Supplement encyclopedia

Magnesium

Magnesium is an essential mineral obtained from food. Supplements and medicines use magnesium salts and other compounds. The weight of the compound is distinct from the amount of elemental magnesium.

Double-check the conclusions and review the sources before choosing a supplement.

Main conclusion

Potential benefits

  • Magnesium bisglycinate with glycine may slightly reduce insomnia complaints. This was observed after four weeks of use in adults with poor sleep; no clear reduction in daytime sleepiness was found.
  • Magnesium L-threonate Magtein may help with mental tasks of this kind. In a study, adults aged 18–45 who were dissatisfied with their sleep had better overall test results after six weeks than those taking placebo.
  • Magnesium oxide may reduce muscle soreness after downhill running. This was observed after one week of use in a small study of male runners with low dietary magnesium intake.
  • Magnesium chloride did not help improve sports performance in a short study: in adults who exercised, maximal oxygen uptake and cycling sprint power were lower than with placebo.
  • In adults aged 50–85 with frequent leg cramps, the studied supplement did not convincingly reduce the number of cramps over four weeks.
  • Magnesium citrate may reduce the frequency of migraine attacks: in one study of adults, attacks became less frequent over 12 weeks. Another trial of a different magnesium preparation found no benefit.
  • Magnesium oxide may relieve chronic constipation: improvement was reported more often than with placebo over four weeks. Here it was used as a laxative at a corresponding dose.

[1] [2] [3] [4] [5] [6] [7] [8]

What was studied

  • Magnesium has been studied for sleep, mental tasks, soreness after running, sports performance, migraine and constipation. Your choice can be guided by the specific goal.
  • Studies included different people, from adults who exercised to people with sleep complaints, cramps or migraine. Compare your situation with the group in which the supplement was studied.

[1] [2] [3] [4] [5] [6] [7] [8]

Forms studied

  • For sleep, researchers studied a preparation that provided 250 mg of elemental magnesium and 1523 mg of glycine daily. Check the ingredients: the result was obtained with this combination.
  • For mental tests, researchers used L-threonate Magtein: 2 g of the compound contained 145 mg of elemental magnesium. Compare the amount of magnesium itself on the label.
  • For migraine, citrate providing 600 mg of elemental magnesium per day was studied; for constipation, 1.5 g of magnesium oxide per day. In the second case, the figure is the mass of the compound.

[1] [2] [3] [4] [5]

How to get magnesium from food

  • A 30 g portion of shelled pumpkin seeds roasted without salt contains about 165 mg of magnesium and 172 kcal. You can add them to everyday meals.
  • A 90 g portion of cooked black beans without salt contains about 63 mg of magnesium, while 85 g of boiled spinach, drained and without salt, contains about 74 mg. These foods help spread magnesium intake across meals.
  • A 40 g portion of dry, unfortified rolled oats contains about 55 mg of magnesium. Cooking increases the portion’s weight because of the added water.
  • The upper intake level for magnesium from supplements and medicines does not apply to the magnesium naturally present in ordinary food. These sources are counted separately.

[1] [2] [3] [4] [5]

Effects and studied questions

How to read the certainty scale 1–4

Certainty is rated separately for each outcome. 1 is very low, 2 is low, 3 is moderate and 4 is high.

Goal or questionResultConfidenceStudiesPopulation studiedKey sources
Insomnia severity (ISI)The ISI score decreased by an additional 1.6 points. The rounded confidence interval includes zero difference. The effect of glycine alone was not studied.
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.
11 publication describedMagnesium Bisglycinate Supplementation in Healthy Adults Reporting Poor Sleep: A Randomized, Placebo-Controlled Trial ↗155 adults aged 18–65 with poor sleep
  1. Magnesium Bisglycinate Supplementation in Healthy Adults Reporting Poor Sleep: A Randomized, Placebo-Controlled Trial ↗
Sleep quality (SQS)No convincing additional difference in SQS scores was found. A small effect cannot be ruled out.
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.
11 publication describedMagnesium Bisglycinate Supplementation in Healthy Adults Reporting Poor Sleep: A Randomized, Placebo-Controlled Trial ↗The same adults with poor sleep
  1. Magnesium Bisglycinate Supplementation in Healthy Adults Reporting Poor Sleep: A Randomized, Placebo-Controlled Trial ↗
Daytime sleepiness (ESS)No convincing difference in daytime sleepiness was found.
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.
11 publication describedMagnesium Bisglycinate Supplementation in Healthy Adults Reporting Poor Sleep: A Randomized, Placebo-Controlled Trial ↗The same adults with poor sleep
  1. Magnesium Bisglycinate Supplementation in Healthy Adults Reporting Poor Sleep: A Randomized, Placebo-Controlled Trial ↗

The study count refers to the cited source. Reviews with overlapping trials and outcomes from one study are not added together.

