Lua Labs Report — Adrenal allostasis in perimenopause: when the adrenal gland compensates, and when it fails

Date: 2026-05-24 Researcher: Lua Labs — Scientist Classification: Neuroendocrine / Metabolomics Line: L2 — HPA-HPO Axis Subtopic: 2.4 — Adrenal allostasis in perimenopause

External sources

Sowers MF, Zheng H, Greendale GA, Neer RM, Cauley JA, Ellis JA, et al. (2012). "Menopausal Transition Stage-Specific Changes in Circulating Adrenal Androgens." Menopause, 19(6), 658-663. PMID: 22415570. PMC3366025.
Grub J, Süss H, Willi J, Ehlert U. (2021). "Steroid Hormone Secretion Over the Course of the Perimenopause: Findings From the Swiss Perimenopause Study." Frontiers in Global Women's Health, 2, 774308. DOI: 10.3389/fgwh.2021.774308. PMC8712488.
Lee AA, Den Hartigh LJ. (2025). "Metabolic impact of endogenously produced estrogens by adipose tissue in females and males across the lifespan." Frontiers in Endocrinology, 16, 1682231. DOI: 10.3389/fendo.2025.1682231. PMC12575157.
Cardillo G. (2026). "Beyond adrenal fatigue: reframing the adrenal stress index through neutrophil-mediated glucocorticoid resistance." Frontiers in Endocrinology, 17, 1785454. DOI: 10.3389/fendo.2026.1785454.
Alebna PL, Maleki N. (2021). "Allostatic Load in Perimenopausal Women With Migraine." Frontiers in Neurology, 12, 649423. DOI: 10.3389/fneur.2021.649423. PMC8100599.
Park SB. (2025). "Cortisol, DHEA-S, and cortisol/DHEA-S ratio in association with oxidative stress in Korean adults." Frontiers in Endocrinology, 16, 1708007. DOI: 10.3389/fendo.2025.1708007. PMC12745230.

Baseline knowledge (what I know before searching)

Adrenopause: the parallel decline no one diagnoses

While all clinical attention focuses on ovarian decline (FSH, AMH, E2, progesterone), a second hormonal decline occurs in parallel and in silence: adrenopause. The adrenal glands produce DHEA and DHEA-S (its sulfated, more stable form) from the zona reticularis. This process peaks around ages 25-30 and then progressively declines — approximately 1-2% per year — without any menstrual event signaling it. At age 47, a woman has approximately 50-60% of the DHEA-S she had at her peak.

DHEA-S is not an active hormone per se: it is a prehormone that is converted intracrinely (inside target cells) into active androgens (testosterone, DHT) or estrogens (E2, E1) depending on local enzyme availability — aromatase (CYP19A1), 17β-HSD, 5α-reductase. Conversion occurs mainly in adipose tissue, liver, skin, muscle, and brain. DHEA does not act directly on androgen or estrogen receptors in a clinically significant way in its unconverted form; its potency comes from its metabolites.

Cortisol, in contrast, does not "decline" with age in the same way. Daily cortisol production (20-25 mg/day) remains relatively stable or even increases during perimenopause, especially in its morning peak (cortisol awakening response, CAR). What does change is:

Circadian amplitude: the difference between the morning peak and the nighttime nadir can flatten in women with a history of sustained chronic stress
Adaptive response: the ability to recover quickly after a stimulus is reduced
Negative feedback sensitivity: the hippocampus, which normally brakes the HPA axis via GR, loses GR receptors under chronic glucocorticoid exposure → perverse positive-gain loop

The net result for Carmen at 47: the DHEA-S:cortisol ratio — her most precise indicator of "anabolic/catabolic balance" — collapses. Not necessarily because cortisol rises a lot, but because DHEA-S falls steadily and inevitably while cortisol is maintained or rises. This ratio is one of the best predictors of accelerated aging, metabolic risk, and immunosuppression in the precision medicine literature.

Cortisol is not "one thing" — the heterogeneity of the diurnal curve

This point is critical and generally ignored in the public conversation (and in many clinics) about stress and menopause. "High cortisol" or "low cortisol" does not capture reality. What matters is the full diurnal pattern — the shape of the curve, not an isolated point.

In perimenopausal women under chronic stress, there are at least two clearly distinguishable phenotypes:

Phenotype A — "Hyperreactive HPA": high CAR (pronounced morning peak, with very elevated cortisol 20-30 min after waking), preserved or increased diurnal amplitude, evening cortisol still relatively high. This phenotype is associated with RECENT or active chronic stress, type A personality, anxiety as the predominant symptom, sleep-onset insomnia (active mind at bedtime), intense hot flashes upon waking, morning irritability. It is the typical "acute overload" phenotype → the adrenal gland actively responds to a perceived state of alarm.

