Magnesium for sleep: what the 2025 RCT data actually shows.

Why magnesium and sleep are plausibly connected

Magnesium (Mg) is the fourth most abundant mineral in the human body and a cofactor in over 300 enzymatic reactions, including nearly every step of ATP (adenosine triphosphate) synthesis, the energy currency that powers cellular work. It is also a structural regulator of neuronal excitability, and that is where the sleep connection begins.

Dietary surveys consistently show that magnesium intake falls short of recommended amounts across modern populations. NHANES (National Health and Nutrition Examination Survey) data from the United States indicates that roughly 48% of adults consume less magnesium than the estimated average requirement. The shortfall is skewed toward older adults, who both absorb magnesium less efficiently and excrete more of it renally. This is one of the more robustly documented nutritional deficits in the Western diet.

The connection to sleep is not direct cause-and-effect in the way that melatonin operates as a circadian signal. It is more nuanced: inadequate magnesium status appears to lower the threshold for neuronal hyperexcitability, elevate cortisol responsiveness, and disrupt the signaling systems that prepare the brain for sleep onset. Repleting a deficiency produces measurable changes in sleep architecture. Whether supplementing a person who is already replete produces the same effect is a different and mostly unanswered question.

That distinction is where most content around magnesium and sleep fails. The evidence supports magnesium supplementation in adults with low magnesium status, older adults with documented insomnia, and populations at high risk of depletion. It does not straightforwardly support the idea that adding magnesium to an already-adequate diet will meaningfully shift sleep in a healthy young adult. The trials reviewed below help clarify who benefits and by how much.

The mechanistic case: GABA, NMDA, melatonin

Three primary pathways link magnesium status to sleep physiology.

GABA-A receptor potentiation. GABA (gamma-aminobutyric acid) is the primary inhibitory neurotransmitter in the CNS (central nervous system), and GABA-A receptors are the target of most pharmaceutical sleep aids. Benzodiazepines and Z-drugs such as zolpidem both work here. Magnesium potentiates GABA-A receptor activity, meaning adequate intracellular magnesium helps the brain's natural inhibitory system function. When magnesium is low, GABAergic tone decreases and the CNS tends toward hyperexcitability, producing the racing-thoughts, difficult-to-quiet-arousal pattern characteristic of sleep-onset difficulty [1].

NMDA receptor antagonism. NMDA (N-methyl-D-aspartate) receptors are the major excitatory counterpart to GABA-A, driven by glutamate. Magnesium ions sit within NMDA receptor channels at resting membrane potential, physically blocking them. This voltage-dependent block is a fundamental mechanism of neuronal inhibition, impaired when intracellular magnesium is low. In animal models, magnesium deficiency increases wakefulness, reduces slow-wave sleep (N3), and is reversed by magnesium repletion [1].

Melatonin and cortisol modulation. Magnesium participates in the enzymatic conversion of serotonin to melatonin via AANAT (arylalkylamine N-acetyltransferase), the rate-limiting enzyme in melatonin synthesis. Animal studies show that dietary magnesium deficiency measurably decreases plasma melatonin, and repletion restores it. Separately, magnesium attenuates HPA (hypothalamic-pituitary-adrenal) axis activity, the cortisol-producing stress response system. Elevated evening cortisol is one of the most consistent findings in hyperarousal insomnia, and magnesium's role in blunting this cascade is a plausible contributor to its sleep effects [2].

The mechanistic case is solid. What it tells us is that magnesium status matters for sleep physiology, not that flooding the system with supplemental magnesium produces proportional benefit in already-replete individuals.

The 2025 RCT: 155 adults, bisglycinate, placebo-controlled

In 2025, Schuster and colleagues published what is currently the largest randomized, double-blind, placebo-controlled trial of magnesium bisglycinate specifically for sleep in healthy adults reporting poor sleep quality [3]. The trial enrolled 155 participants, randomized to either 250 mg of elemental magnesium delivered as bisglycinate (which also delivers approximately 1,500 mg of glycine as the chelating amino acid) or a matched placebo.

