Magnesium — Body Handbook

Supplements · §521

Magnesium

About half of US adults eat less magnesium than their body needs, and the shortage rarely lands as a felt crisis. It shows up sideways — sleep that takes a bit longer to come, a blood pressure number that drifts up across a decade, more migraine days if you're prone, a mood that's harder to keep level, bowels that don't quite move. Hitting your daily target from food first, then a cheap supplement in the right form, closes most of that gap. The trap: the five common forms — glycinate, citrate, oxide, threonate, malate — do meaningfully different things, and picking the wrong one is the most common reason people decide magnesium "didn't work for me."

Do · Daily Evidence Moderate Chapter Supplements

The wins are quiet but real. Falling asleep around 17 minutes faster if you were sleeping badly (Mah & Pitre 2021); a small blood pressure drop within three months if you were running high (Argeros et al. 2025); bowels that move on time; fewer migraine days if you get them. None of it transforms a Tuesday. Eat magnesium-rich food where you can, supplement where you can't, and match the form to the job.

Magnesium is the most-used mineral cofactor in your cells. More than 600 enzymes need it to work, every reaction that burns ATP needs it, and it sits in the doorway of two key brain receptors (Workinger et al. 2018). One of those enzymes activates vitamin D, which is why a vitamin D supplement can quietly underdeliver if your magnesium is low. At rest, a magnesium ion plugs the NMDA channel that lets calcium into excited neurons; when you're short on magnesium, that door doesn't shut as well, and neurons stay revved. At the same time, magnesium makes GABA-A receptors more responsive — the same target benzodiazepines hit. That dual move quiets the nervous system, which is why people fall asleep faster and feel less wound up once their levels come up.

In the vascular bed, magnesium is a natural calcium antagonist: it relaxes the smooth muscle wrapped around your arteries, the same family of mechanisms behind the calcium-channel blockers a cardiologist might prescribe for high blood pressure. That's the engine behind the blood pressure effect. In the gut, the story is different — whatever magnesium doesn't get absorbed sits in the lumen and pulls water in by osmosis, softening stool at low doses and producing diarrhoea at high ones. That's why magnesium oxide is the cheap OTC laxative and why magnesium citrate is what colonoscopy prep tastes like.

What the evidence actually says

The picture differs by effect. Blood pressure has the cleanest data: a 2025 meta-analysis in Hypertension pulled together 38 randomized trials and 2,709 people. People with already-high blood pressure on medication saw the biggest drop; people whose blood pressure was already normal saw essentially nothing (Argeros et al. 2025). That's the pattern across most of the magnesium literature — bigger benefit if you were running low, smaller benefit if you weren't.

Sleep is real but modest. A meta-analysis of three trials in older adults with insomnia found people fell asleep about 17 minutes sooner on magnesium than on placebo (Mah & Pitre 2021). The largest modern trial — 155 healthy adults sleeping badly, given 250 mg of magnesium glycinate nightly for four weeks — found their insomnia severity dropped more on magnesium than on sugar pills, but the gap was small (Held et al. 2025). Mood: across seven trials in people with depression, magnesium beat placebo on standard depression scores (Moabedi et al. 2023). Migraine: the American Academy of Neurology rates it probably effective for prevention (Holland et al., AAN/AHS 2012). Type 2 diabetes: pooled trials show modest reductions in fasting blood sugar (Veronese et al. 2016).

Two negative results worth naming. Cochrane looked at magnesium for muscle cramps across 11 trials and concluded it almost certainly doesn't help the kind of leg cramp most older adults complain about (Garrison et al. 2020). No good trials exist on exercise cramps either way. And the long-term mortality picture, while consistent across cohorts of more than a million people — lower stroke, lower heart failure, lower all-cause death as intake rises from 150 to about 400 mg/day (Fang et al. 2016) (Bagheri et al. 2022) — comes from observational data. Magnesium-rich food is also fibre-rich, less processed, and usually paired with a healthier overall diet, so some of the credit belongs to the diet around it.

Pick the form that matches the job

Adult daily targets: 420 mg/day for men, 320 mg/day for women (NIH ODS 2022). Food first — an ounce of pumpkin seeds delivers 156 mg, a cooked cup of spinach about 157 mg, an ounce of almonds 80 mg. Black beans, cashews, dark chocolate, salmon all bring real numbers too. About a third of what you eat actually gets absorbed; the rest gets eliminated. Some of the magnesium in whole grains and legumes stays locked up by phytic acid, too — soaking or fermenting them frees a little more. If your daily plate puts you near the target, you don't need a pill.