Phenotype B — "Collapsed Allostatic Load": flattened or absent CAR, flat curve all day (low or low-normal morning cortisol, with no reactive peak), relatively high evening cortisol (does not fall at night). This phenotype is associated with PROLONGED chronic stress or a history of burnout, fatigue as the predominant symptom, brain fog, hypersomnia or non-restorative sleep, subclinical depression, diffuse hot flashes with no clear pattern, decreased immune response. The "flat curve" does not mean exhausted adrenals — it means the system set point was recalibrated downward.

This distinction is not cosmetic: it implies completely different interventions, completely different risks, and completely different evolutionary trajectories. A woman with Phenotype A who receives "adaptogens for stress" may worsen if they raise her adrenocortical tone. A woman with Phenotype B who receives passive "stress reduction" may not improve because her problem is lack of rhythmic amplitude, not excess cortisol.

CYP19A1 in fat: the source of E2 no one maps

When the ovary starts to fail in perimenopause, ovarian E2 production becomes erratic. The body does not suddenly run out of estrogens — there is a backup system: peripheral aromatization. The enzyme CYP19A1 (aromatase) converts androgens (mainly androstenedione and testosterone) into estrogens (E1 and E2, respectively) in adipose tissue, liver, muscle, skin, and brain.

The problem lies in the substrate and in the site of conversion:

The main source of androgens in late perimenopause is no longer the ovary (which declines), but the adrenal gland via DHEA → androstenedione → E1
The main site of conversion shifts from the ovary (controlled, pulsatile, regulated by LH/FSH) to visceral fat (VAT, visceral adipose tissue) — continuous, unregulated, determined by tissue mass
VAT predominantly produces estrone (E1), not estradiol (E2). E1 has ~10x less estrogenic potency than E2 at classical receptors, and its metabolism generates 16α-hydroxylated metabolites with greater proliferative activity in breast tissue

The result for Carmen: as her ovary fails, the source of estrogens shifts from pulsatile ovarian E2 to continuous adipose E1. If Carmen has excess weight or accumulates visceral fat (common in perimenopausal women due to the change in fat distribution mediated by the decline in P4 and E2), her "peripheral estrogen reserve" exists — but it is qualitatively inferior and potentially riskier over the long term.

Chronically high cortisol (Phenotype A) worsens this: glucocorticoids upregulate CYP19A1 in VAT → more visceral aromatization → more E1 → more substrate for "lower-quality" estrogens. Insulin resistance, also exacerbated by chronic cortisol, does the same. The loop cortisol → IR → VAT aromatase → E1 is a self-feeding circuit that the literature treats in separate silos but that in Carmen operates as a single system.

The real concept of "allostasis" vs the myth of "adrenal fatigue"

"Adrenal fatigue" is a term popularized in 2001 (Wilson) that has no support in clinical endocrinology. The adrenal glands of a woman under chronic stress do NOT "burn out" — they continue producing cortisol. The real problem is upstream and systemic:

Downregulation of hippocampal GR — chronic exposure to glucocorticoids reduces GR density in the hippocampus → the brake on the HPA axis weakens → positive-gain loop → chronically higher or dysregulated cortisol
Setpoint recalibration — the HPA axis is a proportional + integral control system; under sustained stress, the reference point shifts, resulting in the flat curve (not "less production" but "disorganized production")
Tissue glucocorticoid resistance — some tissues stop responding normally to cortisol (Cardillo 2026 reaches this point: neutrophils with glucocorticoid resistance → cortisol cannot exert its anti-inflammatory function → low-grade systemic inflammation despite "normal" cortisol)

The correct concept is adrenal allostasis: the adrenal-HPA system has modified its operational set point (sometimes upward, sometimes downward) to accommodate historical stress load. The cost of this accommodation — the loss of flexibility, amplitude, and precision in the system — is the allostatic load.

Findings from recent papers

Finding 1 — DHEA-S has a transient rebound in early perimenopause (Sowers et al., 2012, SWAN)

The most counterintuitive finding in the literature on adrenal androgens during the menopausal transition comes from the longitudinal SWAN study. Sowers and collaborators, using longitudinal data from 120 stored samples from a multiethnic cohort, documented that mean DHEA-S concentrations in midlife women exhibit a positive inflection — that is, a transient rise — starting in early perimenopause and continuing into early postmenopause, before returning to baseline perimenopausal levels in late postmenopause.

