MUMultiple speciesBmoderateT1hypothesis only
Serum magnesium captures under two per cent of body stores. What human and equine data really show on cramps, sleep and intestinal absorption.
Executive summary
Magnesium acts as a cofactor in more than 600 enzyme reactions, yet less than two per cent of body stores circulates in the blood. Serum values therefore stay normal for a long time, even when intake is tight. In humans, magnesium lowers blood pressure measurably; in older adults with nocturnal calf cramps it does not. In horses, ionised magnesium is the more sensitive measure. Intestinal uptake routes differ fundamentally between species.
Key points
Magnesium is the second most abundant positively charged particle inside the cell and acts as a cofactor in more than 600 enzyme reactions. Without bound magnesium, ATP, the energy carrier of every cell, is biochemically all but inert. A comprehensive physiological review from 2015 describes three levers that govern body stores: absorption in the gut, reclamation in the kidney and storage in bone. No hormone controls magnesium as tightly as parathyroid hormone controls calcium, in any of the species considered here.
That missing hormonal control is the key to the whole subject. The body balances magnesium essentially as a ledger: what comes in, set against what is lost through the gut, the kidney and, in lactating animals, the milk. If intake falls below outflow, stores decline with nothing to counteract it. With calcium it would be different.
At the nerve and muscle membrane, magnesium competes with calcium and thereby dampens excitability. When it drops, nerve cells fire spontaneously more readily. A matching pattern has been produced deliberately in horses: when ionised calcium was lowered to 0.48 and ionised magnesium to 0.44 millimoles per litre, all seven animals showed electrical signs of nerve overexcitability without appearing clinically abnormal. Note that both values fell together in this experiment, so the finding cannot be attributed to magnesium alone.
Less than one to two per cent of all magnesium sits in the blood and the extracellular space. The rest is in bone and inside cells, where a blood sample cannot reach. A 2026 review of status markers therefore concludes that no single laboratory value reliably reflects whole-body stores, and that the serum value can miss a subclinical depletion.
Just how thin the choice is emerges from a systematic review that examined 20 candidate markers drawn from 21 publications on supplementation and depletion experiments. Only three responded measurably to a change in intake at all: magnesium in serum or plasma, magnesium in red blood cells and urinary excretion. For the remaining 17, the data supported no verdict. Newer approaches such as the magnesium depletion score therefore combine laboratory values with clinical features rather than leaning on a single number.
In equine medicine, yes, at least for detecting a deficiency. In 35 horses undergoing colic surgery, total magnesium before the operation was below the reference range in 17 per cent of animals, ionised magnesium in 54 per cent. The total value was thus considerably less sensitive. In humans the ionised measurement is described as well, but it is not established in routine practice.
Magnesium crosses the intestinal wall by two routes. One passes through the cell, via the TRPM6 and TRPM7 channels, and is saturable. The other passes between the cells, simply follows the concentration gradient and has no upper limit. A classic human absorption study from 1991 separated the two components: as intake rose, the absorbed fraction fell from 65 per cent for the smallest portion to 11 per cent for the largest, while around 7 per cent continued to be taken up irrespective of the amount.
The same study yields a side observation that is often overlooked. Magnesium from almonds was just as available as magnesium from dissolved magnesium acetate, whereas enteric-coated capsules performed markedly worse. The dosage form therefore altered uptake more than the question of whether the magnesium came from food or from a supplement.
Because the cellular route has a capacity limit. Once it is saturated, only the route between the cells remains, and that lets through a constant small fraction. The absolute amount absorbed therefore keeps rising while the percentage falls. Medicines can disturb both routes: with long-term use of proton pump inhibitors, both a throttled transport through the cell and an altered permeability between the cells have been described.
The most striking difference between species concerns where absorption happens. In ruminants, magnesium is taken up predominantly across the rumen wall, by a transport process that depends on the electrical voltage at the membrane. When the potassium concentration in the rumen rises, that uptake collapses. This is precisely the described mechanism of grass tetany on potassium-rich spring pasture, and precisely why a higher magnesium intake works as prevention: above a certain concentration in the rumen, a second, voltage-independent uptake route takes over.
