Every entry below was independently verified against the primary source — not just checked for a working link, but read to confirm it actually says what it's cited for. Where our own research surfaced a mismatch between a claim and its source, we corrected the claim on the page in question rather than publish a citation that doesn't hold up; see How We Source Claims. This bibliography currently covers the sourced claims already published across the site and will grow as new investigations are added.
Search sourcesFilters the topics below as you type — no 'Enter' needed.
The evidence behind What the Label Doesn't Tell You and the How to Read a Health Claim research guide.
The central case study of Book One. Found an 18% increase in lung cancer diagnoses and an 8% increase in deaths among smokers given beta-carotene — the opposite of the hypothesized protective effect.
Cited in: What the Label Doesn't Tell You, Health Claims guide
Read the source →Run independently of ATBC on a comparable population of smokers and asbestos-exposed workers, CARET reached the same negative conclusion and was stopped about 21 months early; ATBC had run to its planned end.
Cited in: What the Label Doesn't Tell You
Read the source →A pooled analysis across multiple trials and national brands, confirming the beta-carotene risk signal held up beyond any single study.
Cited in: What the Label Doesn't Tell You
Read the source →The federal task force's review found an overall 20% increased lung cancer risk associated with beta-carotene supplementation, concentrated among smokers.
Cited in: What the Label Doesn't Tell You
Read the source →Book One's second confirming example: early observational enthusiasm for a supplement reversing under a large, rigorous trial.
Cited in: What the Label Doesn't Tell You
Read the source →The agency's own consumer-facing guidance, used to confirm exactly what the FDA does and does not claim to verify before a supplement reaches a shelf.
Cited in: What the Label Doesn't Tell You
Read the source →The NIH database behind Product Lookup and SuppleFacts. It catalogs the information printed on U.S. supplement labels — label images, ingredient forms and amounts, and every label statement — across more than 200,000 labels. It records what a label says, not what a laboratory finds in the bottle.
Cited in: Product Lookup, About Product Lookup
Read the source →The foundational statute of this investigation — the law that classifies supplements as a food category and removed premarket safety and efficacy review.
Cited in: The Hollow Aisle, What Should Change
Read the source →The regulatory definition distinguishing a structure/function claim from a disease claim — the line nearly all supplement marketing copy is written against.
Cited in: The Hollow Aisle, Regulatory Gap guide
Read the source →The source of the 79,000+ adverse event figure cited in Book Two. The authors report 79,071 dietary supplement adverse events reported to FDA's food adverse event reporting system between 2004 and 2021.
Cited in: The Hollow Aisle
Read the source →FDA's system for adverse event reports and complaints about foods and dietary supplements, earlier published as the CFSAN Adverse Event Reporting System (CAERS). Voluntary reports and the serious adverse event reports that manufacturers must send to FDA are both entered. It is the underlying record for the figure above; FDA's page itself does not state the 79,000 total.
Cited in: The Hollow Aisle
Read the source →The FDA's only use of its DSHEA authority to remove an entire ingredient category from the market — the regulatory record behind Case One in Book Three.
Cited in: The Hollow Aisle, Prove It
Read the source →The federal court case over whether red yeast rice containing concentrated lovastatin should be regulated as a drug — reversed on appeal, then won by the FDA on narrow grounds. The basis of Case Two in Book Three.
Cited in: The Hollow Aisle, Prove It
Read the source →The FDA's original administrative determination that Cholestin's concentrated drug compound disqualified it from supplement status.
Cited in: The Hollow Aisle, Prove It
Read the source →The ongoing federal registry of products found spiked with undisclosed pharmaceutical ingredients — the evidentiary basis for Case Three in Book Three, built from the pattern rather than a single company.
Cited in: The Hollow Aisle, Prove It
Read the source →The FDA's original 2022 exclusion determination for NMN and its 2025 reversal — the timeline behind Case Four in Book Three.
Cited in: Prove It
Read the source →The evidence behind The Hollow Aisle and the When Natural Isn't Safe research guide.
The DILIN registry study behind Book Two's central risk finding — the rise from 7% to 20% of documented liver injury cases attributed to supplements, and the finding that 40% of implicated green tea extract products didn't list the compound on their label.
Cited in: The Hollow Aisle, Risk & Interactions guide
Read the source →The documented case of heart transplant rejection linked to a St. John's Wort–cyclosporine interaction — the clearest single interaction case in the entire investigation.
Cited in: The Hollow Aisle, Risk & Interactions guide
Read the source →The widely discussed meta-analysis behind the cardiovascular risk caveat attached to calcium supplementation, even where fracture-risk benefit is genuine.
Cited in: The Hollow Aisle, Prove It
Read the source →Tested 23 protein powders and shakes; more than two-thirds exceeded CR's 0.5-microgram lead benchmark per serving. Naked Nutrition's Vegan Mass Gainer and Huel's Black Edition were highest, at roughly 1,570% and 1,290% of that benchmark.
Cited in: Heavy Metals in Protein Powder, Top 25, FAQ
Read the source →Tested 90 European protein supplements using validated lab methods. Animal-based products had slightly higher mean lead; plant-based had slightly higher mean cadmium — neither difference statistically significant. No health risk found at a typical 30g/day serving.
Cited in: Heavy Metals in Protein Powder, FAQ
Read the source →Tested 22 commercially available protein powders. Toxic elements were low or undetectable across the sample — described by the authors as "a reassuring snapshot, but not a reassuring quality guarantee."
