Does magnesium help you sleep?

Probably a little, and the evidence is thinner than the shelf space suggests. Every randomised trial of magnesium for insomnia ever pooled comes to three trials in 151 older adults, and the reviewers who pooled them rated the quality low to very low. They found people fell asleep about 17 minutes sooner. Total sleep time did not change in a way that reached significance.

Unsettled. The honest sentence from that review is worth quoting rather than paraphrasing: the quality of the literature is substandard for physicians to make well-informed recommendations about oral magnesium for older adults with insomnia. That is the state of the evidence behind one of the best-selling supplements in the world.

What the three trials were

Three randomised controlled trials, in three countries, comparing oral magnesium with placebo in older adults who had insomnia. All were at moderate to high risk of bias. The pooled effect on how long it took to fall asleep was 17.36 minutes, with a confidence interval running from 7 to 27 minutes. The pooled effect on total sleep time was 16 minutes and did not reach statistical significance.

The reviewers still ended up mildly in favour, and their reasoning is refreshingly practical rather than statistical. Magnesium is cheap and widely available, so a trial of it costs the patient very little even if the evidence is weak. That is a fair argument. It is not the same argument as "clinically proven", which is what the packaging tends to imply.

A separate review of 31 randomised trials of supplements for sleep quality reached the same place from the other direction. It found enough evidence to pool for amino acids, melatonin and vitamin D, and not enough for magnesium.

Myth. Magnesium is not a sedative and it does not work like one. If you are looking for something with a measurable effect on getting to sleep, melatonin has far more trial evidence behind it, and the behavioural treatment for insomnia has more again. Magnesium is the cheapest thing to try and the weakest thing to rely on.

What magnesium does do

This is the part that gets lost, because the sleep claim is louder than the results that are actually solid.

Blood pressure, in the people who have a problem

Across 38 randomised trials in 2,709 people, magnesium lowered systolic blood pressure by 2.81 mmHg and diastolic by 2.05. In people already taking blood pressure medicine the systolic drop was 7.68 mmHg, and in people with low blood magnesium it was 5.97. In people with normal blood pressure nothing happened.

That pattern repeats everywhere in this nutrient and it is the single most useful thing to understand about it. Magnesium does something measurable in people who are short of it or ill, and close to nothing in people who are neither.

Migraine, modestly

Three of four small short placebo-controlled trials found modest reductions in migraine frequency at up to 600 mg a day. Small, short and four is a fair description of that evidence.

Cramps, no

Myth. Night cramps are the second reason people buy magnesium and the claim with the clearest negative answer. A Cochrane review of five trials in 307 older adults found the number of cramps a week changed by 0.18 against placebo, on moderate-certainty evidence. Restless legs is worse served still: a review of everything published found one randomised trial, three case series and four case reports, and could reach no conclusion at all.

Mood, where the trials are small but real

Seven randomised trials in 325 adults who already had a depressive disorder found magnesium lowered depression scores, with a standardised difference of 0.92. That is a large effect on a small base, and the reviewers say plainly that the trials are too small to interpret clinically. It is the one claim in this family where randomised evidence exists at all, and it is still not a treatment.

Mood in the wider population, only in observational data

Across 13 observational studies in 63,214 adults, people eating the most magnesium had 34 per cent less depression than those eating the least, and each extra 100 mg a day tracked with 7 per cent less. That is a real pattern in the data and it is not evidence that a tablet fixes mood. People eating more magnesium are eating more nuts, beans, whole grains and greens, and living differently in a dozen other ways.

Why the shortfall is real even when the supplement is not the answer

Around 48 per cent of Americans take in less magnesium from food and drink than the estimated average requirement. That is not a fringe group. The reason is structural: magnesium lives in the bran of grains, the germ, the skins of seeds, and refining takes it out.

Unsettled. The requirement that 48 per cent is measured against is itself contested, and the challenge comes from the same kind of study the requirement was built on. Pooling every controlled balance experiment run at the US Department of Agriculture laboratory in Grand Forks, 150 women and 93 men on weighed diets for at least 28 days, the body stopped losing magnesium at an intake of 165 mg a day. The researchers who ran those wards put the average requirement near 175 mg and the recommended intake near 250 mg for a 70 kg adult. Against the official 400 to 420, that is a wide disagreement, and nobody has resolved it.

