Magnesium glycinate vs citrate vs oxide: which form is actually better?

There is no best magnesium. There is a salt, a label, a gut, and a job.

The aisle sells glycinate as sleep, citrate as clean, oxide as cheap, and L-threonate as a brain upgrade. NIH’s Office of Dietary Supplements is blunter. Supplement Facts list elemental magnesium. Forms that dissolve well tend to absorb better. Aspartate, citrate, lactate, and chloride usually beat oxide and sulfate. Food still does the first job. The adult tolerable upper intake for supplemental magnesium is 350 mg a day, because unabsorbed salt pulls water into the bowel.

This page compares forms. It is not a Best of rank and not a brand league table. The site’s score chips are Excellent, Fair, or Poor. Where the evidence is a reasonable default rather than a clear win, the table below says Good. That word is the brief. It is not a fourth chip.

How we evaluated this

We are reading absorption papers, an ODS fact sheet updated 6 January 2026, and a short list of sleep, cramp, migraine, and Magtein trials. We did not assay bottles. We did not invent a glycinate-versus-citrate sleep RCT that the literature does not have.

Scores weigh three things: bioavailability literature, tolerance, and use-case fit. Excellent is a clear win on that dimension with human data. Good is a reasonable default. Fair is mixed or thin. Poor is the wrong tool, or the claim is weak.

Serum magnesium is a poor tissue-status marker. ODS is explicit about that. A consumer “before and after” fingerstick is not a repletion certificate.

The label trap

A “500 mg magnesium oxide” tablet is a salt weight until the elemental line says otherwise. Atomic magnesium is 24.305. The percentage is chemistry. Hydrate form changes the denominator. Read the elemental milligrams.

Approximate elemental magnesium by weight of the salt:

Those percentages are molecular-weight arithmetic, not an ODS table and not a promise your hydrate matches the row.

What the absorption studies actually show

Lindberg 1990. Oxide dissolves poorly. Citrate does not. After an oral load, the urinary magnesium increment is larger with citrate.

Walker 2003 (PMID 14596323). Double-blind, placebo-controlled, 46 people, 300 mg elemental magnesium a day for 60 days. Organic forms (citrate and an amino-acid chelate) beat oxide on 24-hour urine at day 60. Citrate had the highest mean serum magnesium after acute and chronic dosing. Oxide was not different from placebo on several markers. This is the strongest consumer-facing chronic comparison among common pills.

Firoz and Graber 2001 (PMID 11794633). Commercial US preparations. Oxide fractional absorption about 4%. Chloride, lactate, and aspartate were higher and similar to each other. Inorganic is not automatically worse. Oxide is.

Kappeler 2017. Crossover in 20 healthy men after a saturation design. Single 300 mg elemental doses. 24-hour urinary magnesium was higher with citrate than oxide (adjusted mean difference +0.565 mmol, 95% CI 0.212 to 0.918, p=0.0034). Serum magnesium was higher for citrate at 2 to 6 hours. Funded by Verla-Pharm, which sells a citrate product. The design is still useful. The funding is not invisible.

Schuette 1994 (PMID 7815675). Magnesium-26 labels in 12 people with ileal resection. Overall absorption: diglycinate 23.5%, oxide 22.8%. In the four poorest oxide absorbers, diglycinate did better (23.5% vs 11.8%) and was better tolerated. That is a niche gut paper. It is not a finding that glycinate always absorbs more in healthy adults.

Coudray 2005 (PMID 16548135). Ten salts in magnesium-depleted rats. Organic slightly above inorganic; gluconate highest. Animal ranking context. Not a human dose.

ODS summarizes the human file the way we will: dissolve well, absorb better; aspartate, citrate, lactate, chloride above oxide and sulfate.

Scores by form and job

The chips above are the claims we will defend on the three-point scale. This table keeps the brief’s full wording, including Good.

Form by form

Oxide. Cheap. About 60% elemental by weight. Poorly absorbed. Often a laxative or an antacid. Excellent when loose stool is the point. Poor when the point is raising status. Carbonate and hydroxide sit in the same osmotic neighborhood.

Citrate. Better absorbed than oxide in Lindberg, Walker, and Kappeler. Can loosen stool. Good for general supplementation and for constipation. The default pill if you want elemental milligrams you will actually take up.

Glycinate / bisglycinate. Generally well tolerated. Good for raising status as a comfortable daily salt. Good to Excellent for GI comfort. Poor as a laxative. Fair for sleep or anxiety superiority: the ads got there first. Head-to-head glycinate versus citrate RCTs in healthy adults for status plus sleep are sparse. We will not invent them.

Chloride and lactate. On the ODS “better absorbed” list. Firoz put both above oxide and next to aspartate. Good for status. Gut comfort is Fair, sometimes Fair to Good for lactate.

Malate. About 15 to 16% elemental. “Energy” copy is mostly marketing and mechanism talk. Human form-specific trials are thin. Good for status only with that caveat.

