“Adrenopause” and menopause: two decoupled hormonal processes that converge—or do not—in tissue

Stage: REPORT_EN
Cutoff date: 2026-08-31
Classification: adrenal androgens, menopause, intracrinology, women's health and longevity
Project: L8-3
Scientific status: complete report for QA; it is not medical advice, diagnosis, or a prescription
Evidence used: human longitudinal, observational, interventional, and ex vivo evidence; animal/primate; in vitro; public computational evidence; and explicitly labeled inference. No current Lua data, PHI, PII, or private genomes were used.

1. Executive scientific abstract

The premise in the title does not hold in its strong form. Menopause and “adrenopause” are neither equivalent hormonal clocks nor intraindividual declines demonstrated to be parallel. The menopausal transition is a nonlinear reorganization of the hypothalamic–pituitary–ovarian (HPO) axis: the probability of ovulation and of a robust luteal phase falls, E2 fluctuates and then declines rapidly around the final menstrual period, and FSH changes earlier and with a different temporal shape. So-called “adrenopause” is a descriptive label—not a diagnosis—for the age-related decline of the adrenal Δ5 branch, especially DHEA and DHEA-S. That decline may reflect lower DHEA production, lower sulfation, redistribution of steroid carbon, altered uptake/use, or clearance; serum DHEA-S alone cannot distinguish these possibilities.

Human longitudinal evidence shows different clocks. In SWAN, E2 began its rapid decline about 2.03 years before the final menstrual period and stabilized about 2.17 years afterward; in another SWAN cohort, DHEA-S had a large cross-sectional age effect (−2.81%/year) but a small residual longitudinal slope (−0.33%/year, P=0.06) and stage changes of about 4% [1,2]. The statement that “84.5% of women increased” was an estimate from a random-effects model, not a count of observed trajectories. A subcohort with nine samples per woman showed a modest median decline in DHEA-S around the final menstrual period [3]. Among 1,104 women assessed by LC–MS/MS in 2025, DHEA was 33% lower in the 65–69 than the 40–44 age group, but DHEA and testosterone did not differ by menopausal stage at ages 48–53 [4]. The defensible result is decoupling of magnitude and clock, not universal adrenal compensation.

Source mechanisms are also asymmetric. The ovary supplies active ligands and cyclic architecture; the adrenal zona reticularis supplies mainly precursors that must still be transported, desulfated, and converted intracrinologically. The adrenal cortex ages selectively: DHEA, DHEA-S, and A4 decline, whereas 11-oxygenated C19 steroids and glucocorticoid production may be preserved [5,6]. Human histology supports zona reticularis remodeling and lower CYB5A/SULT2A1 signal, but it does not determine how much reflects lower zonal mass, lower flux per cell, or altered sulfation.

The obligatory connection between the two systems, if one exists, lies in target tissue. Human adipose tissue contains transport machinery, STS, and downstream steps; homogenate STS activity was higher in 7 postmenopausal than 18 premenopausal women [7,8]. But homogenization removes the transport gate and does not demonstrate net product, ER/AR activation, or function. Exogenous DHEA can increase DHEA-S, E1, E2, and testosterone, yet the audited trials were mostly null for strength, VO₂, body composition, insulin sensitivity, mood, and quality of life; BMD signals were small and did not demonstrate fracture prevention or an interaction with ovarian loss [9–12].

The project's decisive correction concerns bone. The +0.49% in Ghebre et al. was baseline femoral BMD per μmol/L DHEA-S, not 0.49% less annual loss. The DHEA-S×time term was negative (−0.020%/year per μmol/L), so the baseline advantage diminished [13]. In SWAN (n=2,311), DHEA-S, testosterone, and free androgen index did not predict four-year BMD loss; FSH did [14]. This weakens the principal apparent evidence for endogenous adrenal protection.

The computational analysis of GSE253849 likewise did not reproduce architectural loss of zona reticularis. Across four conservatively defined female units aged 18, 53, 59, and 64 years, ZR fractions were 33.4%, 42.7%, 34.7%, and 39.8%; the slope was +1.19 percentage points per decade and became negative after removing the only young donor. A Δ5 module within ZR declined weakly in women (0.928×/decade), but not uniformly by gene and reversed direction when sexes were combined. Computation did not elevate the mechanism to an H2 signal; it required parking the architectural hypothesis and separating reduced 17,20-lyase from reduced sulfation/rerouting [15].

The best-supported explanation today is tissue additivity plus common causes of aging and measurement (L8-3-AR-H2A/H2B, maturity H1). The primary hypothesis with the highest discriminating value—not the greatest prior probability—is nonlinear ER/AR pharmacodynamic convergence (L8-3-AR-H1 v1, H0): two additive molecular inputs could cross a steep region of the ligand–function curve and look like a “double hit” without communication between ovary and adrenal. The translational hypothesis (L8-3-AR-HT1 v1) remains parked because no repeatable observable yet connects ex vivo tissue flux with a human trajectory.