Established. The upper limit for magnesium applies only to supplements and medicines. Magnesium in food and drink is not capped, because the gut absorbs proportionally less as intake rises and the kidneys clear the rest. You cannot overdo it from pumpkin seeds.

What to do with this

If you are not sleeping

Two numbers put the cost of trying it in proportion. Across the cramp trials, minor side effects were about half as common again on magnesium as on placebo, almost all of them in the gut, while serious events were no different at two in 72 against three in 68. In a six-month trial of 450 mg a day in overweight adults with working kidneys, the single serious event reported was judged to have nothing to do with the treatment. So the downside of a short trial of it is small, and that is the honest argument in its favour.

Two things make an ordinary dose easier to live with, and neither of them is printed on the bottle. Split it across the day instead of swallowing it in one go, and take it with food. The splitting advice has a measurement under it from outside the sleep literature: twelve men drinking the same 126 mg of magnesium in mineral water absorbed 50.7 per cent of it in seven small servings against 32.4 per cent in two large ones.

Magnesium is a reasonable cheap experiment and a poor plan. Give it two weeks at an ordinary dose and judge it honestly. If your sleep problem has lasted more than a month, the treatment with the strongest evidence is cognitive behavioural therapy for insomnia, and it is worth asking about before another bottle.

If you take it for cramps

The trials say it does not work for that. Eating the food is still worth it for other reasons.

If you want the benefit that is real

Eat the magnesium rather than swallowing it. A handful of seeds or nuts most days, wholemeal instead of white where you do not mind the swap, beans and leafy greens. The requirement is 400 to 420 mg a day for men and 310 to 320 for women.

Two medicines that change the answer

Long-term loop or thiazide diuretics increase magnesium losses in urine and can deplete it, and prescription proton pump inhibitors taken for more than about a year can push blood magnesium below normal. If you take either and your sleep is poor, magnesium is at least a sensible thing to have measured rather than guessed. Two more interactions are about timing: magnesium binds tetracycline and quinolone antibiotics, and it reduces absorption of oral bisphosphonates, so those tablets want a few hours of distance.

Not for everyone. Two cautions if you do supplement. High doses commonly cause diarrhoea, nausea and cramping, which is the body refusing the excess and also the reason magnesium is the active ingredient in some laxatives and antacids. And if your kidney function is reduced, magnesium accumulates instead of clearing, which turns a harmless-looking supplement into a real risk. That one is a conversation with a doctor rather than a shelf decision.

What would change this answer

• A properly powered randomised trial of magnesium for insomnia in adults of any age. Three small trials in 151 people is not a literature, and this is the rare case where one good trial would settle a question that millions of people are already spending money on.
• A trial that recruits people with low blood magnesium specifically. Everything in this nutrient points to the effect living there, and the sleep trials did not select for it. That is harder than it sounds. In one group of adults, neither serum magnesium nor the ionised magnesium in whole blood tracked how much magnesium they were eating, so the test you would screen with does not measure the thing you want to select on.
• Head-to-head comparison of the salts people are sold. Glycinate, citrate, oxide and threonate are marketed as different products for different purposes, and the trials above mostly did not compare them at all. What exists is narrative rather than trial. Gluconate and chloride are preferred for oral replacement because they cause diarrhoea less often than the other salts, carbonate is avoided as too insoluble to absorb properly, and citrate dissolves readily enough to move the bowels more than oxide does. That is a summary written by people reading single studies, which is why it sits in this list of what is missing rather than in the answer above.
• A reason to distrust the word on the label even when the salt is the same. Thirteen men took 450 mg of supplemental magnesium oxide on a standardised diet, once as an effervescent tablet dissolved in water and once as a capsule. Urinary magnesium rose 40 per cent after the tablet and 20 per cent after the capsule. Urine is a proxy for absorption rather than absorption itself, and thirteen young men is not a literature. It is still enough to say that two bottles reading "magnesium oxide, 450 mg" are not necessarily the same dose.

How we searched

Sources

1. mg-s01 · Meta-analysis or systematic review · systematic review and meta-analysis of randomised trials · n = 151
   Pooled analysis showed that post-intervention sleep onset latency time was 17.36 min less after magnesium supplementation compared to placebo (95% CI - 27.27 to - 7.44, p = 0.0006). Total sleep time improved by 16.06 min in the magnesium supplementation group but was statistically insignificant.
   
