Magnesium

The VerdictMODERATE CONVICTIONVerdict Score 80Worth-It: Solid ROI (77/100)

CONDITIONAL — Highly effective for people with deficiency, on high-dose Vitamin D, or with insulin resistance.

Summary: Magnesium is one of the few supplements where the evidence is genuinely compelling — but only if you're taking the right form, for the right reason. It's essential if you're on high-dose Vitamin D (your body can't activate D3 without it), meaningful for blood sugar control and sleep if you'

Vitamin D cofactor rescue: STRONG — obligate cofactor for CYP2R1/CYP27B1 hydroxylation steps
Insulin sensitivity / T2D: STRONG — HOMA-IR WMD -0.67, p=0.013 (Guerrero-Romero 2016 meta, 22 arms)
Sleep quality: MODERATE — Cohen's d ~0.2, ~17 min faster onset (deficient/stressed populations only)

That's the general answer. Your stack is different.

Check your whole stack

The Marketing

What People Claim

"The master mineral — improves sleep, stops cramps, lowers blood pressure, boosts testosterone, calms anxiety. One pill, every problem solved."

Magnesium is marketed as the ultimate health mineral. TikTok and Instagram push it as a stress cure and sleep hack, often without specifying which of the eight commercially available forms they're recommending.

The most prominent claim is that magnesium "relaxes muscles and stops cramps." Athletes reach for it post-workout, endurance runners add it to their recovery stack, elderly people take it for night cramps. This claim is so pervasive it's treated as settled science.

A quieter but more important claim — common in functional medicine — is that most of the population is chronically magnesium-deficient, running on depleted intracellular stores despite normal blood tests. If true, this would make magnesium one of the highest-impact supplements available.

The Science

What the Evidence Actually Shows

Claimed Benefit	Conviction	Effect / Key Study
Vitamin D cofactor rescue STRONG What would change this: Evidence of D3 hydroxylation proceeding normally at low intracellular Mg without enzyme dysfunction.	HIGH	Obligate cofactor for CYP2R1 + CYP27B1. High-dose D3 without Mg → functional D resistance + vascular calcification risk. (Deng 2013, NHANES III)
Insulin sensitivity / T2D STRONG What would change this: Large RCT in magnesium-replete T2D patients showing no HOMA-IR reduction.	HIGH	HOMA-IR WMD −0.67, p=0.013 (Guerrero-Romero 2016 meta, 22 treatment arms). Requires ≥4 months minimum duration.
Sleep quality MODERATE What would change this: RCT in well-nourished, stress-free adults showing meaningful effect beyond ceiling.	MODERATE	Cohen's d ~0.2; ~17 min faster sleep onset. Restricted to deficient or high-cortisol populations — replete adults hit ceiling. (Abbasi 2012; Briskey 2024)
Cognitive function (L-threonate) EMERGING What would change this: Independent multi-site replication of Hausenblas 2024 findings.	MODERATE	CSF Mg +54%; deep + REM sleep improved; "brain age" reversal in early data. (Hausenblas 2024, N=80). Independent replication required.
Blood pressure reduction MODERATE What would change this: RCT in normotensive adults showing meaningful BP reduction.	MODERATE	SBP −2.81 to −4.18 mmHg. Strictly conditional — hypertensive + hypomagnesemic adults only. Zero benefit in normotensives. (Dibaba 2017; Zhang 2016)
Muscle cramping (idiopathic) DEBUNKED What would change this: High-quality RCT in magnesium-deficient cramp sufferers showing benefit after screening by RBC Mg.	LOW	No significant reduction over placebo — high-certainty evidence. Cochrane 2012 + Cochrane 2020 (two separate reviews). Cramps are neurological, not mineral-driven.

The Biochemistry

How It Works

600+ Enzymatic Reactions

Magnesium is an essential intracellular cation — it lives primarily inside cells, not the bloodstream. It acts as a cofactor for over 600 enzymatic reactions, including ATP synthesis (every energy-producing reaction in the body requires it), DNA/RNA transcription, and cell signalling.

Sleep: NMDA Antagonist + GABA-A Agonist

Magnesium antagonises NMDA receptors, blocking the calcium influx that drives excitatory glutamate activity — calming neuronal excitability. Simultaneously, it acts as a GABA-A receptor agonist, promoting the inhibitory signalling needed for sleep onset. Adequate magnesium status correlates with lower cortisol and increased melatonin production.