The smallest decisive test is neither another serum association study nor another systemic DHEA trial. It is a stop-early sequence: qualify metrology and function in adipose explants (Q0); map an acute E2×DHEA-S surface with Adiol, isotopic flux, STS, ER/AR, and a frozen function (D0-A); test flux adaptation only if function exists (D0-B); and, in an independent adrenal branch, use two isotopic inputs to distinguish reduced lyase from reduced sulfation/rerouting and architecture (E0-ZR). Only then would a single-organ, FMP-aligned cohort be justified. Longevity remains unidentified: there is no “hormonal capital” that can be calculated by adding E2 and DHEA-S.

2. Scientific question and relevance

Central question

In women undergoing natural menopause, does lower adrenal Δ5 availability modify beyond additivity the effect of ovarian loss on intracrine product, ER/AR signaling, and a proximal organ function, or is the apparent convergence explained by additivity, common aging, compartment, and measurement error?

The scientific relevance arises from an artificial historical separation. Reproductive endocrinology usually describes the HPO transition through FSH, E2, ovulation, and menstruation; adrenal endocrinology describes DHEA/DHEA-S by age. Yet postmenopausal tissues receive a mixture of ligands from residual ovarian, adrenal, adipose, and local sources. Treating each silo in isolation could conceal a real tissue mechanism; merging them as “two declines” could manufacture a relationship that does not exist.

The question has three levels that must not be conflated:

1. Source: do ovarian and adrenal production, sulfation, or architecture change?
2. Intracrinology and receptor: how much precursor becomes active ligand, and how do ER/AR respond?
3. Health: does that proximal function precede a persistent organ change and, later, years free of disease or disability?

The work is relevant to women's longevity precisely because it sets a high bar: an association with BMD, mood, IMT, predicted risk, or mortality in a selected sample cannot replace the causal chain. The same steroid exposure may favor bone while harming hormone-sensitive tissue or metabolism; directions must be preserved as a vector of organs and competing risks.

3. Scope, population, and life stage

The primary inferential population is women aged 35–70 years with ovaries and adrenals present at baseline, centered on ages 40–60: late reproductive stage, early and late menopausal transition, early postmenopause, and the first years after the final menstrual period. Final menstrual period date, STRAW+10 stage, age, and years since FMP are distinct variables.

Informative but non-interchangeable populations include:

• surgical menopause or ovarian suppression as perturbations of the ovarian source;
• adrenalectomy, adrenal insufficiency, or rare sulfation defects as source experiments with limited transportability;
• non-tumoral human female tissue as a system for proximal causality;
• animal/primate and in-vitro models only as mechanistic possibility;
• a human longitudinal cohort only after establishing a causal and repeatable measure.

Outside the primary estimand are pregnancy, lactation, POI before age 40, PCOS, congenital adrenal hyperplasia, secreting tumors, Cushing/Addison disease, exposures that preclude estimation of endogenous hormones, and the use of “adrenal fatigue” as an entity. Cortisol/HPA is a confounder or modifier; it does not define “adrenopause.” MHT, DHEA, androgens, and glucocorticoids are not mixed with natural exposure without explicit separation.

4. Background knowledge and mechanism map

4.1 Two non-equivalent sources

The ovary supplies E2, E1, and P4 within a cyclic architecture. With reproductive aging, follicular reserve, ovulation probability, luteal duration/mass, and feedback decline; FSH rises, E2 fluctuates, and later falls. Natural menopause is not equivalent to a single E2 value, and FMP is not an intervention.

The adrenal zona reticularis favors the Δ5 branch:

cholesterol → STAR → CYP11A1 → pregnenolone
pregnenolone → CYP17A1 → 17OH-pregnenolone
17OH-pregnenolone --CYP17A1 + CYB5A + POR--> DHEA
DHEA --SULT2A1 + PAPS/PAPSS2--> DHEA-S

In a purified system, CYB5A increased conversion of 17OH-pregnenolone to DHEA by about 13-fold, demonstrating catalytic potency, not that it is the only age-related bottleneck [16]. Lower DHEA-S can result from lower lyase activity, lower sulfation, greater peripheral STS, greater tissue use, or altered clearance. Topology matters: lower CYB5A/POR should reduce total C19 carbon; lower SULT2A1/PAPSS2 may reduce DHEA-S while preserving or diverting DHEA toward A4/T. The two signatures share a low serum precursor but have opposite consequences.

4.2 The intracrine gate

In tissue t, the relevant flux can be expressed as:

J_t ≈ S_free × uptake_t × STS_t × downstream_conversion_t − reconjugation_t − export_t.

Hydrophilic DHEA-S must enter or be delivered; STS releases DHEA; HSD3B, AKR1C/HSD17B, CYP19A1, and SRD5A distribute carbon toward A4, T, DHT, E1, E2, and other products. Androstenediol (Adiol) adds complexity because it has ER and AR activity and may arise from circulation or peripheral conversion. High tissue content does not equal flux: it may reflect sequestration, reconjugation, or slow clearance.