If it doesn't — and for most American adults it doesn't (NIH ODS 2022) — the form decides what you get:

Glycinate (or bisglycinate) — the default daily form. Well absorbed, doesn't loosen stool, gentle. 200–400 mg of elemental magnesium in the evening if your main goal is sleep, stress, or general topping-up. This is the form behind the 2025 sleep trial (Held et al. 2025).
Citrate — well absorbed, mildly stool-softening. The work-horse choice if you also want a friendlier bowel. 200–400 mg/day. Walker found it more bioavailable than oxide or amino-acid chelate over two months (Walker et al. 2003). At pharmacy-bottle doses (a full 296 mL bottle of citrate solution), it's the standard osmotic laxative — kicks in inside six hours.
Oxide — cheap, lots of magnesium per pill, but only a few percent of it actually gets in. The form used in most migraine prevention trials at 400–500 mg/day (Holland et al. 2012). Also sold as Milk of Magnesia (chemically similar) for constipation.
L-threonate — the only form shown to push brain magnesium up in animals, on the strength of which a few small human trials report memory and reaction-time improvements (Slutsky et al. 2010) (Danielson et al. 2022). Typical dose 1.5–2 g/day of Magtein, which delivers only ~144 mg of elemental magnesium — too little to fix a body-wide deficit, so it's an add-on to a standard form, not a replacement. Costs roughly 5× as much.
Malate — marketed for fatigue and fibromyalgia; the human evidence is thin. Reasonable if you tolerate it well, but no edge over citrate or glycinate on what's been measured.

Split the dose across the day if you're taking more than 300 mg — the gut absorbs it better in halves than as one bolus. Take with food.

Two timing notes for people on common medications. Magnesium binds to certain antibiotics (tetracyclines, ciprofloxacin and its cousins) and to bisphosphonates for osteoporosis, blocking their absorption — separate the doses by two to four hours. And if you're on a proton pump inhibitor (omeprazole, esomeprazole, the rest of the family) plus a diuretic, you're at meaningfully higher risk of low magnesium and may need monitoring rather than just a supplement (Kieboom et al. 2015).

What most magnesium guides get wrong

"My blood test was normal, so I'm fine." Probably not. Only about 1% of the magnesium in your body floats in your blood; the rest sits inside cells and bone. Your kidneys defend the blood number aggressively — pulling magnesium out of bone to keep it in range — long after the rest of you has run low. A normal serum magnesium does not rule out a real shortage; what does is a long-term diet that hits the daily target (Workinger et al. 2018) (DiNicolantonio et al. 2018).

"Any magnesium will do." Form matters more than the supplement aisle advertises. Magnesium oxide is mostly an osmotic laxative with a low fraction actually absorbed; taking it for anxiety mostly produces loose stools. Glycinate is gentle but doesn't move bowels. Threonate gets to your brain in rats but is underdosed for body-wide repletion. Match the form to the goal — see Pick the form that matches the job above.

"More is better." The cap on supplemental magnesium is 350 mg a day from pills (food doesn't count toward this), and the limiting factor is diarrhoea, not toxicity (NIH ODS 2022). Some experts argue the cap is outdated and could safely be higher (Costello et al. 2023), but for most people the body tells you when you've gone too far long before anything dangerous happens.

Why people decide it didn't work

Wrong form for the goal. Taking oxide for anxiety. Taking glycinate for constipation. Taking threonate for body-wide repletion.
Underdosed. The trials that show effects use 200–500 mg of elemental magnesium. A bottle labelled "500 mg magnesium glycinate" usually means 500 mg of the whole compound — only about 100 mg of which is the magnesium itself. Read the elemental number on the back.
Already replete. If your diet already covers the daily target, an extra pill mostly does nothing. The strongest effects across the literature show up in people whose intake was low or whose magnesium status was depleted; people who were fine to begin with don't notice much.
Something is depleting you faster than the pill replaces it. Long-term proton pump inhibitor use combined with a diuretic is the textbook offender (Kieboom et al. 2015). Heavy alcohol does the same through urine loss. Uncontrolled diabetes drives magnesium out through the kidneys too.
Whole dose at once. Past about 300 mg in a single sitting, more of the dose hits the bowel as a laxative instead of the bloodstream. Split it.

What chronic shortage looks like across a decade

Magnesium shortage isn't dramatic. It's the version of you that lies in bed for an extra fifteen minutes most nights and chalks it up to "I just don't sleep great." It's the blood pressure cuff at your annual physical reading 132/85 instead of 122/78, a number your doctor flags but you mostly forget about. It's the third Friday-evening migraine of the month if you're prone, the kind of mood week that sits a little heavier than it should, the bowel pattern that needs coffee plus a particular morning ritual to work.

Stretch that across a decade and the cohort data starts to bite. Across more than a million people followed for years, the bottom of the magnesium intake range carries higher rates of stroke, heart failure, type 2 diabetes, and earlier death than the top of the range (Fang et al. 2016) (Bagheri et al. 2022). None of that is the kind of thing you notice in your forties. It's the kind of thing your sixties hand you a bill for.