L-threonate (Magtein). About 8% elemental. Products often deliver on the order of 70 to 150 mg elemental magnesium a day at gram-level servings. Liu 2016 (PMID 26519439): 12-week RCT, 44 completers, ages 50 to 70. Composite cognition improved versus placebo. Sleep (PSQI) and anxiety (HAM-A) did not beat placebo. Funded by Neurocentria, which co-designed the work and analyzed cognitive endpoints. Hausenblas 2024 (PMID 39252819) reported sleep and daytime-functioning gains with AIDP / Magtein-marketer funding. Fair for cognitive claims. Fair as a costly way to buy elemental milligrams.

Sulfate, oral. Poorly absorbed; can loosen stool. Fair to Poor for status. Good as a laxative salt.

Sulfate, topical (Epsom). Gröber, Schmidt, and Kisters (Nutrients 2017, PMC5579607): evidence is insufficient to recommend transdermal magnesium to raise status. Poor for that job. A warm bath can still help you sit still. That is not a serum protocol.

Carbonate. Antacid-class. Fair across the board.

Use-cases, not personalities

Constipation. Citrate, oxide, or hydroxide. You are buying unabsorbed salt on purpose.

A sensitive gut that still needs a pill. Glycinate. Tolerance is the feature. Do not promote it as a proven sleep drug.

Migraine prevention. AAN/AHS 2012 (PMID 22529202): oral magnesium is Level B, probably effective for episodic migraine prevention. Doses in that literature are often about 300 to 600 mg elemental a day, which can sit above the 350 mg supplemental UL. That is a clinician conversation. Oxide and citrate both appear in the migraine file. We will not crown glycinate as the migraine form without trials that say so.

Sleep and anxiety marketing. Thin. Not form-proven. See the next section.

Cognition. Magtein is Fair, once, with the sponsor named. It is not “brain magnesium” as a class win.

Cramps and sleep, without the caption

Cramps. Garrison and colleagues, Cochrane 2020 (PMID 32956536): magnesium is unlikely to be a clinically meaningful prophylaxis for idiopathic cramps in older adults (moderate certainty). Pregnancy data are uncertain. Diarrhea is common. That is the osmotic effect of the salt you did not absorb, form-agnostic.

Sleep. Mah and Pitre 2021 (PMC8053283): three RCTs, 151 older adults. Sleep-onset latency about −17 minutes (95% CI −27 to −7). Total sleep time +16 minutes, not significant. GRADE low to very low. The pills in those trials were oxide and citrate, including Abbasi’s 500 mg elemental as oxide. That is not a glycinate victory. It is barely a magnesium-for-insomnia victory.

Food first, then the UL

The adult body holds about 25 g of magnesium. Half to 60% sits in bone. Less than 1% sits in serum. Normal serum is about 0.75 to 0.95 mmol/L. Food absorption is typically about 30 to 40%.

RDA, adults (ODS): men 400 to 420 mg; women 310 to 320 mg. Pregnancy and lactation run higher on the ODS table. Pumpkin seeds, chia, almonds, spinach, beans, and whole grains do real work. Fortified foods help. A capsule is for a gap, a restriction, a drug interaction, or a clinician-directed dose. It is not a personality.

UL: 350 mg a day from supplements and medications, not from food, for ages 9 and up. Diarrhea, nausea, and cramping are the usual first signals. ODS names carbonate, chloride, gluconate, and oxide as salts that often show up in that story.

If you want the mineral file next door, the vitamin D guide is about testing and dose honesty, not a secret magnesium synergy. We will not invent one.

Safety and interactions

Do not start a repletion dose if you have kidney disease or a clinician has limited magnesium. Unabsorbed salt that you cannot clear is not a wellness habit.

ODS interaction list, short:

• Oral bisphosphonates: magnesium can blunt absorption. Separate them.
• Tetracyclines and quinolones: timing. Do not swallow them as a stack.
• Loop and thiazide diuretics raise losses. Potassium-sparing diuretics can reduce excretion.
• Long-term PPIs can mean hypomagnesemia.
• Very high zinc (Spencer 1994, as ODS cites it: 142 mg a day) can impair magnesium balance.

A daily electrolyte habit is a sodium conversation first. We filed that at electrolytes, sports-drink hype, and the creatine habit. Adjacent supplement literacy sits in creatine loading vs 5 g a day and why we do not recommend creatine gummies. Those pages are not magnesium form trials.

Sister brand / OCN LLC. Rephora Labs electrolyte + creatine sticks list 75 mg magnesium (as magnesium oxide) per stick. That dose can contribute to daily intake. It is not optimized as a high-bioavailability repletion form. This article compares magnesium forms. It is not a Rephora product review.

The short version

Read elemental milligrams. Eat food. Stay under 350 mg supplemental unless a clinician is running a migraine or deficiency protocol.

Want a laxative? Oxide. Want a default pill? Citrate. Want a pill your gut will finish? Glycinate. Want sleep or a brain upgrade because the label said so? The trials are smaller and noisier than the ads. Magtein is Fair, with the company in the sentence. Epsom on the skin is Poor for raising status.

Nutritionist Guide is not your clinician. This page is not a dose table for kidney disease, migraine prevention, or a drug interaction. Do not treat a bathroom accident as proof the form “is working.”