The horse has no forestomach. Absorption is attributed to the small intestine, resting above all on balance trials from the 1970s; at the molecular level, equine transport remains barely investigated to this day. The horse's large colon is a fermentation chamber, not a rumen: fermentation sits downstream of the main site of mineral absorption, not upstream of it.
| Species | Main segment | Transport principle | Documented influencing factor |
|---|---|---|---|
| Human | Small intestine, supplemented by the large intestine | saturable route through the cell plus a constant fraction between the cells | size of the single portion, enteric coating, proton pump inhibitors |
| Horse | Small intestine | captured in balance trials, barely investigated at the molecular level | body type: ponies appear to utilise magnesium better than large horses |
| Ruminant | Rumen as forestomach | voltage-dependent transport across the rumen wall, supplemented at high concentrations by a second route | high potassium in the rumen, sodium deficiency, ammonia peaks |
The comparison also pays off within a single species. A 2024 meta-analysis pooled 42 feeding studies in horses and ponies and found a clear difference: ponies lost less magnesium in the faeces and apparently utilised it better than large horses. On calculation, ponies needed about half the amount recommended until then, large horses 1.9 times that amount. The mean of the two groups matched the old recommendation. It is an object lesson in how an average can conceal two opposing findings.
Here the evidence is unusually clear, and it comes out negative. A 2020 Cochrane review analysed eleven randomised trials with 735 participants in total. In older adults with nocturnal calf cramps, the difference from placebo after four weeks was 0.18 cramps per week and was statistically indistinguishable from zero. The reviewers consider a clinically meaningful benefit in this group unlikely. Mild side effects, mostly diarrhoea, were reported more often on magnesium than on placebo, though with low certainty.
In pregnancy the picture is inconsistent. A second Cochrane review from the same year found eight small trials with 576 women whose results were reported so differently that pooling was impossible; the certainty of the evidence was rated low to very low. A 2026 meta-analysis was able to pool four pregnancy trials and found a benefit there, but none for nocturnal cramps in adults.
No, it means something more precise: an effect and the outcome it is measured against belong together. For blood pressure, a documented but small effect exists. An analysis of 34 double-blind trials with just over 2,000 adults produced an average reduction of 2.0 millimetres of mercury in the upper reading and 1.8 in the lower one. A further synthesis from 2024 found the larger reductions at relatively low daily amounts and with durations of use beyond three months.
A systematic review of insomnia in older people found only three randomised trials with 151 participants in total. Time to fall asleep was around 17 minutes shorter on magnesium than on placebo, but total sleep duration was not reliably different. All three trials had a moderate to high risk of bias, and the certainty of the evidence was rated low to very low.
A broader analysis of 31 randomised trials of dietary supplements concluded that amino acids, vitamin D and melatonin improved self-reported sleep quality, while the magnesium data were simply too thin to pool. More recent individual trials shift the picture only by degrees. In a study of 155 adults with poor sleep, insomnia severity fell by 3.9 points after four weeks against 2.3 on placebo, a small difference; strikingly, people with a low dietary magnesium intake benefited more. A three-week trial in 80 adults found better readings for deep and REM sleep using a measurement ring.
The most robust equine data come from the clinic, not from performance enhancement. In a prospective study of 75 sick horses, ionised magnesium was too low in 47 per cent of animals with an obstructive gastrointestinal disorder, against 4 per cent of the healthy comparison animals. Unlike in human intensive care patients, a low value here was not linked to mortality. That is a clean example of a prognostic association that cannot be copied across species.
That inflammation alone moves the blood value is shown by a controlled experiment with endotoxin. One hour of infusion lowered ionised magnesium from 0.53 to 0.43 millimoles per litre and at the same time throttled the fraction excreted in the urine from around 37 to 12 per cent. The body was therefore holding magnesium back while the blood value fell. A low reading does not automatically mean that too little was fed.
Yes, a small one. In six geldings with trigeminal-mediated headshaking, the frequency of episodes fell by 52 per cent on magnesium and by 64 per cent on magnesium with boron, compared with hay alone. The third number is what matters for interpretation: the pelleted base feed without added magnesium already reduced frequency by 44 per cent. Six affected animals and an active comparison arm make this a signal, not a statement of efficacy.
For the use of magnesium salts as a calming agent, widespread in competition, the literature so far offers pharmacology but no demonstrated efficacy. An intravenous dose raised plasma ionised magnesium in healthy mares 3.7-fold within five minutes, lowered ionised calcium in the process and measurably altered the calcium-regulating hormones. In another study, in five standing, unsedated horses, no change in pain thresholds for heat, pressure or electrical stimulation could be detected after a magnesium infusion.
The mechanism transfers. In humans, horses and ruminants, magnesium is the same enzymatic cofactor, dampens neuromuscular excitability everywhere and is not tightly regulated by a dedicated hormone in any of these species. One methodological insight transfers too: in all three species the blood value is a weak marker of body stores, because it does not reflect the large reservoirs in bone and cells. Anyone who has learnt to distrust the serum value in one species is right to do so in the others.
