Homeopathy’s documented harms came from the factory, not the dilution

Picture a parent in a pharmacy aisle holding a box of teething tablets labeled Belladonna 12X. The number signals heavy dilution. When the Food and Drug Administration tested tablets like these, the agency said its analysis should not have detected any belladonna at all. It did, and in some tablets the amount of the plant’s toxic alkaloids far exceeded what the label stated.

That finding cuts against both usual stories about homeopathy. Critics say the remedies are only sugar and water, so at worst they do nothing. Fans call them natural and gentle. The documented harms fit neither account. They came from factories that could not control what went into a tablet, from a nasal gel labeled homeopathic that delivered a real dose of zinc, from contaminated eye drops, and from remedies used where effective treatment was needed.

This article starts with that harm record, then reports what the largest government reviews concluded, including an Australian dispute usually told from one side only. It grades specific claims rather than the whole practice, using the five evidence verdicts this collection applies throughout. And it keeps two questions apart: whether homeopathy’s explanation of itself is plausible, and what the trials measured.

The harm record starts in manufacturing

FDA’s laboratory tested eight lots of teething tablets from two manufacturers. The belladonna content was not uniform from tablet to tablet, and levels of scopolamine or atropine in some tablets far exceeded the label. The agency’s reading was blunt: a toxic ingredient present in inconsistent amounts pointed to a poorly controlled manufacturing process and an unnecessary risk to children under two.

FDA worded its adverse event data carefully. A 2016 search of its reporting database found 99 cases consistent with belladonna toxicity, including reports of infant deaths and seizures that the agency described as possibly related to teething products. That is not proof of cause, and a voluntary reporting system cannot give a rate.

The second case shows that a homeopathic label does not guarantee a diluted product. By 2009, FDA had received more than 130 reports of anosmia, the loss of the sense of smell, linked to intranasal zinc products labeled homeopathic. Used as directed, the agency concluded, they delivered a significant daily dose of zinc inside the nose. A 2010 paper in the Archives of Otolaryngology, drawing on 25 patients from a University of California, San Diego nasal dysfunction clinic, applied the standard causation criteria epidemiologists use and concluded that intranasal zinc gluconate causes loss of smell.

Other findings fill out the pattern:

• Contamination. In 2019 FDA sent warning letters to four companies over a homeopathic eye drop whose samples were non-sterile and had a dangerously high pH that could injure the eye. The agency reports more than 20 warning letters since 2020, including for sterility and contamination problems.
• Toxic starting materials. FDA’s consumer page, current as of July 2026, notes that nux vomica, used in some remedies, contains strychnine, and that weak manufacturing controls can leave more of an ingredient in a product than the label says.
• Alcohol. The National Center for Complementary and Integrative Health notes that liquid products may contain alcohol at levels FDA permits to be higher than in conventional drugs.

Then there is harm from substitution. In March 2015 FDA warned consumers not to rely on over-the-counter asthma products labeled homeopathic, which it had never evaluated, because poorly treated asthma can lead to life-threatening attacks. NCCIH notes that products called nosodes have been promoted as replacements for vaccines, and that the Centers for Disease Control and Prevention finds no credible scientific evidence for that claim.

A 2012 review of published case reports by Posadzki and colleagues at the Peninsula Medical School in Exeter, England, found 38 reports covering 1,159 patients with adverse effects, including four deaths. Eight reports involved homeopathy used in place of conventional care. Case reports show that harm happens, not how often. The broader evidence on what postponed treatment costs is laid out in our report on why people set aside proven care.

What homeopathy holds, in its own terms

Homeopathy was developed in Germany in the late 1700s. FDA and NCCIH summarize its two founding principles. The first is like cures like: a substance that produces symptoms in a healthy person can, in diluted form, treat similar symptoms in a sick one. The second, which FDA calls the law of infinitesimals, holds that the more a substance is diluted, the more potent it becomes.

As the European Academies’ Science Advisory Council, known as EASAC, describes the method, a starting substance goes through repeated rounds of dilution and vigorous shaking, called succussion, which the practice holds increases potency. FDA defines a homeopathic drug as one whose active ingredients and dilutions, written on labels as codes such as 12X, appear in the Homeopathic Pharmacopeia of the United States. Many products are diluted so far that no molecule of the original substance remains, NCCIH notes. One proposed explanation, weighed by the official bodies below, is that the water retains an imprint of the substance.

Official scientific bodies have judged that explanation, and their verdicts belong to them. A UK House of Commons committee concluded in 2010 that like cures like is theoretically weak and called the imprint idea scientifically implausible. EASAC said in 2017 that any memory effects in water are far too short-lived and short-range to account for claimed effects.