What changes when you fix it

First two weeks. If you were running low, the bedside experience changes first. You're closing your eyes and you're out, instead of lying there cycling through tomorrow's meeting (Held et al. 2025). The morning bathroom visit lands on time without a fight. If you'd been taking citrate, the second part is on day three.

One to three months. If your blood pressure was running high, the numbers at your next check shift down a few points — not enough to feel, but enough that a partner who watches your home cuff notices the trend (Argeros et al. 2025). If you get migraines, the month before the appointment may have one or two fewer days lost to a dark room (Holland et al. 2012). If you've been carrying a low mood and your magnesium intake was poor, the floor of it lifts a little — your partner stops asking what's been off for the last few weeks (Moabedi et al. 2023).

Across the years. The cohort data suggests this is the part you don't feel and won't be able to attribute. Hitting the daily target across decades sits inside the cluster of choices that decides whether your sixties are the active version or the cardiologist-visit version. Magnesium is one tile in that picture, not the whole picture; it's also one of the cheapest tiles to lay (Fang et al. 2016).

For someone whose diet was already covering the target, the payoff is honest: small. The point isn't to push higher than adequate; it's to stop being short.

If you got here because of bad sleep, magnesium is one lever; the bigger ones are a dark cool bedroom, a consistent wake time, and ruling out sleep apnea. If you got here because of blood pressure, the dietary potassium and sodium balance and aerobic exercise move more numbers than magnesium does. If migraines are the question, riboflavin and CoQ10 also hold AAN/AHS endorsements for prevention. And if you're already on a proton pump inhibitor and a diuretic — that combination is worth a separate look on its own.

Related in the handbook

Worsens / aggravates

— Some of the magnesium in whole grains and legumes stays locked up by phytic acid; soaking or fermenting frees it.
— Long-term acid-blocker use is a recognized cause of low magnesium; if you're on a PPI for years, test it.

Helps

— The same magnesium that loosens stool at high doses — citrate or oxide — is a genuinely useful tool for sluggish bowels.
— Magnesium is rated probably-effective for migraine prevention — a cheap first thing to try if you get them.
— If your legs crawl at night, low magnesium is one of the checkable causes — worth a trial before stronger drugs.

Alternatives

— If you're eyeing ashwagandha for sleep, plain magnesium is the cheaper, better-proven first move.
— Magnesium is one of the low-cost migraine preventives people try before prescription CGRP inhibitors.

— Magnesium pairs with CoQ10 as the evidence-backed supplement duo for cutting migraine frequency.
— If potassium stays low despite a good diet, low magnesium is often the hidden reason; fixing one helps the other stick.
— Magnesium is a cofactor your body uses to activate vitamin D — low magnesium can quietly limit the benefit of a D supplement.
— Boron is a quiet partner here — it slows how fast magnesium leaves the body.
— Magnesium is one leg of the electrolyte balance that the Western diet runs short on.
— Magnesium nudges blood pressure down if you were running high — track it at home to see whether it's doing anything.
— Magnesium shows up in most sleep stacks; the effect is small but defensible for occasional use.
— If you're taking K2 and D for your bones, magnesium is the third leg most people are quietly short on.
— Drinking reverse-osmosis water removes the magnesium hard tap quietly gave you — remineralizing the water is one small way to top it back up.
— Soaking in Epsom salt won't meaningfully raise your magnesium; if you're low, that's a job for diet or a cheap pill, not a float.

Substance and claimed effects

Magnesium is the fourth most abundant cation in the human body and an obligatory cofactor in over 600 enzymatic reactions — including every reaction that uses ATP, every step of DNA and protein synthesis, parathyroid hormone activity, vitamin D activation, and ion-channel gating in nerve and muscle (Workinger et al. 2018; NIH ODS 2022). Roughly 60% of body magnesium sits in bone, 27% in muscle, ~12% in soft tissue, and less than 1% in serum, where it is tightly defended by the kidneys until depletion is severe (Workinger et al. 2018). The catalogue's scope here is dietary plus supplemental magnesium — RDA 420 mg/day for men, 320 mg/day for women — and the consequences that follow from raising intake from below-RDA toward sufficiency: sleep onset and architecture, blood pressure, mood and anxiety, bowel function, muscle / cramp behaviour, glucose handling, and long-term cardiovascular and all-cause mortality. Form selection (glycinate, citrate, oxide, threonate, malate) modifies which of those consequences is hit hardest and which side effects appear. Acute intravenous magnesium for eclampsia, torsades, and severe asthma is out of scope; this is the oral / dietary supplement reading.

