The Most-Cited Magnesium Sleep Study Used Oxide — the Form You're Told to Avoid
The internet's #1 magnesium rule says only glycinate absorbs — about 4x better than "useless" oxide. The famous insomnia trial dosed oxide. The human isotope study measured 23.5% vs 22.8%. And the only three-way head-to-head crowned citrate.
Sterling eyes a single glowing magnesium capsule against pure black
If you have shopped for magnesium, you have met the rule: buy glycinate, because it is the only form that actually absorbs — roughly four times better than that useless magnesium oxide. It is repeated so often it reads as settled science. We went looking for the study behind the number.
The famous sleep trial used oxide
The trial everyone cites for magnesium and sleep is Abbasi 2012, a double-blind, placebo-controlled study in 46 older adults with insomnia. Its own methods section is unambiguous: magnesium was administered as magnesium oxide tablets twice a day, each containing 414 mg of oxide as 250 mg of elemental magnesium, for eight weeks. The study that launched a thousand magnesium-for-sleep recommendations ran on the form the internet tells you to skip.
The "4x absorbed" number is not a sleep number
The human study most often invoked for absorption is Schuette 1994, which used isotope-labelled magnesium to compare a glycine chelate against oxide directly. Its result, verbatim: absorption "was not different for the two supplements" — 23.5% versus 22.8%. That is a ratio of 1.03, not 4. And it measured a mineral in the blood, not a single minute of sleep.
The only head-to-head crowned citrate
Walker 2003 ran the one randomised three-way comparison of amino-acid chelate, citrate and oxide — 300 mg/day of elemental magnesium in 46 healthy adults over 60 days. Its conclusion was that citrate showed superior bioavailability. The cheap-ish middle option beat the premium chelate, and oxide produced no difference from placebo.
So is glycinate a scam? No
Glycinate finally got its own sleep trial in 2025: 155 adults with poor sleep, 250 mg elemental as bisglycinate. Insomnia scores did fall further than placebo — 3.9 points versus 2.3, p = 0.049. Real, and worth knowing. But the authors themselves describe the effect as small, with a Cohen's d of 0.2, and it was measured at week four only.
The honest read is not that glycinate is worthless. It is a well-tolerated, gentle-on-the-stomach form with a modest but genuine sleep signal. Buy it for that. The four-times multiplier is the part with no human evidence behind it — and paying a large premium for a 1.03x difference is the mistake worth avoiding.
SOURCES
- Abbasi B et al. — J Res Med Sci 2012: double-blind, placebo-controlled RCT in 46 elderly with primary insomnia; dosed magnesium OXIDE (414 mg oxide as 250 mg elemental) twice daily for 8 weeks (PMID 23853635)
- Schuette SA et al. — JPEN 1994: 26Mg-labelled human absorption study; "26Mg absorption was low but was not different for the two supplements (23.5% vs 22.8% for magnesium chelate and MgO, respectively)" (PMID 7815675)
- Walker AF et al. — Magnes Res 2003: the only randomised three-way comparison of amino-acid chelate vs citrate vs oxide, 300 mg/day elemental for 60 days in 46 adults; citrate showed superior bioavailability (PMID 14596323)
- Schuster J et al. — Nat Sci Sleep 2025: randomised placebo-controlled trial of magnesium bisglycinate (250 mg elemental) in 155 adults with poor sleep; ISI -3.9 vs -2.3 (p = 0.049), effect size small (Cohen's d = 0.2) (PMID 40918053)