The global supplement industry hit $209 billion in 2025 — but the largest 2024 study of 390,000 adults found multivitamins do not extend life. An evidence-tier audit of what works, what fails, and how DSHEA 1994 built the world's least regulated health-claims market.
The $209 Billion Confidence Game
How a single 1994 statute built the world’s least regulated health-claims market
The global dietary-supplement industry reached $209.52 billion in 2025 and is projected to hit $414.5 billion by 2033 ✓ Established [1]. It is one of the fastest-growing consumer health categories on earth, and the only one in which the manufacturer is permitted to sell a product without first proving it works, or that it is safe at the dose on the label.
This is not a story about vitamins. It is a story about a regulatory architecture that has decoupled health claims from health evidence. The Dietary Supplement Health and Education Act of 1994 — DSHEA — reclassified supplements as a category of food and stripped the Food and Drug Administration of its pre-market approval authority over them ✓ Established [12]. Under DSHEA, a manufacturer may bring a new pill, powder, or gummy to market without demonstrating either efficacy or safety to anyone. The burden of proof, uniquely in the modern American consumer-protection system, is reversed: the FDA must prove that a product on the shelf is harmful before it can be removed.
The economic consequence is visible in the market figures. The US supplements industry alone reached $68.74 billion in 2025, with over-the-counter products accounting for 75.6% of revenue [1]. Vitamins represent the largest segment at 28.2% [1], followed by herbals, sports nutrition, and a rapidly growing category of “adaptogens” and nootropics that did not exist as a marketing category fifteen years ago. Tablets remain the dominant format at 31.9% of revenue [1], though gummy and liquid formats are expanding rapidly — formats that, as we will see, have produced some of the most severe label-accuracy [8] and pediatric-poisoning [9] failures of the past decade.
That growth rate matters. At 8.9% per annum compounded over a decade, the supplement industry is expanding nearly three times faster than US healthcare spending overall, and roughly seven times faster than pharmaceutical R&D output [1]. Capital is flowing into a category that does not require capital to demonstrate, before sale, that its products do anything [12]. The result is a market structurally biased toward marketing investment over evidence investment — toward influencer reach [13], toward proprietary blends, toward formulation novelty without trials, and away from the slow, expensive work of randomized testing [2].
The comparison with the pharmaceutical sector is instructive. A new prescription drug requires, on average, ten to fifteen years of development and roughly $1.3 billion in costs before approval, with three phases of randomized human trials and a post-approval safety surveillance system. A new supplement requires none of these [12]. A US manufacturer needs only to file a New Dietary Ingredient notification with the FDA — for ingredients introduced after 1994 — and many do not [12]. The result is that the average shopper in a Walgreens aisle is choosing among products whose level of clinical evidence ranges from “multiple Phase III randomized trials” to “an in vitro experiment performed on rat cells in 1987.” The label is not required to disclose where any individual product sits on that spectrum.
Industry trade associations describe DSHEA as a consumer-empowerment statute. That framing is half right. DSHEA did expand consumer access to a category of products that had been informally regulated. It also dismantled the principal mechanism by which consumers could distinguish products that work from products that do not — namely, mandatory pre-market evidence review [12]. The thirty years since the law was passed have functioned, in effect, as a natural experiment on what happens when health claims are largely decoupled from evidence requirements. The empirical results of that experiment are now in [3] [2] [10]. They form the substance of this report.
What Actually Works
The short list of supplements with Tier-A evidence behind them
A small number of supplements have accumulated sufficient randomized-trial evidence to justify use in defined populations. They share a common feature: they correct a documented deficiency, or they target a specific physiological deficit, rather than promising general “wellness.” ◈ Strong Evidence The honest list is shorter than the supplement aisle suggests.
Begin with creatine monohydrate. A 2024 systematic review and meta-analysis pooled 16 randomized controlled trials with 492 participants aged 20.8 to 76.4 years, all using creatine monohydrate at 3–5 g/day ✓ Established [6]. Creatine improved muscle strength and power output, reproducing decades of established sports-science findings. More striking, the same analysis reported significant gains in memory and attention domains, particularly in older adults and in conditions of acute cognitive stress such as sleep deprivation, hypoxia, and severe exertion. Five of six studies in older adults reported a positive cognition signal. Creatine is one of the most-studied molecules in the supplement space, with safety established for continuous use up to two years and no serious adverse events at standard dosing.
