Magnesium forms compared: glycinate, citrate, threonate, oxide

There is no single best magnesium. The right form depends on what you want and how your gut copes. So match the form to the purpose instead of hunting for one winner.

Think of it as four tools, not a ranking.

• Bisglycinate (also written glycinate): the gentle all-rounder. Easy on the stomach, the sensible default if your goal is sleep or stress.
• Citrate: well-absorbed but mildly laxative. That side effect is a feature if you also want help with constipation.
• Oxide: cheap and everywhere, but mostly a laxative. It barely raises your body's magnesium.
• Threonate (sold as Magtein): premium-priced and marketed for the brain. The human evidence is thin, so treat it as experimental.

Form	Absorption	Laxative effect	Best for	Price	Evidence
Bisglycinate	Good; an amino-acid chelate beat oxide in the key trial [1]	Lowest	Sleep, stress, sensitive gut	Mid	One RCT in poor sleepers, small effect [18]
Citrate	Highest in the key head-to-head trial [1]	Mild	Constipation plus topping up	Mid	Absorption RCT [1]; migraine trials used dicitrate [15]
Oxide	About 4% [2]; no better than placebo in Walker 2003 [1]	Strong	Cheap occasional laxative	Low	Wrong pick to raise your magnesium
Threonate (Magtein)	Bioavailable, but only ~8% magnesium by weight [17]	Well tolerated in trials [4]	Brain use, as an experiment	Premium	Small, maker-linked trials [4, 13, 16]

Why spell this out? Because most magnesium marketing does two dishonest things. It oversells threonate as a brain upgrade based on rat studies. And it quietly sells oxide as a default, when oxide is the worst pick if absorption is your aim.

The rest of this guide separates the real effects from the inflated ones using the actual numbers. You will see what randomised controlled trials genuinely show for sleep, cramps, migraine, blood pressure and anxiety. RCTs are the gold-standard design: people get either the supplement or a dummy pill, assigned by chance. Some of those claims hold up. Several do not.

A quick reality check before you read on. For most healthy people eating a normal diet, true magnesium deficiency is uncommon. So a lot of the question is not which pill, but whether you need one at all. We cover that at the end with the D-A-CH reference intakes. For now, the takeaway is simple: pick the form that fits your goal and your gut, and ignore the hype around the expensive one.

Oxide absorbs badly mainly because it barely dissolves in your gut. Being inorganic is not the problem: magnesium chloride, also inorganic, absorbed as well as the organic forms in Firoz and Graber 2001 ^[2]. Organic and chelated forms (citrate, bisglycinate, malate) dissolve better and get into your bloodstream more readily. A chelate just means the magnesium is bound to an amino acid, which helps it cross the gut wall.

The number you see everywhere is that oxide is only about 4% absorbed. That figure traces to the same Firoz and Graber study ^[2], which compared US magnesium products and found oxide markedly less bioavailable than chloride, lactate and aspartate. Those three performed similarly to each other and left oxide trailing.

The key head-to-head trial is Walker 2003 ^[1], a 60-day RCT in 46 people taking 300 mg of elemental magnesium a day. Citrate came out on top. It produced the highest serum magnesium (the level in your blood) and the highest salivary magnesium. Both citrate and an amino-acid chelate beat oxide on urinary magnesium by day 60, which signals more magnesium actually entered the body. And oxide? No better than placebo.

So in a proper trial, the cheapest and most common form did not beat a sugar pill on the absorption marker. That is the case against using oxide to raise your magnesium.

A 2021 systematic review (Pardo et al. ^[12]) points the same way. Inorganic forms appear less bioavailable than organic ones, and how much you absorb depends on the dose. The same review adds a fair counterpoint: for healthy people without a deficit, any magnesium supplement can keep blood levels normal. So for most healthy people, gut tolerance matters more than chasing the top absorber.

The practical takeaway: stop thinking of oxide as an absorption vehicle. Think of it as an osmotic laxative. It pulls water into the bowel, which is why it relieves constipation and why it gives people loose stools. If your goal is more magnesium in your tissues, citrate or bisglycinate get you there. If your goal is a cheap, occasional laxative, oxide does that job. Just do not confuse the two.

One real warning if you use oxide as a regular laxative. When your kidneys work less well, which is more common in older age, magnesium can build up in your blood. In a Japanese hospital study of 320 patients prescribed oxide, 23% developed raised blood magnesium, and reduced kidney function was one of the main risk factors ^[20]. Check with your doctor before you take it long term.

Match the form to the goal. Bisglycinate for sleep and stress, because it is gentle on the gut. Citrate for constipation, because it is well-absorbed and mildly laxative. Oxide only as a cheap laxative, since it barely raises body magnesium (Walker 2003 ^[1]). Threonate for the brain only as an experiment: premium-priced, with thin human evidence. Here is each one in more detail.