Daily intake requirements and upper limits

GroupAmountSource
0–6 months · men · US30 mg/dayAITotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
0–6 months · women · US30 mg/dayAITotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
7–12 months · men · US75 mg/dayAITotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
7–12 months · women · US75 mg/dayAITotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
1–3 years · men · US80 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
1–3 years · women · US80 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
4–8 years · men · US130 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
4–8 years · women · US130 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
9–13 years · men · US240 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
9–13 years · women · US240 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
14–18 years · men · US410 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
14–18 years · women · US360 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
14–18 years · pregnancy · US400 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
14–18 years · breastfeeding · US360 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
19–30 years · men · US400 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
19–30 years · women · US310 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
19–30 years · pregnancy · US350 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
19–30 years · breastfeeding · US310 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
31–50 years · men · US420 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
31–50 years · women · US320 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
31–50 years · pregnancy · US360 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
31–50 years · breastfeeding · US320 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
51+ years · men · US420 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
51+ years · women · US320 mg/dayRDATotal intake from food, beverages, supplements, and medicationsFNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
7–11 months · men · EU80 mg/dayAITotal dietary intake. The value for pregnancy and breastfeeding is the same as for women aged 18 and older.EFSA NDA Panel 2015Scientific Opinion on Dietary Reference Values for magnesium ↗
7–11 months · women · EU80 mg/dayAITotal dietary intake. The value for pregnancy and breastfeeding is the same as for women aged 18 and older.EFSA NDA Panel 2015Scientific Opinion on Dietary Reference Values for magnesium ↗
1–3 years · men · EU170 mg/dayAITotal dietary intake. The value for pregnancy and breastfeeding is the same as for women aged 18 and older.EFSA NDA Panel 2015Scientific Opinion on Dietary Reference Values for magnesium ↗
1–3 years · women · EU170 mg/dayAITotal dietary intake. The value for pregnancy and breastfeeding is the same as for women aged 18 and older.EFSA NDA Panel 2015Scientific Opinion on Dietary Reference Values for magnesium ↗
3–10 years · men · EU230 mg/dayAITotal dietary intake. The value for pregnancy and breastfeeding is the same as for women aged 18 and older.EFSA NDA Panel 2015Scientific Opinion on Dietary Reference Values for magnesium ↗
3–10 years · women · EU230 mg/dayAITotal dietary intake. The value for pregnancy and breastfeeding is the same as for women aged 18 and older.EFSA NDA Panel 2015Scientific Opinion on Dietary Reference Values for magnesium ↗
10–18 years · men · EU300 mg/dayAITotal dietary intake. The value for pregnancy and breastfeeding is the same as for women aged 18 and older.EFSA NDA Panel 2015Scientific Opinion on Dietary Reference Values for magnesium ↗
10–18 years · women · EU250 mg/dayAITotal dietary intake. The value for pregnancy and breastfeeding is the same as for women aged 18 and older.EFSA NDA Panel 2015Scientific Opinion on Dietary Reference Values for magnesium ↗
18+ years · men · EU350 mg/dayAITotal dietary intake. The value for pregnancy and breastfeeding is the same as for women aged 18 and older.EFSA NDA Panel 2015Scientific Opinion on Dietary Reference Values for magnesium ↗
18+ years · women · EU300 mg/dayAITotal dietary intake. The value for pregnancy and breastfeeding is the same as for women aged 18 and older.EFSA NDA Panel 2015Scientific Opinion on Dietary Reference Values for magnesium ↗
0–12 months · USnot establishedOnly elemental magnesium from supplements and medications counts. Naturally occurring magnesium in food and beverages is excluded.FNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
1–3 years · US65 mg/dayOnly elemental magnesium from supplements and medications counts. Naturally occurring magnesium in food and beverages is excluded.FNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
4–8 years · US110 mg/dayOnly elemental magnesium from supplements and medications counts. Naturally occurring magnesium in food and beverages is excluded.FNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
9–18 years · US350 mg/dayOnly elemental magnesium from supplements and medications counts. Naturally occurring magnesium in food and beverages is excluded.FNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
19+ years · US350 mg/dayOnly elemental magnesium from supplements and medications counts. Naturally occurring magnesium in food and beverages is excluded.FNB/NASEM via NIH ODSMagnesium - Health Professional Fact Sheet ↗
1–3 years · EUnot definedNo upper limit has been established for this age groupSCF 2001, repeated in EFSA 2025 summaryOverview on Tolerable Upper Intake Levels as derived by the Scientific Committee on Food (SCF) and the EFSA Panel on Dietetic Products, Nutrition and Allergies (NDA) ↗
Ages 4 and older, including during pregnancy and lactation · EU250 mg/dayReadily dissociable salts and compounds, including magnesium oxide, from supplements, water, and fortified foods. Naturally occurring magnesium in food and beverages is excluded.SCF 2001, repeated in EFSA 2025 summaryOverview on Tolerable Upper Intake Levels as derived by the Scientific Committee on Food (SCF) and the EFSA Panel on Dietetic Products, Nutrition and Allergies (NDA) ↗

Forms

Bisglycinate with glycine: In the sleep study, two capsules provided 250 mg of elemental magnesium and 1,523 mg of glycine per day. The trial did not assess the effect of glycine separately. Citrate: The study of migraine frequency used 600 mg of elemental magnesium per day. The dose refers to the magnesium content alone. Oxide: The chronic constipation study tested 1.5 g of magnesium oxide per day, measured by compound mass, as a laxative. Other magnesium oxide trials used different doses, and some did not specify the basis for the reported amount. Magtein L-threonate: In the cognition trial, the daily dose of 2 g of the compound contained 145 mg of elemental magnesium. Chloride: The study of stress in people with fibromyalgia used magnesium chloride that provided 100 mg of elemental magnesium per day. Some exercise trials reported the mass of the preparation or salt, which cannot be treated as the same amount of magnesium. Hydroxide: In the kidney stone recurrence study, the doses of 650 and 1,300 mg/day refer to the mass of the magnesium hydroxide compound. Both doses were studied in the same trial.

Guidance for each goal

Insomnia severity (ISI)2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

The ISI score decreased by an additional 1.6 points. The rounded confidence interval includes zero difference. The effect of glycine alone was not studied.

Study findings. The ISI score decreased by an additional 1.6 points. The rounded confidence interval includes zero difference. The effect of glycine alone was not studied.

Studied regimen250 mg of elemental magnesium and 1,523 mg of glycine per day, taken as two capsulesMagnesium bisglycinate complex with glycine
Duration and population28 days155 adults aged 18–65 with poor sleep

Comparison: placebo or control under comparable conditions

Sleep quality (SQS)2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

No convincing additional difference in SQS scores was found. A small effect cannot be ruled out.

Study findings. No convincing additional difference in SQS scores was found. A small effect cannot be ruled out.

Studied regimen250 mg of elemental magnesium and 1,523 mg of glycine per day, taken as two capsulesMagnesium bisglycinate complex with glycine
Duration and population28 daysThe same adults with poor sleep

Comparison: placebo or control under comparable conditions

Daytime sleepiness (ESS)2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

No convincing difference in daytime sleepiness was found.

Study findings. No convincing difference in daytime sleepiness was found.

Studied regimen250 mg of elemental magnesium and 1,523 mg of glycine per day, taken as two capsulesMagnesium bisglycinate complex with glycine
Duration and population28 daysThe same adults with poor sleep

Comparison: placebo or control under comparable conditions

Migraine attack frequency1/4 Very lowRead moreCollapse
Confidence
1/4 Very lowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: High heterogeneity or inconsistency among individual controlled trials. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

In the citrate trial, attack frequency fell by 41.6% from baseline versus 15.8% with placebo. A separate trial of a different magnesium preparation did not detect an advantage.

Study findings. In the citrate trial, attack frequency fell by 41.6% from baseline versus 15.8% with placebo. A separate trial of a different magnesium preparation did not detect an advantage.

