Polysomnography (Sleep Study): Sex- and Cycle-Related Differences, Normal Ranges, and Clinical Optimization

What a Polysomnography Actually Measures

Polysomnography (PSG) is the reference-standard overnight test for sleep disorders. It records electroencephalography (EEG), electrooculography, chin and leg electromyography, oral-nasal airflow, thoracic and abdominal effort, pulse oximetry, and body position simultaneously. The resulting data let a sleep technologist score every 30-second epoch by sleep stage and flag each respiratory event.

The American Academy of Sleep Medicine (AASM) 2023 scoring manual defines four stages: N1 (light), N2 (intermediate), N3 (slow-wave or deep), and REM. Together these stages cycle roughly every 90 minutes across the night, with more N3 in the first half and more REM in the second half.

Core PSG Parameters and Their Reference Ranges

| Parameter | Normal Adult Range | Clinical Action Threshold | |---|---|---| | AHI (total) | <5 events/hr | 5+ triggers OSA diagnosis | | Sleep efficiency | 85 to 100% | <80% warrants evaluation | | Sleep onset latency | 10 to 20 minutes | >30 min suggests insomnia | | REM onset latency | 90 to 120 minutes | <20 min suggests narcolepsy or depression | | N3 (slow-wave) | 13 to 23% | <10% linked to metabolic risk | | Oxygen nadir (SpO2) | 95% or higher | <88% for 5+ min drives CPAP urgency | | Arousal index | <15 per hour | 15+ associated with daytime impairment |

The AASM notes that these ranges are population-derived and that a single night of PSG carries night-to-night variability of roughly 20 to 30% in AHI. A second study may be indicated when clinical suspicion and the first result diverge. [1]

Why OSA Classification Matters Beyond Snoring

OSA is not simply a nuisance. Moderate-to-severe OSA (AHI 15 or higher) carries a hazard ratio of 2.14 for incident hypertension over 4 years (Sleep Heart Health Study, N=6,132) [2] and associates with a 65% higher rate of type 2 diabetes compared to non-OSA controls in the Wisconsin Sleep Cohort. [3] Those downstream metabolic effects interact directly with hormone therapy decisions for every HealthRX patient category.

Sex Differences in PSG Architecture

Men and women show consistent, reproducible differences in PSG results even when matched for age and BMI. Women obtain more slow-wave sleep, have longer total sleep time, and report subjectively worse sleep quality despite objectively better architecture. This mismatch contributes to under-referral for formal testing.

Slow-Wave Sleep and REM Distribution

A meta-analysis of 65 PSG studies (Reyner and Horne, updated by Ehlers et al.) found that women spend, on average, 5 to 8 percentage points more of the night in N3 compared to age-matched men. [4] REM percentage is comparable between sexes in young adults, but men lose REM faster with age: by age 60, men average 17% REM vs. 21% in women of the same age.

Testosterone appears to reduce N3 directly. Studies in transmen receiving testosterone therapy show a measurable drop in slow-wave sleep within 3 to 6 months of starting treatment. [5] This finding has direct implications for any patient on high-dose TRT.

OSA Prevalence, Symptom Presentation, and Under-Diagnosis

Before menopause, OSA affects roughly 3 to 5% of women compared to 9 to 14% of men in community samples. [6] After menopause, female OSA rates approach or match male rates. The hormonal mechanism is largely progesterone-dependent: progesterone stimulates hypoglossal nerve activity, stiffens pharyngeal dilator muscles, and raises the hypercapnic ventilatory response threshold, all of which reduce collapsibility.

Women with OSA are also more likely to present with insomnia complaints, fatigue, and depression rather than loud snoring, and they produce more hypopneas (partial airflow reduction) than frank apneas (complete cessation). Standard AHI scoring historically under-counted hypopneas using the older 4% desaturation rule. The AASM shifted to a 3% desaturation threshold in 2012, which increased female OSA detection by roughly 33% in one re-scoring study. [7]

The Menstrual Cycle and Sleep Architecture

Sleep is not static across the menstrual cycle. PSG studies in healthy eumenorrheic women demonstrate phase-dependent shifts in every major parameter.

Follicular Phase (Days 1 to 13)

Estrogen is rising but progesterone remains low. Sleep efficiency averages 87 to 89% in this phase, N3 is near its monthly nadir, and core body temperature is at its lowest point before ovulation. Some women report lighter, more fragmented sleep in the late follicular phase, which PSG confirms as a modest increase in wake-after-sleep-onset (WASO). [8]

Luteal Phase (Days 15 to 28) and Progesterone's Effect

Progesterone peaks at 15 to 25 ng/mL in the mid-luteal phase. PSG studies show a concurrent increase in N3 by 5 to 12 percentage points and a decrease in REM by 4 to 8 percentage points. [9] Core body temperature rises 0.3 to 0.5°C in the luteal phase, which reduces sleep onset latency paradoxically but fragments the second half of the night as temperature stays elevated.