Evidence by addressing question

mechanism

Three mechanism strands carry most of the effects covered. First, magnesium gates the NMDA glutamate receptor: a Mg²⁺ ion sits in the channel pore at resting membrane potential and blocks calcium influx until depolarization removes it. Low intracellular magnesium leaves the channel chronically permissive, raising neuronal excitability — the biophysical basis for migraine prophylaxis, the proposed mechanism for the anxiolytic effect, and a contributor to insomnia (Workinger et al. 2018; Holland et al., AAN/AHS 2012). Second, magnesium is a positive allosteric modulator of GABA-A receptors, the same target as benzodiazepines; raising intracellular Mg amplifies inhibitory tone, lowering sleep latency and stress reactivity. Third, magnesium is a physiological calcium antagonist at the smooth-muscle level — it relaxes vascular smooth muscle (lowering systemic vascular resistance, the BP mechanism) and modulates skeletal-muscle contraction (the cramp story).

Bowel effects are osmotic, not receptor-mediated: unabsorbed magnesium in the gut lumen pulls water in, producing softer stools at low doses and outright diarrhoea at high doses. This is dose- and form-dependent — magnesium oxide is poorly absorbed and largely retained in the lumen (the classical laxative form), whereas glycinate is absorbed well and rarely loosens stool (Workinger et al. 2018; Walker et al. 2003).

evidence

Sleep. A 2021 systematic review of three randomized trials in older adults with insomnia (n=151) found magnesium supplementation reduced sleep onset latency by 17.4 minutes versus placebo (95% CI −27.3 to −7.4, p=0.0006); total sleep time gained 16 minutes but did not reach statistical significance, and the authors graded the overall evidence quality as low to very low because of small samples and high risk of bias (Mah & Pitre 2021). The largest single-trial signal is Abbasi et al. 2012: 46 elderly subjects, 500 mg/day for 8 weeks, with significant improvement in PSQI, sleep latency, total sleep time, serum melatonin, and morning cortisol — but published in a small regional journal, no allocation concealment, and the magnesium-bisglycinate vehicle delivered ~1.5 g of glycine alongside, so the magnesium-specific effect is confounded. Held et al. 2025 is the largest placebo-controlled trial to date: 155 adults with self-reported poor sleep, magnesium bisglycinate 250 mg elemental Mg nightly for 4 weeks, Insomnia Severity Index improved 3.9 vs 2.3 points (p=0.049, Cohen's d ≈ 0.2 — small effect). Sleep efficiency and nighttime awakenings improved in the magnesium arm. Net: real but modest signal, larger in deficient older populations than in supplemented adequately-fed adults.

Blood pressure. The 2025 Hypertension meta-analysis pooled 38 RCTs (n=2,709, median dose 365 mg/day, median 12 weeks) and reported pooled reductions of systolic BP −2.81 mmHg (95% CI −4.32 to −1.29) and diastolic −2.05 mmHg (95% CI −3.23 to −0.88). Effect was much larger in hypertensives on BP-lowering medication (systolic −7.68 mmHg) and in subjects with documented hypomagnesemia (systolic −5.97 mmHg); in normotensives, the effect did not reach significance (Argeros et al. 2025). Mechanism (vascular-smooth-muscle calcium antagonism) and dose-response are both consistent. This is the entry's strongest evidence-quality block.

Mood / depression. Moabedi et al. 2023 pooled 7 RCTs (n=325) in adults with depressive disorders and reported a standardized mean difference of −0.92 (95% CI −1.44 to −0.40, p=0.001) on depression scores — a moderate-to-large effect, though sample sizes are small and at least one trial failed to find benefit when magnesium was added to fluoxetine. Boyle et al. 2017 reviewed magnesium for subjective anxiety and stress: signals exist but trials are heterogeneous and small. Evidence is suggestive rather than settled.

Cardiovascular / mortality. Dose-response meta-analysis of 40 prospective cohorts (>1 million participants) showed inverse associations between dietary magnesium and stroke, heart failure, type 2 diabetes, and all-cause mortality; greatest risk reduction landed between 150 and 400 mg/day of intake (Fang et al. 2016). A 2022 meta-analysis covering 19 cohorts (n≈1.17 million, 52,378 all-cause and 23,478 CV deaths) confirmed the inverse mortality association (Bagheri et al. 2022). These are observational — confounded by overall dietary quality (magnesium-rich foods are also fibre-rich and minimally processed) — but the dose-response, the biological plausibility from BP / glucose / arrhythmia mechanisms, and replication across cohorts make the signal robust.

Migraine. The AAN/AHS 2012 guideline rates magnesium Level B (probably effective) for episodic migraine prevention, based on multiple positive RCTs typically dosing 400–600 mg/day of magnesium oxide or citrate (Holland et al. 2012). The American Migraine Foundation continues to endorse this rating.