For sleep or stress: bisglycinate. It is gentle on the gut, so you can take a useful dose without running to the toilet. That tolerance is the main reason to pick it, not better sleep data than other forms. Schuster 2025 ^[18], a university-funded trial from Leibniz University Hannover, gave 155 adults with poor sleep 250 mg of magnesium a day as bisglycinate. After 4 weeks their insomnia score dropped 3.9 points, versus 2.3 on placebo. That just cleared significance (p=0.049), and the effect was small (d=0.2). People who ate little magnesium to begin with seemed to gain the most. We break down the other sleep trials in the section below. If sleep is your real goal, our deep sleep guide covers the levers that move the needle more than any pill.

For constipation (with some absorption): citrate. It is well-absorbed, so it raises your magnesium, and it is mildly laxative, so it loosens stools. That dual action makes it the genuine two-in-one pick. If you are prone to loose stools already, this is the form most likely to give you diarrhoea.

For a cheap laxative only: oxide. What it does well is relieve occasional constipation on a budget. Do not buy it expecting to correct low magnesium, because Walker 2003 ^[1] showed it performs no better than placebo on absorption markers. Wrong tool for that job. If your kidneys are weak, read the warning in the absorption section before you use it regularly.

For the brain or memory: threonate, with heavy caveats. Sold as Magtein, it is premium-priced and marketed as a cognition and sleep upgrade. Treat it as experimental, not proven. The human evidence is small and industry-linked, and we unpack exactly why in the next section.

A few practical notes:

• If you have no specific goal and just want to nudge a low-ish intake upward, bisglycinate or citrate are both reasonable. Bisglycinate if your gut is sensitive, citrate if you would not mind the laxative nudge.
• Malate is another well-absorbed organic option, sometimes marketed for energy or muscle, though the goal-specific RCT evidence is thinner than for the forms above.
• Whatever you pick, check the elemental magnesium per dose on the label, not the total compound weight. They are not the same number, and that trips up a lot of buyers.

Short version: probably not worth the premium. Threonate is the most over-marketed and least-substantiated form, and the brain claim rests on shaky ground.

Start with the famous mechanism. The whole "raises brain magnesium and improves memory" story comes from Slutsky 2010 ^[3] in Neuron. It is a beautiful study. It is also a rat study. In rodents, magnesium-L-threonate raised brain magnesium, increased synaptic connections and improved memory in young and aged animals. None of that proves the same happens in your head. Animal results do not prove a human benefit, so this can never be presented as a proven cognitive upgrade. The marketing presents it that way anyway.

So what do human trials show? The best one to date is Lopresti and Smith 2026 ^[4], a double-blind RCT in 100 adults taking 2 g a day for 6 weeks. The effects were modest. A combined cognition score improved more than with placebo. So did a score for sleep-related daytime impairment. Both results were borderline (around p=0.04), far from the strong p<0.001 you would want for a confident claim. Core sleep disturbance did not change. Neither did sleep architecture, the pattern of light, deep and REM sleep, measured objectively with an Oura ring. What that means: people felt slightly less foggy and slightly less impaired by poor sleep during the day, but their measured sleep did not improve.

There is a bigger asterisk. The manufacturer, Threotech Inc., funded this trial and supplied the product. That does not make it fraudulent. But it is a known source of bias. It means the most positive human data on threonate comes from the company selling it.

One number the labels bury: threonate is a heavy carrier. EFSA's 2024 safety review puts 3,000 mg of magnesium L-threonate at about 250 mg of actual magnesium, roughly 8 percent ^[17]. So the 2 g a day in the Lopresti trial delivered only around 170 mg of magnesium. EFSA judged it safe as a novel food at up to 3 g a day for adults, under conditions of use that exclude pregnant and breastfeeding women ^[17].

The earlier trials are small and run by people tied to the product. Liu 2016 ^[13] (the MMFS-01 trial) tested 44 older adults and reported a large cognition gain, but three of its five authors, including the lead author, worked for the developer, Neurocentria. Hausenblas 2024 ^[16] tested 80 adults, with half its authors from the ingredient supplier AIDP, and reported better Oura deep and REM sleep scores. The larger 2026 trial found no such change in measured sleep.

The verdict: do not pay a threonate premium expecting brain or sleep results that the human evidence does not yet support. If you want to try it, fine, but go in clear-eyed. For most people chasing better sleep, bisglycinate at a fraction of the price is the more sensible bet. Its sleep trial was also not paid for by a manufacturer (Schuster 2025 ^[18]).

Migraine prophylaxis (daily prevention) has the strongest goal-specific evidence: about 600 mg a day, rated Grade C ^[8]. Blood pressure shows a small real effect, around 2 mmHg (Zhang 2016 ^[7]). Sleep gains are modest, about 17 minutes faster to fall asleep (Mah 2021 ^[5]). For leg cramps, the best review finds no meaningful benefit. For anxiety, the evidence is too poor to lean on. Here is each goal in turn.