Vitamin D is the second clear case — but only for documented deficiency. The 2019 JAMA Cardiology meta-analysis of 21 randomized trials, 83,291 patients, found no reduction in major adverse cardiovascular events, myocardial infarction, stroke, or all-cause mortality from vitamin D supplementation ✓ Established [4]. A 2024 update incorporating subsequent trials confirmed the null finding [4]. Yet vitamin D unquestionably works for what it is mechanistically designed to do: prevent rickets, osteomalacia, and severe deficiency-related bone disease, and correct the suppressed parathyroid-hormone signaling that follows from sustained 25-hydroxyvitamin D below 12 ng/mL. The error of the past decade has been generalising the deficiency-correction evidence into a population-wide longevity claim [2]. The trials do not support it [4].
The Frontiers in Nutrition 2024 meta-analysis of 16 RCTs (492 participants) found significant cognitive improvements, particularly in memory and attention, with the largest effects in older adults and under acute cognitive stress [6]. Combined with decades of sports-performance evidence, creatine has the strongest evidence base of any supplement currently sold in mainstream retail.
Omega-3 fatty acids occupy a more complicated middle ground. The 2019 Journal of the American Heart Association meta-analysis pooled 13 randomized trials with 127,477 participants and found marine omega-3 supplementation associated with roughly an 8% reduction in myocardial-infarction risk and a 7% reduction in coronary-heart-disease death ◈ Strong Evidence [5]. The REDUCE-IT trial of icosapent ethyl at 4 g/day demonstrated a 25% relative reduction in major adverse cardiovascular events in patients already on statins [5]. Yet the VITAL primary-prevention trial of EPA+DHA at 840 mg/day in 25,871 healthy adults over 50 found no significant reduction in major cardiovascular events [5]. The pattern is consistent: omega-3 works at high EPA doses in secondary prevention, in patients who already have established cardiovascular disease. The 1 g standard fish-oil capsule sold in retail does not reproduce the trials that justify the marketing.
Magnesium is a third defensible case, again limited to specific subgroups. A 2025 Hypertension meta-analysis of 38 randomized trials found magnesium supplementation reduced systolic blood pressure by 2.81 mmHg and diastolic by 2.05 mmHg compared with placebo ✓ Established [15]. The effect size grew with doses above 400 mg/day and durations over 12 weeks [15], and was largest in hypomagnesemic and hypertensive populations. Magnesium also shows modest sleep benefits in individuals with low baseline status [15]. For a normomagnesemic adult eating a varied diet, the supplementation effect is small to negligible. For a person on chronic proton-pump-inhibitor therapy, with type 2 diabetes, or with documented insufficient intake — the effect is meaningful and clinically relevant.
Vitamin B12 belongs on the short list, but for narrowly defined groups: strict vegans and vegetarians, individuals over 65 with declining intrinsic-factor production, patients on metformin or chronic acid-suppressing medication, and those with documented megaloblastic anaemia. Iron supplementation is similarly conditional: it is unequivocally beneficial for menstruating women with iron-deficiency anaemia or pre-menopausal heavy bleeders, and unequivocally harmful when taken by men or post-menopausal women without deficiency, where it elevates cardiovascular and oxidative-stress markers. Iodine, folate during pregnancy, and vitamin K1 in coagulation contexts round out the legitimate indications.
The pattern is consistent and worth stating plainly. Supplements that have survived randomized testing succeed when they correct a measurable deficit — a deficiency, a sub-clinical insufficiency, or a specific stressor — in a defined population [4] [15] [6]. They do not produce general wellness in healthy people with adequate diets [2]. The evidence base does not support the marketing claim that “everyone should be taking a daily multivitamin” or that “modern soils are depleted, so we all need supplementation.” Those claims are commercial, not scientific. They are the connective tissue of an industry whose growth depends on maintaining them.
The Marketing Theatre
When the bottle says one thing and the meta-analysis says another
For the largest categories of supplements sold by revenue, the randomized-trial evidence ranges from null to actively contradictory of the marketing claim. ✓ Established The gap between the science and the shelf is not a minor calibration error; it is the business model. [2]
The cleanest test case is the daily multivitamin. In June 2024, the National Cancer Institute and a team of NIH researchers published in JAMA Network Open the largest cohort analysis ever conducted on multivitamin use and mortality ✓ Established [2]. The study pooled three prospective US cohorts — the NIH-AARP Diet and Health Study, the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial, and the Agricultural Health Study — for a combined sample of 390,124 healthy adults followed for up to 27 years. Over the follow-up, 164,762 participants died, including 49,836 cancer deaths and 35,060 heart-disease deaths.