The primary outcome was the ISI (Insomnia Severity Index), a validated 7-item questionnaire scored 0 to 28. A reduction of 3 or more points is the accepted threshold for clinical meaningfulness. The trial reported a statistically significant reduction in ISI scores in the magnesium group versus placebo, with most improvement appearing within the first 14 days and maintained through the end of the treatment period.

Several features of this trial deserve close examination rather than a headline-level read:

• The glycine contribution. Magnesium bisglycinate delivers approximately 1,500 mg of glycine alongside the elemental magnesium. Glycine has its own independent sleep-promoting data: a separate 3 g glycine dose before bed reduces sleep-onset latency and improves subjective sleep quality in controlled trials [4, 5, 8]. It is not possible from this trial design to disentangle how much of the observed improvement was attributable to magnesium specifically versus glycine specifically. This materially affects how to interpret the conclusion.
• Self-reported outcomes only. The ISI is validated but subjective. PSG (polysomnography), the gold standard measuring sleep architecture via EEG (electroencephalography), was not used. Self-reported improvements are meaningful for quality-of-life outcomes but cannot confirm changes in sleep stages or objective sleep efficiency.
• Participant selection. The trial enrolled healthy adults who self-reported poor sleep; those with clinically diagnosed insomnia disorder or sleep apnea were excluded. Results apply to the worried-well end of the sleep difficulty spectrum, not to people with severe or diagnosable insomnia.
• Duration limitation. The treatment period was relatively short. Whether improvement is maintained over months or attenuates as the body reaches a new equilibrium is not answered by this design.

The trial is the best evidence yet for magnesium bisglycinate as a sleep intervention. It is not a definitive demonstration that magnesium fixes sleep. It is a well-designed RCT showing a statistically significant improvement in subjective sleep quality in adults reporting poor sleep, with an important confound from concurrent glycine delivery that the design cannot resolve.

The 2025 trial is the best evidence yet for magnesium bisglycinate and sleep. It is not proof: it is a well-designed signal in a specific population, with a glycine confound the design cannot resolve.

What came before: the 2012 trial and the 2021 meta-analysis

Abbasi et al. (2012) is the most-cited trial in this space: a double-blind, placebo-controlled trial in 46 elderly adults with insomnia [6]. Participants received 500 mg per day of elemental magnesium as oxide or placebo for 8 weeks. The magnesium group showed statistically significant improvements across multiple outcomes: sleep onset latency decreased, total sleep time increased, sleep efficiency improved, ISI score fell, and serum melatonin rose. Effect sizes were clinically meaningful. This was a trial in older adults with documented insomnia, not healthy young adults with occasional poor sleep.

Important context: magnesium oxide has relatively poor bioavailability compared to chelated forms, roughly 4% in some estimates. Despite this, the high dose was effective, possibly because a large dose of even a poorly absorbed form still delivers meaningful elemental magnesium in a population where baseline status was likely depleted. The trial population's age profile is the most important context: older adults are the group where magnesium deficiency is most prevalent, and where the sleep benefit from repletion is largest.

Mah and Pitre published a systematic review and meta-analysis in BMC Complementary Medicine and Therapies in 2021, pooling three RCTs in 151 older adults [7]. The pooled analysis found sleep onset latency reduced by a weighted mean difference of 17.36 minutes versus placebo. Sleep efficiency improved. Total sleep time increased by approximately 16 minutes.

The 2021 meta-analysis is frequently cited as robust evidence. Its limitations are equally frequently omitted: only three qualifying trials, 151 total participants, two of three trials in elderly populations specifically, and heterogeneous formulations that complicate pooled interpretation. The authors themselves note the evidence is suggestive but limited.

