Benefits
Blood pressure
A 2025 meta-analysis of 38 randomized trials (2,709 adults; median 365 mg a day for 12 weeks) found magnesium lowered blood pressure by 2.81/2.05 mmHg versus placebo. Drops were larger in people with high blood pressure on medication (7.68 mmHg systolic) and in people with low blood magnesium (5.97 mmHg), and not significant in people with normal blood pressure. Results varied widely between trials.
Fasting glucose and insulin sensitivity
A meta-analysis of 25 double-blind placebo-controlled trials of oral magnesium (oxide most often) found lower fasting glucose in people with diabetes, and lower fasting glucose, 2-hour glucose and HOMA-IR in people at high risk of diabetes. HbA1c and insulin did not change, and results varied between trials. An earlier review found fasting glucose improved only in trials of 4 months or more.
Sleep onset latency in older adults
A pooled analysis of 3 randomized trials in 151 older adults with insomnia found magnesium shortened the time to fall asleep by 17.36 minutes versus placebo. Total sleep time was 16 minutes longer, which was not significant. All three trials had moderate to high risk of bias and the review rated the evidence low to very low quality, so this is limited evidence.
Sleep quality with magnesium L-threonate
In a 21-day placebo-controlled trial in 80 adults aged 35 to 55 with self-reported sleep problems, 1 g a day of magnesium L-threonate improved ring-measured deep and REM sleep scores and self-rated morning alertness and mood versus placebo. Sleep held steady on magnesium while it worsened on placebo. It is a single trial of one branded form, funded by the ingredient supplier.
Anxiety and mood scores
In an open-label trial in 112 adults with mild to moderate depression, magnesium chloride (248 mg a day) for 6 weeks lowered depression scores by 6.0 points and anxiety scores by 4.5 points compared with no treatment. Nobody was blinded and both scores were self-reported. In a magnesium L-threonate trial, effects on anxiety could not be judged because placebo also improved. This is preliminary evidence.
Cognitive scores with magnesium L-threonate
In a 12-week placebo-controlled trial in 44 adults aged 50 to 70 with cognitive impairment, a magnesium L-threonate product (MMFS-01, 1.5 to 2 g a day) improved an overall cognitive score versus placebo (p=0.003). Effects on sleep and anxiety could not be judged because placebo also improved. The developer funded this one small trial.
Headache frequency
In a 12-week placebo-controlled trial in 81 adults with migraine, 600 mg a day of magnesium (trimagnesium dicitrate) cut attack frequency by 41.6%, versus 15.8% on placebo, but attack length and intensity did not differ significantly and 18.6% had diarrhea. The dose is above the 350 mg upper limit for supplements, so use it only with a doctor's advice.
Hip bone mineral gain in girls with low magnesium intake
In a one-year placebo-controlled pilot trial in 50 girls aged 8 to 14 whose diets were low in magnesium (44 finished), 300 mg a day of magnesium raised hip bone mineral content more than placebo (P = 0.05). Spine gains and bone markers did not differ. Adults were not studied.
Muscle soreness after hard exercise
Two small placebo-controlled trials gave magnesium before hard exercise: 350 mg a day for 10 days in 22 young adults, and 500 mg a day (above the 350 mg upper limit) for 7 days in 9 runners on low magnesium diets. Soreness eased in both, but muscle force and performance did not clearly improve. A pooled analysis of 14 trials found no significant gain in strength or power.
Dietary magnesium in long-term cohort studies
A meta-analysis of 13 cohort studies (536,318 people) found that those eating the most magnesium were about 22% less likely to develop type 2 diabetes. This observational link concerns diet, not supplements, and magnesium-rich diets are also high in whole grains, nuts and vegetables.
Mechanism of action
Cofactor for hundreds of enzymes
Magnesium is a cofactor in more than 300 enzyme systems (one major review counts over 600 reactions), including those for energy production and protein synthesis. ATP, the cell's energy molecule, works mainly bound to magnesium.
Calcium channel and blood vessel effects
Magnesium regulates calcium movement into smooth muscle cells. This is the proposed explanation for the blood pressure effect seen in trials. It is a mechanism, not proof of an effect in any one person.