The result was unambiguous. Daily multivitamin use was not associated with reduced all-cause mortality, cancer mortality, or cardiovascular mortality [2]. The hazard ratio in the first follow-up phase was 1.04 (95% CI 1.02–1.07) — pointing the wrong way, though the authors caution that residual confounding and sicker users may explain the small directional signal [2]. The honest reading of the largest, longest, best-controlled cohort study on the question is straightforward: the daily multivitamin does not extend life [2]. It does not prevent cancer or heart disease in healthy adults. It is, on average, a $30-per-bottle nutritional placebo whose cumulative US revenue tops $8 billion per year [1].
These findings suggest that multivitamin use to improve longevity is not supported.
— Loftfield et al., JAMA Network Open, June 2024Vitamin C is the second category in which the marketing has decoupled from the evidence. Linus Pauling’s 1970 hypothesis that high-dose ascorbic acid prevents and treats the common cold has been examined in 53 placebo-controlled trials. The Cochrane meta-analysis is consistent across iterations: regular vitamin C supplementation does not reduce the incidence of colds in the general population ◈ Strong Evidence. It modestly reduces severity and duration in those who do catch one — by roughly 8% in adults — and may reduce cold incidence in extreme physical-stress populations such as marathon runners and soldiers in sub-Arctic training. Taking vitamin C once cold symptoms have begun shows almost no effect. The 1,000 mg “immune-support” gummy bottle on the front display has, for the typical adult, no clinically meaningful preventive effect.
Glucosamine and chondroitin sulfate represent perhaps the most expensive null result in supplement history. The NIH-funded Glucosamine/Chondroitin Arthritis Intervention Trial — GAIT — randomized 1,583 patients with knee osteoarthritis to glucosamine, chondroitin, the combination, celecoxib, or placebo over six months ✓ Established [11]. The primary outcome — a 20% or greater reduction in knee pain — was not statistically better than placebo for any supplement arm [11]. Glucosamine alone produced a 3.9-percentage-point response advantage over placebo; chondroitin 5.3 points; combination therapy 6.5 points [11]. None reached statistical significance. Subsequent Cochrane reviews have repeatedly found glucosamine ineffective for pain on validated WOMAC measures [11]. The category nonetheless generates over a billion dollars in annual US sales [1].
Branched-chain amino acids — BCAAs — sold heavily in fitness retail represent a near-textbook case of context-stripping. Isolated BCAA ingestion does stimulate muscle protein synthesis, but at a magnitude approximately six times lower than equivalent BCAAs delivered as part of a complete whey-protein matrix that contains all nine essential amino acids. The International Society of Sports Nutrition does not currently recommend BCAA supplementation to maximise muscle protein synthesis, given that whey protein is cheaper per dose, contains the same BCAAs, and additionally provides the other essential amino acids necessary for the full anabolic response. The BCAA tub on the supplement-store wall is, in a literal biochemical sense, an inferior version of a cheaper product sold one shelf over.
Collagen peptides illustrate a subtler problem: funding-source bias. A 2025 American Journal of Medicine meta-analysis stratified hydrolyzed-collagen trials by industry funding and found that the apparent benefits on skin hydration, elasticity, and wrinkles were concentrated in industry-funded trials. Independent studies showed null effects ⚖ Contested. Mechanistically, oral collagen is digested into amino acids and short di- and tri-peptides before absorption — much of the marketed mechanism (“collagen rebuilding skin collagen”) cannot proceed in the form imagined. A meaningful peptide-signaling effect may exist; the funding-stratified meta-analysis suggests that what is currently published has not yet established it convincingly.
Apple-cider-vinegar pills, biotin for hair growth in non-deficient adults, elderberry for influenza, BCAAs in well-fed athletes, fasting-mimicking weight-loss supplements, glutamine for general “gut health,” CLA for fat loss, saw palmetto for benign prostatic hyperplasia, and the vast majority of multi-ingredient “immunity blends” populate the same evidentiary territory: trials that did not replicate, mechanisms inferred from cell cultures, and survey data dressed up as clinical evidence. The bottle assumes the buyer cannot distinguish those evidentiary tiers. The bottle is usually right.
What Is Actually Inside the Bottle
Label accuracy, adulteration, and the substitution economy
Even when a supplement has solid evidence, the consumer’s exposure to that evidence depends on whether the bottle actually contains the labeled ingredient at the labeled dose. Independent testing has repeatedly shown that, for large categories of supplements, it does not. ✓ Established [10]
The cleanest single demonstration came from the New York Attorney General’s office in February 2015. Investigators commissioned DNA-barcoding analysis of store-brand herbal supplements sold at Walmart, Walgreens, Target, and GNC — Echinacea, Ginseng, St John’s Wort, Garlic, Saw Palmetto, Valerian, Ginkgo Biloba ✓ Established [10]. The result, in aggregate: only 21% of products tested positive for the labeled plant species [10]. Walmart’s store brand returned positive identifications in 4% of tests. GNC: 22%. Target: 41%. Walgreens: 18% [10]. Substitutes detected included rice powder, asparagus, beans, daucus carota — and, in several cases, a common houseplant [10].
