Ingredient · Mineral
Magnesium
Also called Mg, elemental magnesium, dietary magnesium. Appears on labels as one of 20 salts and chelates, each absorbed differently.
DietarySupplement node and tells a reader in one glance what kind of page they are on. The alias line is doing real work in three directions: it declares synonyms on-page the way IncIDecoder does, it feeds alternateName in schema, and it is the same alias table the label scanner uses to resolve "Mg bisglycinate" on a bottle to this entity. The fact bar replaces the live build's Source of: N/A and Nutrient: N/A. Empty fields are suppressed, never printed. Every number here is text, never an image, so it is crawlable.The live build shows a bare "GREAT" pill with nothing behind it. Here the rank sits next to the four facts that justify it, which is the difference between a claim and a claim with evidence.
Which form of magnesium is best?
For most people, magnesium glycinate is the best-scoring form of magnesium, rated Tier A with a 1.5 multiplier in the SuppScore system. It outperforms magnesium oxide, the cheapest and most common form on UK shelves, which sits at Tier D with under 4% fractional absorption (Firoz 2001, Schuette 1994). Form choice can matter more than dose, because a large dose of a poorly absorbed form delivers less usable magnesium than a moderate dose of a well-absorbed one.
Tiers are assigned from published absorption evidence. See how form tiers are set.
Why each sentence exists. Sentence one gives a named answer, not a hedge, because "it depends" is never quoted. Sentence two names a competing entity and attaches a real statistic to a citation, which is what makes the block verifiable rather than promotional. Sentence three qualifies the claim, and that qualification is exactly what an AI assistant reaches for when it wants to avoid sounding absolute. Hedged writing does not get cited. Confident writing with an honest caveat does.
Score decomposition is exposed three ways: as prose in the sentence, as text pills below it, and as
PropertyValue nodes in the graph. The methodology is one click away from every score mention, and the link goes to the anchor for the exact rule, not the top of the page.
What is magnesium?
The corrected parent definition
Magnesium is an essential dietary mineral that acts as a cofactor in more than 300 enzymatic reactions, including energy production, protein synthesis, muscle and nerve function, blood glucose control and blood pressure regulation. It is one of the most commonly under-consumed nutrients in UK diets, largely because the richest sources are seeds, nuts, pulses and leafy greens.
The body cannot absorb magnesium on its own. It is always bound to something else, an acid, an amino acid or an oxide, and that binding partner decides how much of it reaches the bloodstream. This is why SuppScore grades every magnesium form separately rather than treating magnesium as a single ingredient, and why two products showing "300 mg magnesium" on the label can deliver very different amounts.
Final SuppScore Ingredient.xlsx is a form, and the build is rendering the first one alphabetically, Magnesium Acetate, as if it were the element. This will be true of every parent ingredient on the site, so it is a data contract fix, not a copy fix.What this replacement does differently. Paragraph one is the definition, entity-first, no form-level claims, written to be quotable on its own. Paragraph two is the bridge: it explains why forms matter and hands the reader straight into the form table below. That second paragraph is the most commercially important writing on the page, because it converts a generic "what is magnesium" visitor into someone who now needs our proprietary data to answer their next question.
NLP note. The primary entity "magnesium" appears at a natural density with supporting entities (cofactor, enzymatic, bioavailability, chelate) that co-occur in the authoritative corpus for this topic. We are not keyword stuffing, we are matching the entity neighbourhood a retrieval model expects to see around this subject. Salience target: primary entity above 0.30, supporting entities 0.05 to 0.15 on rendered text.
Every magnesium form, graded
20 forms scored on published absorption evidence. This is the comparison no supplement brand will publish about its own products.
