Walk into any supplement shop and the magnesium shelf reads like a personality test. Glycinate for sleep, citrate for digestion, oxide for the budget conscious, threonate for the brain. The labels promise calm nights and deeper rest, and the internet has already picked a winner: glycinate, usually, by a wide margin. The trials tell a flatter story. No study has ever compared one magnesium form against another for sleep, and the best review of the field, published in 2026, concluded that the evidence does not support magnesium as a routine treatment for insomnia at all. That does not make magnesium useless. It makes the question sharper than the marketing: which form has the best evidence, how big is the effect, and who is most likely to notice anything?

The honest answer is that glycinate has the strongest single trial, citrate has the best absorption data, oxide has the oldest insomnia trial, and threonate has the loudest brain claim with the thinnest backing. None of them has been proven best, because the head to head trial that would settle it has never been run. What follows is what each form can actually claim, with the numbers attached.

The short answer: no proven winner, but glycinate has the best trial

If you want one sentence: no magnesium form has been proven best for insomnia, because no trial has compared forms against each other for sleep, but the strongest single trial tested magnesium bisglycinate, the form sold as glycinate, and found a small improvement in insomnia scores after four weeks. In that 2025 trial, 155 adults who reported poor sleep took 250 mg of elemental magnesium as bisglycinate or a placebo each night. Insomnia Severity Index scores fell by 3.9 points with magnesium and 2.3 points with placebo, a statistically significant difference with a small effect size. The benefit was larger in people whose diets were low in magnesium. That is the best evidence any form currently has, and it is modest: a 1.6 point difference on a 28 point scale, with no change in daytime sleepiness, fatigue, stress or mood.

Everything else on the shelf rests on weaker ground. Citrate beats oxide for absorption in small trials, but its one sleep trial found no advantage over placebo. Oxide is the form in the oldest positive insomnia trial, from 2012, and also the form with the lowest absorption. Threonate has two small sleep trials with mixed results, both tied to the company that sells it. So the practical choice is not about picking a proven winner. It is about matching the form to the evidence, the dose to the official limits, and your own gut tolerance to the chemistry.

What the trials actually show about magnesium and sleep

Before judging any form, it helps to see how thin the whole field is. The best pooled analysis is a 2021 systematic review and meta-analysis by Mah and Pitre, published in BMC Complementary Medicine and Therapies. It found only three randomised trials of oral magnesium for insomnia in older adults, covering 151 people in total. Pooled together, magnesium shortened the time to fall asleep by about 17 minutes compared with placebo, while total sleep time rose by about 16 minutes, a change that was not statistically significant. Every trial carried a moderate to high risk of bias, and the reviewers graded the evidence low to very low quality. Their conclusion was blunt: the literature was not good enough for doctors to make well informed recommendations.

The picture did not improve much by 2026. A systematic review by Lopresti, Smith and Drummond in the Journal of Dietary Supplements looked at 12 placebo controlled trials of oral magnesium alone for sleep, insomnia or poor sleep, searched to June 2026. The trials differed too much in who took part, which form was used, the dose and the outcome measured for the results to be pooled. Some reported improvements in insomnia severity, sleep quality or daytime functioning. Several found no magnesium specific benefit at all, particularly for overall sleep quality and for objective measures such as actigraphy and EEG. The certainty of evidence was again low to very low, and the authors concluded that current evidence does not support oral magnesium as a routine treatment for insomnia. They also noted that no trial had compared one form with another, so no form could be called superior.

Two of the older trials deserve a closer look, because they show how much the form matters to the result. Abbasi and colleagues ran a double blind trial in 2012 in 46 older adults with primary insomnia. They took 500 mg of magnesium a day as magnesium oxide, or a placebo, for eight weeks. The magnesium group fell asleep faster, slept longer and woke less during the night, with better sleep efficiency. That is the most encouraging single result in the older literature, and it used oxide, the form most often dismissed as useless. Then there is Nielsen and colleagues in 2010: 100 adults aged 51 to 85 with poor sleep took 320 mg of magnesium a day as magnesium citrate, or a placebo, for seven weeks. Sleep quality scores improved in both groups, from 10.4 to 6.6 on the Pittsburgh Sleep Quality Index, with no advantage for magnesium. Same mineral, different form, different dose, different population, opposite headline. That is the field in miniature.

