Metabolic health on long-term feminizing HRT¶
The metabolic literature for transfeminine adults past the redistribution window is thinner than the cardiovascular or bone literature, but the signal is consistent. Feminizing HRT shifts body composition gynoid in the first one to two years, modestly worsens insulin sensitivity, mildly improves the lipid panel, and — past year two — lets ordinary adult metabolic drift take over.
The dominant clinical risks at 2+ years aren't exotic estrogen effects. They are weight regain in newly-shaped depots, age- and lifestyle-driven T2DM, hepatic steatosis (especially with CPA or bicalutamide history), and the metabolic side effects of psychiatric medications.
Foundational data: Klaver 2017 (meta-analysis), Klaver 2018 (ENIGI prospective), Auer 2014 (lipid trajectory), Hembree 2017 (Endocrine Society guideline), and WPATH SOC-8 for screening cadence.
Two phases of body composition change¶
Phase one is redistribution, roughly months 0–24. Fat mass rises and shifts gynoid; lean mass falls. Klaver 2017 (n=171) found pooled fat mass +3.0 kg and lean body mass −2.4 kg at 12 months. Klaver 2018 (ENIGI, n=179 on CPA + estradiol) refined the regional picture:
| Region | Change at 12 months on HRT |
|---|---|
| Total body fat | +18% to +27% |
| Android (waist) | +18% (CI 13–23%) |
| Gynoid (hip) | +34% (CI 29–38%) |
| Leg | +42% (CI 37–46%) |
| Hip circumference | +3.2 cm (CI 2.3–4.0) |
| Waist-to-hip ratio | −0.03 (CI −0.04 to −0.02) |
Visceral adipose tissue (VAT) on modern 17β-estradiol regimens is small or unchanged — Klaver 2022 found the VAT/total-fat ratio decreased — in contrast with older ethinyl-estradiol literature.
Phase two is post-redistribution drift, year 2 onward. Trans-specific signals get progressively harder to separate from ordinary adult weight gain. The Wiepjes 2018 Amsterdam cohort (n=2,627) showed mean BMI rose from 24.0 at HRT start to 25.4 at three years, then continued upward at roughly the general-population rate (+0.2–0.3 kg/m² per year).
The clinically useful heuristic: what hasn't redistributed by year two probably won't, and weight gained after year two distributes more like a cis-female age-matched pattern than a continued feminization curve. Trans women who weight-cycled or gained heavily during phase one often see disproportionate persistence of pre-HRT android fat that resists later loss. See Weight Cycling for the evidence on deliberate cycling and its metabolic costs.
Adolescent-onset patients who blocked endogenous puberty achieve composition indistinguishable from cis-female norms (Klaver 2018, Boogers 2024). Their metabolic trajectories track cis-female aging, but this doesn't change the core monitoring schedule.
Insulin sensitivity and T2DM¶
Feminizing HRT modestly worsens insulin sensitivity, and the effect is largely attributable to anti-androgen choice rather than estradiol itself.
Estradiol's effect is mixed-to-favorable. Mauvais-Jarvis 2013 and the cis-female menopausal HRT literature show transdermal estradiol is either neutral or modestly insulin-sensitizing in non-diabetic adults. Oral estradiol's first-pass hepatic effects are less clean but still don't consistently raise diabetes incidence.
CPA is the dominant insulin-resistance signal. Auer 2014 (n=98, 1 year on transdermal estradiol + CPA) found HOMA-IR rose from 1.6 to 2.1 (+31%) and fasting insulin +22%, with fasting glucose unchanged — a pattern consistent with anti-androgen rather than estrogen effects. Gava 2020 (5-year RCT, CPA vs. leuprolide, n=80) found equivalent body-composition endpoints but CPA produced significantly higher fasting glucose and HOMA-IR across the trial, with several CPA participants crossing into prediabetic ranges. See Anti-androgens.
