The Compound Is Real. The Bottle Is a Different Question.

What you'll learn

When a supplement is the right tool rather than a substitute for food

What the law requires of a supplement before it reaches a shelf, and what it does not

Why "rich in magnesium" depends entirely on a number nobody prints

Why better absorption is not automatically better

What quality certification proves, and what it cannot

How SmartBar applies the deficiency framework

SmartBar assesses products using three things we can verify: ingredient composition, level of processing, and nutritional data. We use two organizing terms throughout our education. Deficiency means something the body needs and isn't getting. Toxicity means a burden the body must process and clear, not a poison.

Supplements sit across both. Sometimes supplementation is exactly the appropriate tool: for a documented deficiency, a predictable dietary gap, or a preventive use with strong evidence. Other times it creates a burden through concentration, contamination or enhanced absorption. Which one you are holding depends less on whether the ingredient is natural than on whether a need was established before the product was.

Start here: what a supplement actually is

A supplement takes a dietary ingredient and packages it outside an ordinary food: isolated, concentrated, extracted, synthesised, or simply dried and powdered. That can change the dose, the food matrix around it, the contaminants that travel with it, or how much of the compound reaches your bloodstream.

Sometimes that is exactly the point. The question is never whether something is natural. It is what is in the bottle, how much of it reaches you, whether you need it, and what evidence exists for the outcome being promised.

When the bottle is actually the point

Food first is a useful default. It is not a law of biology, and an article that pretended otherwise would be arguing rather than explaining.

The clearest counterexample is folic acid. The US Preventive Services Task Force recommends that all persons planning to or who could become pregnant take a daily supplement containing folic acid, and grades it A, its strongest grade, on high certainty of substantial net benefit in preventing neural tube defects.[1]

The recommendation does not wait for a measured deficiency, and it does not apply only to people trying to conceive. CDC notes that folic acid, the form used in supplements and fortified foods, is the only type of folate shown in studies to prevent neural tube defects,[2] which occur in the first weeks of pregnancy, often before anyone knows they are pregnant.

Common MTHFR variants do not change that. CDC says people carrying them can process folic acid, that intake matters more than genotype, and that the variants are not a reason to avoid it.[2] Methylated folate is marketed hard on exactly that claim.

Vitamin B12 is the second clean case. Plants do not naturally supply vitamin B12, so a vegan diet needs a reliable source from fortified foods, a supplement, or both.[44] No amount of choosing different unfortified plants solves that particular gap.

Those cases are different in kind from taking an extract because a label promises better sleep, younger skin or detox. Supplementation makes its strongest case when the need is defined first and the bottle second. That sequence is the organising idea of this page.

What the label and the law actually promise

SmartBar scores Nutrition Facts panels. A supplement carries a Supplement Facts panel, which is a different regulatory object, so supplements are not in our database, not in the Whole Foods Library, and cannot be logged in Track.

But the more consequential difference is not the panel. By statute, FDA does not approve dietary supplements for safety and effectiveness, and many can be marketed without even notifying the agency.[3] Companies are responsible for ensuring their own products are safe and properly labelled. FDA inspects facilities, monitors adverse event reports, and can act against products after they are on sale.

That is not the same as unregulated. Federal manufacturing rules require specifications for identity, purity, strength, composition and contamination, with verification under FDA's sampling and testing rules.[4] What they do not require is FDA testing and approving every bottle before sale.

Certain new dietary ingredients also require a safety notification at least 75 days before marketing. FDA estimates that with 80,000 or more products on the market, more than 4,600 such notifications should have been submitted, against roughly 1,200 received.[5]

So supplements are neither pre-approved nor unregulated. The rules exist; premarket approval generally does not, and FDA's own estimate suggests one notification requirement has often been missed.

The number nobody prints

One habit before any specific compound.

Nutrition tables are published per 100 grams. Almost nothing is eaten in 100 gram portions.

Pumpkin seeds contain 550 mg of magnesium per 100 g, 131% of the Daily Value.[6] That looks like a supplement. But 100 g is 574 kcal. A one-ounce (28 g) serving gives about 154 mg, roughly 37% of the Daily Value: unusually magnesium-dense for a common food, and not the same claim.

Chia is the same, at 34.4 g of fibre and 335 mg of magnesium per 100 g, or seven tablespoons and 486 kcal.[7] Dates make the point sharpest: 696 mg of potassium and 54 mg of magnesium per 100 g, which is four dates and 277 kcal.[8] Per calorie, pumpkin seeds deliver roughly five times as much magnesium as Medjool dates.

None of that makes any of these worse foods. It means "rich in" is a claim about a denominator, and the denominator is usually chosen by whoever is selling. Track resolves it because it counts what you actually logged.

Fibre, where whole food competes hardest

Fibre is a standalone macronutrient in Track, most people fall short, and the supplement market is large.

Psyllium has the best regulatory standing here. FDA authorised a health claim linking diets low in saturated fat and cholesterol that include soluble fibre from psyllium seed husk with reduced coronary heart disease risk, identifying 7 g or more per day of soluble fibre from psyllium husk as the associated intake.[9] A meta-analysis of 28 randomized trials and 1,924 participants supports the LDL effect, at a median dose of about 10 g.[10]

Its glycaemic evidence has strengthened too. A 2024 GRADE-assessed meta-analysis of 19 randomized trials and 962 participants found an HbA1c reduction of −0.75 percentage points, 95% CI −1.21 to −0.29, alongside reductions in fasting blood sugar and HOMA-IR, with insulin unchanged.[43] That is not evidence psyllium treats diabetes on its own, but it is a meaningful randomized-trial signal.

Its safety issue follows from how it works. Psyllium forms a gel on contact with liquid, and the FDA regulation itself carries a choking warning for inadequately hydrated product and warns people with difficulty swallowing.[9] The same gel can reduce absorption of oral medications taken at the same time; how far apart to space them is a question for a pharmacist.

Inulin repeats a pattern this library keeps finding. A randomized trial gave 24 adults at elevated diabetes risk 10 g/day of inulin or placebo for six weeks. Bifidobacteria increased significantly. Peripheral insulin sensitivity, the primary outcome, did not improve, and the authors concluded the findings question the need for larger investigations in this population.[12] The marker moved and the outcome did not follow.