   Who: older adults with insomnia in three trials in three countries
   Effect: sleep onset latency 17.36 minutes shorter (95% CI 7.44 to 27.27); total sleep time 16.06 minutes longer and not statistically significant
   Certainty: all trials at moderate to high risk of bias, evidence rated low to very low
   Mah and Pitre, BMC Complementary Medicine and Therapies, 2021 · checked 2026-09-16 · we read the abstract
2. mg-s02 · Meta-analysis or systematic review · systematic review and meta-analysis of randomised trials · n = 151
   This review confirms that the quality of literature is substandard for physicians to make well-informed recommendations on usage of oral magnesium for older adults with insomnia.
   
   Who: older adults with insomnia
   Effect: quality of the literature called substandard; cheapness and availability offered as the practical argument
   Certainty: GRADE rating low to very low
   Mah and Pitre, BMC Complementary Medicine and Therapies, 2021 · checked 2026-09-16 · we read the abstract
3. mg-s03 · Meta-analysis or systematic review · systematic review and meta-analysis of randomised trials · n = 31 trials
   Although not all studies provided adequate data for meta-analysis, we also discussed how magnesium, zinc, resveratrol and nitrate supplementation may improve sleep quality.
   
   Who: adults taking dietary supplements for subjective sleep quality, 31 trials
   Effect: amino acids mean difference -1.27, melatonin -1.21, vitamin D -1.63 on sleep quality indices; magnesium not poolable
   Certainty: high heterogeneity for melatonin and vitamin D
   Chan and Lo, Postgraduate Medical Journal, 2022 · checked 2026-09-16 · we read the abstract
4. mg-s04 · Meta-analysis or systematic review · systematic review, meta-analysis and dose-response analysis of randomised trials · n = 2,709
   Magnesium intake resulted in a reduction in systolic BP of -2.81 mm Hg (95% CI, -4.32 to -1.29) and diastolic BP by -2.05 mm Hg (95% CI, -3.23 to -0.88) compared with placebo.
   
   Who: hypertensive and normotensive adults, trials of at least four weeks, median dose 365 mg
   Effect: systolic -2.81 mmHg (95% CI -4.32 to -1.29); diastolic -2.05 (-3.23 to -0.88); systolic -7.68 in treated hypertensives and -5.97 in people with low blood magnesium; no effect in normotensive groups
   Certainty: high heterogeneity, no dose-response relationship found
   Zhang et al., Hypertension, 2025 · checked 2026-09-16 · we read the abstract
5. mg-s05 · Observational data · systematic review and dose-response meta-analysis of observational studies · n = 63,214
   Pooling 15 effect sizes from 12 studies (including 50 275 participants) revealed that individuals with the highest Mg intake had a 34% lower risk of depression, compared with those with the lowest Mg intake (RR: 0.66; 95% CI: 0.57, 0.78).
   
   Who: adults in 10 cross-sectional and 3 cohort studies
   Effect: highest against lowest intake relative risk 0.66 (95% CI 0.57 to 0.78); each extra 100 mg a day 0.93 (0.90 to 0.96)
   Certainty: observational only, mostly cross-sectional, GRADE assessed
   Noormohammadi et al., Nutrition Reviews, 2025 · checked 2026-09-16 · we read the abstract
6. mg-s06 · Meta-analysis or systematic review · systematic review · n = 8 studies
   After quality appraisal and synthesis of the evidence we were unable to make a conclusion as to the effectiveness of magnesium for RLS/PLMD.
   