4.3 Receptor convergence

Effective cellular exposure is a mixture, not DHEA-S:

L_ER = E2 + weighted E1 + Adiol + other ligands
L_AR = T + weighted DHT + smaller contributions.

Function is F = f(L_ER, L_AR, ERα/ERβ, AR, chromatin, composition, and metabolic state). If f contains a threshold, saturation, cooperation, or antagonism, two additive inputs can appear nonadditive on the observed scale. An E2×DHEA-S interaction does not demonstrate compensation. Intracrine adaptation requires an alteration in absolute precursor-to-product flux before function changes; pharmacodynamic convergence requires stable flux plus a reproducible, receptor-dependent ligand–function surface.

4.4 Multiscale map

aging/inflammation/adiposity/disease/clearance
        ├───────────────┬──────────────────┐
        v               v                  v
 ovary/HPO          adrenal ZR        target tissue
 E2/P4/ovulation    DHEA/DHEA-S       transport/STS/downstream
        └───────────────┴───────────────> ER/AR ligands
                                             v
                                 cellular/organ function
                                             v
                              persistent disease trajectory
                                             v
                       years free of disease/disability
                              with competing risks

The obligatory convergence occurs in the target tissue, not in the source organs.

5. Evidence method

The report builds on the complete SCOPING, EVIDENCE_MAP, EVIDENCE_VERIFICATION, MECHANISTIC_SYNTHESIS, COMPUTE_DECISION, COMPUTE_OPTIONAL, HYPOTHESIS_GENERATION, ADVERSARIAL_REVIEW, and EXPERIMENT_DESIGN artifacts and the lifetime memory packet. Primary sources were prioritized and decisive studies were re-audited. Recent 2025–2026 claims were checked against PubMed/PMC and GEO primary records.

Classification:

• H-L: human longitudinal;
• H-OBS: human observational/cross-sectional;
• H-RCT: human randomized trial;
• H-IV: human tracer/intervention;
• H-EV: human ex vivo;
• A: animal or primate;
• V: in vitro/purified enzyme;
• C: computational;
• INF: inference across studies.

Rules: donor/woman is the unit, not cell or well; concentration is not production; RNA is not protein or flux; product is not function; P>0.05 does not establish equivalence; adjustment for downstream E2 can block mediation; BMD is not fracture; proximal function is not healthspan.

6. Evidence map

Strength hierarchy by link

1. Strong: the HPO transition changes E2/FSH; the Δ5 branch ages selectively; DHEA can undergo peripheral conversion.
2. Moderate: age weighs more than FMP for DHEA/T; compartmentalization and adipose STS capacity exist.
3. Low: postmenopausal intracrine adaptation preserves function; a CYB5A/POR defect dominates sulfation; Adiol explains function.
4. Gap: a human ovary–adrenal interaction from flux through function; a joint effect on healthspan or own survival.

7. Contradictory evidence and null findings

7.1 A DHEA-S rise is not an individual law

Crawford modeled a positive stage effect and estimated that 84.5% would rise; Kim observed a modest median decline around FMP [2,3]. BSO did not abolish a 5–8% rise in 14/20 selected women, excluding ovarian necessity but not establishing adrenal production or prevalence [25]. Explanations include HPO–HPA signaling, clearance, tissue use, selection, or error. Smallest discriminator: repeated DHEA/DHEA-S/products with production/clearance estimation and HPO hormones.

7.2 STS capacity is not buffering

Homogenate STS increases after menopause [7], but homogenization removes transport; downstream product, receptor, and function were not measured. The finding could reflect adaptation, cellular composition, adiposity, surgical indication, or capacity without substrate. Smallest discriminator: labeled DHEA-S in intact explant, STS loss/rescue, complete balance, and function at human-range supply.

7.3 Bone: the correction that changes causal weight

Ghebre appeared to contradict Guthrie/SWAN, but 0.49% was baseline BMD, not less loss; the time interaction was negative [13]. Guthrie found DHEA-S null and final E2 predictive [26]; SWAN confirmed the larger null [14]. Exogenous DHEA produced small signals, possibly statistically mediated by E2, but the analysis was secondary/per protocol and did not localize bone synthesis [10,17]. The contradiction no longer supports robust longitudinal adrenal protection.

7.4 Large products, small or null function

Systemic DHEA raised several steroids by ~60% to >600% in audited studies without consistent benefits in strength, VO₂, composition, insulin sensitivity, or QoL [10–12]. Explanations include wrong tissue, window/dose, saturation, conversion without function, nonequivalent oral exposure, or a true absence of effect. Product without function must count as falsification of that tissue bridge, not proof that conversion did not occur.

7.5 “More androgen” is not “more health”

In 2026, an androgenicity index was cross-sectionally associated with worse predicted cardiometabolic risk, while individual components did not show the same signal and SHBG had the opposite direction [24]. The index may capture adiposity, HDL, and SHBG; it does not establish hormonal harm. It does refute any universally favorable valence.

