Progesterone and Allopregnanolone in Perimenopause: The Mechanism Behind Midlife Anxiety

MARKABLE Research Team · August 2026 · 9 min read

Most clinical conversations about the menopause transition begin with estrogen. For a large proportion of women in their forties, that is not where the story starts. Progesterone declines earlier, it declines for a structural reason, and its neuroactive metabolite has a direct route to the symptoms most often labelled as anxiety.

Summary. Progesterone is produced by the corpus luteum only after ovulation. Ovulatory disturbances become common early in the transition, so progesterone falls before estradiol does. Its metabolite allopregnanolone is a positive allosteric modulator of GABA-A receptors, so its withdrawal reduces inhibitory tone. The resulting presentation of hyperarousal, sleep fragmentation, and premenstrual mood volatility is readily attributed to a primary anxiety disorder. This article is educational content for clinicians and informed readers, not clinical guidance.

The paradigm question

In a review that explicitly calls for a paradigm shift, Prior and Hitchcock describe perimenopause as characterised by three concurrent hormonal changes rather than one: erratically higher estradiol levels, decreased progesterone in normally ovulatory short luteal phase or anovulatory cycles, and disturbed ovarian to pituitary to hypothalamic feedback (Prior & Hitchcock, 2011).

The same review describes luteal out of phase events, in which a major estradiol surge occurs de novo during the luteal phase, and reports that approximately a third of perimenopausal cycles show this pattern.

Two implications follow directly. First, a normal or elevated estradiol does not argue against the transition. Second, the analyte most likely to be informative is the one least likely to be measured correctly.

Why progesterone falls first

The mechanism is structural rather than gradual. Progesterone is produced by the corpus luteum, which exists only after ovulation has occurred. Anovulatory cycles and short luteal phases therefore produce low progesterone directly, and both become more frequent through the transition. Prior notes that ovulatory disturbances occur within cycles that appear clinically normal in length, which is why cycle regularity is a poor proxy for luteal adequacy (Prior, 2018).

Allopregnanolone and GABAergic tone

Progesterone is metabolised to allopregnanolone, a neurosteroid characterised as a potent and efficacious positive allosteric modulator of the GABA-A receptor. Belelli and colleagues review this literature and note that these steroids are synthesised locally in the central nervous system, that they exhibit anxiolytic, anticonvulsant, analgesic and sedative properties in preclinical work, and that receptors incorporating the delta subunit appear to be an important contributor to their behavioural effects (Belelli et al., 2021).

The clinical translation of this mechanism is already established in an adjacent condition. The same review notes the approval of allopregnanolone, as brexanolone, for postpartum depression. A subsequent phase 3 randomised controlled trial of the oral neuroactive steroid zuranolone in severe postpartum depression reported a least squares mean difference in Hamilton Depression Rating Scale change of -4.0 versus placebo at day 15, with significant improvement also at days 3, 28 and 45 (Deligiannidis et al., 2023).

Why this matters for perimenopause. Postpartum depression is a model of abrupt progesterone and allopregnanolone withdrawal. Perimenopause is a model of repeated and unpredictable withdrawal. The mechanism is shared even though the trial evidence is specific to the postpartum state, and extension to perimenopause is a reasonable hypothesis rather than an established indication.

The premenstrual signature

Timing is often the most useful discriminator available without testing. Using a daily symptom diary in mid life women, Hale and colleagues found breast tenderness maximal in the premenstrual window and night sweats maximal premenstrually except in anovulatory cycles (Hale et al., 2003). The late luteal phase is when progesterone and allopregnanolone fall most steeply, and symptoms that cluster there are more consistent with a cyclical neurosteroid pattern than with a continuous anxiety disorder.

The misattribution pathway

The clinical sequence is predictable. A woman in her forties presents with new onset anxiety, initial insomnia or early waking, irritability, and palpitations. There is no obvious psychosocial precipitant. Cycles are still present, so the transition is not considered. An SSRI is initiated. Where the underlying driver is cyclical neurosteroid withdrawal, the response may be partial and the cyclical pattern typically persists.

Relevant national guidance, including NICE guideline NG23, does not position antidepressants as first line for low mood arising as part of the menopause transition where hormonal management has not been considered. That does not make an antidepressant wrong, but it does make the hormonal question one that should be asked before the file is closed.