Vitamin D: The Unavoidable Dependency

CYP2R1 (liver) and CYP27B1 (kidneys) — the two enzymes that convert inactive Vitamin D into its active form — are both magnesium-dependent. High-dose D3 drives these pathways hard, rapidly consuming intracellular magnesium. Without sufficient Mg, conversion stalls: serum 25(OH)D stays low despite massive D3 intake. Unactivated Vitamin D metabolites can also promote vascular calcification in the absence of magnesium's counterbalancing effect.

Insulin Sensitivity: GLUT4 Translocation

Intracellular magnesium is required for insulin receptor function and GLUT4 translocation — the process by which cells take up glucose from the bloodstream. Deficiency impairs both, raising HOMA-IR. Restoration of intracellular Mg reverses these impairments, explaining the strong meta-analysis signal for T2D.

The Controversy

Where Studies Disagree

The Sleep Effect: Impressive vs. Modest

Abbasi 2012 (N=46, elderly insomniacs)

ISI score dropped significantly; melatonin increased; serum cortisol fell. Magnesium dramatically improved sleep quality.

Meta-analysis aggregate (multiple RCTs)

Cohen's d ~0.2 across meta-analyses — modest effect size. ~17 minutes faster sleep onset. Effect disappears in magnesium-replete healthy adults.

Abbasi used elderly insomniacs with probable baseline deficiency — the ideal responder. Healthy, replete adults hit a ceiling effect. Population selection drives the apparent contradiction.

The Deficiency Paradox: Epidemic vs. Normal Tests

Dietary intake data + intracellular deficit theory

Most people consume less than the RDA. Up to 50% of the population may be subclinically deficient despite normal serum levels. A hidden deficiency epidemic.

Standard clinical practice

Routine serum magnesium returns normal in most patients. Clinicians and patients treat "normal serum Mg" as evidence of adequacy.

Serum Mg represents <1% of total body stores. The body tightly defends serum levels by pulling Mg from bone and muscle — "normal" serum hides intracellular depletion. RBC Mg is the correct functional test.

Blood Pressure: Meaningful vs. Zero Effect

Dibaba 2017 / Zhang 2016 meta-analyses

SBP reduced −2.81 to −4.18 mmHg. Meaningful dose-dependent effect. Magnesium lowers blood pressure.

Trials in normotensive subjects

Zero antihypertensive effect in healthy-BP adults. Benefit is entirely restricted to hypertensive + hypomagnesemic participants.

Both are correct. The effect is population-conditional, not universal — hypertensive + hypomagnesemic people respond; normotensives don't. Pooled analyses that mix both populations obscure this.

Translational Limits

Real World vs Lab

The Serum Test Fallacy

Researchers stratify by serum Mg. "Normal" serum Mg used as inclusion criterion.

Practitioners and patients use serum Mg as the primary diagnostic. "Normal" serum masks intracellular depletion in up to 50% of cases.

UNDERESTIMATED

Clinical effect sizes are likely underestimated in trials that didn't screen for true intracellular deficiency.

Baseline Deficiency Dictates Response

Impressive effect sizes in elderly and diabetic cohorts with documented hypomagnesemia.

Results are marketed to healthy adults expecting the same benefits. Ceiling effect is real — replete adults don't respond.

MORE CONSERVATIVE

Form Extrapolation

Mechanistic and PK studies conducted on specific forms (oxide, glycinate, L-threonate) with known absorption parameters.

Manufacturers attribute physiological benefits to any salt form, including poorly-absorbed oxide. "Magnesium is magnesium" is the implicit claim.

FORM CRITICAL

Oxide is 4% absorbed. Always verify the salt form — elemental content on the label is misleading for oxide.

Evidence-Based Dosing

The Protocol

Dosing by Population

Population	Dose	Timing	Form	Duration
General adult (maintenance/deficiency)	200–400 mg elemental	Split AM + PM with meals	Glycinate or Citrate	Ongoing
Sleep optimisation (deficient/stressed)	250–500 mg elemental	30–60 min before bed	Glycinate (bisglycinate)	Ongoing
T2D / Insulin resistance	300–400 mg elemental	Split AM/PM	Glycinate or Malate	≥4 months minimum
Hypertension (hypomagnesemic)	365–450 mg elemental	Split AM/PM	Glycinate or Taurate	Ongoing
Athletes (high-demand training)	400–600 mg elemental	Split pre- and post-workout	Glycinate or Malate	Ongoing
Cognitive / deep sleep (targeted)	1,000 mg L-threonate	Evening	L-threonate (Magtein®)	≥3 weeks
High-dose Vitamin D users (>2,000 IU D3)	300–400 mg elemental	Alongside D3 dose	Glycinate	Ongoing (obligate)

Forms Ranked

Glycinate

~25–30% absorbed

Sleep, anxiety, general deficiency, high-dose protocols

£15–25/month

L-Threonate

Brain-penetrant

Cognitive decline, deep sleep, CSF Mg elevation

£35–60/month

Citrate

~16% absolute

Budget-friendly deficiency correction

£8–15/month

Malate

High bioavail.

