Magnesium: Forms, Doses and What the Label Actually Means
TL;DR: Magnesium supplements differ more in form than in dose. The same 500 mg capsule can deliver 302 mg of actual magnesium as the oxide or 41 mg as L-threonate, and real L-threonate labels declare less again. Choose on elemental content and tolerance, then check the total against the 350 mg supplemental limit.
Magnesium is the ingredient where the label is most likely to mislead you, because almost every product is sold under the name of a compound rather than the mineral. A bottle marked “Magnesium 500 mg” and one marked “Magnesium Bisglycinate 500 mg” can differ fourfold in what they provide.
This page covers the forms and what each delivers, what named trials measured for each use magnesium is sold for, the intake figures, and how to read the bottle — plus what we found on 446 on-market US labels.
Quick facts
| What it is | |
|---|---|
| RDA, adults 19+ | 400 mg men 19–30, 420 mg men 31+, 310 mg women 19–30, 320 mg women 31+. This is an RDA rather than an AI, and it counts magnesium from every source (Linus Pauling Institute) |
| RDA in pregnancy | 400 mg at age 18 or under, 350 mg at 19–30, 360 mg at 31+; 310–360 mg while breastfeeding |
| Tolerable Upper Intake Level | 350 mg/day of supplemental magnesium from age 9 up, 110 mg at 4–8, 65 mg at 1–3. It is defined as the level “likely to pose no risk of diarrhea”, and it does not apply to magnesium in food |
| Forms sold in the US | 16 distinct magnesium salts appeared across the 446 on-market labels we surveyed. Ten are common enough that a shopper will meet them |
| Is deficiency common in the US? | Low intake is common; measured deficiency is a separate question. In NHANES 2017–March 2020, 57% of men and 52% of women consumed less than the Estimated Average Requirement, with mean intakes of 339 mg/day and 269 mg/day (Linus Pauling Institute) |
Forms, and what each actually delivers
Every row below is arithmetic from molecular weights, not a claim needing a citation. Check any of it with a periodic table.
| Form | Compound, molecular weight | Elemental magnesium | From a 500 mg capsule |
|---|---|---|---|
| Oxide | MgO, 40.30 | 60.3% | 302 mg |
| Carbonate | MgCO₃, 84.31 | 28.8% | 144 mg |
| Chloride | MgCl₂, 95.21 (anhydrous) | 25.5% | 128 mg |
| Sulfate | MgSO₄, 120.37 (anhydrous) | 20.2% | 101 mg |
| Citrate, trimagnesium | Mg₃(C₆H₅O₇)₂, 451.11 (anhydrous) | 16.2% | 81 mg |
| Malate | MgC₄H₄O₅, 156.38 | 15.5% | 78 mg |
| Bisglycinate | Mg(C₂H₄NO₂)₂, 172.42 | 14.1% | 70 mg |
| Citrate, monomagnesium | MgC₆H₆O₇, 214.41 | 11.3% | 57 mg |
| Taurate | Mg(C₂H₆NO₃S)₂, 272.59 | 8.9% | 45 mg |
| L-threonate | Mg(C₄H₇O₅)₂, 294.50 (anhydrous) | 8.3% theoretical, 7.2% as declared on US labels † | 41 mg theoretical, 36 mg at the declared figure |
† The L-threonate gap. The anhydrous molecular weight gives 8.3%. Real US labels declare less: of the 20 that state both the compound weight and the elemental magnesium, every one fell between 7.0% and 7.34%, clustering at 7.2%, and the published trials land in the same region — Ni 2023 calls 1.2 g “91.8 mg” (7.65%) and Lopresti and Smith 2026 call 2 g 145 mg (7.25%). The gap is what water of crystallisation would do, about two water molecules’ worth. No label we read says so, so that is our inference, not a stated fact. Use 7.2% on a real bottle; the full survey is in L-threonate versus glycinate.
Hydrated salts change the arithmetic, and labels rarely say which they used: chloride hexahydrate is 12.0% not 25.5%, sulfate heptahydrate (Epsom salt) 9.9% not 20.2%, trimagnesium citrate 11.9% or 10.4% hydrated against 16.2% anhydrous.
Higher elemental content is not automatically better. Oxide carries the most magnesium per milligram and is also the least soluble: virtually insoluble in water, and only 43% dissolved even under maximal gastric acid (Lindberg 1990). That trade-off is the reason multiple forms exist.