   Who: people with restless legs syndrome or periodic limb movement disorder, 8 studies
   Effect: no conclusion possible; the single randomised trial found no significant effect and may have been underpowered
   Certainty: mostly case reports
   Marshall et al., Sleep Medicine Reviews, 2019 · checked 2026-09-16 · we read the abstract
7. mg-t1 · Meta-analysis or systematic review · Cochrane systematic review and meta-analysis of randomised trials · n = 307 in the main comparison, 735 across the whole review
   the difference in the number of cramps per week at four weeks (MD -0.18 cramps/week, 95% CI -0.84 to 0.49; 5 studies, 307 participants; moderate-certainty evidence)
   
   Who: older adults with idiopathic night cramps, mean age 61.6 to 69.3 years, 5 trials
   Effect: number of cramps a week at four weeks, mean difference -0.18 (95% CI -0.84 to 0.49)
   Certainty: moderate-certainty evidence, little heterogeneity
   Garrison et al., Cochrane Database of Systematic Reviews, 2020 · checked 2026-09-16 · we read the abstract
8. mg-01 · Position of an expert body
   Magnesium is a cofactor in more than 300 enzyme systems that regulate diverse biochemical reactions in the body, including protein synthesis, muscle and nerve function, blood glucose control, and blood pressure regulation.
   
   Who: general
   Effect: physiological role
   NIH Office of Dietary Supplements, Health Professional Fact Sheet: Magnesium · checked 2026-09-15
9. mg-02 · Position of an expert body
   19–30 years: 400 mg [male], 310 mg [female]; 31–50 years: 420 mg [male], 320 mg [female]
   
   Who: adults
   Effect: 400–420 mg men, 310–320 mg women
   NIH Office of Dietary Supplements, Health Professional Fact Sheet: Magnesium · checked 2026-09-15
10. mg-03 · Position of an expert body
    The RDAs include magnesium from all sources—food, beverages, dietary supplements, and medications. The ULs only include magnesium from dietary supplements and medications.
    
    Who: general
    Effect: UL covers supplements only
    NIH Office of Dietary Supplements, Health Professional Fact Sheet: Magnesium · checked 2026-09-15
11. mg-04 · Observational data
    An analysis of data from the National Health and Nutrition Examination Survey (NHANES) of 2013–2016 found that 48% of Americans of all ages ingest less magnesium from food and beverages than their respective EARs.
    
    Who: US population, NHANES 2013–2016
    Effect: 48% below EAR
    NIH Office of Dietary Supplements, Health Professional Fact Sheet: Magnesium · checked 2026-09-15
12. mg-05 · Position of an expert body
    Loss of appetite, nausea, vomiting, fatigue, and weakness are some of the early signs and symptoms of magnesium deficiency.
    
    Who: people with deficiency
    Effect: early symptoms
    NIH Office of Dietary Supplements, Health Professional Fact Sheet: Magnesium · checked 2026-09-15
13. mg-06 · Meta-analysis or systematic review
    Magnesium supplementation only marginally lowers blood pressure
    
    Who: adults
    Effect: marginal reduction in blood pressure
    NIH Office of Dietary Supplements, Health Professional Fact Sheet: Magnesium · checked 2026-09-15
14. mg-07 · Position of an expert body
    The American Diabetes Association states that there is insufficient evidence to support the routine use of magnesium to improve glycemic control in people with diabetes.
    
    Who: people with diabetes
    Effect: insufficient evidence
    NIH Office of Dietary Supplements, Health Professional Fact Sheet: Magnesium · checked 2026-09-15
15. mg-08 · Randomised controlled trial(s)
    Three out of four small, short-term, placebo-controlled trials found modest reductions in the frequency of migraines in patients who were given up to 600 mg/day magnesium.
    
    Who: people with migraine
    Effect: modest reduction in frequency
    NIH Office of Dietary Supplements, Health Professional Fact Sheet: Magnesium · checked 2026-09-15
16. mg-09 · Position of an expert body
    High doses of magnesium from dietary supplements or medications often result in diarrhea that can be accompanied by nausea and abdominal cramping.
    
    Who: people taking high-dose supplements
    Effect: diarrhoea
    NIH Office of Dietary Supplements, Health Professional Fact Sheet: Magnesium · checked 2026-09-15
17. mg-10 · Position of an expert body
    The risk of magnesium toxicity increases with impaired renal function or kidney failure because the ability to remove excess magnesium is reduced or lost.
    
    Who: people with renal impairment
    Effect: toxicity risk
    NIH Office of Dietary Supplements, Health Professional Fact Sheet: Magnesium · checked 2026-09-15
18. mg-11 · Position of an expert body
    Magnesium-rich supplements or medications can decrease the absorption of oral bisphosphonates, such as alendronate (Fosamax), that are used to treat osteoporosis.
    