AHI in healthy pre-menopausal women drops measurably in the luteal phase. One controlled crossover study found a 25 to 40% reduction in AHI during mid-luteal vs. Mid-follicular recordings, directly attributable to progesterone-driven upper airway tone. [10]

Premenstrual and Menstrual Phase

In the late luteal phase, progesterone and estrogen both fall sharply. Sleep efficiency declines 4 to 6 percentage points, slow-wave sleep drops, and REM increases. Women with premenstrual dysphoric disorder (PMDD) show a more pronounced REM rebound, with REM onset latency shortening to under 70 minutes in some PSG studies. [11]

Clinically, this means a sleep study scheduled in the late luteal or menstrual phase may over-estimate baseline OSA severity and under-estimate slow-wave sleep capacity. Scheduling notes should document cycle day for any pre-menopausal woman undergoing PSG.

Testosterone, Estrogen, Progesterone, and Sleep Architecture: Mechanistic Summary

Hormones act on sleep through multiple overlapping pathways. Each one carries different clinical implications.

Testosterone

Testosterone suppresses progesterone-driven ventilatory drive and relaxes pharyngeal dilator muscles during REM. Men on testosterone replacement therapy (TRT) show dose-dependent increases in AHI. A 2023 study in the Journal of Clinical Endocrinology and Metabolism (N=118, mean age 52) found that men initiating TRT at 200 mg testosterone cypionate every 2 weeks had a 28% mean increase in AHI at the 6-month mark. [12] TRT at physiologic replacement doses (targeting total testosterone 500 to 700 ng/dL) showed smaller but still statistically significant increases.

The AASM and the Endocrine Society both note that untreated moderate-to-severe OSA is a relative contraindication to TRT initiation in men, and that PSG should be considered in any TRT patient who develops new snoring, nocturia, morning headache, or excessive daytime sleepiness. [13]

Estrogen

Estrogen modulates serotonin and norepinephrine tone in brainstem nuclei that regulate sleep stage transitions. Declining estrogen at perimenopause associates with longer sleep onset latency, reduced N3, and increased arousal frequency. Women with confirmed estradiol below 20 pg/mL on PSG nights show arousal indices averaging 18.3 per hour vs. 11.7 per hour in women with estradiol above 80 pg/mL in one cross-sectional analysis from the Study of Women's Health Across the Nation (SWAN) Sleep Study. [14]

Postmenopausal HRT containing estradiol reduces WASO by an average of 14 minutes per night and increases total sleep time by 11 minutes per night compared to placebo in a 2022 Cochrane review of 11 randomized trials (N=1,147). [15]

Progesterone

Natural progesterone (not synthetic progestins) binds GABA-A receptors via its neurosteroid metabolite allopregnanolone. This mechanism is sedating and explains why the luteal phase typically improves sleep depth. Micronized progesterone 100 to 200 mg taken orally at bedtime demonstrates a statistically significant increase in N3 percentage compared to placebo in peri-menopausal women (P<0.01), while medroxyprogesterone acetate shows no such effect, suggesting receptor-pathway specificity. [16]

The HealthRX clinical framework below maps the hormone-sleep interaction matrix for patient triage decisions.

Hormone-Sleep Interaction Matrix for PSG Interpretation

| Hormone State | AHI Effect | N3 Effect | REM Effect | Clinical Action | |---|---|---|---|---| | High testosterone (TRT supraphysiologic) | Increases AHI | Decreases N3 | Minimal | Baseline PSG before TRT; repeat at 6 months | | Low estrogen (menopause, post-oophorectomy) | Modest AHI increase | Decreases N3 | Increases fragmented REM | Consider HRT; repeat PSG at 12 weeks on therapy | | High progesterone (luteal, oral P4) | Decreases AHI | Increases N3 | Decreases REM | Schedule PSG follicular phase for baseline AHI | | GLP-1 agonist (weight loss context) | Decreases AHI with weight loss | Increases N3 | Minimal direct effect | Recheck AHI at 10% body weight loss milestone |

Polysomnography Normal Ranges in the Context of Hormone Optimization

Interpreting a PSG result without knowing the patient's hormone status, cycle phase, and current medications yields an incomplete picture.

