Magnesium: What It Actually Does, and Why Form and Cofactors Decide Everything

Magnesium is a mineral your body uses as a tool rather than a building block. More than 600 enzymes require it as a cofactor and about 200 more use it as an activator. Its largest single job is making ATP usable: adenosine triphosphate only works when bound to magnesium. Which form you take, and what you take alongside it, decides how much of that happens.

What Does Magnesium Actually Do?

Magnesium is the fourth most abundant cation in the body and the second most abundant inside cells, after potassium. It carries two positive charges, and that charge density is the whole story. Magnesium is useful because it grips negatively charged things — phosphate groups, nucleic acid backbones, enzyme active sites — and holds them in a working shape.

The familiar number is "300+ enzyme systems," from the NIH Office of Dietary Supplements. It isn't wrong. It is conservative, traces to reviews written in the 1980s and 90s, and counts systems rather than enzymes. The better figure comes from de Baaij, Hoenderop and Bindels in Physiological Reviews (2015), derived from curated enzyme databases: over 600 enzymes list Mg²⁺ as a cofactor, and roughly 200 more list it as an activator. You will also see "800+ enzymes." That is those two numbers added together, and it misrepresents both.

What "magnesium-dependent enzyme" actually means

The phrase describes three different relationships, and marketing collapses them into one.

1. Substrate-level dependence. The largest class. The enzyme's true substrate is not ATP but Mg-ATP²⁻. Kinases, ATPases, synthetases. Magnesium isn't assisting; it is part of the thing being acted on.
2. Catalytic cofactor. Mg²⁺ sits in the active site and does chemistry. DNA and RNA polymerases use a two-metal-ion mechanism positioning two magnesium ions.
3. Allosteric activator. Free Mg²⁺ binds a regulatory site and raises activity. Alpha-ketoglutarate dehydrogenase works this way. This is roughly the 200-enzyme group.

Because most magnesium-dependent enzymes sit in the first category, magnesium status isn't a peripheral modifier of metabolism. It is a supply constraint on it. Per Fiorentini's 2021 review in Nutrients, that constraint shows up in at least six of the ten steps of glycolysis, including the rate-limiting phosphofructokinase-1; in three mitochondrial dehydrogenases; and at the F₀/F₁-ATPase, where magnesium is the activator of the terminal step that physically makes ATP. It also stabilizes the DNA double helix by neutralizing its phosphate backbone, and it runs the Na⁺/K⁺-ATPase, which consumes 20–40% of basal cellular ATP.

Two further roles are worth naming because they explain how magnesium feels. Magnesium competes with calcium at calmodulin and troponin C, modulates L-type calcium channels and promotes calcium sequestration — contraction is a calcium event, and magnesium sits on the relaxation side of it, which is why the Linus Pauling Institute calls it "a calcium antagonist and potent vasodilator." And at rest, a single Mg²⁺ ion physically plugs the pore of the NMDA glutamate receptor; when magnesium falls, fewer channels are blocked and neurons become hyperexcitable. Structural work in Nature Neuroscience (2026) has now resolved how that block works.

Why Is ATP Useless Without Magnesium?

This is the most under-explained fact in the category, and it is what makes "magnesium for energy" a real statement rather than a vague one.

ATP carries three phosphate groups in a row, each negatively charged. Those charges repel one another — part of why the molecule stores energy — but they also make naked ATP a poor substrate, interfering with enzyme binding and with the geometry of phosphate transfer.

Magnesium fixes that. Mg²⁺ chelates the β- and γ-phosphates, neutralizing charge and producing, in Fiorentini's words, "an adequate conformation that allows weakening of the terminal O–P bond of ATP, thereby facilitating the transfer of phosphate." The functional molecule in every cell is MgATP²⁻. Free ATP is, for most enzymatic purposes, inert.

Hence the literature's flat statement that magnesium is a cofactor in all reactions involving the utilization and transfer of ATP. A human body turns over roughly its own weight in ATP daily — not by making that much, but by recycling one small pool thousands of times. Magnesium is required at every turn. The full cycle is laid out in how your body actually makes energy.

Two boundaries. This is a mechanism, not an outcome: adequate magnesium doesn't make you feel energetic, it makes the machinery capable of running. And intracellular magnesium of 10–30 mM sounds generous until you learn only 0.5–1.2 mM is free. The rest is already bound. The free pool is small, and it is the pool that regulates.

Where Does Your Magnesium Sit, and Why Doesn't a Blood Test Find It?

An adult holds roughly 25 g of magnesium, about 1,000 mmol. The distribution explains why status is so hard to measure.

Compartment	Share of body magnesium	Note
Bone	50–60%	About a third is exchangeable — a real reserve, not inert mineral
Muscle and soft tissue	34–39%	The functionally active pool. No routine test measures it
Blood and extracellular fluid	Under 1%	The only pool your lab draws

Serum sits between 0.75 and 0.95 mmol/L. When intake falls, the kidney sharply cuts excretion and bone releases exchangeable magnesium, so serum holds steady while tissue draws down — potentially for years. Fiorentini states it without hedging: "a normal level of serum magnesium does not rule out magnesium deficiency." That condition has a name — chronic latent magnesium deficiency, normal serum with a depleted total-body pool, detectable only by a loading and retention test. A 2014 study found it in apparently healthy university students.

The reference range itself is disputed. Costello and colleagues in Advances in Nutrition (2016) argued the 0.75 mmol/L lower bound came from population distributions rather than health outcomes, and that many people inside the "normal" 0.75–0.85 band are functionally deficient. A 2024 meta-analysis in Advances in Nutrition gave that its first outcome anchor: serum magnesium shows a U-shaped relationship with dementia risk, optimum near 0.85 mmol/L, hazard ratio 1.43 below 0.75 and 1.30 above 0.95.