    Who: people taking alendronate or another oral bisphosphonate
    Effect: decreased absorption of the bisphosphonate
    NIH Office of Dietary Supplements, Magnesium - Health Professional Fact Sheet · checked 2026-09-16 · we read the full text
19. mg-12 · Position of an expert body
    Magnesium can form insoluble complexes with tetracyclines, such as demeclocycline (Declomycin) and doxycycline (Vibramycin), and with quinolone antibiotics, such as ciprofloxacin (Cipro) and levofloxacin (Levaquin).
    
    Who: people taking doxycycline, ciprofloxacin or similar antibiotics
    Effect: insoluble complexes with the antibiotic
    NIH Office of Dietary Supplements, Magnesium - Health Professional Fact Sheet · checked 2026-09-16 · we read the full text
20. mg-13 · Position of an expert body
    Chronic treatment with loop diuretics, such as furosemide (Lasix) and bumetanide (Bumex), and thiazide diuretics, such as hydrochlorothiazide (Aquazide H) and ethacrynic acid (Edecrin), can increase the loss of magnesium in urine and lead to magnesium depletion.
    
    Who: people on furosemide, bumetanide, hydrochlorothiazide or ethacrynic acid
    Effect: increased urinary magnesium loss and depletion
    NIH Office of Dietary Supplements, Magnesium - Health Professional Fact Sheet · checked 2026-09-16 · we read the full text
21. mg-14 · Position of an expert body
    Prescription proton pump inhibitor (PPI) drugs, such as esomeprazole magnesium (Nexium) and lansoprazole (Prevacid), can cause hypomagnesemia when taken for prolonged periods (typically more than a year).
    
    Who: people on prescription proton pump inhibitors for prolonged periods
    Effect: hypomagnesaemia after typically more than a year
    NIH Office of Dietary Supplements, Magnesium - Health Professional Fact Sheet · checked 2026-09-16 · we read the full text
22. mg-s08 · Meta-analysis or systematic review · systematic review and meta-analysis of randomised trials · n = 325
    Our findings from the meta-analysis showed a significant decline in depression scores due to intervention with magnesium supplements [standardized mean difference (SMD): -0.919, 95% CI: -1.443 to -0.396, p = 0.001].
    
    Who: adults with a diagnosed depressive disorder, average ages 20 to 60, 7 trials
    Effect: depression score standardised mean difference -0.919 (95% CI -1.443 to -0.396)
    Certainty: small trials and a wide interval; the authors call for larger high-quality trials before using this clinically
    Moabedi et al., Frontiers in Psychiatry, 2023 · checked 2026-09-17 · we read the abstract
23. mg-s09 · Meta-analysis or systematic review · systematic review and meta-analysis · n = 325
    Future high-quality RCTs with larger sample sizes must be run to interpret this effect of magnesium on depression in clinical settings.
    
    Who: adults with a depressive disorder, 7 trials
    Effect: authors call for high-quality trials with larger samples before interpreting the effect in clinical settings
    Certainty: stated limitation of the review
    Moabedi et al., Frontiers in Psychiatry, 2023 · checked 2026-09-17 · we read the abstract
24. mg-r2-1 · Meta-analysis or systematic review · pooled cross-sectional statistical re-analysis of individual balance data from 27 tightly controlled metabolic ward studies, random coefficient models · n = 243 subjects, 664 balance data points
    The models predicted neutral magnesium balance [defined as magnesium output (Y) equal to magnesium intake (M)] at magnesium intakes of 165 mg/d [95% prediction interval (PI): 113, 237 mg/d; Y = 19.8 + 0.880 M]...Neither age nor sex affected the relation between magnesium intake and output.
    
    Who: healthy adults: 150 women averaging 51 years and 93 men averaging 28 years, on controlled diets of at least 28 days with magnesium intakes between 84 and 598 mg a day
    Effect: neutral magnesium balance predicted at 165 mg a day (95% prediction interval 113 to 237 mg/d), equivalently 2.36 mg per kg per day; neither age nor sex changed the intake-to-output relation
    Certainty: this is the average requirement at which intake equals losses, not a recommended intake, and balance is a crude endpoint that says nothing about long-term health; the authors state plainly that it implies a lower requirement than previously estimated
    Hunt & Johnson, American Journal of Clinical Nutrition, 2006 · checked 2026-09-17 · we read the abstract
25. mg-r2-2 · Observational data · synthesis of data from controlled metabolic ward balance and depletion-repletion studies conducted since 1997 · n = not pooled statistically; drawn from the metabolic unit study series
    Balance data obtained since 1997 indicate that the EAR and RDA for 70-kg healthy individuals are about 175 and 250 mg/day, respectively, and these DRIs decrease or increase based on body weight.
    
