Understanding Your Nights · 11 min read

Perimenopause: Why Sleep Gets Worse Before Your Period

A hand holding a pen over an open paper planner, with a run of days marked in yellow highlighter.
Once you can see which week it is, the week stops feeling like a personal failure.

Who this article is for: women who are still having periods, even irregular ones. Everything below depends on a menstrual cycle still happening, so if your periods have stopped for good it will not apply to you.

If you have gone twelve months without a period, you are postmenopausal — your sleep problems are real but they are not cyclical, and our complete sleep guide or menopause insomnia and anxiety at night will be more use to you than this page.

The short answer

If you still get periods, the week before one is measurably your worst sleep week. In a sleep-lab study of perimenopausal women, deep sleep fell and awakenings rose in the days before bleeding started. In midlife women tracked at home, sleep efficiency dropped and total sleep fell by about 25 minutes in that final week.

For about a year I was convinced my sleep had become completely random. Four decent nights, then three where I was awake at two in the morning for no reason I could name, then back to normal. I looked at caffeine, at wine, at the temperature of the room, at stress. None of it lined up.

What finally lined up was my period. Not the period itself — the five or six days before it. Once I put the two things on the same page, the randomness disappeared. It had been a pattern the whole time; I had just been looking at the wrong calendar.

This turns out to be one of the better-documented things in menopause sleep research, and one of the least talked about. So here is what has actually been measured.

What is the luteal phase, and why does it matter?

The luteal phase is the second half of your cycle, from ovulation until your period starts. After ovulation, progesterone rises, peaks, and then falls sharply if you do not conceive. That rise and fall is what your sleep is reacting to, which is why the bad nights cluster in the same place each month.

The first half of the cycle — the follicular phase, from the first day of bleeding up to ovulation — is the calmer half for sleep. Estrogen climbs steadily, progesterone stays low, and most women's nights are unremarkable.

Then you ovulate, and the second half behaves completely differently. Progesterone climbs to levels many times higher than anything in the first half, holds for a few days, and then collapses. Your period follows that collapse. So does, for a lot of women, the worst sleep of the month.

Follicular phaseLuteal phase
WhenFirst day of bleeding to ovulationOvulation to the next period
ProgesteroneLow and flatRises, peaks, then falls sharply
Body temperatureLowerAbout half a degree higher
Typical sleepMore settledLess deep sleep, more awakenings

That temperature row deserves a note of its own. Progesterone raises your core body temperature slightly, and falling asleep depends on your core temperature dropping. So in the luteal phase you are trying to cool down from a higher starting point, in a body that is already prone to running hot at night. The two problems stack.

Does sleep really get worse before your period?

Yes, and it has been measured two different ways. In a sleep laboratory, perimenopausal women had less deep sleep and more awakenings in the luteal phase than in the follicular phase. Tracked at home with wrist monitors, midlife women lost about five percent of their sleep efficiency and 25 minutes of sleep in the final week.

The laboratory study is the more precise of the two. Researchers brought women in the early menopause transition into a sleep lab twice — once in each phase of the same cycle — and recorded their brain activity overnight. In the luteal phase, the women had a lower percentage of slow wave sleep, woke more often, and had more arousals per hour of sleep.

The detail that makes this study genuinely useful is who it applied to. Eleven of the twenty women had an insomnia disorder; nine did not. Both groups showed the same luteal-phase deterioration. So this is not something that only happens to poor sleepers. It is a physiological pattern that sits underneath whatever else is going on with your nights.

25 min
The drop in average total sleep time in the fourth week of the cycle compared with the third, in 163 midlife menstruating women tracked at home. Sleep efficiency also fell by about 5 percentage points in that final premenstrual week. Source: Zheng H, Harlow SD, Kravitz HM, et al. Actigraphy-defined measures of sleep and movement across the menstrual cycle in midlife menstruating women: SWAN Sleep Study. Menopause. 2015;22(1):66–74.

Twenty-five minutes may not sound like a catastrophe, and on its own it is not. But that is an average across a group, and averages hide the women at the rough end. It is also cumulative: five or six consecutive nights of that, every month, is a meaningful amount of sleep to keep losing on a schedule.

Why is this worse now than it was at thirty?

Because perimenopause amplifies a swing your body used to absorb. In your thirties, hormones followed a fairly predictable arc and sleep stayed reasonably stable across the month. Now the peaks are higher, the drops are steeper and less predictable, and hot flashes can arrive on top of them. Same mechanism, much rougher ride.

