Working bibliography
The evidence, before the episode
Every source checked against a primary record and held against its topic on the map — 135 citations across 82 topics. This is groundwork, not an investigation: nothing here has been through the editorial pass, and a citation appearing on this page is not the show endorsing what it found. 40 topics also carry a blueprint — the spine an episode would follow, written from the sources above it and no further.
Sleep & circadian
Adenosine
Rises through every waking hour, falls in sleep; caffeine blocks its receptors.
- Reichert CF, Deboer T, Landolt HP. Adenosine, caffeine, and sleep-wake regulation: state of the science and perspectives.10.1111/jsr.13597 · PMID 35575450
Review. Adenosine is now accepted as an endogenous sleep-regulatory substance and caffeine acts as its receptor antagonist. The authors are careful that a DIRECT role in sleep homeostasis is still argued, and that acute and chronic caffeine differ — worth keeping when the episode is written.
Orexin
Narcolepsy type 1 is the loss of these neurons; the newest sleep drugs block their receptors.
- Jacobson LH, Hoyer D, de Lecea L. Hypocretins (orexins): The ultimate translational neuropeptides.10.1111/joim.13406 · PMID 35043499
Review. Confirms both halves of the node: loss of hypocretin cells in humans causes narcolepsy type 1 with cataplexy, and that causal link produced a drug class — suvorexant and lemborexant approved for insomnia, daridorexant then filed.
Sleep-onset latency
Minutes between lights out and actual sleep; melatonin trials shorten it by about seven.
- Ferracioli-Oda E, Qawasmi A, Bloch MH. Meta-analysis: melatonin for the treatment of primary sleep disorders.10.1371/journal.pone.0063773 · PMID 23691095
19 trials, 1683 subjects. Sleep latency fell by 7.06 minutes (95% CI 4.37–9.75) — this is the source of the 'about seven minutes' figure. The authors say plainly the absolute benefit is smaller than other insomnia drugs, and argue for melatonin on its side-effect profile rather than its effect size.
- Yue JL, Chang XW, Zheng JW, et al. Efficacy and tolerability of pharmacological treatments for insomnia in adults: a systematic review and network meta-analysis.10.1016/j.smrv.2023.101746 · PMID 36701954
69 studies, 17,319 patients, 20 drugs ranked. Orexin receptor antagonists came top for sleep latency, wake after sleep onset and sleep efficiency. Useful counterweight: it also states the long-term adverse effects of ALL these medications are unclear.
Suprachiasmatic nucleus
About twenty thousand neurons above the optic chiasm, set by light through the eye.
- Hastings MH, Smyllie NJ, Patton AP. Molecular-genetic manipulation of the suprachiasmatic nucleus circadian clock.10.1016/j.jmb.2020.01.019 · PMID 31996314
Review from the MRC LMB. The SCN is the principal circadian timekeeper of mammals, built on a transcription–translation feedback loop (Per/Cry against Clock/Bmal1) that exists in ALL major tissues — the SCN's job is co-ordinating them, not keeping the only clock. Also flags astrocytes in the SCN network, which is newer than most popular accounts.
Immune
Autoimmunity
In lupus and RA, self-antibodies appear years before symptoms; boosting is the wrong verb.
- Yamout BI, Alroughani R. Multiple Sclerosis.10.1055/s-0038-1649502 · PMID 29791948
Review, used here as the worked example of an autoimmune disease rather than as a source on autoimmunity generally — replace with a broader epidemiological source before broadcast. Two things to carry forward: MS prevalence is rising for two different reasons that get conflated (earlier diagnosis and longer survival, AND a true rise in incidence), and the disease is framed as a gene-by-environment interaction rather than either alone.
Common-cold trial design
Every trial defines a cold differently, and few confirm that a virus was there.
- Hemilä H, Chalker E. Vitamin C for preventing and treating the common cold.10.1002/14651858.CD000980.pub4 · PMID 23440782
The cleanest worked example of why trial DESIGN decides the answer. 29 comparisons, 11,306 people: regular supplementation did not cut cold incidence in the general population (RR 0.97) — but the same vitamin cut incidence by half (RR 0.48) in marathon runners, skiers and soldiers, shortened colds by 8% in adults and 14% in children, and did nothing at all when taken AFTER symptoms began. Prevention versus treatment, and general population versus extreme physical stress, are four different questions.
Helminth immunomodulation
Worm exposure tracks LOWER rates of autoimmune and metabolic disease. The immunology runs opposite to the folklore about parasites.
- Sipahi AM, Baptista DM. Helminths as an alternative therapy for intestinal diseases.10.3748/wjg.v23.i33.6009 · PMID 28970717
EDITORIAL, not a trial or a systematic review — treat as orientation only. Its own honest line is the one to keep: there is some evidence on safety and tolerability, but 'evidence about their impact on disease activity is lacking'. Species, dose, timing and mechanism are all still open. This is the cautionary half of the hygiene-hypothesis story.
Immune boosting
Marketing sells a single quantity that no laboratory test actually measures.
- Hemilä H, Chalker E. Vitamin C for preventing and treating the common cold.10.1002/14651858.CD000980.pub4 · PMID 23440782
The counter-example to 'immune boosting'. Cochrane's own conclusion: routine vitamin C supplementation is NOT justified for the general population. What survives is narrower and more interesting — a duration effect, and a real halving of incidence under severe physical stress. The integrative reading is not 'it works', it is 'it works for a specific person in a specific state', which is a different claim from the one on the bottle.
Innate immunity
Most cold symptoms come from the host response, not from the virus itself.
- Bekkering S, Blok BA, Joosten LAB, Riksen NP, van Crevel R, Netea MG. In vitro experimental model of trained innate immunity in human primary monocytes.10.1128/CVI.00349-16 · PMID 27733422
IN VITRO, human primary monocytes — a methods paper, not a clinical result. Worth having because it pins down what 'trained immunity' actually means physically: β-glucan, BCG and oxidised LDL all leave monocytes responding harder to unrelated bacteria days later, with a switch to glycolysis. Innate immunity having memory at all is the part that overturns the textbook.
Rhinovirus
More than 150 types circulate, and immunity to one buys little against the rest.
- Murray CS, Simpson A, Custovic A. Allergens, viruses, and asthma exacerbations.10.1513/pats.2306027 · PMID 16113420
Review, and an old one — replace before broadcast. Kept for the synergy point, which still holds: rhinovirus and allergen exposure together drive asthma exacerbations harder than either alone, which is why 'just a cold' is not just a cold in a sensitised asthmatic.
Gut & microbiome
Bile acids
Detergents that also act as hormones, and gut bacteria rewrite them en route.