    Who: healthy adults of about 70 kg; requirements scaled up or down with body weight
    Effect: estimated average requirement about 175 mg a day and recommended dietary allowance about 250 mg a day for a 70 kg person, against the official US and Canadian allowances of 310 to 420 mg
    Certainty: a narrative synthesis by the investigators who produced much of the underlying data, not a systematic review and not an independent body; it is one reading of the evidence, and it disagrees with the official one
    Nielsen & Johnson, Biological Trace Element Research, 2017 · checked 2026-09-17 · we read the abstract
26. mg-r2-4 · Randomised controlled trial(s) · analysis within a randomised, controlled, three-arm crossover bioavailability study, linear mixed-effects model adjusted for age, gender, race and body mass index · n = 23 participants, 105 paired measurements
    Values for dietary magnesium, iMg2+ and s-Mg were 303.8 ± 118.9 mg/day, 1.3 ± 0.1 mg/dL and 2.2 ± 4.1 mg/dL, respectively. No association was found between dietary magnesium intake and iMg2+ -125 ± 176.95 (p = .49) or s-Mg -9.33 ± 5.04 (p = .08).
    
    Who: adults in a pilot supplement bioavailability trial; intake measured by three-day food records with the Nutrition Data System for Research
    Effect: mean dietary magnesium 303.8 ± 118.9 mg a day; no association with whole blood ionised magnesium (p = 0.49) or with serum magnesium (p = 0.08)
    Certainty: a pilot study of 23 people, and it tests only short-term self-reported intake, which is itself imprecise; it does not prove blood magnesium is useless, only that it does not follow recent diet
    Ansu Baidoo et al., Annals of Medicine, 2023 · checked 2026-09-17 · we read the abstract
27. mg-r6-01 · Meta-analysis or systematic review · Cochrane systematic review and meta-analysis of randomised controlled trials, pooling adverse events across all settings · n = 735 individuals in 11 trials; 254 participants in the 4 trials that reported minor adverse events, and 140 in the major-event analysis
    Major adverse events (occurring in 2 out of 72 magnesium recipients and 3 out of 68 placebo recipients), and withdrawals due to adverse events, were not significantly different from placebo. However, in the four studies for which it could be determined, more participants experienced minor adverse events in the magnesium group than in the placebo group (RR 1.51, 95% CI 0.98 to 2.33; 4 studies, 254 participants; low-certainty evidence).
    
    Who: people taking oral magnesium for skeletal muscle cramps: pregnant women, older adults and people with idiopathic cramps
    Effect: minor adverse events risk ratio 1.51 (95% CI 0.98 to 2.33); major adverse events in 2 of 72 magnesium recipients against 3 of 68 on placebo, no significant difference, and withdrawals for adverse events also no different
    Certainty: low-certainty evidence for the minor events, and the confidence interval touches 1.0; only four of eleven trials reported minor events at all
    Garrison et al., Cochrane Database of Systematic Reviews, 2020 · checked 2026-09-17 · we read the abstract
28. mg-r6-04 · Randomised controlled trial(s) · randomised, double-blind, placebo-controlled trial with four arms, magnesium citrate, oxide, sulfate or placebo, for 24 weeks · n = 164 participants
    Magnesium citrate increased plasma (+0.04 mmol/L; 95% CI, +0.02 to +0.06 mmol/L) and urine magnesium (+3.12 mmol/24 h; 95% CI, +2.23 to +4.01 mmol/24 h) compared with placebo. ... One serious adverse event was reported, which was considered unrelated to the study treatment.
    