| Form | Tier | Multiplier | Confidence | Evidence rationale |
|---|---|---|---|---|
| Magnesium glycinateAlso sold as bisglycinate. Same compound, one page. | A | ×1.5 | High | Multiple RCTs and stable isotope studies confirm superior absorption versus oxide. Walker 2003, Siebrecht 2013. |
| Magnesium TRAACS®Patented bisglycinate chelate | A | ×1.5 | High | Patented bisglycinate chelate. Superior absorption confirmed, with additional quality assurance on the raw material. |
| Magnesium citrate | B | ×1.25 | Moderate | Lindberg 1990 showed superior absorption versus oxide. Firoz 2001 supports. Limited direct comparison against bisglycinate. |
| Magnesium L-threonate | B | ×1.25 | Moderate | Central nervous system penetration evidence from Liu 2016, a human cognitive RCT. Limited systemic absorption comparison against bisglycinate. |
| Magnesium taurate | B | ×1.25 | Moderate | Good gastrointestinal tolerability evidence. No stable isotope comparison studies available. |
| Magnesium malate | B | ×1.25 | Moderate | Proxy evidence comparable to citrate. Russell 1995. Limited direct RCTs. |
| Magnesium glycerophosphate | B | ×1.25 | Moderate | Moderately absorbed organic salt. Limited comparative RCTs. |
| Magnesium chloride | C | ×1.0 | Reference | Well-absorbed inorganic salt. Firoz 2001. This is the standard reference-tier form against which others are compared. |
| Magnesium ascorbate | C | ×1.0 | Reference | Standard organic salt. No clear RCT superiority over chloride. |
| Magnesium oxide | D | ×0.85 | Low | Lowest bioavailability of any graded form. Firoz 2001 and Schuette 1994 show under 4% fractional absorption. |
| Magnesium carbonate | D | ×0.85 | Low | Lower absorption than chloride or citrate. Antacid properties reduce bioavailability. Firoz 2001. |
| Magnesium sulphateEpsom salts | D | ×0.85 | Low | Poor oral bioavailability. Used mainly intravenously or as a laxative. |
| Magnesium hydroxideMilk of magnesia | D | ×0.85 | Low | Antacid. Poor oral bioavailability and an osmotic laxative at higher doses. |
| NutriGenesis® magnesium | D | ×0.85 | Low | Food-grown matrix. Limited independent RCT comparison against standard forms. |
| Aquamin™ magnesiumSeawater-derived | D | ×0.85 | Low | Seawater-derived. Limited independent RCT data on bioavailability. |
| Magnesium acetate | ? | Pending | Not graded | This form appears in our ingredient data but has not yet been assigned an evidence tier. It is shown here rather than hidden, and it does not contribute to any product score until it is graded. |
Promoted versus held. Ten forms link out to their own page. The rest are plain text. That difference is decided by search demand, not by how well the form scores. Magnesium oxide is Tier D and gets a page because 9,900 people a month search for it. Magnesium glycerophosphate is Tier B and does not, because almost nobody does. Every held form still appears in full here, so the page answers "how does magnesium glycerophosphate compare" completely without a thin page existing to answer it badly. Promotion is a flag on the record, never a code change.
The bisglycinate alias line prevents a duplicate page. Glycinate and bisglycinate are the same compound with the same tier. Two pages would compete with each other for 288,950 monthly searches and neither would win. One canonical page, one alias declared on-page and in
alternateName. This is the highest-risk de-duplication call on magnesium and it is handled in one line of copy.The ungraded acetate row is deliberate. Showing a gap honestly is a stronger trust signal than a table that looks complete. It also demonstrates the completeness bar working: an ungraded form contributes nothing to any product score.
Mobile. A 20-row, 5-column table is unusable on a phone, and roughly two-thirds of this traffic will be mobile. At 640px the table reflows to cards with the column name as an inline label. No horizontal scroll, no pinch-zoom, and the DOM order is unchanged so the crawler sees the same content either way.
How much magnesium should you take?
SuppScore grades doses across five evidence zones rather than against a single recommended figure. Scored information, not a recommendation.
Below the average requirement. Not enough to meet the needs of most people, even as a top-up alongside diet.
Sits between the average requirement and the UK reference nutrient intake of 270 mg for women and 300 mg for men. Suitable for general maintenance.
Above the reference intake, at the level used in most research. Aimed at correcting a shortfall or meeting raised demand. This is the band most well-formulated UK magnesium products land in.
Supported by systematic review or two or more independent randomised controlled trials. Typically used under supervision rather than chosen off a shelf.
The framing is doing legal work as well as SEO work. Every sentence describes what the evidence supports at a dose band. Not one sentence tells the reader what to take. That distinction is what keeps this on the information side of the line while the clinical layer stays switched off, and it is why the section is safe to ship before a medical reviewer is appointed.
The empty Zone 4 is a deliberate trust signal. Every instinct says to smooth it over. Publishing the gap, and explaining it, does more for credibility than a tidy five-row table would, and it demonstrates that the model is evidence-driven rather than cosmetic. This is the kind of detail an AI assistant quotes when it wants to show a source is rigorous.