For constipation, what inulin is mostly bought for, there is an umbrella review of randomized trials.[13] And there is a real tension: inulin is a fructan, which makes it a FODMAP, and fructans are recognised symptom triggers in irritable bowel syndrome.[14] The same fibre sold for gut health is a documented trigger for a common gut condition.

Chia and flax carry a useful exception to the idea that less processing is always better. Chia leads on fibre, calcium, iron and zinc; flax leads on alpha-linolenic acid at 22.8 g per 100 g.[7] Milling flax substantially improves access to its ALA and lignans, although whole seed is not inert: human studies show unground flax can still raise omega-3 and lignan markers.[41]

Here, breaking the structure improves access to what is inside it. That does not contradict our food-structure argument. It demonstrates it: structure changes what your body can reach, sometimes for the worse and sometimes for the better.

The micronutrients the app counts

Magnesium is where a marketing hierarchy runs ahead of a small literature.

Forms do differ. The most cited human comparison gave 16 adults a single 288 mg elemental dose and used 24-hour urinary excretion as the absorption proxy: oxide came in around 4% against significantly higher figures for chloride, lactate and aspartate.[15] An independent systematic review of 14 comparative studies concluded that inorganic formulations appear less bioavailable than organic ones while emphasising the evidence limitations.[16]

But the literature is small and methodologically heterogeneous: different doses, meals, formulations and indirect measures make precise rankings difficult.[16] The evidence supports differences between forms. It does not support arranging every bottle into a clinically proven league table.

That distinction matters because glycinate is marketed as though better absorption also proves better sleep or anxiety outcomes. A 2025 trial in 155 adults found a small improvement in insomnia severity with magnesium bisglycinate, but the product also supplied glycine, the result sat at the edge of statistical significance, and one author disclosed nutraceutical industry ties.[46] One modest trial does not turn an absorption hierarchy into a clinical one.

On quantity: the US tolerable upper intake level for magnesium from supplements and medications is 350 mg/day for adults, and it does not include magnesium naturally present in food. Higher supplemental intakes can cause diarrhoea, nausea and abdominal cramping.[17]

Omega-3 is where our own app number needs an explanation, and where the evidence does not divide as neatly as it is usually told.

Plant sources supply alpha-linolenic acid. Humans can convert ALA to EPA and then to DHA, but conversion is limited, especially to DHA, and estimates vary substantially with sex and with how conversion is measured.[18]

Track's omega-3 target is the adequate intake figure for ALA. A user who fills that dial entirely from seeds has met the target as measured but has not thereby met an EPA or DHA target. The number is correct; the inference is not.

Now the trial evidence, carefully. VITAL tested 1 g/day of marine omega-3 in 25,871 generally healthy adults and did not significantly reduce its primary cardiovascular composite.[19] But a meta-analysis of 13 trials and 127,477 participants found modest reductions in several cardiovascular endpoints even after REDUCE-IT was excluded.[20] REDUCE-IT itself tested something different: 4 g/day of purified prescription EPA in high-risk, statin-treated patients. Its result cannot simply be transferred to an over-the-counter capsule.

Dose changes the other side of the ledger too. A 2026 analysis of 35 trials and 114,592 participants found a significant increase in atrial fibrillation only in high-cardiovascular-risk patients receiving more than 1.5 g/day of EPA and DHA: odds ratio 1.43, an absolute difference of 0.8 percentage points. The other three dose and risk groups showed no significant increase.[45]

Food, ordinary-dose supplements and high-dose therapeutic formulations are not one category.

Vitamin D already has its own article. The trials, the Daily Value, the threshold dispute and why the Track dial reads low for almost everyone are in our vitamin D and K2 article. Beet nitrates are in our juicing article.

Absorption, as biology rather than instruction

Fat affects the absorption of fat-soluble compounds, including vitamins A, D, E and K. Their handling involves bile, micelles and lipid transport, and for vitamin D human studies generally show better absorption alongside food containing some fat. But absorption still occurs without it, and a systematic review concludes the process involves membrane carriers, particularly cholesterol transporters, rather than diffusion alone.[42] Water-soluble compounds do not depend on that machinery.

That is a mechanism, not a schedule. The more important lesson about absorption is in the next section, and it runs the other way.

The ones sold hardest, which the app does not score

Turmeric shows why better absorbed is not automatically better.

Curcumin is poorly absorbed and rapidly metabolised. Poor oral absorption had been proposed as one explanation for turmeric's historically low apparent hepatotoxicity: perhaps too little reached the circulation to produce either the claimed benefits or much harm.[21]

Manufacturers then engineered around the limitation. One small human pharmacokinetic study reported that adding 20 mg of piperine to 2 g of curcumin increased measured bioavailability twenty-fold.[22] Other formulations use lipids and nanoparticles for the same purpose.

Turmeric-associated liver injury has since become a recognised, rare phenomenon, and its cause is not settled. The US Drug-Induced Liver Injury Network published ten adjudicated cases: eight women, median age 56, five hospitalised, one death from acute liver failure. Chemical analysis confirmed turmeric in all seven products tested and three of those also contained piperine. But seven of the ten patients carried the HLA-B*35:01 allele, an allele frequency of 0.450 against population controls of 0.056 to 0.069.[23]

So the strongest signal in that series is host susceptibility, not formulation. The authors propose enhanced bioavailability as a possible contributor; the series cannot establish that piperine caused the injuries. The question is live in print: a letter in the same journal accepts the genetic finding and argues that the newer high-bioavailability technologies are themselves a putative risk factor, noting they were developed to get around European limits on oral curcumin intake, and the authors published a reply.[24]

The larger lesson survives all of that intact. Changing bioavailability changes exposure, and exposure is not the same as benefit.

Cinnamon has some of the firmest numbers in the article, and they test one of our own arguments.

