Is Magnesium Glycinate Better? A Narrative Review of Human Evidence on Oral Magnesium Forms, Clinical Outcomes, Drinking-Water Magnesium and Food-First Intake

Author: Ori Scott, M.Sc. (Nutrition), Registered Dietitian (Israel), Healthy Habits Coaching, New York, USA. Contact: healthyhabitssimply.com

Key takeaways :

•             No human study shows magnesium glycinate is absorbed or tolerated better than citrate. Citrate has the best human absorption data among organic forms.

•             The one glycinate sleep trial was small, the effect borderline, and the product also contained glycine, so the benefit cannot be credited to glycinate alone.

•             Leg cramps: no proven benefit in older adults; pregnancy evidence is low quality.

•             Blood pressure: about 3/2 mmHg lower on average, mostly in people with hypertension or low serum magnesium.

•             No reliable single blood test shows true magnesium status; drinking-water findings are small and observational.

•             Food first: pumpkin seeds, legumes, spinach and nuts cover most of the gap; supplements are for documented need, within the upper limit of 350 mg/day (US) for supplemental magnesium.

Abstract

Background. Magnesium glycinate (bisglycinate) is widely promoted as better absorbed and better tolerated than other oral magnesium forms, and as helpful for sleep, stress and mood. Much of this promotion rests on laboratory data and marketing rather than on comparative human trials.

Objective. To appraise, in a narrative review, whether human evidence supports the claim that magnesium glycinate is superior to citrate or other forms, whether any form has a demonstrated benefit beyond placebo for sleep, stress, mood, cramps, migraine or blood pressure, and how dietary intake and drinking-water magnesium bear on supplement decisions.

Methods. A narrative review of about 80 sources opened, 45 of them cited (evidence window 2021–2026; older items kept only where they are the sole human data), identified through web searching, reference-list searching and papers supplied by the author. PubMed, Embase and Medscape could not be searched directly. No formal risk-of-bias or certainty grading was performed. Evidence was classified as direct, indirect or emerging.

Results. No human trial compared glycinate with citrate for sleep, stress, mood or tolerability. The only crossover absorption study including both (n = 10, single 350 mg dose, manufacturer-linked authors) showed no clear difference. Citrate has repeated human evidence of better absorption than oxide. The one placebo-controlled glycinate trial (n = 155, 250 mg elemental magnesium plus about 1.5 g glycine) found a small improvement in insomnia severity (−3.9 vs −2.3 points, p = 0.049, d = 0.2), confounded by glycine, with no difference in stress or mood. Cochrane reviews found no cramp benefit in older adults and no proven benefit in pregnancy; a pregnancy-only cramp meta-analysis was non-significant, with one small bisglycinate-chelate trial pointing toward benefit. A 2025 meta-analysis found an average blood-pressure reduction of about 3/2 mmHg, concentrated in people with hypertension or low serum magnesium. Desalinated water was associated with a very small fall in population serum magnesium; heart outcomes remain unestablished. No reliable single status test exists.

Conclusions. Evidence is insufficient to claim that magnesium glycinate is effective for sleep, stress or mood, or that it is superior to citrate. Food-first intake, targeted testing for people with risk factors, and supplementation within upper limits for documented need or for clinician-guided indications are better supported than form-specific claims.

Keywords: magnesium glycinate; magnesium bisglycinate; magnesium citrate; bioavailability; sleep; blood pressure; desalinated water; dietary intake

1. Introduction

Magnesium is a cofactor in hundreds of enzymatic reactions, and dietary adequacy is a recurring public-health concern. In the United States, an analysis of 2013–2016 survey data found that about 48% of people consumed less magnesium from food and beverages than their Estimated Average Requirement, while serum magnesium has not been measured in the national survey since 1974, so the true prevalence of deficiency is unknown [1].

Consumer interest has shifted toward specific supplement forms, and magnesium glycinate (bisglycinate) is currently promoted as the best-absorbed, gentlest option for sleep, stress and anxiety. Online claims commonly cite improved absorption, better stomach tolerance and benefits for sleep, cramps, migraine, mood and blood pressure. Whether these claims are supported by human evidence is not obvious, because magnesium forms differ in chemistry, but clinical outcomes depend on dose, baseline status, population and trial design.

A second question concerns where magnesium comes from. Food normally supplies most intake, while drinking water contributes a smaller share; desalinated and reverse-osmosis water contains little magnesium, which has raised questions about heart disease and about whether supplementation is warranted (Section 6).

Objective. This review asks: (1) Is there human evidence that magnesium glycinate is better absorbed or better tolerated than citrate or other forms? (2) Is there evidence, beyond placebo, that any magnesium form improves sleep, stress, mood, cramps, migraine or blood pressure? (3) What do deficiency, testing and drinking-water studies add? (4) What can be said about food sources, dose limits and safety?

2. Methods

2.1 Review design. This is a narrative review with structured source documentation. It is not a systematic review. No PRISMA process, duplicate independent screening, protocol registration, formal risk-of-bias assessment, certainty-of-evidence grading or meta-analysis was performed. Narrative synthesis was chosen because the question integrates heterogeneous evidence: absorption pharmacokinetics, randomized and open-label trials, meta-analyses, regulator documents, and population-level water studies that cannot be pooled.

2.2 Information sources and dates. Sources were identified through web searching, reference-list searching of retrieved papers, and full-text papers supplied by the author. PubMed, PubMed Central, Embase and Medscape could not be accessed by the tools available; therefore, this review does not claim database-level coverage of the literature. The last source was opened on 18 October 2026. The target window was 2021–2026; older publications are included only where they are the principal human data on the question (for example, absorption trials) and are labelled as earlier work.

2.3 Search concepts. Concepts were magnesium (including MeSH descriptors Magnesium and Magnesium Oxide), form names (glycinate, bisglycinate, citrate, oxide, chloride, lactate, aspartate, threonate, taurate, malate), bioavailability and absorption, sleep and insomnia, stress, anxiety, depression, muscle cramps, migraine, blood pressure, pregnancy, hypomagnesemia, magnesium status testing, drinking water, desalination, adverse effects and tolerable upper intake levels.

