Magnesium

What magnesium is

Magnesium is an essential mineral involved in more than 300 enzyme reactions. Food sources include pumpkin seeds, almonds, spinach, dark chocolate and avocado. As a food supplement it is sold in various salt forms, including glycinate, citrate, malate, threonate and the poorly absorbed oxide.

The most robust figures on intake in Germany come from the German National Nutrition Survey II by the Max Rubner Institute (collected 2005 to 2007): median intake in men and women is above the recommendation in almost all age groups, while 26 % of men and 29 % of women do not reach it. Being below the recommendation is not the same as a deficiency. The Lower Saxony Chamber of Pharmacists states that there is no general magnesium undersupply in Germany; people who regularly eat whole grains, legumes, nuts or seeds are usually adequately supplied. A deficiency can be promoted by, among other things, diseases of the digestive tract, diabetes, heavy alcohol consumption, diuretics and proton pump inhibitors.

In the BK-Score it gets 7 out of 10 points for human evidence and 10 out of 10 for track record of use. The track-record axis only measures how widely and how long something has been used in humans – not whether it works.

How it works

As a cofactor of hundreds of enzymes, magnesium is involved in, among other things, ATP production, muscle function and nerve signal transmission. This chain of action is well understood and scores 8 out of 10 points in the BK-Score.

The mechanism explains why magnesium is marketed for cramps and sleep. But it does not replace proof of efficacy: this is where plausibility and study results diverge.

What the studies show

Blood pressure: meta-analysis of 38 studies

Argeros 2025 pooled 38 randomized controlled trials with 2,709 participants and a duration of at least four weeks in Hypertension. Systolic blood pressure under magnesium was on average 2.81 mmHg lower than under placebo (−4.32 to −1.29), diastolic 2.05 mmHg (−3.23 to −0.88). The effect was larger in people with high blood pressure already on blood-pressure-lowering medication (systolic −7.68 mmHg) and in people with low blood magnesium levels (−5.97 mmHg). In people with normal blood pressure, the difference was not significant. The authors point to the high heterogeneity of the studies; they found no relationship between amount and effect.

Cochrane review: primary endpoint for calf cramps missed

Garrison 2020 analyzed 11 RCTs with 735 participants for Cochrane. The main analysis covered idiopathic, mostly nocturnal calf cramps in people aged 61.6 to 69.3 over four weeks. The primary endpoint, the percentage change in cramps per week, was not reached: mean difference −9.59 % (−23.14 to +3.97), moderate certainty of evidence. Cramps per week (−0.18; −0.84 to +0.49) and the proportion with at least a 25 % reduction (RR 1.04; 0.84 to 1.29, high certainty) also showed no effect. For leg cramps in pregnancy, the review describes the evidence as conflicting; all five trials on this were at high risk of bias.

New RCT: compression stockings worked, magnesium did not

Kuusipalo 2026 ran a three-arm multicenter randomized trial with 121 enrolled participants (109 completers) aged 50 to 85 with at least two cramps per week, with four weeks of use and the endpoint at week 8. The primary endpoint, cramp frequency, was not reached for magnesium: adjusted mean difference −0.20 per week (−1.49 to +1.09; p = 0.929). The third arm provides the contrast: compression stockings reduced the frequency by 1.43 per week (−2.36 to −0.50; p = 0.001).

Sleep: small effect, questionnaire only

Schuster 2025 studied, in a double-blind RCT, 155 people aged 18 to 65 with self-reported poor sleep who received 250 mg of elemental magnesium as bisglycinate for four weeks. The Insomnia Severity Index fell by 3.9 points (−5.8 to −2.0) compared with 2.3 under placebo (−4.1 to −0.4), p = 0.049, with an effect size of d = 0.2. It was measured solely by questionnaire. A review by Lopresti 2026 of 12 RCTs rates the evidence as low to very low and makes no routine recommendation for insomnia.