7.6 Negative architecture computation

GSE253849 did not reproduce lower p_ZR with age. The direction changed after excluding F18; the two F53 libraries disagreed and their independence was unresolved. A weak female module concentrated in SULT2A1/CYB5A/POR but reversed direction when sexes were combined. This does not refute histology because scRNA-seq captures cells, not mass; it does prevent escalation of AP-1/renewal or architecture from that atlas.

8. Multiscale mechanism synthesis

Molecular level

Menopause reduces one source of active ligand. Adrenal aging heterogeneously reduces a precursor branch. The tissue consequence depends on sulfate uptake, STS, 3β/17β-HSD, aromatase, 5α-reductase, reconjugation, and the potency of E1/E2/T/DHT/Adiol. The same low DHEA-S may mean lower C19 carbon or rerouting toward more active ligands.

Cellular level

The cell translates a mixture through ERα/ERβ/AR in a context of chromatin, insulin, inflammation, and composition. Chronic E2 change may alter receptors without altering flux; calling this “intracrine adaptation” would be incorrect. Adaptation requires an early change in the absolute substrate-to-product slope.

Tissue level

Adipose offers the most accessible human biochemical chain, but it is not yet a validated causal organ. VAT and SAT are not equivalent; STS also processes E1-S, and CYP19A1 loss reroutes toward AR. Perturbation must therefore measure the entire mixture and separate ER from AR. A null in adipose kills that bridge, not all adrenal biology.

Organ and systemic level

Bone shows E2/FSH predominance and small exogenous signals; muscle and metabolism show nulls; brain cannot be inferred from serum; CVD and mortality are sensitive to prior disease. The evidence does not converge on a systemic direction.

Longevity level

The chain stops before product→function. Advancement would require:

independent flux/product → receptor → persistent function → organ trajectory → disease/disability → event-free years, with temporality and competing risks. A cross-organ composite created after observing divergent signs would be scientifically invalid.

9. Computational layer

Question

Does the age-related reduction of the Δ5 branch reflect lower ZR representation or a lower program per ZR cell?

Data and method

Public counts from GSE253849, released in GEO in 2026, were analyzed with donor/pool as the unit; the two female age-53 libraries were conservatively merged because independence could not be resolved. All 60,146 processed cells linked to the Cell Ranger H5 files; the primary analysis used provider annotations and pseudobulk. BioNeMo was not used. The deposition had no linked peer-reviewed publication at cutoff.

Result

• Female p_ZR: slope +0.0119/decade (95% CI −0.0402 to +0.0640); changed to −0.0293 after removing F18.
• Female ZR Δ5 module: −0.1078 log2/decade (95% CI −0.3678 to +0.1523), equivalent to 0.928×/decade; combined-sex model +0.0788 log2/decade.
• CYB5A, SULT2A1, POR, and LDLR were negative in the female fit, but five of seven genes changed sign in at least one leave-one-out analysis.
• ZF AP-1 showed no reproducible relationship with lower ZR.

Interpretation

The donor-stability gate failed. Spatial/GTEx validation was not run post hoc. The correct decision was to weaken and park MECH-H4, not to call the result a biological null. Computation only prioritizes the experiment: measure stereology, protein, and flux per cell, and distinguish lyase from sulfation.

What cannot be inferred

Age does not identify menopause; RNA is not protein; captured proportion is not mass; pseudobulk is not flux; the dataset contains no DHEA-S, E2, FMP, ACTH, function, or longevity.

10. Primary hypothesis

L8-3-AR-H1 v1 — nonlinear pharmacodynamic convergence

Falsifiable statement: in qualified human adipose explants with measured receptor state, a complete free-ligand mixture including E2, E1, T, DHT, and Adiol will show a functionally material deviation from additivity when ovarian and adrenal inputs are low, without a change in absolute DHEA-S-to-product flux.

Lineage: MECH-H3 v1 → HG-H1 v1 → AR-H1 v1.
Status: weakened_narrowed.
Maturity: H0.
Confidence: 0.24.

Mechanism: ovarian and intracrine inputs add chemically; ER/AR translate them through a nonlinear surface. The “double hit” is tissue pharmacodynamics, not source cross-talk or compensation.

Predictions: equivalent absolute flux and dose–product slopes across acute conditions; functional interaction outside the margin on the absolute scale; improved out-of-donor prediction; ER/AR dependence; persistence after including Adiol; STS shifts dose but not surface shape.

Evidence for: human peripheral conversion; adipose machinery; variation in E2 response by ERα/ERβ [8,9,27].

Evidence against: no direct interaction; null RCTs; scale dependence; incomplete mixture; possibility of a flat function over the human range.

Kill criteria: interaction within equivalence under a sensitive benchmark; effect on only one scale; disappearance after incorporating mixture/receptor; nonreplication; nonhuman exposure; labeled product without function.

11. Competing hypothesis

L8-3-AR-H2A v1 + L8-3-AR-H2B v1 — tissue additivity and common causes

These two layers are preserved under one competing hypothesis but are not experimentally conflated.

