Magnesium

Overview

Magnesium is the mineral that about 300 enzymes need before they will run, including the ones that make ATP, relax a muscle fiber, and quiet a nerve signal. Roughly half of Americans eat less than the estimated requirement, which is why it sits in every pharmacy aisle. The catch is the form: in the one 60-day trial that compared three forms head to head, magnesium oxide, the cheapest and most common, did no better than placebo.

What is Magnesium?

Key takeaways

• In the one 60-day trial that put three forms head to head at 300 mg of elemental magnesium, oxide did not separate from placebo on serum, urine, or saliva, while citrate and amino acid chelate both raised urinary magnesium and citrate raised serum magnesium the most (Walker 2003, Magnesium Research, 46 adults).
• The sleep evidence rests mostly on 46 adults over 60 who slept poorly and took 500 mg a day for 8 weeks: they fell asleep faster, scored lower on the ISI sleep-difficulty scale, and had more serum melatonin, but total sleep time did not change (Abbasi 2012). Pooling that trial with two others, a 2021 meta-analysis found sleep onset about 17 minutes faster and graded the evidence low to very low (Mah 2021).
• Magnesium oxide is about 60 percent magnesium by weight, glycinate about 14 percent, and L-threonate about 8 percent. A label that reads "magnesium glycinate 500 mg" without the word elemental may be delivering about 70 mg of magnesium.
• Across 34 double-blind trials in 2,028 adults, a median 368 mg a day for three months lowered systolic blood pressure by 2.0 mmHg and diastolic by 1.8 mmHg (Zhang 2016, Hypertension). Real, repeatable, and small.
• The share of a dose you absorb falls as the dose rises, from about 65 percent at small amounts to 11 percent at large ones (Fine 1991, Journal of Clinical Investigation), which is why two 200 mg doses deliver more than one 400 mg dose.
• The Tolerable Upper Intake Level is 350 mg a day from supplements and medications, lower than the RDA, because the UL excludes food and the first thing that goes wrong above it is diarrhea (NIH ODS).
• About 48 percent of Americans eat less magnesium than the estimated average requirement, yet serum magnesium is not on a routine blood panel and the national nutrition survey has not measured it since 1974 (NIH ODS).

Magnesium is an essential mineral that the body cannot make. An adult carries about 25 grams of it, 50 to 60 percent in bone and most of the rest inside cells; less than 1 percent is in blood serum, where it is held between 0.75 and 0.95 mmol/L (NIH ODS). It is a cofactor in more than 300 enzyme systems, meaning those enzymes do not work until a magnesium ion is seated in them, and the list covers protein synthesis, muscle and nerve function, blood glucose control, blood pressure regulation, energy production, and the building of DNA, RNA, and glutathione (NIH ODS). That list is why magnesium gets attached to every health claim on the shelf. It is also why most of those claims are only half wrong.

There is no reserve tank. Bone holds the bulk and lends it slowly, and the kidneys are the thermostat: they excrete about 120 mg a day and cut that number when intake drops (NIH ODS). A healthy person on a low-magnesium diet drifts short over months, not days. That is the setup for the surprise: about 48 percent of Americans eat less than the estimated average requirement, and nobody is checking, because serum magnesium is not on a standard electrolyte panel and the national nutrition survey stopped measuring it in 1974 (NIH ODS).

Magnesium forms compared

Supplement labels sell the compound, and the compound is mostly not magnesium. The Supplement Facts panel is required to list elemental magnesium, the mineral itself; front labels and marketing copy often list the weight of the whole compound instead. The table below is what each form delivers per 100 mg of compound and what the human data actually show.

Two things in that table deserve a sentence each. Glycinate's reputation as the best-absorbed form is not built on a head-to-head human trial; it is built on chemistry, on tolerance, and on a 1994 study of twelve people with part of their intestine removed. Citrate is the form with the receipts. And the elemental math is not a footnote. Delivering 400 mg of magnesium from pure bisglycinate takes roughly 2,850 mg of raw material, which is why single-capsule products at 200 mg or more are almost always a buffered chelate or an oxide blend, and why the compound weight on a front label tells you nothing.

Where it comes from. Green leafy vegetables, legumes, nuts, seeds, and whole grains carry the most; an ounce of pumpkin seeds or almonds, a half cup of cooked spinach or black beans, and a bowl of oatmeal are the realistic daily sources (NIH ODS). Hard water adds some. Refined grains lose most of theirs in milling. Magnesium from almonds was absorbed as well as magnesium from a soluble supplement salt in the one study that tested a food directly (Fine 1991), so food first is not a slogan here. It works.

How the body handles it. Absorption happens in the small intestine by two routes: a passive route between cells that handles most of a large dose, and an active route through TRPM6 and TRPM7 channels that handles small amounts efficiently (Schuchardt 2017). The active route saturates. That is why fractional absorption drops from 65 percent of a small dose to 11 percent of a large one (Fine 1991), and why split doses outperform one big one. Protein, medium-chain fats, and fermentable carbohydrates in the same meal raise uptake; phytate from bran, oxalate from spinach, large doses of calcium or zinc, and insoluble fiber lower it (Schuchardt 2017). Whatever gets in and is not needed leaves in urine within a day or two.