- Jia W, Xie G, Jia W. Bile acid-microbiota crosstalk in gastrointestinal inflammation and carcinogenesis.10.1038/nrgastro.2017.119 · PMID 29018272
Review. The useful frame for this node: bile acids are not just detergents for fat, they are SIGNALS, read by FXR and TGR5, and gut bacteria chemically rewrite them before the host receives them. That makes bile acid metabolism a two-way conversation between liver and microbiome that touches lipid and carbohydrate handling, insulin sensitivity and innate immunity.
Intestinal permeability
Much of the leaky-gut literature rests on a zonulin ELISA that does not measure zonulin.
- Seethaler B, Basrai M, Neyrinck AM, et al. Biomarkers for assessment of intestinal permeability in clinical practice.10.1152/ajpgi.00113.2021 · PMID 34009040
The measurement problem, which is the real story under 'leaky gut'. Against the established lactulose/mannitol test in 78 people, plasma LBP tracked permeability in every cohort — but ZONULIN, the marker sold direct to consumers, only correlated in the overweight and obese group, and only when measured in stool. A zonulin result in a lean person is being asked to carry more than this data supports.
Serotonin
Ninety per cent is made in the gut, and that fraction never reaches the brain.
- Yano JM, Yu K, Donaldson GP, et al. Indigenous bacteria from the gut microbiota regulate host serotonin biosynthesis.10.1016/j.cell.2015.02.047 · PMID 25860609
The landmark, from Mazmanian's lab. Spore-forming gut bacteria drive 5-HT synthesis in colonic enterochromaffin cells, which supply the mucosa, the lumen and circulating platelets — and germ-free mice given the right microbial metabolites raise their colonic and blood serotonin. MOUSE work with human-derived bacteria. The honest framing for air: this is where 'the gut makes most of your serotonin' stops being a slogan and becomes a mechanism, but it is gut serotonin, which does not cross into the brain.
Short-chain fatty acids
Butyrate, propionate, acetate — the cells lining the colon run mostly on the first.
- Fernando MR, Saxena A, Reyes JL, McKay DM. Butyrate enhances antibacterial effects while suppressing other features of alternative activation in IL-4-induced macrophages.10.1152/ajpgi.00440.2015 · PMID 27012776
MOUSE macrophages in culture — narrow, and a better human source should replace it. Kept for two things it states directly: butyrate comes from bacterial fermentation of dietary fibre and is the PRIMARY energy source of colonocytes, and it acts on immune cells partly by inhibiting histone deacetylation. A fibre metabolite changing gene expression is the mechanism worth an episode.
SIBO
Too many bacteria in the small bowel; the breath test and the culture often disagree.
- Quigley EMM, Murray JA, Pimentel M. AGA Clinical Practice Update on Small Intestinal Bacterial Overgrowth: Expert Review.10.1053/j.gastro.2020.06.090 · PMID 32679220
The AGA's own position, and unusually blunt. 'The definition of SIBO as a clinical entity lacks precision and consistency'; true prevalence is therefore undefined; controversy remains over its role in IBS; and the antibiotic evidence base is thin enough that therapy is 'for the most part, empiric'. It also names the real problem — we do not know what a NORMAL small-intestinal bacterial population looks like, so there is no baseline to call overgrowth against. Anyone selling a SIBO protocol is ahead of this document.
Soluble fibre
Viscous types bind bile acids and lower LDL-C; fermentable types feed the bacteria.
- Gibb RD, Sloan KJ, McRorie JW. Psyllium is a natural nonfermented gel-forming fiber that is effective for weight loss: a comprehensive review and meta-analysis.10.1097/JXX.0000000000000882 · PMID 37163454
READ THE AFFILIATION FIRST. The lead author is at The Procter & Gamble Co., which sells psyllium as Metamucil, and this is a meta-analysis of six studies totalling 354 people concluding their product works. That does not make the result false — the mechanism is real and specific: psyllium is soluble but NOT fermented, so it forms a viscous gel that slows nutrient absorption rather than feeding bacteria, which is the opposite of how most fibre is sold. It does mean the effect sizes (-2.1 kg, -0.8 BMI) need an independent replication before this show quotes them. A useful teaching case: industry authorship is a reason to check, not a reason to dismiss.
Tryptophan
Only a small fraction becomes serotonin; most runs down the kynurenine pathway.
- Correia AS, Vale N. Tryptophan metabolism in depression: a narrative review with a focus on serotonin and kynurenine pathways.10.3390/ijms23158493 · PMID 35955633
Narrative review — the weakest review type, so treat as a map rather than evidence. The map itself is the point: one amino acid feeds two pathways that pull in opposite directions, and inflammation shifts the traffic from the serotonin branch toward kynurenine. That is the mechanistic bridge between 'inflammation' and 'low mood' that most popular accounts skip.
Glycemic control
Chromium
Trials are small and concentrated in people who were deficient to start with, and the effect shrinks as the methods get better.
Blueprint
- Open with the headline that sells the supplement: 28 randomised trials, HbA1c down 0.71%, fasting glucose down 19 mg/dl. On paper that rivals a drug.
- Then show the number nobody quotes: I-squared is 99.8%.
- Explain what that means without jargon — heterogeneity is a measure of whether the trials agree. At 99.8% they do not merely disagree, they are effectively measuring different things, and an average across them describes no real patient.
- So the honest statement is: SOME chromium trials found large effects. Not: chromium lowers glucose by 19 mg/dl.
- Ask what would explain such spread: baseline chromium status, the chemical form used, whether participants were deficient to begin with.
- Land on the general rule, because it applies far beyond chromium: a meta-analysis is only as meaningful as the similarity of the trials inside it, and the heterogeneity statistic is where that is declared.
Define on air: meta-analysis · heterogeneity / I-squared · HbA1c · HOMA-IR · confidence interval
Where it lands: A pooled effect too heterogeneous to act on. Most plausible in people who were deficient to start with, which is a different claim from the one on the label.
Still needed: A trial stratified by baseline chromium status, using one defined form.
- Asbaghi O, Naeini F, Rezaei Kelishadi M, et al. Effects of chromium supplementation on glycemic control in patients with type 2 diabetes: a systematic review and meta-analysis of randomized controlled trials.10.1016/j.phrs.2020.105098 · PMID 32730903
28 RCTs. Every glycaemic index moved: fasting glucose -19.0 mg/dl, insulin -12.35 pmol/l, HbA1c -0.71%, HOMA-IR -1.53. Then read the heterogeneity: I-squared is 99.8% for fasting glucose and 89.9% for HOMA-IR. At 99.8% the trials are not estimating the same quantity, and a pooled mean across them is close to meaningless — the honest statement is that SOME chromium trials show large effects, not that chromium lowers glucose by 19 mg/dl. This is the single most important caveat on the node, and it is the sort of thing supplement marketing quotes without.