Numbers and efficacy claims do not transfer. The main site of absorption differs between rumen and small intestine. Reference ranges differ between laboratories and between species. And a negative result from a randomised human trial on calf cramps says nothing about a horse with a tight back, because no comparable trial exists in horses. The reverse holds as well: the headshaking study in six geldings is no argument for humans.
Above all, an order in which to ask the questions. First establish which species is at issue and what evidence exists there at all. Then check whether a deficiency is plausible and what it would be pinned to, because a single serum value is not enough for that. Only then does the question of supplementation become meaningful, and that decision belongs with the treating doctor or veterinary surgeon. This page does not replace an examination and deliberately names no amounts.
Original analysis
| Guiding question | Human | Horse | Ruminant | Strongest study type in the corpus |
|---|---|---|---|---|
| Absorbed fraction measured quantitatively | yes, depending on portion size | yes, in balance trials | yes, in the rumen model | controlled absorption experiment |
| Blood value tested as a marker of body stores | yes, 20 markers compared | yes, ionised versus total | indirectly, via tetany cases | systematic review |
| Effect on muscle cramps tested in randomised trials | yes, eleven trials | no | no | meta-analysis |
| Effect on sleep tested in randomised trials | yes, three trials in older people | no | no | meta-analysis |
| Consequences of a deliberately induced deficiency described | only to a limited extent | yes, electromyographically | yes, clinically as grass tetany | controlled animal experiment |
| Requirement recommendation recalculated quantitatively | not in this corpus | yes, separately for pony and horse | not in this corpus | meta-analysis |
It answers a narrower question than many people expect. The serum value captures the under two per cent of body stores that circulates in the blood, and it can stay normal even as the reservoirs shrink. It is useful when an acute fall is suspected, for instance in bowel disease, on certain medicines or during inflammation. As a screen for a marginal diet it says little. A systematic review found that, of 20 markers examined, only three responded to a change in intake at all.
In older adults with ordinary nocturnal calf cramps, the best available synthesis points to no. The Cochrane analysis of eleven randomised trials found a difference of 0.18 cramps per week against placebo after four weeks, which amounts to practically no difference. Mild side effects such as diarrhoea were reported more often on magnesium, though the certainty of that comparison is low. In pregnancy the situation is different and inconsistent: there, a meta-analysis of four trials found a benefit, while a Cochrane review rated the certainty of the evidence as low to very low.
That is not established. Magnesium salts are used for this purpose in competition, and the pharmacological effect is well documented: an intravenous dose multiplied plasma ionised magnesium in healthy mares within five minutes and lowered ionised calcium at the same time. In standing, unsedated horses, by contrast, no change in pain thresholds could be demonstrated. Controlled studies on the behaviour of healthy horses are missing from the corpus analysed here. The only controlled dietary trial concerned headshaking and involved six animals.
Because the site of absorption is different. In ruminants, magnesium is taken up predominantly across the rumen wall, by a transport process that depends on the electrical voltage at the membrane. Potassium-rich spring grass shifts that voltage and slows uptake, so blood levels fall and overexcitability and cramps can appear. The horse has no forestomach; its magnesium uptake is attributed to the small intestine, where this potassium-sensitive mechanism has not been described in that form.
The most robust answer concerns the portion size and the coating rather than the salt. In a human absorption study, the absorbed fraction fell from 65 per cent for a small single dose to 11 per cent for a large one, because transport through the intestinal cell is saturable. Magnesium from almonds was just as available as magnesium from a dissolved magnesium compound, whereas enteric-coated capsules performed markedly worse. Comparisons between individual organic salts are far less robust in the literature than the advertising suggests.
For the mechanism yes, for numbers and efficacy no. That magnesium dampens neuromuscular excitability holds equally in humans and horses, and in horses it has even been made visible electromyographically. Everything beyond that is species-specific: the site of absorption, the reference ranges, the requirement levels and the clinical consequences. A low ionised magnesium said nothing about mortality in hospitalised horses, whereas in human intensive care patients it is regarded as an unfavourable sign. The same measurement therefore does not carry the same meaning.
Comparative Health
Radiocarbon data suggest the Achilles tendon core forms in growth and is barely renewed thereafter: what that means for healing, scar and the equine t
MUMultiple speciesBmoderateT2limited target data Open access
Comparative Health
Serum amyloid A rose about 227-fold in an equine experiment, haptoglobin only 1.1-fold. How inflammation starts, how it ends, what CRP and SAA measure
MUMultiple speciesBmoderateT2limited target data Open access
Equine Health
What counts is not visible dust but the respirable fraction in the horse's breathing zone: figures from trials on hay, steaming and haylage.
MUMultiple speciesBmoderateT1hypothesis only Open access
Once a month: new syntheses, checked studies, and what has changed in the state of the evidence.