2.4 Evidence selection and prioritization. Primary trials and systematic reviews or meta-analyses were preferred. Regulatory and reference documents (NIH Office of Dietary Supplements, EFSA, Linus Pauling Institute) were used for requirements, limits and status-testing background. Manufacturer white papers, trade news and seller pages were treated as context only and not as evidence of effect. Each source was marked as read in full text, abstract or summary only, or secondary; findings from abstract-level sources are described as such.

2.5 Evidence classification and interpretation. Conclusions were classified as direct (the relevant form, population and outcome were studied against a suitable comparator), indirect (related forms, populations or outcomes), or emerging (small, exploratory, uncontrolled or industry-linked studies). Statistical significance was separated from clinical meaning. Designs without a placebo or active comparator were not used to infer superiority, equivalence or inferiority.

2.6 Methodological limitations of the approach. Because database searching was not possible, relevant studies may have been missed. The review leans on a limited set of full-text papers. Some sources were available only as abstracts or summary pages, which is noted where relevant.

3. Magnesium forms and absorption

3.1 What the human data show

Human absorption comparisons are few, small and usually based on short-term serum or 24-hour urine measures. Magnesium oxide is the poorest-absorbed common form, with fractional absorption of about 4% in one early study [2]. Citrate was absorbed better than oxide in three studies [3–5], and in one trial of 46 participants it produced higher serum magnesium than an amino-acid chelate (not specifically glycinate) and oxide [4]. The earliest of these gave 17 healthy volunteers a single 608 mg elemental-magnesium load of each salt and a water control in random order [3]. Over 4 hours the urinary magnesium increment was greater with citrate (0.022 vs 0.006 mg/mg creatinine, p < 0.05, driven by the second 2 hours) and the serum increment was larger (0.126 vs 0.028 mEq/L, p < 0.05), whereas the first 2 hours did not differ. In vitro, oxide was almost insoluble in water (0.8%) versus 55% for citrate. The dose is well above usual supplement doses, the outcome is a urinary and serum surrogate over a few hours, the paper’s abstract and Table 2 disagree on one urinary value (0.22 vs 0.022), and the tablets were supplied by the citrate manufacturer [3]. One crossover trial in healthy adults found oxide superior to citrate for intracellular magnesium, with unequal doses limiting interpretation [6].

The only human study identified that included both bisglycinate and citrate is a single-dose crossover in 10 healthy adults given 350 mg of magnesium [7]. At 24 hours, blood magnesium rose 5.13% with citrate and 4.88% with bisglycinate; red-cell magnesium changed +6.44% vs −2.63%, and urinary magnesium +10.55% vs +18.18%. The paper reports no clear citrate–bisglycinate difference, the study is underpowered, the description of which comparisons were significant is internally inconsistent, and two authors are employed by the maker of the sucrosomial product that performed best. This study cannot show that glycinate is better absorbed than citrate.

3.2 Where “magnesium glycinate is better” comes from.

The claim appears to derive from three sources, none of which compares glycinate with citrate in healthy people: Caco-2 cell data in a manufacturer white paper (bisglycinate 48% vs citrate 16%) [8]; an in-vitro digestion model of 12 commercial products in which a bisglycinate product reached 66.7% [9], which is laboratory bioaccessibility rather than human absorption; and one small trial in people with bowel resection, described next [10]. A 2025 systematic review that discusses forms relies on the 2003 citrate trial, that same bowel-resection study and a rat study, rather than on a direct human comparison [11]. The “gentler on the stomach” claim is traced in that review to a rat study of intestinal motility [11]; no opened source compared gastrointestinal tolerability of glycinate with citrate in humans.

The bowel-resection trial (Schuette 1994) was a double-blind randomized crossover in 12 patients with ileal resection, 10 of them with Crohn’s disease. It compared magnesium diglycinate with magnesium oxide, not citrate, at a 100 mg dose labelled with a stable isotope; the chelate was synthesized by its manufacturer under the supervision of one of the authors [10]. For the whole group, absorption was the same (23.5% with the chelate vs 22.6% with oxide). Absorption was higher with the chelate (23.5% vs 11.8%, p < .05) only in the four patients who absorbed oxide worst, a subgroup defined by the results of one arm, which makes regression to the mean a possible explanation. In the other eight patients absorption tended to be lower with the chelate, without a significant difference. Fewer stools were passed in the first 24 hours with the chelate (2.4 vs 3.7, p < .05), the peak in plasma isotope appeared about 3 hours earlier, and 5-day urinary excretion and retention did not differ. The authors concluded that for most patients the two were similarly bioavailable and that the chelate may help those with the most impaired absorption; their suggestion of intact dipeptide absorption is an inference from timing, not a direct measurement [10]. The result applies to patients with intestinal resection and does not show that glycinate is better absorbed than citrate or oxide in healthy adults.

Animal data add limited, indirect information. In a mouse study (abstract only) of four organic forms given at three doses, magnesium acetyl taurate increased brain magnesium in all treated animals, and citrate raised muscle and brain magnesium independent of dose, while splitting a high daily dose into two did not raise tissue levels enough [12]. The abstract does not report a glycinate-specific result, and mouse tissue data at 24 hours cannot be translated into human absorption or clinical benefit; these findings are hypothesis-generating only.

3.3 Evidence-based hierarchy of forms

Table 1 ranks forms by the strength of human evidence, not by proven clinical benefit. Ranking could change if a well-powered, independent glycinate–citrate crossover trial were published.

Table 1. Hierarchy of oral magnesium forms by strength of human evidence.