Falsifiable statement: with active mixture, receptor state, and intracellular dose measured, ovarian and adrenal inputs will have additive main effects and the interaction will fall within equivalence; in humans, age, prior health, adiposity, liver/kidney function, HPA, activity, and error will explain most distal covariance, with no external temporal gain from a hormonal interaction.

Status: strengthened_competing.
Maturity: H1.
Confidence: 0.68 for tissue additivity; 0.58 for common causes/measurement.

Mechanism: age and inflammaging partly independently modify HPO, ZR, tissue composition, and function. DHEA-S may mark somatic reserve or clearance; residual hormonal effects are small and organ-specific.

Predictions: ex vivo interaction equivalent with demonstrated sensitivity; no material E2/P4–DHEA/DHEA-S synchrony; human attenuation after prior function and time-varying confounding; isolated DHEA-S does not improve prediction; directions are not uniform across organs.

Evidence for: discordant clocks, DHEA-S null for BMD and mood, null functional RCTs, nonmonotonic mortality, unstable ZR computation [2–4,11,14,19–24].

Evidence against: real peripheral conversion, higher postmenopausal STS, and small exogenous BMD signals leave open a narrow window.

Kill criteria: a material donor-level, receptor-mediated, rescuable interaction robust to mixture/scale; then independent human temporality with external improvement and replication.

12. Translational hypothesis

L8-3-AR-HT1 v1 — conditional dynamic research assay

Falsifiable statement: only after validating function and metrology, a repeatable human observable derived from conversion–response will improve out-of-sample prediction of an organ-specific function beyond age, stage, prior function, adiposity, kidney/liver status, and isolated concentrations.

Status: parked_no_observable.
Maturity: H0.
Confidence: 0.04.
Governance: HUMAN_QA_REQUIRED before human evaluation or biomarker language.

Mechanism: a dynamic may distinguish uptake, conversion, and receptor susceptibility better than a static concentration. Today the assay requires surgical tissue and cannot be repeated longitudinally in the same woman; it is therefore not a biomarker.

Predictions for reactivation: functional causality; interlaboratory assay; accessible and repeatable human exposure; adequate ICC/CV/drift; external gain beyond prior function and covariates.

Dominant evidence against: no longitudinal observable exists, aliquot test–retest would measure technical precision, and surgical tissue selects indication/comorbidity.

Kill criteria: absent causality, no repeatable proxy, only technical reproducibility, signal absorbed by donor/batch/composition, or no external gain.

13. Falsifiable predictions and kill criteria

Equivalence margins must be anchored to the minimum detectable change and a functional benchmark, not selected after seeing data.

14. Discriminating experiment

Minimal program Q0 → D0-A → D0-B

Q0: metrology and sensitivity

Paired VAT/SAT explants from 12 postmenopausal women, donor as unit. Isotopic LC–MS/MS panel for DHEA-S, DHEA, Adiol, A4, T, DHT, E1-S, E1, E2, conjugates, and isotopologues; free fraction, adsorption, medium, tissue, and effluent. Three human-range doses and 0/2/6/24-hour time points. Irosustat as an STS tool; bypass with labeled DHEA. Functional benchmark: isoproterenol increases glycerol/NEFA and insulin suppresses them.

Gates: target analytical recovery 80–120%, carryover <1%, technical CV ≤15%; ≤20% unassigned labeled carbon; STS and bypass in the correct direction; viability ≥85%; at least 9/12 qualified VAT samples. These are experimental rules, not clinical standards.

D0-A: acute surface

Recruit up to 30 donors to obtain 24 qualified VAT samples. A 3×3 surface of free E2 × labeled DHEA-S, four sentinel Adiol conditions, inhibited STS, DHEA bypass, perturbed ER/AR, and matched active mixture. Primary endpoints: absolute product-formation slope and percentage insulin suppression of glycerol after isoproterenol. Mixed model with donor intercept; absolute primary scale; leave-one-donor-out validation; TOST margin derived from Q0.

Adjudication: robust nonlinearity strengthens AR-H1; equivalence strengthens AR-H2A; Adiol with blocked STS strengthens AR-H4; product without function closes adipose; donor/batch dependence or nonhuman exposure makes the result unidentifiable.

D0-B: adaptation, only if D0-A demonstrates function

Low/high E2 preconditioning for up to 48 hours, an acute arm, measured washout, and three labeled DHEA-S doses. Primary endpoint: difference in absolute dose–product slopes before function. Confirmatory set of 20 qualified donors. Adaptation dies if the slope is equivalent, change is only uptake/retention, bypass fails, or STS changes without function.

Independent E0-ZR branch

Non-tumoral female adrenal tissue, two age anchors of 18–40 and 60–80 years; QZ0 with six donors per anchor and a confirmatory target of 40 valid tissues. 3D stereology; independent ZR identity; proteins CYB5A/POR/CYP17A1/SULT2A1/PAPSS2/HSD3B2/LDLR/SCARB1.

Two isotopic inputs:

1. [13C]17OH-pregnenolone → DHEA/DHEA-S/A4/T/total C19 for lyase;
2. [13C]DHEA → DHEA-S vs DHEA/A4/T for sulfation/rerouting.

