Substance and claimed effects

Perimenopause is the multi-year transition from regular ovulatory cycles to the final menstrual period (FMP) and the 12 months that follow it. The STRAW+10 staging system formalises it as two stages — early menopausal transition (cycle length differing by ≥7 days from personal normal, recurrent within 10 cycles) and late menopausal transition (an interval of amenorrhea ≥60 days) — driven by accelerating depletion of the ovarian follicle pool and increasingly erratic estradiol with rising FSH Harlow et al. 2012. Median duration is approximately 4 years, with substantial individual variability (range 1–10+ years); typical age at FMP in the United States is 51–52 Santoro et al. 2021. The transition produces a coordinated symptom pattern across six domains tracked in this entry: menstrual cycle changes, vasomotor symptoms (hot flashes, night sweats), sleep disruption, mood changes (depression risk roughly doubles), cognitive changes (verbal memory and processing speed dips), and body composition shifts (central adiposity, lean mass loss). The entry is scoped as a literacy / recognition entry (action: know) — naming what the transition is, what it does, and how its symptom pattern reads in real time. Downstream treatment entries (menopausal hormone therapy, non-hormonal vasomotor pharmacotherapy, sleep apnea evaluation in midlife women) are not in scope here; they belong in their own entries.

Evidence by addressing question

Mechanism

The underlying physiology is ovarian follicular depletion. Across reproductive life the antral follicle count declines roughly log-linearly; by the early 40s the pool is small enough that the hypothalamic-pituitary-ovarian axis can no longer maintain stable cyclic estradiol production Harlow et al. 2012. Two physiological hallmarks distinguish perimenopause from later reproductive aging: (1) estradiol becomes more variable rather than uniformly low — peak follicular-phase estradiol is often higher than premenopausal values for years before declining, producing wide swings within and across cycles; (2) FSH rises in response to falling inhibin B, but remains variable rather than monotonically elevated until late perimenopause Santoro et al. 2021. This variability — not pure deficiency — is the mechanism that ties the symptom map together. Vasomotor symptoms appear to track estradiol withdrawal episodes rather than absolute levels (the central thermoregulatory window narrows as estrogen falls, so small body-temperature elevations trigger flushing) Maslow & Boggs 2018. Mood vulnerability tracks the magnitude of estradiol fluctuation rather than its mean, with progesterone withdrawal contributing in late luteal phases Joffe et al. 2020. Sleep fragmentation is partly direct (hot flashes wake the sleeper) and partly indirect (estrogen modulates GABA and serotonin signalling in sleep-regulatory nuclei) Kravitz et al. 2008. Cognitive symptoms appear to reflect the transition itself — verbal memory transiently dips during perimenopause and recovers in postmenopause in longitudinal data, consistent with a state effect of estradiol fluctuation rather than permanent damage Greendale et al. 2009.

Evidence

The symptom map is anchored by the Study of Women's Health Across the Nation (SWAN), a 25+ year multi-ethnic longitudinal cohort of ~3,300 women followed from premenopause through postmenopause. SWAN established the duration, prevalence, and ethnic patterning of vasomotor symptoms: median total VMS duration is 7.4 years, with median post-FMP persistence of 4.5 years Avis et al. 2015. Approximately 75–80% of women experience VMS at some point in the transition; African American women have the longest median duration (~10.1 years), Japanese and Chinese American women the shortest (~4.8–5.4 years) Avis et al. 2015. The Penn Ovarian Aging Study independently replicated long VMS duration in a separate cohort Freeman et al. 2014. SWAN data also documented the depression-risk elevation: women with no prior depression history have approximately 2–4× increased odds of a new depressive episode during perimenopause compared to premenopause, with risk concentrated in the late perimenopausal stage Bromberger et al. 2011. The cognitive dip — small but consistent decrements in verbal memory and processing speed — was documented in the SWAN cognitive substudy across ~2,300 women, with effect sizes typically 0.1–0.2 standard deviations and recovery in postmenopause Greendale et al. 2009. Body composition trajectories were quantified in SWAN's DXA substudy: lean mass declines ~0.2 kg/year accelerating around the FMP, fat mass increases especially centrally, and the rate of bone loss roughly doubles in the 12 months before through 24 months after FMP Greendale et al. 2019 Matheson et al. 2003. The El Khoudary SWAN progress report integrates these findings into the cardiovascular and metabolic trajectory, including unfavourable LDL/HDL shifts and increased visceral adiposity over the transition El Khoudary et al. 2016. Clinical practice guidelines summarising the evidence — Endocrine Society Stuenkel et al. 2015, NAMS 2022 position statement NAMS 2022, and the perimenopausal depression guideline Maki et al. 2019 — all converge on the same symptom map and the same broad treatment framework.