Coumarin occurs naturally in cassia cinnamon and is hepatotoxic in sensitive people. EFSA set a tolerable daily intake of 0.1 mg per kg of body weight. Germany's Federal Institute for Risk Assessment translates it: for a 60 kg (132 lb) adult the limit is reached at about 2 g of cassia cinnamon of average coumarin content per day, and for a 15 kg (33 lb) child about 0.5 g.[25] Analysis of 60 ground cinnamon samples across twelve brands found mean coumarin from 2,650 to 7,017 mg/kg, while the sample from Sri Lanka — the Ceylon species — was coumarin-free.[26]

BfR also puts it in proportion: exceeding the limit slightly for one to two weeks can be considered safe, and the risk concerns large quantities over long periods.[25]

And here is the part we are not going to bury. BfR conducted bioavailability studies and found coumarin in the plant matrix of cinnamon is absorbed to a similar degree as isolated coumarin, so the limit applies to the spice as to the pure compound.[25] Our library argues repeatedly that the food matrix changes how a compound behaves. A regulator tested that question here and the matrix made no meaningful difference. A principle is more credible when you show where it was tested and did not hold.

Collagen is where the evidence changed shape when someone sorted it.

A 2025 systematic review pooled 23 randomized trials and 1,474 participants and found significant improvements in skin hydration, elasticity and wrinkles. Then the authors ran a subgroup analysis by funding source. Studies not funded by industry showed no effect on hydration, elasticity or wrinkles. High-quality studies showed no significant effect in any category. Excluding two outliers, the wrinkle effect disappeared. Their conclusion was that there is currently no clinical evidence supporting collagen supplements for skin aging.[27]

The finding is disputed in print. A letter in the same journal argues that several studies classified as non-funded were internal trials run by collagen manufacturers.[28] If that reclassification is correct, it would not obviously rescue the non-industry subgroup; it would move additional manufacturer-run studies into the industry-associated category. The meta-analysis has not been rerun on that basis.

The exercise-and-musculoskeletal literature is separate: a systematic review of collagen supplementation alongside exercise found its clearest signals in joint function and joint pain, although the studies were heterogeneous and the evidence was not uniformly positive.[29]

Spirulina and moringa share a problem, and it is not the active compound.

Aphanizomenon flos-aquae supplements have a documented history of microcystin contamination. Health Canada reports that its testing of marketed spirulina products did not detect microcystins, so routine testing is usually unnecessary,[30] although contamination has occasionally been detected in products presented as spirulina-only.[31] Heavy metals are a separate concern.[32]

Moringa's human evidence is small and population-specific,[33] while product testing has identified heavy metals and, in smaller datasets, other contaminants, with systematic safety surveillance remaining sparse.[34]

Turmeric supplies the clearest supply-chain example. Investigators analysing 524 samples across Bangladesh found lead chromate adulteration in seven of the nine major turmeric-producing districts, with lead concentrations up to 500 times the national limit.[35] The practice was economic, and subsequent enforcement substantially reduced it.[36]

The active compound and the contamination question are different problems. Contaminants can enter or accumulate through cultivation, soil and water, harvesting, processing, or deliberate adulteration.

Creatine, briefly

Creatine is not foreign to the body: it occurs naturally in meat and fish, and the body synthesises it from amino acids. Commercial creatine powder is manufactured, and it carries a Supplement Facts panel. SmartBar scores Nutrition Facts panels, so it falls outside what we assess, exactly as every other supplement does. We score the steak; we do not score the tub.

The evidence deserves more room than a paragraph. Read more on creatine.

When the bottle contains something that isn't on the label

There is a version of the quality problem more serious than a mineral sitting above or below its stated amount.

FDA maintains a Health Fraud Product Database of products cited for undeclared ingredients. Its 2026 record includes honey products containing undeclared sildenafil and tadalafil, a joint-pain capsule containing meloxicam, a metabolism product containing DMAA and tianeptine, and a product adulterated with yellow oleander, a poisonous plant. FDA states plainly that the list covers only a small fraction of the potentially hazardous products on the market, and that absence from it does not mean a product is safe.[40]

Two things keep this in proportion. The categories are concentrated: sexual enhancement, weight loss, energy and performance, and body building. This is not a finding about magnesium or psyllium. And the danger is specific rather than abstract. FDA warns that hidden PDE5 inhibitors such as sildenafil and tadalafil can interact dangerously with nitrate medicines and lower blood pressure to unsafe levels.[40] When the active ingredient is missing from the label, the person taking it does not know there is an interaction to avoid.

Which is why independent verification can matter even when it tells you nothing at all about whether the advertised benefit is real.

What a quality certification proves, and what it cannot

A certification can answer a quality question without answering an efficacy question, and the difference matters more here than almost anywhere else in food.

Independent testing programmes exist precisely because identity and contamination are recurring problems. USP's Verified Mark indicates a product contains the labelled ingredients in the declared amounts, is free of harmful levels of specified contaminants including heavy metals, microbes and pesticides, will break down and release into the body, and was made under good manufacturing practices. USP is explicit about why: testing has shown some supplements contain significantly less or more of the key ingredient than the label states.[37] NSF certifies against a standard covering label claims, toxicology review, and screening for undeclared ingredients and contaminants, with an additional tier for substances banned in sport.[38]

None of those marks evaluates whether the supplement works. A verified vitamin D tablet is confirmed to contain the right dose of real vitamin D. The seal says nothing about whether you need vitamin D.

That is this article in one sentence. What is in the bottle and whether the bottle should exist are two different claims, and they are routinely sold as one.

One distinction about SmartBar's own scoring

Supplement-testing marks are not certifications SmartBar scores, because supplements are outside our database. The certifications that do affect SmartBar scoring, including USDA Organic and Non-GMO Project Verified, apply to products we do score and verify different things about how those products were produced.[39] We award credit for the independently verified standard, not for the burden of obtaining a seal.

Read more on certifications, including what each one verifies and why some affect the score while others do not.

Key takeaways

Some supplements solve a defined problem better than food can. Folic acid around conception carries the strongest possible recommendation grade and does not wait for a measured deficiency. Food first is a default, not an ideology.