This DHEA-S rebound in early perimenopause is compensatory: when the ovary starts to fail, the adrenal gland tries to fill the androgenic-estrogenic gap by transiently increasing its DHEA-S output. This compensation is concomitant with rises in testosterone, DHEA, androstenedione, and androstenediol.

Direct implication: Carmen at 47, in early perimenopause, may have DHEA-S that appears "normal" or even elevated on labs — masking the fact that the adrenal system is in active compensation, not stability. Cross-sectional analysis from a single measurement does not capture this compensatory effort. Allostatic load is detected in the TRAJECTORY, not in a point.

Finding 2 — Cortisol increases during perimenopause while E2 and P4 fall (Grub et al., 2021, Swiss Perimenopause Study)

The Swiss Perimenopause Study followed 127 perimenopausal women aged 40-56 years longitudinally for 13 months with saliva samples at multiple time points. The central finding: perimenopause is characterized by a decline in E2 and progesterone accompanied by an increase in cortisol.

This has a direct implication for the 4-arm model of luteal failure from L2.3, which now extends to anovulatory perimenopausal women: the woman who no longer has regular luteal phases loses the P4 that was the main functional antagonist of endometrial GR and systemic GR. The cortisol that rises in perimenopause now encounters a GR receptor without its natural competitor. Without P4 to balance GR, glucocorticoid signaling dominates in all sensitive tissues. The endometrium, bone, brain, immune system — all are targets of cortisol that no longer has luteal P4 modulation.

The cortisol curve in the Swiss Study was a population average — with high interindividual variance that the study did not stratify sufficiently. The two phenotypes (A and B described above) are superimposed in that average.

Finding 3 — CYP19A1 in visceral fat as a source of E1, and its upregulation by cortisol and insulin (Lee & Den Hartigh, 2025, Frontiers Endocrinology)

The most up-to-date paper on the topic (2025) integrates the mechanisms of adipose aromatization across the female life cycle. In postmenopausal/perimenopausal women, WAT (white adipose tissue) becomes the dominant source of endogenous estrogen production. Specifically:

HSD17B7 in VAT correlates positively with local E2 concentration and the E2:E1 ratio in VAT — women with more HSD17B7 in visceral fat produce a more favorable local estrogenic profile
In women with visceral obesity and adipose tissue dysfunction (metabolically active adiposity), there is a specific increase in CYP19A1 in VAT and an increased estrogen:androgen ratio
Adipose aromatase is upregulated by cortisol, insulin, and proinflammatory cytokines (IL-6, TNF-α) — all elevated in the context of perimenopausal allostatic load

The mechanism from Lee & Den Hartigh 2025 creates the bridge between adrenal allostasis (cortisol) → insulin resistance → low-grade inflammation → CYP19A1 upregulation in VAT → E1 as the predominant estrogenic source → a hormonal profile qualitatively inferior to ovarian E2.

For Carmen with visceral fat accumulated due to chronic cortisol (centripetal redistribution of fat in perimenopause), the endogenous estrogen profile has silently shifted toward E1-dominance, without any clinical analysis detecting it if only total serum E2 is measured.

Finding 4 — "Adrenal fatigue" is a false diagnosis; the real mechanism is tissue glucocorticoid resistance (Cardillo, 2026, Frontiers Endocrinology)

The Cardillo 2026 paper proposes a radically different framework: the "Adrenal Stress Index" — the salivary cortisol and DHEA-S test that functional medicine uses to diagnose "adrenal fatigue" — does not measure adrenal secretory capacity. It measures inflammatory adaptation state and allostatic load, mediated by neutrophil function.

The proposed mechanism: cortisol circulates mainly bound to corticosteroid-binding globulin (CBG). At sites of inflammation, neutrophil elastase releases active cortisol from CBG — this is the normal mechanism of local anti-inflammatory delivery. Under sustained chronic stress, neutrophil azurophilic granules are depleted ("neutrophil exhaustion") → local cortisol delivery deteriorates → inflamed tissues do not receive the anti-inflammatory signal → low-grade systemic inflammation persists despite "normal" serum or salivary cortisol levels.

The implication is fundamental: a perimenopausal woman with "normal cortisol on labs" may be in tissue glucocorticoid resistance if her neutrophil function is compromised by chronic stress. Her inflammatory symptoms (arthralgia, fatigue, brain fog, recurrent infections) do not respond to "relaxing" because the problem is not the amount of cortisol she produces — it is that this cortisol does not functionally reach where it is needed.

