venlafaxine — cut hot flashes by 40–60% on paper [53], but most of that is the placebo response these trials generate. Pooling six randomised trials of paroxetine in 1,486 women, 79% of the improvement in hot-flash frequency was matched in the placebo arms, leaving at most a fifth of the effect attributable to the drug [54]. The mood benefit is real and separate; the vasomotor benefit is smaller than the headline number, and smaller than MHT's.
Mind-body therapies earned a stronger evidence base than many clinicians realize. A 2025 systematic review and meta-analysis of RCTs in >1,500 menopausal women found that mind-body interventions (yoga, mindfulness, qigong, music, dance) significantly improved sleep quality, depression, and anxiety — all large effects (standardized mean differences of −0.86, −0.79, and −1.13, where roughly −0.8 counts as large).[55] These effect sizes are larger than for most pharmacological mood interventions in the same population, though the trials are small and short.
Skin and the genitourinary syndrome (Moderate for skin; Strong for the genitourinary syndrome)
Estrogen drives synthesis and turnover of dermal collagen, elastin, and hyaluronic acid. Postmenopausal hypoestrogenism produces dryness, thinning, wrinkling, and epidermal barrier degradation. A meta-analysis of 15 trials in 1,589 women aged 45–55 found large effects on the structural measures and much smaller ones on how the skin behaves: dermal collagen content improved by a standardized mean difference of 2.01 and skin thickness by 1.27 — both well above the ~0.8 "large" threshold — while skin elasticity moved only 0.28 (95% CI 0.03–0.54), a small effect that barely clears statistical significance, and skin dryness did not move at all [56]. The comparator in these trials was placebo or another treatment, and the trials are small.
For the genitourinary syndrome (vaginal dryness, dyspareunia, recurrent UTIs, overactive bladder), local low-dose vaginal estrogen and intravaginal DHEA (dehydroepiandrosterone; sold as prasterone) are the highest-leverage interventions in all of menopausal medicine. In the placebo-controlled registration trial of daily intravaginal prasterone 6.5 mg, 12 weeks of treatment significantly improved vaginal cell maturation, lowered vaginal pH, and reduced the severity of dyspareunia and dryness versus placebo.[57] Histology series report restored epithelial thickness plus increased elastin and nerve-ending density in the lamina propria, though those are small uncontrolled samples and the widely quoted percentage gains in nerve-fibre density do not come from a controlled trial. Systemic absorption is minimal; the safety profile is permissive even in many women with contraindications to systemic MHT, including selected breast-cancer survivors with oncology co-management.
All-cause mortality and continuation beyond age 65 (Moderate)
A nationwide Danish register-based cohort of more than 800,000 women (over 104,000 MHT users) followed through 2023 found no increase in all-cause mortality — in fact a small reduction (adjusted hazard ratio 0.96), concentrated in mid-duration users.[58] That aligns with the trial meta-analysis above, where the early-initiation subgroup showed roughly 30% lower mortality (RR 0.70) even though MHT overall was mortality-neutral. The 2024 update of the NICE menopause guideline reaches the same conclusion — combined HRT does not raise cardiovascular or overall mortality.[59] The important caveat: this registry cohort was born between 1950 and 1977, so most participants were still only in their fifties at the close of follow-up — the mortality window is too early for a definitive lifetime conclusion, and the effect grade here is best read as Moderate rather than settled. Over 18 years of WHI follow-up, all-cause mortality did not move in either arm — estrogen alone gave a hazard ratio of 0.94, a 6% difference well inside the range expected from chance (95% CI 0.88–1.01). What moved were individual causes of death: breast cancer, and deaths coded to Alzheimer's disease or other dementia, which were about a quarter lower on estrogen alone than on placebo (2.4% vs 3.2%; HR 0.74, 95% CI 0.59–0.94) [60]. Read that dementia result carefully: it is a cause-of-death classification rather than an adjudicated diagnosis, it sits inside a composite "other mortality" category, and the investigators themselves set it against the WHI Memory Study's finding of more incident dementia on hormone therapy (significant for estrogen plus progestin, not for estrogen alone).
Larger and more dramatic figures circulate — a 2025 Menopause Society conference presentation on more than 120 million health records was widely reported as showing roughly 60% lower odds of breast cancer, heart attack, or stroke in long-term early initiators. That analysis has not been peer-reviewed or published, so it is not treated as evidence here. Healthy-user bias in particular pervades this literature: women prescribed and retained on MHT tend to be healthier, wealthier, and more engaged with preventive care than never-users, and no amount of adjustment removes that.