Rank	Form	Human absorption evidence	Human clinical evidence	Evidence class
1	Citrate	Better than oxide in 3 studies [3–5]; highest serum magnesium among citrate, oxide and an amino-acid chelate [4]	Migraine prevention at 600 mg/day (low-certainty systematic review) [13]	Direct for absorption vs oxide; indirect for others
2 (tie)	Chloride, lactate, aspartate	Reported as equal to each other and better than oxide (abstract) [2]	Chloride: open-label depression trial [14]; aspartate-HCl: no migraine benefit [13]	Indirect to emerging
2 (tie)	Glycinate / bisglycinate	Same as oxide overall, higher only in the four poorest absorbers, in one 12-patient bowel-resection crossover [10]; better in vitro; one 10-person manufacturer-linked crossover vs citrate with no clear difference [7–9,15]	One placebo-controlled sleep RCT, small effect, confounded by added glycine [16]; one small pregnancy leg-cramp RCT of a bisglycinate chelate (n = 86), not replicated in pooled data [17]	Emerging
3	L-threonate	No human absorption data found	Two small industry-linked RCTs on cognition or sleep (n = 100 and 80) [18,19]	Emerging
3	Taurate, malate	No human data found; one mouse study (abstract only) found higher brain magnesium with acetyl taurate [12]	None found	Preclinical only
4	Oxide	Lowest absorption [2]	Laxative responders about 68% vs 19% (meta-analysis, abstract level) [20]; one small restless-legs trial of magnesium plus vitamin B6, so not oxide alone [21]	Direct for laxative effect
4	Sulfate	Listed as poorly absorbed [1]	Not for supplementation; severe poisoning reported after large oral doses [22,23]	Case reports

A general review of the literature finds no consistent effect of form on outcomes [24]. The practical conclusion is that citrate has the strongest human absorption evidence, glycinate is plausible but unproven against citrate, and oxide is a rational choice for constipation rather than for raising magnesium status.

4. Sleep, stress, anxiety and mood

4.1 Sleep

The most relevant glycinate trial was a registered (DRKS00031494), double-blind, placebo-controlled, home-based trial in Germany that randomized 155 adults aged 18–65 who reported poor sleep for more than 4 weeks (Regensburg Insomnia Scale score above 12) and no chronic disease [16]. Participants took two capsules 30–60 minutes before bed for 4 weeks: magnesium bisglycinate supplying 250 mg elemental magnesium plus 1,523 mg glycine per day (product manufactured by Biogena), or cellulose placebo. In the intention-to-treat analysis (153 analysed; 134 completed) the Insomnia Severity Index fell by 3.9 points with the supplement and 2.3 with placebo, a between-group difference of 1.6 points (95% CI 0.0 to 3.3; p = 0.049; Cohen’s d = 0.20) [16]. The lower confidence limit touches zero, the effect is small, and the trial was powered for d = 0.45, more than twice the effect observed. A per-protocol analysis (n = 134) gave −5.0 vs −3.1 (p = 0.035), and placebo recipients also improved by 18% against 28% with the supplement. Outcomes were self-reported only, sleep diaries were too incomplete to analyse (under 10% completed them), and end-of-trial scores remained in the subthreshold insomnia range. Because glycine was co-administered, the benefit cannot be attributed to magnesium. Secondary outcomes did not differ from placebo: the Regensburg Insomnia Scale (−2.0 vs −1.1, p = 0.186), single-item sleep quality (p = 0.069), perceived stress (d = 0.04), mood and anxiety (PHQ-4, PANAS) [16]. An exploratory correlation within the supplement group between lower self-reported magnesium intake (one non-validated question) and larger ISI improvement (rho −0.25, p = 0.036) is hypothesis-generating only. Self-reported adverse events were 2 with the supplement and 7 with placebo, none serious. The study was funded by the authors’ university institute; one author directs a contract research organization funded by nutraceutical companies [16].

A meta-analysis of three trials in 151 older adults with insomnia, using different magnesium forms, found sleep onset about 17 minutes shorter and total sleep time about 16 minutes longer, the latter not statistically significant; certainty was low to very low and risk of bias moderate to high [25]. Overall, direct evidence for glycinate and sleep consists of one small-effect trial, and evidence for magnesium generally is inconclusive.

The Regensburg Insomnia Scale (RIS) is a 10-item self-report instrument for the thoughts, emotions and behaviours of insomnia, developed to track cognitive-behavioural treatment outcomes (Cronbach’s alpha 0.890 in 218 patients and 94 controls) [26]. In the glycinate sleep trial above it was used for screening and as a secondary outcome, where it showed no significant between-group difference; the authors noted it was designed mainly to assess CBT-I outcomes and includes an item on sleep-medication use, which makes it less applicable there [16].

4.2 Stress and anxiety

A narrative review of 18 studies concluded that evidence for magnesium in anxiety and stress was “suggestive but inconclusive”, with many poor-quality studies, some manufacturer-funded, and no effect on postpartum anxiety [24]. In a trial of 264 stressed adults with low serum magnesium, depression-anxiety-stress scores fell about 42% with magnesium lactate (300 mg elemental) and 45% with magnesium plus vitamin B6, with no significant between-arm difference (0.72 points, 95% CI −1.15 to 2.59) [27]. There was no placebo group, so the improvement cannot be separated from natural recovery or expectation; the investigators were blinded but participants were not; and the apparent benefit of B6 arose from a severe-stress subgroup analysis added after database lock. The trial was funded by the maker of the combination product, three authors being employees [27]. About 40% of participants in each arm reported at least one adverse event [27].

4.3 Depression

A meta-analysis of randomized trials reported a large but inconsistent effect in depression, with many trials at high risk of bias [28]. The best-known single trial is an open-label crossover of magnesium chloride (248 mg elemental per day for 6 weeks) in 126 adults with mild-to-moderate depression; PHQ-9 improved by a net 6.0 points (95% CI −7.9 to −4.2) and GAD-7 by 4.5 points (−6.6 to −2.4) [14]. It had no placebo and no blinding, serum magnesium was not measured, and the form was chloride. It cannot show an effect beyond expectation. The authors’ conclusion that magnesium is “effective” and “safe” exceeds what an open-label 6-week design can show.