The supplement industry pushed back hard on the methodology, arguing that DNA barcoding cannot reliably detect highly processed botanical extracts in which the source plant’s DNA may be degraded or absent. The methodological caveat is partially correct, and subsequent regulatory guidance now treats DNA barcoding as one of several testing pillars rather than a stand-alone gold standard. But the broader finding has repeatedly survived more rigorous follow-up testing using HPLC, mass spectrometry, and chromatographic profiling [14]. The 2024 USP herbal-supplement quality reports, the German BVL working-group findings [14], and the ConsumerLab.com testing programme all continue to find substantial label-accuracy failures across the herbal category — particularly for cheaper imports and unbranded private-label products.
Cohen et al. tested 25 melatonin gummy products sold in the US market. Three contained melatonin within ±10% of label; the remaining 22 ranged from 74% to 347% of the declared dose. One product contained no detectable melatonin at all. Five CBD-containing products tested at 104%–118% of declared CBD content [8]. The analysis was published in JAMA in April 2023 and has since prompted several state-level pediatric labeling proposals.
The melatonin gummy market is the most consequential current example. A 2023 JAMA analysis tested 25 melatonin gummy products on the US market and found that only three — 12% — contained melatonin within ±10% of the labeled dose ✓ Established [8]. The range was extreme: from 74% to 347% of the declared content [8]. One product contained no detectable melatonin at all [8]. The five gummies that also declared CBD content tested at 104% to 118% of label [8]. For a substance routinely taken by children at parental discretion, with no medical supervision, a 3.5-fold variation in actual dose between bottles labeled identically is a non-trivial safety failure [9].
The FDA’s own inspection record reflects the underlying manufacturing reality. From 2023 to 2024, FDA Form 483 observations of supplement manufacturers under 21 CFR Part 111 increased by 46% ◈ Strong Evidence [12]. The most common single observation in both years was failure to establish product specifications for the identity, purity, strength, or composition of the finished dietary supplement [12]. The second most common: failure to test raw materials for identity [12]. These are not edge-case quality findings. They are the foundational requirements of pharmaceutical-grade manufacturing — and a meaningful share of the supplement industry is failing them at a rising rate.
Third-party certification programmes — USP Verified, NSF Certified for Sport, Informed-Choice — provide a partial remedy. Products carrying these marks have been independently tested for identity, dose accuracy, contamination, and disintegration. The proportion of products in the global market carrying meaningful third-party verification, however, remains a minority. NSF and USP together verify a few thousand SKUs against a market of hundreds of thousands. The presence of such a mark is one of the very few reliable consumer-side signals of genuine quality. Its absence is not proof of failure, but it is proof that no independent party has confirmed what is in the bottle.
Adulteration with unlabeled pharmaceuticals deserves separate mention. The FDA maintains a Tainted Products database documenting supplements — particularly weight-loss, sexual-enhancement, and bodybuilding products — found to contain undisclosed prescription drugs, including sibutramine, sildenafil analogues, anabolic steroids, and beta-agonists. As of 2025 the database lists thousands of products. These adulterants are not labeling errors. They are deliberate inclusions designed to produce the marketed effect — at the cost of unmonitored drug interactions and unmanaged dose escalation in consumers who believe they are taking a botanical or amino-acid product.
The Hidden Harms
Liver injury, pediatric poisoning, and the 23,000 emergency-room visits a year
The framing of supplements as “natural and therefore safe” is the single most successful marketing claim of the past three decades — and the most empirically false. The CDC’s surveillance data quantify the cost in hospital terms. ✓ Established [3]
The foundational study is Geller et al., published in the New England Journal of Medicine in October 2015, using ten years of CDC adverse-drug-event surveillance from 63 hospitals ✓ Established [3]. The result: an estimated 23,005 emergency-department visits per year in the United States are attributable to adverse events from dietary supplements, resulting in approximately 2,154 hospitalizations annually. Twenty-eight percent of visits involved adults aged 20–34. Twenty-one-point-two percent involved unsupervised pediatric ingestion — children opening a bottle they should not have been able to reach. Among non-pediatric visits, weight-loss products were implicated in 25.5% and energy products in 10.0%.
Subsequent surveillance has reinforced rather than softened these numbers. The CDC’s 2022 MMWR report documented that pediatric melatonin ingestions reported to US poison-control centers rose 530% between 2012 and 2021 ✓ Established [9]. Over the period: 27,795 emergency-department visits, 4,097 hospitalizations, 287 ICU admissions, and two deaths [9]. The trajectory is being driven primarily by gummy formulations marketed in flavours and packaging visually indistinguishable from candy [8], sold in bottles of 60 to 240 units within reach of children.