Architecture advances only if ZR mass explains ≥50% of the contrast and flux per cell is equivalent. NCI-H295R may calibrate signatures but may never replace human tissue.

Conditional H0-FMP cohort

Only after mechanism and a repeatable observable: metrology in 30 women; then up to 240 to obtain ≥180 evaluable participants aged 42–55, quarterly visits for 36 months, STRAW+10/FMP, a complete steroid panel, and one standardized adipose outcome. DAG, lags, scale, missingness, and prior function preregistered. No mortality, fracture, dementia, or composite endpoint.

15. Biomarkers and stratification

There is currently no validated biomarker of “functional adrenal reserve” or ovarian–adrenal interaction. DHEA-S is a nonspecific circulating exposure. A research-useful panel must separate:

• source: DHEA, DHEA-S, A4, and, where possible, production/clearance;
• partition: Adiol, T, DHT, E1-S, E1, E2, conjugates, and SHBG;
• HPO: STRAW+10, FMP, E2/E1/P4/FSH, and independent ovulatory/luteal activity;
• context: ACTH/cortisol, liver/kidney status, adiposity, activity, inflammation, and prior disease;
• function: one reproducible organ-specific readout.

Initial stratification is scientific, not clinical: VAT/SAT depot; years since FMP; BMI as a block, not an exploratory subgroup; exogenous hormonal exposure separated; center/ancestry/surgical indication explicit. Depots will not be averaged and post hoc “responders” will not be created.

16. Individual variability

Heterogeneity may arise from:

1. Trajectory: age, FMP timing, anovulation, and luteal exposure.
2. Adrenal source: ZR mass, lyase, sulfation, cholesterol, ACTH, and clearance.
3. Tissue: transporters, STS, aromatase, HSDs, 5α-reductase, reconjugation, cellular composition, and adiposity.
4. Receptor: ERα/ERβ/AR, chromatin, insulin, and inflammation.
5. Metrology: phase, time of day, fasting, batch, LLOQ, free fraction, and matrix.
6. Selection: surgery, comorbidity, treatment, and survival.

Mexican/LATAM populations are underrepresented. This is a transportability gap, not evidence that the mechanism is stronger or different. Ancestry, adiposity, stage, and center should be studied only after mechanism and assay validation.

17. Pharma relevance and maturity

The current opportunity is target validation and tissue pharmacology, not a therapy-ready program.

Class risks include breast/endometrium, bone, thrombosis/vasculature, liver, virilization, brain/mood, and compensatory rerouting. No target reaches H5; HUMAN_QA_REQUIRED before partnering. The report does not select an indication, compound, formulation, or dose.

18. Limitations

1. No cohort repeatedly measures HPO, DHEA/DHEA-S, Adiol/products, production/clearance, and function in the same women around FMP.
2. Many transition studies used immunoassays; DHEA-S stage effects near 4% compete with historical CVs of 10–13%.
3. Nonparallelism is triangulated across cohorts/models, not through a formal paired change-point contrast.
4. Adipose evidence measures fragments, homogenate, or small groups; it does not close end-to-end flux and function.
5. Systemic RCTs alter multiple tissues and steroids and do not localize STS or receptor.
6. Adiol rests on SWAN samples selected by DHEA-S and a historical bioassay; it prevents omission but does not validate compensation.
7. Normal young/older female adrenal tissue is scarce; age, ischemia, pathology, and indication may be confounded.
8. GSE253849 has only one young woman, unresolved F53 dependence, and nonindependent annotation; scRNA-seq does not measure mass.
9. Animal/primate models and NCI-H295R do not reproduce natural transition or complete human physiology.
10. Lipolysis suppression is a proximal function, not diabetes, disability, or longevity.
11. A null adipose result does not refute other tissues; it does preclude extending adipose without independent justification.
12. Cohorts are mainly White/European or White/African American SWAN; transport to Mexico/LATAM is unproven.
13. No net multiorgan direction exists: benefits and risks may coexist.

19. Conclusions

1. Menopause and “adrenopause” are not two demonstrated parallel declines. They share age but differ in organ, topology, magnitude, and clock.
2. DHEA-S is not a summary hormone: it integrates production, sulfation, use, and clearance and does not identify tissue dose.
3. Peripheral conversion is real; its functional importance remains unproven. STS capacity and circulating product do not equal buffering.
4. The main favorable bone evidence was corrected: 0.49% was baseline BMD, not annual protection. The best transition data are null for DHEA-S.
5. The public atlas did not support architectural ZR loss; it left only a selective, unstable signal that requires separation of lyase from sulfation/rerouting.
6. The best-supported hypothesis is additivity plus common causes. Nonlinear convergence is prioritized because it can be falsified quickly, not because it is more probable.
7. The decisive experiment must measure absolute flux, the complete mixture—including Adiol—receptor, and function with equivalence and a benchmark; a serum interaction is insufficient.
8. Pharma relevance is in tissue target validation. There is no therapeutic readiness or repeatable translational observable.
9. Healthspan and longevity remain closed until persistent function, organ-specific mediation, and competing risks are demonstrated.

