Nothing shows that all, or even most people, in the US is magnesium deficient. About 48% of people in the United States take in less magnesium from food and drink than the estimated average requirement for their age and sex, and that gap is real [1]. But eating less than a target is not a deficiency. Outright deficiency is uncommon in otherwise healthy people, because the kidneys hold on to the mineral when intake drops, and how many people are quietly running low is genuinely unknown, because there is no agreed test or cutoff to count them with [1].
Some of the attention is deserved. Magnesium is a cofactor in more than 300 enzyme systems, involved in protein synthesis, muscle and nerve function, blood glucose control and blood pressure regulation. An adult body holds about 25 grams of it, half to two thirds locked in bone and most of the rest inside soft tissue [1].
The distance between those two facts, that a lot of people could eat more of it and that you personally are deficient, is where the selling happens.
Why Is It Difficult to Test Magnesium Deficiency
The obvious way to end an argument like this is to measure it, which turns out to be the hard part.
Less than 1% of the body's magnesium is in blood serum, and the body holds that small amount inside a narrow range, so the figure a blood test returns barely moves even when the rest of the body is running low. That's why a normal result doesn't reliably track total-body or tissue magnesium, and can't on its own rule out low stores. The National Institutes of Health is direct about the problem: assessing magnesium status is difficult, serum is the most commonly used test and has little correlation with total body magnesium, and no single method is considered satisfactory [1].
Because that test is so blunt, there has been a long argument about where the line for normal should sit. A 2016 perspective paper argued that the standard serum reference range came from the distribution in a 1974 survey rather than from clinical outcomes, and that subclinical deficiency can hide inside it. Worth knowing who's arguing: several of its authors are affiliated with the Center for Magnesium Education and Research [2].
For a long time there was no newer national picture to argue with. U.S. surveys stopped measuring serum magnesium after 1974, and it isn't part of routine electrolyte testing [1]. That has changed. The national survey measured it again between 2021 and 2023, and an analysis published in 2026 used 787 children and 5,474 adults to publish the first contemporary reference values in half a century [3].
The same paper is where the headline comes from, the one saying roughly two thirds of American adults are deficient. That figure, 67.8%, did not come out of the reference range the authors had just calculated. They produced it by applying a separately proposed cutoff of 0.85 millimoles per liter, a threshold that isn't universally accepted and that sits inside the normal range published in the same paper. Someone at 0.80 is within the reference interval and below the deficiency cutoff at the same time. And the counting was done by people with a position in the argument: three of the five authors are affiliated with that same center, including its director [3].
Does Soil Depletion Explain the Magnesium Gap?
The popular explanation for why everyone is supposedly short is that the ground ran out. Decades of intensive farming stripped minerals from the soil, so today's vegetables carry less magnesium than the ones your grandparents ate.
The paper usually cited for this compared United States Department of Agriculture nutrient figures for 43 garden crops between 1950 and 1999. What it actually measured is the detail that goes missing: water, energy, protein, fat, carbohydrate, ash, calcium, phosphorus, iron, vitamin A, thiamin, riboflavin, niacin and vitamin C. Magnesium wasn't on the list, so the study most often cited for falling magnesium never measured any [4]. Of the nutrients it did cover, six declined as a group, while about 28% of the individual food and nutrient comparisons went up instead, which the authors called hard to reconcile with minerals running out of the ground. Their own explanation was the shift to higher-yielding cultivated varieties, where breeding for size can trade away nutrient density [4].
A 2017 review took the wider literature apart. Comparing food composition tables printed decades apart is unreliable to begin with, because the data sources, crop varieties, growing regions, ripeness, sampling and laboratory methods all changed underneath. And where researchers checked archived soil from land farmed intensively for decades, the mineral content of that soil hadn't declined [5].
Magnesium does turn up elsewhere in that literature. British food tables compared across 50 years showed magnesium down 35% in vegetables and 11% in fruit, and long-running wheat records showed it falling as semi-dwarf high-yield varieties came in [5]. What didn't survive was the explanation. In the wheat data those declines appeared whether the field got no fertilizer, synthetic fertilizer or manure, while the archived soil from the same sites held steady or gained magnesium [5]. The better-supported reading is dilution: varieties bred for bigger yields carry more carbohydrate without a matching rise in minerals [5]. So the story isn't invented so much as misattributed, and whether changes that size matter to anyone's health is a separate question that hasn't been settled.