5. Other claimed benefits

5.1 Muscle cramps

A Cochrane review of 11 trials (735 participants) found that magnesium does not reduce idiopathic cramps in older adults: the difference in cramps per week at 4 weeks was −0.18 (95% CI −0.84 to 0.49; moderate certainty), and the proportion with at least a 25% reduction showed a relative risk of 1.04 (95% CI 0.84 to 1.29; high certainty), with no dose-response across 200–520 mg of oxide, citrate, lactate or aspartate [29]. Gastrointestinal adverse effects were reported by 11–37% on magnesium versus 10–14% on control. In pregnancy, the review found five trials (408 women) that it judged to be at high risk of bias; none could be pooled, one found benefit, one was uninterpretable, and the only one not rated high risk for bias trended toward placebo, so certainty was very low and no conclusion could be drawn [29]. The trials did not select participants by magnesium status, so they cannot say whether supplementation corrects cramps caused by true deficiency; no cramp trial restricted to deficient people was found.

A 2021 PROSPERO-registered meta-analysis restricted to pregnancy pooled four placebo-controlled trials (332 women; 300–360 mg/day of magnesium lactate/citrate, citrate or a bisglycinate chelate for 2–4 weeks) [17]. Cramps per week did not differ significantly (weighted mean difference −0.47, 95% CI −1.14 to 0.20; three trials, I² = 0%), and neither did recovery from cramps (OR 0.47, 95% CI 0.14–1.52; three trials, I² = 75%); the authors concluded that oral magnesium is not effective. The recovery result was inconsistent across trials: a 1995 Swedish trial (OR 0.13, 95% CI 0.03–0.63) and the only bisglycinate-chelate trial (Supakatisant 2015; 100 mg three times daily for 4 weeks; 86 women; OR 0.41, 95% CI 0.17–0.99) pointed toward benefit, whereas a 2020 Brazilian trial did not (OR 1.24, 95% CI 0.59–2.63). The paper does not state the direction of the odds ratio; I infer it from the authors’ description that the Swedish and Thai trials reported improvement. Three of the four trials were not preregistered, the Brazilian study was a secondary analysis of a published trial, one author of the Norwegian trial developed the tablet and was paid by its manufacturer, and serum magnesium did not rise with supplementation in the two trials that measured it [17]. The Cochrane review rated this bisglycinate trial, which had independent funding and no declared conflicts, at high risk of bias for two reasons: participants with missing data were counted as treatment failures in a way that favoured the supplement (4 missing in the placebo group, 2 in the supplement group), and the reported results were incongruous, because the median number of cramps per week rose in both groups (by 1 with the supplement and 3 with placebo) while a large percentage reduction in cramps was reported [29]. In summary, the single glycinate-type cramp signal comes from a small, non-preregistered trial of 86 women in pregnancy that was not replicated in the pooled data, and it does not support a general cramp claim for glycinate.

5.2 Pregnancy

A Cochrane review of 10 randomized trials in 9,090 women (search ended March 2013) found no significant effect of magnesium supplementation on perinatal death (RR 1.10, 95% CI 0.72–1.67), small-for-gestational-age babies (RR 0.76, 0.54–1.07) or pre-eclampsia (RR 0.87, 0.58–1.32). Possible signals included fewer low Apgar scores and fewer antenatal hospital stays, and a possible increase in neonatal death before discharge (RR 2.21, 1.02–4.75) that was driven by one trial and disappeared after excluding deaths from severe anomalies. Only 2 of 10 trials were high quality, and no trial used glycinate [30]. This is the oldest evidence in the review and has not been updated here.

5.3 Migraine

A 2026 systematic review of 6 trials (about 502 participants) found that citrate at 600 mg/day reduced attacks in trials of that form (low certainty), whereas magnesium aspartate-HCl did not, and oxide did not differ from valproate; no glycinate migraine trial was identified [13]. American neurology and headache societies describe magnesium as “probably effective” for migraine prevention; the trial doses exceed the 350 mg supplemental upper limit for adults and therefore call for clinician supervision [1].

5.4 Blood pressure

A 2025 meta-analysis of 38 randomized, placebo-controlled trials (2,709 adults; median 365 mg/day elemental magnesium; median 12 weeks) found a mean reduction of 2.81 mmHg systolic (95% CI −4.32 to −1.29) and 2.05 mmHg diastolic (−3.23 to −0.88), with high heterogeneity (I² 78% and 88%) [31]. The effect was not significant in people with normal blood pressure or normal serum magnesium, and was larger in people with low serum magnesium (−5.97/−4.75 mmHg) and in hypertensive participants taking antihypertensives (−7.68/−2.96 mmHg), many of whom used diuretics that waste magnesium. There was no dose-response relationship. Chloride (7 trials) and aspartate-HCl (8 trials) showed effects, but all chloride trials enrolled people with low serum magnesium, so form and deficiency are confounded; one diglycinate trial and four citrate trials showed no significant change [31]. The U.S. Food and Drug Administration has allowed a qualified health claim for magnesium and hypertension while describing the evidence as inconsistent and inconclusive [32]. The meta-analysis’s senior author reports speaker honoraria and advisory roles with pharmaceutical companies; the work was publicly funded [31].

5.5 Other outcomes

Oxide is the best-supported form for constipation, with about 68% responders versus 19% in a meta-analysis (abstract level) [20]. Diabetes meta-analyses were opened but are not used to support claims here. In the oxidative-stress and inflammation meta-analysis (28 studies, animal and human), oxidative-stress markers were null and C-reactive protein showed a small effect (SMD 0.21, 95% CI 0.09–0.32) with no form comparison [11].