Pediatric melatonin poisonings reported to US poison-control centers rose 530% from 2012 to 2021. The product category is unregulated for children, sold in candy-like gummy formats, with documented label-accuracy failures of up to 347% of declared dose. Two pediatric deaths were recorded in the surveillance window. This is not a theoretical safety concern; it is an active public-health pattern with hospital-admission data behind it.
Liver injury from supplements is now a recognized clinical category. The NIH Drug-Induced Liver Injury Network — DILIN — has accumulated case series for several botanicals once assumed safe. Turmeric and curcumin produced ten confirmed acute-liver-injury cases in DILIN, six enrolled since 2017 ◈ Strong Evidence [7]. Seven of the ten patients carried the HLA-B*35:01 risk allele, also implicated in green-tea, Polygonum multiflorum, and Garcinia cambogia hepatotoxicity. Critically, most cases involved formulations co-blended with piperine — black-pepper extract, included to boost curcumin bioavailability by approximately 2,000%. The bioavailability boost appears to convert sub-clinical exposure into clinically significant hepatic stress in genetically susceptible individuals.
Other botanicals on the DILIN watchlist include kava, kratom, ashwagandha, hydroxycut formulations, and several weight-loss blends containing usnic acid [3]. Ashwagandha’s liver-injury signal led EU regulators to flag the plant for an Article 8 procedure in 2024, with potential reproductive, thyroid, and immune-system risks contributing to the overall safety concern ✓ Established [14]. Ashwagandha is currently banned in food supplements in Denmark, restricted in France and the Netherlands [14], and unrestricted in the United States — where it is the third-fastest-growing adaptogen category by revenue [1].
| Supplement Category | Risk Level | Assessment |
|---|---|---|
| Weight-loss and stimulant blends | 25.5% of adult supplement-related ED visits. Frequent adulteration with sibutramine, beta-agonists, undisclosed stimulants. FDA Tainted Products list dominated by this category. | |
| Pediatric melatonin gummies | +530% pediatric ingestions 2012–2021. Label accuracy 12% within ±10% of claimed dose. Two confirmed pediatric deaths. Candy-like packaging implicated. | |
| Turmeric/curcumin with piperine | 10 DILIN-confirmed acute liver-injury cases. HLA-B*35:01 carriers at elevated risk. Piperine raises curcumin bioavailability ~2,000% — the bioavailability gain may be the harm. | |
| Ashwagandha extracts | Banned in Denmark; flagged in EU for thyroid, hepatic, reproductive, and immune-system effects. Long-term safety beyond 3 months not established. Unrestricted in US. | |
| Megadose vitamin C / IV vitamin C | Doses >2 g/day increase oxalate kidney-stone risk. IV megadosing in cancer contexts remains experimental, with limited Phase II evidence and case reports of harm in G6PD-deficient patients. |
Megadose vitamin C — Linus Pauling’s legacy claim — has its own toxicity register. Doses above 2 g/day increase urinary oxalate excretion and elevate kidney-stone risk; case reports of acute oxalate nephropathy in megadose users continue to appear in the renal literature. Intravenous vitamin C in cancer contexts, championed by some integrative-oncology practitioners, can produce severe hemolysis in patients with glucose-6-phosphate dehydrogenase deficiency — a relatively common enzymatic variant in populations of Mediterranean, African, and South-East Asian descent. The marketed framing (“natural antioxidant”) does not warn for either the kidney-stone or hemolysis pathway.
Geller et al. analysed ten years of CDC adverse-drug-event surveillance data from 63 hospitals participating in the NEISS-CADES system [3]. Annual estimates: 23,005 ED visits, 2,154 hospitalizations. Weight-loss supplements implicated in 25.5% of non-pediatric visits; energy supplements in 10.0%. The study remains the most rigorous quantification of the supplement-harm signal in the US to date.
The aggregate picture is that “natural” carries no pharmacological meaning. Botanicals are pharmacologically active substances with dose-response curves, drug interactions, organ-specific toxicities, and genetic susceptibility windows — exactly like the prescription drugs from which they are linguistically distinguished [7]. The industry’s safety claims rest on the absence of mandatory adverse-event reporting, not on demonstrated low risk [3]. When adverse events are systematically tracked, as in the CDC NEISS-CADES system, the harms appear at scale [9].