Signs of running low, and the testing problem. Early deficiency looks like loss of appetite, nausea, fatigue, and weakness; as it deepens, numbness, tingling, muscle cramps, personality change, and abnormal heart rhythms appear, and severe deficiency drags calcium and potassium down with it (NIH ODS). The people most likely to get there are those with chronic diarrhea or intestinal malabsorption, people with type 2 diabetes who lose magnesium in urine, people with alcohol dependence, and older adults, whose absorption falls and excretion rises with age (NIH ODS). Testing is the awkward part. Serum magnesium is the standard test, with 0.75 to 0.95 mmol/L called normal and under 0.75 called low, but serum holds less than 1 percent of body magnesium and tracks tissue stores poorly. Red-cell magnesium, urinary magnesium, and a loading test exist, and no single method is considered satisfactory (NIH ODS). A normal serum value does not rule out a low tissue level; a low one is real. Our 2023 overview of magnesium supplements covers the food side.

How it works

Magnesium works by being the ion that ATP needs. Adenosine triphosphate, the molecule every cell spends as energy, is biologically active only as a magnesium-ATP complex; the magnesium ion bridges the phosphate groups so an enzyme can grip and split them. Every kinase, every ATPase pump, and every polymerase that copies DNA or RNA runs on magnesium-ATP, which is the plain reason a magnesium shortfall shows up first as fatigue and weakness rather than as any one organ failing (NIH ODS). The mitochondria that make the ATP depend on it too, which is where magnesium and CoQ10, the electron carrier in the same membrane, overlap in the fatigue conversation.

In nerves, magnesium is a brake. The NMDA receptor, the glutamate-gated channel that excites neurons and underlies learning, is plugged by a magnesium ion at resting voltage and opens only once the cell is already depolarized; magnesium is, in effect, a built-in NMDA antagonist. It also supports signaling at GABA-A receptors, the brain's main inhibitory switch. Less excitation plus more inhibition is the proposed mechanism behind the sleep findings, and the 2012 trial measured what that would predict: higher serum melatonin and lower cortisol in the magnesium group after 8 weeks (Abbasi 2012). Proposed is the operative word. The trial showed the hormones moved, and it did not show that NMDA block is the reason they moved.

In muscle, magnesium is calcium's counterweight. Calcium flooding into a muscle cell triggers contraction; magnesium competes with calcium at the same channels and binding sites and is required by the pumps that clear calcium back out so the fiber can relax (NIH ODS). Low magnesium tilts the balance toward contraction, which is the textbook explanation for cramps, twitching, and the tightness people describe. Smooth muscle in artery walls works the same way, and that is the mechanism behind the small, consistent blood pressure effect in the trials: a little less tone in the vessel wall, a little more give.

In bone, magnesium is both a brick and a regulator. Half to 60 percent of body magnesium is in the skeleton, and magnesium status changes the output of parathyroid hormone and the activation of vitamin D, the two hormones that run calcium traffic (NIH ODS). The enzymes that convert vitamin D to its active form need magnesium as a cofactor, which is why vitamin D3 taken against a magnesium shortfall can underperform, and why the vitamin D guide lists magnesium as a companion nutrient. Vitamin K2 directs where calcium lands; magnesium helps decide how much calcium is in motion to begin with.

The absorption physiology matters because it sets the dose. Magnesium crosses the gut by two routes: a passive leak between cells that scales with how much is dissolved in the lumen, and an active, saturable route through TRPM6 and TRPM7 channels (Schuchardt 2017). The active route is efficient and fills up fast. The passive route is leaky and depends on the compound being in solution. A form that will not dissolve (oxide in water, Lindberg 1990) cannot use either route well. A form that dissolves but arrives all at once overwhelms the active route and leans on the passive one, which is why fractional absorption fell from 65 percent to 11 percent as the dose rose in the Fine study. A chelate adds a third path: the 1994 isotope study found part of a magnesium diglycinate dose absorbed as an intact dipeptide, earlier and higher than the same magnesium from oxide in the patients with the worst absorption (Schuette 1994). Whether that pathway carries a meaningful share in a healthy gut has not been measured.

The mechanism predicts a specific pattern: measurable effects on sleep onset, blood pressure, and the symptoms of deficiency in people who were short to begin with; small or no effects in people who were not; and form differences that show up as absorption and tolerance rather than as different benefits. That is close to what the trials found.

Evidence at a glance

What the evidence doesn't support yet

Sleep for people who already sleep. Every positive sleep trial enrolled people with insomnia or self-reported poor sleep, most of them older, and the 2022 systematic review of nine studies found the randomized trials contradictory once you look past the best one (Arab 2022). No trial shows a good sleeper sleeping better on magnesium, and the mechanism does not predict one. A bottle that promises deep sleep to the general public is extrapolating from 46 retirees in Isfahan, which is a long way to extrapolate. If stress is the real problem, read the ashwagandha evidence instead.