Test	Measures	Verdict
Serum total Mg	Under 1% of the body pool	Standard of care. Detects only overt hypomagnesemia
Ionized Mg2+	The active ~70% of serum magnesium	More meaningful in principle; assays unstandardized, no consensus range
RBC magnesium	Intracellular pool, ~3× plasma	Better tissue proxy; deficiency cut-off <1.6 mmol/L. Reflects red cell lifespan, not current status
24-hour urinary Mg	Renal handling	Under 80 mg/day suggests compensatory conservation
Loading/retention test	Fraction of a load retained	The reference standard. Impractical outside research

The most useful recent advance skips measurement and scores causes instead. The Magnesium Depletion Score, from Fan and colleagues in the Journal of Nutrition (2021), gives one point each for current diuretic use, current PPI use, heavy alcohol intake, and an eGFR of 60 to under 90 — two points if eGFR is under 60. Zero to one is low, two is middle, three to five is high. In US adults with diabetes (NHANES 2003–2018, n = 5,219), high versus low MDS carried an all-cause mortality hazard ratio of 1.58 (95% CI 1.20–2.08) and a cardiovascular mortality hazard ratio of 1.92 (95% CI 1.28–2.88). It has since been validated in established cardiovascular disease and stroke.

The editorial point: your magnesium risk is better predicted by what is depleting you than by a test designed to look normal. The symptom side is covered in signs you might be low on magnesium.

How Much Do You Need, and How Much Are You Getting?

Age	Male	Female	Pregnancy
9–13 y	240 mg	240 mg	—
14–18 y	410 mg	360 mg	400 mg
19–30 y	400 mg	310 mg	350 mg
31+ y	420 mg	320 mg	360 mg

The Daily Value used on labels is 420 mg. Against that: 48% of Americans of all ages consume less magnesium from food than their Estimated Average Requirement, per NIH ODS citing NHANES / What We Eat In America 2013–2016. Mean intakes without supplements in an earlier cycle were 268 mg/day for men and 234 mg/day for women. Highest risk: men over 71, and adolescents.

One correction, because credibility depends on it. That 48% describes inadequate intake relative to a requirement estimate, not diagnosed deficiency, and the claim that "80% of Americans are magnesium deficient" has no basis. The defensible version was made properly by Rosanoff and colleagues in Nutrition Reviews (2012): suboptimal magnesium status is common and clinically underappreciated. We put numbers on the whole picture in the magnesium gap.

What Depletes Magnesium — and What Doesn't

Depleter	Mechanism	Evidence
Chronic stress	Catecholamines and cortisol shift magnesium out of cells; kidneys then excrete it	Strong
Alcohol	Renal tubular magnesium wasting, plus poor intake and GI losses	Strong
Proton pump inhibitors	Impair TRPM6/TRPM7-mediated absorption; FDA safety communication 2011	Strong, dose-related
Loop and thiazide diuretics	Increase renal excretion	Strong
Aging	Lower absorption, higher excretion, more polypharmacy	Strong
Refined diets	Milling wheat to white flour removes ~80% of its magnesium	Strong
High blood glucose	Osmotic diuresis drives urinary loss	Strong
Phytate and oxalate	Bind magnesium in the gut; spinach magnesium is less absorbable than kale	Well demonstrated
Heavy sweating	Magnesium lost in sweat; turnover rises	Real but modest for most people
Caffeine	Acute, roughly 3-hour rise in urinary excretion	Weak. No evidence habitual coffee causes deficiency

That last row deserves a sentence, because the internet has decided otherwise. Massey and Whiting (Journal of Nutrition, 1993) found oral caffeine raises urinary calcium, magnesium, sodium and chloride "for at least 3 h"; a 1994 follow-up found no sustained effect, and coffee itself supplies about 7 mg of magnesium a cup. We do the arithmetic in what really drains your minerals.

Stress is the one that runs in a loop. Pickering and colleagues, "Magnesium Status and Stress: The Vicious Circle Concept Revisited" (Nutrients, 2020), describe both directions. Going out: a stressor activates the HPA axis, magnesium shifts from intracellular to extracellular compartments, and the kidneys excrete it, drawing progressively on the bone reserve. Coming back: as magnesium falls, the NMDA block weakens and glutamatergic excitation rises while magnesium's support of GABA-A inhibition weakens. Less braking, more accelerator. Magnesium-deficient animals show elevated ACTH and CRH and are sensitized to further stressors.

There is human data on the cortisol side. Joris and colleagues (Clinical Endocrinology, 2021) reported a post-hoc analysis of a 24-week randomized trial in 49 overweight adults aged 45–70 taking 350 mg/day of magnesium: 24-hour urinary cortisol fell by 32 nmol/24h (95% CI −59 to −5, p = 0.021) and the cortisol/cortisone ratio by 0.10 (p = 0.005). One post-hoc analysis of one modest trial, but a real one. More in why stress burns through your minerals.

Why 500 mg of Magnesium Citrate Is Not 500 mg of Magnesium

Magnesium doesn't exist on a shelf as magnesium. It exists bonded to something else, and the front of the box may be quoting the weight of the whole compound. Magnesium's atomic weight is 24.305; everything else in the molecule is packaging.