Spironolactone has minimal direct insulin-sensitivity effect at typical 100–200 mg doses, but its hyperkalemia and renal-handling effects complicate T2DM management when patients are also on ACE inhibitors or SGLT2 inhibitors. Bicalutamide has no consistent insulin-sensitivity signal — but the data are sparse.
A1c thresholds and screening cadence¶
Standard ADA thresholds apply. There is no published evidence that HRT distorts A1c interpretation in either direction at typical estradiol levels.
| A1c (%) | Fasting glucose (mg/dL) | Category |
|---|---|---|
| <5.7 | <100 | Normal |
| 5.7–6.4 | 100–125 | Prediabetes |
| ≥6.5 | ≥126 | Diabetes |
Hembree 2017 and SOC-8 recommend baseline fasting glucose or A1c plus periodic monitoring without specifying intervals. Fenway and UCSF protocols converge on:
| Patient profile | A1c cadence |
|---|---|
| No risk factors | Every 1–2 y |
| BMI ≥25, family history T2DM, prior fasting glucose ≥100, weight-active psych meds, oral E2 + CPA history | Annual |
| Established prediabetes, T2DM, MASLD | Per disease guideline |
CPA history persists past discontinuation
Long-term CPA users — even after switching off — should be screened more aggressively during the first year post-switch. Some of CPA's insulin-resistance signal does not immediately reverse on transition to spironolactone or post-orchidectomy estradiol monotherapy.
Hepatic monitoring¶
Liver effects cluster into three concerns: anti-androgen hepatotoxicity, estradiol's hepatic metabolism, and MASLD (formerly NAFLD).
Anti-androgen LFT cadence¶
| Drug | Baseline | Early | Maintenance | Trigger |
|---|---|---|---|---|
| Cyproterone acetate | Yes | Every HRT visit during use | + 3 mo and 12 mo post-discontinuation | ALT >3× ULN |
| Bicalutamide | Yes | 4 wk, 3 mo, 6 mo | Every 6 mo | ALT >3× ULN |
| Spironolactone | Yes | With renal/K+ panel | Annual | Per metabolic panel |
CPA at typical trans-care doses (10–25 mg) carries a real hepatotoxicity signal: case reports of fulminant hepatitis exist, and routine ALT/AST elevation is reported in roughly 5–10% of users. The 2020 EMA restriction (citing meningioma risk) reset much of European practice toward 10 mg or discontinuation after T-suppression is established.
Bicalutamide carries an FDA hepatotoxicity warning, with severe-hepatitis incidence around 1 in 4,000 in the prostate-cancer literature. Trans-care experience (Neal 2021, IC clinic case series) reports lower observed rates but with short follow-up. Asymptomatic ALT/AST rises ≤3× ULN can be observed;
3× warrants discontinuation. Bicalutamide-induced hepatitis is usually idiosyncratic and early (first 3–6 months) but late cases exist.
Estradiol metabolism and route choice¶
17β-estradiol is metabolized primarily via hepatic CYP3A4 and CYP1A2 to estrone and downstream conjugates, then glucuronidated and sulfated for renal/biliary excretion. Oral estradiol undergoes substantial first-pass metabolism, raising hepatic exposure to estrogen metabolites — which explains its larger effect on SHBG, hepatic-derived clotting factors, and triglycerides relative to transdermal or injected routes.
Route matters more with hepatic disease
In patients with established liver disease, transdermal or injected estradiol is preferred. CYP3A4 inducers (carbamazepine, rifampin, St. John's wort) can drop estradiol levels substantially; CYP3A4 inhibitors (azole antifungals, some HIV protease inhibitors) raise them.
MASLD screening¶
MASLD prevalence in cis adults is 25–30% and rises with metabolic-syndrome components. Trans-specific data are limited to small ultrasound cohorts (Maraka 2017; scattered ENIGI sub-analyses) suggesting prevalence broadly tracks BMI and metabolic-syndrome status rather than HRT exposure per se.