NMDA receptor regulation
Magnesium blocks NMDA receptor channels in nerve cells at rest, limiting excess glutamate signaling. This is the usual explanation offered for effects on sleep, anxiety and headache, but the human results for anxiety are open-label or could not be determined.
Vitamin D metabolism
Magnesium is needed by the enzymes that convert vitamin D into its active forms, so low magnesium status can limit how well vitamin D works.
Insulin signaling and glucose metabolism
Magnesium is needed for insulin receptor signaling and for enzymes in glucose metabolism. Low magnesium status is linked with insulin resistance, which is the proposed reason supplements lowered fasting glucose and HOMA-IR in trials.
Clinical trials
Systematic review and meta-analysis of randomized controlled trials of oral magnesium lasting at least 4 weeks, with a dose-response analysis (Argeros et al. 2025, Hypertension).
2,709 participants in 38 trials; doses 82.3 to 637 mg/day (median 365 mg), median 12 weeks.
Systolic blood pressure fell 2.81 mmHg and diastolic 2.05 mmHg versus placebo. Falls were larger in people with high blood pressure on medication (-7.68/-2.96 mmHg) and in people with low blood magnesium (-5.97/-4.75 mmHg), and not significant in people with normal blood pressure. There was no dose-response relationship, and heterogeneity was high.
Systematic review and meta-analysis of double-blind randomized placebo-controlled trials of oral magnesium in people with diabetes or at high risk of diabetes (Veronese et al. 2021, Nutrients).
25 trials; magnesium oxide was the most used form.
Fasting glucose was lower than placebo in people with diabetes. In people at high risk, fasting glucose, 2-hour glucose tolerance and HOMA-IR improved. HbA1c and insulin did not change.
Systematic review and meta-analysis of randomized trials comparing oral magnesium with placebo in older adults with insomnia, graded with GRADE (Mah et al. 2021, BMC Complement Med Ther).
151 older adults with insomnia in 3 trials from three countries.
Time to fall asleep was 17.36 minutes shorter than placebo (95% CI -27.27 to -7.44). Total sleep time was 16.06 minutes longer, which was not significant. All trials were at moderate to high risk of bias, and the evidence was rated low to very low quality.
Randomized, double-blind, placebo-controlled parallel trial of magnesium L-threonate (Magtein) 1 g/day for 21 days, measured with sleep questionnaires and an Oura ring, funded by the ingredient supplier AIDP (Hausenblas et al. 2024, Sleep Med X).
80 adults aged 35 to 55 with self-reported sleep problems.
Deep and REM sleep scores, activity and readiness scores, and self-rated behavior on waking, energy, mood and alertness improved versus placebo. The magnesium group maintained its sleep quality while the placebo group declined. No safety concerns were reported.
Randomized, double-blind, placebo-controlled trial of MMFS-01, a magnesium L-threonate product, at 1.5 to 2 g/day for 12 weeks, funded by its developer Neurocentria (Liu et al. 2016, J Alzheimers Dis).
44 adults aged 50 to 70 with cognitive impairment (23 active, 21 placebo).
The overall cognitive composite score improved versus placebo (p=0.003; Cohen's d 0.91). Effects on sleep and anxiety could not be determined because of strong placebo effects. A single small trial in people with cognitive impairment.
Prospective, multicenter, randomized, double-blind, placebo-controlled trial of trimagnesium dicitrate 600 mg (24 mmol)/day for 12 weeks after a 4-week baseline (Peikert et al. 1996, Cephalalgia).
81 adults aged 18 to 65 with migraine (mean 3.6 attacks a month).
In weeks 9 to 12, attack frequency fell 41.6% with magnesium versus 15.8% with placebo (p<0.05); migraine days and pain medicine use also fell. Attack duration and intensity did not differ significantly. Diarrhea affected 18.6% and stomach irritation 4.7%. The dose exceeds the 350 mg upper limit for supplements.
Open-label, blocked, randomized crossover trial: 6 weeks of magnesium chloride (248 mg elemental magnesium/day) versus 6 weeks of no treatment (Tarleton et al. 2017, PLoS One).
126 adults with mild to moderate depression enrolled; 112 provided analyzable data.
Depression scores (PHQ-9) improved by a net 6.0 points and anxiety scores (GAD-7) by 4.5 points versus no treatment. Nobody was blinded, the control was no treatment rather than placebo, and both outcomes were self-reported.