Research, made readable
Which magnesium is better for sleep?
How to read the bisglycinate and L-threonate placebo-controlled trials without declaring a winner.
Published: · Sources checked:
The short answer
The trials reviewed here compared each product with placebo, not with the other form. They do not establish which is better for sleep. Compare the study outcomes and magnesium amounts separately.
Prepared with AI assistance from the linked sources. Not independently reviewed by a clinician. Source access and study limitations are stated in each article.
What the research says
Bisglycinate: a small change in self-reported insomnia
- Who and what
- Schuster et al. (2025) randomized 155 adults aged 18–65 reporting poor sleep. For four weeks, the active group received 250 mg elemental magnesium daily as bisglycinate, supplying about 1.5 g glycine, versus placebo. The primary outcome analysis included 153 participants who completed the baseline questionnaire (77 in the magnesium group and 76 on placebo); 134 participants completed all four weeks.
- What changed
- The statistical model estimated mean reductions in Insomnia Severity Index scores of 3.9 points with magnesium and 2.3 with placebo. The between-group difference in mean change was 1.6 points (p=0.049; small effect, Cohen’s d=0.2). This average difference is not a prediction of an individual’s improvement or a treatment success rate. Other psychological outcomes did not differ significantly.
- What this cannot tell us
- Sleep was self-reported, without actigraphy or polysomnography. The study cannot separate magnesium’s effect from the glycine supplied with it. It excluded several conditions and medicines, including renal insufficiency and sleep-affecting medication.
- Sources and disclosure
- Trial full texts and the threonate correction were reviewed. Funded by the University of Hannover institute conducting the study. One author disclosed nutraceutical-industry research and honoraria relationships.
L-threonate: selected measures improved, not every outcome
- Who and what
- Hausenblas et al. (2024) randomized 80 adults aged 35–55 with self-reported sleep problems; 76 were analyzed. The 21-day trial compared 1 g/day of magnesium L-threonate (about 75 mg elemental magnesium) with placebo, using questionnaires and an Oura ring.
- What changed
- Some wearable sleep scores and daytime measures favored threonate. For overall insomnia severity, both groups improved and the treatment-by-time interaction was not significant (p=0.39). Selected positive scores do not mean every sleep outcome improved.
- What this cannot tell us
- A short trial in selected adults; wearable scores are not a laboratory sleep study. The 2025 correction states that the trial was registered retrospectively, after it had been conducted.
- Sources and disclosure
- Trial full texts and the threonate correction were reviewed. AIDP commissioned and funded the trial. The correction discloses patents and paid relationships for three authors; it does not report a change to the trial results.
Connect the evidence to your own questions
Neither trial tested these two forms together. Different participants, doses, durations and outcome measures prevent a simple winner-by-numbers comparison. Both studies were short and involved selected adults; they do not settle long-term safety or treatment of a diagnosed sleep disorder. Study amounts are not personal dosing instructions. Compound weight and elemental magnesium are different quantities.
Related ingredient guides
Read the detailed evidence noteThis opens the checker without selecting ingredients or adding doses. It checks the rules in our dataset; it cannot predict whether a supplement will work for you or prove a combination is safe.
Article history
· Clarified the 155 randomized, 153 analysed and 134 completed participants and the meaning of the 1.6-point mean difference. Rechecked the previously obtained full-text record; no new online source access or treatment conclusion.
· First published. Focused source reading; no clinical review.
We update an article when a relevant trial, correction or safety finding changes the interpretation. The history records substantive changes; a routine check alone does not refresh the date.