6. Deficiency, testing and drinking water

6.1 How deficiency is identified

Serum magnesium is the standard test but reflects well under 1% of body magnesium [31,33]. The NIH describes 0.75–0.95 mmol/L as normal and below 0.75 mmol/L as hypomagnesemia, and notes that no single indicator of status is considered satisfactory [1]. The Linus Pauling Institute states that “there is no reliable indicator of magnesium status”, describes the intravenous magnesium-tolerance test as the reference method but impractical, and advises combining diet, serum and urine data [34]. A 2025 proposal would raise the laboratory lower limit to 0.85 mmol/L [35]; it is one group’s recommendation, not a standard. No source identified here establishes how well serum tracks intracellular magnesium, so a strong correlation cannot be assumed. Practically, serum levels, symptoms and risk factors (diuretics, long-term acid-reducing drugs, alcohol, diabetes, gut disease and low intake) should be considered together.

A 2024 narrative review by laboratory-medicine specialists adds detail [36]. More than 99% of body magnesium is intracellular, and total serum magnesium, though the usual test, has been reported to correlate with bone and interstitial-fluid magnesium but not with total body magnesium; red-cell magnesium is theoretically attractive but serum or plasma remain the samples of choice in routine care, and the physiologically active free (ionized) fraction needs ion-selective electrodes that most laboratories do not use. Reference intervals differ: 0.7–1.1 mmol/L in one source and 0.75–0.95 mmol/L in a US sample of 15,820 adults. The review defines hypomagnesemia as serum magnesium below 0.7 mmol/L (mild 0.57–0.70, moderate 0.40–0.56, severe below 0.40) and notes that it typically causes clinically significant signs only below 0.5 mmol/L. When serum is low, urine helps find the cause: a 24-hour urinary excretion above 24 mg or a fractional excretion above 3% suggests kidney wasting, whereas lower values suggest low intake or gut losses; a random urine magnesium-to-creatinine ratio of 0.04–0.12 mg/mg is given as the adult reference range. Reported prevalence of hypomagnesemia is 2.7% in the general population, about 10% in hospitalized patients, 14.7% in chronic kidney disease, 30–80% in alcohol use disorder and 10–60% in diabetes [36]. For treatment, the review reserves oral replacement for asymptomatic patients and intravenous magnesium for severe or symptomatic cases, states that organic magnesium salts are generally best absorbed followed by chloride and then oxide (without naming glycinate), and cites 800–1,600 mg/day for moderate-to-severe hypomagnesemia under clinical care, a range far above the supplemental upper limit that applies to self-treatment [36]. This is a narrative review whose absorption statement rests on an older general source rather than head-to-head trials.

6.2 Low-mineral drinking water

The World Health Organization reviewed ecological, case-control and cohort studies: most ecological studies linked soft water with more cardiovascular death, case-control and cohort studies were less consistent, protection was generally seen at about 10 mg/L magnesium and above, and causation was not proven [37]. In Israel, magnesium is not added to desalinated water; median tap-water magnesium was 9 mg/L in districts supplied mostly by desalinated seawater versus 34 mg/L in groundwater districts [38]. An Israeli analysis reported odds of ischaemic heart disease of 0.96 at baseline and 1.06 at the end of 2004–2013 as desalinated water expanded (abstract only; internally inconsistent) [39].

A health-fund study of 66,764 people compared serum magnesium before (2010–2013) and after (2014–2016) 15 communities switched to desalinated water, with 24 communities that did not [40]. Mean serum magnesium fell from 2.065 to 2.057 mg/dl (p < 0.0001), a 0.4% change that is statistically significant because of sample size but small. The share at or below 1.6 mg/dl rose from 1.62% to 1.91% in 2014–2016 (+0.29 percentage points; +18.5% relative) and to 2.01% in 2015–2016 (+0.39 points; +24% relative); no such change occurred in the comparison communities (2.18% to 1.98%). The study was retrospective, used each person’s first test only, compared communities rather than individuals, and lacked data on diet, drugs and supplements [40]. The authors cite a separate study reporting that 30-day all-cause mortality after myocardial infarction was more than double in desalinated-water regions [40]; I did not open that paper and describe it only as reported.

A comparison of two Israeli cities found a similar small serum change and no higher dementia prevalence in the desalinated-water city (2.37% vs 2.55%, not significant); the authors called the serum changes statistically but not clinically significant for dementia, and estimated that desalinated water removes about 15 mg of magnesium per day, 3–5% of the RDA [41]. The paper has internal inconsistencies in city sizes, dementia counts, study periods and units, and its dementia figures were not age-adjusted [41]. Simple arithmetic (author’s calculation) gives about 18 mg/day from 2 L of water at 9 mg/L versus about 68 mg at 34 mg/L, a gap of roughly 50 mg/day relative to recommended intakes of 310–420 mg.

A 2016 meta-analysis of 10 studies (77,821 coronary heart disease deaths) found an inverse association between drinking-water magnesium and coronary heart disease mortality (RR 0.89, 95% CI 0.79–0.99; I² 70.6%) [42]. The result was driven by case-control studies (7 studies, RR 0.87, 0.76–0.98); the three cohort studies were not significant (RR 0.88, 0.57–1.34), a subgroup of three ischaemic-heart-disease studies showed no association (RR 1.01, 0.96–1.05), and nine of the ten studies came from Europe [42]. This is observational, heterogeneous data that cannot show that magnesium in water, rather than hardness, calcium or other differences between areas, is protective, and it says nothing about supplements.

A 2021 narrative review of magnesium in ageing proposes that chronic “latent” magnesium deficiency (normal serum, low tissue magnesium) is common in older adults and may contribute to chronic low-grade inflammation [43]. I treat this as a hypothesis: it rests on the same limited status tests discussed above, and the authors themselves state that the role of supplements is unclear and that long-term blinded studies are lacking [43].

Taken together, low-mineral water may modestly lower magnesium intake and slightly shift population serum levels; whether this raises heart-disease risk, and whether supplementation would prevent it, is not established. Hard-water associations may also reflect calcium or other co-varying minerals [34].