Magnesium glycinate: the form with the best trial

Magnesium glycinate, more precisely magnesium bisglycinate, is magnesium bound to two molecules of glycine, the simplest amino acid. It is the form most often sold for sleep, and since 2025 it is also the form with the best trial. Schuster and colleagues at Leibniz University Hannover randomised 155 healthy adults aged 18 to 65 who reported poor sleep to 250 mg of elemental magnesium as bisglycinate, or a matching placebo, taken 30 to 60 minutes before bed for four weeks. The trial was published in Nature and Science of Sleep in 2025.

The result was real but small. Insomnia Severity Index scores fell by 3.9 points in the magnesium group and 2.3 points in the placebo group, a difference of 1.6 points that was statistically significant at p = 0.049, with a small effect size of Cohen’s d = 0.2. To put that in context, a drop of 3 points on the ISI is generally treated as the smallest change worth noticing, and both groups cleared that bar, so much of the improvement was the placebo effect plus the passage of time. The magnesium group did a little better on top. Exploratory analysis suggested the benefit was larger in people who reported eating less magnesium, which fits the simplest theory in this field: magnesium is most likely to help when you are short of it. No significant differences appeared in the other measures, including daytime sleepiness, fatigue, stress and mood, and the trial used no objective sleep measurement, so nothing can be said about deep sleep or sleep stages.

There is also a confound that honest coverage should not hide. Each capsule in the trial contained 893 mg of magnesium bisglycinate, supplying 125 mg of magnesium and 761.5 mg of glycine, so the daily dose delivered 1,523 mg of glycine alongside the magnesium. Glycine itself has been tested for sleep at 3 g before bed in small trials, with some positive results. The Schuster trial cannot separate how much of the benefit came from the magnesium and how much from the glycine, and the authors say so in the paper. That does not cancel the result, but it does mean the trial is evidence for bisglycinate as a product, not proof that the magnesium inside it is doing the work.

On tolerability, glycinate has a practical advantage that is widely reported and plausible, though not proven in a head to head trial: it is usually gentle on the stomach. The chelated form does not tend to draw water into the bowel the way citrate and oxide can, so it is the form most often recommended for evening use by people who have had loose stools from other forms. That is a tolerability argument, not an efficacy argument, and it should be stated as one.

Magnesium citrate: better absorbed, weaker sleep evidence

Magnesium citrate is magnesium bound to citric acid. It dissolves well in liquid, and the absorption evidence is the best of any form. The US National Institutes of Health Office of Dietary Supplements states that the aspartate, citrate, lactate and chloride forms tend to be absorbed more completely than magnesium oxide and magnesium sulfate. A small randomised trial by Werner and colleagues in 2019 tested this directly: 14 healthy men, their magnesium pools first saturated, took a single 400 mg dose of citrate or oxide. Urinary magnesium over 24 hours rose significantly after citrate but not after oxide, and blood magnesium was significantly higher after citrate at four and eight hours. The authors concluded citrate showed higher bioavailability. Fourteen people and a single dose is a thin base, but it is a real head to head measurement, which is more than the sleep field has.

Yet citrate’s sleep record is the weakest of the three common forms. Its one dedicated sleep trial is the Nielsen 2010 study described above: 100 older adults, 320 mg a day for seven weeks, and no advantage over placebo. The 2026 systematic review noted that trials using citrate did not show consistent sleep benefits. That creates an awkward gap for anyone shopping on absorption alone: the form that gets into the blood most reliably is not the form with the sleep results. Absorption is a step on the way to an effect, not the effect itself.

Citrate also has the most obvious side effect. Unabsorbed magnesium salts act osmotically, pulling water into the colon, and citrate is the form most likely to loosen stools at higher doses. That is why it is also sold as a laxative. For a sleep supplement taken at night, a laxative effect is a poor trade, and it is the main practical reason citrate is a worse evening choice than glycinate for many people, even though it absorbs better.

Magnesium oxide: cheap, poorly absorbed, still in the old trials

Magnesium oxide is the form most supplement sceptics love to dismiss, and the dismissal is half right. Oxide is about 60 percent elemental magnesium by weight, so a small tablet carries a large dose, and it is the cheapest form to make. Its fractional absorption is low, around 4 percent in the most cited estimate, and the NIH fact sheet places it below citrate, lactate, chloride and aspartate for bioavailability. In the Werner trial above, a single oxide dose did not raise 24 hour urinary magnesium significantly. It is also among the forms most often linked to stomach upset and diarrhoea.

