Magnesium for sleep: what the 2026 review of 12 trials found

Does magnesium help you sleep?

Magnesium may help a small share of adults sleep a bit better, but trials disagree and the evidence is weak. A systematic review published on 2026-08-28 reviewed 12 randomized trials of oral magnesium in its main narrative synthesis (Journal of Dietary Supplements, 2026). The authors searched five databases and registries up to 1 June 2026.

The review found inconsistent results across questionnaires, EEG, research-grade actigraphy, consumer wearables and blood markers. It rated the certainty of the evidence low to very low. Its conclusion was that the evidence “does not support oral magnesium as a routine treatment for insomnia” but that magnesium “may provide modest benefits for selected subjective outcomes in some adults.”

The review does not say magnesium does nothing. It says the trials are too small, too different and too often biased to say.

What did the 12 trials in the 2026 review find?

The trials covered older adults, adults with poor sleep, and people with other conditions, using at least six magnesium forms at very different doses. The review itself says populations, formulations, elemental doses, durations and outcome measures varied substantially, so it summarised results in words rather than pooling them.

The table below shows ten randomized trials from the review’s reference list, each opened on PubMed or the journal site, with numbers taken from its abstract or full text on 2026-10-26. Two of the 12 trials are not in the table.

Trial	Who (n)	Magnesium and dose	Length	What it found
Held 2002	Adults aged 60-80 (12), crossover	Effervescent Mg, built up to 30 mmol a day	20 days per phase	More slow-wave sleep (16.5 vs 10.1 min, p ≤ 0.05) on EEG
Nielsen 2010	Adults over 51 with poor sleep (100; 96 finished)	320 mg a day as citrate vs sodium citrate	7 weeks	Sleep scores improved (PSQI 10.4 to 6.6), but the authors suspect a placebo effect
Abbasi 2012	Older adults with insomnia (46)	500 mg a day	8 weeks	Longer sleep time (p = 0.002) and better efficiency (p = 0.03) than placebo
Liu 2016	Ages 50-70 with cognitive impairment (51 randomized; 44 completed)	MMFS-01, a magnesium L-threonate compound, 1.5-2 g a day	12 weeks	Cognition improved; sleep effect “could not be determined” because of placebo response
Macian 2022	Fibromyalgia	100 mg a day as chloride, slow release	1 month	Sleep not different from placebo
Gholizadeh-Moghaddam 2023	Women with PCOS (64)	250 mg magnesium oxide tablet a day	10 weeks	No effect on sleep quality (p = 0.85)
Khalid 2024	Adults with diabetes and insomnia (320; 290 finished), single-blind	Gluconate, two 250 mg tablets a day, alone or with potassium	2 months	Lower Insomnia Severity Index (ISI) and changes in melatonin and cortisol
Hausenblas 2024	Ages 35-55 with self-assessed sleep problems (80)	Magnesium L-threonate, 1 g a day	21 days	Better questionnaire and Oura-ring scores than placebo, which declined
Lopresti 2026	Ages 18-45 with dissatisfied sleep (100)	Magnesium L-threonate, 2 g a day	6 weeks	Better sleep-related impairment (p = 0.043); no significant advantage over placebo for sleep disturbances (p = 0.316) or ring sleep outcomes
Schuster 2025	Ages 18-65 with poor sleep (155)	Bisglycinate, 250 mg elemental magnesium a day	To week 4	ISI down 3.9 vs 2.3 on placebo (p = 0.049), effect size 0.2

Sources: Held 2002 (Pharmacopsychiatry), Nielsen 2010 (Magnesium Research), Abbasi 2012 (J Res Med Sci), Liu 2016 (J Alzheimers Dis), Macian 2022 (Nutrients), Gholizadeh-Moghaddam 2023 (Health Sci Rep), Khalid 2024 (Front Endocrinol), Hausenblas 2024 (Sleep Med X), Lopresti 2026 (Front Nutr), Schuster 2025 (Nat Sci Sleep).

Two patterns stand out. The trials in older adults that found gains had 12 to 100 people. And the PCOS and fibromyalgia trials found no sleep benefit.

What does “low certainty” mean for magnesium?

Low certainty means further research is very likely to change the estimate of the effect, and very low means any estimate is highly uncertain. These are labels from the GRADE system the review used, which rates how far you can trust a result.

In practice, the review gives you a direction and no reliable size. Magnesium might help, might not, and a trial run next year could flip the answer. The reasons are visible in the table: tiny samples (12, 44, 46), short follow-up (20 days to 12 weeks), different forms and doses, and sleep measured mostly by questionnaire.