Annual ALT/AST in patients with BMI ≥25, T2DM, or dyslipidemia is reasonable. Persistent unexplained transaminase elevation should trigger FibroScan or ultrasound, not reflex HRT discontinuation. Estradiol does not consistently worsen MASLD and may modestly improve it (cis-female menopausal HRT literature).
Lipid panel and metabolic syndrome¶
Auer 2014 remains the cleanest lipid trajectory in trans women. At one year on transdermal estradiol + CPA:
| Marker | Change at 12 months |
|---|---|
| Total cholesterol | −5% |
| LDL | −7% |
| HDL | +3% |
| Triglycerides | +24% |
The triglyceride rise is route-sensitive — oral estradiol raises TGs more than transdermal — and stacks with metabolic-syndrome components. Wierckx 2014 and ENIGI lipid sub-analyses replicate the broad pattern: estrogen mildly improves the LDL/HDL ratio while raising TGs, with the net cardiovascular effect still debated.
Metabolic syndrome screening at 2+ years¶
Standard ATP-III / harmonized IDF criteria (any 3 of 5):
| Component | Threshold |
|---|---|
| Waist circumference | See note below |
| Fasting TG | ≥150 mg/dL |
| HDL | <50 mg/dL |
| Blood pressure | ≥130/85 mmHg |
| Fasting glucose | ≥100 mg/dL |
Waist threshold is the ambiguous one
Adult-onset trans women retain higher absolute waist than cis-female norms even with good gynoid redistribution. Some clinicians use the AMAB threshold (≥102 cm) for 2–5 years and shift to the cis-female threshold (≥88 cm) afterward. No consensus exists.
The highest-risk metabolic-syndrome phenotype at 2+ years is the trans woman with CPA history, BMI drift past 27, oral estradiol, and family-history T2DM — overrepresented in clinic populations because CPA is cheap, oral estradiol is a common starter regimen, and weight regain is common. Switching to transdermal estradiol and off CPA addresses two modifiable contributors.
Psychiatric medications and metabolic drift¶
The trans population is overrepresented on psychiatric medications, many of which are independently weight-active. See Comorbidities for prescribing context.
| Drug class | Metabolic signal |
|---|---|
| SSRIs / SNRIs | Weight-neutral short-term. Paroxetine, mirtazapine most weight-positive. Long-term (>2 y) ~1–3 kg gain |
| Atypical antipsychotics — olanzapine, quetiapine, risperidone | Strongest signal: weight gain, insulin resistance, dyslipidemia, frank T2DM induction |
| Aripiprazole, lurasidone | Weight-neutral alternatives where clinically appropriate |
| Lithium | Weight gain in ~25%. Renal-handling interaction with spironolactone — levels can drift on spiro changes |
| Stimulants | Modestly weight-suppressive. Cardiovascular not metabolic concern |
| Valproate, gabapentinoids, mirtazapine | Substantially weight-positive; combined with HRT can produce 10+ kg gain in 18 months |
Highest-risk regimen this page addresses
Oral estradiol + CPA + olanzapine (or another high-metabolic-burden antipsychotic) is the single highest-risk combination for metabolic syndrome. Patients on weight-active psych drugs need metabolic-syndrome screening every 6 months regardless of HRT status. Estrogen also amplifies antipsychotic-induced hyperprolactinemia.
The practical screening rule: patients on any weight-active psych drug should have A1c, lipids, and LFTs annually rather than every 1–2 years, and the medication contribution should be explicitly considered before attributing metabolic drift to HRT.
Weight cycling and metabolic markers¶
The cis-population literature is unambiguous that weight cycling worsens cardiometabolic outcomes independent of net weight change. Zou 2019 meta-analysis (n=245,109) found weight fluctuation associated with CVD mortality RR 1.36; the Vanderbilt EHR cohort (Swartz 2024–25) found cycling vs. stability gave HR 1.54 for heart failure, 1.28 for OSA and MASLD, 1.23 for T2DM. Mechanistically, repeated cycles produce catch-up fat (preferential fat regain over lean mass), progressively higher fat:lean ratios, and persistent low-grade inflammation.