Misconceptions

Three recurring misconceptions distort symptom recognition. (1) "Perimenopause is when your periods stop." The transition begins with cycle changes (shorter cycles initially as the luteal phase contracts, then increasingly variable intervals), not amenorrhea — amenorrhea defines the late stage. Median time from first cycle irregularity to FMP is ~4 years Harlow et al. 2012. (2) "Hot flashes mean it's almost over." SWAN data flip this: hot flashes typically persist a median of 4.5 years after the FMP, with one-third of women symptomatic 10+ years post-FMP Avis et al. 2015. (3) "Mood symptoms in midlife are situational, not hormonal." The doubled depression risk in the transition is observed after controlling for life-stress measures and prior depression history; the hormone-fluctuation contribution is independent Bromberger et al. 2011. A fourth, weaker misconception: that an FSH blood test can confirm perimenopause. Because FSH is highly variable within and between cycles in the transition, a single elevated value can occur premenopausally and a single normal value can occur in late perimenopause; STRAW+10 deliberately bases staging on bleeding pattern rather than FSH for this reason Harlow et al. 2012.

Audience and population variability

Audience is anatomically female; the entry's primary readership is ages 40–59, with onset in the late 30s for a minority and persistence into the 60s for some. Symptom prevalence and duration vary substantially by demographic and baseline factors. Race/ethnicity: African American women report longer VMS duration and higher symptom burden; East Asian American women shorter and lower Avis et al. 2015. Body composition: higher BMI is associated with more VMS in the perimenopausal stage (the older "obesity protects" framing was a postmenopausal artifact) Maslow & Boggs 2018. Prior reproductive history: women with prior premenstrual mood disorders or postpartum depression have higher risk of perimenopausal depressive episodes Maki et al. 2019. Surgical and induced menopause: bilateral oophorectomy before natural FMP produces an acute estrogen-withdrawal syndrome with more severe VMS and accelerated bone loss, and is sometimes confused with natural perimenopause but follows a different trajectory Stuenkel et al. 2015. Smoking: associated with earlier FMP (~1–2 years) and higher VMS burden Santoro et al. 2021.

Failure modes

The most common failure of recognition is symptom mis-attribution. A woman in early perimenopause with new-onset insomnia, anxiety, and brain fog presenting to primary care is frequently worked up for thyroid dysfunction, primary depression, or "stress" without the menopause transition being raised — particularly when she is in her early 40s and still cycling. Cognitive symptoms get pathologised as early dementia rather than recognised as the SWAN-documented transient transition effect Greendale et al. 2009. Sleep symptoms get attributed to sleep hygiene or stress when the underlying mechanism is nocturnal VMS or increased obstructive sleep apnea risk (OSA prevalence rises post-FMP, partly via fat redistribution and partly via loss of progesterone's upper-airway tone) Park et al. 2020. The second failure mode is over-attribution: assuming every symptom in a 45-year-old is perimenopause when an actual primary cause (thyroid disease, sleep apnea, iron deficiency, depression unrelated to hormonal fluctuation) is present. The third is the FSH test trap — clinicians or patients order serum FSH, get a normal value mid-cycle, and conclude perimenopause has been ruled out Harlow et al. 2012. The fourth is conflating perimenopause with menopause: treatment-decision thinking that defers all action until the FMP misses the highest-symptom window (late perimenopause) and the bone-loss acceleration window (12 months pre-FMP through 24 months post-FMP) Greendale et al. 2019 Matheson et al. 2003.

Stakes

The stakes of un-recognized perimenopause are concrete and tractable. Untreated vasomotor symptoms persist a median of 7.4 years, with attendant sleep fragmentation, daytime fatigue, and emerging evidence of cardiovascular signal: frequent or persistent hot flashes are associated with greater carotid intima-media thickness and adverse subclinical cardiovascular markers in SWAN Thurston et al. 2016. The depression-risk window is most pronounced in late perimenopause and resolves in postmenopause in most women, but a missed episode in this window often becomes a chronic recurrent depression that outlasts the transition Maki et al. 2019. Bone loss is the highest-stakes silent consequence: trabecular bone loss accelerates to ~2× the premenopausal rate in the late perimenopause / early postmenopause window, and bone density lost in this window is harder to restore later Matheson et al. 2003. Genitourinary syndrome of menopause (vaginal dryness, dyspareunia, recurrent UTI, urinary urgency) is progressive and does not spontaneously remit; without recognition it produces a slow erosion of sexual function and urinary continence over years Portman & Gass 2014.