What each form is sold for, and what has been tested
| Form | Typical positioning | What has actually been measured |
|---|---|---|
| Oxide | Cheapest elemental magnesium; also sold as a laxative and antacid | Poorly soluble: only 43% dissolved even under maximal gastric acid conditions (Lindberg 1990) |
| Citrate | General supplementation, bowel preparation | Raised urinary magnesium significantly more than oxide in two human studies (Lindberg 1990; Werner 2019) |
| Bisglycinate | Marketed for tolerance and daily use | No randomized trial in healthy adults compares it with other forms. The one head-to-head trial, in 12 ileal-resection patients, found no significant overall difference versus oxide (Schuette 1994) |
| Chloride, lactate, aspartate | General supplementation | Higher fractional absorption than oxide in one comparison (Firoz & Graber 2001) |
| L-threonate | Marketed for cognition | Lowest elemental content of the common forms. Five human trials, four funded by sellers of the ingredient, none longer than 12 weeks |
| Malate, taurate | Marketed for energy, cardiovascular use | Little comparative human data on either positioning |
A caution for the whole table: most absorption comparisons test one or two forms against each other, in small groups, by urinary excretion rather than clinical outcomes. Ranking all ten by absorption would need evidence that does not exist.
Going deeper: glycinate vs oxide traces the widely quoted “4% vs 18.8%” figures to their sources. Glycinate vs citrate covers why “magnesium citrate” is not one compound. L-threonate vs glycinate follows the brain claim to its source.
What the US market actually stocks
Earlier surveys counted how much magnesium each product declares. This one asks which salts are in the bottle, how often they are mixed, and what the pill is, across 446 on-market records from 186 brands in the NIH Dietary Supplement Label Database.
Sixteen distinct magnesium salts appeared. 241 products (54.0%) name a single salt, 123 (27.6%) are blends of two to five magnesium compounds, and for the remaining 82 the record named no compound at all. The chart below gives the count for each salt.
Oxide usually arrives with company. Of the 112 products containing oxide, only 41 contain nothing else; 71 are blends. Read the other way, 57.7% of multi-form products contain oxide — the cheapest, least soluble salt — and citrate-plus-oxide is the commonest pairing by far, in 53 of the 123 blends. If you chose a form deliberately, the form list under the magnesium line on the Supplement Facts panel is the only place that tells you whether you got it.
Oxide is a tablet; citrate is everything.
| Delivery format | All 446 products | Of the 112 oxide products | Of the 202 citrate products |
|---|---|---|---|
| Capsule | 176 (39.5%) | 28 | 89 |
| Tablet or pill | 139 (31.2%) | 64 | 49 |
| Powder | 47 (10.5%) | 0 | 22 |
| Liquid | 32 (7.2%) | 0 | 17 |
| Softgel | 26 (5.8%) | 8 | 15 |
| Other, incl. gummies | 26 (5.8%) | 12 | 10 |
Not one oxide product is a powder or a liquid, which is what a dense, barely soluble compound gives you. Citrate appears in every row.
A third of records do not name the salt where you would look for it. The panel’s magnesium line named a compound for 300 of the 446 products (67.3%); for 64 (14.3%) only the product name gave it away, and for 82 (18.4%) no form appeared anywhere — including products sold as “Magnesium Complex”. That describes the database record, not the printed bottle, which we did not inspect.
What the evidence shows
Each use carries a grade — Strong, Limited or None — defined in our methodology. The right-hand column says what a named trial measured, not what magnesium does.