Vanadium
Insulin-mimetic in animals, and toxic at the doses that did it. The clearest example on this map of a mechanism that never became a medicine.
Blueprint
- Lead with the pharmacology, because it is genuinely impressive: vanadium inhibits protein tyrosine phosphatases, the enzymes that switch the insulin signal OFF. Block them and the signal runs longer.
- The downstream effects are real insulin-mimicry — more GLUT-4 at the cell membrane, more glycogen synthesis, less gluconeogenesis.
- Now the sentence that ends the supplement conversation: the therapeutic window is VERY narrow.
- At only a few micromoles — not far above the useful range — the same compounds stop cell proliferation and cause apoptosis, necrosis and inflammation.
- Draw the general lesson about dose. A molecule that is a drug at one concentration and a cytotoxin slightly above it is a molecule that needs monitoring, not a bottle on a shelf.
- Be explicit that this show is not recommending it, and why that is a pharmacological judgement rather than a squeamish one.
Define on air: protein tyrosine phosphatase · insulin-mimetic · GLUT-4 · gluconeogenesis · therapeutic window · apoptosis vs necrosis
Where it lands: Real mechanism, unacceptable margin. The interesting question is not whether vanadium works but what it teaches about dose.
Still needed: An English-language primary source; the current citation is a Polish review with no DOI. Human dosing and toxicity data.
- Korbecki J, Baranowska-Bosiacka I, Gutowska I, Chlubek D. [Insulin-mimetic property of vanadium compounds].PMID 28132446
Review, Polish language, no DOI — find an English primary source before broadcast. The pharmacology is the story and it cuts both ways. Vanadium inhibits protein tyrosine phosphatases, which is genuinely insulin-mimetic: more GLUT-4 at the membrane, more glycogen synthesis, less gluconeogenesis. But the authors state the therapeutic window is VERY NARROW, and that at only a few micromoles the same compounds inhibit cell proliferation and cause apoptosis, necrosis and inflammation. A supplement whose effective dose sits next to its cytotoxic dose is not a supplement.
Zinc
Zinc status tracks glycaemic control and repletion helps the deficient — which is not the same thing as a treatment for diabetes.
Blueprint
- Open on quality: 32 placebo-controlled trials, 1700 people, 14 countries, published in the American Journal of Clinical Nutrition. This is a better literature than most supplement stories get.
- The numbers: fasting glucose down 14.15 mg/dl, two-hour postprandial down 36.85, HbA1c down 0.55%, HOMA-IR down 0.73, and hs-CRP down 1.31 mg/l.
- The inflammation result matters — it suggests the glucose effect is not the whole story.
- Then the subgroup finding that should change how anyone shops: the effect depended on diabetic status AND on the FORM of zinc, with inorganic salts outperforming.
- So 'zinc' on a label is not one intervention, and a trial of one salt does not license a claim about another.
- Close by contrasting this with chromium, deliberately: same category, same promise, very different evidence quality. That comparison is the episode.
Define on air: placebo-controlled · postprandial · HOMA-IR · hs-CRP · inorganic vs organic salt · subgroup analysis
Where it lands: The strongest mineral evidence in the glycaemic cluster. Real, replicated, modest, and form-dependent.
Still needed: Head-to-head comparison of zinc forms, and whether the CRP effect tracks the glucose effect or runs independently.
- Wang X, Wu W, Zheng W, et al. Zinc supplementation improves glycemic control for diabetes prevention and management: a systematic review and meta-analysis of randomized controlled trials.10.1093/ajcn/nqz041 · PMID 31161192
32 placebo-controlled trials, 1700 people, 14 countries, in the AJCN. Fasting glucose -14.15 mg/dl, 2-hour postprandial -36.85, HbA1c -0.55%, HOMA-IR -0.73, and hs-CRP -1.31 mg/l. Better quality than the chromium literature and the subgroup analysis is the useful part: the effect on fasting glucose depended on diabetic status AND on the FORM of zinc, with inorganic salts outperforming. 'Zinc' on a label is not one intervention.
Thyroid
Vitamin E
The thyroid runs on trace elements in balance, not in isolation — selenium, iodine and the antioxidants that protect the gland while it works.
Blueprint
- Start with the gland's chemistry problem: making thyroid hormone means handling hydrogen peroxide on purpose. Oxidative stress is not a side effect here, it is the process.
- Hence the trace elements. Iodine is the raw material; selenium runs the enzymes that both activate the hormone and clean up the peroxide.
- The systems point, which is the reason this node exists: these elements sit in a DYNAMIC BALANCE. Excess or deficiency of one disturbs the others — which is why correcting iodine without selenium, or the reverse, can go wrong.
- Where vitamin E is supposed to fit: as the lipid-phase antioxidant working alongside the selenium-dependent ones.
- And then the admission. The trace-element review does not cover vitamin E, and the authors say the trace-element/thyroid relationship is itself 'still unclear'.
- So the episode is about the shape of the system and the danger of correcting one element in isolation — not about a vitamin E recommendation, which the evidence here does not support.
Define on air: thyroid peroxidase · hydrogen peroxide · selenoprotein · deiodinase · glutathione peroxidase · lipid peroxidation
Where it lands: The balance argument is sound and clinically useful. The vitamin E claim specifically is not yet evidenced on this node and should not be made until it is.
Still needed: A trial of vitamin E in thyroid disease, and a source for the selenium-iodine correction sequence. Until then this episode is about interactions, not supplements.
- Zhou Q, Xue S, Zhang L, Chen G. Trace elements and the thyroid.10.3389/fendo.2022.904889 · PMID 36353227
Review. The frame this node exists for: thyroid hormone synthesis depends on several trace elements at once — iodine and selenium first, then iron, zinc, magnesium, manganese and others — and they sit in a DYNAMIC BALANCE where an excess or deficiency of one disturbs the rest. That is why correcting iodine without selenium, or the reverse, can go wrong. The authors are careful that the relationship between trace elements and thyroid disorders 'is still unclear' and needs more work. Vitamin E specifically is NOT covered here and still needs its own source — this is the trace-element scaffolding the vitamin E question hangs off.
Cardiometabolic
Betaine (TMG)
Trimethylglycine donates a methyl group to recycle homocysteine back to methionine — a second route when the folate one is blocked.
