HRT Side Effects Overview: What the Evidence Actually Shows

Why Formulation and Route Change Everything

The side effects of HRT are not uniform across all products. Oral estrogen, transdermal estrogen patches, vaginal estradiol, synthetic progestogens, and micronized progesterone each carry a distinct risk footprint. Treating them as a single category is responsible for most of the confusion that followed the 2002 WHI publication.

The Women's Health Initiative (WHI) enrolled 16,608 postmenopausal women with a mean age of 63 and randomized them to conjugated equine estrogen (CEE) 0.625 mg plus medroxyprogesterone acetate (MPA) 2.5 mg or placebo. [1] The combination arm was stopped early at 5.2 years because of a statistically significant increase in invasive breast cancer (hazard ratio 1.26 to 95% CI 1.00 to 1.59). Cardiovascular and VTE signals also emerged. The estrogen-alone arm (women with prior hysterectomy) was stopped at 6.8 years for stroke risk, but that arm showed no increase in breast cancer and a possible reduction. [2]

Those two arms behaved very differently. Conflating them is a persistent problem in popular media coverage of hormone therapy.

Route matters as much as molecule. A 2007 nested case-control study in the British Medical Journal (N=30,099) showed that transdermal estradiol was not associated with increased VTE risk (adjusted odds ratio 0.9 to 95% CI 0.6 to 1.5), while oral estrogen carried an odds ratio of 3.5 (95% CI 2.8 to 4.3). [3] That distinction never made the same headlines the WHI did.

Breast Cancer Risk: Progestogen Type Is the Central Variable

The breast cancer question is the one most women ask first. The answer depends on which hormone combination is prescribed and for how long.

The WHI EPT arm (CEE plus MPA) produced an absolute excess of 8 extra breast cancer cases per 10,000 women per year after 5 years of use. [1] That is the real number behind the headlines. For context, drinking one alcoholic drink per day raises breast cancer risk by a similar magnitude.

Estrogen alone tells a different story. In the WHI estrogen-only arm, CEE alone in women without a uterus was associated with a hazard ratio of 0.77 (95% CI 0.59 to 1.01) for invasive breast cancer, suggesting possible risk reduction rather than elevation. [2] A 2022 UK Biobank analysis (N=97,376 women) published in PLOS Medicine confirmed estrogen-only therapy was not associated with increased breast cancer incidence. [4]

The progestogen type matters specifically. The E3N cohort study (N=54,548 French women) compared EPT regimens and found that estrogen combined with synthetic progestins raised breast cancer risk (relative risk 1.69 for CEE plus synthetic progestin), while estrogen combined with micronized progesterone did not significantly raise risk (relative risk 1.00 to 95% CI 0.83 to 1.22). [5] Micronized progesterone (the active ingredient in Prometrium and Utrogestan) appears to have a more favorable breast tissue profile than medroxyprogesterone acetate or norethisterone.

The Menopause Society (formerly NAMS) states in its 2022 position statement: "Hormone therapy remains the most effective treatment for vasomotor symptoms and has been shown to prevent bone loss and fracture. The risks of HT differ for younger versus older women." [6] Individualized risk assessment, not population-level avoidance, is the recommended approach.

Women with a uterus who need progestogen coverage should ask their prescriber specifically about micronized progesterone. Women with a prior hysterectomy can use estrogen alone, which carries no documented breast cancer excess.

VTE and Stroke Risk: Why Transdermal Estrogen Changes the Calculus

Venous thromboembolism (VTE, meaning deep vein thrombosis or pulmonary embolism) and ischemic stroke are the two cardiovascular side effects most relevant to choosing an HRT route.

Oral estrogen undergoes first-pass hepatic metabolism, which raises clotting factors and C-reactive protein. Transdermal estrogen bypasses that mechanism entirely. The ESTHER study (N=881 VTE cases, 1,452 controls) showed an adjusted odds ratio for VTE of 4.2 (95% CI 1.5 to 11.6) with oral estrogens versus 0.9 (95% CI 0.4 to 2.1) for transdermal estrogens. [7] In plain terms, transdermal estradiol did not raise VTE risk above baseline in that study.