| Use | Grade | What a named trial or review measured |
|---|---|---|
| Raising serum magnesium | Strong | Across 19 randomized trials in type 2 diabetes, supplementation raised serum magnesium by a weighted mean difference of 0.69 (95% CI 0.32 to 1.06, P < 0.01) (Al Maqrashi 2025) |
| Relieving chronic constipation, as magnesium oxide | Strong | 34 women took 1.5 g/day of magnesium oxide or placebo for 28 days: 70.6% were responders against 25.0% on placebo (P = 0.015), spontaneous bowel movements rose 6.07 per week against 2.86 (P = 0.002), and whole-colon transit fell from 75.5 to 41.6 hours (Mori 2019). A second trial randomized 90 patients to senna, magnesium oxide 1.5 g or placebo: response was 68.3%, 69.2% and 11.7% (P < 0.0001) (Morishita 2021) |
| Lowering blood pressure | Limited, and only in one group | A categorized review of 49 trials found 240 mg/day or more lowered blood pressure in people whose hypertension was uncontrolled while on medication, and found no effect in people whose hypertension was controlled or whose pressure was normal, at doses up to 632 mg/day. Its authors state the main limitation is “the lack of quantification of the BP changes” (Rosanoff 2021). A pooled analysis of 34 double-blind trials at a median 368 mg/day reported 2.00 mmHg systolic and 1.78 mmHg diastolic (Zhang 2016) |
| Glycemic control in type 2 diabetes | Limited | 23 randomized trials, 1,345 participants. Fasting blood glucose fell (WMD −0.58, 95% CI −0.87 to −0.28, P < 0.01), but HbA1c did not reach significance (−0.16%, 95% CI −0.32 to 0.00, P = 0.05) and insulin, HOMA-IR and BMI showed no effect (Al Maqrashi 2025) |
| Falling asleep faster | Limited | Three randomized trials in 151 older adults, pooled: time to fall asleep was 17.36 minutes shorter than placebo (95% CI −27.27 to −7.44, P = 0.0006); total sleep time was not significantly different (+16.06 minutes, P = 0.15). Doses were 320–729 mg/day as oxide or citrate. The reviewers graded the certainty low and wrote that “the quality of literature is substandard” for firm recommendations (Mah & Pitre 2021) |
| Migraine prophylaxis | Limited | 81 patients took 600 mg/day of trimagnesium dicitrate or placebo for 12 weeks; attack frequency in weeks 9–12 fell 41.6% against 15.8% on placebo (P < 0.05), with diarrhea in 18.6% (Peikert 1996). A systematic review applying the American Academy of Neurology’s own evidence classification rated the body of trials Grade C, possibly effective (von Luckner & Riederer 2018) |
| Preventing leg cramps in older adults | None | Pooled across 5 trials and 307 participants, magnesium changed cramp frequency by −0.18 cramps per week (95% CI −0.84 to 0.49) at four weeks, on moderate-certainty evidence. The reviewers concluded it is “unlikely that magnesium supplementation provides clinically meaningful cramp prophylaxis to older adults” (Garrison 2020) |
| Cognition, from the L-threonate form | Limited, and the funding matters | Five randomized trials, 431 analysed participants, none longer than 12 weeks. Four were paid for by companies selling the ingredient. The one funded by a public research agency, in 109 patients, found no difference in persistent pain (P = 0.67) or cognition (Ni 2023) |
| Choosing one form over another for a health outcome | None | No trial has compared magnesium forms against a clinical outcome. The comparisons that exist measure urinary or plasma magnesium in small groups (Lindberg 1990; Schuette 1994; Werner 2019) |
Two things there are worth saying out loud. The two uses with the clearest placebo-controlled evidence are the two almost nobody advertises: raising a low blood level, and moving the bowels (Al Maqrashi 2025; Mori 2019). And the blood-pressure finding is not a general one: it appeared in people already on antihypertensive medication whose pressure had not come under control, not in people whose pressure was normal.
On funding: magnesium salts are commodities with no dominant supplier, with one exception. The L-threonate category rests on a single patented raw material, named by 28 of 32 labels in our earlier survey, whose owners funded four of its five human trials.
How much, and when
| Group | RDA — counts all sources | Upper limit — supplements only |
|---|---|---|
| Children 1–3 | 80 mg | 65 mg |
| Children 4–8 | 130 mg | 110 mg |
| Children 9–13 | 240 mg | 350 mg |
| Age 14–18 | 410 mg males, 360 mg females | 350 mg |
| Men 19–30 | 400 mg | 350 mg |
| Men 31+ | 420 mg | 350 mg |
| Women 19–30 | 310 mg | 350 mg |
| Women 31+ | 320 mg | 350 mg |
| Pregnancy 19–30 | 350 mg | 350 mg |
| Pregnancy 31+ | 360 mg | 350 mg |
Reference intakes as tabulated by the Linus Pauling Institute. No upper limit was set under 12 months, where intake is expected from food and formula only.