But oxide is not useless for sleep, because the oldest positive insomnia trial used it. The Abbasi 2012 trial gave 500 mg a day of magnesium oxide to older adults with insomnia for eight weeks and found better sleep scores, longer sleep and better sleep efficiency than placebo. An older crossover trial from 1994, in 12 patients who had part of their bowel removed, found overall absorption of a diglycinate chelate and oxide was effectively the same, about 23.5 percent versus 22.8 percent, though the chelate was better tolerated and absorbed about twice as well in the four poorest oxide absorbers. The lesson is that absorption percentages do not decide the sleep question. A poorly absorbed form at a high dose can still deliver enough magnesium to matter, and a well absorbed form can still fail to move sleep. Oxide’s real drawback is tolerability and dose: at the 500 mg used in the Abbasi trial, you are above the US supplemental upper limit of 350 mg, and loose stools are likely. If oxide is what you have, it is not a wasted purchase, but it is the worst evening choice of the three.

Magnesium L-threonate: the brain claim, examined

Magnesium L-threonate, sold under the trademark Magtein, is marketed as the form that best raises magnesium in the brain. That claim rests largely on animal research, and the human sleep trials are small and mixed. It is also the most expensive form by a wide margin, and each gram of the compound carries only about 75 mg of elemental magnesium, so the doses in the trials were small in magnesium terms.

The first sleep trial, by Hausenblas and colleagues in 2024, randomised 80 adults aged 35 to 55 with self reported sleep problems to 1 g of Magtein a day, about 75 mg of elemental magnesium, or a placebo, for 21 days. The headline result was mixed: the main insomnia score showed no significant difference between groups, with a group by time interaction of p = 0.390, but some proprietary Oura ring scores, including deep sleep and REM sleep scores, were better with threonate, and daytime energy and mood measures improved. There was no significant difference in total, deep or REM sleep minutes. The trial was funded by AIDP, the company that supplies the ingredient, and a 2025 corrigendum disclosed that the trial had been registered retrospectively and added complete statements about AIDP funding and three authors’ financial relationships with the company.

The second trial, by Lopresti and Smith in 2026, randomised 100 adults aged 18 to 45 who were dissatisfied with their sleep to 2 g of Magtein a day, supplying 145 mg of elemental magnesium, or a placebo, for six weeks. Sleep related impairment improved, but the Oura ring showed no significant difference in sleep outcomes, and for total sleep time the p value was 0.461. The primary outcome of that trial was cognition, not sleep. Taken together, the threonate evidence is two small trials, one with a null main outcome and one with no objective sleep difference, both using doses far below the magnesium amounts in the glycinate and oxide trials. That is not a case for paying several times more.

The glycine question nobody mentions

There is a wrinkle in the glycinate story that most product pages skip. The best glycinate trial delivered 1,523 mg of glycine a day alongside the magnesium, and glycine on its own has sleep evidence. Three small trials tested 3 g of glycine before bed, Inagawa in 2006, Yamadera in 2007 and Bannai in 2012, and reported better sleep quality and less daytime fatigue. The dose in the bisglycinate trial was about half that, so it is not a straight comparison, but it is enough to stop anyone claiming the trial proves magnesium itself improves sleep. It proves that bisglycinate capsules, magnesium plus glycine together, improved insomnia scores a little more than placebo over four weeks. If you already take glycine, or you eat a protein rich diet, the magnesium specific benefit may be smaller still. This is the kind of detail that matters for self experimenters, because it changes what you are actually testing.

How much to take, and when

Dose is where the label games start, so read the elemental number, not the compound weight. Supplement labels declare elemental magnesium, the actual magnesium, not the weight of the whole compound, and the gap is large: bisglycinate is only about 14 percent magnesium by weight, which is why the Schuster trial needed two 893 mg capsules to deliver 250 mg of elemental magnesium. Citrate is about 16 percent, oxide about 60 percent. A 1,000 mg tablet of oxide carries roughly 600 mg of magnesium, while a 1,000 mg tablet of glycinate carries roughly 140 mg. Comparing products by tablet weight is meaningless.