Questionnaires matter because sleep trials have strong placebo responses. In Liu 2016 the authors said strong placebo effects on sleep stopped them judging the treatment, and in Nielsen 2010 they suspect a placebo effect. Schuster 2025 (Nature and Science of Sleep) reported an effect size of 0.2, which its authors call modest.

How much faster could magnesium help you fall asleep?

The best pooled estimate is about 17 minutes faster, but it comes from two small trials with moderate-to-high risk of bias. A 2021 meta-analysis included three trials with 151 older adults. Its pooled sleep-onset estimate came from two trials with 55 participants: 17.36 minutes shorter on magnesium than placebo (95% CI -27.27 to -7.44) (BMC Complementary Medicine and Therapies, 2021). Its authors rated the evidence low to very low quality.

Here is the arithmetic for someone who takes 45 minutes to fall asleep.

• Point estimate: 45 - 17.4 = about 28 minutes.
• Best end of the range: 45 - 27.3 = about 18 minutes.
• Worst end of the range: 45 - 7.4 = about 38 minutes.

That range still assumes the average trial result applies to you and that the trials were unbiased. Total sleep time rose by 16.06 minutes, but that gain was not statistically significant. The base is 55 people.

Which magnesium is best for sleep?

No trial has shown which magnesium form is best for sleep, because none of the ten trials above compared two forms directly. The review also notes that results differed across formulations without a form-by-form ranking. For the head-to-head view, see magnesium glycinate vs citrate for sleep.

The table shows what has been tested: glycinate (in its bisglycinate form), L-threonate, citrate, gluconate, oxide and chloride. Each has one to three small trials, in different people, at different doses. Web pages that name glycinate or citrate the winner are drawing on absorption claims, and the trials above do not test that.

If you decide to try a supplement, the form is a lower priority than the daily amount, whether it contains other ingredients, and whether it suits your medicines and kidneys.

How much magnesium for sleep is safe?

The NHS says 400 mg or less a day of magnesium from supplements is unlikely to cause harm. Taking more than that for a short time can cause diarrhoea, and the NHS says there is not enough evidence on high doses over a long time (NHS, page last reviewed 2020-08-03).

No dose is proven for sleep. The table’s listed amounts run from 100 mg to 500 mg a day, and the L-threonate trials used 1 g and 2 g of the compound, which holds less elemental magnesium than that. Held 2002 used 30 mmol a day, which is about 729 mg of elemental magnesium (30 x 24.3). That is above the NHS figure, and the study was 12 people for 20 days.

The NHS puts the amount you need from all sources at 300 mg a day for men and 270 mg for women aged 19 to 64. A supplement adds to what you already eat.

If you have kidney disease, check with your doctor before starting, because the kidneys clear excess magnesium.

Who might still try magnesium for sleep?

The people with the most reason to try it are those who eat little magnesium, and older adults, though neither group has proof. The 155-person Schuster 2025 trial ran an exploratory analysis suggesting larger gains in people with lower dietary magnesium, and its authors say this needs confirming. In Nielsen 2010, 58% of participants were eating less than the US Estimated Average Requirement for magnesium.

Older adults dominate the earliest positive trials: Held 2002 (60-80), Abbasi 2012 (older adults with insomnia) and Nielsen 2010 (over 51). The 2021 meta-analysis of older adults rated its three trials at moderate-to-high risk of bias and called the evidence substandard for firm recommendations.

If you are in midlife and wake at night, magnesium is one of several options, and not the best supported. Compare it with other options in sleep aids for menopause, compared by evidence, or weigh it against melatonin for older adults. If the problem is repeated waking, sleep restriction therapy is a behavioural treatment for insomnia, and the guide to why you wake at 3am covers causes a supplement will not touch. If cramps are what wake you, start with leg cramps at night. For hormone-related sleep loss, see perimenopause insomnia.

See a doctor if poor sleep has lasted more than a few weeks, if you snore loudly or stop breathing at night, or if you feel very sleepy in the day. Those can point to causes a supplement will not fix.

Is magnesium for sleep worth trying?

Magnesium is a low-cost experiment with a small upside: stay under 400 mg a day and judge it on your own numbers. The 2026 review does not support routine use, and its low certainty leaves room for a real benefit in some people.

A simple way to judge it for yourself is to keep a sleep diary for two weeks before you start and two weeks after, so you compare numbers rather than impressions. Remember that the placebo arms in these trials often improved too. If nothing changes after a few weeks, stop.