Compound	Elemental magnesium	Magnesium in 500 mg of compound
Oxide	60.3%	302 mg
Carbonate (anhydrous)	28.8%	144 mg
Chloride (anhydrous / hexahydrate)	25.5% / 12.0%	128 / 60 mg
Sulfate (anhydrous / heptahydrate)	20.2% / 9.9%	101 / 50 mg
Trimagnesium dicitrate (anhydrous)	16.2%	81 mg
Citrate (hydrated or monobasic)	11–12%	55–60 mg
Malate (anhydrous / dihydrate)	15.5% / 12.6%	78 / 63 mg
Bisglycinate (anhydrous / dihydrate)	14.1% / 11.7%	71 / 59 mg
Lactate (anhydrous / dihydrate)	12.0% / 10.2%	60 / 51 mg
Marine / ionic seawater-derived	~10–12%	50–60 mg
Taurate	8.9%	45 mg
L-threonate	8.3%	42 mg
Orotate (anhydrous / dihydrate)	7.3% / 6.6%	37 / 33 mg

Two things fall out. "500 mg magnesium citrate" is somewhere between 55 and 81 mg of magnesium depending on the salt and hydration state, and most front labels don't say which. The Supplement Facts panel is required to state elemental magnesium — read that, not the marketing, which is what our guide to reading an electrolyte label is for. And hydration state matters more than anyone admits: bisglycinate anhydrous is 14.1% magnesium, the dihydrate 11.7%. Sources quoting one figure per form are quietly picking the flattering one.

The crucial caveat: elemental percentage is a denominator, not a benefit. It tells you how much powder you need. It says nothing about how much you absorb. Oxide leads this table at 60.3% and is the worst-absorbed common form in the literature.

This arithmetic is the reason Current uses three forms rather than one. Magnesium bisglycinate for tolerance, magnesium malate for its Krebs-cycle carrier, and ionic magnesium from inland seawater concentrate, which arrives already dissociated — 220 mg of elemental magnesium in total, 52% of the Daily Value.

Which Form of Magnesium Is Best?

There is no single best form, and the honest answer depends on the goal. But the forms are not interchangeable, and four studies do most of the work. Firoz and Graber (Magnesium Research, 2001) gave volunteers about 21 mEq/day of four commercial preparations and measured urinary increment: oxide came in at 4% fractional absorption, while chloride, lactate and aspartate were significantly higher and statistically equivalent. Walker and colleagues (Magnesium Research, 2003) randomized 46 adults double-blind to 300 mg elemental magnesium daily for 60 days as amino-acid chelate, citrate or oxide: citrate produced the greatest serum concentration acutely (p = 0.026) and chronically (p = 0.006), while oxide "resulted in no differences compared to placebo." Kappeler (BMC Nutrition, 2017) replicated citrate over oxide in a randomized crossover. And Blancquaert, Vervaet and Derave (Nutrients, 2019) screened 15 commercial products and found in vitro dissolution at pH 6.8 predicted in vivo bioavailability at r = 0.867 (p < 0.001).

Form	Elemental Mg	Absorption evidence	GI tolerance	Best use
Oxide	60.3%	Worst in class: 4% absorbed, no better than placebo on serum	Poor — undissolved magnesium reaches the colon	Genuinely good laxative and antacid; defensible inside a blend, not as standalone repletion
Citrate	16.2% / 11–12%	Best-evidenced form (Walker 2003, Lindberg 1990, Kappeler 2017)	Laxative above ~300–400 mg; a licensed bowel-prep agent at high dose	General repletion; constipation; citrate independently inhibits calcium oxalate stones
Glycinate / bisglycinate	14.1% / 11.7%	Thinner than the marketing implies — a small 1994 isotope study in ileal-resection patients is the primary citation, and Walker 2003 found a chelate inferior to citrate on serum	Best tolerated in practice, and this part is well supported: a neutral chelate leaves little free ionic magnesium in the lumen	Evening dosing, higher doses, sensitive stomachs
Malate	15.5% / 12.6%	Limited direct human data; grouped with the well-absorbed organic acid salts	Well tolerated, less laxative than citrate	Daytime dosing; carrier is a genuine Krebs cycle intermediate
Chloride	25.5% / 12.0%	Good — better than oxide, equivalent to lactate and aspartate	Soluble; can be laxative, tastes bitter and salty	Liquid and ionic formats; used in several better clinical trials
Lactate	12.0% / 10.2%	Good; the form used in the largest magnesium stress trial	Well tolerated, low laxative effect	General repletion; common in European pharmaceutical products
Ionic / marine	~10–12%	Better than most realize. Dowley 2024: more soluble than bisglycinate formulations in gastric and intestinal phases with food. McFarlin 2025 (n=20 crossover): both seawater forms significantly beat placebo on 18-hour urinary excretion	Well tolerated over 12 weeks in older adults	Drink mixes and liquids; delivers ionic Mg2+, the direct substrate for both absorption routes
Threonate	8.3%	Weakest human evidence in the category — one small industry-affiliated RCT plus an open-label trial. Brain penetration is rodent-derived; no human study has measured brain or CSF magnesium after oral threonate	Fine	Niche cognitive positioning; a 2,000 mg dose delivers only ~144–168 mg magnesium
Taurate	8.9%	Essentially none. No published human RCTs. The "heart form" idea traces to a 1996 Medical Hypotheses paper — a rationale, by its own title	Well tolerated	A reasonable gentle form; treat cardiovascular claims as extrapolation
Sulfate (Epsom)	20.2% / 9.9%	Among the lower-bioavailability oral forms	Powerful osmotic laxative by mouth	Oral saline laxative. Bath absorption is not established
Orotate	7.3% / 6.6%	Small cardiac literature, single-center and methodologically criticized	Fine	Niche; lowest elemental density in common use, and expensive
Carbonate	28.8% / ~20–25%	Modest; converts to chloride in stomach acid, so absorption is acid-dependent — a real problem for PPI users and older adults	Laxative at dose	Antacid

Two places where marketing has outrun evidence deserve naming, because they are the two most confidently sold forms in the category. Threonate and the brain: the original work showing raised brain and CSF magnesium was done in rats and has never been replicated in humans, because nobody has measured human brain magnesium after oral threonate. Taurate and the heart: taurine has its own cardiovascular literature and magnesium has its own blood pressure literature, but no trial has tested the salt of the two. That is a syllogism, not a finding. The three forms most people actually choose between are compared in glycinate vs citrate vs malate.