Trans-specific cycling data are essentially absent. No prospective study has measured A1c, lipids, or hepatic markers across deliberate cut-and-bulk cycles in transfeminine adults. The mechanistic concern is that cycling stacks weight-fluctuation metabolic costs on top of an already-modestly insulin-resistant baseline (especially on CPA), and that catch-up fat may still raise total fat mass and worsen markers even when regain is gynoid-biased.
Patients pursuing intentional cycling should have A1c, full lipid panel, and LFTs at baseline, end of cut phase, and 6 months into bulk phase. See Weight Cycling for the broader evidence picture and Training & Nutrition for the lower-risk alternatives.
Consolidated screening schedule at 2+ years¶
A reasonable cadence for a transfeminine adult past redistribution, on stable HRT, without metabolic comorbidity:
| Cadence | Tests |
|---|---|
| Annually | Weight, BP, fasting lipid panel, A1c, ALT/AST, BMP including potassium |
| Every 1–2 years | Estradiol/testosterone levels, prolactin |
| At any anti-androgen change | LFTs at 4 and 12 weeks |
| At route change (PO ↔ TD/IM) | Lipids and A1c at 3 and 12 months |
Patients with BMI ≥25, family history T2DM, oral estradiol, CPA history, or weight-active psych medications should compress the lipid/A1c/LFT cadence to every 6–12 months. Patients with established metabolic syndrome, T2DM, or MASLD follow disease-specific guidelines that supersede HRT-monitoring defaults.
Where the evidence runs out¶
The honest gaps:
- Almost no prospective metabolic data extend past five years on HRT in adult-onset transitioners.
- Bicalutamide metabolic data are sparse and mostly hepatic.
- Intentional weight-cycling effects in trans women are unstudied.
- MASLD prevalence in trans women is essentially unmeasured outside small clinic series.
- The interaction between orchidectomy → estradiol monotherapy and metabolic markers is studied only in narrow Amsterdam cohorts.
Anyone making confident long-term claims is extrapolating. The responsible posture is monitor-and-treat against general-population thresholds, with anti-androgen choice and estradiol route as the two HRT levers most worth adjusting when markers drift.
Sources¶
- Klaver M et al. Andrologia 2017. Cross-sex hormone therapy body composition meta-analysis (n=171 trans women).
- Klaver M et al. J Sex Med 2018. Regional body fat changes, ENIGI cohort (n=179).
- Klaver M et al. Eur J Endocrinol 2022. Visceral adipose tissue on modern 17β-estradiol regimens.
- Auer MK et al. PLoS ONE 2014. 12-month endocrine/metabolic profile in trans women on HRT (n=98).
- Hembree WC et al. J Clin Endocrinol Metab 2017. Endocrine Society guideline.
- WPATH Standards of Care, version 8 (2022). Coleman E et al.
- Gava G et al. Clin Endocrinol 2020. CPA vs. leuprolide 5-year RCT.
- Wierckx K et al. J Sex Med 2014. Lipid and metabolic effects in trans women on long-term HRT.
- Wiepjes CM et al. J Bone Miner Res 2018. Amsterdam Cohort of Gender Dysphoria (n=2,627) — body composition and weight trajectory.
- Mauvais-Jarvis F, Clegg DJ, Hevener AL. Endocr Rev 2013. Estrogens and insulin sensitivity.
- Zou H et al. Obes Rev 2019. Weight fluctuation and CVD mortality meta-analysis (n=245,109).
- ADA 2024 Standards of Care. A1c thresholds.
- UCSF Center of Excellence for Transgender Health. 2nd edition guidelines.
- Fenway Health and Callen-Lorde informed-consent monitoring protocols.
- EMA 2020. Cyproterone acetate restriction announcement.