Those judgments answer one question and leave the other open. An implausible mechanism raises the bar a positive trial must clear, but it does not by itself prove a trial wrong, and a positive trial would not prove the mechanism. The authors of the most-cited positive meta-analysis opened their paper by acknowledging that homeopathy seems scientifically implausible, and then analyzed the trials anyway.

What thirty years of trial reviews found

The record reads best in order, because each major review answered the one before it.

1991. Kleijnen and colleagues in the Netherlands, writing in the BMJ, found 81 positive results among 105 interpretable controlled trials. Most trials seemed to be of very low quality, and the authors judged the evidence positive but insufficient for firm conclusions, partly because nobody could tell how much negative work had gone unpublished.

1997 and 1999. Linde and colleagues in Munich pooled 89 placebo-controlled trials in The Lancet and found an odds ratio of 2.45 in homeopathy’s favor, falling to 1.66 in the 26 better-quality trials. They concluded the results were not compatible with homeopathy being entirely placebo, yet found insufficient evidence that it clearly worked for any single condition. Two years later the same team reanalyzed the same trials and found clear evidence that better-designed studies produced less positive results. Quoting the 1997 number without the 1999 paper misreports both.

2005. A University of Berne team led by Aijing Shang matched 110 homeopathy trials with 110 trials of conventional treatments. In both groups, smaller and weaker trials showed bigger effects. Among the largest higher-quality trials, homeopathy’s odds ratio was 0.88, with a range that included no effect, while conventional medicine‘s showed a clear benefit. The authors called this compatible with homeopathy’s effects being placebo effects, a more cautious claim than proof.

A 2008 reanalysis by Lüdtke, of a German foundation that funds complementary medicine research, and Rutten, of a Dutch association of homeopathic physicians, found that Shang’s answer shifted depending on where the size cutoff was drawn, and leaned heavily on a single trial of muscle soreness in 400 long-distance runners.

2014 and 2017. Robert Mathie, then of the British Homeopathic Association and later the Homeopathy Research Institute, led two reviews. For individualized prescribing, built around a long consultation, only 3 of 32 trials were judged reliable; together they produced an odds ratio of 1.98, which the authors presented cautiously as possible small, specific effects. For standardized products, again only 3 of 75 trials were reliable, and their pooled result was not statistically significant.

2022. An independent check came from Gerald Gartlehner and colleagues at Danube University Krems in Austria. Since 2002, almost 38 percent of registered homeopathy trials had gone unpublished, half of published randomized trials had never been registered, and a quarter of primary outcomes had been changed from the registry. Unregistered trials showed effects nearly four times larger than registered ones, whose pooled effect did not clearly differ from zero.

None of that depends on anyone’s view of mechanism. It is ordinary reporting bias, which also distorts conventional medicine, and it explains why the pooled positive numbers shrink as the filters tighten. Sham controls carry complications of their own, argued in detail in our acupuncture coverage, but homeopathy has the easy version: an inert pellet is simple to make and hard to tell apart from the real product.

Australia’s two verdicts, and the fight between them

The most quoted modern verdict came from Australia’s National Health and Medical Research Council in 2015: “NHMRC concludes that there are no health conditions for which there is reliable evidence that homeopathy is effective.” According to NCCIH, the overview behind it drew on 57 systematic reviews covering 176 studies. The council added that homeopathy should not be used for conditions that are chronic, serious or could become serious.

The dispute began earlier. A University of South Australia team under contract to the council had produced a 2012 draft that rated the overall evidence poor to moderate and found encouraging evidence, at its Grade C, for five conditions, among them fibromyalgia, middle-ear infection and upper respiratory infections in adults. The council never published it. In August 2019 its chief executive, Professor Anne Kelso, released an annotated version, saying a significant amount of misinformation had built up about it. The draft, she explained, was never endorsed by the council or its committee, had not been through standard quality assurance, and had been superseded.

Advocates read the history differently. A 2021 paper in the journal Cureus by Dana Ullman, whose listed affiliation is a homeopathy education company in Berkeley, argued that the 2015 overview treated trials as reliable only above 150 participants without saying so in its executive summary, and described the 2012 draft as concealed until freedom-of-information requests surfaced it. Those are the practitioner side’s stated objections.

