Magnesium Side Effects: Diarrhea, Dose, and the 350 mg Line
Informational summary of published trials, case series and the NIH fact sheet, not medical advice or a diagnosis. Muscle weakness, facial flushing, drowsiness or a slow heartbeat while taking magnesium, especially with any kidney problem, needs a clinician the same day, not a web page.
Quick answer
The common side effects of magnesium are loose stools, diarrhea, nausea and abdominal cramping, and they track the dose and the chemical form rather than magnesium itself: the part of the dose that is not absorbed stays in the gut and pulls water in behind it, which is how magnesium oxide works as a laxative. The threshold that matters is 350 mg a day of supplemental elemental magnesium, the NIH tolerable upper intake level, and it is a supplement number: magnesium from food carries no upper limit. The one effect that is more than discomfort, hypermagnesemia, is a kidney-clearance story, which is why an eGFR under 60 changes the answer.
Per serving the shelf looks compliant. Follow the label’s own directions and it is not. We read the Supplement Facts panel of 833 magnesium labels: the median serving declares 250 mg of elemental magnesium, but the median day as directed is 399 mg, past the 350 mg line. Both medians recomputed in both halves. See the label read →
Looking for how much to take rather than what went wrong? The magnesium dosage guide owns the dose, and the forms comparison ranks all eight forms on absorption and cost. Gummies have their own read →
On this page
The common effects: a gut effect, at a dose, in a form
The NIH Office of Dietary Supplements lists diarrhea, nausea and abdominal cramping as what too much supplemental magnesium does in an adult with normal kidney function, and sets the tolerable upper intake level at 350 mg a day of supplemental elemental magnesium (NIH ODS). The endpoint that ceiling was set against, when the Institute of Medicine established it in 1997, was diarrhea: a bowel effect, not organ damage (Costello 2023, PMID 37487817). It is a tolerability ceiling with a safety margin, not the point where magnesium turns harmful.
How often does it happen? The largest pooled safety read is the Cochrane review of magnesium for muscle cramps, which collected adverse events from all 11 included trials (735 people). Magnesium went with mostly gastrointestinal events, 11% of participants in one trial (10% of controls) up to 37% in another (14%). Minor events were more common on magnesium but the pooled estimate crossed 1 (risk ratio 1.51, 95% CI 0.98 to 2.33; 4 studies, 254 participants, low certainty), and major events did not differ: 2 of 72 against 3 of 68 on placebo (Garrison 2020, PMID 32956536). Common enough to be why people stop, mild enough that trials rarely lose anyone to it.
There is a live argument that the 350 mg line is too low. A 2023 perspective found 7 of 10 qualifying studies, at intakes from 128 to 1,200 mg a day, reporting no significant difference in diarrhea between magnesium and control, and only 40 attributable gastrointestinal cases in the FDA adverse-event system for single-ingredient magnesium products (Costello 2023). Its authors are the Center for Magnesium Education and Research, which advocates for magnesium, so weigh it accordingly. The direction survives the conflict: diarrhea is a minority experience at the doses the shelf sells.
Why the form decides so much of it
Magnesium the small intestine does not absorb stays in the lumen, holds water osmotically and speeds transit. That is not a theory, it is the mechanism behind the laxative use of the same compound, which has been studied in placebo-controlled trials (Morishita 2021, PMID 32969946; Mori 2019, PMID 31587548). Ask your clinician before using magnesium that way.
How much is left behind depends on the salt. Magnesium oxide showed a fractional absorption of about 4% in a controlled comparison of four US commercial preparations, while chloride, lactate and aspartate were higher and equivalent (Firoz 2001, PMID 11794633). In 46 healthy adults taking 300 mg of elemental magnesium a day for 60 days, citrate and an amino-acid chelate raised 24-hour urinary magnesium more than oxide, and oxide was no different from placebo (Walker 2003, PMID 14596323); a single 400 mg dose in 14 saturated men repeated it (Werner 2019, PMID 32162607).