Why magnesium loosens some people's stools

Not idiosyncrasy. Osmosis, and entirely predictable. Absorption is saturable and gets less efficient as dose rises. Whatever isn't absorbed stays in the lumen as an osmotically active solute, pulls water in, distends the bowel and accelerates transit. So three variables drive tolerance: total unabsorbed magnesium, the solubility profile of the salt, and whether magnesium is present as a free ion or a neutral chelate. Note that a soluble salt that isn't absorbed and an insoluble salt that never dissolves both end up in the same place. The Cochrane review of magnesium for cramps quantified it: minor GI events in 11–37% of magnesium recipients versus 10–14% of controls. The levers are staying under 350 mg supplemental daily, taking magnesium with food, splitting the dose, and not concentrating the whole load into one osmotically aggressive species.

We audited what eight popular electrolyte mixes actually deliver in The Magnesium Gap — the doses run from 0 mg to 220 mg. For the two closest matchups on magnesium specifically, see Current vs Magna and Current vs Nuun.

Why Can a Blend of Forms Beat a Single Form?

This is the intellectual core, and it rests on physiology almost no consumer page covers: magnesium is absorbed by two different routes with two different kinetics.

	Transcellular (active)	Paracellular (passive)
Route	Through the enterocyte	Between enterocytes, via tight junctions
Machinery	TRPM6 and TRPM7 channels; CNNM4 for basolateral exit	Claudin-16 and claudin-19
Site	Distal small intestine, cecum, colon	Jejunum and ileum
Kinetics	Saturable — there are only so many transporters	Non-saturable, concentration-driven, plus solvent drag
Dominant when	Luminal magnesium is low	Luminal magnesium is high; ~80–90% of absorption at typical intakes
Regulated by	1,25(OH)2D, EGF, insulin, fermentable fiber and SCFA, luminal pH	Concentration gradient, transit time, solubility

Now add the dose-response data. Fine, Santa Ana, Porter and Fordtran (Journal of Clinical Investigation, 1991) tested supplemental increments of 0, 10, 20, 40 and 80 mEq and found fractional absorption fell from 65% at the lowest intake to 11% at the highest. The curve was "almost perfectly represented by an equation containing a hyperbolic function plus a linear function" — a saturable component that maxes out, plus a non-saturable component picking up a roughly constant 7% of what passes through. That result is the whole argument: a large bolus of one form pushes on a route that is already saturated, and the surplus is left to passive flow or to the colon.

Three findings build on it. Solubility differs, and solubility predicts absorption. Blancquaert found dissolution across 15 products ranged from over 80% release in 10 minutes to more than 120 minutes to reach 80%; different salts dissolve at different pH and become available in different gut segments, and Dowley 2024 found a marine complex and a bisglycinate chelate behaved differently in gastric versus intestinal phases and fed versus fasted. A blend widens the window over which magnesium is present in absorbable form.

Different species plausibly engage different machinery. Ionic magnesium is the direct substrate for TRPM6/TRPM7 and for paracellular flux. Organic acid salts keep magnesium in solution longer and buffer luminal pH. Chelated bisglycinate is hypothesized to travel intact via peptide and amino-acid transporters, sidestepping competition for mineral channels — a strong hypothesis, widely repeated, and not demonstrated in healthy humans. It should be labeled that way wherever it appears, including here.

One direct demonstration exists. In Blancquaert's human phase, 30 randomized crossover participants took either a two-salt product delivering 196 mg elemental magnesium or a single-salt oxide product delivering 450 mg. Incremental serum AUC was 6.87 mM·min for the blend versus 0.31 mM·min for the single salt (p ≤ 0.011) — a 2.3-fold smaller blended dose outperforming by more than twentyfold. The authors' conclusion is the sentence this category should have memorized: "The solubility of a magnesium supplement is of greater relevance for in vivo bioavailability than the loading." The same study found bioavailability plateaued between one and two tablets, a modern confirmation of Fine's saturation curve.

The honest limitation

No published head-to-head randomized trial has compared a three-form magnesium blend against a single form at equal elemental dose with tissue-level endpoints. It does not exist. What exists is an inferential case from four independent lines: saturation kinetics, dissolution-bioavailability coupling, dual-pathway physiology, and one blend-beats-single-form comparison. That is a good mechanistic argument. It is not proof, and anyone claiming their blend is clinically proven superior is describing their marketing department. There is counter-evidence too: a 2019 mouse study found that splitting a high daily dose into two administrations did not meaningfully raise tissue magnesium. That tests split timing rather than split chemistry, and it is rodent data — but a careful reader should know it is there.

What Does Magnesium Need in Order to Work?

Everything above is about getting magnesium in. This is about what happens next, and it is the part the category ignores. Magnesium sits inside a web of dependencies: nutrients whose activation requires magnesium, and nutrients that decide whether magnesium is absorbed, retained or usable.

Vitamin D cannot be activated without magnesium

The most consequential and least-discussed relationship in the field. Uwitonze and Razzaque (Journal of the American Osteopathic Association, 2018) laid it out: every enzymatic step of vitamin D metabolism is magnesium-dependent. The liver's 25-hydroxylase (CYP2R1/CYP27A1), which makes 25(OH)D. The kidney's 1-alpha-hydroxylase (CYP27B1), which makes the active hormone 1,25(OH)₂D. The 24-hydroxylase that inactivates it. And the vitamin D-binding protein that carries it in blood.

Their words: "All of the enzymes that metabolize vitamin D seem to require magnesium." And the line with real consequences: "Magnesium supplementation was shown to markedly reduce the resistance to vitamin D treatment." A magnesium-deficient person can be functionally vitamin-D-resistant, and taking D without adequate magnesium may not produce the expected rise in 25(OH)D. The relationship runs back the other way too: 1,25(OH)₂D upregulates TRPM6 and stimulates intestinal magnesium absorption. A shortfall in either holds both down.