Type 2 diabetes / glycemic control. Veronese et al. 2016 meta-analysed 18 double-blind RCTs and reported a significant reduction in fasting plasma glucose in people with or at risk of diabetes; effect on HbA1c was favourable but smaller. Effect is largest in subjects with low baseline magnesium status.

Muscle cramps. The 2020 Cochrane review (11 trials, n=735) concluded magnesium is unlikely to provide clinically meaningful prophylaxis for idiopathic skeletal-muscle cramps in older adults at any dose tested (moderate-certainty evidence) — and no RCTs exist for exercise-associated cramps in athletes (Garrison et al. 2020). The exception, not directly trialled but clinically obvious: cramps in documented hypomagnesemia resolve when the deficit is corrected. So "magnesium for cramps" is true for the small deficient minority and false for the average eucalcaemic adult — a separation the supplement industry largely ignores.

Cognition (threonate). The substrate study is Slutsky et al., Neuron 2010: magnesium L-threonate raised brain magnesium concentration in rats (other forms do not cross the blood-brain barrier efficiently), increased hippocampal synaptic density, and improved learning and memory. Human trials are early. Danielson et al. 2022: 109 Chinese adults 18–65, 2 g/day Magtein vs placebo, improvements across all five Clinical Memory Test subscales. A 2025 Frontiers in Nutrition trial reported improvements in working and episodic memory and a 7.5-year reduction in estimated brain cognitive age in healthy young adults; both human trials carry industry-affiliation footnotes (Magtein is a single-source patented ingredient).

protocol

Three patterns separate by form (Workinger et al. 2018; Walker et al. 2003):

Glycinate / bisglycinate — high bioavailability (~80% absorption typical estimates), low GI loosening, the standard "any-purpose" daily form. Typical daily dose 200–400 mg elemental Mg/day, taken evening for sleep and stress benefit. The form used in the 2025 Held trial.
Citrate — well absorbed (~30–40%), mild osmotic effect at higher doses; the work-horse form for people who want both magnesium status and gentle bowel regularity. Typical 200–400 mg/day. Walker et al. found citrate more bioavailable than amino-acid chelate or oxide over 60 days.
Oxide — high elemental Mg per pill (60%) but low fractional absorption (Firoz estimate ~4%; Ranade ~23%); cheap, the form behind most migraine RCTs at 400–500 mg/day; also the form sold OTC for constipation (Milk of Magnesia is magnesium hydroxide, the chemical cousin) where the unabsorbed fraction is the point.
Citrate (high dose) — at 5–10 oz of oral solution, magnesium citrate is a clinical osmotic laxative producing a bowel movement within 30 min to 6 hr (used as bowel prep).
L-threonate — Magtein, the patented form; expensive (~$30+/month). Crosses the blood-brain barrier in animal data; trials at 1.5–2 g/day (~144 mg elemental Mg) — too little to correct systemic deficiency. Reasonable as a cognition-targeted addition on top of a standard form, not a replacement.
Malate — modest absorption; the rat data in Workinger et al. 2018-cited literature suggest preferential muscle / brain accumulation. Marketed for fibromyalgia / fatigue; human evidence is thin.

Practical defaults the dossier supports: for a generic adult who eats limited green vegetables and nuts and wants sleep / stress / BP coverage, glycinate 200–400 mg with dinner. For chronic constipation, citrate. For migraine prophylaxis, oxide 400 mg/day per the AAN-cited trials. For cognition-specific use, threonate as an add-on. Food-first is always the better lever — almonds (80 mg/oz), pumpkin seeds (156 mg/oz), spinach cooked (~157 mg/cup), black beans, cashews, dark chocolate (NIH ODS 2022).

contraindications

Renal function is the gatekeeper. Magnesium is cleared by the kidneys; CKD stages 4–5 lose the ability to excrete excess, and supplemental Mg can accumulate to dangerous levels (hypermagnesemia causes hypotension, bradycardia, neuromuscular depression). Standard advice: consult a nephrologist before supplementing in CKD (NIH ODS 2022). Magnesium also interacts with several drug classes — bisphosphonates and tetracycline / quinolone antibiotics (chelation; separate dosing by 2–4 hours), and high-dose loop / thiazide diuretics increase urinary magnesium loss (the "second hit" with PPIs — see below). Pregnancy: dietary and standard supplemental doses are safe; magnesium sulfate IV is itself the standard eclampsia treatment.

misconceptions

Three common errors:

"Take any magnesium for any benefit" — form matters. Oxide for sleep is mostly osmotic effect with little systemic delivery; glycinate for constipation is the wrong tool; threonate for sleep is underdosed.
"Serum magnesium rules out deficiency" — only ~1% of body magnesium is in serum; the kidneys defend it aggressively. A normal serum can sit on top of substantial intracellular depletion. RBC magnesium is a more sensitive marker but rarely ordered; intake assessment is often the practical proxy (Workinger et al. 2018; DiNicolantonio et al. 2018).
"More is better" — the supplemental UL is 350 mg/day elemental Mg from supplements, not the daily total intake; the limiting factor is osmotic diarrhoea, not toxicity. The Council for Responsible Nutrition's 2023 reanalysis argues the human data supports raising the UL to 500 mg supplemental/day; Costello et al. 2023 formally calls for re-evaluation. Food magnesium has no UL because the kidneys eliminate excess.

failure-modes

Common reasons "magnesium didn't work for me":

Wrong form for the goal (oxide expected to treat anxiety; glycinate expected to fix constipation).
Underdosed — taking 100 mg when the trial dose was 400 mg.
Already replete — additional magnesium in someone hitting RDA from diet produces little benefit on BP, mood, or sleep, per the consistent subgroup pattern across meta-analyses (the strongest effects come in deficient or hypertensive populations).
Ongoing depletion — chronic PPI use plus a diuretic depletes magnesium faster than oral supplementation replaces it; the FDA issued a 2011 safety communication for PPI-associated hypomagnesemia, and the effect is concentrated in patients also on diuretics (Kieboom et al. 2015). Heavy alcohol use depletes magnesium through urinary loss and reduced GI absorption.
Tolerance / GI ceiling — split doses produce better absorption than single boluses; once-daily 500 mg oxide produces more diarrhoea than 2 × 250 mg with meals.

practicalities

All forms are OTC and inexpensive (the exception is threonate at ~$30–60/month). Generic citrate at supermarket prices is under $0.10/day; glycinate runs $5–15/month for a daily dose. The supplement industry sells novel forms (taurate, orotate, ascorbate) at premium prices with negligible head-to-head bioavailability evidence over citrate or glycinate. Food magnesium per dollar dominates — a $0.30 ounce of pumpkin seeds delivers 156 mg, roughly half the female RDA.

stakes

The stakes are subtle. The headline effect of chronic mild magnesium insufficiency isn't a felt crisis — it's a slow accumulation of suboptimal vascular tone, slightly worse sleep architecture, modestly higher migraine vulnerability, and over decades a higher mortality hazard. Fang et al. 2016 and Bagheri et al. 2022 both show inverse dose-response with all-cause and cardiovascular mortality. Acute symptomatic hypomagnesemia (arrhythmia, tetany, seizures) is a hospital problem and rare; chronic subclinical insufficiency is common and shapes baseline metabolic and vascular health over decades.

payoff

For someone genuinely deficient or hypomagnesemic (low intake plus a depleting drug or condition), payoff is meaningful and fast: better sleep latency within weeks (Abbasi et al. 2012; Held et al. 2025), measurable BP reduction within 12 weeks (Argeros et al. 2025), reduced migraine frequency by ~12 weeks (Holland et al. 2012), softer bowel function within days (citrate dose), modest mood improvement in 4–8 weeks if baseline status was poor (Moabedi et al. 2023). For someone already replete via diet, payoff is small to nil. The long-term mortality signal is real but not perceptible week-to-week.

audience

Subgroups where the entry's relevance jumps:

Adults > 65. 83% of US adults ≥65 fail to meet the RDA from diet (NHANES 2005–2016); absorption declines, intake declines, drug-related losses rise (NIH ODS 2022).
People on chronic PPIs. Especially when combined with diuretics — the Kieboom et al. 2015 Rotterdam cohort showed OR 1.54 for hypomagnesemia in concomitant users.
People with type 2 diabetes / insulin resistance. Both for the small glycemic benefit (Veronese et al. 2016) and because hyperglycemia drives urinary magnesium loss.
Migraineurs. Level B AAN/AHS rating; specifically the oxide form at trial doses.
Heavy drinkers and people on chronic loop diuretics. Both deplete magnesium aggressively.

The credibility range

The optimist case. Magnesium is the most underappreciated single nutrient intervention in mainstream medicine. Subclinical deficiency is prevalent (~48% of US adults below EAR; NIH ODS 2022; Rosanoff et al. 2022), serum testing misses it routinely, and the downstream costs are everywhere the catalogue's readers feel them — sleep latency, blood pressure trajectory, migraine, mood, glucose handling, cardiovascular mortality. The dose-response meta-analyses on mortality, the AHA-journal BP meta-analysis with named subgroup effects, the AAN/AHS Level B for migraine, and the NMDA / GABA mechanism story all point the same direction. DiNicolantonio et al. 2018 argues subclinical deficiency is a principal driver of cardiovascular disease. The intervention is cheap, OTC, well tolerated within the UL, and the worst common side effect is loose stools. The pragmatic call: most adults who aren't eating 4+ servings of magnesium-rich foods daily should supplement, and the form-to-goal match (glycinate for sleep / anxiety, citrate for constipation, oxide for migraine) is real and useful.