A homeopath took the matter to Australia’s Commonwealth Ombudsman, whose investigation ran from September 2017 to July 2023. As reported by the Australian outlet Medical Republic, Ombudsman Iain Anderson said it produced no adverse findings about the review or the council, which had improved several processes along the way. He also said that, without independent scientific expertise, his office could not conclusively determine the questions of scientific method at the heart of the complaint. The council was not found to have acted improperly, and no one with authority ruled that its method was right.

A newer evaluation came next. Prepared by a contracted review firm for the council’s Natural Therapies Working Committee and approved on 16 August 2024, it assessed 93 studies across 20 priority conditions using the GRADE certainty system. Against placebo, homeopathy appeared to offer little or no benefit for most outcomes, at certainty ranging from moderate to very low.

The exceptions were specific. Low-certainty evidence suggested it might reduce medication use in allergic rhinitis, from one trial of 108 people, and disease severity in atopic dermatitis, from three trials with 172 people. Against no treatment, it was probably better at cutting infection frequency in recurrent upper respiratory infections, at moderate certainty, from one trial of 256 people. The report cautioned that when usual care is part of a comparison it is often impossible to isolate homeopathy’s own effect, and it did not assess safety.

Insurance rules reflect the earlier verdict. Since 1 April 2019 Australian private insurers have been barred from covering 16 natural therapies, homeopathy among them. When seven became eligible again on 1 July 2025, homeopathy was not one of them.

How US law treats a product labeled homeopathic

In the United States, homeopathic products are drugs in law, and the reason traces to one senator. According to FDA, when the Federal Food, Drug, and Cosmetic Act passed in 1938, its Senate sponsor, Royal Copeland, himself a homeopathic practitioner, added a provision recognizing the homeopathic pharmacopeia alongside the standard US drug compendia.

NCCIH’s homeopathy page, last updated in April 2021, reports that FDA withdrew its 1988 compliance policy for these products in 2019. That page predates FDA’s final guidance of December 2022, which says homeopathic drugs face the same legal requirements as other drugs, premarket approval included. FDA states that no product labeled homeopathic is FDA-approved. Being regulated as a drug and being approved as one are different things.

Because nearly everything on the shelf is technically unapproved, the guidance names where enforcement will concentrate: products tied to injury reports or risky ingredients, anything taken other than by mouth or on the skin, products aimed at serious conditions or vulnerable groups, and products with significant quality problems.

Advertising falls to the Federal Trade Commission. Its Enforcement Policy Statement, announced on 15 November 2016, holds homeopathic claims to the same standard as claims for other drugs. An unsupported claim may avoid being deceptive, the FTC said, if the marketing effectively tells consumers there is no scientific evidence the product works and that its claims rest only on “theories of homeopathy from the 1700s” not accepted by most modern medical experts. The disclosure must stand out close to the claim.

That statement is an enforcement policy describing when the FTC would not treat a claim as deceptive. It is not a labeling rule requiring particular words on every box. Homeopathy also turns up on a professional exam, and the naturopathy report explains where it sits on the board examination for naturopathic doctors.

Other countries reached different decisions

Governments weighing similar evidence have acted differently, which shows how far policy can sit from science.

• United Kingdom. A House of Commons committee concluded in February 2010 that homeopathic products perform no better than placebos. On 30 November 2017 NHS England advised that primary-care prescribers should not start homeopathic prescriptions for new patients, and the British Homeopathic Association’s court challenge failed on 5 June 2018. It was prescribing guidance, not a ban.
• European science academies. EASAC, formed by the national science academies of EU member states, found in September 2017 no known disease with reliable, reproducible evidence of an effect beyond placebo, and urged plain labeling of ingredients and amounts.
• France. On 28 June 2019 the Haute Autorité de Santé, the national health technology agency, assessed some 1,200 products against more than 1,000 publications and gave an unfavorable opinion on continued reimbursement.
• Switzerland. Compulsory insurance covers classical homeopathy, according to the federal health office’s February 2026 guide, but only from a qualified medical specialist trained in it. Coverage followed a 2009 vote in which two-thirds of voters backed adding complementary therapies to the constitution. In 2016 the interior ministry said proving efficacy and cost-effectiveness for these disciplines as a whole was impossible, and granted them equal insurance standing anyway, a policy choice rather than an efficacy finding.

Grading the claims you are likely to hear

Each claim gets one of this collection’s five verdicts. The reasoning behind the scale, and the base rates showing that inconclusive results are common across all of medicine, sit in the evidence-verdicts explainer.