Leg cramps. The oldest magnesium claim, and the Cochrane review ended it for the people who buy the product: in older adults with night cramps, magnesium was indistinguishable from placebo across five trials, with moderate certainty (Garrison 2020). Pregnancy cramps remain an open question, and exercise cramps have never been tested at all.

Glycinate as the absorption champion. The claim rests on a 12-patient study in people missing part of their intestine, and on chemistry. The one 2024 crossover that measured plasma after a single bisglycinate dose in healthy adults saw no significant rise (Pajuelo 2024). Glycinate earns its place on tolerance. The absorption lead belongs to citrate until someone runs the trial.

Threonate and brain magnesium. The brain-penetration story comes from rodent work. The human trials run three to six weeks, are company-funded, and the second found better cognition scores with no change in objective sleep (Lopresti 2026). Promising, early, and priced like a conclusion.

Clinical evidence

• The effect of magnesium supplementation on primary insomnia in elderly: A double-blind placebo-controlled clinical trial — 46 older adults with primary insomnia in Iran, 8 weeks, 500 mg of magnesium or placebo once daily. Measured the Insomnia Severity Index, sleep log, and serum magnesium, renin, melatonin, and cortisol. Magnesium cut sleep-onset latency and ISI score, raised sleep efficiency, sleep time, serum renin, and melatonin, and lowered cortisol versus placebo; total sleep time did not differ between groups. Limitation: small, single-site, 8 weeks, and the sleep measures were questionnaires rather than polysomnography.
• Oral magnesium supplementation for insomnia in older adults: a Systematic Review & Meta-Analysis — Three RCTs, 151 older adults with insomnia in three countries, oral magnesium versus placebo at under 1 gram up to three times a day. Pooled sleep-onset latency was 17.4 minutes shorter on magnesium (95% CI 7.4 to 27.3); total sleep time was 16 minutes longer but not significant. All trials were at moderate to high risk of bias and the authors graded the evidence low to very low quality. Limitation: three small trials, and the pooled benefit is driven by one of them.
• The Role of Magnesium in Sleep Health: a Systematic Review of Available Literature — Systematic review of 9 cross-sectional, cohort, and randomized studies with 7,582 adults, searching to November 2021, on dietary or supplemental magnesium and sleep duration, onset, awakenings, and stages. Observational studies linked better magnesium status with better sleep quality; the randomized trials gave contradictory results and the authors called the supplementation evidence uncertain. Limitation: narrative synthesis without pooling, and short trials under 12 weeks.
• Magnesium-L-threonate improves sleep quality and daytime functioning in adults with self-reported sleep problems: A randomized controlled trial — 80 adults aged 35 to 55 with self-assessed sleep problems, 21 days, 1 g/day of magnesium L-threonate (about 80 mg elemental) or placebo. Measured the Insomnia Severity Index, Leeds and Restorative Sleep questionnaires, mood, and Oura ring sleep and activity scores. Threonate improved Oura deep-sleep and REM scores, light-sleep time, readiness and activity scores, and self-rated energy, alertness, and mood versus placebo. Limitation: three weeks, authors employed by the ingredient supplier, and the elemental magnesium dose was far below the sleep trials in older adults.
• The effects of magnesium L-threonate (Magtein) on cognitive performance and sleep quality in adults: a randomised, double-blind, placebo-controlled trial — 100 adults aged 18 to 45 with self-reported dissatisfied sleep, 6 weeks, 2 g/day of magnesium L-threonate or placebo. Measured the NIH Cognition Toolbox, Raven's matrices, reaction time, self-rated sleep and wellbeing, and Oura ring sleep, heart rate, and heart rate variability. Threonate improved the total cognition composite, working and episodic memory, reaction time, and self-rated sleep-related impairment, with lower sleeping heart rate and higher HRV; Oura sleep outcomes and most self-rated sleep scales did not differ from placebo. Limitation: six weeks, industry-funded, and objective sleep did not move.
• Mg citrate found more bioavailable than other Mg preparations in a randomised, double-blind study — 46 healthy adults in the UK, 60 days, 300 mg/day of elemental magnesium as amino acid chelate, citrate, or oxide, or placebo, double-blind and parallel. Measured 24-hour urinary magnesium, serum magnesium, erythrocyte magnesium, and salivary magnesium at baseline, 24 hours, and 60 days. Citrate and chelate raised urinary magnesium more than oxide at 60 days; citrate produced the highest serum magnesium after both the first dose and 60 days and the highest salivary magnesium; oxide did not differ from placebo on any measure. Limitation: about 11 people per arm and surrogate markers rather than clinical outcomes.
• Magnesium bioavailability from magnesium citrate and magnesium oxide — Healthy volunteers given a 25 mmol (about 608 mg elemental) oral load of magnesium citrate or magnesium oxide, with urinary magnesium measured for 4 hours, plus in vitro solubility testing across simulated stomach acid levels. Citrate was 55 percent soluble in plain water and more soluble than oxide at every acid level; oxide was essentially insoluble in water and 43 percent soluble at peak acid. Urinary magnesium rose far more after citrate than after oxide (0.22 versus 0.006 mg per mg creatinine over 4 hours). Limitation: single large dose, small group, urinary surrogate.
• Bioavailability of US commercial magnesium preparations — Healthy volunteers given about 21 mEq/day (about 255 mg elemental) of four commercial preparations, with bioavailability measured as the rise in urinary magnesium. Magnesium oxide showed fractional absorption of about 4 percent; magnesium chloride, lactate, and aspartate were significantly higher and equivalent to one another. The authors concluded that an inorganic salt can match an organic one depending on the preparation, and that oxide is the outlier. Limitation: small sample, urinary measure, no citrate or glycinate arm.