Blueprint
- Set up the fork in the road: homocysteine has two ways back to methionine. One runs on folate and B12. The other runs on betaine, and it is the one nobody mentions.
- Name the molecule plainly — betaine IS trimethylglycine, glycine carrying three methyl groups, and it hands one over.
- The trial: 100 adults with raised homocysteine, 12 weeks, double-blind. Low-dose B vitamins plus 1 g betaine cut homocysteine by 10.1% against placebo.
- Why that number is smaller than it sounds, and why it is still interesting: the fall tracked the rise in BOTH folate and betaine, which is evidence the betaine arm did work rather than riding along.
- Two caveats, said out loud. It was a combination, so betaine alone is not isolated. And the population had no mandatory folic acid fortification — in a fortified country the folate route is already running, so there is less for betaine to add.
- Declare the conflict: one author is affiliated with a supplement company.
Define on air: methyl group · methylation · remethylation · betaine · folic acid fortification
Where it lands: A plausible second route with one small positive combination trial behind it. Worth knowing about when the folate route is blocked; not yet worth a standalone recommendation.
Still needed: Betaine alone against placebo, in a fortified population, with a clinical endpoint rather than a number.
- Lu XT, Wang YN, Mo QW, et al. Effects of low-dose B vitamins plus betaine supplementation on lowering homocysteine concentrations among Chinese adults with hyperhomocysteinemia: a randomized, double-blind, controlled preliminary clinical trial.10.1007/s00394-023-03087-y · PMID 36717385
100 adults, 12 weeks, randomised and double-blind. 400 µg folic acid + 8 mg B6 + 6.4 µg B12 + 1 g betaine cut plasma homocysteine by 10.1% against placebo, and the fall tracked the rise in BOTH folate and betaine — evidence the betaine arm contributed rather than riding on the folate. Two things to keep honest: this was a COMBINATION, so it cannot isolate betaine; and the population was Chinese adults free of mandatory folic acid fortification, so the effect in a fortified country may be smaller. Note also an author affiliation with BYHEALTH, a supplement company.
Hyperhomocysteinaemia
Raised homocysteine tracks clots and strokes. Whether it CAUSES them, or just marks the damage, is still the open question.
Blueprint
- Start with the clot, not the number. Raised homocysteine tracks arterial disease, stroke AND venous thromboembolism — the last one is the part that gets left out.
- Then the honest uncertainty, in the reviewer's own words: it is still uncertain whether this is causative or a marker of vascular disease. The strongest causal evidence is in ANIMAL models.
- The mechanism, because it is specific and teachable: the endothelium loses nitric oxide — either oxidised away, or never made, because ADMA blocks the enzyme. Add thiolation of circulating and endothelial proteins, and endoplasmic-reticulum stress that switches on inflammation and apoptosis.
- The turn: lowering it did not help. B12 given to people with raised homocysteine and cardiovascular disease did not reduce heart attack or stroke, and did not change cognitive decline.
- So what is the number for? Probably as a readout of B-vitamin status and one-carbon metabolism — useful for finding a deficiency, not as a target to chase.
- End on who should still be tested: the young stroke, the unprovoked clot, the family history that does not fit.
Define on air: homocysteine · endothelium · nitric oxide · ADMA · venous thromboembolism · surrogate endpoint
Where it lands: A real marker with a real mechanism and a failed intervention. Treat a high homocysteine as a question about B12, folate, glycine and betaine — not as a disease to normalise.
Still needed: Whether the trial failures reflect the wrong population (already-diseased), the wrong agent, or a marker that was never causal.
- Lentz SR. Mechanisms of homocysteine-induced atherothrombosis.10.1111/j.1538-7836.2005.01364.x · PMID 16102030
Review, and admirably direct about the thing most coverage skips: raised homocysteine is a risk factor for cardiovascular disease, stroke and VENOUS THROMBOEMBOLISM, but 'it is still uncertain whether hyperhomocysteinemia is a causative factor or a marker of vascular disease'. The causal evidence is strongest in ANIMAL models. The mechanism, where it has been worked out, is endothelial: less bioavailable nitric oxide, either through oxidative inactivation or through inhibition of NO synthase by ADMA, plus thiolation of plasma and endothelial proteins and endoplasmic-reticulum stress driving inflammation and apoptosis. Read alongside the B12 node, where LOWERING homocysteine did not lower events — the two together are the whole argument.
Mood & neurochemistry
Abeta*56
The 2006 oligomer that shaped a field. The paper was retracted in 2024 after its images were found to have been altered.
Blueprint
- Tell it as a detective story with a documentary record, and source it from the record rather than from the reporting.
- 2006, Nature. A 56-kilodalton soluble assembly of amyloid-beta, purified from the brains of memory-impaired mice, impaired memory when given to young rats — and did so INDEPENDENTLY of plaques or neuron loss.
- Why it mattered so much: it appeared to explain why plaque burden correlates so poorly with symptoms. The toxic species was soluble and invisible on the slide.
- Sixteen years of building on it.
- Then the retraction in 2024, after image alteration was identified. PubMed itself now classifies the paper as a Retracted Publication — the listener can check that in one click, and should.
- Now the discipline the episode must show. What was retracted is THIS oligomer result. Not amyloid plaques, which are observable pathology described in 1906 and visible in any neuropathology lab. Getting that boundary right is the difference between criticism and conspiracy.
Define on air: oligomer · kilodalton · soluble vs fibrillar · retraction · image duplication · Tg2576 mouse
Where it lands: A specific, documented, checkable research failure with real consequences for a field. Precisely bounded — and the precision is the point.
Still needed: How much downstream work depended on it, and what the field's self-correction actually looked like.
- Lesné S, Koh MT, Kotilinek L, et al. A specific amyloid-beta protein assembly in the brain impairs memory. [RETRACTED]10.1038/nature04533 · PMID 16541076
PubMed itself classifies this paper as a Retracted Publication — that is checkable in one click and is how the episode should source the claim, rather than by citing journalism about it. The original claim: a 56-kDa soluble assembly, Abeta*56, purified from Tg2576 mouse brain impaired memory in young rats INDEPENDENTLY of plaques or neuron loss. It ran for 16 years and shaped a field. Retracted in 2024 after image alteration. Precision matters here: what was retracted is this specific oligomer result, not the existence of amyloid plaques, which have been observable pathology since 1906.
GSK-3beta
The enzyme joining lithium, insulin signalling and tau phosphorylation — and the least famous node on this map.
- Duthie A, et al. Pilot Trial of Lithium in Humans With Mild Cognitive Impairment (GSK-3 target engagement).10.3389/fnmol.2019.00163 · PMID 31316348
Kynurenine pathway
Inflammation induces IDO, which pulls tryptophan away from serotonin.