7. Dose, upper limits and safety

The U.S. Recommended Dietary Allowance is 310–320 mg/day for adult women and 400–420 mg/day for adult men; pregnancy values are 350–360 mg/day [1]. The NIH tolerable upper intake level for supplemental magnesium (supplements and medicines, not food) is 350 mg/day for adults [1]; the European Food Safety Authority sets 250 mg/day [44]. Trial doses were 250 mg (glycinate sleep trial) [16], 248 mg (depression) [14], 300 mg (stress) [27], 200–520 mg (cramps) [29], median 365 mg (blood pressure) [31] and 600 mg (migraine) [13]; the last exceeds the upper limit.

Common adverse effects are diarrhoea, nausea and abdominal cramping; people with kidney disease are at higher risk of accumulation. Reported interactions include bisphosphonates, tetracycline and quinolone antibiotics (to be separated by hours), diuretics and long-term acid-reducing drugs [1]. Hypermagnesemia and poisoning are documented mainly as case reports and reviews involving kidney disease, laxatives or bowel-prep preparations [22,23,45], which indicate a signal rather than an incidence.

8. Food-first intake

Foods differ widely in magnesium content (Table 2) [1]. Reported average intakes in US adults who do not use supplements are about 234 mg/day (women) and 268 mg/day (men) [1], so the typical gap to the RDA is roughly 50–150 mg. About 30–40% of dietary magnesium is absorbed [1]. A combination such as 1 oz pumpkin seeds (156 mg), half a cup of cooked spinach (78 mg), half a cup of black beans (60 mg) and 1 oz almonds (80 mg) provides about 374 mg. Nuts alone are inefficient: about 4 oz of almonds would be needed to reach roughly 320 mg (author’s arithmetic). Refined grains lose magnesium (brown rice 42 mg versus white 10 mg per half cup) [1].

Table 2. Magnesium content of selected foods (NIH Office of Dietary Supplements) [1].

Food	Serving	Magnesium (mg)
Pumpkin seeds, roasted	1 oz (28 g)	156
Chia seeds	1 oz	111
Almonds, dry roasted	1 oz	80
Spinach, boiled	½ cup	78
Cashews, dry roasted	1 oz	74
Peanuts, oil roasted	¼ cup	63
Soymilk	1 cup	61
Black beans, cooked	½ cup	60
Edamame, cooked	½ cup	50
Peanut butter	2 tbsp	49
Brown rice, cooked	½ cup	42
Plain low-fat yogurt	8 oz	42
Banana	1 medium	32
Salmon, cooked	3 oz	26
White rice, cooked	½ cup	10

9. Comparing products fairly

No source opened in this review shows that any named commercial magnesium product is effective. The only glycinate-specific placebo-controlled trial studied a bisglycinate-plus-glycine regimen and found a small effect on one measure [16]. Product comparison should therefore be based on label facts: the form named, elemental magnesium per serving (not compound weight), total daily dose relative to the upper limit, added ingredients such as glycine, melatonin, herbs or high-dose vitamin B6 that blur attribution, third-party testing disclosed on the testing body’s own site, and price per 100 mg elemental magnesium. Laboratory tests of products and bioaccessibility in vitro [9] indicate content and dissolution, not clinical effect.

10. Strengths and limitations

Strengths include the focus on human evidence, explicit separation of statistical and clinical significance, labelling of industry links and design limits, and the use of several full-text primary papers. Limitations include the absence of direct database searching, the lack of formal risk-of-bias or certainty grading, reliance on abstract or secondary summaries for some sources, a small number of absorption trials with short follow-up, and heterogeneity of forms, doses and populations. Some papers contained internal inconsistencies (notably the Brilli crossover and the Israeli dementia study) that limit numerical certainty. Water studies are observational and cannot establish causation. The author is a dietitian with a nutrition practice and a supplement store (see Declarations).

11. Implications and research gaps

For practice, the evidence supports a food-first approach, targeted testing in people with risk factors, and supplementation for documented low magnesium or clinician-guided indications (migraine, hypertension with low magnesium), staying within upper limits and avoiding use in kidney disease unless supervised. Choosing glycinate rather than citrate is not supported by human comparative data; tolerability and cost may reasonably guide the choice.

Research needs include adequately powered, independent crossover trials of glycinate versus citrate (absorption and gastrointestinal tolerability); placebo-controlled sleep and anxiety trials using glycinate without added glycine and with validated outcomes such as the RIS; cramp trials restricted to people with documented low magnesium; validation of status indicators; and trials testing whether magnesium supplementation changes cardiovascular outcomes in people drinking low-mineral water.

12. Conclusions

There is no human evidence that magnesium glycinate is better absorbed or better tolerated than citrate, and the evidence is insufficient to claim that it improves sleep, stress or mood beyond placebo. Citrate has the strongest human absorption data among organic forms; oxide is best supported for constipation. Cramp and pregnancy supplementation lack proven benefit (one small bisglycinate-chelate trial in pregnancy cramps favored magnesium but was not replicated), blood-pressure effects are small and concentrated in people with hypertension or low magnesium, and drinking-water studies show a small serum shift without established clinical consequences. Dietary intake and individual risk factors should guide decisions, and claims about a specific form should be limited to what trials have shown.

Declarations

Funding: to be completed by the author. Conflicts of interest and affiliate disclosure: The author runs a nutrition coaching practice and a supplement store, and may recommend magnesium supplements as an Amazon affiliate, earning a commission on qualifying purchases. This commercial interest could bias the choice and interpretation of evidence; to limit that, the review reports effects only where trials show them and makes no product recommendation. Other relationships: to be completed by the author. Ethics approval and consent: not applicable (review of published literature). Data availability: no new data were generated. AI-assisted technology disclosure: An AI assistant (Claude, Anthropic) was used to help locate, read and organize sources and to draft text. The author is responsible for source selection, interpretation, verification of facts and references, and the final content. AI is not an author. Medical disclaimer: This review is educational and not individual medical advice. People who are pregnant, have kidney disease or take interacting medicines should consult a clinician before using magnesium supplements.