Payoff

Recognition itself produces immediate reduction in alarm and self-mis-attribution (brain fog is not early dementia; mood instability is not "I've become someone unstable"; cycle chaos is not a missed cancer). It opens the door to evidence-based treatment of each symptom domain: menopausal hormone therapy for vasomotor and genitourinary symptoms in appropriate candidates within the timing-hypothesis window (within ~10 years of FMP, before age 60) NAMS 2022 Manson et al. 2017; non-hormonal VMS pharmacotherapy (SSRIs, gabapentin, fezolinetant) for women with contraindications Stuenkel et al. 2015; CBT and antidepressants for perimenopausal depression with stronger evidence than in premenopausal depression Maki et al. 2019; vaginal estrogen (essentially no systemic absorption, safe even in many breast cancer survivors) for GSM Portman & Gass 2014; baseline DXA and bone-density preservation planning in the high-loss window Matheson et al. 2003. The timing-hypothesis literature suggests cardiovascular and possibly cognitive benefits of MHT initiated early in the transition that are not present when initiated >10 years post-FMP NAMS 2022.

Out-of-scope

Not covered in this entry but adjacent: the protocol-level detail of menopausal hormone therapy (formulations, routes, dosing, candidate selection) belongs in a dedicated MHT entry; the bone-density screening protocol (DXA cadence, T-score thresholds, FRAX) belongs in a screening entry; sleep apnea evaluation in midlife women belongs in the apnea entry; the genitourinary syndrome of menopause as a specific treatable condition belongs in its own entry. Premature ovarian insufficiency (POI; menopause before age 40) is a distinct clinical entity with different management and is out of scope.

The credibility range

Optimist case

The symptom map is settled science. STRAW+10 staging is the international standard and is reproducible across cohorts Harlow et al. 2012. SWAN provides 25+ years of longitudinal data documenting the prevalence, duration, and ethnic patterning of each symptom domain El Khoudary et al. 2016. Treatment of the recognised symptoms is high-evidence: MHT reduces VMS by ~75% in RCTs and treats GSM completely NAMS 2022; the perimenopausal depression guideline is built on multiple replicated trials Maki et al. 2019. The literacy intervention this entry represents is highly leveraged: a single 10-minute read can shift years of trajectory by getting the reader to the right diagnostic question.

Skeptic case

Symptom-attribution to perimenopause has expanded faster than the evidence in popular media. Many "perimenopause" symptoms — fatigue, weight gain, brain fog, low libido, joint aches — also have base rates that rise in midlife independent of hormonal transition, and longitudinal studies have variable success isolating the transition's contribution from aging-per-se. The cognitive dip is statistically real but small (effect sizes 0.1–0.2 SD) and recovers postmenopausally Greendale et al. 2009; popular framing as "brain fog that lasts forever" is not what SWAN actually shows. Treatment benefits — particularly MHT's cardiovascular and cognitive effects — remain contested between the WHI generation of trials and the timing-hypothesis re-analyses; the field is closer to consensus on safety in the early-transition window than on benefit for indications other than VMS and GSM Manson et al. 2017. Body-composition and bone-loss changes are partly attributable to chronological aging and are not exclusively a transition phenomenon, though SWAN does isolate a transition-specific acceleration Greendale et al. 2019.

Author's call

The core symptom map — vasomotor, sleep, mood, cognition, cycle, body composition, GSM, bone — is well-evidenced, replicated across cohorts, and not seriously contested. Quantitative claims in this entry stay close to SWAN-documented numbers (median VMS duration 7.4 years, ~2× depression risk, small cognitive dip with recovery). The entry is conservative on which symptoms are causally attributable to the transition versus midlife co-occurring, and conservative on downstream treatment specifics (which are scoped to other entries). evidence rates a 5 (multiple large longitudinal cohorts plus replicated guidelines). controversy rates 2 — the symptom map itself is non-controversial; MHT use is contested but is not the substance of this entry.