The consistent signal across the 2012 trial, the 2021 meta-analysis, and the 2025 RCT is that magnesium supplementation improves subjective and some objective sleep measures, particularly in older adults and in adults with documented poor sleep. The signal is genuine. The effect sizes are real but modest. The evidence base is substantially smaller and narrower than the marketing around magnesium for sleep implies.

Why form matters: glycinate vs. oxide vs. citrate

Not all magnesium supplements are equivalent. Formulation differences affect bioavailability, GI (gastrointestinal) tolerance, and in the case of bisglycinate, the concurrent delivery of glycine.

Magnesium oxide is the most common form sold due to low cost and high elemental magnesium content by weight (approximately 60% elemental). Its absorption is poor relative to chelated forms, roughly 4% in some direct comparison studies. Despite this, high-dose oxide was effective in the 2012 Abbasi trial. GI side effects, particularly osmotic laxation at higher doses, are a genuine limiting factor.

Magnesium citrate has substantially better bioavailability than oxide, is generally well tolerated, and is inexpensive. At doses of 200 to 400 mg elemental, it is an effective magnesium source. For sleep specifically, it lacks the glycine co-delivery advantage of bisglycinate.

Magnesium glycinate and bisglycinate (these terms are used interchangeably in the supplement literature; bisglycinate refers to two glycine molecules chelated to one magnesium ion) has good bioavailability, minimal GI side effects at typical doses, and delivers glycine alongside elemental magnesium. The glycine component is not trivial. Glycine independently promotes sleep onset via lowering core body temperature and modulating NMDA receptor activity in the SCN (suprachiasmatic nucleus), the brain's master circadian clock [4, 5, 8]. The 2025 Schuster trial used bisglycinate delivering 250 mg elemental magnesium plus approximately 1,500 mg glycine, close to the 3 g glycine dose used in the independent glycine-for-sleep trials.

Magnesium threonate (sold as MagTein) has preclinical and early human data suggesting preferential CNS penetration, which is mechanistically interesting for sleep but the human trial data for sleep outcomes specifically is thin and the cost is substantially higher. It does not have the same clinical trial footprint as bisglycinate for sleep outcomes.

For the specific use case of supporting sleep, bisglycinate is the best-evidenced choice. The combination of good bioavailability, low GI impact, and concurrent glycine delivery maps most closely to the intervention tested in the best available trial.

What magnesium will not do for sleep

Calibrating expectations is as important as establishing what the evidence supports.

Magnesium will not meaningfully address OSA (obstructive sleep apnea), the partial or complete collapse of the upper airway during sleep. OSA is a structural and neuromuscular problem. If unrefreshing sleep, witnessed apneas, loud snoring, or morning headaches are present, an HSAT (home sleep apnea test) is the correct next step. CPAP (continuous positive airway pressure) therapy addresses OSA. No supplement does.

Magnesium will not replicate the effect of CBT-I (cognitive behavioral therapy for insomnia) in chronic insomnia. CBT-I produces medium-to-large effect sizes on sleep-onset latency, WASO (wake-after-sleep-onset), and sleep efficiency, with durability at one-year follow-up. The magnesium effect sizes, roughly 17 minutes on sleep onset latency in the meta-analysis, are real but not comparable in magnitude or durability to structured behavioral intervention.

Magnesium will not correct circadian misalignment. If the issue is a late chronotype, social jet lag from inconsistent schedules, or evening light exposure suppressing melatonin onset, magnesium supplementation will not recalibrate the circadian clock. Low-dose melatonin used as a chronobiotic, morning light timing, and behavioral schedule consistency address those problems.

The honest framing: magnesium bisglycinate is a low-risk intervention with a real signal for improving subjective sleep quality in adults with depleted status or documented poor sleep. It is one component of a sleep optimization approach, not a standalone solution for structural or behavioral sleep problems.

A tiered framework

These tiers are frameworks for situating the evidence in a practical decision structure. Every specific decision is one to take to a clinician.