References

1.          National Institutes of Health, Office of Dietary Supplements. Magnesium: Fact Sheet for Health Professionals. Updated January 2026. https://ods.od.nih.gov/factsheets/Magnesium-HealthProfessional/

2.          Firoz M, Graber M. Bioavailability of US commercial magnesium preparations. Magnes Res. 2001;14:257–62. (Earlier work; abstract only.)

3.          Lindberg JS, Zobitz MM, Poindexter JR, Pak CYC. Magnesium bioavailability from magnesium citrate and magnesium oxide. J Am Coll Nutr. 1990;9(1):48–55. doi:10.1080/07315724.1990.10720349. (Earlier work.)

4.          Walker AF, Marakis G, Christie S, Byng M. Mg citrate found more bioavailable than other Mg preparations in a randomised, double-blind study. Magnes Res. 2003;16(3):183–91. PMID 14596323. (Earlier work; abstract only. DOI not retrieved.)

5.          Kappeler D, et al. Higher bioavailability of magnesium citrate as compared to magnesium oxide shown by evaluation of urinary excretion and serum levels after single-dose administration in a randomized cross-over study. BMC Nutr. 2017;3:7.

6.          Shechter M, et al. Comparison of magnesium status using X-ray dispersion analysis following magnesium oxide and magnesium citrate treatment of healthy subjects. Magnes Res. 2012;25:28–39. (Earlier work.)

7.          Brilli E, Khadge S, Fabiano A, Zambito Y, Williams T, Tarantino G. Magnesium bioavailability after administration of sucrosomial® magnesium: results of an ex-vivo study and a comparative, double-blinded, cross-over study in healthy subjects. Eur Rev Med Pharmacol Sci. 2018;22(6):1843–51. doi:10.26355/eurrev_201803_14605.

8.          Bortz J. Understanding magnesium and bioavailability studies. Albion/Balchem white paper; 2024. (Manufacturer document; context only.)

9.          Bawiec P, Jaworowska A, Sawicki J, Czop M, Szalak R, Koch W. In vitro evaluation of bioavailability of Mg from daily food rations, dietary supplements and medicinal products from the Polish market. Nutrients. 2025;17(5):748. doi:10.3390/nu17050748.

10.      Schuette SA, Lashner BA, Janghorbani M. Bioavailability of magnesium diglycinate vs magnesium oxide in patients with ileal resection. JPEN J Parenter Enteral Nutr. 1994;18:430–5. (Earlier work; read in full.)

11.      Cepeda V, Ródenas-Munar M, García S, Bouzas C, Tur JA. Unlocking the power of magnesium: a systematic review and meta-analysis regarding its role in oxidative stress and inflammation. Antioxidants. 2025;14(6):740. doi:10.3390/antiox14060740.

12.      Ates M, Kizildag S, Yuksel O, Hosgorler F, Yuce Z, Guvendi G, Kandis S, Karakilic A, Koc B, Uysal N. Dose-dependent absorption profile of different magnesium compounds. Biol Trace Elem Res. 2019;192:244–51. doi:10.1007/s12011-019-01663-0. (Mouse study; abstract only.)

13.      Jagunmolu H, Oyetola E, Lawal Y, Oyelude S, Ogunsola O, Oladeji A, Ogundipe S, Ibrahim M, Ajibade D, Ariyo O, Ajagbe K. Efficacy of magnesium supplementation in the prophylaxis of migraine: a systematic review of clinical trials comparing oral magnesium to placebo or standard care. Egypt J Neurol Psychiatr Neurosurg. 2026;62:81. doi:10.1186/s41983-026-01150-z. (Abstract level.)

14.      Tarleton EK, Littenberg B, MacLean CD, Kennedy AG, Daley C. Role of magnesium supplementation in the treatment of depression: a randomized clinical trial. PLoS One. 2017;12:e0180067. (Earlier work; open-label.)

15.      Siebrecht S. Magnesium bisglycinate as safe form for mineral supplementation in human nutrition. OM & Ernährung. 2013;144. (Single-author overview; used only to see which human studies exist.)

16.      Schuster J, Cycelskij I, Lopresti A, Hahn A. Magnesium bisglycinate supplementation in healthy adults reporting poor sleep: a randomized, placebo-controlled trial. Nat Sci Sleep. 2025;17:2027–40. doi:10.2147/NSS.S524348.

17.      Liu J, Song G, Zhao G, Meng T. Effect of oral magnesium supplementation for relieving leg cramps during pregnancy: a meta-analysis of randomized controlled trials. Taiwan J Obstet Gynecol. 2021;60:609–14. doi:10.1016/j.tjog.2021.05.006.

18.      Lopresti AL, Smith SJ. The effects of magnesium L-threonate (Magtein®) on cognitive performance and sleep quality in adults: a randomised, double-blind, placebo-controlled trial. Front Nutr. 2025;12:1729164. doi:10.3389/fnut.2025.1729164. (A correction to this article has been issued; see the journal page.)

19.      Hausenblas HA, Lynch T, Hooper S, Shrestha A, Rosendale D, Gu J. Magnesium-L-threonate improves sleep quality and daytime functioning in adults with self-reported sleep problems: a randomized controlled trial. Sleep Med X. 2024;8:100121. doi:10.1016/j.sleepx.2024.100121.

20.      Van Der Schoot A, Katsirma Z, Whelan K, Dimidi E. Systematic review and meta-analysis: foods, drinks and diets and their effect on chronic constipation in adults. Aliment Pharmacol Ther. 2023. doi:10.1111/apt.17782. (The 68% vs 19% figure was taken from a conference abstract, Proc Nutr Soc 2023;82(OCE1), and has not been checked against the full paper.)