Supplements are neither pre-approved nor unregulated. Manufacturing rules require contamination limits; FDA does not assess safety or effectiveness before sale.

Psyllium has unusually strong evidence for a supplement sold as fibre, with an authorised health claim and randomized-trial support for LDL and HbA1c.

Magnesium forms differ in absorption, but the clean hierarchy advertised online runs ahead of a small and heterogeneous literature, and better absorption is not a clinical outcome.

Better absorbed is not automatically better. Turmeric's rare liver injury tracks most strongly with a genetic susceptibility marker, and enhanced-absorption formulations are a plausible contributor rather than an established cause.

Collagen's skin effect survived pooling and did not survive sorting by funding source.

Quality certification tells you what is in the bottle. It does not tell you the bottle works.

How to think about this in thirty seconds

Ask what the serving is before you believe a nutrient number.
Ask what problem the supplement solves: a measured deficiency, a predictable dietary gap, an evidence-based preventive use, or a marketing claim.
Treat enhanced absorption as a change in dose, not a free upgrade.
Ask who funded the study that persuaded you.
Judge the stack, not only the bottle: an ingredient repeated across a multivitamin, a powder, an electrolyte mix and a single-nutrient supplement can turn ordinary amounts into a combined exposure nobody intended.
If the worry is what is actually in the container, look for independent quality verification, and remember that quality testing does not prove efficacy.

What SmartBar does with this

We sell no supplements and score none, which is why this page can be written the way it is. We have nothing in the category to protect.

What we do is count what you eat. If Track shows a micronutrient sitting low, the first useful question is which foods contain it in servings you would plausibly eat. That is why "high in magnesium" returns pumpkin seeds rather than a product.

Whole food is our default, not our ideology. When the goal is ordinary nourishment, food brings the nutrient with other nutrients, a structure, and a dose shaped by eating rather than scooping. But some supplements solve problems food does not solve reliably or in time: a documented deficiency, a predictable dietary gap, or a preventive use backed by strong evidence. In those cases the bottle is not a failed version of food. It is the appropriate tool.

The sequence is what matters. Define the need, then judge the product. Not the other way around.

This article is educational and is not medical advice. Nothing here is a reason to start, stop or change any medication. Several substances discussed can interact with prescription medicines or affect their absorption, and whether any of them is appropriate for you, in what amount, is a question for a pharmacist or clinician who knows your history and what you already take. If you are pregnant, breastfeeding, planning a pregnancy, have liver or kidney disease, a bleeding disorder, or are scheduled for surgery, that conversation matters more, not less.

Keep exploring: Nutrition · Vitamin D and K2 · Juicing and blending · Front-of-package claims · Longevity