- Correia AS, Vale N. Tryptophan metabolism in depression: a narrative review with a focus on serotonin and kynurenine pathways.10.3390/ijms23158493 · PMID 35955633
Narrative review. Places the kynurenine branch at the junction of neuroinflammation, stress, the microbiota and BDNF regulation. Needs primary trial evidence alongside it before any claim that shifting this pathway changes symptoms.
Low-dose lithium
Doses far below the psychiatric range inhibit GSK-3 and track lower dementia rates. The trials so far are small, and two of them disagree.
- Forlenza OV, Coutinho AMN, Aprahamian I, et al. Long-term lithium treatment reduces glucose metabolism in the cerebellum and hippocampus of nondemented older adults: an [18F]FDG-PET study.10.1021/cn5000315 · PMID 24730717
RCT, 4 years low-dose lithium in amnestic MCI. Found REDUCED regional glucose metabolism — an awkward result for the neuroprotection story, and the reason this is not a clean win.
- Duthie A, van Aalten L, MacDonald C, et al. Recruitment, Retainment, and Biomarkers of Response; A Pilot Trial of Lithium in Humans With Mild Cognitive Impairment.10.3389/fnmol.2019.00163 · PMID 31316348
Well tolerated; GSK-3 biomarkers did not move — authors suggest the dose may have been too low to engage the target.
- Elefante C, Brancati GE, Torrigiani S, et al. Bipolar Disorder and Manic-Like Symptoms in Alzheimer's, Vascular and Frontotemporal Dementia: A Systematic Review.10.2174/1570159X20666220706110157 · PMID 35794767
Systematic review noting low-dose lithium may exert neuroprotective activity in AD.
- Kessing LV, Gerds TA, Knudsen NN, et al. Association of lithium in drinking water with the incidence of dementia.10.1001/jamapsychiatry.2017.2362 · PMID 28832877
73,731 dementia cases against 733,653 matched controls in Denmark — and the shape of the result is the story. It is NOT a dose-response line. Against 2.0–5.0 µg/L, dementia was LOWER above 15.0 µg/L (IRR 0.83) but HIGHER at 5.1–10.0 µg/L (IRR 1.22). A U-shape that turns the wrong way in the middle is what confounding usually looks like, and the authors say outright that confounding from other things associated with municipality of residence cannot be excluded. Use this as the evidence that the QUESTION is serious, not as evidence that lithium in water prevents dementia.
Nitrous oxide
It oxidises the cobalt atom at the centre of vitamin B12 and switches the vitamin off. Your blood level can look completely normal while the spinal cord is failing.
Blueprint
- The chemistry: N2O irreversibly oxidises the cobalt ion in cobalamin, which shuts down methionine synthase.
- The consequence: a functional B12 deficiency that produces subacute combined degeneration of the cord.
- The trap: serum B12 can be normal. Methylmalonic acid and homocysteine are the tests that see it.
- The population: teenagers and people in their twenties, buying canisters legally.
Define on air: N · i · t · r · o · u · s · · o · x · i · d · e · · — · · l · a · u · g · h · i · n · g · · g · a · s · . · · A · · l · e · g · i · t · i · m · a · t · e · · a · n · a · e · s · t · h · e · t · i · c · · a · n · d · · a · n · a · l · g · e · s · i · c · , · · a · n · d · · t · h · e · · f · a · s · t · e · s · t · - · g · r · o · w · i · n · g · · r · e · c · r · e · a · t · i · o · n · a · l · · i · n · h · a · l · a · n · t · · i · n · · t · h · e · · U · K · .
Where it lands: A genuine, preventable, rising harm that most people using it have never been told about — and one where the standard blood test reassures you wrongly.
Still needed: Whether the antidepressant trials of a single sub-anaesthetic dose carry any of this risk. Different exposure entirely, and it has not been settled.
- Mair D, Paris A, Zaloum SA, White LM, Dodd KC, Englezou C, Patel F, Abualnaja S, Lilleker JB, Gosal D, Hayton T, Liang D, Allroggen H, Pucci M, Keddie S, Noyce AJ. Nitrous oxide-induced myeloneuropathy: a case series.10.1136/jnnp-2023-331131 · PMID 37253616
119 patients across London, Birmingham and Manchester — the largest series to date. Pins and needles was the presenting complaint in 85%, lower limbs before upper; gait ataxia common; bladder and bowel involvement frequent. The cord lesion sits at C3–C5 on T2 imaging. Canisters consumed per week correlated with methylmalonic acid (ρ=0.44), which is the functional B12 marker — and the reason a normal serum B12 does not exclude this.
- Xiang Y, Li L, Ma X, Li S, Xue Y, Yan P, Chen M, Wu J. Recreational nitrous oxide abuse: prevalence, neurotoxicity, and treatment.10.1007/s12640-021-00352-y · PMID 33770366
The review that frames the public health side: rising recreational use in adolescents and young adults, most of whom have no idea of the risk. Stopping and replacing B12 early gives a good prognosis; long-term use without treatment ends in irreversible neurological damage. The mechanism is still argued over, which is worth saying out loud.
Saffron
Seven randomised trials put it level with fluoxetine for mild-to-moderate depression. It is also the most adulterated spice on earth.
Blueprint
- The claim: a kitchen spice treats depression as well as an SSRI.
- The trials: seven RCTs, large effect against placebo, no detectable difference against fluoxetine or imipramine.
- The catch nobody prices in: saffron is the most adulterated spice in the world, and the trials used standardised extracts, not the jar.
- The dose that keeps recurring is 30 mg a day of a standardised extract — not a pinch in rice.
Define on air: S · a · f · f · r · o · n · · i · s · · t · h · e · · d · r · i · e · d · · s · t · i · g · m · a · · o · f · · t · h · e · · C · r · o · c · u · s · · s · a · t · i · v · u · s · · f · l · o · w · e · r · · — · · t · h · r · e · e · · t · h · r · e · a · d · s · · p · e · r · · b · l · o · o · m · , · · p · i · c · k · e · d · · b · y · · h · a · n · d · , · · w · h · i · c · h · · i · s · · w · h · y · · i · t · · c · o · s · t · s · · m · o · r · e · · b · y · · w · e · i · g · h · t · · t · h · a · n · · g · o · l · d · .
Where it lands: Genuinely promising and genuinely under-powered. The gap between the trial material and the retail material is the whole episode.
Still needed: A trial run outside Iran, with an independently assayed extract, powered for non-inferiority against a named SSRI.
