The 46-Person Trial Behind the Magnesium-for-Sleep Claim
Almost every article recommending magnesium for sleep eventually cites one trial: 46 elderly people in Tehran, 500 mg a day, eight weeks. It is a reasonable small trial that has been asked to carry far more than it can. Here is exactly what it did.
Evidence strength
Level 1b
Individual RCT
Peer-reviewed refs
6
Reading time
8 min
Key Takeaways
- Forty-six adults around 65 with primary insomnia were randomized to 500 mg of elemental magnesium a day, given as two magnesium oxide tablets, or placebo for eight weeks. Forty-three finished.
- The magnesium group improved on the Insomnia Severity Index (P=0.006), sleep-onset latency (P=0.02) and sleep efficiency (P=0.03). Serum cortisol fell (P=0.008) and melatonin rose (P=0.007).
- Total sleep time did not differ between groups (P=0.37), and serum magnesium rose only marginally (P=0.06). The abstract reports P values; effect sizes with confidence intervals are not given there.
- The trial is small, single-centre, eight weeks long, and relies on questionnaires and diaries rather than measured sleep. The participants were elderly Iranians whose baseline intake was probably low.
- The form was magnesium oxide, the one most guides tell you to avoid. Nothing in this trial supports a special form for sleep.
- Pooled with two other small trials, the effect is about 17 minutes less time to fall asleep. That is a real, modest effect in older adults with insomnia, not evidence for magnesium as a sleep aid in replete younger adults.
If you have read that magnesium "improves sleep efficiency, sleep onset latency and early morning awakening" and "lowers cortisol while raising melatonin", you have read a summary of one paper. It was published in 2012 in the Journal of Research in Medical Sciences by a group at the Faculty of Nutrition and Food Technology in Tehran, and it is the largest placebo-controlled trial of magnesium alone for insomnia that exists. [1] That is why it is everywhere.
It is also 46 people, eight weeks and a questionnaire. Both facts matter. This breakdown goes through who was studied, what changed, what the paper cannot tell you, and what it should change about how you use magnesium. The wider trial landscape is in Magnesium for Sleep and Anxiety: What the Trials Show.
Who Was Studied and What They Took
The trial recruited elderly volunteers with primary insomnia through the health divisions of cultural centres in three districts of Tehran between January and March 2012. Participants passed a psychiatric examination to exclude insomnia with another cause. The sample was 46 people, 23 women and 23 men, with a mean age of 65 and a standard deviation of 4.6 years. [1]
The sample size was planned. The authors calculated that 21 people per group would give 80% power to detect the expected difference at a significance level of 5%, then added 10% for dropouts to reach 23 per group. Two people were excluded for not taking the tablets regularly and one for missing the second blood draw, so 43 completed, 21 men and 22 women. [1]
The intervention was 500 mg of elemental magnesium a day, given as two tablets of 414 mg magnesium oxide, each containing 250 mg of magnesium, for eight weeks. The placebo was matched. Everything was double-blind. [1]
Measurements were taken at baseline and at eight weeks:
- Insomnia Severity Index, the primary subjective outcome
- A sleep log, giving sleep-onset latency, sleep time, sleep efficiency and early-morning awakening
- Three days of 24-hour dietary recall, to control for intake of magnesium, calcium, potassium, caffeine and calories
- Blood samples for serum magnesium, renin, melatonin and cortisol
- Physical activity, weight and body mass index as confounders
The two groups did not differ on any of these at baseline.
What the Trial Found
Compared with placebo, the magnesium group showed:
| Measure | Direction | P value |
|---|---|---|
| Insomnia Severity Index | Lower | 0.006 |
| Sleep-onset latency | Shorter | 0.02 |
| Sleep efficiency | Higher | 0.03 |
| Sleep time (diary) | Longer | 0.002 |
| Early-morning awakening | Less | 0.08 (not significant) |
| Total sleep time | No difference | 0.37 |
The abstract reports P values and directions. The group means and percentage changes are in the paper's tables, and the abstract does not give confidence intervals or effect sizes. [1] One way to put a number on the effect comes from the 2021 meta-analysis that pooled this trial with two others, 151 older adults in total: sleep-onset latency was 17.36 minutes shorter with magnesium (95% CI −27.27 to −7.44), and total sleep time was 16.06 minutes longer, which was not significant. [2] That matches the pattern here: people fell asleep faster and rated their insomnia as less severe, but did not clearly sleep longer.
The sleep-time result deserves a second look, because the paper reports diary sleep time as significantly longer and total sleep time as unchanged. The two come from different calculations in the sleep log. A trial of this size cannot resolve that kind of inconsistency, which is one reason the pooled analysis is a better guide to the size of the effect than any single line in the abstract.