For the daily amount, the UK reference nutrient intake is 300 mg a day for men and 270 mg for women aged 19 to 64, from all sources, and the NHS advises that most people can get that from food. The US recommended dietary allowance is higher, 400 to 420 mg for men and 310 to 320 mg for women. For supplements specifically, the UK Expert Group on Vitamins and Minerals sets a safe upper level of 400 mg a day, and the NHS says having 400 mg or less a day from supplements is unlikely to cause any harm. The US tolerable upper intake level for supplemental magnesium is 350 mg a day. The trial doses sit inside those limits, except for the Abbasi oxide trial at 500 mg: glycinate 250 mg, citrate 320 mg, threonate 75 to 145 mg. A reasonable self experiment, if you choose to run one, is 200 to 250 mg of elemental magnesium as bisglycinate, 30 to 60 minutes before bed, which matches the best trial. Do not stack a high dose supplement on top of a magnesium rich diet and a multivitamin without adding the numbers up.

Food comes first in every guideline, and it is worth a sentence on why. About 30 to 40 percent of the magnesium in food is absorbed, and good sources are pumpkin seeds, leafy greens, nuts, beans and whole grains. Survey data in the US found 48 percent of Americans eat less magnesium than their estimated average requirement, and the people most likely to be low are older men and teenagers. If your diet is already rich in those foods, a supplement has less room to help, which is consistent with the Schuster finding that low intake participants improved most. Before buying anything, a week of eating magnesium rich foods is a cheaper test than a bottle of capsules.

Who should skip it, and the drug interactions

Magnesium is generally safe in healthy adults, but there are three groups who should talk to a clinician first. The first is anyone with kidney disease: the kidneys control magnesium excretion, and when they cannot, magnesium can build up to dangerous levels. The second is anyone taking medicines that interact with magnesium. The NIH fact sheet names oral bisphosphonates such as alendronate, tetracycline antibiotics and quinolone antibiotics, all of which magnesium can stop from being absorbed properly, and notes that diuretics and proton pump inhibitors can affect magnesium status. Separating magnesium from those medicines by at least two hours is the standard advice for bisphosphonates. The third is anyone pregnant, breastfeeding, or taking medicines that affect heart rhythm or blood pressure, where even a low risk supplement deserves a conversation rather than a self experiment.

The common side effect is the one the upper limits are built around: loose stools. Taking more than 400 mg of supplemental magnesium for a short time can cause diarrhoea, according to the NHS, and that is usually the first and only sign you have taken too much. It settles when the dose drops. Serious toxicity, hypermagnesaemia, is rare in people with healthy kidneys and is not a realistic worry at the doses in the sleep trials. Magnesium applied to the skin, sprays and oils, deserves a brief note: no trial in the 2026 review tested it for sleep, and it has not been shown to raise body magnesium reliably, so it is not a substitute for the oral evidence above.

A four week self test that actually tells you something

Because the expected effect is small, a casual try will not tell you whether magnesium helped. A structured four week test will. First, spend a week tracking your sleep without changing anything: bedtime, time to fall asleep, wake ups, morning rating out of ten, and a simple insomnia score if you want one, the seven question Insomnia Severity Index is free to use and takes two minutes. Second, pick one form, bisglycinate at 250 mg elemental, taken 30 to 60 minutes before bed, and keep everything else the same: same bedtime, same caffeine cut off, same evening light. Third, track for four weeks, the length of the Schuster trial, and compare the last two weeks with the baseline week. A drop of 3 or more points on the ISI, or falling asleep noticeably faster on most nights, is a signal worth keeping. A change of one point, or a good week followed by a bad one, is noise. Fourth, set a stop rule in advance: if nothing has changed after four weeks, stop, because the trials say the effect, if it exists, appears within that window. If loose stools appear, drop the dose or switch forms rather than pushing through. And do not test magnesium at the same time as a new sleep routine, a new pillow, or a new supplement, or you will never know which change did the work.

If magnesium does nothing

Most people who try magnesium for sleep will notice little or nothing, and that is the expected outcome from the trials, not a failure of the experiment. The treatment with the strongest evidence for chronic insomnia is cognitive behavioural therapy for insomnia, CBT-I, which the Schuster paper notes improves sleep efficiency by 10 to 20 percent and reduces insomnia severity in up to 60 percent of patients, with benefits that often last. It is the first line treatment in clinical guidelines, ahead of any supplement or sleeping pill. If poor sleep has lasted more than a few weeks, is affecting your days, or involves loud snoring, pauses in breathing, or severe daytime sleepiness, that is a conversation with a GP or a sleep clinic, not a supplement aisle decision. Magnesium is a reasonable, low cost, low risk experiment for mild sleep difficulty, particularly if your diet is low in it. It is not a treatment for insomnia, and no form on the shelf has earned the title of best.