The Influencer Economy of Wellness
Affiliate kickbacks, proprietary blends, and rat studies dressed as human evidence
The supplement industry has always been a marketing-led category. The current generation of marketing — built on social-media influencer affiliate networks — has compressed the distance between health claims and commercial conflicts to near zero. The Federal Trade Commission’s 2024 enforcement turn is the regulatory consequence. ✓ Established [13]
The mechanics are now routinized. A supplement brand recruits influencers through affiliate programmes that pay 15% to 40% commission on each sale generated through a personalized discount code or tracking link. The influencer features the product in lifestyle content — a morning routine, a workout, a wellness day — and discloses the relationship at varying levels of compliance with FTC guidance. The audience sees a personal endorsement; the influencer earns per click and per conversion; the brand acquires customer-acquisition data at far below the cost of paid media. The structure is rational for all three parties. It is also a textbook conflict of interest in any context where the endorsed product makes a health claim.
The Federal Trade Commission’s 2024 final rule under 16 CFR Part 465 — effective October 2024 — explicitly prohibits fake reviews, undisclosed insider testimonials, and AI-generated endorsements, with civil penalties up to $51,744 per violation in 2025 inflation-adjusted dollars ✓ Established [13]. The rule was preceded by November 2023 warning letters to twelve health influencers and two trade associations for failing to disclose connections in social-media posts about aspartame and sugar [13] — a separate matter, but a clear signal of the agency’s focus. Public Citizen has formally requested an FTC investigation of specific high-profile wellness figures for potential violations of influencer-marketing standards in supplement promotion [13].
The FTC considers it deceptive when creators promote products without disclosing they were paid, received a gift, or have a personal relationship with the brand. Affiliate links are a material connection and require disclosure every time they are used.
— FTC Endorsement Guides, 2024 updateBeyond the affiliate channel, the industry deploys a small library of evidence-laundering techniques that have become familiar to anyone reading supplement labels carefully. The first is the proprietary blend: a list of ingredients followed by a single combined milligram total, without disclosing the dose of each individual ingredient. The blend allows the manufacturer to include a clinically tested ingredient at a sub-clinical dose while listing it on the label, then add cheaper filler ingredients to reach the headline weight. The buyer cannot reconstruct from the label whether they are receiving the studied dose of any active compound.
The second technique is the species-of-evidence substitution. The product label cites “clinical studies” or “peer-reviewed research” without specifying that the cited research consists of in-vitro experiments on cultured cells, single-arm uncontrolled human studies funded by the manufacturer, animal trials in rodents, or surrogate-marker outcomes that have not been validated against clinical endpoints. The phrase “clinically proven” on a supplement label is largely meaningless under DSHEA, which permits structure/function claims with substantially weaker evidence than would be required for a drug claim.
The third is the structure/function escape hatch itself. A supplement may claim to “support healthy joint function,” “promote restful sleep,” or “help maintain immune health” without demonstrating efficacy, provided the label includes a small disclaimer stating that the FDA has not evaluated the claim and that the product is not intended to diagnose, treat, cure, or prevent any disease. Consumer-research data show that the average shopper does not distinguish between “supports immune health” and “treats colds.” The disclaimer language was designed to make that distinction; its empirical effect on consumer interpretation is approximately zero.
Watch for: (1) proprietary blends that hide individual ingredient doses; (2) “clinically studied” claims that cite in-vitro, animal, or unblinded single-arm research; (3) structure/function language (“supports,” “promotes,” “maintains”) used as a regulatory escape hatch from disease-treatment claim standards. These are not edge cases. They are core to how a substantial fraction of the industry markets products.
The compounding effect of these techniques is that an evidence-literate consumer cannot, from the label alone, distinguish a product backed by Phase III randomized trials from a product whose entire evidence base is a 2003 mouse study. The price difference at retail typically does not reflect the evidence difference either. A $60 bottle of resveratrol with a glossy label backed by anti-aging marketing may have weaker human-trial evidence than a $12 bottle of generic creatine monohydrate sold in plain packaging. The shelf signals the inverse.
16 CFR Part 465, finalised in August 2024 and effective October 2024, prohibits fake reviews, undisclosed insider testimonials, and AI-generated endorsements [13]. The supplement and wellness sectors were named explicitly in the rule’s Federal Register preamble as core enforcement targets. Brand-side liability extends to companies that direct, finance, or benefit from non-compliant influencer endorsements.
This is why the FTC’s 2024 rule matters disproportionately for the supplement category. By making fake reviews and undisclosed affiliate relationships subject to civil penalties, it raises the cost of the marketing techniques on which much of the modern supplement-influencer economy depends [13]. The agency’s capacity to enforce the rule against the long tail of TikTok and Instagram creators is limited; the deterrent value of a single high-profile prosecution is, however, considerable. The industry has noticed.