The Hormone Numbers Everyone Quotes
Three blood results get repeated most: serum cortisol fell (P=0.008), serum melatonin rose (P=0.007), and serum renin rose (P below 0.001). Serum magnesium itself rose only marginally between groups (P=0.06), despite 500 mg a day for eight weeks. [1]
The hormone findings are consistent with an older German study, in which twelve adults aged 60–80 took a rising dose of magnesium for 20 days in a crossover design and showed more slow-wave sleep, higher renin and lower nocturnal cortisol on EEG and overnight blood sampling. [3] Two small trials pointing the same way is more than most supplements have. It is still two small trials.
What the hormone data cannot do is prove a mechanism. Cortisol and melatonin were measured once, at a single time point, in people whose sleep had improved. Better sleep lowers cortisol and raises melatonin on its own; the measurements are compatible with magnesium acting through the HPA axis and equally compatible with magnesium improving sleep by some other route and the hormones following.
What the Study Cannot Tell You
Whether it applies to you. The participants were Iranians in their sixties with diagnosed primary insomnia. Magnesium intake in older adults is often below requirement, and in the US about 48% of the population consumes less than the estimated average requirement. [6] A 100-person US trial in adults over 51 found that magnesium raised serum levels only in those who started low, and that sleep improved in the placebo group as much as in the magnesium group. [4] The Tehran result may be what repletion looks like in a population that was short, not what magnesium does to anyone's sleep.
How big the effect is. P values tell you the difference was unlikely to be chance. They do not tell you whether insomnia scores fell by two points or ten. The pooled 17-minute figure is the best available estimate and carries a wide interval.
Whether sleep actually changed. No polysomnography, no actigraphy. Every sleep outcome came from a questionnaire or a diary, which are the outcomes most sensitive to expectation. Blinding limits that, but blinding is imperfect when the active tablet can cause loose stools.
Anything about form. The trial used magnesium oxide, which in healthy volunteers has about 4% fractional absorption and performed no better than placebo on blood and urine markers over 60 days. [5] That it worked at all in this trial is another hint that the participants were depleted, and it is a direct contradiction of the idea that a special form is needed for sleep.
What happens after eight weeks. No follow-up. Whether the effect persists, fades or needs a maintenance dose is unknown.
Whether the findings replicate. The trial was single-centre, and in thirteen years no larger trial of magnesium alone for insomnia has been published that we could find.
What This Changes, and What It Does Not
It changes the claim "there is no evidence magnesium helps sleep" to "there is one reasonable small trial, supported by a twelve-person EEG study and a low-quality meta-analysis, showing a modest benefit in older adults with insomnia." That is enough to justify an eight-week trial in someone with insomnia and a diet low in magnesium, which is how The Deep Sleep Stack uses it: as a second layer, added on its own, so the effect can be judged against a diary.
It does not change the picture for younger adults with adequate intake, for whom no placebo-controlled trial shows a benefit. It does not support any form over any other, and it specifically undercuts the claim that oxide is useless for sleep. It does not justify "magnesium lowers cortisol" as a general statement; it justifies "cortisol fell in a trial where sleep improved".
What would change our reading: a placebo-controlled trial of 200 or more adults under 60, with actigraphy or polysomnography, a single prespecified primary endpoint and independent funding. If that trial found a 17-minute reduction in sleep-onset latency, the grade would move from C to B. If it found nothing, the 2012 result would be filed as a repletion effect in a depleted population, which is where the current evidence already leans. Which form to use if you run the trial on yourself is covered in Magnesium Glycinate vs Threonate vs Citrate.
Scientific References
- [1]Abbasi B, Kimiagar M, Sadeghniiat K, Shirazi MM, Hedayati M, Rashidkhani B. The effect of magnesium supplementation on primary insomnia in elderly: A double-blind placebo-controlled clinical trial — Journal of Research in Medical Sciences (2012)Oxford 1bPMID 23853635
- [2]Mah J, Pitre T. Oral magnesium supplementation for insomnia in older adults: a Systematic Review & Meta-Analysis — BMC Complementary Medicine and Therapies (2021)Oxford 1aPMID 33865376
- [3]Held K, Antonijevic IA, Künzel H, et al.. Oral Mg(2+) supplementation reverses age-related neuroendocrine and sleep EEG changes in humans — Pharmacopsychiatry (2002)Oxford 2bPMID 12163983
- [4]Nielsen FH, Johnson LK, Zeng H. Magnesium supplementation improves indicators of low magnesium status and inflammatory stress in adults older than 51 years with poor quality sleep — Magnesium Research (2010)Oxford 1bPMID 21199787
- [5]Walker AF, Marakis G, Christie S, Byng M. Mg citrate found more bioavailable than other Mg preparations in a randomised, double-blind study — Magnesium Research (2003)Oxford 1bPMID 14596323
- [6]Rosanoff A, Weaver CM, Rude RK. Suboptimal magnesium status in the United States: are the health consequences underestimated? — Nutrition Reviews (2012)Oxford 5PMID 22364157
Follow us on Google Search
Add BiohackingHub as a preferred source to see more of our research in Google results.
Related protocols
Partner with BiohackingHub
Editorial placements, sponsorships, and product features for evidence-minded brands.