This is the part I found genuinely clarifying, because it explains something that had been bothering me: I have had periods for thirty-five years and this is new. If progesterone withdrawal disrupts sleep, why did it not do this at twenty-five?

The answer is that perimenopause is not a smooth decline in hormones. It is a period of increasing volatility before the decline. Estrogen can spike higher than it ever did in your reproductive years and then crash. Cycles shorten, then lengthen. Ovulation becomes unreliable. The swings your body handled easily when they were small and regular are now large and erratic.

The useful reframe: the premenstrual bad week is not a new problem that arrived with perimenopause. It is an old, mild pattern that perimenopause turned up the volume on. That is why it feels like it came out of nowhere — and also why identifying it explains so much at once.

What is progesterone actually doing to your sleep?

Progesterone has a calming, sedative-like effect through its breakdown products, which act on the same brain receptors as some sleeping medications. That sounds helpful, and while the level is high it often is. The trouble is the withdrawal: when progesterone drops in the days before bleeding, that chemical support disappears fairly abruptly.

Progesterone is converted in the body into neurosteroids, and one of them, allopregnanolone, acts on GABA receptors — the same braking system in the brain that benzodiazepines and several sleeping tablets work on. Your own body has been running a mild sedative for a week or so.

Then it stops. Not gradually, in the last days of the cycle, but on a fairly steep slope. And there is a finding that makes this concrete: in women tracked with overnight monitoring, a steeper late-luteal rise and fall in progesterone was linked to significantly more time awake after falling asleep. It is not the level that predicts the bad nights so much as the size of the swing.

A full moon visible through a dark window, with a warm candle glowing on the sill inside the room.
The withdrawal happens over a few days. That is why the bad nights arrive in a run rather than one at a time.

This also explains the emotional weather that often comes with the bad sleep week. The same neurosteroid withdrawal is one of the leading explanations for premenstrual mood symptoms. So if you find that the week you cannot sleep is also the week you feel tearful, irritable and thin-skinned, those are not two separate failings. They are the same hormonal event showing up in two places.

If you want to go further into what progesterone does and does not do for sleep — including what happens when it is prescribed rather than produced — that is covered in does progesterone help you sleep in menopause.

Why is one month terrible and the next one fine?

Most likely because in the good month you did not ovulate. Cycles without ovulation become much more common in perimenopause, and no ovulation means no progesterone rise in the second half. No rise means no drop, so the premenstrual sleep crash never happens. You still bleed, so the month looks normal from the outside.

This is the single most useful thing I learned writing this article, because it dissolves the objection that stops most women from believing the pattern exists. If it were really hormonal, you think, it would happen every month. It does not happen every month. Therefore it must be something else — stress, or the wine, or you.

But an anovulatory cycle is a cycle where the ovary does not release an egg. No egg means no corpus luteum, and the corpus luteum is the structure that produces progesterone. So the entire second-half hormone surge simply does not occur. You may still bleed roughly on schedule, which is why you would never know.

What this means in practice: a month with no bad sleep week is not evidence against the pattern. It is more likely to be evidence of a cycle where you did not ovulate — which is completely normal in perimenopause. Judge the pattern across three or four cycles, not one.

Anovulatory cycles happen occasionally at every age, but they cluster at the two ends of reproductive life: the years just after periods begin, and the years before they stop. Cycles that come unusually early or unusually late are more likely to be anovulatory. If your cycles have become unpredictable in length, some of them are almost certainly happening without ovulation.

How do you predict the bad week when cycles are irregular?

Count backwards, not forwards. Forward counting fails in perimenopause because cycle length keeps changing, so "day 21" lands somewhere different every month. Instead, log your sleep and your periods for three cycles and mark the days before each bleed. The pattern shows up relative to the period, not to the calendar.

Here is the method, and it takes about ten seconds a day:

  1. Every morning, write one number. How you slept, from 1 to 5. That is the whole daily commitment. Nothing else is required.
  2. Mark the first day of every period. Circle it, star it, whatever you like — but mark day one specifically, not "sometime that week".
  3. After three cycles, line them up backwards. Write the day before each period as −1, the day before that as −2, and so on back to −10. Then compare your sleep scores at each position.
  4. Look for where the bad scores cluster. Most women who have this pattern find it somewhere between −7 and −2. Yours will have its own shape.