The nuance a careful reader deserves: Dai and colleagues (American Journal of Clinical Nutrition, 2018) ran a double-blind 2×2 factorial trial in 180 adults aged 40–85 with doses personalized to baseline intake, and found the effect bidirectional — magnesium raised 25(OH)D₃ in participants starting near 30 ng/mL and lowered it in those between 30 and 50, with interactions significant after Bonferroni correction. So magnesium isn't a vitamin D amplifier. It looks more like a regulator pushing the system toward homeostasis. Full treatment in vitamin D and why magnesium comes first.

Thiamine cannot become a coenzyme without magnesium

Thiamine is inactive as swallowed. It must be converted to thiamine pyrophosphate by thiamine pyrophosphokinase, which transfers a pyrophosphate group from ATP — a Mg-ATP-dependent reaction. Then the enzymes that use TPP, including pyruvate dehydrogenase and transketolase, require magnesium in their own right. A clean two-step dependency: magnesium to make the active form, magnesium again for it to do anything. The same logic runs across the B vitamins, nearly all of which are cofactor precursors requiring ATP-dependent activation. Which forms matter is covered in the eight B vitamins and their forms.

B6, P5P, and the claim that needs hedging

You will read everywhere that B6 helps magnesium get into cells. That claim is weakly sourced in humans, and the best test of it went the other way. Noah and colleagues (Magnesium Research, 2020) analyzed status data from a 264-person randomized trial: in participants with low baseline erythrocyte magnesium, magnesium alone raised RBC magnesium by 0.21 mmol/L (p = 0.0003) while magnesium plus B6 raised it by 0.13 mmol/L (p = 0.0233). Their conclusion: "vitamin B6 supplementation did not further increase magnesium levels." The uptake idea has rodent support — a 1989 rat study found tissue magnesium rose with pyridoxine specifically on low-magnesium diets — but not human support. The better-supported direction is the reverse: pyridoxal kinase, the enzyme converting B6 into its active coenzyme pyridoxal-5'-phosphate, is itself magnesium-dependent. Magnesium helps activate B6, not the other way round.

Which doesn't make the pairing pointless — its clinical evidence is the strongest in this article. Pouteau and colleagues (PLoS ONE, 2018) ran a Phase IV randomized, investigator-blinded, parallel-group trial in 264 adults over 8 weeks. Entry required a DASS-42 stress score above 18 and serum magnesium of 0.45–0.85 mmol/L, deliberately including the "normal but suboptimal" band. One arm took 300 mg elemental magnesium as magnesium lactate dihydrate; the other took the same plus 30 mg pyridoxine HCl.

The whole-population result is the part marketers omit: both arms improved enormously (−44.9% versus −42.4%) and the between-group difference of 0.72 points was not significant. In the pre-specified severe and extremely severe stress subgroup (DASS-42 ≥ 25, n = 162) it was: −16.36 points with Mg+B6 versus −13.20 with magnesium alone, a difference of 3.16 points (95% CI 0.50 to 5.82, p = 0.0203) — roughly 24% greater relative improvement, already significant at week 4. Treatment-related adverse events were lower in the combination arm (12.1% versus 17.4%). A real finding, in stressed people who were also short on magnesium, with no effect in the general population: a pharmacodynamic partnership, not a transport one. More in magnesium and B6.

Potassium: the relationship every nephrology textbook teaches

Two mechanisms at once. First the pump: Na⁺/K⁺-ATPase runs on Mg-ATP, and without magnesium it cannot hold the intracellular potassium gradient. Second the kidney: Huang and Kuo (Journal of the American Society of Nephrology, 2007) showed intracellular Mg²⁺ normally plugs the ROMK potassium channel in the distal nephron from the cytoplasmic side; when intracellular magnesium falls, that inhibition lifts and potassium secretion rises. A 2025 paper in The Journal of Physiology refined it: the effect requires activation of both ENaC and ROMK.

The clinical upshot is taught on every critical care rotation: hypokalemia that will not correct with potassium replacement is a classic sign of untreated magnesium deficiency. Correct the magnesium first. One accuracy note from Huang and Kuo themselves — magnesium deficiency alone doesn't necessarily cause hypokalemia; it generally requires concurrent elevated aldosterone or increased distal sodium delivery. Where these minerals sit relative to one another is mapped in electrolytes vs minerals vs trace minerals.

Boron helps you keep what you take in

Boron isn't glamorous and its literature is small, but it is real, and it comes from one careful program: Forrest Nielsen and colleagues at the USDA's Grand Forks Human Nutrition Research Center. In human depletion-repletion experiments, postmenopausal women and men over 45 ate a low-boron diet and were then repleted with 3 mg/day. Nielsen's synthesis in Environmental Health Perspectives (1994) reports that boron reduced plasma calcium and the urinary excretion of both calcium and magnesium, and that its effects were more pronounced when dietary magnesium was deliberately held low. Boron is a retention nutrient, not an absorption nutrient — it reduces what you lose in urine. Note the dose: the classic evidence uses 3 mg/day, against typical US dietary intake of roughly 1–1.5 mg.

Glycine and malate are not just packaging

Glycine is a genuine inhibitory neurotransmitter — an agonist at strychnine-sensitive glycine receptors and an obligatory co-agonist at the NMDA receptor's glycine-B site — plus a precursor to glutathione and collagen. Trials of 3 g before bed have improved subjective and polysomnographic sleep measures, apparently via NMDA receptors in the suprachiasmatic nucleus driving peripheral vasodilation and a drop in core temperature. The caveat that matters: 220 mg of magnesium delivered entirely as bisglycinate carries at most about 1.3 g of glycine, and far less if the magnesium comes from several forms. Glycine is a mechanistically aligned carrier, not a sleep dose — the same discipline we apply to magnesium and sleep generally in magnesium and sleep.

