The form of magnesium matters

Form makes a measurable difference to how much magnesium is absorbed from a dose, and the clearest case is magnesium oxide, which is absorbed less completely than soluble forms such as citrate, chloride, lactate and aspartate. In small studies that track magnesium in urine after a dose, oxide usually comes out lowest, with estimates from urine of about 4 to 10 percent of the dose. Beyond oxide, differences between forms are smaller and less consistent.

What has not been shown is that this difference changes anything a person takes magnesium for. The largest trial comparing forms gave 164 adults citrate, oxide, sulfate or a placebo at 450mg a day for 24 weeks. Citrate raised urinary magnesium somewhat more than the other two, blood magnesium rose by about the same amount on all three, and none of them changed artery stiffness or blood pressure, the outcomes the trial was built to test. People taking oxide reported fewer stomach and bowel complaints than people taking citrate. A search of the medical literature found no trial comparing forms on sleep, anxiety, cramps or migraine.

The absorption studies are small and use different yardsticks. In a 1990 study citrate dissolved readily even in plain water while oxide barely dissolved, and the rise in urinary magnesium over the four hours after a single large dose was several times higher for citrate. A 2001 study of four US products put oxide at about 4 percent absorbed and found chloride, lactate and aspartate higher and equal to one another; chloride is an inorganic salt, so that result separates oxide from the rest rather than organic forms from inorganic ones. A 2003 trial in 46 healthy adults found citrate and an amino-acid chelate raised urinary magnesium more than oxide over 60 days, with no difference in red-cell magnesium. Two systematic reviews, from 2021 and 2025, each gathered 14 comparison studies and read them the same way, with oxide the least available form, without pooling the results into a single figure.

Not every result points the same way. In a 2012 crossover trial, 41 healthy adults took oxide and citrate for a month each; magnesium inside cells rose on oxide and not on citrate, though the oxide dose was larger, 520mg against 296mg a day. In a 2024 trial, 40 healthy adults took single 375mg doses of oxide, citrate and bisglycinate; blood magnesium over the next six hours did not differ between the three, and bisglycinate produced no significant rise at any time point. A 2017 review of this literature describes the type of salt as less relevant than often thought, and dose and existing magnesium status as more important. Magnesium levels are tightly regulated and higher absorption is usually followed by higher excretion in urine, so better absorbed does not necessarily mean more retained.

For glycinate and L-threonate, human comparisons remain scarce. Besides the 2024 trial, one of the few other human comparisons of a glycinate form used an isotope tracer in 12 patients who had part of their small intestine removed. Across the group, absorption matched oxide, 23.5 percent against 22.8 percent; in the four patients who absorbed oxide worst, diglycinate was absorbed about twice as well, and it was better tolerated by all of them. For L-threonate, the European Food Safety Authority concluded in 2024 that magnesium is bioavailable from it. The comparative data in that opinion come from rats, and the one human study submitted indicated that magnesium was absorbed, not that it was absorbed better than other forms.

One labelling detail from the NIH Office of Dietary Supplements: the magnesium amount on a Supplement Facts panel is elemental magnesium, not the weight of the whole compound, so the figures on labels for different forms describe the same thing. The same fact sheet summarises small studies as showing aspartate, citrate, lactate and chloride absorbed more completely than oxide and sulfate, and does not mention glycinate or threonate.