The hormonal, neurological, and behavioural science of sleep in midlife women.
Poor sleep is one of the most universal complaints of perimenopause — and one of the most inadequately treated. It is not just tiredness. Chronic sleep disruption accelerates cognitive decline, raises cardiovascular risk, and drives weight gain. Understanding why it happens is the first step to fixing it properly.
Select a Topic
A normal sleep night consists of 4–6 cycles, each approximately 90 minutes. Each cycle moves through three NREM stages (N1 light sleep, N2 consolidated sleep, N3 deep slow-wave sleep) followed by REM sleep. N3 slow-wave sleep is the most physically restorative stage — when growth hormone is secreted, cellular repair occurs, immune function is consolidated, and the brain's glymphatic system clears metabolic waste including amyloid. REM sleep is when emotional memory processing and creativity occur.
Slow-wave sleep (N3) declines progressively from the late 30s — accelerated in women by the hormonal changes of perimenopause. Estrogen and progesterone both support slow-wave architecture and REM sleep. Their loss leads to more time in lighter N1/N2 sleep and more frequent awakenings. This is why perimenopausal women often feel unrefreshed despite sleeping 7–8 hours — the architecture is fragmented even when duration appears adequate.
Sleep is governed by two interacting systems: Process S (homeostatic sleep pressure — adenosine accumulates during waking and drives sleep need) and Process C (circadian rhythm — the 24-hour clock driven by the suprachiasmatic nucleus, light exposure, and cortisol/melatonin timing). Insomnia typically involves a disruption in one or both processes. Understanding which process is primarily disrupted guides treatment — circadian misalignment responds to light and melatonin; homeostatic disruption often requires CBT-I and sleep restriction.
Progesterone metabolises to allopregnanolone, which acts as a positive allosteric modulator of GABA-A receptors — the same receptors targeted by benzodiazepines and sleeping pills. Natural progesterone has a mild anxiolytic and sleep-promoting effect. As progesterone drops in perimenopause, this built-in sleep support disappears. Women who were always good sleepers often begin experiencing insomnia for the first time in their 40s — this is not psychological. It is the loss of a neurosteroid.
Estrogen receptors are expressed in the hypothalamus, which controls both the thermoregulatory centre and the suprachiasmatic nucleus (circadian clock). Estrogen suppresses REM sleep latency (helps initiate REM), supports slow-wave sleep depth, and stabilises body temperature regulation. As estrogen fluctuates and falls, the thermoregulatory set-point becomes unstable — small rises in core temperature trigger vasodilatory responses (hot flushes and night sweats) that repeatedly fragment sleep architecture.
Estrogen regulates serotonin and noradrenaline neurotransmission. As estrogen fluctuates, these systems become less stable — producing anxiety, mood disruption, and hyperarousal that are incompatible with sleep initiation. Many perimenopausal women describe lying in bed with a "switched-on" brain that cannot settle — this is in part a direct consequence of neurotransmitter dysregulation driven by hormonal change, not anxiety as a primary diagnosis.
When a woman in her late 40s comes to me with new-onset insomnia and anxiety, my first question is not "what is she stressed about" — it is "where are her hormones?" Treating the hormonal driver of perimenopausal insomnia with MHT produces better and faster results than starting a woman on an SSRI for anxiety that is fundamentally endocrine in origin.
— Dr KD · The Longevity ShiftVasomotor symptoms (hot flushes and night sweats) occur when the hypothalamic thermoregulatory set-point becomes unstable due to estrogen loss. At night, a small rise in core body temperature — triggered by this instability — produces a vasodilatory flush, sweating, and subsequent compensatory cooling that together force a brief awakening. These awakenings often pull women out of deep N3 sleep into lighter stages, and repeated disruption across the night produces the non-restorative sleep that characterises perimenopausal insomnia.