• Bioavailability of magnesium diglycinate vs magnesium oxide in patients with ileal resection — 12 patients with prior ileal resection, double-blind randomized crossover, a single 100 mg dose of 26Mg-labeled magnesium diglycinate (chelate) or magnesium oxide. Measured isotope absorption and enrichment over time. Absorption was similar overall (23.5 versus 22.8 percent) but twice as high from the chelate in the four patients with the poorest oxide absorption (23.5 versus 11.8 percent); the chelate peaked about 3 hours earlier with a larger area under the curve and was better tolerated by all patients. Limitation: 12 people with surgically shortened intestines, a single 100 mg dose, and no healthy comparison group.
• Comparative Clinical Study on Magnesium Absorption and Side Effects After Oral Intake of Microencapsulated Magnesium (MAGSHAPE Microcapsules) Versus Other Magnesium Sources — 40 healthy adults, double-blind randomized crossover after 7 days on a low-magnesium diet, a single dose each of microencapsulated magnesium, magnesium oxide, magnesium citrate, and magnesium bisglycinate. Measured plasma magnesium at 1, 4, and 6 hours and digestive side effects. Plasma magnesium rose significantly after oxide at 1 hour and citrate at 4 hours; bisglycinate produced no significant rise at any time point; the microencapsulated product rose at all three and caused fewer digestive complaints. Limitation: single doses, plasma only, and the study was run for the microcapsule's manufacturer.
• Intestinal absorption of magnesium from food and supplements — Healthy adults given a standard meal plus 0, 10, 20, 40, or 80 mEq of magnesium acetate, with net absorption measured directly. Absorbed magnesium rose with each increment but fractional absorption fell from 65 percent at the lowest intake to 11 percent at the highest, fitting a saturable process plus a linear 7 percent leak. Magnesium from almonds was absorbed as well as magnesium acetate; enteric-coated magnesium chloride was absorbed far worse. Limitation: single-meal balance study in a small number of men, one salt.
• Effects of Magnesium Supplementation on Blood Pressure: A Meta-Analysis of Randomized Double-Blind Placebo-Controlled Trials — 34 randomized double-blind placebo-controlled trials, 2,028 normotensive and hypertensive adults, oral magnesium at a median 368 mg/day for a median 3 months. Pooled systolic blood pressure fell 2.00 mmHg (95% CI 0.43 to 3.58) and diastolic 1.78 mmHg (0.73 to 2.82), with serum magnesium up 0.05 mmol/L; a spline analysis found 300 mg/day or 1 month was enough to raise serum magnesium and lower pressure, and effects were larger in higher-quality trials. Limitation: residual heterogeneity across trials and a small absolute effect.
• Effects of Intravenous and Oral Magnesium on Reducing Migraine: A Meta-analysis of Randomized Controlled Trials — 21 RCTs searched to February 2015: 11 of intravenous magnesium for acute migraine (948 participants) and 10 of oral magnesium for prevention (789 participants). Oral magnesium reduced migraine frequency and intensity versus control (odds ratios 0.20 and 0.27); intravenous magnesium relieved acute attacks at 15 to 45 minutes, 2 hours, and 24 hours. Limitation: several included trials used inadequate randomization, doses and forms varied, and oral prevention doses were often above the supplemental UL.
• Oral Magnesium Supplementation for Treating Glucose Metabolism Parameters in People with or at Risk of Diabetes: A Systematic Review and Meta-Analysis of Double-Blind Randomized Controlled Trials — Double-blind RCTs of oral magnesium versus placebo in people with diabetes or at high risk of it, pooled as standardized mean differences for glucose and insulin-sensitivity outcomes. Magnesium lowered fasting plasma glucose in people with diabetes, and in high-risk people improved fasting glucose, 2-hour glucose after an oral tolerance test, and insulin-sensitivity markers. Limitation: trials were short and heterogeneous in dose and form, and the review did not test people with normal magnesium or normal glucose.
• Magnesium for skeletal muscle cramps — 11 RCTs, 735 people, oral magnesium in any form versus placebo or other treatment for cramps: 5 trials in pregnancy, 5 in idiopathic cramps in older adults (mean age 62 to 69), 1 in cirrhosis. In older adults, the change in cramps per week at 4 weeks was 0.18 fewer than placebo (95% CI 0.84 fewer to 0.49 more), not significant, with moderate-certainty evidence and no difference in intensity or duration; pregnancy results conflicted. Gastrointestinal side effects affected 11 to 37 percent on magnesium versus 10 to 14 percent on placebo; major adverse events did not differ. Limitation: pregnancy trials at high risk of bias, no exercise-cramp trials.
• Dose-Dependent Absorption Profile of Different Magnesium Compounds — BALB/c mice given oral magnesium citrate, malate, acetyl taurate, or glycinate at 45, 135, or 405 mg per 70 kg of elemental magnesium, with a split-dose arm, and serum, muscle, and brain magnesium measured at 24 hours. Citrate raised muscle and brain magnesium at every dose; acetyl taurate raised brain magnesium; malate and glycinate did not stand out; splitting the high dose did not raise tissue levels further. Limitation: mice, single day, and tissue levels rather than any functional outcome.
• Intestinal Absorption and Factors Influencing Bioavailability of Magnesium-An Update — Review of human and mechanistic data on magnesium absorption. Magnesium crosses the gut by a passive paracellular route and an active transcellular route through TRPM6 and TRPM7 channels; bioavailability spans a wide range set mainly by dose and meal composition. Phytate, oxalate, insoluble fiber, and high doses of other minerals reduce uptake; protein, medium-chain triglycerides, and fermentable carbohydrates increase it. Several small doses through the day are absorbed better than one large one, and the authors judged the salt form less important than marketing implies. Limitation: narrative review without pooled estimates, and it predates the 2024 crossover data.
• Magnesium: Fact Sheet for Health Professionals — The reference source on this page for the RDA (310 to 420 mg/day for adults from all sources), the Tolerable Upper Intake Level (350 mg/day from supplements and medications), body content and serum ranges, the 48 percent of Americans below the estimated average requirement in NHANES 2013 to 2016, the deficiency symptom sequence, the medication interactions with tetracyclines, quinolones, bisphosphonates, diuretics, and proton pump inhibitors, and the hypermagnesemia case reports. Limitation: a summary document, not a study; its underlying citations are listed on the fact sheet.