Malate is a real Krebs cycle intermediate, oxidized by malate dehydrogenase to oxaloacetate and central to the malate-aspartate shuttle. That biochemistry isn't marketing. The leap to fatigue relief is: the clinical basis is a 1995 crossover pilot in 24 people, and a 2019 review of seven systematic reviews found the evidence insufficient. Claim the biochemistry, not the outcome.

What not to pair, and one absorption claim to drop

Calcium. Rosanoff, Dai and Shapses (Advances in Nutrition, 2021) note that high calcium-to-magnesium intake ratios have been associated with increased risk across multiple chronic conditions, and that the US ratio has drifted toward 3:1 or higher as calcium fortification rose while magnesium intake fell. A proposed optimum sits nearer 2:1. Leaving calcium out of a magnesium product is a defensible choice, not an omission.

Vitamin C does not improve magnesium absorption. There is no credible evidence for it, and it gets asserted constantly. The enhancers with real support are vitamin D via TRPM6, adequate protein (absorption is lowest below about 30 g/day), and fermentable fibers such as inulin-type fructans, which raise short-chain fatty acid production, acidify the lumen and upregulate TRPM6. The broader logic of nutrient dependency is in nutrient cofactors and synergy.

How Current Is Formulated, and Why

This section is formulation rationale, not a superiority claim. Read it as design choices set against the evidence above.

Current delivers 220 mg of magnesium per stick — 52% of the Daily Value — from magnesium bisglycinate, magnesium malate, and ionic inland seawater concentrate. Alongside it: 2 mg of vitamin B6 as P5P, 25 mcg (1,000 IU) of vegan D3, 0.3 mg of boron, 330 mg of potassium, and six bioactive B vitamins (B1, B3, B5, B6, B9 as 5-MTHF, B12 as methylcobalamin — not a full B-complex; there is no riboflavin or biotin). No added calcium.

• Three forms, not one. A chelate with the best-documented GI tolerance, an organic acid salt whose carrier is a Krebs cycle intermediate, and an ionic marine source that arrives already dissociated as the direct substrate for TRPM6/TRPM7 and paracellular flux. Three solubility profiles across a wider pH and segment window.
• 220 mg, not 500 mg. That is 63% of the 350 mg supplemental upper limit, inside the range where fractional absorption is still favorable and below the dose where osmotic load dominates the experience.
• Vitamin D in the same serving, because every enzyme that metabolizes vitamin D is magnesium-dependent and activated D reciprocally upregulates magnesium absorption.
• Potassium in the same serving, because the Na⁺/K⁺-ATPase runs on Mg-ATP and intracellular magnesium gates ROMK.
• P5P rather than high-dose pyridoxine. P5P is the active coenzyme, skipping a magnesium-dependent activation step, and 2 mg sits far below the range where chronic pyridoxine raises neuropathy concerns. Note Pouteau used 30 mg of pyridoxine HCl, and no trial shows P5P outperforms it for stress.
• Boron at 0.3 mg is a dietary top-up, one-tenth of Nielsen's 3 mg research dose. It should not be presented as replicating that intervention.

What this is: a formulation argument grounded in mechanism, published absorption physiology and documented cofactor dependencies. What it is not: evidence that this product outperforms another, because no such comparison has been run — on this formula or, as far as the published literature goes, on any three-form magnesium blend. If that changes, we will write about it.

Is Magnesium Safe?

Magnesium from food has no upper limit in people with normal kidney function. The Tolerable Upper Intake Level applies to supplemental magnesium only: 350 mg/day for anyone 9 and older (65 mg ages 1–3, 110 mg ages 4–8). Why that limit exists changes how you read it. It isn't based on toxicity. The Institute of Medicine set it at the lowest dose where osmotic diarrhea was reliably observed — one of the very few ULs in the entire framework derived from a nuisance endpoint that reverses when you stop.

Real toxicity requires very large doses or impaired renal excretion. Symptoms begin above serum concentrations of roughly 1.74–2.61 mmol/L and progress from nausea and flushing through hypotension and lethargy to muscle weakness and loss of deep tendon reflexes. Most reported cases involve magnesium-containing laxatives or antacids taken by people with renal impairment or bowel obstruction, not ordinary supplements.

The single most important safety point: the kidney is the sole regulator of magnesium excretion. Anyone with chronic kidney disease, particularly an eGFR below 30 mL/min/1.73 m², should not take supplemental magnesium without physician supervision. Also check with a clinician if you have heart block or a bradyarrhythmia, myasthenia gravis, bowel obstruction, or you are pregnant. Separate magnesium from bisphosphonates, tetracyclines and fluoroquinolones by at least two hours and from levothyroxine by four. In the other direction: loop and thiazide diuretics increase magnesium loss, potassium-sparing diuretics reduce its excretion and can allow accumulation, long-term PPI use is associated with hypomagnesemia, and magnesium deficiency potentiates digoxin toxicity.