Studied regimen600 mg of elemental Mg per daymagnesium citrate
Duration and population12 weeks81 adults with migraine

Comparison: Placebo

Systolic blood pressure2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: High heterogeneity or inconsistency among individual controlled trials. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: The limited conclusion does not warrant a separate downgrade. There was no further downgrade in confidence on this basis. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

Across the mixed groups, systolic blood pressure was 2.81 mmHg lower on average. No convincing difference was found in participants with normal baseline blood pressure.

Study findings. Across the mixed groups, systolic blood pressure was 2.81 mmHg lower on average. No convincing difference was found in participants with normal baseline blood pressure.

Studied regimen82.3–637 mg of elemental Mg per day, median 365 mgvarious magnesium salts
Duration and population4–24 weeks, median 12 weeks2,709 adults in 38 RCTs with varying baseline blood pressure

Comparison: Placebo

Diastolic blood pressure2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: High heterogeneity or inconsistency among individual controlled trials. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: The limited conclusion does not warrant a separate downgrade. There was no further downgrade in confidence on this basis. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

Across the mixed groups, diastolic blood pressure was 2.05 mmHg lower on average. No convincing difference was found in participants with normal baseline blood pressure.

Study findings. Across the mixed groups, diastolic blood pressure was 2.05 mmHg lower on average. No convincing difference was found in participants with normal baseline blood pressure.

Studied regimen82.3–637 mg of elemental Mg per day, median 365 mgvarious magnesium salts
Duration and population4–24 weeks, median 12 weeks2,709 adults in 38 RCTs with varying baseline blood pressure

Comparison: Placebo

Weekly frequency of leg cramps2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

No convincing difference was found in the number of leg cramps per week. Cramps at any time of day were counted.

Study findings. No convincing difference was found in the number of leg cramps per week. Cramps at any time of day were counted.

Studied regimen250 mg of elemental magnesium per day according to the article. The exact salt remains disputed.the article describes the magnesium salt ambiguously
Duration and population4 weeks of treatmentAdults aged 50–85 with leg cramps at least twice a week. Cramps at any time of day were counted.

Comparison: Placebo

Reduction in cramp frequency by at least half1/4 Very lowRead moreCollapse
Confidence
1/4 Very lowRisk of bias: Some participants dropped out. The counts and percentages in the abstract need to be reconciled with the full text. Consistency across studies: According to the Cochrane review, trials of cramps during pregnancy produced conflicting results. Applicability to the stated group: The evidence applies directly only to pregnant women with the stated symptoms. There was no further downgrade in confidence on this basis. Precision of the estimate: One small study. The precision of the estimate and the safety data are limited. Potential publication bias: No specific publication bias was identified. There was no further downgrade in confidence on this basis.

In a small study of pregnant women with leg cramps, a reduction in cramp frequency by at least half was more common. Evidence from other studies is mixed, so this result should be interpreted cautiously.

Study findings. In a small study of pregnant women with leg cramps, a reduction in cramp frequency by at least half was more common. Evidence from other studies is mixed, so this result should be interpreted cautiously.

Studied regimen300 mg per day according to the publication. The abstract does not clarify whether this refers to the weight of elemental magnesium or of the compound.Magnesium bisglycinate chelate
Duration and populationFour weeks86 healthy pregnant women at 14–34 weeks of gestation with leg cramps at least twice a week. 80 completed the study.

Comparison: Placebo

Relief of chronic constipation2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

A response was observed in 32/47 versus 9/47. This is a laxative dose of a specific salt and differs from a typical nutritional dose of magnesium.

Study findings. A response was observed in 32/47 versus 9/47. This is a laxative dose of a specific salt and differs from a typical nutritional dose of magnesium.

Studied regimen1.5 g of MgO per day, measured as compound weightmagnesium oxide as a laxative
Duration and population4 weeks94 adults with chronic constipation in two RCTs

Comparison: placebo or control under comparable conditions

Fasting glucose1/4 Very lowRead moreCollapse
Confidence
1/4 Very lowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: High heterogeneity or inconsistency among individual controlled trials. Applicability to the stated group: The reported pooled estimate explicitly includes different groups and magnesium salts. It cannot predict the effect in a healthy person or the effect of any single salt. There was no further downgrade in confidence on this basis. Precision of the estimate: Small average effects and pooled intervals across different groups do not provide a reliable basis for expecting a meaningful change in any one person. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

Across the mixed groups, fasting glucose was 3.60 mg/dL lower on average. The outcome was fasting glucose. Diabetes incidence was not assessed by this measure.

Study findings. Across the mixed groups, fasting glucose was 3.60 mg/dL lower on average. The outcome was fasting glucose. Diabetes incidence was not assessed by this measure.

Studied regimenvarious dosesvarious forms of magnesium
Duration and populationvarying durationsVarious groups of adults, including people with diabetes and prediabetes

Comparison: placebo or control under comparable conditions

Glycated hemoglobin (HbA1c)1/4 Very lowRead moreCollapse
Confidence
1/4 Very lowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: High heterogeneity or inconsistency among individual controlled trials. Applicability to the stated group: The reported pooled estimate explicitly includes different groups and magnesium salts. It cannot predict the effect in a healthy person or the effect of any single salt. There was no further downgrade in confidence on this basis. Precision of the estimate: Small average effects and pooled intervals across different groups do not provide a reliable basis for expecting a meaningful change in any one person. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

The average change in HbA1c in the review was small. A separate one-year trial in adults with type 2 diabetes found no convincing difference versus placebo.

Study findings. The average change in HbA1c in the review was small. A separate one-year trial in adults with type 2 diabetes found no convincing difference versus placebo.

Studied regimenvarious dosesvarious forms of magnesium
Duration and populationvarying durationsVarious groups of adults, including people with diabetes

Comparison: placebo or control under comparable conditions

Fasting insulin1/4 Very lowRead moreCollapse
Confidence
1/4 Very lowRisk of bias: Concerns include reporting, dropout, disclosure of treatment assignment, or selection. Consistency across studies: Results vary across studies by participant group, intervention, or outcome, or individual studies contradict one another. Applicability to the stated group: The evidence applies directly only to the stated group, formulation, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: The number of participants is small, the interval is wide, or comparative data are incomplete. Potential publication bias: No specific indication of publication bias was identified. There was no further downgrade in confidence on this basis.

The review found no convincing overall change in fasting insulin. HOMA-IR is a separate measure.