The Types of Magnesium, and What the Label Tells You
Nothing in the bottle is plain magnesium. It's magnesium bound to something else, and the Supplement Facts panel counts only the elemental magnesium rather than the weight of the whole compound, so the big number on the front of the jar isn't necessarily the number that matters [1].
Oxide. Cheap, high in magnesium by weight, and among the least well absorbed. It's also one of the forms most often reported to cause diarrhea, because the salt that doesn't get absorbed pulls water into the gut [1].
Citrate. Dissolves well, and absorbs more completely than oxide [1].
Chloride, lactate, aspartate. Different partners, and the ones grouped with citrate: they dissolve well and absorb more completely than oxide or sulfate [1].
Glycinate, or bisglycinate. Magnesium bound to the amino acid glycine, and the form the sleep internet prefers. The trial usually cited for it delivered about 1.5 grams of glycine alongside the magnesium, which makes it hard to say which half of the capsule was doing the work [6].
Malate, taurate. Sold widely, and absent from that absorption comparison [1]. Confident claims about how well either one gets in are running ahead of the reference.
L-threonate. The newest arrival, and the one sold for the brain rather than the muscles. That's a marketing position rather than a finding: it isn't part of the federal absorption comparison either, and nothing cited here puts it ahead of the older forms [1].
Sulfate. Epsom salts. Poorly absorbed when you take it by mouth [1].
None of that is a ranking, and it isn't advice about which one to buy. How well a form absorbs tells you how much of the mineral gets into you, which is a different question from whether you needed more of it in the first place. Whatever's on the label, high doses from supplements can cause diarrhea, nausea and cramping. The upper limit for adults is 350 mg a day, and that ceiling counts only what comes from supplements and medications, not the magnesium already in food, which healthy kidneys clear without trouble [1]. That last part depends on the kidneys working, though. When they don't, magnesium builds up instead of being flushed out, so the same dose that's harmless for most people can be too much [1]. Before you take magnesium for any medical reason, ask your clinician first.
What Magnesium Deficiency Has to Do With Sleep
Most people meet this argument through sleep. Magnesium became the default nightstand supplement, and the deficiency story is what makes it feel safe to keep taking: you're not medicating, you're topping up something you were missing.
Whether it does anything for your nights is a separate question, and the evidence is weaker than the trend suggests: a small pooled set of trials in older adults [7], plus one newer trial with a modest and borderline result [6]. That argument is the subject of its own piece: why magnesium replaced melatonin on the nightstand.
And if the reason you're reading about minerals at all is months of bad nights, magnesium isn't what the guidelines reach for. For insomnia that has lasted that long, the recommended starting point is CBT-I (Cognitive Behavioral Therapy for Insomnia) [8]. Rest is a program built on the behavioral principles of CBT-I, delivered as a short daily conversation, by text or by voice. Start with Rest »
Disclaimer. This article is published for general information only. It is not medical advice or a diagnosis, and nothing here is a recommendation to take, avoid, or change any supplement or medication. Deciding whether to test or supplement is a conversation for a qualified health professional who knows your history. Do not start, stop, or adjust a treatment, including a prescribed sleep medication, without talking to your clinician.
Citations
National Institutes of Health, Office of Dietary Supplements. Magnesium: Fact Sheet for Health Professionals. https://ods.od.nih.gov/factsheets/Magnesium-HealthProfessional/ (accessed August 2026) — Cofactor in more than 300 enzyme systems; adult body holds ~25 g, 50%–60% in bone, less than 1% in blood serum; 48% of Americans consume less than the estimated average requirement (NHANES 2013–2016); symptomatic deficiency from diet alone uncommon in otherwise healthy people because the kidneys conserve magnesium; assessing status is difficult, serum is the most common test and has little correlation with total body magnesium or tissue concentrations, and no single method is considered satisfactory; the aspartate, citrate, lactate and chloride forms tend to have higher bioavailability than oxide and sulfate (malate, taurate, glycinate and L-threonate are not part of that comparison); Supplement Facts declares elemental magnesium; high supplemental doses cause diarrhea, nausea and abdominal cramping, and risk rises with impaired kidney function; Tolerable Upper Intake Level for supplemental magnesium is 350 mg for adults and covers supplements and medications only, not magnesium naturally present in food and beverages.
Costello, R. B., Elin, R. J., Rosanoff, A., Wallace, T. C., et al. (2016). Perspective: The case for an evidence-based reference interval for serum magnesium: The time has come. Advances in Nutrition, 7(6), 977–993. https://doi.org/10.3945/an.116.012765 — Argues subclinical magnesium deficiency can exist within the current serum reference interval (0.75–0.95 mmol/L), which derives from the NHANES I distribution in 1974 rather than clinical outcomes. Several authors are affiliated with the Center for Magnesium Education and Research; it is a perspective paper, not a trial.