Practical points for assessment

Ask about phase, not just presence. Whether symptoms cluster in the luteal phase is more informative than their absolute severity.
Time progesterone properly or not at all. Mid luteal sampling, roughly seven days post ovulation, is the only interpretable window. Where cycles are irregular this is difficult, and a single low value should not be over read.
Do not treat a normal estradiol as exclusionary. Given luteal out of phase events, it is an expected finding.
Exclude the mimics. Thyroid dysfunction, iron deficiency, and low B12 reproduce this presentation closely. See our review of micronutrient depletion in perimenopause.
Consider the obstetric history. Recent pregnancy and lactation extend the low progesterone window considerably. See the postpartum and perimenopause overlap.

Track the pattern, not the single visit

MARKABLE produces a personal baseline across facial, cognitive, sensory and symptom signals that can be repeated over time, so cyclical and progressive change becomes visible between appointments.

Schedule a Demo

Frequently Asked Questions

Which hormone declines first in perimenopause?

Progesterone, in most women. It is produced by the corpus luteum only after ovulation, and ovulatory disturbances become common well before estradiol declines. Prior and Hitchcock describe perimenopause as erratically higher estradiol combined with decreased progesterone in anovulatory or short luteal phase cycles, rather than as a state of falling estrogen.

What is allopregnanolone and why does it matter clinically?

Allopregnanolone is a neuroactive metabolite of progesterone and a positive allosteric modulator of GABA-A receptors, the same receptor family targeted by benzodiazepines. When progesterone falls, allopregnanolone falls with it and GABAergic inhibitory tone is reduced. This is a plausible mechanistic route to the anxiety, hyperarousal, and sleep fragmentation reported in perimenopause.

Is there direct evidence that this mechanism drives mood symptoms?

The strongest supporting evidence comes from postpartum depression, another state of abrupt progesterone withdrawal. A phase 3 randomised controlled trial published in the American Journal of Psychiatry found that zuranolone, a neuroactive steroid acting as a positive allosteric modulator of GABA-A receptors, produced statistically significant improvement in depressive symptoms compared with placebo. The perimenopausal extension of this mechanism remains an area of ongoing research.

Why can a normal or high estradiol result be misleading?

Because in perimenopause estradiol is not simply declining. Prior and Hitchcock report that roughly a third of perimenopausal cycles include a luteal out of phase event, a de novo estradiol surge during the luteal phase. A normal or elevated estradiol therefore does not exclude the transition, and the more informative absence is often progesterone.

How should progesterone be measured?

Serum progesterone is only interpretable in the mid luteal phase, approximately seven days after ovulation. Sampling at any other point will return a low value that carries no information. Where cycles are irregular, timing is difficult and a single measurement should not be over interpreted.

Does MARKABLE measure hormones?

No. MARKABLE is a general wellness product, not a medical device. It does not measure hormone levels and it does not diagnose, treat, or cure any condition. It scans patterns across facial features, cognition, vision, hearing, and symptom reporting to produce a personal baseline that can be tracked over time and shared with a clinician.

References

Prior JC, Hitchcock CL. The endocrinology of perimenopause: need for a paradigm shift. Frontiers in Bioscience (Scholar Edition). 2011;3(2):474-486. doi:10.2741/s166
Belelli D, Phillips GD, Atack JR, Lambert JJ. Relating neurosteroid modulation of inhibitory neurotransmission to behaviour. Journal of Neuroendocrinology. 2021;34(2):e13045. doi:10.1111/jne.13045
Deligiannidis KM, Meltzer-Brody S, Maximos B, et al. Zuranolone for the treatment of postpartum depression. American Journal of Psychiatry. 2023;180(9):668-675. doi:10.1176/appi.ajp.20220785
Prior JC. Progesterone for the prevention and treatment of osteoporosis in women. Climacteric. 2018;21(4):366-374. doi:10.1080/13697137.2018.1467400
Hale GE, Hitchcock CL, Williams LA, Vigna YM, Prior JC. Cyclicity of breast tenderness and night-time vasomotor symptoms in mid-life women: information collected using the Daily Perimenopause Diary. Climacteric. 2003;6(2):128-139.

Sources were located and verified against PubMed. DOI links are provided so the primary literature can be read directly.

This article is educational content and does not constitute medical advice, diagnosis, or a treatment recommendation. Progesterone is a prescription medicine and any decision regarding it rests with a qualified clinician. MARKABLE is a general wellness product, not a medical device, and is not intended to diagnose, treat, cure, or prevent any disease.