Finding 5 — DHEA-S is a component of the allostatic load index in perimenopausal women; migraineurs have 63% higher load (Alebna & Maleki, 2021, SWAN)

This paper is methodologically important because it formalizes the use of DHEA-S as a component of the allostatic load index (AL score) in perimenopausal women. In 2,105 perimenopausal women from SWAN, the AL score included: systolic/diastolic blood pressure, CRP, HDL, total cholesterol, waist-hip ratio, fasting glucose, triglycerides, and DHEA-S. Migraineurs had a 63% greater probability of elevated AL (OR=1.63, 95% CI 1.17-2.29).

The inclusion of DHEA-S in the AL index is scientifically solid because DHEA-S functions as a functional antagonist of the HPA axis: the higher the DHEA-S, the greater the glucocorticoid counterweight. Low DHEA-S contributes directly to high AL independently of absolute cortisol. Headache/migraine frequency thus takes on new meaning: it operates as a clinical proxy for total perimenopausal allostatic load.

Finding 6 — Elevated cortisol/DHEA-S ratio is associated with oxidative stress markers independently of age and BMI (Park, 2025, n=1,341)

The Park 2025 study in 1,341 Korean adults (mean age 52.6 ± 11.9 years — which includes perimenopausal women) documents that an elevated cortisol/DHEA-S ratio is negatively associated with hemoglobin and positively associated with GGT (gamma-glutamyl transferase, a marker of hepatic oxidative stress), independently of age and BMI.

Relevance: the association cortisol/DHEA-S ratio → oxidative stress → hepatic/systemic damage is a direct link between adrenal allostasis and accelerated aging. In perimenopausal women where the ratio deteriorates through the double pathway (DHEA-S in decline + sustained or elevated cortisol), systemic oxidative stress increases — which has direct consequences for:

Residual ovarian mitochondrial quality (even though the ovary is already failing, the remaining follicles are more vulnerable)
Granulosa function in the ovulatory cycles Carmen still has
Speed of global allostatic load accumulation

Complete molecular/endocrine mechanism

The "pentagon of buffers" in perimenopause

Perimenopause in Carmen is not the failure of one system — it is the coordinated failure of five interconnected buffers (extension of the 4 arms from L2.3):

BUFFER 1 — OVARIAN (collapsing erratically)
Pulsatile ovarian E2 ↓↓ → luteal P4 ↓↓ → irregular LH/FSH pulses
FSH ↑↑ (intermittent) → dysregulated KNDy (L2.1) → erratic cycles

BUFFER 2 — ADRENAL-DHEA (compensatory but limited and in decline)
DHEA-S ↑ transiently in early peri → then declines
DHEA → (CYP19A1, VAT) → E1 [10x less potent than E2]
DHEA → (17β-HSD, tissues) → Testosterone → (CYP19A1) → E2 [smaller amount]
DHEA → (5α-reductase) → DHT [androgenic, not estrogenic]
DHEA-S:Cortisol ratio ↓↓ → decreased net anabolism

BUFFER 3 — CORTISOL-HPA (variable: hyperreactive OR flattened)
CAR → follows pattern A (high) or pattern B (flat)
In Pattern A: cortisol ↑↑ → GR available without P4 competition → amplified catabolic effects
In Pattern B: disorganized cortisol → corrupted circadian signal → distinct metabolic consequences
Glucocorticoids → CYP19A1 in VAT ↑↑ → peripheral E1 ↑ (without ovarian P4 to balance it)

BUFFER 4 — MICROBIAL-VAGAL (compromised by perimenopausal dysbiosis)
↓ Lactobacillus/Bifidobacterium (L1.3) → ↓ progesterobolome (L1.2)
Biliary cortisol NOT converted to endogenous P4 → systemic cortisol:P4 ratio ↑
Low vagal-tone phenotype → silent vagal afference → PVN-CRH tonically disinhibited (L1.6)
↓ bacterial GABA (L1.6) → background anxiety/insomnia without apparent ovarian mechanism

BUFFER 5 — PHASE-GR-SENSOR (dysregulated without regular luteal P4)
Without regular luteal cycles, the HSD11B2→HSD11B1 switch (L2.3)
does not occur synchronously in endometrium/corpus luteum
FKBP51-PR (L2.3) induced by sustained cortisol → 
P4 that IS STILL produced in ovulatory cycles cannot exert its function correctly