• Homeopathy works for health conditions in general. Tested and not supported. Australia’s two reviews, the UK committee and EASAC reached that conclusion by different routes.
• It is harmless because it is so diluted. Tested and not supported, by FDA’s own laboratory and report data.
• It cuts medication use in hay fever or eases eczema better than placebo. Insufficient evidence to assess. The low-certainty signals rest on one and three small trials, a claim not yet tested at adequate scale rather than one too vague to test.
• Individualized homeopathy beats placebo. Insufficient evidence to assess. Three reliable trials covering different conditions, against a documented record of selective publication, are too thin to count as a reliably measured effect.
• It reduces repeat respiratory infections compared with doing nothing. Positive trial evidence, contested interpretation. Moderate certainty, but from one trial with no placebo arm and usual care mixed in, so what produced the difference is open to argument.
• Nosodes protect like vaccines. Insufficient evidence, in the sense that no credible test supports it, according to CDC as relayed by NCCIH. The federal concern is a nosode standing in for a vaccine.

When the question belongs with a clinician

Some situations need medical care whatever product is involved. A seizure, labored breathing, or an infant who is unusually hard to wake is an emergency, and keeping the package helps clinicians and poison control see what was given. Asthma that is getting worse, such as needing a rescue inhaler more often or waking at night short of breath, belongs with the clinician who manages it.

A loss of smell after using any nasal product needs a doctor. So does any condition that is chronic, serious or likely to become serious, and any question about whether a nosode can take the place of a childhood vaccine, which is a conversation for your child’s pediatrician.

The Takeaway

Two findings stand, and they answer different questions. On effectiveness, the largest independent reviews converge on little or no benefit beyond placebo, with a handful of narrow exceptions graded low certainty or open to dispute. On safety, the documented record points away from dilution and toward two things: whether a product contains what its label says, and whether it is being used where effective care is needed. The word homeopathic on a box answers neither question, and it does not mean FDA has approved what is inside.