Study findings. The review found no convincing overall change in fasting insulin. HOMA-IR is a separate measure.

Studied regimenVarious doses, with at least 300 mg per day in most RCTsVarious magnesium salts
Duration and populationVarying durations, with most RCTs lasting at least 12 weeksAdults with varying baseline health status and cardiometabolic risk

Comparison: Placebo

Depressive symptoms (final HDRS score)2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

Final HDRS scores were 10.7 versus 10.4, with no convincing difference at the endpoint. The scores reflect endpoint values and cannot be read as a difference in change from baseline.

Study findings. Final HDRS scores were 10.7 versus 10.4, with no convincing difference at the endpoint. The scores reflect endpoint values and cannot be read as a difference in change from baseline.

Studied regimen40 mg of elemental Mg three times per dayMagnesium aspartate
Duration and population8 weeks37 adults with depression taking fluoxetine, with 32 assessed at week eight

Comparison: placebo with the same fluoxetine regimen

Perceived stress (PSS)2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The stated question specifically concerns people with poor sleep and their PSS stress or PHQ anxiety scores. No general conclusion about treating anxiety follows from the scope of that question. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

The PSS comparison showed no convincing difference. Participants were not selected based on stress.

Study findings. The PSS comparison showed no convincing difference. Participants were not selected based on stress.

Studied regimen250 mg of elemental magnesium and 1,523 mg of glycine per day, taken as two capsulesMagnesium bisglycinate complex with glycine
Duration and population28 daysAdults with poor sleep

Comparison: placebo or control under comparable conditions

Anxiety (PHQ-4 subscale)2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The stated question specifically concerns people with poor sleep and their PSS stress or PHQ anxiety scores. No general conclusion about treating anxiety follows from the scope of that question. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

No difference was detected on the PHQ-4 anxiety subscale. Small effects cannot be ruled out.

Study findings. No difference was detected on the PHQ-4 anxiety subscale. Small effects cannot be ruled out.

Studied regimen250 mg of elemental magnesium and 1,523 mg of glycine per day, taken as two capsulesMagnesium bisglycinate complex with glycine
Duration and population28 daysAdults with poor sleep

Comparison: placebo or control under comparable conditions

Stress (DASS-42)2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

Mean scores on day 28 were 21.8 versus 21.6, with no convincing overall benefit.

Study findings. Mean scores on day 28 were 21.8 versus 21.6, with no convincing overall benefit.

Studied regimen100 mg of elemental Mg per daymagnesium chloride
Duration and population28 daysPatients with fibromyalgia and stress

Comparison: placebo or control under comparable conditions

NIH composite cognitive test score2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

Mean scores increased by 8.40 versus 5.60 points, a difference in change of +2.80 points. The link with daily functioning was not established.

Study findings. Mean scores increased by 8.40 versus 5.60 points, a difference in change of +2.80 points. The link with daily functioning was not established.

Studied regimen2 g of the compound, 145 mg of elemental Mg/dayMagtein magnesium L-threonate
Duration and population6 weeks100 adults aged 18–45 who were dissatisfied with their sleep

Comparison: placebo or control under comparable conditions

Soreness after downhill running1/4 Very lowRead moreCollapse
Confidence
1/4 Very lowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the specified group. The group of interest is the nine male runners with low magnesium intake, and the outcome is soreness. There was no further downgrade in certainty for this domain. Precision of the estimate. The crossover study had only nine participants and provided no reliable interval for the paired effect. The direction and magnitude of the effect are therefore very uncertain. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

A small crossover study reported reduced soreness. Evidence on tolerability and functional recovery is limited.

Study findings. A small crossover study reported reduced soreness. Evidence on tolerability and functional recovery is limited.

Studied regimen500 mg of elemental magnesium per day in three capsules, as calculated by the authorsMagnesium oxide. The capsules also contained magnesium stearate and microcrystalline cellulose.
Duration and population7 daysNine male runners were prescribed a diet containing less than 260 mg of magnesium per day during the study

Comparison: Corn flour capsules

Maximal oxygen uptake (VO₂max)1/4 Very lowRead moreCollapse
Confidence
1/4 Very lowRisk of bias. Of the 22 participants enrolled, only 15 completed the oxygen uptake test and 13 completed the sprint. Substantial missing outcome data and analysis limited to completers raise serious concerns about bias. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

Maximal oxygen uptake was lower with magnesium (41.3 versus 44.4 mL/kg/min, p=0.005).

Study findings. Maximal oxygen uptake was lower with magnesium (41.3 versus 44.4 mL/kg/min, p=0.005).

Studied regimen300 mg of magnesium chloride twice daily as reported in the publication. The amount of elemental magnesium was not established.magnesium chloride
Duration and population9 days per regimen, with a 3-week break between regimens15 young adult completers who were training

Comparison: Placebo

Mean sprint power1/4 Very lowRead moreCollapse
Confidence
1/4 Very lowRisk of bias. Of the 22 participants enrolled, only 15 completed the oxygen uptake test and 13 completed the sprint. Substantial missing outcome data and analysis limited to completers raise serious concerns about bias. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

Mean power during the 30-second cycling sprint was lower with magnesium (415 versus 439 W, p=0.03).

Study findings. Mean power during the 30-second cycling sprint was lower with magnesium (415 versus 439 W, p=0.03).

Studied regimen300 mg of magnesium chloride twice daily as reported in the publication. The amount of elemental magnesium was not established.magnesium chloride
Duration and population9 days per regimen, with a 3-week break between regimens13 of the 22 participants enrolled completed the sprint test

Comparison: Placebo

Combined bone mineral content across three hip sites2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Concerns include reporting, dropout, disclosure of treatment assignment, or selection. Consistency across studies. This outcome comes from a single double-blind, one-year trial, so consistency across studies could not be assessed. There was no further downgrade in certainty for this domain. Applicability to the stated group: The evidence applies directly only to the stated group, formulation, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: The number of participants is small, the interval is wide, or comparative data are incomplete. Potential publication bias: No specific indication of publication bias was identified. There was no further downgrade in confidence on this basis.

Over one year, combined bone mineral content across three hip sites increased by about 1.05 g with magnesium and 0.97 g with placebo. The result was borderline (p=0.0534 in the table).

Study findings. Over one year, combined bone mineral content across three hip sites increased by about 1.05 g with magnesium and 0.97 g with placebo. The result was borderline (p=0.0534 in the table).