Stroke risk follows a similar pattern. A 2012 BMJ meta-analysis of 28 trials and observational studies found oral estrogen increased ischemic stroke risk by about 32% (relative risk 1.32 to 95% CI 1.14 to 1.53), while transdermal estrogen showed a relative risk of 0.95 (95% CI 0.75 to 1.20), which is statistically neutral. [8]

In absolute terms, the WHI reported approximately 8 extra strokes per 10,000 women per year in the CEE plus MPA arm among women aged 60 to 69. For women aged 50 to 59, no statistically significant stroke increase was observed. [1] Age and baseline cardiovascular risk, not hormone therapy alone, determine who actually faces meaningful elevated stroke risk.

Women who carry Factor V Leiden, prothrombin G20210A, or have a personal or first-degree family history of VTE should be considered for transdermal estrogen specifically, as the neutral VTE profile makes the risk-benefit calculus considerably more favorable.

HRT and Dementia Risk: The Timing Hypothesis

Cognitive outcomes with HRT depend almost entirely on when therapy starts relative to menopause. This is called the critical window hypothesis and it is the reason the WHI Memory Study findings cannot be applied to perimenopausal women starting HRT in their late 40s or early 50s.

The WHI Memory Study (WHIMS) enrolled women aged 65 to 79 and found that CEE plus MPA increased the risk of probable dementia (hazard ratio 2.05 to 95% CI 1.21 to 3.48) and that CEE alone showed a neutral effect. [9] These were women starting hormone therapy an average of 13 years after menopause. Using these results to counsel a 49-year-old perimenopausal woman is a direct misapplication of the data.

Observational data from the Cache County Study (N=1,357 women) found that women who initiated HRT near menopause and continued for 10 or more years had a significantly lower risk of Alzheimer's disease (relative risk 0.59 to 95% CI 0.36 to 0.96). [10] A 2022 systematic review in Maturitas covering 51 studies concluded that initiation of HRT before age 60 or within 10 years of menopause was associated with reduced dementia risk, while initiation after age 65 was associated with increased risk. [11]

The clinical implication is direct: HRT started in the perimenopausal window or within 5 to 10 years of the final menstrual period does not appear to raise dementia risk and may reduce it. Starting HRT for the first time at age 65 or older, particularly with older formulations, carries a different risk profile and is not generally recommended outside of specialist care for specific indications.

Surgical menopause before age 45 is a separate category. Women who undergo bilateral oophorectomy before natural menopause and do not receive estrogen replacement have a doubled risk of dementia compared to women with intact ovaries, according to a Mayo Clinic cohort study (N=1,837). [12] For these women, HRT until at least the age of natural menopause (approximately 51) is standard of care in most guidelines.

Gallbladder Disease: The Overlooked Side Effect

Gallbladder disease is one of the most consistently documented HRT side effects and one of the least discussed in popular coverage of hormone therapy risks.

Oral estrogen increases biliary cholesterol saturation, reduces bile acid synthesis, and slows gallbladder emptying, all of which promote gallstone formation. The WHI reported a hazard ratio of 1.67 (95% CI 1.45 to 1.91) for cholecystitis requiring surgery in women taking CEE plus MPA, representing approximately 55 additional gallbladder surgeries per 10,000 women over 5 years. [13] That number is larger in absolute terms than the breast cancer excess in the same trial.

Transdermal estrogen again shows a more favorable profile. A nested case-control study in Pharmacoepidemiology and Drug Safety (N=7,672 gallbladder disease cases) found that transdermal estrogen was associated with a significantly lower risk than oral estrogen (odds ratio 1.17 for transdermal vs. 1.74 for oral). [14] The transdermal route still carries a small increase, but the magnitude is materially lower.

Women with a history of gallstones, prior cholecystitis, or biliary colic should discuss transdermal or vaginal estrogen as a first-line route with their prescriber. Oral estrogen is not contraindicated in women with intact gallbladders and no prior disease, but the gallbladder signal is a legitimate factor in route selection.

Common Early Side Effects: What to Expect in the First 12 Weeks

Beyond the major long-term risk categories above, most women encounter a set of early adaptation side effects in the first 8 to 12 weeks of HRT. These are dose-dependent and typically self-limiting.