The upper limit being lower than the requirement confuses almost everyone who notices it. The two numbers count different things. The RDA counts magnesium from everything: food, water, supplements, medicines. The Tolerable Upper Intake Level of 350 mg counts only supplements and magnesium-containing medicines such as some antacids and laxatives. Food magnesium sits outside it, because healthy kidneys clear the excess and, per the Linus Pauling Institute, no adverse effects have been identified from magnesium occurring naturally in food. The limit marks the level “likely to pose no risk of diarrhea” — a tolerance threshold, not a toxicity one.
Two things follow. A magnesium-containing antacid or laxative counts toward the 350 mg, which people frequently do not realise. And in our earlier label surveys close to half of bisglycinate and oxide products declared more than 350 mg in one serving.
With food, split or not, and at what hour. Absorption is roughly 30–40% of what is swallowed and falls as the single dose rises, which is the usual argument for splitting it. That argument is an inference from physiology rather than a tested one: we found no trial randomising split dosing against a single daily dose, and none comparing morning against evening. The trials each simply picked a schedule, from once a day to three times a day after meals. Products marketed for sleep are taken at night because of the positioning, not because anyone tested the hour.
How to read the label
The front of the bottle names the compound; the Supplement Facts panel names the mineral. Only the second tells you what you are taking, and US labelling requires it to be there.
- “Magnesium (as magnesium citrate) 100 mg” — 100 mg of elemental magnesium. The useful number, nothing to convert.
- “Magnesium citrate 1,000 mg” on the front, with no elemental figure — this is the compound weight, and the honest answer is a range: 162 mg as anhydrous trimagnesium citrate, 119 mg as the nonahydrate, 113 mg as the monomagnesium salt, 104 mg as the tetradecahydrate. Not one of the 45 citrate labels in our citrate survey said which salt it used, so 1,000 mg narrows down only to somewhere between 104 and 162 mg.
- “Magnesium L-threonate 2,000 mg” — multiply by the declared 7.2%, not the theoretical 8.3%: about 144 mg.
- “% Daily Value” — against the 420 mg Daily Value, so 50% DV is about 210 mg of elemental magnesium.
- The form list beneath the magnesium line — 71 of the 123 multi-form products in our survey included oxide. If the form mattered to your choice, check here.
Where a label gives both numbers the ratio should land on the figure in the table above, but most do not: in our earlier surveys only 3 of 45 bisglycinate products and none of 45 oxide products gave both, against 22 of 32 L-threonate products.
Safety, interactions, and who should be careful
Loose stools are the first sign of too much. The Cochrane cramps review recorded gastrointestinal adverse events, mainly diarrhea, in 11% to 37% of participants on oral magnesium against 10% to 14% on control (Garrison 2020). The migraine trial at 600 mg/day reported diarrhea in 18.6% (Peikert 1996). The widely quoted “19 to 40%” range comes from the magnesium migraine trials specifically, not from general use (Linus Pauling Institute).
Reduced kidney function is the serious risk. Among patients on daily magnesium oxide for constipation, 38% of those with CKD stage 4–5 had elevated serum magnesium and 19% met criteria for hypermagnesemia (Mori H 2019 — a different paper from the constipation trial above). Nobody in that study was symptomatic, but reduced renal function and magnesium dose were each independently associated with hypermagnesemia. Do not supplement magnesium with impaired kidney function without medical supervision, whatever the form.
Interactions run in two directions, and the fix differs:
- Magnesium blocks the drug — separate the doses in time. Bisphosphonates: the FDA alendronate label says to wait at least 30 minutes before any other oral medicine, and the Linus Pauling Institute advises two hours. Quinolone and tetracycline antibiotics, nitrofurantoin, digoxin and penicillamine are also absorbed less well with magnesium. MedlinePlus gives two hours as the general rule for magnesium oxide.
- The drug depletes your magnesium — spacing does not help. Proton pump inhibitors: the Linus Pauling Institute records that three months or more of use may raise the risk of low blood magnesium, and the FDA’s 2011 Drug Safety Communication reported the same, mostly beyond a year. Loop diuretics such as furosemide, and some thiazides, impair renal magnesium reabsorption at long-term high doses.