Two caveats keep that honest. A systematic review that screened 433 studies and kept 14 concluded only that inorganic forms look less bioavailable and that the absorbed percentage is dose-dependent (Pardo 2021, PMID 34111673). A detailed absorption review argues the salt matters less than commonly thought, because dose and existing magnesium status control more of the variance, and the same daily amount split across the day is absorbed better than one large dose (Schuchardt 2017, PMID 29123461). We found no randomized trial comparing diarrhea rates between glycinate and oxide at an equal elemental dose, so the absorption data make the switch reasonable rather than measured.
| Effect | What the evidence shows | Where it shows up |
|---|---|---|
| Loose stools and diarrhea | 11% to 37% of trial participants against 10% to 14% of controls (Garrison 2020); osmotic, and the laxative effect of magnesium oxide (Morishita 2021). | Rises with the elemental dose and with poorly absorbed forms. The 350 mg ceiling was set against this endpoint. |
| Hypermagnesemia | 75 of 320 hospital patients on magnesium oxide (23%), driven by low eGFR, dose and duration (Wakai 2019). Low blood pressure, nausea, flushing, muscle weakness (Ayuk 2014). | Needs impaired kidney clearance plus a steady intake. Rare with normal renal function. |
Hypermagnesemia: the serious one, and it is a kidney story
The kidney can raise fractional magnesium excretion to nearly 100% once the renal threshold is passed, which is why a healthy adult can take a large dose, feel it in the bathroom, and never move their serum magnesium. Clinically significant hypermagnesemia generally occurs only where renal insufficiency and excessive intake meet. The symptoms include hypotension, nausea, facial flushing and muscle weakness, and stopping the magnesium usually restores normal concentrations (Ayuk 2014, PMID 24402002).
The quantified picture comes from hospital prescribing, not from supplements. Among 320 inpatients prescribed magnesium oxide, 75 (23%) developed hypermagnesemia and 13 (4%) reached grade 3. Four factors were independently associated with it: an eGFR at or below 55.4 mL/min (odds ratio 3.105), blood urea nitrogen at or above 22.4 mg/dL (3.49), a dose at or above 1,650 mg a day (1.914) and 36 days or more of use (2.198), and the rate climbed as they stacked (Wakai 2019, PMID 30805197). Read it as a shape, not a personal risk: hospital patients on a prescribed laxative dose, not people taking a glycinate capsule.
The case reports follow one template. A case series described 4 patients with symptomatic hypermagnesemia from magnesium oxide taken as a laxative: all over 65, all with renal dysfunction, all unable to report symptoms clearly because of dementia or a previous stroke (Yamaguchi 2019, PMID 30136128). Case reports describe a route, not a rate: older, kidneys already impaired, a magnesium laxative taken daily, nobody checking a serum magnesium. If that describes someone you look after, the check is one blood test.
Impaired kidneys are not an absolute bar, which is worth saying because the internet treats them as one. A randomized double-blind trial gave 36 people with CKD stages 3 and 4 placebo or 15 or 30 mmol a day of elemental magnesium for 8 weeks; 34 completed and no serious adverse events were attributed to the study medication (Bressendorff 2017, PMID 29142966). The difference between that trial and the case series is monitoring. If your eGFR is below 60, magnesium is a conversation with your clinician and a serum magnesium result, not a decision made from a label.
Interactions run in two directions, and most people hear about one
Magnesium reducing the absorption of a drug you need
Magnesium is a divalent cation, and several drug classes chelate to it in the gut and stop being absorbed. The effect is not small. In 10 healthy volunteers, an aluminum-magnesium hydroxide antacid cut the exposure of a 200 mg doxycycline dose from 38.6 to 6 mg·h/L, which the authors called subtherapeutic, while doing nothing to amoxicillin or cephalexin (Deppermann 1989, PMID 2610502). A 400 mg moxifloxacin dose taken with the same kind of antacid lost roughly 60% of its exposure and peak, but giving the antibiotic 2 hours before the antacid or 4 hours after it made the interaction clinically irrelevant (Stass 2001, PMID 11352441). That is the fix: separate them.
Levothyroxine belongs on the list, and the evidence is newer than most guidance. A crossover trial in 15 healthy euthyroid adults found magnesium aspartate reduced thyroxine exposure by 12% (geometric mean ratio 0.88, 95% CI 0.81 to 0.95) and magnesium citrate by a non-significant 7%; the authors conclude levothyroxine should be taken separately from any magnesium formulation, particularly where TSH has to stay in a narrow band (Attinger 2025, PMID 41221788). Oral bisphosphonates are the fourth class: a review of osteoporosis dosing found supplements containing multivalent cations including magnesium reduced the bioavailability of every oral bisphosphonate examined (Wiesner 2021, PMID 33805435), which is why those labels insist on an empty stomach and a clear window.
Drugs that push magnesium down, not up
The second direction gets missed on side-effect pages, and it matters because low magnesium looks nothing like too much. Proton pump inhibitors are the documented case: a meta-analysis of 9 observational studies covering 109,798 patients put the pooled risk ratio for hypomagnesemia in PPI users at 1.43 (95% CI 1.08 to 1.88), rising to 1.63 among the highest-quality studies (Cheungpasitporn 2015, PMID 26108134). The FDA reached the same conclusion in 2011 and required a class warning, which still reads that low magnesium has been reported on prolonged treatment, most often after a year (FDA-approved label, 2011).