Fatigue, fibromyalgia, athletes (Krebs cycle)

£12–20/month

Taurate

High (limited PK)

Hypertension, cardiovascular

£15–22/month

Oxide

~4% absorbed

Constipation/antacid only — NOT for deficiency

£4–8/month

Absorption Tips

⚡ Split doses — divide into 2–3 smaller amounts to avoid transporter saturation and GI distress
🍽️ Take with food — reduces GI side effects for citrate and lower-quality forms
⏱️ Separate from fluoroquinolones/tetracyclines by minimum 2 hours — chelation causes therapy failure
💊 PPI users need higher doses — 30–40% absorption impairment from long-term omeprazole/lansoprazole
🚫 Avoid high-dose zinc (>40 mg) or calcium (>1,000 mg bolus) at the same time — competitive absorption

Risk Profile

Safety & Interactions

⚠ Severe Interactions

Fluoroquinolones (Cipro, Levofloxacin)

Chelation in gut → insoluble complex → antibiotic therapy failure. Separate by minimum 2 hours; ideally avoid co-prescription entirely.

Tetracyclines (Doxycycline, Minocycline)

Same chelation mechanism — absorption of antibiotic blocked at any dose. Separate by minimum 2 hours.

High-Impact Interactions

Bisphosphonates (Alendronate/Fosamax)

Mg significantly reduces absorption of osteoporosis medication. Take alendronate 2+ hours before magnesium.

Proton Pump Inhibitors (Omeprazole, Lansoprazole)

Long-term use lowers gastric acid → 30–40% Mg absorption impairment → hypomagnesemia with prolonged use. Monitor Mg status; increase dose or switch to glycinate.

High-dose Vitamin D (>2,000 IU)

D3 drives hepatic + renal hydroxylases that consume Mg reserves. Co-supplement 300–400 mg elemental Mg when on high-dose D3.

Moderate Interactions

Loop Diuretics (Furosemide) + Thiazide Diuretics

Increase urinary Mg excretion → systemic depletion. Monitor and supplement in long-term users.

High-dose Zinc (>40 mg) / High-dose Calcium (>1,000 mg bolus)

Competitive intestinal absorption. Separate timing or spread doses through the day.

Contraindicated Populations

Do NOT supplement (or consult physician first)

End-stage renal disease (ESRD): Kidneys can't excrete excess Mg → hypermagnesemia → cardiac arrhythmia, respiratory depression. CONTRAINDICATED.
Myasthenia gravis: Mg has neuromuscular blocking properties — can worsen muscle weakness.
Severe bradycardia: Mg acts as physiologic calcium channel blocker — may exacerbate slow heart rate.

NIH Tolerable Upper Limit (supplemental): 350 mg/day elemental — established for osmotic diarrhea risk only, NOT toxicity. Dietary Mg has no UL. True toxicity threshold is very high (>5g/day or IV in renal failure).

What the Simple Answer Misses

The Nuance

Magnesium nuance and population stratification

What Doesn't Work

Magnesium for idiopathic muscle cramps: Cochrane reviewed every trial twice (2012 + 2020) — high-certainty null evidence. Cramps are predominantly neurological and hydration-driven, not mineral-deficiency-driven. Pickle juice (1ml/kg bodyweight, ~85 sec resolution) outperforms Mg for acute cramping via neurological reflex, not electrolyte replacement.
Magnesium oxide as a health supplement: ~4% absolute bioavailability. Most elemental Mg passes through unabsorbed, acting as an osmotic laxative. The high elemental percentage on the label is misleading. Buy only as a laxative or antacid.
"Normal" blood test = adequate magnesium: Serum Mg is <1% of total body stores. The body tightly defends serum levels by pulling Mg from bone and muscle. "Normal" serum is not evidence of adequacy — request RBC Mg for true functional status.