Dosage & forms

• Magnesium glycinate (bisglycinate):
• Magnesium citrate:
• Magnesium amino acid chelate (including buffered chelate):
• Magnesium L-threonate:
• Magnesium malate:
• Magnesium oxide:

Safety & interactions

How much to take

The research dose is 300 to 500 mg of elemental magnesium a day, and most products deliver that, provided the number on the label is elemental. The sleep trial used 500 mg a day for 8 weeks (Abbasi 2012); the three trials pooled in the 2021 meta-analysis gave under 1 gram up to three times a day (Mah 2021); the blood pressure trials averaged 368 mg a day, and the dose-response curve flattened near 300 mg (Zhang 2016); the absorption comparison used 300 mg (Walker 2003). A 200 mg capsule twice a day or a 400 mg scoop once a day sits inside that window. A 100 mg tablet does not, and neither does a multivitamin with 50 mg.

Split the dose if you can. Fractional absorption falls from about 65 percent of a small dose to 11 percent of a large one (Fine 1991), and the absorption review reached the same conclusion: several small doses through the day beat one large one (Schuchardt 2017). Two 200 mg doses with meals will deliver more magnesium than one 400 mg dose at bedtime. If sleep is the goal, the compromise is one dose with dinner and one at bedtime; no trial has tested timing directly, so evening dosing is convention, not evidence.

Take it with food. Protein, medium-chain fats, and fermentable carbohydrates raise uptake; bran, spinach-level oxalate, and large simultaneous doses of calcium or zinc lower it (Schuchardt 2017). One study found that 142 mg a day of zinc, far above any sensible supplement, disrupted magnesium balance (NIH ODS). Keep magnesium 2 hours away from any antibiotic or bone drug on the table below. Pair it with vitamin D3 if you take one; the enzymes that activate vitamin D need magnesium, and the D3 and K2 pairing works better on an adequate magnesium intake.

Expect nothing for the first few weeks. Serum magnesium rises within 24 hours of a dose and plateaus by about a month (Walker 2003; Zhang 2016). Sleep-onset changes appeared at 3 weeks in the threonate trial and 8 weeks in the trial in older adults. Blood pressure moved by 1 to 3 months. If nothing has changed by 8 weeks at a research dose, more magnesium is unlikely to help, and the honest next question is whether magnesium was ever the bottleneck.

The ceiling is the Tolerable Upper Intake Level, 350 mg a day from supplements and medications (Food and Nutrition Board, via NIH ODS). It is lower than the RDA because the RDA counts food and the UL does not, and because the endpoint the UL is built on is diarrhea, not toxicity. Doses above 350 mg are common in trials and in products and are tolerated by most healthy adults. Above it, the marginal milligram mostly ends up in the toilet.