Diuretics are the other common cause, and here the usual advice is half wrong. In 9,820 participants in the Rotterdam Study, thiazide use went with lower serum magnesium and with hypomagnesemia (odds ratio 3.14 below one defined daily dose, 2.74 at or above it), mostly in long-term users. Loop diuretics did not: they went with slightly higher serum magnesium, and adding a potassium-sparing agent to a thiazide removed the association altogether (Kieboom 2018, PMID 30095199). A long-term thiazide user with cramps and fatigue may need magnesium checked rather than avoided; our deficiency signs page covers how poorly a serum test finds it.
We read 833 labels: the serving is compliant, the directions are the problem
Original research Every on-market magnesium supplement label in the NIH Dietary Supplement Label Database whose product name contains the term (39,884 on the market, 833 read, 825 declaring a parseable elemental quantity, 779 stating a daily serving direction, 276 brands), with the elemental dose taken from the Supplement Facts panel and the day computed from the label’s own serving instruction.
The finding that held is a gap between two numbers on the same bottle. The median label declares 250 mg of elemental magnesium in one serving, comfortably under the 350 mg supplemental ceiling. Follow what that same label tells you to take and the median day is 399 mg, past it. The gap recomputed in both halves of the read order, and both medians carry a stability flag. The mechanism is mundane: of the 826 labels that state a serving schedule at all (a wider set than the 779 that also declare a dose), 592 direct one serving and 234 direct more than one, and that multiplication appears nowhere on the front of the bottle.
| Item | Value ( mg) |
|---|---|
| Median serving on panel | 250 mg |
| Supplemental upper limit | 350 mg |
| Median day, as directed | 399 mg |
One thing this read can say and one it cannot. It can say the direction: more labels cross the 350 mg line when you follow the daily instruction than when you count a single serving, and that held in both halves. It cannot say by how much. The share of labels over the limit swung between the two halves of the sample, so no percentage is printed, here or anywhere else on the site.
The form mix connects back to the diarrhea question, and it also held in both halves: magnesium oxide outnumbers magnesium glycinate on the panel, 244 labels to 68. Add every chelated variant (26 bisglycinate and 17 taurinate on top of the 68 glycinate, 111 in all) and oxide still outnumbers the lot, while citrate, the other osmotic form, is the largest single group at 351. The two forms with the strongest osmotic pull are the two this shelf is built from.
| Item | Value |
|---|---|
| Citrate | 351 |
| Oxide | 244 |
| Glycinate | 68 |
| Malate | 61 |
| Bisglycinate | 26 |
| Taurinate | 17 |
One descriptive line, not a finding: the median oxide label declares 300 mg of elemental magnesium a serving, against 232.5 mg for glycinate and 250 mg for citrate. The form most likely to loosen your stools is also sold at the higher serving. These per-form medians did not go through the split-half check.
Method: 833 labels retrieved from the NIH Dietary Supplement Label Database (DSLD) v9 on 2026-08-29 for the term “magnesium”, out of 39,884 on-market products matching it; 825 declare elemental magnesium with a parseable quantity, 779 state a daily serving direction, across 276 brands. Doses are elemental, not compound weight. What it cannot tell you: which products sell most, whether the capsule matches the panel, or anything about products not filed with the NIH.
What to do, by symptom
Loose stools or cramping in the first few days, or after a dose increase. Cheapest fix first. Read the panel for the elemental amount and the directed servings per day, and multiply. Split what you take across the day, which also improves absorption (Schuchardt 2017). Drop the elemental dose. Only then change form, from an osmotic salt (oxide, citrate, sulfate) to a chelated one: the forms comparison ranks them and glycinate vs citrate is the switch most people make.
Still loose on a chelated form at a lower, split dose. Talk to your clinician; persistent diarrhea has many causes, and swapping brands can delay finding the reason.
Nausea, facial flushing, unusual drowsiness, muscle weakness, low blood pressure or a slow heart rate. Stop the magnesium and get assessed the same day, particularly with any kidney impairment or a magnesium-containing laxative in the mix. That is the symptom set of hypermagnesemia (Ayuk 2014), and serum magnesium is an ordinary blood test.
An eGFR below 60, dialysis, or a bowel obstruction. Do not start or continue magnesium without your clinician. The CKD trial above shows it can be done under monitoring (Bressendorff 2017).
On levothyroxine, doxycycline, a fluoroquinolone or an oral bisphosphonate. Keep the magnesium, move it. A gap of several hours restored moxifloxacin to normal exposure (Stass 2001), and levothyroxine is the one to be strictest with (Attinger 2025). Ask your pharmacist for the interval that fits your medicine.