Who Benefits Most

High-Dose Vitamin D Users

Anyone on >2,000 IU D3/day without Mg is working against themselves. Functional D resistance is the outcome without adequate Mg. This is non-negotiable.

T2D / Insulin-Resistant Adults

Strongest metabolic evidence. Requires >4 months. Don't expect results in 4 weeks — this isn't a quick fix.

Subclinically Deficient Adults

Stress, poor diet, PPI/diuretic use, high alcohol intake — all deplete Mg. All endpoints become more responsive in this state.

Insomniacs with High Cortisol

Sleep onset improvement, melatonin increase, ISI improvement. Effect is real but restricted to deficient or high-stress populations.

Hypertensives (+ Hypomagnesemia)

Dose-dependent SBP/DBP reduction. Zero benefit in normotensives — don't extrapolate the effect to healthy-BP adults.

Cognitive Decline (L-Threonate)

CSF Mg elevation + deep/REM sleep improvement. Promising 2024 RCT. Needs independent replication — and at £35–60/month, only worth it for targeted cognitive goals.

Cost-Effectiveness

Form	Monthly Cost	Food Alternative	Value Verdict
Glycinate / Citrate	£8–25/month	Pumpkin seeds (28g=156mg), dark chocolate, spinach, black beans	WORTH IT (if deficient/D3/IR)
L-Threonate	£35–60/month	No food equivalent for brain-targeted delivery	CONDITIONAL (cognitive goal)
Malate / Taurate	£12–22/month	Same food sources as glycinate	CONDITIONAL (specific use case)
Oxide	£4–8/month	—	SKIP (for systemic use)

Quick Reference Card

Verdict Conditional — highly effective in the right populations; overhyped for cramps

Best Form Glycinate/bisglycinate (general); L-Threonate (cognitive/sleep target)

Dose 200–400 mg elemental (glycinate); 1,000 mg L-threonate; split doses with food

Key Interaction Fluoroquinolones + tetracyclines — antibiotic therapy failure; separate 2 hrs min

Test to Order RBC Magnesium — NOT serum Mg (serum misses 50% of deficient cases)

Critical Flag If on D3 >2,000 IU/day, magnesium is non-negotiable — not optional

Overall Evidence Confidence

MODERATE

HIGH for Vitamin D cofactor rescue + T2D/insulin sensitivity

MODERATE CONVICTION

What would change this verdict?

A 6-month, independently funded, multi-centre RCT in healthy adults stratified by RBC magnesium (not serum), comparing 500 mg/day glycinate vs 1g/day L-threonate vs placebo, using objective polysomnography for sleep and NIH Cognitive Toolbox for cognition.

If benefits emerge specifically in the magnesium-replete subgroups, the paradigm shifts from deficiency correction to pharmacological enhancement — and sleep/cognition conviction would upgrade to HIGH. The high-conviction endpoints (Vitamin D cofactor, insulin sensitivity) already have the evidence density needed; they won't change.

Evidence Base

Key References

Abbasi B et al. (2012) — J Res Med Sci — N=46

The effect of magnesium supplementation on primary insomnia in elderly: ISI and melatonin improved significantly in elderly insomniacs with probable deficiency.

Guerrero-Romero F et al. (2016) — Pharmacol Res — Meta-analysis, 22 treatment arms

Oral magnesium supplementation improves insulin sensitivity: HOMA-IR WMD −0.67, p=0.013. Duration dependency confirmed.

Dibaba DT et al. (2017) — Hypertension — Meta-analysis, 11 RCTs, N=543

The effect of magnesium supplementation on blood pressure: SBP −4.18 mmHg in hypertensive + hypomagnesemic adults.

Garrison SR et al. (2020) — Cochrane Database Syst Rev

Magnesium for skeletal muscle cramps: high-certainty evidence of no benefit for idiopathic cramps in older adults. Second Cochrane review confirming 2012 finding.

Hausenblas HA et al. (2024) — Sleep Medicine: X — N=80, 21 days

Magnesium L-threonate supplementation and sleep quality: deep sleep and REM improved; CSF Mg +54% in earlier mechanistic work. Independent replication required.

Briskey D et al. (2024) — Food and Nutrition Sciences — N=80, 71 completed

Magnesium supplementation and sleep quality: 32% vs 16% sleep quality improvement rate over placebo.

Deng X et al. (2013) — BMC Medicine — NHANES III large cohort

Magnesium, vitamin D status and mortality: Mg intake modulates vitamin D status; both low together → higher all-cause mortality. Mechanistic confirmation of CYP2R1/CYP27B1 Mg dependency.