The full pathway of perimenopausal adrenal allostasis

ADRENARCHE (25-30 years): DHEA-S at peak — maximum anabolic buffer
           ↓ (silent adrenopause 1-2%/year)
EARLY PERIMENOPAUSE (42-47 years):
  Ovary fails → FSH ↑ → adrenal gland "tries to compensate" → transient DHEA-S ↑ (Sowers 2012)
  Cortisol ↑ (Swiss Perimenopause Study) while E2/P4 fall
  DHEA-S:Cortisol ratio collapses (numerator falls, denominator rises)
           ↓
STORM ZONE (Carmen, 47):
  Cortisol without P4 competition → free GR in all tissues
  Glucocorticoids → CYP19A1 in VAT ↑ → DHEA → peripheral E1 (not E2)
  Low DHEA-S → estrobolome with less androgenic substrate for residual E2
  Cortisol:DHEA-S ratio ↑ → oxidative stress (Park 2025) → residual follicular mitochondrial damage
  Neutrophils under chronic stress → depleted azurophilic granules → tissue glucocorticoid resistance (Cardillo 2026)
  At the same time: peri dysbiosis ↓ progesterobolome → biliary cortisol not converted to P4
           ↓
BIFURCATED SYMPTOM PROFILE according to cortisol phenotype:
  Phenotype A (high CAR): anxiety, sleep-onset insomnia, AM hot flashes, irritability, pressure
  Phenotype B (flat CAR): fatigue, brain fog, subclinical depression, hypersomnia, diffuse hot flashes
           ↓
POSTMENOPAUSE (Rosa, 55): if no intervention occurs
  DHEA-S at 30-40% of original peak
  VAT as the ONLY source of peripheral E1 (ovarian E2 = 0)
  Cortisol without P4 or DHEA-S counterweight → high total allostatic load
  AL score (including low DHEA-S) predicts cardiovascular, cognitive, immunological risk

Peripheral aromatization: from DHEA to visceral fat E1

DHEA (adrenal) → sulfatase → DHEA (free)
                              ↓ CYP19A1 (VAT)
              Androstenedione (A4) → Estrone (E1)
                              ↓ 17β-HSD (VAT)
                         Estradiol (E2) [lower efficiency]

Regulators of CYP19A1 in VAT:
  Cortisol → ↑ (upregulates CYP19A1 promoter PI.4)
  Insulin → ↑ (via IGF-1R → PI3K → Sp1 in promoter)
  IL-6 → ↑ (via STAT3 → promoter PI.3)
  Tumor necrosis factor → ↑ (NF-κB)
  Androgens → ↑ (positive feedback)

HSD17B7 in VAT:
  Low expression → more E1, less E2 (unfavorable profile)
  High expression → more E2 (relatively more favorable profile)
  Modulated by: progesterone, DHEA, local microbiome

Cross-synthesis with previous findings

Connection with L1 (Estrobolome + Progesterobolome + Neurobolome)

The perimenopausal dysbiosis documented in L1.3 (Munyoki 2025, Kim/Benayoun 2026) now has a third direct endocrine consequence — beyond reducing E2 recycling (estrobolome, L1.1) and P4 recycling (progesterobolome, L1.2):

When adrenal DHEA-S is the main source of androgenic precursors, and when those androgens are converted to E2 in peripheral tissues, the intestinal estrobolome remains relevant — but its substrate has changed. It no longer recycles the pulsatile ovarian E2 of the young woman. Now it recycles the continuous peripheral E1 from VAT. An intact estrobolome in a perimenopausal woman with high VAT aromatase could, paradoxically, maintain high circulating E1 levels — an effect that is not necessarily protective.

Additionally: the biliary cortisol that was substrate for the progesterobolome (L1.2 — conversion of biliary cortisol to P4 by Eggerthella lenta) now has greater availability because more cortisol is produced. But perimenopausal dysbiosis simultaneously reduces the capacity to carry out that conversion. The system has more cortisolic substrate but less bacterial machinery to convert it into intestinal endogenous P4. Double loss.

Connection with L2.1 (KNDy and GnRH)

The perimenopausal "KNDy storm" (L2.1): GnRH pulses increased in frequency but erratic in amplitude, hot flashes, sweating, insomnia — now has a new additional modulator. DHEA has a partially inhibitory effect on the HPA axis (functional antiglucocorticoid). With DHEA-S in decline, the tonic inhibition that DHEA exerts on the PVN is reduced → more active CRH-PVN → more KNDy suppression when there are cortisol peaks, and more reactivity when cortisol falls. Carmen's KNDy oscillates more because one of its dampers (DHEA) is weakening.

Connection with L2.2 (Ovarian CRH)

L2.2 documented the ovarian inverted U (Gershon 2025): mild chronic stress facilitates ovulation via CRHR1; severe stress blocks it. In perimenopause, the CRHR1 of residual follicles is already in high-sensitivity compensatory mode (because there are fewer follicles and each one "matters more"). The cortisol additionally elevated by perimenopausal allostasis can derail even the few ovulatory cycles Carmen still has — pushing those cycles from the inverted-U zone toward the descending arm (suppression).