Studied regimen300 mg of elemental magnesium per dayMagnesium oxide
Duration and populationOne year50 girls aged 8–14 with low magnesium intake, of whom 44 completed the study

Comparison: Placebo

Total hip bone mineral content2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Concerns include reporting, dropout, disclosure of treatment assignment, or selection. Consistency across studies. This outcome comes from a single double-blind, one-year trial, so consistency across studies could not be assessed. There was no further downgrade in certainty for this domain. Applicability to the stated group: The evidence applies directly only to the stated group, formulation, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: The number of participants is small, the interval is wide, or comparative data are incomplete. Potential publication bias: No specific indication of publication bias was identified. There was no further downgrade in confidence on this basis.

Over one year, total hip bone mineral content increased by about 1.33 g with magnesium and 1.29 g with placebo. No convincing difference between groups was found (p=0.577).

Study findings. Over one year, total hip bone mineral content increased by about 1.33 g with magnesium and 1.29 g with placebo. No convincing difference between groups was found (p=0.577).

Studied regimen300 mg of elemental magnesium per dayMagnesium oxide
Duration and populationOne year50 girls aged 8–14 with low magnesium intake, of whom 44 completed the study

Comparison: Placebo

Quality of life measured with the SF-36 questionnaire2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

The overall SF-36 quality-of-life score was higher with magnesium than with placebo in this group of women with PCOS (p<0.001).

Study findings. The overall SF-36 quality-of-life score was higher with magnesium than with placebo in this group of women with PCOS (p<0.001).

Studied regimen250 mg of magnesium oxide per day as reported in the publication. The amount of elemental magnesium was not reported separately.magnesium oxide
Duration and population10 weeks64 women with PCOS from a fertility clinic were randomized, and 54 completed the study

Comparison: Placebo containing 5 mg of starch

Acne2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

No convincing difference between groups was found. Small effects cannot be ruled out.

Study findings. No convincing difference between groups was found. Small effects cannot be ruled out.

Studied regimen250 mg of magnesium oxide per day as reported in the publication. The amount of elemental magnesium was not reported separately.magnesium oxide
Duration and population10 weeks64 women with PCOS from a fertility clinic were randomized, and 54 completed the study

Comparison: Placebo containing 5 mg of starch

Alopecia2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

No convincing difference between groups was found. Small effects cannot be ruled out.

Study findings. No convincing difference between groups was found. Small effects cannot be ruled out.

Studied regimen250 mg of magnesium oxide per day as reported in the publication. The amount of elemental magnesium was not reported separately.magnesium oxide
Duration and population10 weeks64 women with PCOS from a fertility clinic were randomized, and 54 completed the study

Comparison: Placebo containing 5 mg of starch

Abnormal uterine bleeding2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

No convincing difference between groups was found. Small effects cannot be ruled out.

Study findings. No convincing difference between groups was found. Small effects cannot be ruled out.

Studied regimen250 mg of magnesium oxide per day as reported in the publication. The amount of elemental magnesium was not reported separately.magnesium oxide
Duration and population10 weeks64 women with PCOS from a fertility clinic were randomized, and 54 completed the study

Comparison: Placebo containing 5 mg of starch

Restless legs symptoms (IRLS)1/4 Very lowRead moreCollapse
Confidence
1/4 Very lowRisk of bias. Unblinded researchers assessed subjective outcomes, and the handling of missing data and the pairwise comparison were insufficiently clear. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the specified group. The stated group includes people aged 15–50 taking the supplement alongside pramipexole. There is no separate estimate for adults alone. There was no further downgrade in certainty for this domain. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

Mean scores were 16.08 and 24.24. The published p-value applies to the comparison across all three groups. The adjusted pairwise effect and adult-only results were not established.

Study findings. Mean scores were 16.08 and 24.24. The published p-value applies to the comparison across all three groups. The adjusted pairwise effect and adult-only results were not established.

Studied regimen250 mg of magnesium oxide per day, measured as compound weight. The amount of elemental magnesium was not specified.magnesium oxide
Duration and population2 months75 people aged 15–50 with restless legs syndrome who were taking pramipexole

Comparison: placebo with the same pramipexole regimen

Short-term pain on a 0–100 scale2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

Across two trials, the mean difference was −0.55 points on a 0–100 pain scale (95% CI −8.32 to 7.21). No convincing reduction in pain was found.

Study findings. Across two trials, the mean difference was −0.55 points on a 0–100 pain scale (95% CI −8.32 to 7.21). No convincing reduction in pain was found.

Studied regimenvarious dosesoral magnesium in various salt forms
Duration and populationshort term118 adults with various types of chronic noncancer pain across two randomized controlled trials

Comparison: placebo or control under comparable conditions

Kidney stone recurrence2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

No convincing reduction in recurrence was established. The interval is compatible with both benefit and harm.

Study findings. No convincing reduction in recurrence was established. The interval is compatible with both benefit and harm.

Studied regimen650 mg of magnesium hydroxide per day, measured as compound weightmagnesium hydroxide
Duration and population3 yearsPatients with recurrent calcium oxalate stones

Comparison: Matching placebo in the same trial

Kidney stone recurrence2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

The second dose also showed no clear benefit. It was evaluated in the same trial.

Study findings. The second dose also showed no clear benefit. It was evaluated in the same trial.

Studied regimen1,300 mg of magnesium hydroxide per day, measured as compound weightmagnesium hydroxide
Duration and population3 yearsThe same population in the same trial

Comparison: Matching placebo in the same trial

Total cholesterol2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias. Most randomized controlled trials in the review had a low risk of bias, and no serious risk was identified for total cholesterol specifically. The problematic LDL/HDL data cells in the Sobhani study were outside the total cholesterol analysis. There was no further downgrade in certainty for this domain. Consistency across studies. Results varied substantially (I² = 65.7%). The exact source of the variation was not established. Applicability to the specified group. The outcome is total cholesterol pooled across different magnesium salts and patient populations. This outcome cannot establish whether clinical events are prevented. There was no further downgrade in certainty for this domain. Precision of the estimate. The interval from −5.65 to +2.35 mg/dL is compatible with both a decrease and an increase. No claim of equivalence is made. Possible publication bias. The review found insufficient grounds for a separate downgrade for this outcome. Unpublished results cannot be ruled out. There was no further downgrade in certainty for this domain.