Myths Worth Retiring

• "Magnesium is involved in 300 reactions." Conservative and dated. Over 600 enzymes as cofactor, about 200 more as activator — and "800 enzymes" is those two figures illegitimately summed.
• "80% of Americans are magnesium deficient." Not supported. 48% consume less than the EAR from food: inadequate intake, not diagnosed deficiency.
• "Get a blood test." Serum holds under 1% of body magnesium and is defended by withdrawal from bone. The Magnesium Depletion Score is more informative for most people.
• "Magnesium oxide is worthless." As standalone repletion the data are unkind, but it is a legitimate laxative and antacid, and Blancquaert's best-performing product contained oxide blended with a soluble organic salt. Nuance beats absolutism.
• "Higher elemental percentage means a better product." Measured directly: a 196 mg blended product produced more than twenty times the serum AUC of a 450 mg oxide product.
• "Threonate raises brain magnesium." Demonstrated in rats. No human study has measured brain or CSF magnesium after oral threonate, and it has among the lowest elemental density of any form.
• "Taurate is the heart form." There are no human RCTs of magnesium taurate.
• "Magnesium fixes muscle cramps." The Cochrane review (11 trials, 735 participants) concluded magnesium is "unlikely to reduce the frequency or severity of muscle cramps in older adults," at moderate certainty. What the evidence does and doesn't say is in muscle cramps at night.
• "More is always better." Fractional absorption falls 65% to 11% with dose, bioavailability plateaus, and the serum magnesium–dementia relationship is U-shaped with an optimum near 0.85 mmol/L.
• "Epsom baths and magnesium oil replete you." No credible human evidence of meaningful transdermal absorption. Baths are relaxing. Different claim.
• "72 trace minerals make marine magnesium absorb better." The trace-mineral contribution has never been isolated in a controlled comparison. Solubility, comparable bioavailability and tolerability have been shown — claim those. More in sea minerals vs synthetic minerals.

Where the Evidence Is Weak

The best sobriety check in this field is Veronese and colleagues' umbrella review in the European Journal of Nutrition (2020), covering 16 meta-analyses and 55 independent outcomes. Only 12 of 36 RCT-derived outcomes and 9 of 19 observational outcomes reached p < 0.05. Class I evidence was awarded to exactly two things: reducing migraine intensity and frequency, and decreasing hospitalization risk in pregnant women. Everything else is weaker than the marketing suggests.

• Multi-form blends. No head-to-head RCT at equal elemental dose. The case is mechanistic.
• Bisglycinate's absorption route. The dipeptide-transporter hypothesis is plausible, widely repeated, and not demonstrated in healthy humans.
• Sleep. The best recent trial — 155 adults, 250 mg elemental magnesium as bisglycinate, double-blind and placebo-controlled, in Nature and Science of Sleep (2025) — found an Insomnia Severity Index change of −3.9 versus −2.3 for placebo (p = 0.049, Cohen's d = 0.2). Significant, clinically small, concentrated in participants with lower baseline dietary magnesium.
• Blood pressure. A 2025 meta-analysis in Hypertension pooling 38 RCTs and 2,709 participants found −2.81 mmHg systolic overall, with larger effects in medicated hypertensives and people with low magnesium status, and no significant effect in normotensive participants. The authors reported high heterogeneity and no detectable dose-response.
• Cognition. RCT evidence is insufficient to draw conclusions. The brain imaging work is cross-sectional with very small effect sizes, and the widely repeated "one year younger brain" headline is media framing, not a reported result.
• Mortality. A 2021 Advances in Nutrition analysis found higher dietary magnesium associated with reduced all-cause and cancer mortality but not cardiovascular mortality.
• Antioxidant claims. A 2025 meta-analysis found CRP was the only biomarker to reach significance; MDA, nitric oxide, total antioxidant capacity and glutathione were null.

The pattern is consistent, and it is the right way to think about magnesium: it behaves like a repletion intervention, not a drug. It does something measurable in people who are short. It does considerably less in people who aren't. That isn't a disappointing conclusion. It is why status matters more than dose.

Frequently Asked Questions

What does magnesium do for the body?

It acts as a cofactor for over 600 enzymes and an activator for roughly 200 more. Its largest role is making ATP usable, since ATP is enzymatically active only when chelated to magnesium. It also stabilizes DNA, powers the sodium-potassium pump, blocks the NMDA receptor at rest, and acts as a physiological calcium antagonist in muscle and blood vessels.

Which form of magnesium is best?

It depends on the goal. Citrate has the strongest human absorption evidence but is laxative at higher doses. Bisglycinate is the best tolerated. Malate is well tolerated with a metabolically active carrier. Ionic and marine forms are highly soluble, particularly with food. Oxide is 60% elemental magnesium and about 4% absorbed, making it a good laxative and a poor repletion form.

Is it OK to take magnesium every day?

For most healthy adults, yes, within the supplemental upper limit of 350 mg per day. That limit exists because of loose stools rather than toxicity, and it doesn't apply to magnesium from food. Anyone with reduced kidney function should not supplement without medical supervision, since the kidney is the only regulator of magnesium excretion.

Can a blood test tell me if I'm low on magnesium?

Only if you are very low. Serum holds under 1% of body magnesium and is defended by withdrawal from bone, so a normal result doesn't rule out depleted tissue stores. RBC magnesium is a better proxy but still imperfect. For most people the Magnesium Depletion Score — diuretic use, PPI use, heavy alcohol, kidney function — is more informative than a lab value.

Why doesn't 500 mg of magnesium citrate give me 500 mg of magnesium?

Because the label may be quoting the weight of the whole compound. Magnesium citrate is roughly 11–16% magnesium by weight depending on the salt and hydration state, so 500 mg of compound delivers between 55 and 81 mg of magnesium. The Supplement Facts panel must state elemental magnesium; the front of the box need not.

Does magnesium help with sleep?

Modestly, and mostly in people whose intake is low. The best recent randomized trial found an Insomnia Severity Index improvement of −3.9 versus −2.3 for placebo (p = 0.049) with a small effect size, and the benefit was greater in participants with lower baseline dietary magnesium. Magnesium is not a sedative.

Should I take magnesium with vitamin D?

The biochemistry favors it. Every enzyme that metabolizes vitamin D — 25-hydroxylase, 1-alpha-hydroxylase and 24-hydroxylase — is magnesium-dependent, as is the vitamin D-binding protein, and activated vitamin D in turn upregulates intestinal magnesium absorption. One randomized trial found magnesium's effect on 25(OH)D was baseline-dependent, raising it in people starting near 30 ng/mL and lowering it in those starting higher.

What should you not take magnesium with?