On a proton pump inhibitor or a long-term thiazide, with cramps, twitching or fatigue. Your problem may be the opposite one: both are associated with low magnesium (Cheungpasitporn 2015; Kieboom 2018). The dosage guide has the NIH intake by age and sex, and older adults have their own page.
The product links below go to Amazon and we may earn a commission if you buy. It never changes which product we pick or what we say about it. How we choose.
Glycinate products we track, with the elemental amount per serving: Nature Made Magnesium Glycinate 200mg $0.39/day is 200 mg of elemental magnesium in 1 capsule, under the 350 mg ceiling; it is USP Verified, confirmed in the USP registry.
BulkSupplements Magnesium Glycinate Powder (250 g) $0.23/day is a powder (300 mg elemental per 2.2 g powder (3/4 tsp)); NSF/ANSI 173 is confirmed in the NSF registry. Thorne Magnesium Glycinate (90 capsules) $0.96/day is 120 mg elemental per capsule, listed in NSF/ANSI 173 (confirmed in the NSF registry). Every magnesium product we track is ranked in the best magnesium guide.
Frequently asked questions
What are the side effects of taking too much magnesium?
In an adult with working kidneys, loose stools, diarrhea, nausea and abdominal cramping. The unabsorbed part of the dose stays in the gut and pulls water in behind it, the same mechanism that makes magnesium oxide a laxative (Morishita 2021, PMID 32969946). In the cramp trials pooled by Cochrane, gastrointestinal effects were reported by 11% to 37% on magnesium and 10% to 14% of controls (Garrison 2020, PMID 32956536). The one effect that is more than discomfort, hypermagnesemia, needs reduced kidney clearance.
Which form of magnesium is least likely to cause diarrhea?
The chelated forms, glycinate and bisglycinate, and the absorption data point that way: magnesium oxide had a fractional absorption of about 4% in a comparison of four commercial preparations (Firoz 2001, PMID 11794633), and over 60 days raised urinary magnesium no more than placebo while citrate and an amino-acid chelate did (Walker 2003, PMID 14596323). Less absorbed means more left in the gut to pull water. But nobody has published a trial comparing diarrhea rates between glycinate and oxide at the same elemental dose, so "gentler" is an inference, not a measured result. Splitting the dose across the day also helps (Schuchardt 2017, PMID 29123461).
Does the 350 mg magnesium limit apply to the magnesium in food?
No. The 350 mg tolerable upper intake level covers supplemental elemental magnesium only. Food magnesium has no upper limit, because healthy kidneys clear the ordinary excess and a meal does not deliver a concentrated bolus. The two numbers measure different things: the 320 to 420 mg recommended intake is total magnesium from everything, while 350 mg is a ceiling on the supplemental elemental amount on the panel. A third trap sits under that: compound weight is not elemental weight, so a capsule sold as 1,000 mg of magnesium glycinate delivers a fraction of that as magnesium.
Can magnesium supplements hurt your kidneys?
The risk runs the other way: reduced kidney function is what lets magnesium accumulate. The kidney can raise fractional magnesium excretion to nearly 100%, so hypermagnesemia generally happens only where impaired renal function and high intake meet (Ayuk 2014, PMID 24402002). Among 320 hospital patients on magnesium oxide, 75 (23%) developed it; the risk factors were an eGFR at or below 55.4 mL/min, a raised blood urea nitrogen, a dose at or above 1,650 mg a day and 36 days or more of use (Wakai 2019, PMID 30805197). Under monitoring, kidney disease is still not an absolute bar (Bressendorff 2017, PMID 29142966).
Does magnesium interact with medications?
In both directions. Magnesium binds some drugs in the gut and cuts their absorption: an aluminum-magnesium antacid dropped doxycycline exposure from 38.6 to 6 mg·h/L into the subtherapeutic range (Deppermann 1989, PMID 2610502), cut moxifloxacin exposure by roughly 60% (Stass 2001, PMID 11352441), and magnesium aspartate reduced levothyroxine exposure by 12% (Attinger 2025, PMID 41221788). Those are spacing problems, not reasons to stop magnesium. In the other direction, proton pump inhibitors and thiazide diuretics are common causes of low magnesium.