Salt forms: how much magnesium is in them and how much gets absorbed

How much magnesium a compound contains follows from its molar mass: magnesium oxide is about 60 % magnesium, anhydrous citrate about 16 %, bisglycinate about 14 %, Epsom salt (magnesium sulfate heptahydrate) about 10 % and L-threonate about 8 %. What counts is therefore the amount of elemental magnesium stated on the pack, not the weight of the compound. In terms of absorption, the oxide does worse. In Dallas it was virtually insoluble in water, citrate 55 % soluble, and urinary magnesium rose markedly more after citrate (Lindberg 1990). In New York, the absorbed fraction of the oxide was 4 %, while chloride, lactate and aspartate did better and equally well (Firoz 2001). In a double-blind study in Reading with 46 healthy volunteers over 60 days, citrate was the most bioavailable, and the oxide did not differ from placebo (Walker 2003). Each of these measured urine and blood values, not clinical endpoints. The Lower Saxony Chamber of Pharmacists concludes that the effect depends on the amount absorbed and can be achieved with different compounds.

Where the data stop

• Effect on exercise-associated cramps. The Cochrane review found zero randomized controlled trials on this. So there is no negative finding here, but no finding at all – the use is widely advertised nonetheless.
• Objectively better sleep. The significant finding from Schuster 2025 comes from a questionnaire. Neither actigraphy nor polysomnography was used, and the effect size of d = 0.2 is small.
• Superiority of individual salt forms. Glycinate, citrate, malate and threonate are marketed with clearly separated areas of use. Differences in absorption have been measured, mainly to the disadvantage of the oxide, via urine and blood values. Comparative studies with clinical endpoints supporting an assignment such as “glycinate for sleep, malate for energy” do not exist; in the BK-Score this costs points on the hype gap.
• Blood pressure without hypertension. In the meta-analysis by Argeros 2025, the effect in people with normal blood pressure was not significant, the studies differed widely, and no relationship between amount and effect was found.

Status, approval and legal

The German Federal Institute for Risk Assessment (BfR) proposes a maximum level of 250 mg per daily dose for food supplements. It is based on the value of 250 mg of additional intake that the Scientific Committee on Food (SCF) set as the tolerable upper limit in 2001. EFSA keeps this value unchanged in its 2024 overview of upper limits. It covers readily dissociable magnesium salts and magnesium oxide in food supplements, water or added to foods, not magnesium naturally present in food. It is a safety limit, not an intake recommendation.

Safety

Mild diarrhea occurs from about 360 to 365 mg of additional intake. In the Cochrane review, mild side effects were more frequent under magnesium (RR 1.51; 0.98 to 2.33), mainly diarrhea at up to 37 % compared with 14 % under placebo. Caution is advised in severe kidney failure. Magnesium can also weaken the effect of some antibiotics; anyone with heart or kidney disease or taking related medication should discuss additional magnesium with a pharmacist or doctor beforehand. This text is for information and does not replace medical advice.

Sources

• Garrison 2020, Cochrane Database of Systematic Reviews CD009402.pub3 – 11 RCTs, calf cramps (PMID 32956536)
• Kuusipalo 2026, Trials – three-arm multicenter RCT, magnesium and compression stockings (PMID 41680812)
• Schuster 2025, Nature and Science of Sleep – double-blind RCT, 155 participants, sleep (PMID 40918053)
• Lopresti 2026 – review of 12 RCTs on magnesium and sleep (PMID 42661485)
• BfR – Proposed maximum levels for magnesium in foods including food supplements
• Argeros 2025, Hypertension – meta-analysis, 38 RCTs, 2,709 participants, magnesium and blood pressure (PMID 41000008)
• Lindberg 1990, Journal of the American College of Nutrition – solubility and absorption of magnesium citrate and magnesium oxide (PMID 2407766)
• Firoz and Graber 2001, Magnesium Research – bioavailability of four magnesium preparations (PMID 11794633)
• Walker 2003, Magnesium Research – double-blind RCT, 46 healthy volunteers, citrate, amino acid chelate and oxide (PMID 14596323)
• Max Rubner Institute – German National Nutrition Survey II, results report part 2 (2008), magnesium intake in Germany (German)
• aponet.de / pta-Forum, June 8, 2026 – magnesium citrate or magnesium oxide, with statements from the Lower Saxony Chamber of Pharmacists (German)
• EFSA – Overview on Tolerable Upper Intake Levels, Summary Report 2024 (magnesium: 250 mg/day additional, SCF 2001)