Regulators have moved on the same reappraisal. In November 2025 the FDA announced it would strip the class-wide boxed warnings from menopausal hormone therapy, and in February 2026 it approved the resulting labeling changes across all four product categories — removing the cardiovascular-disease, breast-cancer and probable-dementia risk statements, and keeping only the endometrial-cancer boxed warning on systemic estrogen-alone products [61]. That is a labelling decision, not new evidence: the trials it reinterprets are the same ones described above, the agency's own press materials quote benefit percentages that are not traceable to any specific trial, and the Menopause Society's response was markedly more cautious than the announcement. What changes is the warning a reader will see on the box — not what the drug does.
Use after age 65. Up to 40% of women in their 60s and ~15% in their 70s still experience hot flashes; an unpublished 2024 conference abstract from a single menopause clinic — just over 100 women aged 65 and up, mean age 71, nearly 8% over 80 — found that only about a quarter had ever tried stopping, and of those two dozen, 87% restarted because hot flashes came back.[62] The Menopause Society's current position is explicit: there is no universal age limit for continuation, provided lower-dose transdermal regimens are used and surveillance is active.[63] Initiation after 65 is a different and riskier proposition than continuation of long-established therapy [64].
Diagnostic and monitoring framework (Strong on protocol)
Before initiation:
- History. Age at menopause, time since final menstrual period; personal and family history of breast, ovarian, and endometrial cancer; prior VTE, stroke, CHD, migraine with aura, liver disease; known prothrombotic mutations (e.g., factor V Leiden) — ask, but do not test routinely before starting.
- Examination. Blood pressure, breast and pelvic exam, BMI and waist circumference.
- Laboratory. Lipid panel including apoB (apolipoprotein B, the count of atherogenic particles); HbA1c and fasting glucose; liver and renal function; thyroid-stimulating hormone (TSH). Follicle-stimulating hormone (FSH) and estradiol only when diagnosis is uncertain (e.g., perimenopausal symptoms with retained cycles).
- Imaging. Mammography per age-appropriate cadence; a DXA (dual-energy X-ray absorptiometry) bone-density scan in any woman with risk factors for early bone loss; pelvic ultrasound if abnormal bleeding.
Formulation choice:
| Component | Preferred | Acceptable alternative | Avoid where possible |
|---|---|---|---|
| Estrogen | Transdermal estradiol (patch, gel, spray) | Oral estradiol in low cardiometabolic risk | Conjugated equine estrogens (largely historical) |
| Progestogen (intact uterus) | Oral micronized progesterone 100–200 mg | Dydrogesterone | MPA, norethisterone |
| Local GSM | Vaginal estrogen, intravaginal DHEA (prasterone) | Non-hormonal moisturizers | — |
| Newer oral estrogen | Estetrol (EU only: 14.2 mg or 18.9 mg), with a progestogen prescribed separately if the uterus is intact | — | Not FDA-approved for menopause |
Targets and follow-up:
- Lowest effective dose for symptom control; titrate downward over time as tolerated.
- Recheck symptoms, BP, and weight at 3 months, then every 6–12 months.
- Annual breast exam plus mammography per age-based cadence; DXA every 2–3 years if at risk.
- No mandatory stop date. Annual shared-decision review of continuation, with documented assessment of ongoing benefit vs incremental risk.
What's overrated
- The WHI-era "all hormone therapy is dangerous" frame. Generalizing oral CEE plus MPA in women a decade past menopause to transdermal estradiol plus micronized progesterone in a 51-year-old is a categorical error. The modern literature has done the disaggregation; the cultural memory hasn't caught up.
- MHT as primary cognitive prevention. KEEPS Continuation is honest about this — no harm, no independent benefit, when started early. The levers that actually move dementia risk — treating hearing loss (association, not yet proven causal), blood pressure control, exercise, sleep — outperform on absolute terms (dementia prevention).
- Compounded "bioidentical" hormone cocktails. "Body-identical" estradiol and micronized progesterone are available as licensed, regulated, evidence-based products. The compounded pellet / cream / lozenge industry adds dose variability, lacks endometrial-protection guarantees, and circulates a marketing claim that doesn't survive contact with the pharmacology.
- Stopping at an arbitrary age. There is no evidence-based age cutoff for continuation in a stable, symptom-driven regimen under active surveillance.