No convincing difference in total cholesterol was found across the mixed populations. The interval is compatible with both a decrease and an increase.

Study findings. No convincing difference in total cholesterol was found across the mixed populations. The interval is compatible with both a decrease and an increase.

Studied regimenvarious doses across the primary trials, with no single dose established for the subset of studies reporting lipid outcomesvarious salts and forms, reported individually in the primary studies
Duration and populationvaried durations. The review as a whole covered 3–52 weeks, and the distribution of durations for this outcome was not established separately.Various adult populations in the meta-analysis (34 estimates), including healthy adults and people with cardiometabolic diseases

Comparison: placebo or control as specified in each study

Triglycerides2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias. No serious risk was identified for this outcome specifically. The problematic LDL/HDL data cells in the Sobhani study were outside the triglyceride analysis. Most randomized controlled trials in the review had a low risk of bias. There was no further downgrade in certainty for this domain. Consistency across studies. Results varied substantially (I² = 76%). The exact source of the variation was not established. Applicability to the specified group. The outcome is a pooled estimate for triglycerides. It does not guarantee a result for any individual salt or the prevention of clinical events. There was no further downgrade in certainty for this domain. Precision of the estimate. The pooled interval from −16.87 to −1.61 supports a limited conclusion of a small average change. Differences across studies were accounted for separately in the consistency assessment. There was no further downgrade in certainty for this domain. Possible publication bias. The authors reported possible asymmetry or small-study bias for this outcome.

In the 2026 review, triglycerides were lower by an average of 9.24 mg/dL (95% CI −16.87 to −1.61), but trial results varied substantially. The 2025 review found no convincing difference.

Study findings. In the 2026 review, triglycerides were lower by an average of 9.24 mg/dL (95% CI −16.87 to −1.61), but trial results varied substantially. The 2025 review found no convincing difference.

Studied regimenvarious doses across the primary trials, with no single dose established for the subset of studies reporting lipid outcomesvarious salts and forms, reported individually in the primary studies
Duration and populationvaried durations. The review as a whole covered 3–52 weeks, and the distribution of durations for this outcome was not established separately.Various adult populations in the meta-analysis (41 estimates), including healthy adults and people with cardiometabolic diseases

Comparison: placebo or control as specified in each study

Altitude sickness during rapid ascent2/4 LowRead moreCollapse
Confidence
2/4 LowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the stated group: The conclusion is limited to the stated group, dose, and outcome. There was no further downgrade in confidence on this basis. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

Altitude sickness symptoms were absent in 20% of participants taking magnesium and 16.1% taking placebo. No convincing benefit was found. Loose stools were more common with magnesium (38.2% versus 11.8%).

Study findings. Altitude sickness symptoms were absent in 20% of participants taking magnesium and 16.1% taking placebo. No convincing benefit was found. Loose stools were more common with magnesium (38.2% versus 11.8%).

Studied regimen400 mg of magnesium as citrate (16 mmol according to the article) every 8 hoursOral magnesium citrate
Duration and population5 days70 adults undergoing rapid ascent were randomized, and 61 were included in the efficacy analysis

Comparison: Matching placebo

Menstrual pain1/4 Very lowRead moreCollapse
Confidence
1/4 Very lowRisk of bias. The way the small trial was reported and the conflicting salt names in the primary publications make it difficult to attribute the result to a specific product. The exact effect size is unreliable. Consistency across studies: No serious unexplained heterogeneity was identified for this outcome. There was no further downgrade in confidence on this basis. Applicability to the specified group. The outcome of interest is menstrual pain. Findings from another condition do not apply to this outcome. There was no further downgrade in certainty for this domain. Precision of the estimate: A small sample and a wide or unreported interval leave substantial uncertainty. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

The authors reported reduced pain. The publications describe a similar protocol but list different salts, so these data cannot establish the effect of a specific form.

Study findings. The authors reported reduced pain. The publications describe a similar protocol but list different salts, so these data cannot establish the effect of a specific form.

Studied regimen150 or 300 mg per day as reported in the publications. The amount of elemental magnesium was not established.The publications list different salts: magnesium stearate and magnesium sulfate
Duration and populationTwo cycles, from day 15 of each cycle until the pain resolved in the following cycle60 female students with moderate or severe primary menstrual pain

Comparison: Placebo, with ibuprofen allowed as needed

Body weight1/4 Very lowRead moreCollapse
Confidence
1/4 Very lowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: High heterogeneity or inconsistency among individual controlled trials. Applicability to the stated group: The reported pooled estimate explicitly includes different groups and magnesium salts. It cannot predict the effect in a healthy person or the effect of any single salt. There was no further downgrade in confidence on this basis. Precision of the estimate: Small average effects and pooled intervals across different groups do not provide a reliable basis for expecting a meaningful change in any one person. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

Body weight was lower by an average of 0.70 kg (95% CI −1.30 to −0.09). The estimate is sensitive to individual studies and applies to mixed populations.

Study findings. Body weight was lower by an average of 0.70 kg (95% CI −1.30 to −0.09). The estimate is sensitive to individual studies and applies to mixed populations.

Studied regimenvarious dosesvarious forms of magnesium
Duration and populationvarying durationsAdults across a mix of clinical and other populations

Comparison: placebo or control under comparable conditions

Body fat percentage1/4 Very lowRead moreCollapse
Confidence
1/4 Very lowRisk of bias: Participant attrition, incomplete original study plans, or limited assessment of risk of bias in individual trials. Consistency across studies: High heterogeneity or inconsistency among individual controlled trials. Applicability to the stated group: The reported pooled estimate explicitly includes different groups and magnesium salts. It cannot predict the effect in a healthy person or the effect of any single salt. There was no further downgrade in confidence on this basis. Precision of the estimate: Small average effects and pooled intervals across different groups do not provide a reliable basis for expecting a meaningful change in any one person. Potential publication bias: No specific grounds for a separate downgrade were identified. Unpublished data may exist. There was no further downgrade in confidence on this basis.

No convincing difference was found. The mean change relative to control was −0.02 percentage points (95% CI −0.46 to 0.43).

Study findings. No convincing difference was found. The mean change relative to control was −0.02 percentage points (95% CI −0.46 to 0.43).