One item is blocked. The workbook places the "above upper limit" threshold at 750 mg. EFSA's tolerable upper intake level for supplemental magnesium is widely cited as 250 mg per day. Those are not the same basis and we should not publish either number until Gordon's team confirms which applies. The section ships without a hard upper limit figure and uses the laxative threshold, which is well supported, in its place. Flagged as blocking.
What is magnesium good for?
Each use carries its own evidence grade, because the strength of the research is not the same across them.
- Sleep qualityThe most researched use, and the reason most people in the UK buy magnesiumStrong evidence
- Muscle function and crampWell established mechanistically, mixed on cramp specificallyModerate evidence
- Stress and anxietyContributes to nervous system regulation and GABA activityModerate evidence
- Bone healthRoughly 60% of body magnesium is stored in boneModerate evidence
- Energy productionRequired cofactor in ATP metabolismStrong evidence
- PMS and menstrual symptomsSmall trials, consistent direction, limited scaleEmerging evidence
- Blood pressure regulationModest effect sizes in meta-analysisModerate evidence
What replaces it. A controlled benefit taxonomy where every row links to its hub. This is the mechanism that turns navigation into an entity system: magnesium lists every benefit it earns, and
/benefits/sleep lists every ingredient that earns a place there, with magnesium at the top. Examine runs this at roughly 78 outcomes by 129 ingredients and it is the core of their authority. Ours is smaller, but it can be complete for the benefits we cover, and complete beats large.Per-benefit evidence grades are the differentiator. Everyone else publishes an undifferentiated benefits list. Grading each use separately is more honest, more useful, and gives an assistant a reason to cite us rather than a listicle.
Blocked on data. The ingredient-to-benefit mapping exists in the keyword universe but is not yet a field in the data contract. Until it is, this section cannot be generated. It is the single highest-value contract addition on the list.
What are the signs of low magnesium?
Magnesium shortfall is common and hard to spot, because blood tests are a poor guide. Only about 1% of the body's magnesium circulates in blood, with the rest held in bone and muscle, so a normal serum reading does not rule out a shortfall.
The signs most consistently reported alongside low magnesium status are muscle cramp and twitching, fatigue that does not respond to rest, poor sleep quality, and low mood or irritability. Groups more likely to fall short include people with type 2 diabetes, people taking proton pump inhibitors or loop diuretics long term, people who drink heavily, and older adults, in whom absorption declines and losses increase.
These are associations reported in the literature, not a diagnostic checklist. If you recognise several of them, that is a conversation to have with a GP rather than a reason to self-treat.
Why it sits here and not behind the clinical gate. The line we are holding is between describing what the literature reports and telling an individual what is wrong with them. Every sentence here is population-level and sourced. There is no symptom checker, no "you may have", no severity scale. The final paragraph routes the reader to a GP explicitly, which is both the right thing to do and the thing that keeps this defensible.
The serum testing point is our most quotable original insight in this section. It is true, it is counter-intuitive, it is rarely stated plainly by consumer sites, and it is exactly the shape of sentence an AI assistant lifts when answering "how do I know if I'm low on magnesium".
Which foods are highest in magnesium?
Pumpkin seeds are the densest common source, supplying roughly 150 mg in a 30 g portion, around half the UK reference nutrient intake. Chia seeds, almonds, cashews, spinach, black beans, dark chocolate above 70%, and wholegrains follow. Refining strips magnesium out, which is why wholemeal bread carries several times more than white.
Most people who eat seeds, nuts, pulses and leafy greens regularly will reach the reference intake from food alone. Supplementing is worth considering when diet is consistently short, when demand is raised, or when a medication is increasing losses.
The decision: absorb it, do not build for it. This is a section on the parent, not a set of food pages. Food pages would rank, then dilute the entity and pull the site's centre of gravity away from scoring. One well-written section captures the demand and, crucially, the second paragraph routes the reader back to the supplement question, which is where our data is.
Marked as enrichment. This section can ship after launch without any re-architecture. It is in the reference build so the dev team can see where it goes.
When should you take magnesium?
There is no single correct time, and absorption does not differ meaningfully across the day. Evening dosing is common simply because magnesium is most often taken for sleep and muscle relaxation. Taking it with food reduces the chance of loose stools, which matters most with oxide, citrate and sulphate. Splitting a larger dose across two servings is better tolerated than one large one, and improves the proportion absorbed, since magnesium uptake becomes less efficient as a single dose rises.