Connection with L2.3 (P4 vs cortisol in GR / FKBP51)

L2.3 established that GR-mediated FKBP51 can functionally deactivate PR even if serum P4 is "normal." In perimenopausal women, this mechanism becomes more devastating because:

P4 is already intermittent (only in ovulatory cycles)
When P4 IS present (a month when Carmen ovulated), if chronic cortisol induced FKBP51, real P4 does not exert its function → shortened and symptomatic luteal phase even though plasma progesterone is "adequate"
FKBP51 in the perimenopausal endometrium perpetuates a state of P4 insensitivity that blood testing does not detect

This mechanism explains why some perimenopausal women have progesterone-deficiency symptoms with levels "in range" — the problem is not quantity but downstream signaling.

The "fifth arm" of perimenopausal luteal failure

Extending the 4-arm model from L2.3 to the perimenopausal context:

ARM 1: P4 substrate deficit (ovarian + broken progesterobolome + low DHEA-S)
ARM 2: Amplified inflammation (NF-κB, IL-1β, dysbiosis-LPS + gluconeogenic cortisol)
ARM 3: GR-phase dyssynchrony (HSD11B2/HSD11B1 switch without regular luteal cycle)
ARM 4: Collapse of the microbial-progesterobolomic buffer (inherited L1.2)
ARM 5 (NEW): Tissue glucocorticoid resistance (Cardillo 2026) — 
         tissues do not respond normally to cortisol → chronic low-grade inflammation
         which in turn perpetuates arms 2, 3, and 4 in a feedback loop

Lua Labs hypotheses

Hypothesis 16: "The diurnal cortisol pattern defines two biotypes of perimenopausal adrenal allostasis with qualitatively distinct evolutionary trajectories, symptom profiles, and therapeutic responses — discriminable without serum biomarkers"

Statement: In perimenopausal women (42-52 years), the diurnal cortisol-curve phenotype — high (pronounced CAR + preserved amplitude) vs flat (flattened CAR + flat curve) — predicts dominant symptoms, allostatic-load trajectory, and differential response to behavioral-nutritional interventions, and could be discriminated through structured self-report of energy, sleep, stress perception, and eating patterns, without the need for salivary or serum cortisol.

Proposed mechanism:

The diurnal cortisol phenotype is the result of the integration of chronic stress exposure history, negative hippocampal feedback capacity (GR hippocampal density — reduced in prolonged burnout), and the state of the sympathetic-adrenal axis (catecholamines → CRH → ACTH). It is not random: it depends on the nature of the stressor (repetitive acute vs sustained chronic vs mixed), coping profile (active/reactive vs dissociative/passive — inherited L2.1: Klimek 2025 and coping style), and NR3C1 and FKBP5 polymorphisms.

Phenotype A — "Hyperreactive HPA" (Carmen with active caregiving, high load, type A personality):

Elevated CAR (20-30% above age average)
Preserved diurnal amplitude
Symptoms: anxiety, sleep-onset insomnia, intense AM hot flashes, evening irritability, racing mind
Metabolic risk: hypertension, elevated fasting glucose, accumulating visceral adiposity
Optimal response: interventions that REDUCE CAR (magnesium, L-theanine, caffeine restriction, slow breathing) and RESTORE the evening decline (AM Z2 exercise, regular sleep)

Phenotype B — "Collapsed Allostatic Load" (Carmen with a long history, years of exhaustion, caregiving burnout):

Flattened or absent CAR
Flat curve all day (constant low-moderate)
Symptoms: deep fatigue with no apparent origin, brain fog, hypersomnia, subclinical depression, diffuse hot flashes without pattern, recurrent infections
Metabolic risk: insulin resistance, dyslipidemia, increased immunological risk (due to tissue glucocorticoid resistance)
Optimal response: interventions that RESTORE circadian amplitude (morning sunlight 7-8am, timing-specific AM exercise, protein + tyrosine in the morning, consolidated sleep rhythm)

Proposed digital discriminators (without measured cortisol):

Question 1: "Is your energy higher in the morning or does it fall during the day?" 
  High morning/falls late → Phenotype A
  Flat-low all day → Phenotype B

Question 2: "Is your insomnia sleep-onset insomnia (you cannot fall asleep) or early waking/non-restorative sleep?"
  Sleep-onset → Phenotype A
  Early waking/non-restorative → Phenotype B

Question 3: "Does your stress feel like a 'racing mind' or like 'everything overwhelms me but I cannot react'?"
  Racing mind → Phenotype A
  Overwhelmed-shut down → Phenotype B

Question 4: "Are your hot flashes/heat sensations more intense in the morning?"
  Yes → Phenotype A
  Diffuse all day or nocturnal → Phenotype B

vagal-tone phenotype (inherited L1.6) as modulator: low vagal-tone phenotype amplifies phenotype B; high vagal-tone phenotype can attenuate phenotype A.
PSS-4 as signal: high PSS-4 + high AM energy → A; medium PSS-4 + low energy all day → B.