Medical Disclaimer: Read our full Medical Disclaimer

References

US Preventive Services Task Force (2023). "Folic Acid Supplementation to Prevent Neural Tube Defects: US Preventive Services Task Force Reaffirmation Recommendation Statement." JAMA 330(5):454–459. DOI: 10.1001/jama.2023.12876. PMID: 37526713. (Grade A recommendation: for persons planning to or who could become pregnant, high certainty that folic acid supplementation has a substantial net benefit in preventing neural tube defects. The recommendation applies to all persons who could become pregnant, not only those planning pregnancy, and does not require a demonstrated deficiency.)
Centers for Disease Control and Prevention. Folic acid and neural tube defect prevention. (Folic acid is the only type of folate shown in studies to prevent neural tube defects. Neural tube defects occur in the first weeks of pregnancy, frequently before pregnancy is known. CDC also reports that about half of women surveyed named enriched cereal grain products as their only source of folic acid, indicating fortification carries much of the population-level benefit.)
US Food and Drug Administration. "Dietary Supplements" and "FDA 101: Dietary Supplements." (By statute, FDA does not approve dietary supplements for safety and effectiveness, and many dietary supplements can be marketed without notifying FDA. Manufacturers are responsible for ensuring products are safe and comply with labelling and quality requirements. FDA inspects facilities, monitors adverse event reports, and takes action against adulterated or misbranded products after marketing.)
US Food and Drug Administration, current good manufacturing practice for dietary supplements, 21 CFR Part 111, with FDA's small entity compliance guide. (Requires manufacturers to establish specifications for identity, purity, strength and composition, and limits on contamination that may adulterate the product, and to verify that the production system meets them. Finished-batch verification is required for a subset of batches identified through a sound statistical sampling plan, unless the manufacturer elects to verify every batch, and particular finished-product specifications may be exempted from finished-batch testing with documented justification and appropriate upstream controls. v2 of this article described the requirement as testing every finished batch, which is broader than the regulation.)
US Food and Drug Administration, draft guidance on new dietary ingredient notifications (2022). (New dietary ingredients require a premarket safety notification at least 75 days before marketing; a product containing an NDI marketed without the required notification is adulterated. FDA estimates that with 80,000 or more products on the market, more than 4,600 NDI notifications should have been submitted, against roughly 1,200 received.)
Pumpkin and squash seed kernels, roasted. US Department of Agriculture, FoodData Central 170556. (Per 100 g: 574 kcal, magnesium 550 mg or 131% DV, zinc 7.64 mg, iron 8.07 mg, potassium 788 mg, protein 29.84 g, fibre 6.5 g.)
Chia seeds, dried, USDA FoodData Central 170554, and flaxseed, 169414. (Chia per 100 g: 34.4 g fibre, 631 mg calcium, 335 mg magnesium, 7.7 mg iron, 4.6 mg zinc, 407 mg potassium, 17.8 g ALA, 486 kcal. Flax per 100 g: 22.8 g ALA and more protein.)
Dates, Medjool, USDA FoodData Central 171707. (Per 100 g: 277 kcal, 66 g sugar, 6.7 g fibre, 696 mg potassium, 64 mg calcium, 54 mg magnesium. One Medjool date is about 24 g pitted. Magnesium per calorie: pumpkin seeds 0.96 mg/kcal against Medjool dates 0.195 mg/kcal, a ratio of approximately 4.9.)
US Food and Drug Administration. Health claims: soluble fiber from certain foods and risk of coronary heart disease. 21 CFR 101.81. (Authorised for psyllium seed husk. The regulation identifies 7 g or more per day of soluble fibre from psyllium seed husk as the intake associated with the claim, and requires a qualifying food to contain at least 1.7 g of soluble fibre from psyllium per reference amount customarily consumed. The regulation also carries a choking warning for inadequately hydrated psyllium and warns people with difficulty swallowing.)
Jovanovski, E., Yashpal, S., Komishon, A., Zurbau, A., Blanco Mejia, S., Ho, H.V.T., et al. (2018). "Effect of psyllium (Plantago ovata) fiber on LDL cholesterol and alternative lipid targets, non-HDL cholesterol and apolipoprotein B: a systematic review and meta-analysis of randomized controlled trials." American Journal of Clinical Nutrition 108(5):922–932. DOI: 10.1093/ajcn/nqy115. PMID: 30239559. (28 trials, 1,924 participants, median dose approximately 10.2 g; LDL mean difference −0.33 mmol/L.)
(Reference removed at v6. The 2020 overview of meta-analyses previously cited for psyllium's HbA1c effect became redundant when reference 43 superseded it. The number is left vacant rather than renumbered mid-cycle; renumbering happens once, manually, at production.)
Randomized controlled pilot trial of inulin supplementation in adults at elevated risk of type 2 diabetes. PMCID: PMC8471706. (24 adults, 10 g/day inulin against matched maltodextrin placebo, six weeks. Bifidobacteria increased significantly; peripheral insulin sensitivity, the primary outcome, did not improve. Authors state the findings question the need for larger investigations in this population.)
Chung, W.S.F., et al. (2023). "Inulin as an agent to alleviate constipation: an umbrella review and meta-analysis of randomized controlled trials." Journal of Human Nutrition and Dietetics. DOI: 10.1111/jhn.13110. PMID: 37410051.
Gibson, P.R., Shepherd, S.J. Evidence-based dietary management of functional gastrointestinal symptoms: the FODMAP approach. PMID: 20136989. (Identifies fructans among fermentable carbohydrates associated with symptoms in irritable bowel syndrome. Inulin is a fructan.)
Firoz, M., Graber, M. "Bioavailability of US commercial magnesium preparations." Magnesium Research. PMID: 11794633. (16 healthy adults, single 288 mg elemental dose of oxide, chloride, lactate or aspartate, 24-hour urinary excretion as absorption proxy. Oxide approximately 4% fractional absorption, significantly lower than the other three. Small sample, indirect measure, and the panel did not include glycinate or malate.)
Pardo, M.R., Garicano Vilar, E., San Mauro Martín, I., Camina Martín, M.A. (2021). "Bioavailability of magnesium food supplements: A systematic review." Nutrition 89:111294. DOI: 10.1016/j.nut.2021.111294. PMID: 34111673. (433 records screened, 14 studies retained. Concludes that inorganic formulations appear less bioavailable than organic ones, with absorption also dose-dependent, while emphasising the limitations of the evidence base. Note: "organic" here means organic chemical salts, not USDA organic certification. This reference replaces a commercially adjacent source used in v1.)
National Institutes of Health, Office of Dietary Supplements. Magnesium fact sheet. (Tolerable upper intake level of 350 mg/day for adults from supplements and medications, excluding magnesium naturally present in food. High supplemental intakes can cause diarrhoea, nausea and abdominal cramping. Forms that dissolve well tend to be more completely absorbed.)
Burdge, G.C., Calder, P.C. (2005). "Conversion of alpha-linolenic acid to longer-chain polyunsaturated fatty acids in human adults." Reproduction Nutrition Development 45(5):581–597. DOI: 10.1051/rnd:2005047. PMID: 16188209; with National Institutes of Health, Office of Dietary Supplements, omega-3 fatty acids fact sheet. (Conversion of ALA to EPA is limited and further conversion to DHA very low, with reported sex differences. NIH describes conversion as very limited, citing rates under 15%, and under 1% for DHA in its immune function fact sheet. The body states conversion as limited and variable rather than as fixed percentages, because published estimates vary substantially with the compartment and method measured.)
Manson, J.E., Cook, N.R., Lee, I-M., et al. (2019). "Marine n−3 Fatty Acids and Prevention of Cardiovascular Disease and Cancer." New England Journal of Medicine 380(1):23–32. (The omega-3 arm of VITAL; 25,871 participants; 1 g/day; primary composite of major cardiovascular events not significantly reduced. The vitamin D arm is cited in our D3/K2 article.)