Related guides
- Magnesium dosage guide: the NIH intake by age and sex
- Magnesium forms compared: eight forms on absorption and cost
- Glycinate vs citrate: the switch most people make
- Magnesium deficiency signs: the other direction
- Magnesium for seniors: where the kidney margin narrows
- Vitamin D side effects and melatonin side effects
- All magnesium guides
Sources
- NIH Office of Dietary Supplements. "Magnesium: Fact Sheet for Health Professionals." Recommended intakes, the 350 mg tolerable upper intake level for supplemental magnesium, the effects of excess intake, and the drug interactions summarized on this page. ods.od.nih.gov
- Garrison SR, et al. "Magnesium for skeletal muscle cramps." Cochrane Database Syst Rev. 2020. PMID: 32956536
- Costello R, et al. "Perspective: Call for Re-evaluation of the Tolerable Upper Intake Level for Magnesium Supplementation in Adults." Adv Nutr. 2023. Authors are affiliated with the Center for Magnesium Education and Research. PMID: 37487817
- Morishita D, et al. "Senna Versus Magnesium Oxide for the Treatment of Chronic Constipation: A Randomized, Placebo-Controlled Trial." Am J Gastroenterol. 2021. PMID: 32969946
- Mori S, et al. "A Randomized Double-blind Placebo-controlled Trial on the Effect of Magnesium Oxide in Patients With Chronic Constipation." J Neurogastroenterol Motil. 2019. PMID: 31587548
- Firoz M, Graber M. "Bioavailability of US commercial magnesium preparations." Magnes Res. 2001. PMID: 11794633
- Walker AF, et al. "Mg citrate found more bioavailable than other Mg preparations in a randomised, double-blind study." Magnes Res. 2003. PMID: 14596323
- Werner T, et al. "Assessment of bioavailability of Mg from Mg citrate and Mg oxide by measuring urinary excretion in Mg-saturated subjects." Magnes Res. 2019. PMID: 32162607
- Pardo MR, et al. "Bioavailability of magnesium food supplements: A systematic review." Nutrition. 2021. PMID: 34111673
- Schuchardt JP, Hahn A. "Intestinal Absorption and Factors Influencing Bioavailability of Magnesium: An Update." Curr Nutr Food Sci. 2017. PMID: 29123461
- Ayuk J, Gittoes NJ. "Contemporary view of the clinical relevance of magnesium homeostasis." Ann Clin Biochem. 2014. PMID: 24402002
- Wakai E, et al. "Risk factors for the development of hypermagnesemia in patients prescribed magnesium oxide: a retrospective cohort study." J Pharm Health Care Sci. 2019. PMID: 30805197
- Yamaguchi H, et al. "Severe hypermagnesemia induced by magnesium oxide ingestion: a case series." CEN Case Rep. 2019. PMID: 30136128
- Bressendorff I, et al. "Oral Magnesium Supplementation in Chronic Kidney Disease Stages 3 and 4: Efficacy, Safety, and Effect on Serum Calcification Propensity." Kidney Int Rep. 2017. PMID: 29142966
- Deppermann KM, et al. "Influence of ranitidine, pirenzepine, and aluminum magnesium hydroxide on the bioavailability of various antibiotics, including amoxicillin, cephalexin, doxycycline, and amoxicillin-clavulanic acid." Antimicrob Agents Chemother. 1989. PMID: 2610502
- Stass H, et al. "Evaluation of the influence of antacids and H2 antagonists on the absorption of moxifloxacin after oral administration of a 400mg dose to healthy volunteers." Clin Pharmacokinet. 2001. PMID: 11352441
- Attinger MC, et al. "Single Center, Open-Label, Randomized Crossover Trial on Drug-Drug Interactions of Levothyroxine/Magnesium-Citrate and Levothyroxine/Magnesium-Aspartate in Healthy Subjects: The ThyroMag Trial." Clin Transl Sci. 2025. PMID: 41221788
- Wiesner A, et al. "Optimal Dosing Regimen of Osteoporosis Drugs in Relation to Food Intake as the Key for the Enhancement of the Treatment Effectiveness." Foods. 2021. PMID: 33805435
- Cheungpasitporn W, et al. "Proton pump inhibitors linked to hypomagnesemia: a systematic review and meta-analysis of observational studies." Ren Fail. 2015. PMID: 26108134
- Kieboom BCT, et al. "Thiazide but not loop diuretics is associated with hypomagnesaemia in the general population." Pharmacoepidemiol Drug Saf. 2018. PMID: 30095199
- US Food and Drug Administration. Approved prescribing information carrying the class hypomagnesemia warning added after FDA's 2011 review of low magnesium in long-term proton pump inhibitor users. accessdata.fda.gov
- NIH Office of Dietary Supplements. Dietary Supplement Label Database. dsld.od.nih.gov (833 magnesium labels retrieved 2026-08-29; 3 of 6 findings held split-half validation and only those are printed).