It also does a job for the user. Someone who has just decided which form to buy immediately wants to know how to take it. Placing this after the form table and dose zones follows the real decision sequence rather than the keyword volume order.
Interactions, side effects and personalised dosing
SuppScore does not currently cover these. We are not publishing guidance on drug interactions, contraindications or individual dosing until a named, credentialled medical reviewer has signed it off. For anything in this category, speak to a pharmacist or GP.
- Interactions with medicines
- Contraindications and who should avoid supplementing
- Side effects beyond the tolerability notes above
- Dosing for specific conditions or during pregnancy
MedicalIndication and MedicalContraindication begin emitting. No rebuild, no re-architecture, one flag.Why show it rather than hide it. An absent section reads as an oversight. A section that says plainly what we do not cover and why reads as discipline, and it is one of the strongest trust signals available to a young YMYL site. It also pre-empts the obvious criticism that a supplement scoring site is quietly practising medicine.
All 988 magnesium questions in our keyword universe carry the medical-claim flag. This gate is doing real work, and the named reviewer is the critical-path dependency for the entire clinical layer across the site, not a nice-to-have.
Common questions about magnesium
Seven questions answered here. Broader questions have their own pages.
What is magnesium good for?
Magnesium acts as a cofactor in more than 300 enzymatic reactions. The uses with the strongest evidence behind them are sleep quality, muscle function and cramp, stress and anxiety regulation, and bone health. SuppScore grades each of these uses separately, because the strength of evidence differs between them.
Which magnesium is best for sleep?
Magnesium glycinate is the highest-scoring form for sleep, rated Tier A with a 1.5 multiplier. The glycine it is bound to is itself associated with sleep onset, and the chelate is well tolerated at the doses used in the evening. Magnesium L-threonate is the alternative most often cited for cognitive effects, at Tier B. See every ingredient graded for sleep.
When is the best time to take magnesium?
There is no single correct time, and absorption does not differ meaningfully across the day. Evening dosing is common because magnesium is most often taken for sleep and muscle relaxation. Taking it with food reduces the chance of loose stools, particularly with oxide, citrate and sulphate.
How long does magnesium take to work?
Correcting a genuine shortfall typically takes several weeks of consistent intake, because magnesium is stored largely in bone and muscle rather than circulating freely. Effects some people notice sooner, such as bowel changes from oxide or citrate, are a property of the form rather than a sign of magnesium status improving.
Is magnesium glycinate the same as magnesium bisglycinate?
Yes. They are the same compound, magnesium bound to two glycine molecules. SuppScore grades both at Tier A with a 1.5 multiplier and treats them as one entity. Products may use either name on the label.
Can you take too much magnesium?
Magnesium from food is not a concern, because healthy kidneys clear the excess. Supplemental magnesium is different. Doses above roughly 350 mg per day from supplements are associated with a laxative effect, and that threshold is lower for loosely bound forms such as oxide and sulphate. Anyone with reduced kidney function should speak to a doctor before supplementing, because clearance is the mechanism that makes magnesium self-limiting.
Which foods are highest in magnesium?
Pumpkin seeds, chia seeds, almonds, spinach, cashews, black beans, dark chocolate and wholegrains are the densest common UK sources. A 30 g portion of pumpkin seeds supplies roughly 150 mg, around half the UK reference nutrient intake.
The selection rule. A question goes in this block when it has real demand but no natural home in the body copy, and when the answer fits in one paragraph. A question gets its own page when it has standalone volume or its own People Also Ask presence. "Can magnesium help you sleep" (14,800, confirmed AI Overview) is a standalone Q&A page, not a row here, because it deserves a full treatment and it can rank in its own right.
Answers are complete in the first sentence. Every one of these leads with the direct answer and then qualifies. That is a retrieval decision: an assistant reading this block should be able to stop after one sentence and still be correct.
Accordions are safe here. The content is in the DOM at page load, not injected on click, so the crawler sees all of it. The first item is open by default so the pattern is obvious.
details and summary give keyboard and screen reader behaviour for free, which a custom accordion would have to rebuild.FAQPage schema covers these seven only. Never the questions answered in body sections, or the same content is claimed twice.