The skeptic case. Pooled effect sizes are mostly small. Sleep: 17 minutes faster onset is a real signal but Cohen's d ≈ 0.2 in the largest modern RCT — well below benzodiazepine or behavioural CBT-I effect sizes (Held et al. 2025; Mah & Pitre 2021). BP: pooled 2.8 mmHg systolic across all comers (Argeros et al. 2025) — the meaningful effect lives only in hypertensives and hypomagnesemics, and in normotensives the signal disappeared. Muscle cramps: Cochrane found magnesium unlikely to produce clinically meaningful relief in idiopathic cramps (Garrison et al. 2020). Cognition (threonate): rat data is striking, human trials are small and industry-funded. Mood: small samples, heterogeneous, one negative add-on trial. The mortality cohort signal is observational and confounded by overall diet quality. The form market has been built on bioavailability claims that often weren't tested head-to-head — Walker 2003 found citrate > oxide and chelate, but the literature is thin on direct glycinate-vs-citrate comparisons in humans. The supplement industry's revenue motive on "specialized forms" (threonate at 5× citrate price; taurate / orotate / ascorbate at premium with no comparative trials) deserves named skepticism.

The author's call. The real entry is: fix dietary intake first, supplement if you can't. The catalogue's reader probably eats below RDA and probably benefits from food-first plus a low-cost daily form. Effects are real, mostly modest, biggest in the subgroups whose intake or pharmacology is most depleted. The form story is honestly useful — glycinate for daily / sleep, citrate for bowel-friendly daily, oxide for migraine, threonate as a cognition flier — without being magical. Score evidence at 4 (multiple meta-analyses across multiple endpoints; the question is effect size not direction). Score the benefit dimensions as small-to-meaningful, not transformative. Controversy is low: there's near-universal consensus that adequate magnesium is necessary and inexpensive; debates are at the margin (UL re-evaluation, form-specific bioavailability) rather than over whether the substance works.

Stakeholder + incentive map

Supplement industry. Magnesium is one of the largest single-nutrient supplement categories; "specialized forms" (threonate, taurate, orotate, ascorbate, malate) command 3–10× the price of citrate or oxide. Commercial incentive to over-claim form-specific effects. Magtein (L-threonate) is single-source patented and behind nearly every published threonate trial — author affiliations matter.
Mainstream nutrition / public health. NIH ODS, USDA dietary guidelines, EFSA — all flag widespread inadequate intake. Generally aligned with the optimist case, conservative on supplemental UL (US UL has been criticized as outdated; Costello et al. 2023).
Neurology / cardiology guidelines. AAN/AHS endorses magnesium for migraine prevention; AHA does not formally recommend supplementation for BP in absence of deficiency, despite the 2025 Hypertension meta-analysis.
Functional / integrative medicine. Heavy push of RBC magnesium testing and aggressive supplementation; sometimes ahead of the evidence on specific forms, sometimes correctly flagging serum-testing's blindness to intracellular depletion.
Skeptic / counter-incentive. Evidence-based medicine voices noting effect-size modesty and heterogeneity (Cochrane on cramps; sleep meta-analyses graded low-certainty); supplement-skeptic clinicians who reasonably point at the food-first lever.

Population variability

Baseline intake. US mean intakes — men 340–344 mg/day, women 256–273 mg/day — sit below RDA across both sexes (NIH ODS 2022). ~48% of all-age Americans fall below EAR. Older adults are the worst-served subgroup: 83% of US adults ≥65 below recommended levels.
Sex. Lower RDA in women (320 vs 420 mg/day) but lower intake too; net deficit prevalence is comparable.
Age. Absorption declines and renal loss rises with age; older adults need more, get less.
Diabetes / insulin resistance. Hyperglycemia drives osmotic urinary magnesium loss; deficiency prevalence elevated.
Drug exposure. Chronic PPI + diuretic (Kieboom et al. 2015), loop diuretics, prolonged proton pump suppression, certain antibiotics (aminoglycosides), and alcohol all depress status.
Pregnancy. Demand rises; deficiency associated with pre-eclampsia risk in observational data.
Genetic. Rare familial hypomagnesemia syndromes (TRPM6 mutations) exist but are not the population story.