Frequently asked questions

Which magnesium is best for insomnia?

No form has been proven best, because no trial has compared magnesium forms against each other for sleep. The strongest single trial tested magnesium bisglycinate, the form sold as glycinate, at 250 mg of elemental magnesium a day for four weeks in 155 adults, and found a small improvement in insomnia scores over placebo. That makes glycinate the best evidenced choice, not a proven winner.

Is magnesium glycinate better than citrate for sleep?

There is no trial comparing glycinate with citrate for sleep, so the question cannot be answered directly. Glycinate has the better sleep trial, while citrate has the better absorption data: a small trial found citrate raised blood and urinary magnesium more than oxide, and citrate’s own sleep trial found no advantage over placebo. Glycinate is also usually gentler on the stomach, which matters for a supplement taken at night.

How much magnesium should I take for sleep?

The trials used 250 mg of elemental magnesium a day as bisglycinate, 320 mg as citrate, and 500 mg as oxide. The UK safe upper level for magnesium from supplements is 400 mg a day, and the US upper level is 350 mg. Check the elemental magnesium on the label, not the compound weight, and add up everything you take, including multivitamins.

How long does magnesium take to help sleep?

In the best trial, the difference from placebo appeared by week four, with most improvement in the first two weeks. The older trials ran three to eight weeks. If you run a self test, give it four weeks at a fixed dose, then stop if nothing has changed.

Can magnesium cause diarrhoea?

Yes. Loose stools are the most common side effect of magnesium supplements, and the NHS notes that taking more than 400 mg for a short time can cause diarrhoea. Citrate and oxide are more likely to cause it than glycinate. Lowering the dose usually settles it.

Is magnesium L-threonate better for sleep?

Not on current evidence. Its two sleep trials were small and mixed: one found no significant difference in its main insomnia score, and the other found no difference in wearable measured sleep. Both used low elemental magnesium doses, 75 mg and 145 mg a day, and the first was funded by the company that supplies the ingredient. The brain claim rests largely on animal research.

Sources

• Mah J, Pitre T. Oral magnesium supplementation for insomnia in older adults: a systematic review and meta-analysis. BMC Complementary Medicine and Therapies. 2021;21:125. PubMed 33865376
• Schuster J, Cycelskij I, Lopresti A, Hahn A. Magnesium bisglycinate supplementation in healthy adults reporting poor sleep: a randomized, placebo-controlled trial. Nature and Science of Sleep. 2025;17:2027-2040. PMC12412596
• Lopresti AL, Smith SJ, Drummond PD. Magnesium supplementation for sleep in adults: a systematic review of randomized controlled trials. Journal of Dietary Supplements. 2026;23(5):553-581. DOI: 10.1080/19390211.2026.2719670
• 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;23(4):158-168.
• 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;17(12):1161-1169.
• Hausenblas HA, Lynch T, Hooper S, Shrestha A, Rosendale D, Gu J. Magnesium-L-threonate improves sleep quality and daytime functioning in adults with self-reported sleep problems: a randomized controlled trial. Sleep Medicine: X. 2024;8:100121. PMC11381753
• Lopresti AL, Smith SJ. The effects of magnesium L-threonate (Magtein) on cognitive performance and sleep quality in adults: a randomised, double-blind, placebo-controlled trial. Frontiers in Nutrition. 2026;12:1729164. Frontiers in Nutrition
• Werner T, Kolisek M, Vormann J, Pilchova I, Grendar M, Struharnanska E, Cibulka M. Assessment of bioavailability of Mg from Mg citrate and Mg oxide by measuring urinary excretion in Mg-saturated subjects. Magnesium Research. 2019;32(3):63-71. PubMed 32162607
• National Institutes of Health, Office of Dietary Supplements. Magnesium: Health Professional Fact Sheet. NIH ODS magnesium fact sheet
• NHS. Vitamins and minerals: magnesium. Guidance on daily amounts and the 400 mg supplement limit.

Related reading

• The 6.4 to 7.8 Hour Sleep Window: Dialling In Your Longevity Sweet Spot
• Can Sleep Slow Biological Ageing?
• What Are the Top 3 Supplements for Longevity?