Stakeholder and incentive map

Pushing the topic: NAMS and the broader menopause-specialist clinical community (legitimate; pushing for under-recognized condition awareness); a growing online "perimenopause influencer" sector and direct-to-consumer telehealth (mixed incentives; many are useful but some over-attribute and over-prescribe); pharmaceutical interests in MHT and newer non-hormonal VMS drugs (fezolinetant launch 2023). Pushing back: a generation of clinicians trained during the post-WHI MHT pullback who under-treat menopausal symptoms; primary care time pressure that makes hormonal symptom workup less attractive than handing out an SSRI; lingering cultural framing of menopause as "natural, just endure it." The reader needs to be aware that both over- and under-attribution failure modes exist.

Population variability

See "Audience and population variability" subsection above. Key variables: race/ethnicity (substantial VMS-duration differences), BMI, smoking, prior reproductive psychiatric history, surgical vs natural menopause, age at onset. The entry's central tendencies (~4 year transition, ~75–80% VMS prevalence, ~2× depression risk, small cognitive dip) are reasonable defaults but each individual reader's experience will vary along these axes.

Knowledge gaps

Several gaps remain. (1) The mechanistic link between perimenopausal estradiol fluctuation and mood is biologically plausible and observed but not as cleanly mapped as the VMS pathway; whether estradiol-stabilisation therapies (e.g. transdermal estradiol patches in late perimenopause) prevent depressive episodes in high-risk women is still being evaluated. (2) Long-term cognitive trajectory after the transition is mostly reassuring (SWAN shows recovery) but whether subgroups exist who do not recover — and whether early MHT alters that trajectory — is not settled. (3) The contribution of perimenopausal vasomotor burden to long-term cardiovascular risk is suggestive (Thurston SWAN data) but not yet at the level of guideline action. (4) Best biomarker for staging late perimenopause beyond bleeding pattern (AMH, INSL3, transvaginal antral follicle count) — useful in research, not yet standard of care.

Scope and the brief. The brief named cycles, sleep, mood, cognition, body composition, and vasomotor symptoms. All six are covered as the six-domain map in the evidence section. The entry is framed as a recognition entry (action: know); downstream treatment specifics are deliberately out of scope to avoid blurring the literacy job with hormone-therapy prescribing, which warrants its own entry.

Hard scoping calls.

• No protocol section. A literacy entry has no protocol to give. The closest thing — "raise it with your doctor, get a bleeding-pattern history, treat by symptom" — is woven through payoff rather than given its own section, because pulling it into a protocol would imply a do-this-now action the entry isn't actually authorising.
• No contraindications section. The recognition act has no contraindications. Treatment contraindications belong with the treatment entries.
• Genitourinary syndrome of menopause (GSM) is mentioned but not given its own subsection in the symptom map, because it's progressive rather than transition-bounded and deserves a standalone entry. It appears in stakes and payoff where it's load-bearing.

Separate-entry candidates flagged for the backlog.

• Menopausal hormone therapy (formulations, candidate selection, timing hypothesis, contraindications).
• Non-hormonal vasomotor treatment (SSRIs, gabapentin, fezolinetant).
• Genitourinary syndrome of menopause as its own treatable condition.
• Bone density screening cadence in midlife women (DXA timing, T-score thresholds, FRAX).
• Premature ovarian insufficiency (menopause before 40).
• Obstructive sleep apnea in midlife women specifically — different presentation and lower awareness than the male-coded sleep-apnea entry.

Rating difficulties.

• longevity: 2 was the hardest call. The bone-loss window and cardiovascular trajectory are real and consequential, but the literacy substance itself doesn't change mortality — downstream interventions do. A 2 ("small additive") felt honest; a 3 would have implicitly claimed credit for treatments scored on their own entries.
• focus: 2 reflects the SWAN cognitive dip's modest effect size (0.1-0.2 SD) and its recovery in postmenopause. Tempting to score higher because brain fog dominates lived-experience reports, but the actual measured dip is small and the score has to reflect the substance honestly.
• cost_burden: 0 and effort_burden: 1 reflect the literacy entry specifically — recognising the pattern is free and light. The downstream treatments carry real burden and will be scored where they live.

Audience scoping. Set to female / 40-59. Excluding 18-39 because the entry's substance is normal-onset perimenopause; premature ovarian insufficiency is flagged separately. Late-30s onset exists but is rare enough that scoping to 40-59 reflects the actual primary audience without misleading. 60+ excluded because that population is postmenopausal, not in transition — though they may read for retrospective recognition.

Future link candidates. Once the entries above exist, wire cross-links from the symptom map into each.