Hu, Y., Hu, F.B., Manson, J.E. (2019). "Marine Omega-3 Supplementation and Cardiovascular Disease: An Updated Meta-Analysis of 13 Randomized Controlled Trials Involving 127,477 Participants." Journal of the American Heart Association 8(19):e013543. DOI: 10.1161/JAHA.119.013543. PMID: 31567003. PMCID: PMC6806028. (Excluding REDUCE-IT, marine omega-3 supplementation was associated with significantly lower risk of myocardial infarction (RR 0.92), CHD death (0.92), total CHD (0.95), CVD death (0.93) and total CVD (0.97). Including REDUCE-IT strengthened the associations and introduced heterogeneity; risk reductions were linearly related to dose. Note for future reviewers: v1 of this article cited this paper in support of the claim that over-the-counter fish oil had simply failed. The paper concludes close to the opposite, and the body was rewritten.)
National Institute of Diabetes and Digestive and Kidney Diseases. "Turmeric." LiverTox. NCBI Bookshelf NBK548561. (Records that poor oral absorption was offered as a reason for turmeric's apparent lack of hepatotoxicity, with uncertainty whether systemic exposure was ever adequate to produce beneficial or adverse effects, and that bioavailability enhancement was subsequently developed, after which case reports of liver injury were published.)
Shoba, G., Joy, D., Joseph, T., Majeed, M., Rajendran, R., Srinivas, P.S. (1998). "Influence of piperine on the pharmacokinetics of curcumin in animals and human volunteers." Planta Medica 64(4):353–356. DOI: 10.1055/s-2006-957450. PMID: 9619120. (20 mg piperine with 2 g curcumin increased measured human bioavailability by 2000%. A small pharmacokinetic experiment; it does not establish that every piperine-containing product produces twenty-fold exposure.)
Halegoua-DeMarzio, D., Navarro, V., Ahmad, J., Avula, B., Barnhart, H., Barritt, A.S., et al. (2023). "Liver Injury Associated with Turmeric — A Growing Problem: Ten Cases from the Drug-Induced Liver Injury Network [DILIN]." The American Journal of Medicine 136(2):200–206. DOI: 10.1016/j.amjmed.2022.09.026. PMID: 36252717. PMCID: PMC9892270. (Ten adjudicated cases 2004–2022, all enrolled since 2011 and six since 2017. Eight women, median age 56. Hepatocellular injury in nine; five hospitalised; one death from acute liver failure. Latency one to four months. Chemical analysis confirmed turmeric in all seven products tested; three also contained piperine. Seven of ten patients carried HLA-B*35:01, allele frequency 0.450 against population controls of 0.056 to 0.069. The authors describe the HLA linkage as strong and phrase the piperine contribution as a possibility.)
Chrubasik-Hausmann, S. (2023). "Are New Technologies a Putative Risk Factor for Curcumin Liver Injury?" The American Journal of Medicine 136(8):e167. DOI: 10.1016/j.amjmed.2023.03.014. PMID: 37481332. With the authors' reply: Halegoua-DeMarzio, D., Navarro, V., Khan, I. (2023). The American Journal of Medicine 136(8):e168. DOI: 10.1016/j.amjmed.2023.04.003. PMID: 37481333. (The letter accepts the HLA-B*35:01 genetic risk finding and argues that high-bioavailability turmeric formulations are themselves a putative risk factor, noting that such technologies were developed to overcome the European Food Safety Authority's restriction on oral curcumin intake of 3 mg/kg body weight, which is 2 mg/kg for food supplements. Correction note: v2 of this article attributed a different set of arguments to a letter it described as sceptical of the prevalence framing. That characterisation came from a search summary rather than a retrieved document and could not be confirmed; it was removed and this verified correspondence substituted.)
German Federal Institute for Risk Assessment (BfR). "Cassia cinnamon with high coumarin contents to be consumed in moderation." (EFSA tolerable daily intake 0.1 mg coumarin per kg body weight. For a 60 kg adult the TDI is reached at 2 g of cassia cinnamon of average coumarin content per day; for a 15 kg child, 0.5 g. Exceeding the TDI slightly for one to two weeks can be considered safe; risk concerns large quantities over long periods. BfR bioavailability studies demonstrated that coumarin in the plant matrix of cinnamon is absorbed to a similar degree as isolated coumarin.)
Assessment of coumarin levels in ground cinnamon available in the Czech retail market. PMCID: PMC3385612. (60 samples, twelve brands. Mean coumarin 2,650 to 7,017 mg/kg; the sample from Sri Lanka was coumarin-free. Notes the human metabolic pathway for coumarin differs from the rodent pathway associated with hepatotoxicity in long-term animal studies.)
Myung, S-K., Park, Y. (2025). "Effects of Collagen Supplements on Skin Aging: A Systematic Review and Meta-Analysis of Randomized Controlled Trials." The American Journal of Medicine 138(9):1264–1277. (23 RCTs, 1,474 participants, searched to June 2024. Pooled across all trials, significant improvement in hydration, elasticity and wrinkles. In subgroup analysis, studies not funded by industry showed no effect on any of the three; high-quality studies showed no significant effect in any category; excluding two outliers, the wrinkle effect disappeared. Authors conclude there is currently no clinical evidence supporting collagen supplements for skin aging.)
Rajan, R., Divyalakshmi, C. (2025). "Comments on Effects of Collagen Supplements on Skin Aging." The American Journal of Medicine 138(11):e249. DOI: 10.1016/j.amjmed.2025.07.018. PMID: 41167924. (Argues that several studies classified as non-funded were internal research trials conducted by collagen manufacturers and should have been counted as industry-sponsored. The meta-analysis has not been rerun on that basis.)
Khatri, M., Naughton, R.J., Clifford, T., Harper, L.D., Corr, L. (2021). "The effects of collagen peptide supplementation on body composition, collagen synthesis, and recovery from joint injury and exercise: a systematic review." Amino Acids 53(10):1493–1506. DOI: 10.1007/s00726-021-03072-x. PMID: 34491424. PMCID: PMC8521576. (15 randomized trials selected from 856 records: 12 studies of recreational athletes, 2 of elderly participants, 1 of untrained premenopausal women. Authors judged the clearest benefits to be joint functionality and joint pain, in conjunction with exercise. A separate randomized trial in 167 physically active people with knee pain found no benefit over placebo, which is why the body describes the literature as heterogeneous and not uniformly positive.)
Health Canada. Quality of Natural Health Products Guide. (Products containing Aphanizomenon flos-aquae have a history of microcystin contamination so testing is necessary, with adopted tolerance limits. Health Canada tests on a significant number of marketed spirulina products have not detected such contamination, so microcystin testing for spirulina products is usually not necessary. A risk-based testing judgement, not a statement that spirulina products are contamination-free.)
French Agency for Food, Environmental and Occupational Health and Safety (ANSES), opinion on cyanobacteria-based food supplements; and Detection of Cyanotoxins in Algae Dietary Supplements, PMCID: PMC5371831. (Most studies conclude an absence of microcystins in spirulina, but detections have occurred in batches presented as spirulina-only at generally around 0.1 µg/g, against up to 18 µg/g in Aphanizomenon products. Cultivation in natural waters without appropriate quality controls allows contamination by toxin-producing species present in the environment.)
Rzymski, P., Niedzielski, P., Kaczmarek, N., Jurczak, T., Klimaszyk, P. (2015). "The multidisciplinary approach to safety and toxicity assessment of microalgae-based food supplements following clinical cases of poisoning." Harmful Algae 46:34–42. DOI: 10.1016/j.hal.2015.05.003. (The two affected individuals had been taking both spirulina and chlorella, and the implicated products had elevated metal content; the study cannot attribute the symptoms specifically to spirulina.)
Moringa oleifera human trials: a three-month unblinded randomized controlled trial in 45 Sahrawi women with type 2 diabetes, 30 receiving 10 g/day leaf powder against 15 controls; and a parallel-group randomized trial in 40 Nigerian adults with type 2 diabetes receiving 20, 40 or 60 g daily for 14 days, PMID 37229639. (Reviewers note trial doses range from about 14 to 30 g/day for one to six months and that the bitter taste can make effective doses unpalatable outside a research setting.)
Food-chemistry review of Moringa oleifera products. Frontiers in Nutrition (2021) 8:751076. (Identifies concerning contamination findings in some sampled products, including heavy metals — for example 24 leaf powders from 12 producers in Benin, of which 58.3% exceeded the cited EU lead maximum for leafy vegetables — and, in smaller datasets, polycyclic aromatic hydrocarbons and mycotoxins, while emphasising that systematic risk-assessment studies remain scarce. Also discusses uptake from contaminated soils.)