    Who: overweight and slightly obese adults, mean age 63.2 years, 63.4 per cent women
    Effect: magnesium citrate raised plasma magnesium by 0.04 mmol/L (95% CI 0.02 to 0.06) and 24-hour urinary magnesium by 3.12 mmol (95% CI 2.23 to 4.01) against placebo; one serious adverse event was reported and judged unrelated to treatment
    Certainty: 24 weeks in people with normal kidney function, so it says nothing about kidney disease; the trial was designed for arterial stiffness, which it did not change, and safety was a secondary observation
    Zhang et al., Journal of the American Heart Association, 2022 · checked 2026-09-17 · we read the abstract
29. mg-e1-07 · Randomised controlled trial(s) · two-day crossover study with two stable isotopes labelling the two regimens, absorption by faecal monitoring and retention by urinary isotope excretion · n = 12 men
    Higher Mg absorption (50·7 (SD 12·7) v. 32·4 (SD 8·1) %; P = 0·0007) and retention (47·5 (SD 12·9) v. 29·0 (SD 7·5) %; P = 0·0008) from Mg-rich mineral water were observed when it was consumed in seven servings compared with larger servings.
    
    Who: 12 healthy men drinking 1.5 litres of magnesium-rich natural mineral water, 126 mg of magnesium in both arms
    Effect: absorption 50.7% (SD 12.7) with seven servings of 212 ml against 32.4% (SD 8.1) with two servings of 750 ml (P = 0.0007); retention 47.5% against 29.0% (P = 0.0008)
    Certainty: 12 healthy young men and one water, so the size of the gain is uncertain; the paper is described as crossover but the abstract does not say the order was randomised, which is why this is not graded higher. Full text not available to us (not in PMC)
    Br J Nutr, 2011 · checked 2026-09-18 · we read the abstract
30. mg-e1-08 · Randomised controlled trial(s) · crossover clinical trial on a standardised diet, 450 mg supplemental magnesium from each preparation, fractional urine collection and six blood samples per test day · n = 13 men
    With standardized conditions, urinary magnesium excretion increased by 40% after ingestion of the effervescent tablets, and by only 20% after intake of the capsules.
    
    Who: 13 healthy male volunteers aged 22-31
    Effect: urinary magnesium excretion rose 40% after effervescent tablets and 20% after capsules of the same magnesium oxide dose; the authors attribute the gap to the tablet being dissolved and ionised before swallowing
    Certainty: urinary excretion is a proxy for absorption, not absorption itself; 13 young healthy men, no placebo arm, and the abstract does not state that the order was randomised. It compares two products of the same salt, so it says nothing about oxide against citrate or glycinate. Full text not available to us (not in PMC)
    Urological Research, 2011 · checked 2026-09-18 · we read the abstract
31. mg-e1-15 · Position of an expert body · perspective article summarising the properties of the salt forms on sale · n = narrative summary, no pooled estimate
    Mg gluconate and Mg chloride have been preferred for oral replacement because they cause diarrhea less often than other salt forms [47].
    
    Who: adults taking oral magnesium supplements
    Effect: magnesium gluconate and magnesium chloride cause diarrhoea less often than other salt forms; magnesium carbonate is avoided as too insoluble to absorb well; citrate is highly water-soluble and may affect stools more than oxide, which is poorly soluble even in acid
    Certainty: a narrative summary with single citations behind each claim, not a head-to-head comparison; our own corpus has no trial comparing oxide, citrate and glycinate directly, which is why this narrative is the best answer available and is graded E
    Costello & Rosanoff, Advances in Nutrition, 2023 · checked 2026-09-18 · we read the fulltext
32. mg-e1-16 · Position of an expert body · perspective article summarising practical factors affecting tolerance · n = narrative summary
    As indicated above, taking a high amount of Mg in divided doses instead of in a single dose and taking supplements with food can mitigate the chances of gastrointestinal symptoms [40].
    
    Who: adults taking high oral magnesium doses
    Effect: taking a high amount of magnesium in divided doses rather than one, and taking it with food, can reduce the chance of gastrointestinal symptoms; the authors also note that in one of the source trials some participants reported symptoms easing when the supplement was taken with food
    Certainty: advice drawn from observations inside trials rather than from a trial designed to test it; it points the same way as the mineral-water finding that the same magnesium split into seven servings was absorbed at 50.7% against 32.4% in two
    Costello & Rosanoff, Advances in Nutrition, 2023 · checked 2026-09-18 · we read the fulltext