Breast tenderness is the most frequently reported early symptom, occurring in roughly 40% of women starting combined estrogen-progestogen therapy. It usually resolves within 6 to 8 weeks. Reducing the estrogen dose or switching from synthetic progestogen to micronized progesterone often eliminates it.

Nausea occurs in approximately 15 to 20% of women starting oral estrogen. Taking oral estradiol with food or switching to a transdermal patch (such as Vivelle-Dot, delivering 0.05 mg/day) eliminates first-pass exposure and resolves nausea in most cases.

Irregular or breakthrough bleeding is common in the first 3 to 6 months of combined continuous HRT (meaning estrogen and progestogen taken daily without a break). The endometrium takes time to suppress. Bleeding that persists beyond 6 months, or that is heavy, should be evaluated with a pelvic ultrasound or endometrial biopsy to exclude hyperplasia.

Bloating and fluid retention are reported more often with MPA than with micronized progesterone. The androgenic activity of some synthetic progestogens (norethisterone, MPA) can also cause acne and mood changes. Switching to a progestogen with a more neutral androgenic profile, such as dydrogesterone or micronized progesterone, generally resolves these symptoms.

Mood changes in the first few weeks, including low mood or anxiety, sometimes reflect progestogen sensitivity rather than underlying depression. A progestogen holiday of 2 to 4 weeks under clinical supervision can identify sensitivity, though this approach requires careful endometrial monitoring.

Who Should Not Start Standard HRT

Absolute contraindications to estrogen-containing HRT include a personal history of estrogen receptor-positive breast cancer, unexplained vaginal bleeding, active or recent arterial thromboembolic disease (myocardial infarction or stroke within the past 6 months), active VTE not on therapeutic anticoagulation, and known or suspected estrogen-sensitive malignancy other than breast cancer (such as endometrial carcinoma). [6]

Relative contraindications, where risk-benefit assessment is required, include a strong family history of VTE, controlled hypertension, active gallbladder disease, hypertriglyceridemia (serum triglycerides above 5 mmol/L represent a specific risk for pancreatitis with oral estrogen), migraine with aura, and liver disease with impaired synthetic function.

The BRCA1 or BRCA2 carrier who has undergone risk-reducing bilateral salpingo-oophorectomy before natural menopause is a population where HRT is generally considered protective rather than harmful. The Menopause Society's guidance explicitly notes that for BRCA carriers who have had prophylactic oophorectomy, HRT until the natural age of menopause is appropriate. [6]

How to Read HRT Risk Numbers Without Misleading Yourself

The single most common error in interpreting HRT research is confusing relative and absolute risk. A 26% increase in relative risk (the WHI breast cancer figure) sounds alarming. Eight extra cases per 10,000 women per year (the corresponding absolute figure) places the risk in clinical context.

For comparison, a 10-year smoking history increases breast cancer risk by approximately 6% in relative terms. Obesity (BMI above 30) raises postmenopausal breast cancer risk by 30 to 60% in relative terms, through endogenous estrogen production in adipose tissue. Neither of those exposures generates the same anxiety as HRT.

The NNH (number needed to harm) for breast cancer with EPT over 5 years, based on WHI data, is approximately 250. Meaning 250 women must take combined HRT for 5 years to produce one additional breast cancer case. The NNH for VTE with oral estrogen is lower (around 100 over 5 years), which is why transdermal delivery is preferred in women with any VTE risk factors.

These numbers do not mean HRT is riskless. They mean that risk must be weighed against benefit. Hot flash severity, sleep disruption, genitourinary syndrome of menopause, and osteoporosis prevention are meaningful clinical endpoints. A woman with severe vasomotor symptoms who declines HRT because of the breast cancer headline and then sustains a hip fracture at age 72 has also experienced a harm, one that was potentially preventable.

Prescribers using the HealthRX platform assess baseline 10-year cardiovascular risk (using the ACC/AHA Pooled Cohort Equations), FRAX fracture risk score, personal and family cancer history, current medications (particularly anticoagulants and CYP3A4 inducers), and route preference before initiating any hormone therapy regimen. Formulation decisions are revisited at 3 months and 12 months using a structured side effect checklist.