Optimal vs. Normal: A Functional Medicine Distinction

Conventional "normal" AHI below 5 events per hour was derived from population samples that were predominantly male. For women, some longevity-medicine clinicians argue the functional optimum is an AHI below 2 events per hour, given women's higher sensitivity to sleep fragmentation on cognitive outcomes. The AASM official position defines normal as below 5, mild as 5 to 14.9, and moderate as 15 to 29.9. [1]

For N3, the functional target for patients on TRT or HRT is 18 to 23% of total sleep time. Below 13% in a patient under 60 warrants investigation of testosterone excess, thyroid dysfunction, alcohol use, benzodiazepine use, or untreated OSA as contributing causes.

OSA Severity and Testosterone Candidacy

The Endocrine Society 2018 Clinical Practice Guideline on testosterone therapy states: "We suggest that testosterone therapy not be initiated in patients with severe untreated obstructive sleep apnea." [13] Specifically, an AHI at or above 30 events per hour should be treated with CPAP to documented efficacy before TRT begins. An AHI of 15 to 29.9 events per hour requires shared decision-making and close follow-up.

For women beginning TRT or high-dose DHEA (which converts peripherally to testosterone), baseline PSG or a validated home sleep test is appropriate if BMI exceeds 30 kg/m, neck circumference exceeds 40 cm, or STOP-BANG score is 3 or higher.

GLP-1 Agonists and OSA Reversal

Semaglutide's SURMOUNT-OSA trial (N=469, presented at ATS 2024) showed that semaglutide 2.4 mg weekly reduced AHI by 55% at 52 weeks in adults with moderate-to-severe OSA, with a mean AHI reduction from 51.5 to 23.8 events per hour in the non-CPAP arm. [17] This PSG-confirmed reduction is the largest pharmacologic effect on AHI reported to date and has immediate clinical relevance for HealthRX GLP-1 patients who also carry an OSA diagnosis.

The mechanism is primarily weight-loss mediated (mean 15.3% body weight reduction in that arm), though direct GLP-1 receptor effects on brainstem respiratory centers may contribute.

How Cycle Phase and Hormonal Status Should Inform PSG Scheduling

Timing the PSG correctly improves diagnostic accuracy. These practical scheduling principles apply across HealthRX patient types.

For Pre-Menopausal Women

Schedule PSG in the follicular phase (cycle days 5 to 12) to obtain a baseline AHI unaffected by progesterone suppression. If the clinical question is CPAP titration pressure optimization rather than initial diagnosis, the luteal phase reading is adequate. Document the estimated cycle day on the requisition.

For Peri-Menopausal Women

Cycle irregularity makes phase-based scheduling unreliable. Measure serum progesterone on the morning of the study. A level above 5 ng/mL suggests the patient is in a progesterone-dominant state; interpret AHI conservatively (true baseline may be higher). Estradiol below 30 pg/mL on the test night predicts higher arousal index regardless of AHI.

For Patients on Hormone Therapy

Men on TRT: schedule the PSG trough (72 to 96 hours after last injection or patch dose) for the most conservative AHI estimate. Peak testosterone states may transiently worsen AHI.

Women on cyclic HRT (estradiol plus oral micronized progesterone on days 12 to 25): schedule PSG during the progesterone-off week for unmodified baseline AHI.

Women on continuous combined HRT: PSG can be scheduled at any time; document medication dose and timing in the referral.

OSA Screening Tools and When PSG Is Required vs. Optional

Not every patient needs full in-lab PSG. The decision tree below reflects AASM and Endocrine Society guidance.

Home Sleep Apnea Testing (HSAT)

Type III HSAT (4-channel: airflow, effort, oximetry, heart rate) is appropriate for adults with high pre-test probability of moderate-to-severe OSA (STOP-BANG score 5 to 8) without comorbid central sleep apnea, heart failure, or neuromuscular disease. HSAT cannot stage sleep, so it provides AHI but not N3, REM, or arousal index. A negative or low HSAT in a symptomatic patient requires full PSG. [1]

Women with primary insomnia complaints and suspected OSA should go directly to PSG, not HSAT, because their hypopnea-predominant OSA can be missed without EEG staging and full respiratory channel analysis.

PSG Indications Specific to Hormone Therapy Patients

• Any TRT patient with new-onset snoring, witnessed apnea, or STOP-BANG score 3 or higher
• Any GLP-1 patient with known OSA who has lost 10% or more of body weight (to document AHI improvement and assess CPAP de-escalation)
• Any peri-menopausal patient on HRT with persistent insomnia unresponsive to estradiol optimization
• Any patient with unexplained low morning testosterone (in men) despite adequate TRT dosing (OSA-driven cortisol spikes suppress LH and free testosterone by 20 to 30%)

A PSG AHI of 30 events per hour or higher in a patient with low total testosterone should prompt the treating clinician to prioritize CPAP initiation before adjusting TRT dose upward, per Endocrine Society guidance. [13]