- Yang X, Chen X, Fu Y, Luo Q, Du L, Qiu H, Qiu T, Zhang L, Meng H. Comparative efficacy and safety of Crocus sativus L. for treating mild to moderate major depressive disorder in adults: a meta-analysis of randomized controlled trials.10.2147/NDT.S157550 · PMID 29849461
Seven RCTs. Against placebo, SMD -1.22 (95% CI -1.94 to -0.49) — a large effect. Against synthetic antidepressants, SMD 0.16 (-0.25 to 0.57), which is the statistical way of saying no detectable difference. Remission, response and drop-out rates all comparable. The authors rate overall study quality as MODERATE and note moderate heterogeneity, and almost every trial is small and Iranian; that is the thing to check before this becomes an episode, not the effect size.
- Lopresti AL, Drummond PD. Saffron (Crocus sativus) for depression: a systematic review of clinical studies and examination of underlying antidepressant mechanisms of action.10.1002/hup.2434 · PMID 25384672
The review that catalogues dosages, extract sources and standardisation alongside the effect — which is the part that matters for a spice sold by the gram. Proposed mechanisms are serotonergic, antioxidant, anti-inflammatory and neuroendocrine, i.e. four plausible stories rather than one demonstrated one.
Tau
Tangles track cognitive decline better than plaques do, which has always been an awkward fact for the amyloid story.
- De-Paula VJ, Radanovic M, Diniz BS, Forlenza OV. Alzheimer's disease.10.1007/978-94-007-5416-4_14 · PMID 23225010
Standard review, useful for defining terms on air without jargon. Two hallmarks, two different proteins: neuritic PLAQUES are extracellular amyloid-beta; neurofibrillary TANGLES are intracellular, and are cytoskeletal damage caused by hyperphosphorylation of the microtubule-associated protein tau. Conflating them is the commonest error in popular coverage. Also lays out the three clinical phases — pre-symptomatic, pre-dementia, dementia — which is why 'early intervention' arguments carry so much weight.
The amyloid cascade
Thirty years of drug development aimed at plaques. The plaques are real and visible; that they are the CAUSE is the part still contested.
Blueprint
- Open on the strange fact the hypothesis has never explained well: plaque burden and cognitive impairment correlate badly. People with heavy plaque who were fine; people with modest plaque who were not.
- State the cascade fairly — amyloid accumulates, triggers downstream damage, tangles and death follow — because a critique of a straw man is worthless.
- Then the drugs. Anti-amyloid antibodies clear amyloid on PET and the FDA approved them.
- The challenge, from a peer-reviewed review of eight antibodies: the fall in PET signal may not be a one-to-one measure of amyloid removal. It may partly reflect therapy-related brain damage — and the raised incidence of ARIA and the reported loss of brain volume are the evidence offered.
- The authors ask the FDA to PAUSE existing and new approvals pending phase 4 data. That is an extraordinary thing for a review to say, and the episode should mark it as extraordinary rather than normalise it.
- Close on what would settle it, and note that one of the authors has spent a career arguing the amyloid hypothesis is wrong — which is context, not disqualification.
Define on air: amyloid PET · ARIA · brain atrophy · phase 4 · accelerated approval · surrogate endpoint
Where it lands: A serious, published, mechanistic challenge to how the benefit of an approved drug class is being measured. Not proof the cascade is wrong — proof that the evidence for the drugs is contested by people qualified to contest it.
Still needed: The phase 4 data the authors ask for; the trial-sponsor response to the volume-loss argument.
- Høilund-Carlsen PF, Revheim ME, Costa T, et al. Passive Alzheimer's immunotherapy: a promising or uncertain option?10.1016/j.arr.2023.101996 · PMID 37414156
The serious, peer-reviewed challenge — and far stronger ground than 'the plaques were made up'. Reviewing RCTs of eight anti-amyloid antibodies, the authors argue the fall in amyloid PET signal is UNLIKELY to be a one-to-one measure of amyloid removal, and may instead reflect therapy-related brain damage, pointing to the raised incidence of ARIA and the reported loss of brain volume. They call on the FDA to PAUSE existing and new approvals pending phase 4 data. George Perry is among the authors. This is the citation for a critical episode; the retraction of Abeta*56 is a separate and narrower point.
Micronutrients
Allithiamine (TTFD)
Fat-soluble thiamine crosses membranes the water-soluble salt cannot. The real question is who is actually deficient, not who feels tired.
Blueprint
- Start with the problem the derivatives were invented to solve: thiamine is water-soluble and absorption is capped, so raising the blood level with ordinary tablets is hard.
- Introduce the lipid-soluble derivatives — benfotiamine, and allithiamine/TTFD — as an absorption strategy, not a different vitamin.
- The mechanism is specific and worth the airtime: benfotiamine activates transketolase, which pulls glucose out of the damaging alternative routes — polyol, hexosamine, protein kinase C, and advanced glycation end products.
- That is a coherent account of why a B vitamin might matter in diabetic nerve damage, rather than a vague 'nerve support' claim.
- Then the boundary this episode must not cross: the evidence reviewed is for BENFOTIAMINE. Allithiamine/TTFD is a different molecule with a different absorption route, and inheriting benfotiamine's trial data is not allowed.
- Say plainly what would change the verdict: TTFD trials with the same endpoints.
Define on air: water- vs lipid-soluble · transketolase · polyol pathway · hexosamine pathway · advanced glycation end products · TTFD
Where it lands: A real mechanistic case for lipid-soluble thiamine derivatives in diabetic neuropathy — held by benfotiamine. Allithiamine currently borrows that credibility without having earned it.
Still needed: Any randomised trial of allithiamine/TTFD itself.
- Várkonyi T, Körei A, Putz Z, et al. Advances in the management of diabetic neuropathy.10.23736/S0026-4806.17.05257-0 · PMID 28541026
CAREFUL: this reviews BENFOTIAMINE, a different lipid-soluble thiamine derivative from allithiamine/TTFD. It is here for the mechanism the two share and for the fact that a derivative was needed at all — benfotiamine activates transketolase and thereby diverts glucose away from the polyol, hexosamine, protein-kinase-C and advanced-glycation pathways. That is the pharmacological case for the derivatives over plain thiamine. Do not let the episode slide from benfotiamine evidence to allithiamine claims; TTFD needs its own trials cited.
Creatine
One of the few supplements where the strength data are solid and the brain data are not.