NIH Office of Dietary Supplements — Magnesium Fact Sheet for Health Professionals

Tolerable UL 350 mg supplemental (osmotic diarrhea risk, NOT toxicity); mechanisms, drug interactions, and population-specific considerations.

Action ROI

Is this worth your time, money, effort, risk, and trust for this goal? Different from Verdict Score (evidence strength) and Leverage Map (relative importance) — Action ROI is the worth-it call once friction is priced in.

Action ROI score

77/100 Solid ROI Trust grade B

Conditional. Genuinely high value if you take high-dose Vitamin D, are insulin resistant, or are subclinically deficient (stress, poor diet, PPI or diuretic use). Near-zero if you are already magnesium-replete and sleeping well. Useless for muscle cramps.

Time

Low

Money

Low

Effort

Low

Risk

Low

Why this score

B-grade evidence (Verdict 80). Trust is split by endpoint: STRONG and near-mechanistically obligate as a Vitamin D cofactor (Deng 2013, NHANES III; both activating enzymes CYP2R1 and CYP27B1 are magnesium-dependent) and STRONG for insulin sensitivity (Guerrero-Romero 2016 meta, 22 arms, HOMA-IR WMD -0.67, p=0.013, needs 4+ months). Only MODERATE for sleep and DEBUNKED for idiopathic cramps (Cochrane 2020, high-certainty null).
The benefit lever is real but population-gated. For high-dose D3 users the cofactor need is obligate, for insulin resistance HOMA-IR drops over months, but the sleep lever is modest (Cohen's d about 0.2, roughly 17 minutes faster onset, Abbasi 2012 / Briskey 2024) and shows up only in deficient or high-stress people. Replete adults hit a ceiling and feel little.
Money cost is Low: about $8-20/month for glycinate or citrate at 200-400 mg elemental, and food can cover it for some (1 oz pumpkin seeds about 150 mg, spinach, black beans).
Time and effort cost is Low: one or two capsules a day with meals, about 30 seconds. The only behavior change is reading the elemental milligrams and avoiding the cheap oxide form (about 4% absorbed, effectively a laxative).
Risk is Low at standard doses. The supplemental upper limit of 350 mg/day is set for GI diarrhea, not toxicity. Real risk is confined to end-stage renal disease (hypermagnesemia) and chelation with fluoroquinolone or tetracycline antibiotics (separate by 2 hours).

Why it didn’t score higher

B-grade trust caps the score at 89, but the practical cap is the population-gating: the sleep and recovery lever named in this goal context is modest (about 17 minutes) and entirely conditional on deficiency, so a large share of buyers will be replete and feel nothing.
The most marketed use, magnesium for muscle cramps, is debunked by two rounds of high-certainty Cochrane evidence, which drags the category's real-world ROI down.
Standard serum testing misses intracellular deficiency in up to 50% of people, so most buyers cannot easily confirm they are in the responder group before spending.

Best for

Adults taking Vitamin D above 2,000 IU/day (obligate cofactor; functional D resistance without it)
Insulin-resistant adults and type 2 diabetics willing to run it 4+ months
The subclinically deficient: high stress, poor diet, regular alcohol, PPI or diuretic users
Insomniacs with a plausible deficiency (high-stress lifestyle, poor intake) rather than replete good sleepers

Lower ROI if

You are a magnesium-replete healthy adult with a good diet and already sleeps well (ceiling effect).
You are taking it for idiopathic muscle cramps (Cochrane high-certainty null).
You are normotensive and expects a blood-pressure benefit (effect is restricted to hypertensive plus hypomagnesemic subjects).
You buy magnesium oxide, where roughly 4% is absorbed and most passes through as a laxative.
You expect a fast or dramatic effect; the metabolic benefits take months and the sleep effect is subtle.

Minimum effective dose

200-400 mg elemental magnesium/day as glycinate or citrate, split with meals. For sleep in a likely-deficient person: 250-500 mg glycinate 30-60 minutes before bed. For high-dose D3 users: 300-400 mg elemental alongside the D3 dose. Dose to elemental magnesium, not compound weight, and avoid oxide for systemic use.

Track this

Subjective sleep onset and sleep quality at 2-4 weeks (expect subtle, not dramatic)
If on high-dose D3: whether serum 25(OH)D finally responds once magnesium is adequate
Fasting glucose or HOMA-IR over 3-4 months if taking it for insulin resistance
RBC magnesium rather than serum if you test status at all