21.      Jadidi A, et al. Therapeutic effects of magnesium and vitamin B6 in alleviating the symptoms of restless legs syndrome: a randomized controlled clinical trial. BMC Complement Med Ther. 2023;23:1. doi:10.1186/s12906-022-03814-8.

22.      Torigoe K, et al. Laxative-induced hypermagnesemia in a peritoneal dialysis patient. Cureus. 2025 (case report).

23.      Si G-F, Ge Y-X, Lv X-P, Li Y-Q, Chen X-M, Yuan X-M. Case report: near-fatal hypermagnesemia resulting from the use of Epsom salts in a patient with normal renal function. Front Med. 2024;11:1416956. doi:10.3389/fmed.2024.1416956.

24.      Boyle NB, Lawton C, Dye L. The effects of magnesium supplementation on subjective anxiety and stress: a systematic review. Nutrients. 2017;9:429. (Earlier work.)

25.      Mah J, Pitre T. Oral magnesium supplementation for insomnia in older adults: a systematic review & meta-analysis. BMC Complement Med Ther. 2021;21:125. doi:10.1186/s12906-021-03297-z. (Record only.)

26.      Crönlein T, Langguth B, Popp R, Lukesch H, Pieh C, Hajak G, Geisler P. Regensburg Insomnia Scale (RIS): a new short rating scale for the assessment of psychological symptoms and sleep in insomnia patients. Health Qual Life Outcomes. 2013;11:65. doi:10.1186/1477-7525-11-65.

27.      Pouteau E, Kabir-Ahmadi M, Noah L, et al. Superiority of magnesium and vitamin B6 over magnesium alone on severe stress in healthy adults with low magnesemia: a randomized, single-blind clinical trial. PLoS One. 2018;13:e0208454. doi:10.1371/journal.pone.0208454. (Earlier work.)

28.      Moabedi M, Aliakbari M, Erfanian S, Milajerdi A. Magnesium supplementation beneficially affects depression in adults with depressive disorder: a systematic review and meta-analysis of randomized clinical trials. Front Psychiatry. 2023;14:1333261. doi:10.3389/fpsyt.2023.1333261.

29.      Garrison SR, Korownyk CS, Kolber MR, Allan GM, Musini VM, Sekhon RK, Dugré N. Magnesium for skeletal muscle cramps. Cochrane Database Syst Rev. 2020;(9):CD009402. doi:10.1002/14651858.CD009402.pub3.

30.      Makrides M, Crosby DD, Shepherd E, Crowther CA. Magnesium supplementation in pregnancy. Cochrane Database Syst Rev. 2014;(4):CD000937. doi:10.1002/14651858.CD000937.pub2.

31.      Argeros Z, Xu X, Bhandari B, Harris K, Touyz RM, Schutte AE. Magnesium supplementation and blood pressure: a systematic review and meta-analysis of randomized controlled trials. Hypertension. 2025;82:1844–56. doi:10.1161/HYPERTENSIONAHA.125.25129.

32.      U.S. Food and Drug Administration. Qualified health claim: magnesium and high blood pressure. 10 January 2022.

33.      Workinger JL, Doyle RP, Bortz J. Challenges in the diagnosis of magnesium status. Nutrients. 2018;10:1202. doi:10.3390/nu10091202. (Read through a summary page; secondary.)

34.      Linus Pauling Institute, Oregon State University. Magnesium. Micronutrient Information Center; reviewed October 2025.

35.      Weisser H, Kisters K, Vormann J, von Ehrlich B, Micke O. Serum magnesium concentration: time for a standardised and evidence-based reference range. Magnes Res. 2025;38(2):41–4. doi:10.1684/mrh.2025.0548. (Summary page only.)

36.      Salinas M, López-Garrigós M, Flores E, Leiva-Salinas C. Improving diagnosis and treatment of hypomagnesemia. Clin Chem Lab Med. 2024;62(2):234–48. doi:10.1515/cclm-2023-0537. (Narrative review; read in full.)

37.      World Health Organization. Calcium and magnesium in drinking-water: public health significance. Geneva; 2009.

38.      Sagie N, Levin R, Groisman L, Carmi-Sasson K, De Falco N, Rachmilevitch S, Novack L. Desalination and calcium, fluoride, and magnesium composition: assessing water quality variations across water sources in Israel. Water. 2026;18(17):2073. doi:10.3390/w18172073.

39.      Shlezinger M, et al. Association between exposure to desalinated sea water and ischemic heart disease, diabetes mellitus and colorectal cancer; a population-based study in Israel. Environ Res. 2018. PMID 29982150. (Abstract only; volume, pages and DOI not retrieved.)

40.      Koren G, Shlezinger M, Katz R, Shalev V, Amitai Y. Seawater desalination and serum magnesium concentrations in Israel. J Water Health. 2017;15(2):296–9. doi:10.2166/wh.2016.164.

41.      Ben Zaken S, Simantov O, Abenstein A, Radomysky Z, Koren G. Water desalination, serum magnesium and dementia: a population-based study. J Water Health. 2020;18(5):722–7. doi:10.2166/wh.2020.132.

42.      Jiang L, He P, Chen J, Liu Y, Liu D, Qin G, Tan N. Magnesium levels in drinking water and coronary heart disease mortality risk: a meta-analysis. Nutrients. 2016;8(1):5. doi:10.3390/nu8010005.

43.      Barbagallo M, Veronese N, Dominguez LJ. Magnesium in aging, health and diseases. Nutrients. 2021;13(2):463. doi:10.3390/nu13020463. (Narrative review; hypothesis-generating.)

44.      European Food Safety Authority. Overview on tolerable upper intake levels as derived by the Scientific Committee on Food and the EFSA Panel on Nutrition, Novel Foods and Food Allergens. Version 11, August 2025.

45.      Qazi M, et al. Causes of hypermagnesaemia: a literature review. EMJ Nephrol. 2021.