- Watson G, Casa DJ, Fiala KA, et al. Creatine use and exercise heat tolerance in dehydrated men.PMID 16619091
THE DEHYDRATION PREMISE DOES NOT SURVIVE THIS ONE EITHER, and the episode has to say so before it says anything else. Double-blind randomised crossover, deliberately hostile conditions: trained men took 21.6 g/day for a week, were dehydrated by 2% of body mass, then exercised 80 minutes at 33.5 C. No difference from placebo in core temperature, heart rate, blood pressure, lactate, thirst, perceived exertion, plasma or urine osmolality, or symptom scores. Conclusion: short-term creatine 'did not increase the incidence of symptoms or compromise hydration status or thermoregulation'. Twelve men, short-term, so it is not the last word — but it is directly on the claim. A sceptical creatine episode is still available; it just cannot be built on dehydration.
Folic acid vs folate
Not the same molecule. The synthetic form needs a reduction step many people run slowly, and it can mask a B12 deficiency while the nerves keep going.
Blueprint
- Open with the win, honestly and first: mandatory fortification from 1998 prevented neural tube defects. That is a real public-health success and the episode does not get to skip it.
- The unintended benefits too — less anaemia, lower homocysteine, lower cardiovascular risk in the review's reading.
- Then turn. The same review names three potential harms: unmetabolised folic acid circulating in the blood, an increased risk of cancer, and the masking of vitamin B12 deficiency.
- Spend the time on masking, because it is the one that hurts people quietly: folic acid corrects the anaemia that would have prompted the B12 test, while the neurological damage continues unopposed.
- Distinguish the molecules carefully. Folate is the form in food; folic acid is synthetic, more stable, more bioavailable, and must be reduced before the body can use it — which is where the unmetabolised fraction comes from.
- End on the population-versus-individual tension: a policy that is right for a population can still be wrong for a particular person, and that is not a contradiction.
Define on air: neural tube defect · fortification · folate vs folic acid · unmetabolised folic acid · dihydrofolate reductase · masking
Where it lands: A genuinely successful policy with a real and under-discussed cost. The strong version of the claim — that synthetic folic acid simply does not work — is not what the literature says; the masking problem is.
Still needed: Primary sources for the cancer signal, and for the size of the unmetabolised fraction at ordinary intakes.
- Ismail S, Eljazzar S, Ganji V. Intended and unintended benefits of folic acid fortification — a narrative review.10.3390/foods12081612 · PMID 37107407
Narrative review of ~60 reports since US mandatory fortification began in 1998. It names BOTH halves without flinching. Intended benefit: fewer neural tube defects. Unintended benefits: less anaemia, lower homocysteine, lower cardiovascular risk. And the potential harms, in the authors' own list — unmetabolised folic acid circulating in the blood, increased cancer risk, and MASKING OF VITAMIN B12 DEFICIENCY. That last one is the clinically dangerous one: the anaemia resolves while the neurological damage continues. Note this is about the synthetic form specifically, which is the distinction the episode turns on.
Homocysteine
B vitamins lower it reliably; the trials found no drop in heart attacks or deaths.
- Langan RC, Goodbred AJ. Vitamin B12 deficiency: recognition and management.PMID 28925645
Carries the negative result that matters for this node: giving B12 to people with raised homocysteine and cardiovascular disease does NOT reduce myocardial infarction or stroke, and does not alter cognitive decline. Homocysteine is a real marker of B-vitamin status; lowering the number is not the same as lowering the risk it marks. A clean worked example of a surrogate endpoint failing.
Inositol
Myo- and D-chiro-inositol shift insulin resistance and ovulation in PCOS trials — one of the few supplements with both a mechanism and a signal.
Blueprint
- Start where the patient is: metformin is the gold-standard insulin sensitiser in PCOS, and a large number of women stop taking it because of what it does to their gut.
- Then the finding: 26 randomised trials, 1691 women. Inositol raised the chance of a regular cycle 1.79-fold over placebo, and was NON-INFERIOR to metformin.
- Everything else that moved: BMI, free and total testosterone, androstenedione, glucose, insulin AUC down; SHBG up.
- Be precise about what non-inferior means — not better. The argument for inositol is tolerability, not superiority.
- Be equally precise about what is missing: no fertility outcome, no live-birth outcome. Cycle regularity is a proxy for the thing most of these women actually want.
- Land it as the clearest example on this map of an integrative option that earned its place by trial, not by tradition.
Define on air: PCOS · insulin sensitiser · non-inferiority · SHBG · androstenedione · area under the curve
Where it lands: Among the best-evidenced integrative interventions on the map. Reasonable first-line alternative where metformin is not tolerated; not a fertility treatment on this evidence.
Still needed: Live-birth and pregnancy outcomes; myo- versus D-chiro-inositol ratios.
- Greff D, Juhász AE, Váncsa S, et al. Inositol is an effective and safe treatment in polycystic ovary syndrome: a systematic review and meta-analysis of randomized controlled trials.10.1186/s12958-023-01055-z · PMID 36703143
26 RCTs, 1691 women. The strongest integrative result in this set: inositol raised the chance of a regular cycle 1.79-fold over placebo and was NON-INFERIOR to metformin, the gold-standard insulin sensitiser, while also lowering BMI, free and total testosterone, androstenedione, glucose and insulin AUC and raising SHBG. The clinical argument is the side-effect profile — metformin's gastrointestinal effects are why people stop taking it. Note what this does NOT show: no fertility or live-birth outcome, and 'non-inferior' is not 'better'.
Magnesium
Blood holds about 1% of the body's supply, so a normal serum level rules out little.
- Zhang X, Li Y, Del Gobbo LC, et al. Effects of magnesium supplementation on blood pressure: a meta-analysis of randomized double-blind placebo-controlled trials.10.1161/HYPERTENSIONAHA.116.07664 · PMID 27402922
34 double-blind placebo-controlled trials, 2028 adults, in Hypertension. Median 368 mg/day for three months lowered systolic BP by 2.00 mmHg and diastolic by 1.78 mmHg. Real, replicated, and SMALL — roughly a tenth of what a first-line antihypertensive does, which is the honest way to present it rather than as either useless or a substitute. One author is at the Center for Magnesium Education and Research, worth noting. The authors themselves flag residual heterogeneity and call for better trials.
Omega-3 (EPA/DHA)
Prescription EPA cut cardiac events; the same dose off a shelf has not been shown to.
- Jia X, Kohli P, Virani SS. Omega-3 fatty acid and cardiovascular outcomes: insights from recent clinical trials.10.1007/s11883-019-0763-0 · PMID 30631963
Why 'fish oil' is not one question. Reading REDUCE-IT alongside ASCEND and VITAL, the authors conclude the benefit tracks FORMULATION and DOSE: high-dose PURE eicosapentaenoic acid reduced atherosclerotic events where mixed EPA/DHA preparations at ordinary doses did not. A trial of a 4 g pure-EPA prescription drug says nothing about a 1 g supermarket capsule, and both get reported as 'omega-3'.