Safety profile

Oral magnesium at supplement doses has one of the longest safety records of any ingredient on this site. The reason is mechanical: a healthy kidney clears the excess within a day (NIH ODS). The largest pooled dataset, 34 double-blind trials in 2,028 adults at a median 368 mg a day for a median 3 months, reported no safety signal beyond the digestive one (Zhang 2016). The Cochrane cramp review pooled adverse events across its trials and found major adverse events no more common than placebo, 2 of 72 on magnesium against 3 of 68 on placebo (Garrison 2020). The threonate trials at 1 to 2 grams of compound a day reported no significant adverse reactions (Hausenblas 2024; Lopresti 2026).

Common. Loose stools and diarrhea, sometimes with nausea and abdominal cramping. In the Cochrane pooled analysis, 11 to 37 percent of magnesium recipients reported gastrointestinal effects against 10 to 14 percent on placebo (Garrison 2020). The forms most often responsible are carbonate, chloride, gluconate, and oxide; the effect is osmotic, unabsorbed salt pulling water into the colon (NIH ODS). The fix is the form, the dose, or the split, in that order.

Uncommon. Lightheadedness in people whose blood pressure runs low or who take medication for it, consistent with the 2 mmHg drop in the meta-analysis (Zhang 2016). A heavy stomach after one large dose (Pajuelo 2024).

Rare. Hypermagnesemia, a serum magnesium high enough to matter, essentially does not occur from supplements in people with working kidneys. The reported cases involved laxative or antacid intakes above 5,000 mg a day, or kidney failure, or both (NIH ODS). Signs, in the order they appear as serum climbs past about 1.74 mmol/L: nausea, flushing, low blood pressure, urinary retention, lethargy, then muscle weakness, breathing difficulty, and irregular heartbeat (NIH ODS).

Drug interactions

Who should ask a clinician first

People with reduced kidney function, kidney disease, or on dialysis. The kidney is the only exit for magnesium, and every reported case of dangerous magnesium excess involved a very large intake, a kidney that could not clear it, or both (NIH ODS). This is the one group for whom an ordinary supplement dose can be unsafe. Ask, and expect a serum level.

People who are pregnant or nursing. The RDA rises to 350 to 400 mg a day in pregnancy, the supplement UL stays at 350, and the trials of magnesium for pregnancy leg cramps conflict (Garrison 2020). A prenatal already carries some. Ask before adding more.

Anyone on the drugs in the table above, and anyone who has taken an acid blocker for more than a year. People with type 2 diabetes, who lose extra magnesium in urine and may benefit, should coordinate rather than self-prescribe, because the glucose effect is real and may change what their medication needs to do (Veronese 2021). People with chronic diarrhea, celiac disease, Crohn's disease, or a history of intestinal surgery are the most likely to be deficient and the least likely to tolerate the forms that correct it; a chelate and a clinician are the right pair (Schuette 1994). Older adults on several medications fit most of these categories at once.

If none of that applies to you and you get loose stools, you have found your dose. Back off by one capsule or half a scoop; that is the whole safety instruction for most people.

Buying guide

The number that matters is elemental magnesium

Read the Supplement Facts panel, not the front label. The panel is required to state elemental magnesium, the mineral itself; the front of the bottle often states the weight of the compound, and the gap is enormous. Magnesium oxide is about 60 percent magnesium by weight, citrate about 16 percent, glycinate about 14 percent, L-threonate about 8 percent. So "magnesium bisglycinate 1,000 mg" on a front label can mean 140 mg of magnesium, a third of the research dose, and the panel is where you find out. "Magnesium (as magnesium bisglycinate) 200 mg" is the honest number. "Magnesium complex 500 mg" with no elemental figure is hiding it.

Match the form to the job

Citrate has the most human absorption data and is the default for a powder or a drink; it will loosen stool in some people at 400 mg. Glycinate is the form for anyone who cannot tolerate citrate, at the cost of capsule count. An amino acid chelate, including the buffered 40 percent chelate we reviewed in full, is the way to get 200 mg into one capsule without reaching for plain oxide. L-threonate is a cognition bet at a premium. Malate is fine and unproven. Oxide is the form to avoid for repletion and the form to use for constipation; a product that leads with oxide and prints "high potency" is selling you the elemental number and keeping the absorption number.

Proof of testing

The raw mineral is cheap and stable; the quality question is whether the capsule contains what the label says and nothing it should not, and the only proof is a certificate of analysis by lot. Ours are made in our own cGMP facility in Mt. Angel, Oregon; every batch is tested in our in-house lab for identity, potency, and purity; and certificates of analysis are published by lot on our lab-testing page. As of this writing, Good Magnesium's certificates are posted there and Calming Mag Powder's are not yet, which is exactly the kind of gap a reader should hold a brand to. Demand the same from any brand, including us: a lot number, a findable certificate, a potency result that matches the label. Never read one? Our guide to supplement testing and certificates of analysis explains every line.