Three cycles matters, and now you know why: one of them may well have been anovulatory. A single month tells you very little. Three months tells you whether there is a pattern and roughly where it sits.

If you would rather track properly than scribble on a calendar, what to monitor for two weeks covers what is worth recording and what is a waste of effort. The short version: fewer variables, recorded consistently, beats an elaborate log you abandon after nine days.

A caution about wearables: a recent study of both young and midlife people found that wearables reliably picked up menstrual cycle variation in skin temperature and heart rate, but not in their sleep metrics. Your ring or watch may well miss this pattern entirely. Your own morning score is, for this specific purpose, the better instrument.

What actually helps during that week?

Nothing has been trialed specifically for premenstrual sleep in perimenopause, so this is borrowed evidence applied sensibly. The most useful move is anticipation: protect that week rather than fighting through it. Cooler bedroom, alcohol out, caffeine earlier, and no important decisions or hard deadlines scheduled into the five days before you bleed.

I want to be straight about the evidence here. There are good studies describing what happens to sleep in the luteal phase. There are essentially none testing what fixes it in perimenopausal women specifically. So what follows is reasoning from mechanism, not from trial results, and I would rather tell you that than dress it up.

  • Drop the room temperature further that week. Your core temperature is already running higher in the luteal phase, and cooling down is how sleep begins. This is the one intervention that follows directly from the mechanism.
  • Cut alcohol during those days specifically. Alcohol fragments the second half of the night, which is precisely the half that is already being disrupted. Alcohol and menopause sleep explains why the "nightcap" backfires.
  • Move your caffeine cutoff earlier. Not necessarily less coffee — earlier coffee. Caffeine and menopause sleep covers the timing.
  • Do not lie there fighting it. If you are awake for more than about twenty minutes, get up. How to fall back asleep at 3am is the practical version of this.
  • Take the pressure off the week itself. If you know which days they are, stop scheduling the difficult conversation, the presentation or the long drive into them. This is not a sleep intervention. It is the one that changes how the week actually feels.

That last point is the one I would keep if I could only keep one. The sleep loss is real, but a good part of the damage comes from the confusion around it — lying awake wondering what you did wrong, and then facing a demanding day on four hours because you did not know it was coming. Knowing removes both.

When should you talk to a doctor about this?

Bring it up when the pattern is costing you function, not just comfort. Worth a conversation: sleeplessness so severe you cannot work that week, a low mood that lifts predictably once bleeding starts, bleeding that is very heavy or lasts longer than a week, or any bleeding at all after twelve period-free months.

The reason to mention the cyclical nature specifically is that it points somewhere. Insomnia that has no pattern gets treated as insomnia. Insomnia that arrives on a hormonal schedule opens up different options, and a clinician who knows it is cyclical will consider things they would not otherwise raise.

  • Severe premenstrual mood symptoms that reliably lift when you bleed can be premenstrual dysphoric disorder, which is a recognized diagnosis with real treatments — not something to endure quietly.
  • Cyclical hormonal treatment is a genuine option for some women. HRT for menopause insomnia covers the landscape, and the answer depends on your bleeding pattern and history.
  • Heavy or prolonged bleeding needs looking at on its own terms. Anovulatory cycles can produce heavy bleeding, and heavy bleeding can cause iron deficiency — which itself worsens sleep and is strongly linked to restless legs.
  • Any bleeding after twelve months without a period should be assessed promptly. That is not part of this pattern and always warrants a check.

Take your three cycles of notes with you. "I sleep badly sometimes" gets a general answer. "My sleep score drops from a 4 to a 2 in the five days before every period, here are three months of it" gets a specific one.

Where to go from here

If the sleepless nights come with a racing mind rather than a hot body, menopause insomnia and anxiety at night is the closer fit. If you are still working out whether what is happening to you is perimenopause at all, start with why can't I sleep at 45.

For the hormone question in depth, read does progesterone help you sleep in menopause. If the daytime exhaustion is the bigger problem, tired all day, awake at night covers that. And for the whole picture in one place, start at the complete perimenopause and menopause sleep guide.