Confidence level: Medium — The differentiation of two diurnal cortisol phenotypes is documented in the burnout, HPA, and perimenopausal literature (multiple sources). Digital discriminability without measured cortisol is the original part that requires validation. The underlying mechanisms of each phenotype are well documented.

How to validate:

The discriminability of the phenotype from self-report is the part that requires formal validation:

Cross-sectional observational, n=200 perimenopausal (42-55), salivary cortisol at 4 points of the day + symptom questionnaire + 2-week actigraphy → sensitivity/specificity of the self-report classification against the real cortisol curve.
Then 12-week intervention: randomize within each phenotype to A-specific vs B-specific vs control intervention → does phenotype-intervention matching predict a better outcome than a generic intervention?

Limitations:

The two phenotypes are extremes of a continuum — many women will have mixed or variable profiles
The 4-question digital discrimination is crude — it needs iterative refinement with real data
Cycle week and situational stress affect the cortisol curve — a single measurement (or questionnaire) may not capture the stable phenotype
NR3C1, FKBP5, CRHR1 polymorphisms modulate the phenotype — without genotyping, there is overlap

Candidate formulation: phenotype-bifurcated adrenal-behavioral approach (Carmen)

Compounds and mechanisms:

Common foundation (both phenotypes):

Magnesium (dark cacao/pumpkin/seeds/mineral water) — cofactor in 300 enzymatic reactions; reduces HPA-axis hyperreactivity, cofactor for adrenal DHEA synthesis, antagonizes NMDA in CNS (reduces neuronal excitability). 300-400 mg/day from foods.
Vitamin C (kiwi, green/red chili, guava, citrus) — the highest ascorbate concentration in the entire organism is in the adrenal cortex; it is a cofactor in cortisol synthesis (dopamine-β-hydroxylase + adrenal enzymes). With low DHEA, adrenal "support" is critical. 200-400 mg/day from food.
B5 (pantothenic acid) (avocado, mushrooms, egg yolk) — CoA cofactor in adrenal hormone synthesis. Without B5, both cortisol and DHEA are reduced.
Fermentable fiber + LATAM ferments — restores progesterobolome (L1.2): biliary cortisol → endogenous P4 if Eggerthella lenta is present. Nopal, beans, traditional fermented food.
Adequate tryptophan (chia, egg, legumes) — serotonin→melatonin precursor. The circadian rhythm of cortisol depends on the nighttime melatonin signal.

Arm A — Hyperreactive HPA Phenotype (reduce peak, restore decline):

L-theanine (matcha, green tea, supplement) — antagonizes glutamate in CNS, reduces CAR without suppressing daytime function. Evidence: reduces cortisol reactive to psychological stress in multiple RCTs.
Glycine (pre-sleep, 2-3g) — potentiates GABA-A inhibition in hypothalamus → reduces nocturnal CRH-PVN activation → decreases morning cortisol elevation.
Post-12pm caffeine restriction — caffeine blocks adenosine and potentiates CAR; in Phenotype A, it amplifies a peak that is already excessive.
Morning Z2 exercise (30-45min) — reduces reactive CAR by using already elevated cortisol; improves insulin sensitivity + visceral fat redistribution → less substrate for CYP19A1 VAT.
Slow breathing 4-6 cycles/min (5-10 min before sleep) — activates parasympathetic tone, reduces sympathetic NE, restores HRV → vagal tone (L1.6/vagal-tone phenotype).

Arm B — Collapsed Allostatic Load Phenotype (restore circadian amplitude):

Morning sunlight 7-8am (10-15 min) — the light signal is the only stimulus that physiologically restores CAR. Without morning light, the cortisol peak does not trigger correctly. Many perimenopausal women in Mexico have insufficient morning exposure (indoor work, screens).
Protein + tyrosine at breakfast (egg, chicken, legumes) — tyrosine is a precursor of catecholamines (DA, NE) that facilitate the morning sympathetic response coordinated with cortisol. A protein-rich breakfast amplifies the physiological CAR response.
AM exercise without fasting — Phenotype B does not tolerate extended fasting (it further reduces morning cortisol); eating + AM exercise restores the morning autonomic axis.
Adaptogens (food-based or herbal, with evidence): Ashwagandha (KSM-66, 300-600mg) — mechanism: reduces ACTH-mediated cortisol, restores HPA amplitude, RCT evidence in burnout and fatigue. Rhodiola rosea (150-200mg salidroside) — activates Nrf2, reduces perceived fatigue, improves cognitive function under allostatic load. Arm B only — in arm A, they may inadvertently amplify.
Consolidated sleep 22:30-06:30 — Phenotype B needs sleep consolidation so nocturnal cortisol (REM phase) restores the CAR signal upon waking.