Final scientific delta of REPORT_EN

This report replaces the narrative of two parallel hormonal declines with a falsifiable model of temporally decoupled sources, alternative adrenal topologies, and conditional tissue convergence. The innovation is not to add E2 and DHEA-S, but to distinguish four outcomes that previously looked alike: true additivity, receptor nonlinearity, flux adaptation, and sulfation with rerouting. The experimental program can kill each with the fewest steps and recognizes conversion without function or low DHEA-S with preserved C19 carbon as positive knowledge outcomes.

20. References

1. Randolph JF Jr, et al. The value of follicle-stimulating hormone concentration and clinical findings as markers of the late menopausal transition. J Clin Endocrinol Metab. 2011;96:746–754. DOI: 10.1210/jc.2010-1746. PMID: 21159842. Human longitudinal, SWAN, n=1,215 with FMP.
2. Crawford S, et al. Longitudinal changes in adrenal androgens during the menopausal transition. J Clin Endocrinol Metab. 2009;94:2945–2951. DOI: 10.1210/jc.2009-0386. PMID: 19470626. PMCID: PMC2730879. Human longitudinal, n=2,886, 15,930 observations.
3. Kim C, et al. Changes in androgens during the menopausal transition: recent longitudinal findings from SWAN. Women's Midlife Health. 2017. DOI: 10.1186/s40695-017-0028-4. PMID: 29333273. Human longitudinal, n=110, nine samples per woman.
4. Wang Y, Islam RM, Bond M, Davis SR. Testosterone and pre-androgens by age and menopausal stage at midlife: findings from a cross-sectional study. EBioMedicine. 2025;121:105972. DOI: 10.1016/j.ebiom.2025.105972. PMID: 41106025. PMCID: PMC12552971. Human cross-sectional LC–MS/MS, n=1,104.
5. Nanba AT, et al. 11-Oxygenated C19 steroids do not decline with age in women. J Clin Endocrinol Metab. 2019;104:3160–3170. DOI: 10.1210/jc.2018-02527. PMID: 30753518. PMCID: PMC6525564. 100 younger women, 100 older women, and 8+8 adrenals.
6. Tezuka Y, et al. Age-dependent changes of the human adrenal cortex and its steroidogenic capacity. J Clin Endocrinol Metab. 2021. DOI: 10.1210/clinem/dgab007. PMID: 33524149. PMCID: PMC8502483. Human histology of 60 adrenals and a separate steroidomics cohort.
7. Paatela H, et al. Steroid sulfatase activity is higher in adipose tissue from postmenopausal than premenopausal women. Eur J Endocrinol. 2016;174:167–175. DOI: 10.1530/EJE-15-0831. PMID: 26553725. Human ex vivo, 18 pre/7 post.
8. Dalla Valle L, et al. Adipose tissue and adipocytes express transporters and steroid sulfatase involved in DHEA-S utilization. J Endocrinol. 2006. DOI: 10.1677/joe.1.06811. PMID: 16837617. Human ex vivo/in vitro.
9. Longcope C, et al. Peripheral conversion of dehydroepiandrosterone to estrogens in postmenopausal women. Maturitas. 1982. DOI: 10.1016/0378-5122(82)90065-2. PMID: 6221177. Human tracer, n=7.
10. Jankowski CM, et al. Effects of dehydroepiandrosterone replacement therapy on bone mineral density in older adults: a randomized, controlled trial. J Clin Endocrinol Metab. 2008. DOI: 10.1210/jc.2007-2614. PMID: 18812486. PMCID: PMC2626446.
11. Nair KS, et al. DHEA in elderly women and DHEA or testosterone in elderly men. N Engl J Med. 2006. DOI: 10.1056/NEJMoa054629. PMID: 17050889. Randomized trial; 57 women.
12. Igwebuike A, et al. Lack of dehydroepiandrosterone effect on a combined endurance and resistance exercise program in postmenopausal women. 2008. PMCID: PMC2729150. Human trial, 31 completed.
13. Ghebre MA, et al. Association between DHEA-S and bone mineral density in postmenopausal women: the Chingford Study. Calcif Tissue Int. 2011;89:295–302. DOI: 10.1007/s00223-011-9518-9. PMID: 21789637. PMCID: PMC3175043.
14. Sowers MR, et al. Amount of bone loss in relation to time around the final menstrual period and follicle-stimulating hormone staging of the transmenopause. J Clin Endocrinol Metab. 2006;91:1261–1267. DOI: 10.1210/jc.2005-1836. PMID: 16403818. SWAN, n=2,311.
15. NCBI Gene Expression Omnibus. GSE253849: human adult adrenal single-cell RNA sequencing. Public 2026-05-11; series updated 2026-06-23. Canonical URL: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE253849. Public dataset; donor-level analysis documented in compute.md.