MHT directly addresses the hormonal cause of night sweats and hot flushes — reducing frequency and severity by 75–90% in most women. For women whose sleep disruption is primarily driven by vasomotor symptoms, HRT produces better sleep outcomes than any other intervention. Transdermal estradiol with micronised progesterone is the preferred formulation — and micronised progesterone itself has an additional sleep-promoting GABA-mediated effect.
Cortisol follows a clear circadian pattern: lowest in the early sleep hours (midnight–2am), then rising sharply between 2am and the cortisol awakening response (CAR) at 30–45 minutes after waking. This early morning cortisol rise prepares the body for the day. In healthy sleepers, it does not intrude on sleep. In perimenopausal women with HPA axis dysregulation, cortisol begins rising earlier — often at 2–3am — producing the characteristic wide-awake, anxious, racing-thoughts awakening that is one of the most recognisable features of perimenopausal insomnia.
Estrogen modulates the sensitivity of the HPA axis — it normally buffers the cortisol response to stressors. As estrogen falls, the HPA axis becomes hypersensitive and less well-regulated. Combined with the drop in progesterone (which has inherent anxiolytic/cortisol-dampening effects), perimenopausal women have a neurochemical environment that is more prone to nocturnal HPA activation. Chronic sleep deprivation itself then worsens HPA dysregulation — creating a cycle that is difficult to break without addressing the hormonal foundation.
Sleep hygiene alone is insufficient for clinical insomnia — it is a foundation, not a treatment. Women who have had disrupted sleep for more than 3 months, regardless of the cause, often develop conditioned arousal (the bedroom becomes a cue for wakefulness) that requires CBT-I to reverse, not just better habits.
Cognitive Behavioural Therapy for Insomnia (CBT-I) is a structured programme that addresses both the cognitive (thoughts, beliefs, and anxiety about sleep) and behavioural (habits, scheduling, stimulus control) factors that perpetuate chronic insomnia. Multiple randomised controlled trials and systematic reviews show CBT-I is more effective than sleeping pills in the long term and produces durable remission — benefits are maintained after the programme ends, unlike medication effects that stop when the drug is stopped.
CBT-I is delivered by trained psychologists or via validated digital programmes. Digital CBT-I apps (Sleepio, Somryst) have strong evidence and are more accessible than in-person therapy. A full programme typically takes 6–8 weeks. For perimenopausal women, CBT-I is most effective when combined with HRT addressing the hormonal substrate — the two are complementary, not competing.
I recommend CBT-I to virtually every woman I see with chronic insomnia — because even when hormones are the primary driver, months of disrupted sleep create learned insomnia patterns that persist even after the hormonal cause is treated. Address both. The combination is far more effective than either alone.
— Dr KD · The Longevity ShiftMelatonin is a circadian signal — not a sedative. It tells the brain "it is dark, prepare for sleep" by shifting the circadian clock and lowering the body's arousal threshold. It does not induce sleep directly. It is most effective for circadian phase disruption (jet lag, shift work, delayed sleep phase) and less effective for sleep maintenance insomnia (waking in the night) where the circadian clock is not the primary problem.
The pharmacologically effective dose for circadian signalling is 0.5–1mg. Most over-the-counter melatonin in the US comes in 5–10mg doses — this is a pharmacological dose that produces supraphysiological blood levels, can cause daytime grogginess, and does not produce better circadian effects than lower doses. Start at 0.5mg. The evidence for sleep latency improvement is consistent at low doses. Higher doses (3–10mg) are sometimes used short-term for jet lag but are not appropriate for nightly use.
For sleep onset: take melatonin 60–90 minutes before desired sleep time, not immediately before bed. This mirrors the natural onset of endogenous melatonin secretion (dim light melatonin onset, DLMO) and produces the best phase-shifting effect. Combining melatonin with morning light exposure produces synergistic circadian strengthening that is more effective than either alone.