Studied regimenvarious dosesvarious forms of magnesium
Duration and populationvarying durationsAdults in mixed populations

Comparison: placebo or control under comparable conditions

Safety and limitations

Check the label for the amount of elemental magnesium. In the US, the upper limit for adults is 350 mg/day from supplements and medications. In the EU, the limit is 250 mg/day for readily dissociable salts and compounds such as magnesium oxide from supplements, water, and fortified foods. Naturally occurring magnesium in food is excluded from these limits. Magnesium supplements can cause diarrhea, nausea, and abdominal cramps. When kidney function is impaired, the body clears magnesium less efficiently, and it can build up. Also count magnesium from laxatives and antacids.

Magnesium - Health Professional Fact Sheet ↗ · Overview on Tolerable Upper Intake Levels as derived by the Scientific Committee on Food (SCF) and the EFSA Panel on Dietetic Products, Nutrition and Allergies (NDA) ↗

Interactions

WithEffectActionSource
Tetracyclines and fluoroquinolonesMagnesium can reduce the absorption of antibiotics such as doxycycline or ciprofloxacin.Take the antibiotic at least 2 hours before magnesium or 4–6 hours after it. Check the instructions for your specific medication to confirm the timing.Magnesium - Health Professional Fact Sheet ↗
Oral bisphosphonatesMagnesium can reduce the absorption of medications such as alendronate.Take magnesium and the bisphosphonate at least 2 hours apart and follow the medication instructions.Magnesium - Health Professional Fact Sheet ↗

Food sources

Food composition

FoodGroupAmountServingCaloriesSources
Pumpkin and zucchini seeds, hulled, roasted without salthulled, roasted without saltFood550 mg / 100 ghulled, roasted without salt30 g · 165 mg · 172.2 kcal574 kcal / 100 gUSDA National Nutrient Database for Standard Reference, Release 28 — Nut and Seed Products ↗
Reference Amounts Customarily Consumed: List of Products for Each Product Category: Guidance for Industry ↗
Chia seeds, drieddriedFood335 mg / 100 gdried30 g · 100.5 mg · 145.8 kcal486 kcal / 100 gUSDA National Nutrient Database for Standard Reference, Release 28 — Nut and Seed Products ↗
Reference Amounts Customarily Consumed: List of Products for Each Product Category: Guidance for Industry ↗
Almonds, dry-roasted without saltdry-roasted without saltFood279 mg / 100 gdry-roasted without salt30 g · 83.7 mg · 179.4 kcal598 kcal / 100 gUSDA National Nutrient Database for Standard Reference, Release 28 — Nut and Seed Products ↗
Reference Amounts Customarily Consumed: List of Products for Each Product Category: Guidance for Industry ↗
Cashews, dry-roasted without saltdry-roasted without saltFood260 mg / 100 gdry-roasted without salt30 g · 78 mg · 172.2 kcal574 kcal / 100 gUSDA National Nutrient Database for Standard Reference, Release 28 — Nut and Seed Products ↗
Reference Amounts Customarily Consumed: List of Products for Each Product Category: Guidance for Industry ↗
Spinach, boiled, drained, without saltboiled, drained, without saltFood87 mg / 100 gboiled, drained, without salt85 g · 73.95 mg · 19.55 kcal23 kcal / 100 gUSDA National Nutrient Database for Standard Reference, Release 28 — Vegetables and Vegetable Products ↗
Reference Amounts Customarily Consumed: List of Products for Each Product Category: Guidance for Industry ↗
Black beans, boiled without saltboiled without saltFood70 mg / 100 gboiled without salt90 g · 63 mg · 118.8 kcal132 kcal / 100 gUSDA National Nutrient Database for Standard Reference, Release 28 — Legumes and Legume Products ↗
Reference Amounts Customarily Consumed: List of Products for Each Product Category: Guidance for Industry ↗
Sunflower seeds, hulled, dry-roasted without salthulled, dry-roasted without saltFood129 mg / 100 ghulled, dry-roasted without salt30 g · 38.7 mg · 174.6 kcal582 kcal / 100 gUSDA National Nutrient Database for Standard Reference, Release 28 — Nut and Seed Products ↗
Reference Amounts Customarily Consumed: List of Products for Each Product Category: Guidance for Industry ↗
Long-grain brown rice, boiledboiledFood39 mg / 100 gboiled140 g · 54.6 mg · 172.2 kcal123 kcal / 100 gUSDA National Nutrient Database for Standard Reference, Release 28 — Cereal Grains and Pasta ↗
Reference Amounts Customarily Consumed: List of Products for Each Product Category: Guidance for Industry ↗
Potatoes, baked with skin, without saltbaked with skin, without saltFood28 mg / 100 gbaked with skin, without salt110 g · 30.8 mg · 102.3 kcal93 kcal / 100 gUSDA National Nutrient Database for Standard Reference, Release 28 — Vegetables and Vegetable Products ↗
Reference Amounts Customarily Consumed: List of Products for Each Product Category: Guidance for Industry ↗
Rolled oats, dry, unfortifieddry, unfortifiedFood138 mg / 100 gdry, unfortified40 g · 55.2 mg · 151.6 kcal379 kcal / 100 gUSDA National Nutrient Database for Standard Reference, Release 28 — Breakfast Cereals ↗
Reference Amounts Customarily Consumed: List of Products for Each Product Category: Guidance for Industry ↗

USDA National Nutrient Database for Standard Reference, Release 28 — Nut and Seed Products ↗