Cautions
- Late initiation in the late-60s and beyond. Increased short-term MI, stroke, and VTE risk in the first 1–2 years. Initiation after 60–65 needs unambiguous indication and full informed consent; this is not the same as continuing existing therapy.
- Oral estrogen in cardiometabolic risk. Hepatic first-pass raises clotting factor synthesis, CRP, and triglycerides. Prefer transdermal with cardiovascular risk factors — obesity, well-controlled diabetes or hypertension, high triglycerides, migraine with aura. Prior VTE needs an individual risk-benefit assessment before any hormone therapy, not just a route switch.
- Synthetic progestins. MPA and norethisterone carry a meaningful breast cancer signal vs micronized progesterone or dydrogesterone; the modern default should be the latter.
- MHT and ovarian cancer. Small but real absolute increase during current or recent use, present even with under five years of use and larger with longer use; revisit the indication annually.
- Hard contraindications. Personal history of hormone-receptor-positive breast cancer, active estrogen-dependent malignancy, active thromboembolic disease, unexplained vaginal bleeding, severe uncontrolled cardiovascular disease, severe active liver disease.
- The Wilder asymmetry. As with TRT and GLP-1s, the biggest absolute benefits probably concentrate in women whose underlying trajectory has begun to deteriorate — a plausible pattern, not a measured one. In a perfectly healthy 50-year-old without symptoms, the net longevity case is more modest than enthusiastic communicators suggest.
Further reading
- KEEPS Continuation — long-term cognitive effects of early MHT.[65]
- Menopausal hormone therapy — risks, benefits, emerging options. Int J Mol Sci 2025 narrative review.[66]
- 2025 MHT guidelines — Korean Society of Menopause.[67]
- ESE clinical practice guideline for menopause and perimenopause management. Eur J Endocrinol 2025.[68]
- Mehta JM, Chester RC, Kling JM. The Timing Hypothesis: Hormone Therapy for Treating Symptomatic Women During Menopause and Its Relationship to Cardiovascular Disease. J Womens Health 2019.[69]
- Manson JE et al. The Women's Health Initiative Randomized Trials and Clinical Practice: A Review. JAMA 2024.[70]
- Benkhadra K et al. Menopausal Hormone Therapy and Mortality: A Systematic Review and Meta-Analysis. J Clin Endocrinol Metab 2015.[71]
- MHT and long-term mortality, nationwide register-based cohort study. BMJ 2026.[72]
- Menopause: identification and management — NG23, 2024 update.[73]
- Vinogradova Y et al. Discontinuation of menopausal hormone therapy and risk of fracture: nested case-control studies using routinely collected primary care data. Lancet Healthy Longev 2025.[74]
- Torgerson DJ, Bell-Syer SEM. Hormone replacement therapy and prevention of nonvertebral fractures — meta-analysis of randomized trials. JAMA 2001.[75]
- Cauley JA et al. Effects of estrogen plus progestin on risk of fracture and bone mineral density: the Women's Health Initiative randomized trial. JAMA 2003.[76]
- Fournier A et al. Unequal risks for breast cancer associated with different hormone replacement therapies: results from the E3N cohort study. Breast Cancer Res Treat 2008.[77]
- Muka T et al. Association of Age at Onset of Menopause and Time Since Onset of Menopause With Cardiovascular Outcomes, Intermediate Vascular Traits, and All-Cause Mortality. JAMA Cardiol 2016.[78]
- Levine ME et al. Menopause accelerates biological aging. PNAS 2016.[79]
- Labrie F et al. Efficacy of intravaginal dehydroepiandrosterone (DHEA) on moderate to severe dyspareunia and vaginal dryness. Menopause 2016.[80]
- Exercise plus nutrition for sarcopenia — meta-analysis of 15 RCTs in older adults (n = 2,491); no hormone-therapy arm.[81]
- Mind-body therapies for menopausal sleep, depression, and anxiety.[82]
- Pivazyan L et al. Skin Rejuvenation in Women using Menopausal Hormone Therapy: A Systematic Review and Meta-Analysis. J Menopausal Med 2023.[83]
- E4COMFORT I — estetrol for moderate-to-severe vasomotor symptoms, phase 3 placebo-controlled trial. Maturitas 2026.[84]
- Estradiol and micronized progesterone — narrative review.[85]
- Menopause Society — no universal age limit for HT continuation.[86]
- McCrory C et al. GrimAge outperforms other epigenetic clocks for age-related clinical phenotypes and mortality. J Gerontol A Biol Sci Med Sci 2021.[87]