Forsyth, J.E., Nurunnahar, S., Islam, S.S., Baker, M., Yeasmin, D., Islam, M.S., Rahman, M., Fendorf, S., Ardoin, N.M., Winch, P.J., Luby, S.P. (2019). "Turmeric means 'yellow' in Bengali: Lead chromate pigments added to turmeric threaten public health across Bangladesh." Environmental Research 179:108722. DOI: 10.1016/j.envres.2019.108722. PMID: 31550596. (152 interviews and 524 samples. Lead chromate pigment adulteration found in 7 of the 9 major turmeric-producing districts; turmeric lead concentrations exceeded the national limit by up to 500 times, maximum 1,152 µg/g in powder; pigments were 2 to 10% lead by weight. Practice reported as beginning more than thirty years ago and driven by consumer preference for bright yellow, with merchants able to sell poor-quality roots by having them adulterated.)
Food safety policy enforcement and associated actions reduce lead chromate adulteration in turmeric across Bangladesh. Environmental Research (2023). (Documents substantial reduction following enforcement. Cited so that reference 35 is not read as a timeless description of turmeric.)
United States Pharmacopeia. USP Verified Mark for dietary supplements. (Indicates the product contains the ingredients listed on the label in the declared potency and amounts; does not contain harmful levels of specified contaminants including heavy metals, microbes and pesticides; will break down and release into the body; and was made under good manufacturing practices. USP states the mark does not establish overall safety, interactions, contraindications or effectiveness.)
NSF certification of dietary supplements under NSF/ANSI 173, covering label claim review, toxicology review of intentionally added ingredients, and contaminant review for undeclared ingredients and unacceptable contaminant levels; with the Certified for Sport programme additionally screening for substances banned by major athletic organisations.
US Department of Agriculture, Agricultural Marketing Service, National Organic Program. (Synthetic substances are generally prohibited in organic production unless specifically allowed on the National List, and natural substances generally allowed unless specifically prohibited; genetic engineering, ionizing radiation and sewage sludge are prohibited, as are antibiotics and growth hormones in livestock. Three years free of prohibited substances required before land qualifies. Administered by USDA-accredited certifying agents inspecting annually, with additional unannounced inspections and testing. The seal is a registered trademark with penalties for false use. The seal requires at least 95% organic ingredients; "made with organic" indicates at least 70% and carries no seal.)
US Food and Drug Administration. Health Fraud Product Database; Medication Health Fraud notifications; and Recalls — Health Fraud. (Lists products cited in warning letters, advisory letters, recalls, public notifications and press announcements for undeclared ingredients and related violations. FDA states the list includes only a small fraction of the potentially hazardous products marketed to consumers online and in retail, and that a product's absence from the list does not mean it is safe, and separately that FDA is unable to test and identify all products marketed as dietary supplements with potentially harmful hidden ingredients. 2026 entries include honey products with undeclared sildenafil and tadalafil, a joint and body ache capsule containing undeclared meloxicam, a metabolism product containing undeclared DMAA and tianeptine, and a product containing yellow oleander. Affected categories are concentrated in sexual enhancement, weight loss, energy and performance, and body building. FDA warns that undeclared PDE5 inhibitors may interact with nitrates taken by people with diabetes, high blood pressure, high cholesterol or heart disease.)
Flaxseed form and bioavailability. Austria, J.A., et al., trial baking 30 g of whole seed, milled seed or oil into muffins, finding significant plasma ALA increases with oil and milled seed but not whole flaxseed, PMID 18689552; Kuijsten, A., et al., randomized crossover in 12 subjects concluding that crushing and milling substantially improve enterolignan bioavailability, PMID 16317125; and Sanghvi, A.V., et al., crossover study giving 16 postmenopausal women 25 g of unground whole flaxseed in food bars for four weeks, in which total serum n-3 fatty acids and urinary lignans increased significantly, PMID 20035554. (The evidence is split on how much whole seed delivers, and the body says so. v1 to v3 of this article stated that flax "has to be ground to deliver anything," which the Sanghvi crossover contradicts; corrected at v4. Note also that in the muffin trial, gastrointestinal effects caused withdrawals in the whole-seed and oil groups but none in the milled group.)
Intestinal absorption of vitamin D: a systematic review. Nutrition Reviews 76(1):60–76. PMID: 29025082. (46 of 2,069 identified articles met inclusion criteria. Vitamin D was better absorbed with fat-containing meals, but absorption also occurred without fat or oily vehicles. Concludes absorption is probably passive diffusion plus a mechanism involving membrane carriers, especially cholesterol transporters, with data scarce; factors modifying cholesterol absorption also alter vitamin D absorption. This finding also prompted a one-sentence correction to the locked D3/K2 article, which had stated micellar transport as obligatory.)
Gholami, Z., Clark, C.C.T., Paknahad, Z. (2024). "The effect of psyllium on fasting blood sugar, HbA1c, HOMA IR, and insulin control: a GRADE-assessed systematic review and meta-analysis of randomized controlled trials." BMC Endocrine Disorders 24:82. DOI: 10.1186/s12902-024-01608-2. PMID: 38844885. PMCID: PMC11155034. (19 randomized trials, 962 participants, searched to 15 July 2022. HbA1c WMD −0.75, 95% CI −1.21 to −0.29, p < .001; fasting blood sugar −6.89 mg/dL, 95% CI −10.62 to −3.16, p < .001; HOMA-IR −1.17, 95% CI −2.11 to −0.23, p < .05; insulin not significantly changed. Majority of included studies rated low risk of bias. Subgroup analysis found significant differences between doses below and above 10 g/day on fasting glucose.)
National Institutes of Health, Office of Dietary Supplements. Vitamin B12 fact sheet. (Plant foods do not naturally contain vitamin B12; fortified foods and supplements are identified as the relevant sources for people consuming little or no animal foods.)
Effects of Omega-3 Fatty Acid Treatment on Risk for Atrial Fibrillation: An Updated Meta-Analysis of 35 Trials including 114,592 Individuals. Circulation: Arrhythmia and Electrophysiology (2026). DOI: 10.1161/CIRCEP.125.014785. PMID: 42517224. PMCID: PMC13489779. (35 randomized trials, 37 datasets, 114,592 participants; doses ≥500 mg/day EPA and/or DHA, ≥12 months treatment, participants ≥50 years. Only high-cardiovascular-risk patients on >1,500 mg/day showed a significant increase in atrial fibrillation: OR 1.43, 95% CI 1.14–1.79, absolute risk difference 0.8%, 0.40–1.1%. The other three strata were not significant at 1.07, 1.06 and 1.03; only 2 of the 35 trials individually found a significant increase. The analysis added 15 trials with previously unpublished AF data and included studies reporting zero events, and its authors note that earlier meta-analyses raising AF concern included no more than eight trials, and that across REDUCE-IT, GISSI-HF, RESPECT-EPA and ASCEND the cardiovascular benefit equalled or exceeded the AF increase. Note: an earlier preprint of this work reports 34 trials and OR 1.48; the published figures are used here and the two are not mixed. Note also that the paper carries disclosed relationships including commercial omega-3 interests among its authors, which is why the article states its findings and does not rely on the authors' interpretive commentary.)
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. DOI: 10.2147/NSS.S524348. PMID: 40918053. (155 adults aged 18–65 with self-reported poor sleep, randomized double-blind to 250 mg elemental magnesium as bisglycinate or placebo for four weeks, conducted across Germany. Insomnia Severity Index reduction −3.9 against −2.3 on placebo, p = 0.049, Cohen's d = 0.2. No significant differences in other psychological outcomes. The supplement also supplied 1,523 mg of glycine daily, so the effect cannot be attributed to magnesium alone. Exploratory analysis suggested greater improvement among those with lower baseline dietary magnesium intake. Competing interests: one author is managing director of a contract research organisation that receives funding from nutraceutical companies and has received presentation honoraria from them; the other authors declare none. Disclosed here for the same reason the collagen funding split is disclosed in the body.)