Key takeaways

  • This applies only if you are still having periods, even irregular ones — the pattern depends on a cycle still happening.
  • In a sleep laboratory, perimenopausal women had less slow wave sleep and more awakenings in the luteal phase, and this held whether or not they had insomnia.
  • Tracked at home, 163 midlife women lost about 25 minutes of sleep and 5 percentage points of sleep efficiency in the premenstrual week.
  • The trigger is progesterone withdrawal: its breakdown products act on the same calming receptors as some sleep medications, and they fall away over a few days.
  • A steeper late-luteal rise and fall in progesterone was linked to more time awake — the size of the swing matters more than the level.
  • A good month is often an anovulatory month: no ovulation means no progesterone rise, no drop, and no crash. Judge the pattern over three cycles, not one.
  • Count backwards from bleeding, not forwards from a date — cycle length changes too much in perimenopause for fixed calendar rules to work.
  • Wearables have been shown to miss menstrual sleep variation; a one-to-five score written each morning is the better instrument here.

Frequently asked questions

Why does my sleep get worse the week before my period?

Progesterone rises after ovulation and then falls sharply in the days before you bleed. Its breakdown products act on the same calming brain receptors as some sleep medications, so when the level drops, that support disappears. Sleep labs show less deep sleep and more awakenings in exactly that window.

Is premenstrual insomnia worse in perimenopause?

It appears to be. In younger women, sleep stays fairly steady across the menstrual cycle. In women in the early menopause transition, sleep measured in a laboratory was clearly more disrupted in the luteal phase, whether or not they had an insomnia disorder to begin with.

Why is one month terrible and the next month fine?

Most likely because you did not ovulate that month. Cycles without ovulation become common in perimenopause, and without ovulation there is no progesterone rise and therefore no drop before bleeding. You still get a period, so the month looks normal, but the sleep crash never arrives.

How can I predict the bad week if my cycles are irregular?

Count backwards from bleeding rather than forwards from a date. Log your sleep and your periods for three cycles, then line up the days before each bleed. The pattern shows up relative to when your period starts, which is why fixed calendar rules stop working in perimenopause.

Does this mean I need progesterone?

Not automatically. The cyclical pattern is a reason to discuss hormonal options with a clinician, not a self-diagnosis. Micronized progesterone taken at night does have evidence for sleep in menopausal women, but whether it suits you depends on your bleeding pattern, symptoms and medical history.

Sources cited

  1. Baker FC, Sassoon SA, Kahan T, et al. Menstrual cycle-related variation in physiological sleep in women in the early menopausal transition. J Clin Endocrinol Metab. 2015;100(8):2918–2926. pubmed.ncbi.nlm.nih.gov
  2. Zheng H, Harlow SD, Kravitz HM, et al. Actigraphy-defined measures of sleep and movement across the menstrual cycle in midlife menstruating women: Study of Women's Health Across the Nation Sleep Study. Menopause. 2015;22(1):66–74. pubmed.ncbi.nlm.nih.gov
  3. Baker FC, Lee KA. Menstrual cycle effects on sleep. Sleep Med Clin. 2018;13(3):283–294. sleep.theclinics.com
  4. Harlow SD, Gass M, Hall JE, et al. Executive summary of the Stages of Reproductive Aging Workshop +10: addressing the unfinished agenda of staging reproductive aging. J Clin Endocrinol Metab. 2012;97(4):1159–1168. pubmed.ncbi.nlm.nih.gov
  5. Alzueta E, Gombert-Labedens M, Javitz H, et al. Menstrual cycle variations in wearable-detected finger temperature and heart rate, but not in sleep metrics, in young and midlife individuals. J Biol Rhythms. 2024. pmc.ncbi.nlm.nih.gov
  6. Anovulatory bleeding. StatPearls. National Library of Medicine. ncbi.nlm.nih.gov
Illustrated avatar of Ellen Hayes smiling with a mug of tea.

Ellen Hayes

Health & Wellness Writer and Researcher

Hi — I'm Ellen. I spent a year treating my own bad nights as random before I put them on the same page as my periods. Three cycles of a single number written each morning told me more than a year of theorizing had.

I'm not a doctor. I'm not an expert — and I'm never going to talk down to you like one. I'm a woman in the middle of perimenopause who got tired of vague answers, started reading the actual research, and writes it down here in plain English, the way I'd explain it to a friend over coffee. More about Ellen →

Medical disclaimer: This article is for educational purposes only and is not a substitute for professional medical advice, diagnosis, or treatment. Heavy, prolonged or irregular bleeding, and any bleeding after twelve months without a period, need to be assessed by a clinician rather than explained away as perimenopause. Do not start or stop any hormonal treatment without medical advice. Read our full medical disclaimer.