Target population: Carmen (47, perimenopause). Rosa (55, postmenopause) when L8 is complete.

Regulatory status: Food-based (magnesium, B5, vitamin C, tryptophan, glycine, L-theanine) = GRAS without discussion. Ashwagandha/Rhodiola = herbal supplement with RCT evidence (requires note "this is research, not prescription"). CYP19A1-VAT pathway: no pharmacological intervention proposed.

Validation required: 2×2 crossover design (phenotype × intervention). N=120 perimenopausal women, 16 weeks. Endpoints: salivary cortisol at 4 points (weeks 0, 8, 16), DHEA-S, Greene Climacteric Score, and nocturnal HRV. The hypothesis to be falsified is that matching cortisol phenotype to intervention outperforms generic intervention in reducing climacteric symptoms and in normalizing the diurnal curve.

Individual variability

Genetics

FKBP5 rs1360780 T/T (inherited L2.1/L2.3): greater HPA reactivity → higher probability of Phenotype A; but also higher risk of "collapse into B" under sustained stress due to deeper feedback suppression
NR3C1 BclI G/G: GR hypersensitive → lower cortisol set point for effects → more symptoms with "normal" cortisol
CYP19A1 variants rs2289666, rs749292: polymorphisms that affect aromatase expression in VAT → variability in the quality of peripheral E1 produced
HSD17B7 variants: determine the E2:E1 ratio in VAT; hypoactive variants → more E1, less E2 in visceral fat
COMT Met/Met: sustained catecholamines → prolonged CRH-PVN → contributes to Phenotype A in stressed women

Epigenetics

ACEs ≥ 4 (childhood adversity): lifelong FKBP5 hypomethylation → greater basal HPA reactivity → predisposes to early Phenotype A, but also to faster B collapse
History of night shifts: persistent circadian dyssynchrony → reduced cortisol amplitude → Phenotype B more likely even without "subjective stress"
LATAM-urban acculturation (inherited L1.4): loss of Prevotella + dysbiosis → compromised progesterobolome → less cortisol→P4 conversion → more "free" cortisol available for systemic effects

LATAM environmental-contextual

Caregiving load (care of dependents): the main chronic stressor documented in perimenopausal LATAM women (L2.1). Determines the DURATION of stress more than its intensity — a critical factor for whether the woman develops Phenotype A (new/recent load) or Phenotype B (years-long load).
Greater visceral adiposity by BMI in Latinas: Latina women tend to accumulate more visceral fat at the same BMI than Caucasian women — potentially implying greater CYP19A1-VAT activity → more peripheral E1 at a lower apparent "metabolic load." Literature gap: there are no specific data on VAT aromatization in Mexican women.
Nutrition transition (L1.4): loss of ancestral diet → loss of fermentable fiber → broken progesterobolome → less endogenous intestinal P4 → more cortisol free in the system.

Lateral connection — thyroid axis

Elevated cortisol in perimenopause (in both phenotypes, but more pronounced in Phenotype A) directly inhibits the thyroid axis at three points:

TSH secretion at the pituitary level — the same CRH-PVN that activates the HPA axis suppresses hypothalamic TRH
T4→T3 conversion — inhibition of type I and II 5'-deiodinase in liver and kidney
T3 binding to its receptor (TRβ) in the presence of high glucocorticoids — interference at the chromatin level, analogous to the FKBP51-PR mechanism described in L2.3

Result: a perimenopausal woman with Phenotype A may present "normal" TSH and clear symptoms of functional hypothyroidism, because cortisol discreetly suppresses TRH and keeps TSH artificially "in range" while free T3 falls. The clinician sees normal TSH and rules out the thyroid; the thyroid axis remains an open front that no isolated marker resolves.

In addition, DHEA has documented immunomodulatory effects that protect against autoimmunity, including potentially autoimmune thyroiditis (Hashimoto). With DHEA-S in decline during perimenopause, that protection also diminishes. Literature gap: there are no controlled human trials directly connecting DHEA-S decline with Hashimoto incidence in perimenopause, but the immunomodulatory mechanism is well characterized. It is a plausible causal connection that future research will need to prove or refute.