What the Regulators Are (and Are Not) Doing
The transatlantic split, country comparisons, and the limits of post-market enforcement
Thirty-one years after DSHEA, the global regulatory picture for supplements has diverged sharply. The European Union, the United Kingdom, Australia, and several individual member states have moved toward pre-market notification and ingredient-level scrutiny. The United States has not. ◈ Strong Evidence [12]
The regulatory split is best understood through specific cases. Ashwagandha is unrestricted in the US, where it is sold as a daily adaptogen in mainstream retail [1]. In Denmark it is banned in food supplements outright [14]. France and the Netherlands have published advisories restricting use in pregnancy, lactation, and specific risk groups, citing potential effects on thyroid function, reproductive hormones, and hepatic toxicity ✓ Established [14]. The EU Heads of Agencies Working Group on Food Supplements has flagged ashwagandha as a priority candidate for an Article 8 procedure that could result in EU-wide restrictions [14]. The same plant, with the same in-vitro evidence base, produces three radically different regulatory outcomes across three regulatory regimes.
Melatonin presents a similar pattern. In the United States it is sold as a dietary supplement in any dose, in any format including pediatric gummies, with no prescription required [8]. In the United Kingdom and most EU member states melatonin is a prescription-only medicine. In Germany it is available over the counter only at low doses for short-term jet-lag indications. In Australia it shifted to over-the-counter in 2021 only at low doses for adults over 55 with insomnia. The transatlantic divergence on melatonin is the proximate cause of the pediatric-poisoning surveillance pattern documented in the CDC’s MMWR report [9] — the harm signal exists where the regulatory regime is permissive.
Australia’s Therapeutic Goods Administration occupies a middle position. Listed supplements (low-risk) require pre-market notification and must use only ingredients on a permitted list, but no efficacy review is required. Registered supplements (higher-risk) require evidence review. The split has produced one of the most coherent regulatory regimes in the OECD: Australian consumers receive both ingredient-level scrutiny and the practical convenience of mainstream retail availability. The Therapeutic Goods Administration’s post-market surveillance and recall data are also more accessible than the FDA’s, providing better empirical visibility into adverse-event patterns.
Canada’s Natural Health Products Regulations, effective 2004, require all natural health products — including supplements — to obtain a Natural Product Number from Health Canada based on a pre-market submission documenting safety, quality, and efficacy. The Canadian system has been the regulatory model most often cited by US reform advocates: it preserves consumer access while imposing meaningful evidence requirements. Implementation has been imperfect, with substantial enforcement backlogs, but the regulatory architecture is structurally distinct from DSHEA.
The compare-grid below sets the structural comparison out plainly.
Pre-market regimes (EU, UK, Canada, Australia partial)
Positive lists of permitted ingredients with maximum doses, reviewed by EFSA, MHRA, Health Canada.
EFSA rejected ~80% of submitted health claims for supplements; surviving claims are evidence-supported.
Pre-market submission required for new ingredients; novel-food regulation closes loopholes.
Mandatory pharmacovigilance reporting in EU/UK; suspended licences for safety failures.
Specific population restrictions for ashwagandha, melatonin, herbals with hormonal activity.
Post-market regime (United States, DSHEA)
No positive list; any ingredient marketed before 1994 is grandfathered. New Dietary Ingredient notifications often skipped.
Structure/function claims permitted with low evidence threshold; “not evaluated by FDA” disclaimer is the only required signal.
FDA must prove harm post-market; no ingredient registry; SKU-level listing not yet mandatory.
Manufacturer reporting required only for serious events; CDC NEISS-CADES is the principal independent surveillance source.
Few category-level restrictions; ashwagandha unrestricted; melatonin available in pediatric gummy formats.
The bipartisan Mandatory Product Listing proposals advancing through the US Congress in 2025–2026 would partially close the DSHEA gap by requiring supplement manufacturers to register each product’s ingredients and labels with the FDA [12]. The legislation does not impose pre-market efficacy review and would not bring the United States into alignment with European or Canadian frameworks. It would, however, create an SKU-level inventory that the FDA presently lacks — the agency cannot currently produce a definitive count of how many supplement products are on the US market [12]. Industry groups have split, with the Council for Responsible Nutrition supporting the listing requirement and the Natural Products Association opposing it as regulatory creep. The legislative odds remain uncertain.