Knowledge gaps

Head-to-head bioavailability trials between modern organic forms (glycinate vs citrate vs malate) in humans are sparse; most form claims rest on indirect / single-form data and assumed equivalence of organic absorption.
Threonate's blood-brain-barrier story is animal-grade; the human cognition trials are small, short, and largely industry-sponsored. A large, independent, long-duration trial would change the call.
Exercise-associated muscle cramps lack any RCT specifically (Garrison et al. 2020). The athletic community uses magnesium reflexively; the evidence for that specific use case is structurally absent.
The dose-response for mortality endpoints saturates near 250–400 mg/day in Fang et al. 2016; whether higher intakes add benefit, or whether ratio with calcium/potassium matters more than absolute Mg, isn't settled.
The 350 mg supplemental UL is based on diarrhoea threshold from a small original dataset; Costello et al. 2023 argues for re-evaluation. Practical impact is real — many studied doses (400, 500 mg) sit above the formal UL despite being routinely tolerated.
RBC magnesium reference ranges aren't standardized; "functional optimal" ranges from integrative medicine sit higher than conventional lab ranges, and rigorous evidence for the higher threshold's clinical superiority is thin.

Scope decisions. The brief named sleep, muscle function, blood pressure, mood, bowel function, and cardiovascular markers, with form-specific differences. All six are covered: sleep in evidence + payoff; blood pressure as the cleanest evidence block with its own science callout; mood under evidence; bowel function woven through mechanism and protocol (citrate / oxide); cardiovascular markers folded into the long-term stakes and payoff sections via the cohort-mortality data. Muscle function is named but handled honestly — the Cochrane review concluded magnesium does not provide clinically meaningful relief for idiopathic skeletal-muscle cramps in the populations studied (Garrison et al. 2020), and there are no RCTs on exercise-associated cramps. Rather than silently dropping the topic, the article surfaces this in evidence as a named negative result. Form-specific differences are the spine of the protocol section.

Excluded. Intravenous magnesium for eclampsia, torsades, and acute severe asthma — different substance, hospital use, separate clinician context. PMS / dysmenorrhoea (mixed literature, niche). Bone health / osteoporosis (real but smaller signal, deserves its own treatment alongside calcium and vitamin D). Premenstrual symptom data and pregnancy-leg-cramp data both excluded to keep scope manageable.

Rating difficulties.

Sleep at 3 rather than 4: pooled effect is consistent but small (Cohen's d ≈ 0.2 in Held et al. 2025), and the older-adult meta-analysis is graded low to very low certainty. A 4 would imply the kind of effect CBT-I or trazodone produces; magnesium doesn't reach that.
Focus at 1 rather than 2: the threonate cognition signal is real but small, short-duration, and almost entirely industry-affiliated (Magtein is single-source). Scoring 2 would credit the threonate trials more than independent replication warrants. Standard forms produce no consistent cognitive effect.
Health_short_term at 3 aggregates four real-but-modest endpoints (sleep, BP, migraine, bowel). Each individually is closer to a 2; together — and given how many readers will experience at least one — 3 is the honest holistic call.
Evidence at 4 rather than 5: multiple meta-analyses across multiple endpoints, including the 38-RCT BP meta-analysis in Hypertension, but Cochrane-negative on cramps, low-certainty on sleep, and form-specific bioavailability data is structurally thin.

Separate-entry candidates.

Magnesium L-threonate / Magtein may eventually warrant its own entry if independent replication of the cognition data lands — currently too thin and too commercially entangled to centre an entry around.
Proton pump inhibitor + diuretic interaction is an underappreciated chronic-depletion vector that touches several entries and could be its own short note.
Subclinical mineral deficiency assessment (RBC vs serum, the kidney's defence of serum levels) is a cross-cutting concern that could grow into its own piece.

Future link candidates. Once they exist: Sleep hygiene fundamentals, Blood pressure home monitoring, Migraine prophylaxis (riboflavin, CoQ10), Potassium intake, RDA vs EAR vocabulary explainer, Proton pump inhibitor long-term use.

Hard calls during writing.

Whether to recommend RBC magnesium testing. Decided against routine recommendation — reference ranges aren't well standardized, integrative-medicine "optimal" thresholds aren't rigorously validated, and most readers benefit more from a $0.30 ounce of pumpkin seeds than a $50 lab draw. The article handles this by reframing the assessment question around dietary intake instead.
How much to credit threonate. The Slutsky et al. 2010 Neuron paper is striking and the human follow-ups are coherent, but Magtein's single-source patent and industry-funded trial pattern warrants caution. The article includes threonate honestly in the form table and explicitly flags it as an add-on to a standard form rather than a replacement.
Whether to push the food-first lever harder. Magnesium-rich foods (pumpkin seeds, spinach, almonds, black beans) are cheaper than supplements per mg and travel with fibre + potassium that the supplement doesn't. The article lands on "food first, supplement where you can't" rather than defaulting to a pill — the right editorial frame for the catalogue's voice.