Studies

PublicationWhat was studiedParticipantsForm and doseResult
Magnesium Bisglycinate Supplementation in Healthy Adults Reporting Poor Sleep: A Randomized, Placebo-Controlled Trial ↗Insomnia severity: ISI: Insomnia severity: ISI Sleep quality: SQS: Sleep quality: SQS Daytime sleepiness: ESS: Daytime sleepiness: ESS Perceived stress: PSS: Perceived stress: PSS Anxiety: PHQ-4 subscale: Anxiety: PHQ-4 subscaleInsomnia severity: ISI: 155 adults aged 18–65 with poor sleep Sleep quality: SQS: The same adults with poor sleep Daytime sleepiness: ESS: The same adults with poor sleep Perceived stress: PSS: Adults with poor sleep Anxiety: PHQ-4 subscale: Adults with poor sleepInsomnia severity: ISI: Magnesium bisglycinate complex with glycine, providing 250 mg of elemental magnesium and 1,523 mg of glycine per day in two capsules for 28 days. Comparison: placebo or control under comparable conditions. Sleep quality: SQS: Magnesium bisglycinate complex with glycine, providing 250 mg of elemental magnesium and 1,523 mg of glycine per day in two capsules for 28 days. Comparison: placebo or control under comparable conditions. Daytime sleepiness: ESS: Magnesium bisglycinate complex with glycine, providing 250 mg of elemental magnesium and 1,523 mg of glycine per day in two capsules for 28 days. Comparison: placebo or control under comparable conditions. Perceived stress: PSS: Magnesium bisglycinate complex with glycine, providing 250 mg of elemental magnesium and 1,523 mg of glycine per day in two capsules for 28 days. Comparison: placebo or control under comparable conditions. Anxiety: PHQ-4 subscale: Magnesium bisglycinate complex with glycine, providing 250 mg of elemental magnesium and 1,523 mg of glycine per day in two capsules for 28 days. Comparison: placebo or control under comparable conditions.Insomnia severity: ISI: The ISI score decreased by an additional 1.6 points. The rounded confidence interval includes zero difference. The effect of glycine was not studied separately. Sleep quality: SQS: No convincing additional difference in SQS scores was found. A small effect cannot be ruled out. Daytime sleepiness: ESS: No convincing difference in daytime sleepiness was found. Perceived stress: PSS: The comparison showed no convincing difference in PSS scores. Participants were not selected based on stress levels. Anxiety: PHQ-4 subscale: No difference was found on the PHQ-4 anxiety subscale. Small effects cannot be ruled out.
Prophylaxis of migraine with oral magnesium: results from a prospective, multi-center, placebo-controlled and double-blind randomized study ↗Migraine attack frequency81 adults with migrainemagnesium citrate, providing 600 mg of elemental Mg per day for 12 weeks. Comparison: placebo or control under comparable conditions.Frequency decreased from baseline by 41.6% versus 15.8%. These percentages do not represent the proportion of patients who responded. Another RCT found no benefit.

Sources

Sources
  1. Magnesium Bisglycinate Supplementation in Healthy Adults Reporting Poor Sleep: A Randomized, Placebo-Controlled Trial ↗
  2. Prophylaxis of migraine with oral magnesium: results from a prospective, multi-center, placebo-controlled and double-blind randomized study ↗
  3. Magnesium Supplementation and Blood Pressure: A Systematic Review and Meta-Analysis of Randomized Controlled Trials ↗
  4. Secondary prevention of leg cramps using compression stockings or magnesium supplements: a three-arm randomized clinical trial ↗
  5. Oral magnesium for relief in pregnancy-induced leg cramps: a randomised controlled trial ↗
  6. Interventions for leg cramps during pregnancy | Cochrane ↗
  7. The effect of food, vitamin, or mineral supplements on chronic constipation in adults: A systematic review and meta-analysis of randomized controlled trials ↗
  8. Comprehensive Effects of Magnesium Supplementation on Cardiometabolic Risk Factors: A Systematic Review and Dose-Response Meta-Analysis ↗
  9. Effects of Magnesium Supplementation on Unipolar Depression: A Placebo-Controlled Study and Review of the Importance of Dosing and Magnesium Status in the Therapeutic Response ↗
  10. Short-Term Magnesium Therapy Alleviates Moderate Stress in Patients with Fibromyalgia: A Randomized Double-Blind Clinical Trial ↗
  11. The effects of magnesium L-threonate (Magtein®) on cognitive performance and sleep quality in adults: a randomised, double-blind, placebo-controlled trial ↗
  12. One week of magnesium supplementation lowers IL-6, muscle soreness and increases post-exercise blood glucose in response to downhill running ↗
  13. Short-Term Magnesium Supplementation Has Modest Detrimental Effects on Cycle Ergometer Exercise Performance and Skeletal Muscle Mitochondria and Negligible Effects on the Gut Microbiota: A Randomized Crossover Clinical Trial ↗
  14. A randomized controlled study of effects of dietary magnesium oxide supplementation on bone mineral content in healthy girls ↗
  15. The effects of magnesium supplementation on abnormal uterine bleeding, alopecia, quality of life, and acne in women with polycystic ovary syndrome: a randomized clinical trial ↗
  16. Therapeutic effects of magnesium and vitamin B6 in alleviating the symptoms of restless legs syndrome: a randomized controlled clinical trial ↗
  17. Ketamine and other NMDA receptor antagonists for chronic pain ↗
  18. Results. In: Recurrent Nephrolithiasis in Adults and Children: Comparative Effectiveness of Preventive Medical Strategies: Systematic Review ↗
  19. Chlorthalidone reduces calcium oxalate calculous recurrence but magnesium hydroxide does not ↗
  20. The effect of magnesium supplementation on serum concentration of lipid profile: an updated systematic review and dose-response meta-analysis on randomized controlled trials ↗
  21. Interventions for preventing high altitude illness: Part 2. Less commonly-used drugs ↗
  22. Effectiveness of Magnesium on Menstrual Symptoms Among Dysmenorrheal College Students: A Randomized Controlled Trial ↗
  23. Magnesium - Health Professional Fact Sheet ↗
  24. USDA National Nutrient Database for Standard Reference, Release 28 — Nut and Seed Products ↗
  25. Reference Amounts Customarily Consumed: List of Products for Each Product Category: Guidance for Industry ↗
  26. USDA National Nutrient Database for Standard Reference, Release 28 — Vegetables and Vegetable Products ↗
  27. USDA National Nutrient Database for Standard Reference, Release 28 — Legumes and Legume Products ↗
  28. USDA National Nutrient Database for Standard Reference, Release 28 — Cereal Grains and Pasta ↗
  29. USDA National Nutrient Database for Standard Reference, Release 28 — Breakfast Cereals ↗
  30. Scientific Opinion on Dietary Reference Values for magnesium ↗
  31. Overview on Tolerable Upper Intake Levels as derived by the Scientific Committee on Food (SCF) and the EFSA Panel on Dietetic Products, Nutrition and Allergies (NDA) ↗
  32. Magnesium for the prevention and treatment of acute mountain sickness ↗
  33. The Effect of Magnesium Sulfate on Pain Intensity and Menstrual Blood Loss in Students With Primary Dysmenorrhea: A Randomized Controlled Trial ↗
  34. Magnesium in the prophylaxis of migraine-a double-blind, placebo-controlled study ↗
  35. The effect of oral magnesium supplementation on glycemic control and metabolic parameters in type 2 diabetes mellitus: a double-blind randomized controlled trial ↗