16. Katagiri M, et al. Cytochrome b5 modulation of human CYP17A1 17,20-lyase activity. Arch Biochem Biophys. 1995. DOI: 10.1006/abbi.1995.1173. PMID: 7893148. Purified human enzyme.
17. Jankowski CM, et al. Effects of DHEA on bone mineral density in women and men: pooled analysis of randomized trials. 2019. PMCID: PMC6336516. Pooled analysis, 295 women.
18. Dayal M, et al. Effects of dehydroepiandrosterone and conjugated equine estrogen on postmenopausal women: a pilot randomized factorial study. J Womens Health. 2005;14:391–400. DOI: 10.1089/jwh.2005.14.391. PMID: 15989411. n=50.
19. Bromberger JT, et al. Longitudinal change in reproductive hormones and depressive symptoms across the menopausal transition. Arch Gen Psychiatry. 2010. DOI: 10.1001/archgenpsychiatry.2010.55. PMID: 20530009. PMCID: PMC3129620. SWAN, n=3,302.
20. Cappola AR, et al. DHEAS levels and mortality in disabled older women. J Gerontol A. 2006;61:957–962. DOI: 10.1093/gerona/61.9.957. PMID: 16960027.
21. Cappola AR, et al. DHEAS levels, change and mortality in older adults. J Gerontol A. 2009. DOI: 10.1093/gerona/glp129. PMID: 19713299. PMCID: PMC2773814.
22. Shufelt CL, et al. DHEA-S and cardiovascular mortality in postmenopausal women with suspected ischemia. J Clin Endocrinol Metab. 2010. DOI: 10.1210/jc.2010-0143. PMID: 20739385. PMCID: PMC2968728.
23. Meun C, et al. High androgens in postmenopausal women and cardiovascular disease: Rotterdam Study. J Clin Endocrinol Metab. 2018. DOI: 10.1210/jc.2017-02421. PMID: 29408955. n=2,578.
24. Niu Z, Wactawski-Wende J, Hovey K, Millen AE, LaMonte MJ. Androgenic profile and cardiometabolic health among post-menopausal women: the Buffalo OsteoPerio Study. Maturitas. 2026;212:109057. DOI: 10.1016/j.maturitas.2026.109057. PMID: 42492442. Human cross-sectional, n=599.
25. Lasley BL, et al. Menopausal transition stage-specific changes in circulating adrenal androgens after bilateral salpingo-oophorectomy. Menopause. 2011. DOI: 10.1097/gme.0b013e3181fb53fc. PMID: 21178790. PMCID: PMC3123411. Quasi-experiment, 20 analyzed.
26. Guthrie JR, et al. Central abdominal fat and endogenous hormones as determinants of bone loss in women during the menopausal transition. Osteoporos Int. 2004. DOI: 10.1007/s00198-004-1624-3. PMID: 15042284. Human longitudinal, n=159.
27. Park YM, et al. Estradiol, ERα/ERβ balance and adipose tissue lipolytic response in postmenopausal women. PLoS One. 2017. DOI: 10.1371/journal.pone.0176446. PMID: 28472101. PMCID: PMC5417515. Human crossover, n=35.
28. Palmieri C, et al. Presurgical window study of irosustat in postmenopausal women with ER-positive early breast cancer. Breast Cancer Res Treat. 2017. PMID: 28795252. STS target-engagement tool, 13 recruited.
29. Stanway SJ, et al. Phase I study of STX64 (irosustat), an irreversible steroid sulfatase inhibitor, in advanced breast cancer. Clin Cancer Res. 2006;12:1585–1592. PMID: 16533785.
30. Lasley BL, et al. Androstenediol and estrogenic bioactivity during the menopausal transition. Menopause. 2012;19:650–657. DOI: 10.1097/gme.0b013e31823df577. PMID: 22415563. PMCID: PMC3366061. 144 selected SWAN samples.
31. McConnell DS, et al. Changes in circulating adrenal androgens and androstenediol during the menopausal transition. Menopause. 2012;19:658–663. DOI: 10.1097/gme.0b013e31823fe274. PMID: 22415570. PMCID: PMC3366025. 120 selected samples.
32. Thomas-Teinturier C, et al. Age at natural menopause and risk of chronic disease: Mendelian-randomization analysis. Int J Epidemiol. 2022. PMID: 36409989. Human genetics; organ-specific trade-offs.
33. Iwahashi Y, et al. A single-cell and spatial transcriptomic atlas of the aging human adrenal gland. Molecular Metabolism. 2024;84:101954. DOI: 10.1016/j.molmet.2024.101954. PMID: 38718896. PMCID: PMC11101872. Human atlas with strong sex/age confounding.
34. Sultana A, et al. Sex steroids and cognitive decline in older women. Climacteric. 2025;28:446–455. DOI: 10.1080/13697137.2025.2470458. PMID: 40085743. Human prospective, n=4,444.
35. Schmid M, et al. Sex hormones and epigenetic clocks in older adults. J Gerontol A. 2025;80:glaf106. DOI: 10.1093/gerona/glaf106. PMID: 40342106. n=1,404, 48% women.