Price per day at the real dose

Do the math at the elemental dose you will actually take, not per bottle. Good Magnesium is 60 capsules at 200 mg of elemental magnesium as an amino acid chelate (the panel reads rice protein chelate) for $21.99, which is $0.37 a day at one capsule and $0.73 a day at the 400 mg research dose. Calming Mag Powder is 60 scoops at 400 mg of elemental magnesium, from magnesium carbonate that reacts with citric acid in the glass to form magnesium citrate, for $47.99, which is $0.80 a day; the scoop also carries 1,000 IU of vitamin D3, 120 mg of vitamin C, 15 mg of zinc, 98 mg of potassium, 500 mg of glycine, and 3 mg of boron, so it is priced as a drink mix rather than as a magnesium capsule. Prices and serving counts checked on the live product pages 2026-10-14. Run the same arithmetic on any other brand at 400 mg elemental; a glycinate at three capsules a serving usually costs more per day than its bottle price suggests.

Red flags

• A proprietary blend or "magnesium complex" that hides the elemental amount.
• A front label in compound weight with no elemental figure on the panel.
• "Clinically proven" with no study named. The studies on this page are the ones a label can honestly point to.
• "100 percent absorbed." No form is, and the best-documented one managed a measurable fraction.
• L-threonate priced by the gram of compound as if it were a gram of magnesium.
• Oxide as the first ingredient in a product marketed for sleep or relaxation.
• More than 500 mg of elemental magnesium per serving with no reason given; the sleep and blood pressure trials needed less, and the rest is excreted.

Where our two products land

Both meet the checklist above and they are for different people. Good Magnesium is the capsule: 200 mg elemental per capsule, one or two a day, a chelate that supports normal muscle relaxation and a calm nervous system without the digestive effect of a cheaper salt.* Calming Mag Powder is the evening drink: 400 mg elemental per scoop, fizzing into citrate, with glycine and vitamin D3 alongside, for people who want a wind-down ritual that supports restful sleep.* A full scoop is 50 mg over the supplemental UL; if loose stools show up, a half scoop is a complete dose for most people. Neither product is for someone whose intake is already adequate and who sleeps fine. The honest fit is a diet that runs short, a form that absorbs, and a dose inside the research range.

Next reads

• Our full review of the chelate form: why a 40 percent buffered chelate exists and what it trades away.
• Vitamin D3 (cholecalciferol): the vitamin that needs magnesium to switch on, and the dose that goes with it.
• How to get the sleep your body needs: the habits that have to be in place before any mineral can show its effect.

The bottom line

If you eat few greens, beans, nuts, or whole grains, or you take a diuretic or a long-term acid blocker, 300 to 400 mg of elemental magnesium a day as citrate or a chelate, split across two meals, is cheap, safe, and inside the research range. Expect faster sleep onset and a point or two off blood pressure, not a transformation. If your diet covers the RDA and you sleep well, skip it. If your kidneys are impaired, that changes everything, and the answer is no until your clinician says otherwise.

From Mt. Angel Vitamins

• Good Magnesium – Gentle, Bioavailable Magnesium Chelate For Muscles, Sleep, And Stress Support

  $21.99

  Shop Now

Frequently Asked Questions

Magnesium glycinate vs citrate vs threonate: what is the difference?

The difference is the molecule magnesium is bound to, and it changes three things: how much magnesium you get per milligram of compound, how well it dissolves, and how your bowel reacts. Citrate is about 16 percent magnesium, dissolves in water, and has the most human absorption data; it loosens stool in some people. Glycinate is about 14 percent magnesium, the gentlest on digestion, and its absorption reputation rests on one 12-patient study rather than a head-to-head trial. L-threonate is about 8 percent magnesium, studied at 1 to 2 grams of compound a day in two short company-funded trials for cognition and sleep, and costs the most per milligram of magnesium. For a general daily dose, citrate or a chelate; for a sensitive gut, glycinate.

How much elemental magnesium is in each form?

Per 100 mg of compound: oxide about 60 mg, citrate about 16 mg (about 11 mg for hydrated powders), malate about 15 mg, glycinate or bisglycinate about 14 mg, a fully reacted amino acid chelate 10 to 20 mg, a buffered chelate about 40 mg, and L-threonate about 8 mg. The Supplement Facts panel is required to list elemental magnesium, so the number there is the one to trust. A front label that says 'magnesium glycinate 500 mg' may be describing 500 mg of compound, which is about 70 mg of magnesium, a fifth of what the sleep and blood pressure trials used.

Does magnesium help with sleep, and what dose did the trials use?

For older adults with insomnia, probably yes, modestly. The best trial gave 46 adults over 60 a 500 mg dose daily for 8 weeks; they fell asleep faster, scored lower on the Insomnia Severity Index, and had higher serum melatonin, but total sleep time did not change (Abbasi 2012). A meta-analysis of three trials put sleep onset about 17 minutes faster and rated the evidence low quality (Mah 2021). In adults 35 to 55 with poor sleep, 1 gram a day of magnesium L-threonate for 3 weeks improved wearable-measured deep sleep (Hausenblas 2024), though a second trial found no objective change. No trial shows benefit in people who already sleep well. The habits in our [sleep guide](/blog/sleep-mood/how-to-get-the-sleep-your-body-needs) come first.

What are the signs of low magnesium, and how is it tested?