Who should ask a clinician first, per the Linus Pauling Institute: anyone with reduced kidney function; anyone whose gut absorbs poorly, including Crohn’s disease, celiac disease, prolonged diarrhea or small-bowel resection; anyone on digoxin; anyone with heart, liver or intestinal disease; and anyone with surgery coming up, who should tell the medical team. The Institute also names alcohol use disorder, type 2 diabetes and age over 50 as situations where intake tends to be low. In pregnancy the RDA is 350–400 mg from all sources, 310–360 mg while breastfeeding, and supplemental use has not been studied in a way that would support deciding alone.
Duration limits on the drug products. MedlinePlus sets limits most supplement articles omit: do not use magnesium oxide as an antacid for longer than 2 weeks, or as a laxative for longer than 1 week, without medical direction.
Seek medical attention for severe or persistent diarrhea, muscle weakness, difficulty breathing or an irregular heartbeat while taking magnesium. In impaired kidney function, hypermagnesemia can progress from a fall in blood pressure to muscle weakness and, in severe cases, cardiac arrest.
Frequently asked questions
Which form of magnesium is best?
No form is best for everyone, because no trial has compared magnesium forms against a clinical outcome. Oxide gives the most magnesium per milligram and is the cheapest; citrate has the clearest human absorption evidence, measured against oxide in two independent trials; bisglycinate is the form people choose when tolerance is the binding constraint, on reasoning rather than on trial data.
How do I know how much magnesium I am actually taking?
Read the Supplement Facts panel, not the front of the bottle. US labelling requires elemental magnesium to be stated there, and that is the number to compare between products. If a label gives only the compound weight, multiply by the percentage in the table above — and note that for magnesium citrate the answer is a range, because labels do not say which citrate salt they used.
Can I take magnesium every day?
Many people do, within the 350 mg supplemental upper limit for adults. That limit counts supplements and magnesium-containing medicines but not food, so an antacid or laxative containing magnesium counts toward it. In our earlier label surveys close to half of bisglycinate and oxide products exceeded 350 mg in a single serving, so the arithmetic is worth doing before adding a second product.
Does magnesium help with sleep?
Three randomized trials in 151 older adults, pooled, found people taking magnesium fell asleep 17.36 minutes faster than those on placebo (95% CI −27.27 to −7.44), with no significant difference in total sleep time. The reviewers rated the certainty low and called the literature substandard for firm recommendations. The trials used 320–729 mg/day of oxide or citrate in adults over 55, so they say little about younger people or the forms marketed for sleep.
Does magnesium prevent leg cramps?
Not in older adults, on the evidence available. A Cochrane review pooling five trials and 307 participants found magnesium changed cramp frequency by −0.18 cramps per week against placebo (95% CI −0.84 to 0.49), and concluded that clinically meaningful cramp prophylaxis in older adults is unlikely. For pregnancy-associated cramps the reviewers judged the evidence too conflicting to draw a conclusion.
Who should not take magnesium supplements without asking a doctor?
Anyone with reduced kidney function, which is the single most important risk factor, because the kidneys are what clear excess magnesium. Also anyone on bisphosphonates, quinolone or tetracycline antibiotics, digoxin, proton pump inhibitors or diuretics; anyone whose gut absorbs poorly; anyone pregnant or breastfeeding; and anyone with surgery scheduled.
Sources
- Linus Pauling Institute, Oregon State University. Magnesium.
- NIH Office of Dietary Supplements. Magnesium — Health Professional Fact Sheet.