Endogenous melatonin production declines with age — by the early 50s, nocturnal melatonin levels are meaningfully lower than in younger women. Low-dose melatonin supplementation in perimenopausal and postmenopausal women restores circadian signalling and can improve sleep onset latency and overall sleep quality. It is not a substitute for HRT in women with significant vasomotor-driven sleep disruption, but it is a safe and useful adjunct.
Magnesium activates GABA receptors and inhibits NMDA (excitatory) receptors — producing neurological calming that supports sleep. Deficiency is common (up to 50% of adults) and directly linked to insomnia, anxiety, and restless legs syndrome. Magnesium glycinate (200–400mg before bed) is the preferred form — better absorbed and less likely to cause loose stools than magnesium oxide. Evidence in older adults shows improved sleep efficiency and reduced awakenings.
L-theanine is an amino acid from green tea that increases alpha brain wave activity (associated with relaxed alertness) and modulates GABA, serotonin, and dopamine. It reduces anxiety without sedation and improves sleep quality in several randomised trials. Dose: 100–200mg before bed. Well-tolerated, no dependency, no morning grogginess. Particularly useful for sleep disruption driven by anxiety and rumination rather than primary circadian disruption.
Glycine is an inhibitory neurotransmitter and thermoregulatory amino acid. It lowers core body temperature by dilating peripheral blood vessels — mimicking the temperature drop needed to initiate sleep. Randomised trials in humans show improved subjective sleep quality and daytime alertness with 3g glycine taken 30–60 minutes before bed. Particularly relevant for perimenopausal women with thermoregulatory instability.
Ashwagandha is an adaptogen with cortisol-lowering and GABA-enhancing properties. The KSM-66 extract has several randomised trials showing improved sleep latency, sleep quality, and reduced anxiety. Dose: 300–600mg KSM-66 before bed. Most useful for sleep disruption primarily driven by cortisol dysregulation and stress. Onset of effect is gradual — 4–8 weeks of consistent use for full benefit.
Supplements are supportive — not curative for clinical insomnia. Do not use them as a reason to avoid addressing the hormonal driver (if relevant) or engaging with CBT-I. They work best as part of a layered approach that treats the underlying cause alongside symptom management.
Benzodiazepines (temazepam, diazepam) and Z-drugs (zolpidem/Stilnox, zopiclone) enhance GABA activity, producing sedation. They are effective short-term. The problems: tolerance develops within 2–4 weeks; physical dependence occurs within 4–6 weeks of nightly use; withdrawal insomnia on stopping is worse than the original insomnia; they suppress slow-wave sleep (the most restorative stage); and in women over 60 they significantly increase fall and fracture risk. Appropriate use: maximum 2–4 weeks for acute insomnia during a clearly defined crisis. Not appropriate for perimenopausal insomnia as ongoing management.
Very low-dose doxepin (3–6mg) is an antihistamine-based sleep agent with a different mechanism to benzodiazepines. It improves sleep maintenance (staying asleep) without the dependency risk or slow-wave suppression of Z-drugs. Approved for sleep maintenance insomnia. A reasonable option for some women with perimenopausal sleep disruption where HRT is not yet effective or contraindicated.
Mirtazapine at 7.5–15mg has potent histamine antagonism producing sedation. It also improves appetite (relevant if weight loss is a concern) and has antidepressant properties at higher doses. Used off-label for sleep in women with co-existing depression or anxiety. Side effect profile includes weight gain and next-day sedation, limiting use in some women.
If you have been on Z-drugs or benzodiazepines for more than 4 weeks, do not stop abruptly. Gradual tapering under medical supervision, combined with CBT-I to manage the withdrawal insomnia, is the safest discontinuation approach. Abrupt cessation of benzodiazepines can cause seizures in long-term users.
Obstructive sleep apnoea (OSA) in women presents differently from the classic male pattern. Women are less likely to report loud snoring (the primary diagnostic trigger in men) and more likely to present with insomnia, fatigue, headache, mood changes, and unrefreshing sleep. Their bed partners may not notice apnoeas. Menopause increases OSA risk 2–3 fold — upper airway muscle tone is partly estrogen-dependent, and adipose tissue redistribution to the neck with weight gain post-menopause further increases airway collapsibility. Despite higher prevalence in perimenopausal women than previously recognised, most are never assessed for it.