Separate magnesium from bisphosphonates, tetracycline and fluoroquinolone antibiotics by at least two hours, and from levothyroxine by at least four, since magnesium reduces their absorption. Very high supplemental zinc (around 142 mg/day, far above normal doses) interferes with magnesium balance. Large added calcium doses are also questionable, given evidence that high calcium-to-magnesium ratios track with worse outcomes.

Does magnesium stop muscle cramps?

Probably not. A Cochrane review of 11 trials and 735 participants concluded magnesium is unlikely to reduce the frequency or severity of muscle cramps in older adults, at moderate certainty, and results for pregnancy-related leg cramps were inconsistent. Magnesium has a real role in muscle relaxation physiology, but that mechanism hasn't translated into cramp relief in trials.

Can low magnesium make you tired?

Mechanistically, magnesium is required for ATP to be usable at all, so a genuine shortfall constrains energy metabolism. That said, fatigue has many causes, and no trial has shown magnesium improves energy in people who are already replete. Persistent fatigue is worth investigating with a doctor rather than self-treating.

Sources

1. NIH Office of Dietary Supplements. "Magnesium: Fact Sheet for Health Professionals." Link
2. Linus Pauling Institute. "Magnesium." Oregon State University. Link
3. de Baaij JHF, et al. "Magnesium in Man." Physiological Reviews, 2015. Link
4. Fiorentini D, et al. "Magnesium: Biochemistry, Nutrition, Detection." Nutrients, 2021. Link
5. Fine KD, et al. "Intestinal absorption of magnesium from food and supplements." J Clin Invest, 1991. Link
6. Firoz M, Graber M. "Bioavailability of US commercial magnesium preparations." Magnes Res, 2001. PubMed
7. Walker AF, et al. "Mg citrate found more bioavailable than other Mg preparations." Magnes Res, 2003. PubMed
8. Blancquaert L, et al. "Predicting and Testing Bioavailability of Magnesium Supplements." Nutrients, 2019. Link
9. Kappeler D, et al. "Higher bioavailability of magnesium citrate compared to oxide." BMC Nutrition, 2017. Link
10. Dowley A, et al. "Bioaccessibility and tolerability of marine-derived magnesium and calcium." Methods, 2024. PubMed
11. McFarlin BK, et al. "Comparing Two Seawater-Derived Magnesium Preparations." J Med Food, 2025. PubMed
12. "Nutritional Factors Affecting Magnesium Bioavailability." Biol Trace Elem Res, 2025. Link
13. Costello RB, et al. "The Case for an Evidence-Based Reference Interval for Serum Magnesium." Adv Nutr, 2016. PubMed
14. Rosanoff A, et al. "Suboptimal magnesium status in the United States." Nutr Rev, 2012. PubMed
15. "Chronic latent magnesium deficiency in apparently healthy university students." 2014. PubMed
16. Fan L, et al. "Magnesium Depletion Score Predicts Risk of Systemic Inflammation and Cardiovascular Mortality." J Nutr, 2021. PMC
17. "Magnesium depletion score and mortality in US adults with diabetes." PLoS ONE, 2024. Link
18. "Association of Magnesium Depletion Score With Cardiovascular Disease and Mortality." JAHA, 2024. Link
19. Pickering G, et al. "Magnesium Status and Stress: The Vicious Circle Concept Revisited." Nutrients, 2020. Link
20. Joris PJ, et al. "Long-term magnesium supplementation improves glucocorticoid metabolism." Clin Endocrinol, 2021. PubMed
21. Massey LK, Whiting SJ. "Caffeine, urinary calcium, calcium metabolism and bone." J Nutr, 1993. PubMed
22. Uwitonze AM, Razzaque MS. "Role of Magnesium in Vitamin D Activation and Function." JAOA, 2018. Link
23. Dai Q, et al. "Magnesium status and supplementation influence vitamin D status and metabolism." Am J Clin Nutr, 2018. PubMed
24. Pouteau E, et al. "Superiority of magnesium and vitamin B6 over magnesium alone on severe stress." PLoS ONE, 2018. Link
25. Noah L, et al. "Impact of magnesium supplementation with vitamin B6 on stress and magnesium status." Magnes Res, 2020. PubMed
26. Huang CL, Kuo E. "Mechanism of hypokalemia in magnesium deficiency." JASN, 2007. PubMed
27. Maeoka Y, et al. "Hypomagnesaemia-associated hypokalaemia requires activation of both ENaC and ROMK." J Physiol, 2025. PubMed
28. Nielsen FH. "Biochemical and physiologic consequences of boron deprivation in humans." Environ Health Perspect, 1994. PMC
29. Rosanoff A, Dai Q, Shapses SA. "Supplements Containing Calcium and Magnesium and Their Ratio." Adv Nutr, 2021. PubMed
30. Veronese N, et al. "Magnesium and health outcomes: an umbrella review." Eur J Nutr, 2020. Link
31. Chen F, et al. "Magnesium and Cognitive Health in Adults." Adv Nutr, 2024. PubMed
32. Argeros Z, et al. "Magnesium Supplementation and Blood Pressure." Hypertension, 2025. PubMed
33. "Magnesium Bisglycinate Supplementation in Healthy Adults Reporting Poor Sleep." Nat Sci Sleep, 2025. PubMed
34. Garrison SR, et al. "Magnesium for skeletal muscle cramps." Cochrane Database Syst Rev, 2020. Cochrane
35. "Molecular mechanism of calcium permeability and magnesium block in NMDA receptors." Nat Neurosci, 2026. Link

Life's better mineralized.

These statements have not been evaluated by the Food and Drug Administration. This product is not intended to diagnose, treat, cure, or prevent any disease. This article is for educational purposes and is not medical advice. Talk to your doctor before starting any supplement, especially if you are pregnant, nursing, have kidney disease, or take prescription medication.