- Lindberg JS, et al. Magnesium bioavailability from magnesium citrate and magnesium oxide. J Am Coll Nutr. 1990;9(1):48-55. PMID 2407766
- Werner T, et al. Assessment of bioavailability of Mg from Mg citrate and Mg oxide. Magnes Res. 2019;32(3):63-71. PMID 32162607
- Schuette SA, et al. Bioavailability of magnesium diglycinate vs magnesium oxide in patients with ileal resection. JPEN. 1994;18(5):430-5. PMID 7815675
- Firoz M, Graber M. Bioavailability of US commercial magnesium preparations. Magnes Res. 2001;14(4):257-62. PMID 11794633
- Mori S, Tomita T, Fujimura K, et al. A randomized double-blind placebo-controlled trial on the effect of magnesium oxide in patients with chronic constipation. J Neurogastroenterol Motil. 2019;25(4):563-75. PMC6786451
- Morishita D, Tomita T, Mori S, et al. Senna versus magnesium oxide for the treatment of chronic constipation: a randomized, placebo-controlled trial. Am J Gastroenterol. 2021;116(1):152-61. PMID 32969946
- Mori H, et al. Clinical features of hypermagnesemia in patients with functional constipation taking daily magnesium oxide. J Clin Biochem Nutr. 2019;65(1):76-81. PMC6667383
- Rosanoff A, Costello RB, Johnson GH. Effectively prescribing oral magnesium therapy for hypertension: a categorized systematic review of 49 clinical trials. Nutrients. 2021;13(1):195. PMC7827637
- Zhang X, Li Y, Del Gobbo LC, et al. Effects of magnesium supplementation on blood pressure: a meta-analysis of randomized double-blind placebo-controlled trials. Hypertension. 2016;68(2):324-33. PMID 27402922
- Al Maqrashi N, Al Busaidi S, Al-Rasbi S, et al. Effect of magnesium supplements on improving glucose control, blood pressure and lipid profile in patients with type 2 diabetes mellitus: a systematic review and meta-analysis. Sultan Qaboos Univ Med J. 2025;25(1):382-94. PMC12244252
- Mah J, Pitre T. Oral magnesium supplementation for insomnia in older adults: a systematic review and meta-analysis. BMC Complement Med Ther. 2021;21:125. PMC8053283
- Peikert A, Wilimzig C, Köhne-Volland R. Prophylaxis of migraine with oral magnesium: results from a prospective, multi-center, placebo-controlled and double-blind randomized study. Cephalalgia. 1996;16(4):257-63. PMID 8792038
- von Luckner A, Riederer F. Magnesium in migraine prophylaxis — is there an evidence-based rationale? A systematic review. Headache. 2018;58(2):199-209. PMID 29131326
- Garrison SR, Korownyk CS, Kolber MR, et al. Magnesium for skeletal muscle cramps. Cochrane Database Syst Rev. 2020;(9):CD009402. Cochrane
- Ni Y, Deng F, Yu S, et al. A randomized, double-blind, placebo-controlled trial of magnesium-L-threonate for persistent pain after breast cancer surgery. Breast Cancer (Dove Med Press). 2023;15:495-504. PMC10386839 — publicly funded; reported no benefit.
- Lopresti AL, Smith SJ. The effects of magnesium L-threonate (Magtein) on cognitive performance and sleep quality in adults. Front Nutr. 2026;12:1729164. Full text — industry-funded.
- MedlinePlus. Magnesium Oxide. US National Library of Medicine.
- US Food and Drug Administration. Daily Value on the Nutrition and Supplement Facts labels — magnesium Daily Value 420 mg for adults and children 4 and older.
- US Food and Drug Administration. Drug Safety Communication: Low magnesium levels can be associated with long-term use of proton pump inhibitor drugs, 2011. The original FDA page has been retired; the communication is cited and summarised by the Linus Pauling Institute.
- NIH Dietary Supplement Label Database — label survey of 446 on-market products across 186 brands, retrieved November 2026, plus our earlier surveys of 45 labels per form for bisglycinate, oxide and citrate and 32 for L-threonate.
About the label survey on this page. The DSLD API rate-limits heavily, and we retrieved nine pages of 50 records before stopping: 450 records, of which 446 were on-market products named for magnesium and carrying a magnesium nutrient line, across 186 brands. It is the database’s own relevance ranking rather than a random draw, so the share of each form within the sample should not be read as market share. Figures about blending, delivery format and whether the panel names a salt are properties of those 446 records. Where a record named no magnesium compound we counted it as such rather than guessing, and we did not inspect physical labels.
Elemental percentages calculated from standard molecular weights (Mg 24.305; MgO 40.30; MgCO₃ 84.31; MgCl₂ 95.21; MgSO₄ 120.37; trimagnesium citrate anhydrous 451.11; magnesium malate 156.38; magnesium bisglycinate 172.42; monomagnesium citrate 214.41; magnesium taurate 272.59; magnesium L-threonate anhydrous 294.50 g/mol).
Information only, not medical advice. Talk to a doctor or pharmacist before starting a supplement, particularly if you have kidney disease, are pregnant or breastfeeding, or take prescription medication.