Sourcing note: this article cites federal regulations and agency guidance, national risk assessments, systematic reviews and meta-analyses of randomized trials, individual randomized trials, government food composition databases, supply-chain investigations and clinical case series, with DOI/PMID where available. Study designs, sample sizes, funding and limitations are stated in the reference notes rather than left implicit, industry funding is disclosed in the body where it bears on how a finding should be read, and where a finding is disputed in print the dispute is reported. This article makes no recommendation, provides no dose, protocol or regimen for the reader, and names no brand or product; doses appearing in the text are the doses trials administered or the intakes regulators have assessed. Individual medications are not catalogued, because interaction risk is product- and patient-specific and belongs with a pharmacist or clinician. Measurements are given in both metric and imperial. SmartBar's scoring criteria are described without individual point values. Cross-link targets (#) resolve once the site structure is live. Reading time is measured, not asserted: body words excluding the reference list and the metadata header, counted at 225 words per minute and rounded to the nearest minute.

PRODUCTION NOTE — DELETE BEFORE PUBLICATION

v6 is a compression pass only. No scientific conclusion, pillar or interpretation was added, removed or changed. Roughly 900 words came out of eight zones identified in review: the folic acid and MTHFR supporting figures, repetition in the regulatory section, the superseded psyllium overview, the flax evidentiary detail, the magnesium sleep qualifications, the omega-3 scaffolding, the contamination study detail, and the SmartBar scoring section. Every pillar remains. All detail removed from the body is still carried in the reference annotations.

Reference 11 is retired, not renumbered. The 2020 overview it held became redundant once reference 43 superseded it. The slot is left vacant with an explanatory line; renumbering happens once, manually, at production, per the standing rule.

Zero bibliographic red flags across the remaining 45 references.

Four corrections are recorded inside the references rather than silently applied, because this article got each wrong first: reference 20, a meta-analysis cited against its own conclusion; reference 24, arguments attributed to a letter that could not be retrieved; reference 4, finished-batch testing described too broadly; reference 41, whole flaxseed described as inert.

One correction propagated outward, from reference 42 into the locked D3/K2 article, now v6.

Two disclosures kept deliberately, both because this article's strongest finding is a funding split and selective disclosure would be indefensible: the nutraceutical industry ties in reference 46, and the commercial omega-3 relationships in reference 45. In the second case the body states the paper's findings and does not rely on the authors' interpretive commentary.

Targeted list for clinical review, so the gate stays a safety check rather than another editorial round: folic acid and pregnancy language; psyllium swallowing and medication-absorption language; the supplemental magnesium upper limit; the omega-3 and atrial fibrillation framing; turmeric hepatotoxicity and the HLA framing; cassia cinnamon and coumarin; hidden PDE5 inhibitors and the nitrate interaction; and the population list in the final medical disclaimer. The reviewer is asked to identify errors, unsafe implications, missing clinically material qualifications or misleading wording, not to restyle a locked article.

Outstanding: clinical review; production cross-links; the single manual renumbering pass; and confirmation at production that reference 40's FDA health-fraud record is current, since that database updates continuously.