The Evidence-Tier Conclusion
Who actually benefits, and what an honest supplement aisle would look like
The randomized-trial evidence does not support the marketing claim that everyone needs supplementation. It supports a narrower, more useful claim: defined supplements correct defined deficits in defined populations. The structural mismatch between what the evidence shows and what the industry sells is the heart of the problem. ◈ Strong Evidence
Pull together the threads of the preceding sections and a clear evidentiary picture emerges. A small number of supplements — creatine monohydrate [6], vitamin D for deficient individuals [4], omega-3 at high EPA doses for secondary cardiovascular prevention [5], magnesium for hypomagnesemic and hypertensive populations [15], vitamin B12 for vegans and the elderly, iron for menstruating women with deficiency, iodine, folate during pregnancy, vitamin K1 in coagulation contexts — have accumulated genuine Tier-A randomized-trial evidence in defined indications ◈ Strong Evidence. Use them appropriately and the benefit is real and clinically meaningful.
A second tier — magnesium for general sleep, ashwagandha at standardized 600 mg/day extract for short-term anxiety, certain probiotic strains for IBS subtypes, hydrolyzed collagen possibly for joint pain on independently funded studies — has weaker but defensible evidence with caveats. The pattern: choose strain-specific or extract-specific products that match the trial protocols, take them for trial-duration windows rather than indefinitely, and accept that long-term safety data are typically absent.
The vast remaining bulk of the industry — multivitamins for general adult longevity [2], vitamin C megadoses for cold prevention, glucosamine for typical knee osteoarthritis [11], BCAAs for the well-fed athlete consuming adequate protein, biotin for hair growth in non-deficient adults, apple-cider-vinegar pills for weight loss, “immunity blends,” most adaptogen multipack products, and the entire weight-loss and bodybuilding adulteration economy [3] — cannot withstand the evidentiary scrutiny their marketing implies. Some are biochemically inert. Some carry meaningful harm. Almost all are sold at price points orders of magnitude above their evidence-tier-justified value [1].
The corollary follows directly. An honest supplement aisle would group products by evidence tier rather than by category, with the Tier-A short list at eye level and the rest required to display a clear evidentiary disclosure rather than a vague structure/function claim. The technical capacity to construct such a label exists. The political and economic capacity to require it has, in the United States, been blocked at every legislative attempt since DSHEA passed. The Mandatory Product Listing proposals are a marginal improvement; they are not the regulatory architecture the evidence base would justify.
What does the evidence imply for the individual reader? Three operational rules emerge from the data presented in this report. First, supplement only to correct a documented deficit. Test for vitamin D and B12 levels before taking either; supplement to within reference range and stop. Eat a varied diet first and supplement second; the cohort data on multivitamins and the meta-analyses on vitamin D and omega-3 are consistent in showing that the marginal supplementation benefit in well-nourished adults is small to absent. Second, prefer single-ingredient, third-party-verified, plain-packaged products at trial-validated doses over proprietary blends and influencer-marketed multi-ingredient stacks. The marketing premium does not buy evidence; it buys marketing. Third, treat botanicals as the pharmacologically active substances they are. Disclose them to your physician. Stop them before surgery. Do not take them in pregnancy or with prescription medications without explicit clinical advice. Particularly do not give them to children outside of pediatric medical guidance.
A short list of supplements has Tier-A randomized-trial evidence in defined deficiency and disease contexts; the vast majority of the $209 billion industry does not, and the gap between the science and the shelf is not a calibration error but the structural feature of a regulatory regime that, uniquely in modern consumer protection, requires the regulator to prove harm before a product can be removed.
The deeper lesson of the post-DSHEA era is methodological. Health markets that decouple claims from pre-market evidence requirements do not, in practice, generate the “consumer-empowered” outcomes industry advocates predict. They generate marketing-led product-development cycles, predictable label-accuracy failures, predictable adulteration, predictable adverse-event accumulation, and predictable concentration of the evidence-laundering techniques described in section 6. These are not bugs of the regime; they are the equilibrium toward which the regime is structurally biased. The regulatory architectures that have produced different outcomes — the EU Food Supplements Directive, Canada’s NHP Regulations, Australia’s TGA, the EU novel-food regime — are different not in detail but in structure. They require evidence before sale rather than harm after sale.
The supplement industry will continue to grow. Demographic ageing, chronic-disease load, declining trust in conventional medicine, and the influencer-affiliate marketing channel all push in the same direction [13]. The question is not whether the industry will reach the projected $414 billion by 2033 [1]. It will. The question is whether, by then, the regulatory architecture in the largest consumer market will have evolved to bring evidence and marketing into closer alignment, or whether the gap documented in this report will continue to widen. On current trajectory the latter is the base case. The cost of that base case — measured in unnecessary ED visits [3], in unrecognised liver injury [7], in pediatric melatonin poisonings [9], in money spent on products that do not work [2] — will continue to be borne, as it always has been, by the consumers furthest from the evidence.