Products containing magnesium
Scored products that list magnesium on the label, highest scoring first. Ordering is by SuppScore only. We have no commercial relationship with any brand shown.
Fix two, the scale. The live module shows scores of 1, 9, 24, 63, 105, 134 and 183 side by side. How-we-score describes the SuppScore as a combined total, so a product with more scored ingredients mechanically outscores a better-formulated product with fewer. To a user this reads as a bug. It also blocks
Rating schema entirely, because Rating needs bestRating and worstRating and an unbounded sum cannot supply them. This mock uses a normalised 0 to 100 display with the raw total kept underneath. Gordon's decision, but it blocks the product template and all score markup.Fix three, duplicate records. The live module shows the same brand's "Daily Multivitamin" at 105 and "Daily Multi-Vitamin" at 95. Either two records for one product or two products with no way to tell them apart. It also renders a brand as "TheStack-Web", which is a source identifier leaking into the brand field.
The decision. On the flagship ingredient page of a site whose entire proposition is independence, the ordering rule must be explicit, disclosed and defensible. The line above the grid states it. This is not decoration, it is the thing a sceptical journalist checks first.
The added form line on each card is a small thing that does a lot: it shows the score is derived from something, and it seeds the internal link from product back to form.
References
- Firoz M, Graber M. Bioavailability of US commercial magnesium preparations. Magnesium Research, 2001. PubMed
- Walker AF, Marakis G, Christie S, Byng M. Mg citrate found more bioavailable than other Mg preparations in a randomised, double-blind study. Magnesium Research, 2003. PubMed
- Schuette SA, Lashner BA, Janghorbani M. Bioavailability of magnesium diglycinate versus magnesium oxide in patients with ileal resection. JPEN, 1994. PubMed
- Siebrecht S. Magnesium bisglycinate as a safe and well absorbed form of magnesium. OM & Ernährung, 2013.
- NIH Office of Dietary Supplements. Magnesium: Fact Sheet for Health Professionals. ods.od.nih.gov
Why this matters more here than almost anywhere. The entire proposition is that we are the source worth trusting. A reference list that looks impressive and is not is the exact failure we are selling against, and it invites the question of whether the scores behind it are equally padded. Deduplicate at the data layer and display two honestly rather than five dishonestly.
Format. Author, year, title, journal, link. Not a bare URL. Full citations are what make the list usable by a human, quotable by a machine, and mappable to
ScholarlyArticle with real author and datePublished values. rel="nofollow" on outbound reference links is deliberate: we cite generously and we do not want a hundred and forty ingredient pages each passing equity to PubMed.
Reviewed by [NAMED REVIEWER REQUIRED]
Registered Nutritionist, Association for Nutrition · Advisory board
Published 4 Mar 2026 · Last reviewed 31 Jul 2026
- Dose zones updated to the Five-Zone model. Zone 4 gap documented.
- Form tiers refreshed. 20 magnesium forms graded, up from 14.
- Two references added for bisglycinate absorption.
The reviewer must be a structured field, not hard-coded copy. Name, credential, awarding body and a link to a corroborable profile, marked up as
Person with hasCredential and sameAs. The sameAs is what lets a knowledge graph verify the person exists independently of our claim about them. A name with no corroboration is barely better than no name.The update history is generated, not written. Entries come from pipeline events: dose zones changed, references added, forms regraded. It costs nothing per page and it is the strongest freshness and rigour signal available. Rendered as list items with
time elements, never as headings. Examine ships eight H1s on their magnesium page because their update history uses heading tags. That bug costs them nothing at their authority. It would cost us.This block is deliberately incomplete in the reference build. The reviewer name is the critical path for the clinical layer, the
Person schema and this byline. It is one name, and it is blocking three things.
What SuppScore does not tell you
Our score measures formulation: which forms a product uses, at what doses, with what added. It does not measure:
- Purity, contamination or heavy metal content. We do not test products.
- Manufacturing quality, or whether the label matches the capsule.
- Whether a supplement is right for you, which depends on your diet, medication and health.
A high score means a product is well formulated for what it claims to do. It is not a recommendation to take it.
One, it is honest, and honesty is the brand. Two, it is precisely the kind of qualification an AI assistant quotes when it wants to hedge, which means we get cited on questions we do not even answer. Three, it heads off the criticism that a supplement score is a safety claim, which is the single most likely way this product gets into trouble.
This costs one short section and it is worth more than any amount of "trusted by" badging.