A home sleep apnoea test (HSAT) is the standard first investigation — comfortable, done at home, provides AHI (apnoea-hypopnoea index). Moderate to severe OSA (AHI >15) warrants CPAP therapy. CPAP is highly effective — most women with daytime fatigue secondary to sleep apnoea experience significant improvement within 1–2 weeks of good CPAP use. Mandibular advancement devices are an alternative for mild-moderate OSA in women who cannot tolerate CPAP.
A woman who comes to me at 53 with fatigue, insomnia, mood changes, and unrefreshing sleep should have sleep apnoea assessed — not just perimenopausal hormonal change assumed. The two commonly coexist. Starting HRT and missing untreated OSA results in partial treatment at best. Both need addressing.
— Dr KD · The Longevity ShiftAlcohol enhances GABA (inhibitory) neurotransmission and suppresses glutamate (excitatory), producing sedation and faster sleep onset. This is real — alcohol does reduce sleep latency. This is why many people use it as a sleep aid. The problem is what happens 3–4 hours later, when alcohol is metabolised and its sedative effect reverses: glutamate rebounds, cortisol rises, and the brain enters a hyperactivated state that fragments the second half of the night.
Even one standard drink within 3 hours of bedtime measurably reduces sleep quality in sleep EEG studies. For perimenopausal women with vasomotor symptoms, the combination of alcohol and hormonal thermoregulatory instability is particularly disruptive. This is not a moral judgement — it is physiology. If sleep is a priority, evening alcohol is one of the highest-leverage things to reduce.
Chronic short sleep duration (less than 6 hours/night) is associated with a 20% increased risk of heart attack and a 15% increased risk of stroke, independent of other cardiovascular risk factors. This is dose-dependent: the shorter the sleep, the higher the risk. The mechanism involves increased sympathetic nervous system activity, elevated blood pressure, worsened insulin resistance, and increased inflammatory markers — all of which are amplified by the hormonal changes of perimenopause. Sleep deprivation and estrogen deficiency compound each other's cardiovascular effects.
During deep NREM sleep (N3), the brain's glymphatic system is maximally active — cerebrospinal fluid flows through interstitial spaces, clearing amyloid-beta and tau protein (the hallmarks of Alzheimer's pathology). Chronic sleep disruption impairs glymphatic clearance, allowing amyloid to accumulate. This is one of the most compelling mechanistic links between sleep and dementia — and a specific reason why treating perimenopausal sleep disruption matters beyond quality of life: it may protect cognitive longevity.
Natural killer (NK) cell activity — critical for immune surveillance against cancer cells — is significantly suppressed by sleep deprivation. A single night of 4 hours sleep reduces NK activity by approximately 70%. Chronic short sleep is associated with increased cancer risk in epidemiological data. Sleep is when the immune system consolidates immunological memory after infection and vaccination — consistently short sleepers have a weaker response to vaccination.
Even partial sleep restriction (6 hours vs 8 hours for 2 weeks) produces insulin resistance equivalent to the effect of 4.5kg of weight gain. Sleep deprivation elevates ghrelin (hunger hormone) and suppresses leptin (satiety hormone), producing increased appetite — particularly for calorie-dense carbohydrate-rich foods. For perimenopausal women already navigating insulin sensitivity changes, metabolic disruption from poor sleep compounds the hormonal metabolic challenge significantly.
Sleep is not a luxury. It is the biological maintenance period without which every other intervention — HRT, GLP-1, nutrition, exercise — produces a fraction of its intended effect. A woman who is sleeping well processes her hormones better, trains more effectively, eats more intelligently, and ages more slowly. Sleep first. Everything else builds on that foundation.
— Dr KD · The Longevity Shift