Sleep: modest and mixed. Mah and Pitre 2021 ^[5] reviewed 3 RCTs in 151 older adults with insomnia. They pooled 2 of those trials (n=55). There, sleep-onset latency, the time it takes to fall asleep, dropped about 17 minutes versus placebo (p=0.0006). Total sleep time rose around 16 minutes, but that was not significant (p=0.15). The evidence was graded low certainty. Abbasi 2012 ^[14] is one of the pooled trials. One catch: both pooled trials used magnesium oxide at 500 to 729 mg of elemental magnesium a day, two to three times the 250 mg supplemental limit. No trial in this review used bisglycinate. So the 17 minutes is an effect of high-dose magnesium, not proof that one form beats another. The newest trial, Schuster 2025 ^[18] in adults aged 18 to 65 taking bisglycinate, fits the pattern: a small drop in insomnia scores that only just reached significance (p=0.049). So magnesium may help you nod off a touch faster, mostly if you are older or low on it. Do not expect a transformation.

Leg cramps: largely null. This one contradicts the popular claim. The Garrison 2020 Cochrane review ^[6] found no significant benefit for idiopathic (no known cause) or nocturnal leg cramps in older adults. The differences versus placebo were small and not significant. If a shop assistant tells you magnesium fixes night cramps, the highest-quality evidence does not back them. Pregnancy cramps are a separate question: the trials there conflict, so talk to your midwife or doctor.

Migraine prophylaxis: the strongest goal-specific signal. Prophylaxis means taking something daily to prevent attacks. Von Luckner and Riederer 2018 ^[8] rate high-dose oral magnesium as Grade C, "possibly effective". The dose behind that rating is 600 mg of elemental magnesium a day (24 mmol), given as trimagnesium dicitrate in trials like Peikert 1996 ^[15]. Chiu 2016 ^[9] found oral magnesium reduced attack frequency and intensity, with odds ratios around 0.20 and 0.27 (below 1 means a better result than placebo). It is cheap and generally well tolerated, though about 1 in 5 people in Peikert 1996 got diarrhoea (18.6%). It is not first-line therapy either, meaning not what a doctor would try first.

Blood pressure: real but small. Zhang 2016 ^[7] pooled 34 RCTs (2028 people, around 368 mg/day). Systolic pressure fell about 2.0 mmHg and diastolic about 1.8 mmHg. That is an adjunct effect, helpful alongside other measures, not a hypertension treatment on its own.

Anxiety and stress: the weakest claim. Boyle 2017 ^[10] reviewed 18 studies and found only suggestive benefit in vulnerable subgroups (mild anxiety, PMS, high blood pressure). The authors rated the evidence quality as poor. So the calming-pill story is the least supported of all the marketed claims.

The pattern: migraine has the best case, blood pressure a small real effect, sleep a modest one. Cramps are largely hype, and the anxiety claim rests on poor evidence.

The D-A-CH reference values (DGE 2021 ^[21]) and EFSA 2015 ^[11] agree: about 300 mg a day for women and 350 mg a day for men. Technically it is an "adequate intake": a best-estimate target, not a strict minimum. In Germany's national nutrition survey, the typical adult already eats about that much or more ^[19]. Genuine deficiency is uncommon in healthy people eating normally. So before buying anything, ask whether you actually need it.

Some people genuinely are at higher risk of running low:

• Long-term use of proton pump inhibitors (PPIs, common acid-reflux drugs)
• Regular high alcohol intake
• Loop or thiazide diuretics (water tablets, often prescribed for blood pressure or heart conditions)
• Gut diseases that impair absorption

This matters for older DACH users, who are more likely to take several of these at once.

Now dosing and tolerance. The amount of usable magnesium differs by form, because the label compound includes the carrier. The dose-limiting side effect is diarrhoea, worst with oxide and citrate. That is why the EU Scientific Committee on Food (SCF, 2001) set a tolerable upper intake level of 250 mg a day for supplemental magnesium, on top of what you eat. Germany's Federal Institute for Risk Assessment (BfR) recommends the same figure as the maximum per supplement, split into two or more doses a day ^[19]. Both aim to avoid the osmotic-laxative effect. It is advice, not law.

Notice the tension here. German, Austrian and Swiss products often sell single doses of 300 to 400 mg, above that 250 mg supplemental guidance. In a 2020 market check, Germany's consumer advice centres found 57 percent of magnesium supplements above it ^[22]. And the migraine dose that the evidence supports, 600 mg of elemental magnesium as dicitrate (von Luckner 2018 ^[8]), sits well above that limit. Discuss that higher dose with your doctor rather than dosing yourself, especially if you take other medications or your kidneys are weak ^[20].

Start with food. You can cover most of your needs from:

• Nuts and seeds (pumpkin seeds are especially rich)
• Legumes (beans, lentils, chickpeas)
• Whole grains
• Leafy greens
• Dark chocolate

If you do supplement, match the form to your purpose and your gut tolerance. Bisglycinate for sleep or a sensitive stomach. Citrate if you also want a constipation nudge. Oxide only as a budget laxative. Threonate only if you are happy to experiment. Do not chase a single best magnesium, because it does not exist. For how magnesium fits next to the handful of supplements with real evidence, see our longevity supplements guide.