Cited in: Heavy Metals in Protein Powder
Read the source →Tested 160 protein powder products from 70 brands; reported 47% exceeded a Prop 65 threshold for at least one contaminant, with plant-based and organic products averaging higher lead and cadmium than conventional whey. Full methodology and funding sources are not public.
Cited in: Heavy Metals in Protein Powder
Read the source →Establishes FDA's interim reference levels for dietary lead: 2.2 micrograms/day for children, 8.8 micrograms/day for women of childbearing age. FDA has not set an action level specific to protein powder.
Cited in: Heavy Metals in Protein Powder, FAQ
Read the source →Announces a state investigation into whether protein powder makers misrepresented product safety or failed to disclose known heavy-metal contamination, citing Consumer Reports' and the Clean Label Project's testing. No specific brands named.
Cited in: Heavy Metals in Protein Powder, FAQ
Read the source →Would have required protein product manufacturers to test for heavy metals and publicly disclose results. Passed two Senate committees before the Senate Appropriations Committee held it in suspense; did not advance out of the Senate.
Cited in: Heavy Metals in Protein Powder, FAQ
Read the source →NIH's plain-language consumer fact sheet. It states that, unlike medicines, supplements are not approved by FDA before they are sold and that FDA does not determine whether they are effective before marketing, and it cautions that independent quality-testing seals do not guarantee a product is safe or effective.
Cited in: The Hollow Aisle
Read the source →NCCIH's clinical digest on eight herbs. It describes St. John's wort as having an overall high risk of drug interaction, with documented interactions including cyclosporine, and reports that ginkgo taken with warfarin is associated with more major bleeding than warfarin alone.
Cited in: The Hollow Aisle
Read the source →NIH's clinician fact sheet on vitamin D. It states that toxicity is almost always the result of excessive intake through supplements and gives a Tolerable Upper Intake Level of 25–100 mcg (1,000–4,000 IU) depending on age.
Cited in: The Hollow Aisle
Read the source →The evidence behind The Hollow Aisle and the Who Profits From Belief research guide.
A review of dozens of MLM companies' own income disclosures, finding most participants earned $1,000 or less annually — and that most participants earned nothing at all in at least 17 companies reviewed.
Cited in: The Hollow Aisle, The Business guide
Read the source →Survey data corroborating the FTC findings — nearly half of MLM participants reported losing money once expenses were accounted for.
Cited in: The Hollow Aisle
Read the source →The original source of the 25-hour minimum nutrition education recommendation that U.S. medical schools have fallen short of for four decades.
Cited in: The Hollow Aisle
Read the source →Survey data showing U.S. medical schools averaging 19–24 hours of nutrition education against the 25-hour recommended minimum.
Cited in: The Hollow Aisle
Read the source →Reporting on the 2025 agreement by more than fifty U.S. medical schools to expand nutrition curricula to roughly 40 hours, confirming the training gap this investigation documents.
Cited in: The Hollow Aisle
Read the source →One half of the VITAL trial (25,871 participants) — found no significant reduction in cancer incidence from vitamin D supplementation.
Cited in: Prove It
Read the source →The companion VITAL result for omega-3 supplementation — no significant reduction in the composite cardiovascular endpoint, but a real 28% reduction in heart attack risk specifically.
Cited in: Prove It
Read the source →The largest trial ever run on ginkgo biloba for dementia prevention (3,069 participants) — found no protective effect.
Cited in: Prove It
Read the source →A 225-patient randomized trial finding saw palmetto did not improve BPH symptoms compared with placebo.
Cited in: Prove It
Read the source →A follow-up 369-patient trial testing escalating doses up to three times standard — reached the same negative conclusion at every dose.
Cited in: Prove It
Read the source →The source of "dysrationalia" — research finding that general intelligence correlates only weakly with resistance to cognitive bias, central to Book Three's chapter on why smart people believe unproven claims.
Cited in: Prove It
Read the source →The independent scientific review commissioned ahead of the FDA's 2004 ephedra ban, cited in Case One of Book Three.
Cited in: Prove It
Read the source →NCCIH's evidence summary: echinacea may slightly reduce the chance of catching a cold, it is unclear whether it shortens a cold, and the findings across many studies are modest and not conclusive.
Cited in: Prove It
Read the source →NCCIH's summary: melatonin may help with jet lag and appears to help delayed sleep-wake phase disorder, but there is not enough strong evidence to recommend it for chronic insomnia. It also cites a 2023 study in which 22 of 25 melatonin gummies were inaccurately labeled.
Cited in: Prove It
Read the source →The 1999 NIH notice announcing a planned double-blind, placebo-controlled prevention trial of ginkgo biloba in older adults, with one award of up to $15 million over five years. Book Three cites it for the scale of public investment behind a single rigorous trial.
Cited in: Prove It
Read the source →The evidence behind Prove It, Chapter 9, and the What Other Countries Do research guide.
The regulatory basis for Canada's premarket licensing system (Natural Product Number), evidence-scaled through a monograph model — one of two working alternatives behind the appendix's recommended framework.
Cited in: Prove It, What Should Change
Read the source →The regulatory basis for Australia's three-tier, risk-based classification system — the second working alternative behind the appendix's recommended model.
Cited in: Prove It, What Should Change
Read the source →The EU's positive-list framework — only listed vitamins, minerals, and forms may be used in a supplement, prohibited otherwise since August 2005.
Cited in: Prove It, International Comparison guide
Read the source →The EU's premarket safety evaluation requirement for any ingredient without a documented history of use before 1997.
Cited in: Prove It
Read the source →No sources match your search.
If you believe a citation on this page has been misrepresented, reach out through Coleman Publishing or brokenpromiseshealthcare.org — it will be reviewed.