Magnesium for sleep in perimenopause: the evidence

You have read it in a group chat, heard it on a podcast, maybe had a friend press a bottle into your hand: take magnesium, it fixed my sleep. When nights break apart at 3 a.m. for months at a time, a cheap supplement with a calm reputation is an easy thing to reach for.

It is a reasonable question to ask. It deserves a more honest answer than the label gives it.

Why magnesium became the midlife sleep supplement

Sleep trouble is not a small side note in perimenopause. SWAN cohort research found sleep problems reported by roughly 16–42% of women before the transition and 39–47% during it. That is a lot of people looking for something to try, and magnesium sits in a comfortable spot: sold without a prescription, familiar from food, described as “relaxing.”

There is a plausible story behind it, too. Magnesium is involved in nerve signaling and muscle relaxation, and low intake is common. But a plausible mechanism is where research starts, not where it ends. The question is what happened when people actually took it and were compared against people who didn’t.

What the trials actually found

The best single summary is a 2021 systematic review and meta-analysis by Mah and Pitre. It looked for randomized trials of oral magnesium for insomnia in older adults and found three, with 151 participants in total, aged 55 and up.

Pooled together, people taking magnesium fell asleep about 17 minutes faster than those on placebo. Total sleep time went up by about 16 minutes, but that difference was not statistically reliable. And the authors were blunt about the foundation: every trial carried moderate-to-high risk of bias, and the evidence was rated low to very low quality. Their conclusion was that the literature is not good enough for physicians to make well-informed recommendations.

One of the three trials, a 2012 Iranian study of 46 older adults, is the one most often quoted online for its improvements in sleep efficiency and time to fall asleep. It is also the one reviewers flagged for unclear allocation concealment — a quiet way of saying we cannot be sure the groups were fairly assigned.

Notice what is missing. None of these studies focused on women in perimenopause, and none looked at the kind of sleep disruption that arrives with night sweats and hormonal swings. The honest summary is not “magnesium doesn’t work.” It is “we don’t know much, and what we know is small.”

The placebo problem, in plain numbers

A newer trial makes the next point better than any review could. In 2025, researchers randomized 155 adults with self-reported poor sleep to magnesium bisglycinate or an identical placebo capsule for four weeks.

The magnesium group improved by 3.9 points on the Insomnia Severity Index. The placebo group improved by 2.3 points. The difference was statistically significant, but only just, and the effect size was small. Put differently: much of the improvement people felt happened whether or not the capsule contained magnesium.

That is not a knock on anyone who felt better. It is a description of how sleep works. Deciding to do something about your nights, paying attention to bedtime and expecting relief all change how nights go and how you rate them. The same trial hinted that people with lower dietary magnesium intake might benefit more — an exploratory finding, worth a follow-up study rather than a conclusion.

What magnesium probably won’t touch: the heat

If what wakes you is heat — a flush, a damp pillow, kicking off the duvet — the evidence points somewhere else. A double-blind trial of 289 women with a history of breast cancer and bothersome hot flashes compared magnesium oxide with placebo. How often and how severely hot flashes came was similar in every group. The authors wrote that their results do not support magnesium for hot flashes.

That matters for sorting your own nights. Perimenopausal sleep loss is often more than one problem layered together: trouble falling asleep, early waking, and night sweats that show up as wrist temperature changes. A supplement studied mainly for sleep onset is, at best, aimed at the first layer.

Magnesium is not risk-free either. It commonly upsets the stomach, it is not suitable for everyone with kidney problems, and it can interfere with how some antibiotics and bone medications are absorbed. A pharmacist or clinician can check it against what you already take — a two-minute question worth asking before, not after.

Your nights are the study that matters

Group averages from small trials can’t tell you whether magnesium does anything for you. But because the effect, if there is one, is small, your memory is the wrong instrument for judging it. One rough week before and one decent week after will feel like proof. It isn’t.

This is where comparing against your own baseline earns its keep. What an Apple Watch records overnight — sleep duration, awakenings, overnight heart rate and heart rate variability, wrist temperature — won’t give you a laboratory answer. Sleep stages from a wrist are estimates with known blind spots. What the watch does offer is consistency: the same imperfect ruler, every night, before and after.

A fair personal comparison looks like a tiny version of a trial:

• A before window. Two to three weeks of your usual nights, recorded, before you change anything.
• One change at a time. If you also start a new bedtime routine, cut caffeine and add magnesium in the same week, you won’t know which did what.
• An after window of the same length, with the same eye on what else changed — travel, illness, alcohol, a stressful month.
• Modest expectations. In the trials, the difference was measured in minutes, not hours. A shift that small can easily hide inside ordinary night-to-night variation.

If your nights look clearly better across weeks, that is useful information. If they look the same, that is useful too — and possibly a reason to spend your energy on what is actually breaking your sleep.

How Perigee reads these signals

Perigee compares each night’s sleep, heart rate and wrist temperature against your own recent weeks, not against a population average, and lets you log what you changed and when. It won’t tell you whether a supplement works; it will show you whether your nights moved after you started one, and by how much relative to your usual range.

Every reading behind that is an estimate, not a lab measurement. That is the premise of what we built: a clearer record of your own nights to bring to a conversation, not a verdict. The sleep signal page sets out what overnight data can and cannot support.

One small thing to try this week

Before you open the bottle, write down today’s date and one sentence describing your nights right now — how long it takes to fall asleep, how often you wake, whether heat is part of it. Then give it two weeks of ordinary recorded nights first.

If you do start, you will have something better than a feeling to compare against. And if you mention it to your clinician, you can say what changed and when — the kind of detail that turns “I think it helps” into a real conversation.