Sources

1. U.S. Food and Drug Administration. “Laboratory Analysis of Homeopathic Teething Tablets (no date on page; eight lots from two manufacturers, Belladonna 12X HPUS labeling, alkaloid levels far exceeding label).”
2. U.S. Food and Drug Administration. “Homeopathic Drug Products: Guidance for FDA Staff and Industry, final guidance, December 2022 (1938 Copeland provision; same statutory requirements and premarket approval; enforcement priorities; footnotes 12 and 13 on belladonna reports, intranasal zinc anosmia reports and 2019 eye drop warning letters).” fda.gov
3. U.S. Food and Drug Administration. “Homeopathic Products (content current as of 17 January 2025).” fda.gov
4. U.S. Food and Drug Administration. “Some Homeopathic Products May Put You at Risk (content current as of 6 July 2026).” fda.gov
5. U.S. Food and Drug Administration. “FDA warns consumers about potential health risks of over-the-counter asthma products labeled as homeopathic (content current as of 20 March 2015).” fda.gov
6. National Center for Complementary and Integrative Health, National Institutes of Health. “Homeopathy (Last Updated April 2021; predates FDA's December 2022 final guidance).” nccih.nih.gov
7. Federal Trade Commission. “Enforcement Policy Statement on Marketing Claims for OTC Homeopathic Drugs.” ftc.gov
8. Federal Trade Commission. “FTC Issues Enforcement Policy Statement Regarding Marketing Claims for Over-the-Counter Homeopathic Drugs (press release, 15 November 2016).” ftc.gov
9. National Health and Medical Research Council (Australia), hosted by Analysis & Policy Observatory. “NHMRC Information Paper: Evidence on the effectiveness of homeopathy for treating health conditions (2015).” apo.org.au
10. Australian Government Department of Health (prepared by HTANALYSTS for the NHMRC Natural Therapies Working Committee). “Natural Therapies Review 2024: Homeopathy Evidence Evaluation (approved 16 August 2024).” health.gov.au
11. Australian Government Department of Health. “Natural therapies review (private health insurance exclusions since 1 April 2019).” health.gov.au
12. Australian Government Department of Health. “PHI 51/25: Private Health Insurance Legislation Amendment Rules (No. 5) 2025 (seven natural therapies eligible from 1 July 2025).” health.gov.au
13. The Limbic. “Woodhead M. NHMRC releases 'hidden' homeopathy review as minister rethinks rebate ban, 27 August 2019.” thelimbic.com
14. The Medical Republic. “Swannell C. Homeopathy loses another fight with reason, 8 August 2023 (Commonwealth Ombudsman's statements on the NHMRC review).” medicalrepublic.com.au
15. UK House of Commons Science and Technology Committee. “Evidence Check 2: Homeopathy, Fourth Report of Session 2009-10, HC 45 (2010).” publications.parliament.uk
16. High Court of England and Wales (The National Archives). “R (British Homeopathic Association) v NHS England [2018] EWHC 1359 (Admin), judgment of 5 June 2018.” caselaw.nationalarchives.gov.uk
17. European Academies' Science Advisory Council (EASAC). “Homeopathic products and practices (statement, September 2017).” easac.eu
18. Haute Autorité de Santé (France). “Homeopathic medicines: insufficient therapeutic efficacy for reimbursement (28 June 2019).” has-sante.fr
19. Swiss Federal Office of Public Health. “Guide to compulsory health insurance, February 2026 edition (legal position as of 1 January 2026; complementary medicine coverage).” bag.admin.ch
20. swissinfo.ch. “Swiss to recognise homeopathy as legitimate medicine, 29 March 2016.” swissinfo.ch
21. Archives of Otolaryngology-Head & Neck Surgery. “Davidson TM, Smith WM. "The Bradford Hill criteria and zinc-induced anosmia: a causality analysis." Arch Otolaryngol Head Neck Surg, 2010. PubMed. DOI: 10.1001/archoto.2010.111.” pubmed.ncbi.nlm.nih.gov
22. International Journal of Clinical Practice. “Posadzki P, Alotaibi A, Ernst E. "Adverse effects of homeopathy: a systematic review of published case reports and case series." Int J Clin Pract, 2012. PubMed. DOI: 10.1111/ijcp.12026.” pubmed.ncbi.nlm.nih.gov
23. BMJ. “Kleijnen J, Knipschild P, ter Riet G. "Clinical trials of homoeopathy." BMJ, 1991. PubMed. DOI: 10.1136/bmj.302.6772.316.” pubmed.ncbi.nlm.nih.gov
24. The Lancet. “Linde K, Clausius N, Ramirez G, Melchart D, Eitel F, Hedges LV, Jonas WB. "Are the clinical effects of homeopathy placebo effects? A meta-analysis of placebo-controlled trials." Lancet, 1997. PubMed. DOI: 10.1016/s0140-6736(97)02293-9.” pubmed.ncbi.nlm.nih.gov
25. Journal of Clinical Epidemiology. “Linde K, Scholz M, Ramirez G, Clausius N, Melchart D, Jonas WB. "Impact of study quality on outcome in placebo-controlled trials of homeopathy." J Clin Epidemiol, 1999. PubMed. DOI: 10.1016/s0895-4356(99)00048-7.” pubmed.ncbi.nlm.nih.gov
26. The Lancet. “Shang A, Huwiler-Müntener K, Nartey L, et al. "Are the clinical effects of homoeopathy placebo effects? Comparative study of placebo-controlled trials of homoeopathy and allopathy." Lancet, 2005. PubMed. DOI: 10.1016/S0140-6736(05)67177-2.” pubmed.ncbi.nlm.nih.gov
27. Journal of Clinical Epidemiology. “Lüdtke R, Rutten ALB. "The conclusions on the effectiveness of homeopathy highly depend on the set of analyzed trials." J Clin Epidemiol, 2008. PubMed. DOI: 10.1016/j.jclinepi.2008.06.015.” pubmed.ncbi.nlm.nih.gov
28. Systematic Reviews. “Mathie RT, Lloyd SM, Legg LA, et al. "Randomised placebo-controlled trials of individualised homeopathic treatment: systematic review and meta-analysis." Syst Rev, 2014. PubMed. DOI: 10.1186/2046-4053-3-142.” pubmed.ncbi.nlm.nih.gov
29. Systematic Reviews. “Mathie RT, Ramparsad N, Legg LA, et al. "Randomised, double-blind, placebo-controlled trials of non-individualised homeopathic treatment: systematic review and meta-analysis." Syst Rev, 2017. PubMed. DOI: 10.1186/s13643-017-0445-3.” pubmed.ncbi.nlm.nih.gov
30. BMJ Evidence-Based Medicine. “Gartlehner G, Emprechtinger R, Hackl M, et al. "Assessing the magnitude of reporting bias in trials of homeopathy: a cross-sectional study and meta-analysis." BMJ Evid Based Med, 2022. PubMed. DOI: 10.1136/bmjebm-2021-111846.” pubmed.ncbi.nlm.nih.gov
31. Cureus. “Ullman D. "An Analysis of Four Government-Funded Reviews of Research on Homeopathic Medicine." Cureus, 2021 (cited only for the practitioner side's stated objections). PubMed. DOI: 10.7759/cureus.15899.” pubmed.ncbi.nlm.nih.gov