Early signs are loss of appetite, nausea, fatigue, and weakness; as the shortfall deepens, numbness, tingling, muscle cramps, personality changes, and abnormal heart rhythms can follow, and severe deficiency pulls calcium and potassium down with it (NIH ODS). Testing is imperfect. Serum magnesium is the standard test, with 0.75 to 0.95 mmol/L called normal, but serum holds under 1 percent of body magnesium and tracks tissue stores poorly; red-cell magnesium, urinary magnesium, and a loading test exist, and no single method is considered satisfactory. A low serum value is real and worth acting on. A normal one does not rule out a low tissue level, which is why diet history matters as much as the lab.

Can magnesium interact with antibiotics or blood pressure medication?

Yes to both, in different ways. Magnesium binds tetracycline antibiotics (doxycycline, minocycline) and quinolones (ciprofloxacin, levofloxacin) in the gut into complexes that are not absorbed, so the antibiotic underdoses; take the antibiotic at least 2 hours before or 4 to 6 hours after any magnesium (NIH ODS). The same spacing applies to oral bisphosphonates. With blood pressure medication the issue is additive effect: supplemental magnesium lowered pressure by about 2 mmHg on its own across 34 trials (Zhang 2016), which is rarely a problem but worth telling the prescriber. Loop and thiazide diuretics drain magnesium; potassium-sparing diuretics retain it; proton pump inhibitors used over a year can push it low.

What is the upper limit for supplemental magnesium?

350 mg a day from supplements and medications combined, for adults, set by the Food and Nutrition Board and reported by NIH ODS. It is lower than the RDA of 310 to 420 mg because the RDA counts food and the UL does not; magnesium from food is not limited because the kidneys clear it. The UL is built on diarrhea as the endpoint, not toxicity, so exceeding it by 50 to 150 mg, as many trials and products do, is a tolerance question for healthy adults rather than a safety one. Dangerous magnesium excess has been reported only at laxative or antacid intakes above 5,000 mg a day or in people whose kidneys cannot clear it.

Which form causes the least digestive upset?

Glycinate (bisglycinate), followed by amino acid chelates, then citrate, with oxide, carbonate, chloride, and gluconate at the loose-stool end (NIH ODS). Digestive upset from magnesium is osmotic: whatever is not absorbed sits in the colon and pulls water in, so the forms that are absorbed least, or dosed highest, cause the most. Glycinate's advantage is partly that the glycine half of the molecule is not an osmotic salt at all. Two other levers matter as much as the form: splitting the dose, since fractional absorption roughly doubles at smaller doses (Fine 1991), and taking it with a meal. If citrate is the only option, 200 mg twice a day with food is usually tolerated where 400 mg at once is not.

Is magnesium oxide useless?

For raising your magnesium status, close to it. Fractional absorption measured about 4 percent in healthy volunteers (Firoz 2001), and in the 60-day head-to-head trial at 300 mg a day, oxide did not separate from placebo on serum, urinary, or salivary magnesium while citrate and chelate did (Walker 2003). It is nearly insoluble in water and less than half soluble even at peak stomach acid (Lindberg 1990). Oxide is useful for one thing: unabsorbed oxide draws water into the colon, which is why it is the standard laxative and antacid form. A sleep or relaxation product built on oxide is selling the big number on the label, not the magnesium.

Do I need a magnesium supplement at all?

Possibly not. About 48 percent of Americans eat less than the estimated average requirement (NIH ODS), so the odds are close to a coin flip, and the answer depends on your diet. An ounce of pumpkin seeds, a half cup of cooked spinach or black beans, a serving of almonds, and whole grains instead of refined ones will cover the RDA for most people, and magnesium from almonds absorbed as well as a supplement salt in the one study that checked (Fine 1991). The people most likely to need a supplement are those with a low-magnesium diet, chronic diarrhea or malabsorption, long-term use of a diuretic or acid blocker, or older age. If none of that describes you and you sleep well, your money is better spent on the groceries.

Who should ask a doctor before taking magnesium?

Anyone with reduced kidney function, kidney disease, or on dialysis, first and without exception: the kidney is the only way out for magnesium, and every reported case of dangerous excess involved a kidney that could not clear it or an intake above 5,000 mg a day (NIH ODS). After that: anyone taking a tetracycline or quinolone antibiotic, an oral bisphosphonate, a diuretic, or a proton pump inhibitor; anyone pregnant or nursing; people with diabetes, whose glucose response may change; and people with celiac disease, Crohn's disease, chronic diarrhea, or prior bowel surgery, who are most likely to be low and least likely to tolerate the fix. For everyone else, the stopping rule is simple: loose stools mean the dose is too high.

Related Mt. Angel Vitamins Products

• Recommended for Magnesium

  Good Magnesium – Gentle, Bioavailable Magnesium Chelate For Muscles, Sleep, And Stress Support

  $21.99

  Shop Now

• Recommended for Magnesium

  Calming Mag Powder – Magnesium Relaxation & Sleep Support Drink

  $47.99

  Shop Now