Jiao, K., Costello, R., Gahche, J., Rosanoff, A., & Wallace, T. C. (2026). Serum magnesium concentrations in the United States: An updated population reference interval in children and adults. The Journal of Nutrition, 156(6), 101539. https://doi.org/10.1016/j.tjnut.2026.101539 — NHANES 2021–2023, 787 children and 5,474 adults; first contemporary U.S. population reference intervals since NHANES I (1971–1974). Reference interval in the metabolically healthy population: 0.71–0.93 mmol/L for males, 0.70–0.91 mmol/L for females. Separately, applying a previously proposed cutoff of <0.85 mmol/L, the paper estimated 67.8% of adults at risk of “chronic latent magnesium deficiency.” That cutoff was not derived from this population and is not universally accepted; it falls inside the reference interval the same paper reports. Three of the five authors are affiliated with the Center for Magnesium Education and Research, including its director.
Davis, D. R., Epp, M. D., & Riordan, H. D. (2004). Changes in USDA food composition data for 43 garden crops, 1950 to 1999. Journal of the American College of Nutrition, 23(6), 669–682. https://doi.org/10.1080/07315724.2004.10719409 — Compared 13 nutrients and water: protein, fat, carbohydrate, ash, energy, calcium, phosphorus, iron, vitamin A, thiamin, riboflavin, niacin and ascorbic acid. Magnesium was not among the nutrients analyzed, so this paper is not evidence about magnesium in either direction. Statistically reliable group-level declines for six nutrients (protein, calcium, phosphorus, iron, riboflavin, ascorbic acid) and no reliable change for seven others; about 28% of the food-and-nutrient R-values exceeded 1 (content increased), which the authors called difficult to reconcile with a broad mineral-depletion hypothesis; they attributed any real declines mainly to changes in cultivated varieties, with possible trade-offs between yield and nutrient content.
Marles, R. J. (2017). Mineral nutrient composition of vegetables, fruits and grains: The context of reports of apparent historical declines. Journal of Food Composition and Analysis, 56, 93–103. https://doi.org/10.1016/j.jfca.2016.11.012 — Critical review of the soil-depletion claim. Comparisons of food composition data published decades apart are not reliable; matching archived soil samples from intensively cultivated sites show soil mineral content has not declined; lower mineral concentrations in high-yield varieties reflect a dilution effect. The review does report magnesium declines in the underlying datasets: Mayer (1997) found magnesium down 35% in U.K. vegetables and 11% in fruit, White and Broadley (2005) found 19% in U.K. vegetables, and long-running wheat data showed magnesium declines coinciding with the introduction of semi-dwarf high-yielding cultivars, independent of whether fields received no fertilizer, inorganic fertilizer or manure, while archived soil magnesium held stable or increased. Its conclusion is that soil depletion is not the explanation, not that magnesium never declined.
Schuster, J., Cycelskij, I., Lopresti, A., & Hahn, A. (2025). Magnesium bisglycinate supplementation in healthy adults reporting poor sleep: A randomized, placebo-controlled trial. Nature and Science of Sleep, 17, 2027–2040. https://doi.org/10.2147/NSS.S524348 — n=155 adults aged 18–65 with self-reported poor sleep, diagnosed sleep disorders excluded; 250 mg elemental magnesium as bisglycinate with ~1.5 g glycine co-delivered; Insomnia Severity Index fell 3.9 points vs 2.3 on placebo (p=0.049), effect size small (d=0.2). Subjective outcomes only.
Mah, J., & Pitre, T. (2021). Oral magnesium supplementation for insomnia in older adults: A systematic review & meta-analysis. BMC Complementary Medicine and Therapies, 21, 125. https://doi.org/10.1186/s12906-021-03297-z — Three RCTs, 151 older adults with insomnia; time to fall asleep shorter than placebo; evidence rated low to very low quality, too weak for physicians to make a well-informed recommendation.
Edinger, J. D., Arnedt, J. T., Bertisch, S. M., et al. (2021). Behavioral and psychological treatments for chronic insomnia disorder in adults: An American Academy of Sleep Medicine clinical practice guideline. Journal of Clinical Sleep Medicine, 17(2), 255–262. https://doi.org/10.5664/jcsm.8986 — CBT-I as first-line treatment for chronic insomnia in adults.