Pyridoxine (B6)
The vitamin that reliably causes harm: at supplement doses pyridoxine causes a sensory neuropathy that is regularly missed for months.
Blueprint
- Open by splitting the word. 'B6' is six different molecules, and only pyridoxal 5'-phosphate — PLP — is the active one.
- The supplement on the shelf is usually pyridoxine, which the body has to convert.
- Then the finding that changes the conversation: the author states PLP-based supplements are PREFERRED over pyridoxine, because pyridoxine showed more neurotoxicity in neuronal viability testing.
- The therapeutic index is narrow. Neurotoxicity appears above roughly 100 nmol/l, which is not a wildly supratherapeutic dose.
- Because B6 metabolites are long-lived, the recommendation is WEEKLY dosing of 50–100 mg rather than daily — a genuinely counterintuitive piece of practice.
- Keep the author's own hedge in the episode: the association between elevated pyridoxine and neuropathy 'is not well established'. State the uncertainty and still act cautiously, because the downside is a peripheral neuropathy.
Define on air: vitamer · pyridoxal 5'-phosphate · therapeutic index · peripheral neuropathy · half-life
Where it lands: The strongest evidence on this map for the claim that a synthetic vitamin form can harm — specific, dose-related, and formulation-dependent rather than a blanket condemnation.
Still needed: Prospective data on neuropathy incidence at ordinary supplement doses; direct PLP-versus-pyridoxine comparison in humans.
- Reddy P. Preventing vitamin B6-related neurotoxicity.10.1097/MJT.0000000000001460 · PMID 36608063
The most direct support in this whole set for the 'synthetic form can harm you' argument, and it is specific rather than sweeping. B6 is six vitamers; only pyridoxal 5'-phosphate (PLP) is biologically active. The author states the therapeutic index is NARROW, that supraphysiologic daily dosing risks neurotoxicity above roughly 100 nmol/L, and that PLP-based supplements are PREFERRED OVER PYRIDOXINE because pyridoxine showed more neurotoxicity in neuronal viability testing. Because B6 metabolites are long-lived he recommends WEEKLY 50–100 mg rather than daily. Note the honest hedge the episode must keep: the author also says the pyridoxine–neuropathy association 'is not well established'.
Thiamine (B1)
Metformin inhibits the transporter that absorbs it; the clinical evidence is thinner.
Blueprint
- Open with the number that should stop the room: the classic triad of Wernicke's — confusion, eye signs, unsteady gait — is complete in only 10% of cases.
- And: only a few cases are diagnosed before death.
- Untreated, roughly 80% go on to Korsakoff syndrome — permanent memory damage, filled in with confabulation.
- Break the association with alcohol, which is what causes the misses. Alcohol is the commonest cause in the US, but hyperemesis gravidarum, bowel obstruction and malignancy all do it — which means it arrives on wards that are not looking for it.
- The Mayo recommendation, quoted directly: overdiagnosis and overtreatment are PREFERABLE, given thiamine's excellent safety profile.
- Close on what that implies about medical decision-making — when a test is unreliable, a treatment is harmless and a miss is catastrophic, the arithmetic says treat.
Define on air: Wernicke encephalopathy · Korsakoff syndrome · confabulation · hyperemesis gravidarum · gait apraxia
Where it lands: A treatable emergency that is routinely missed because clinicians wait for a triad that mostly does not appear. Among the strongest arguments on this map for a low threshold to treat.
Still needed: The dose, route and duration remain contested; the review says so explicitly.
- Sinha S, Kataria A, Kolla BP, Thusius N, Loukianova LL. Wernicke encephalopathy — clinical pearls.10.1016/j.mayocp.2019.02.018 · PMID 31171116
Mayo Clinic Proceedings, and quietly damning about ordinary practice. The classic triad — confusion, eye signs, gait — is complete in only 10% of cases, and only a few cases are diagnosed BEFORE DEATH. Untreated, about 80% go on to Korsakoff syndrome. Their recommendation is that overdiagnosis and overtreatment are preferable given thiamine's excellent safety profile. Also worth saying on air: alcohol is the commonest cause in the US but hyperemesis gravidarum, bowel obstruction and malignancy all do it too, which is how it gets missed.
Vitamin B12
Metformin lowers its absorption, and the deficiency mimics the neuropathy of diabetes.
- Langan RC, Goodbred AJ. Vitamin B12 deficiency: recognition and management.PMID 28925645
No DOI on this one. Practical and specific about who is actually at risk: gastric or small-bowel resection, IBD, METFORMIN for more than four months, PPIs or H2 blockers for more than twelve, vegans and strict vegetarians, and anyone over 75. Two things worth saying on air: serum B12 alone is unreliable, and methylmalonic acid is what confirms deficiency at low-normal levels; and ORAL high-dose B12 (1–2 mg daily) matches intramuscular for correcting both anaemia and neurological symptoms, which contradicts the received wisdom that injections are necessary.
Vitamin D
The VITAL trial gave 25,871 adults 2000 IU daily; cancer and heart events did not move.
Blueprint
- Open on the category error, and make it concrete rather than semantic. A vitamin is something you must eat. This one you make in skin from cholesterol, under ultraviolet light.
- Then the definitive part, which is about mechanism, not naming: the active form binds a NUCLEAR receptor, pairs with RXR, and binds DNA directly — regulating genes many kilobases from where transcription starts.
- That is precisely how a steroid hormone works. Cortisol, oestrogen and testosterone all do the same thing.
- Follow the activation chain so the listener can hold it: skin, then liver (CYP2R1, making 25-OH-D, the form that gets measured), then kidney, making calcitriol, the active one.
- Show that the system can be dangerous when it breaks — inactivating mutations in CYP24A1, the enzyme that switches the hormone off, probably cause idiopathic infantile hypercalcaemia.
- Close on why the framing matters practically: hormones have feedback loops, receptors and toxicity thresholds. Calling it a vitamin is how it ended up being taken casually in doses nobody would take a hormone in.
Define on air: secosteroid · nuclear receptor · VDR / RXR · calcitriol · 25-hydroxyvitamin D · CYP2R1 · CYP24A1 · hypercalcaemia
Where it lands: Not a semantic argument — the mode of action is a hormone's. The practical consequence is that dosing deserves the caution hormones get.
Still needed: The supplementation-outcome trials belong in this episode as the counterweight: mechanism this strong has repeatedly failed to produce endpoint benefits.
- Christakos S, Dhawan P, Verstuyf A, Verlinden L, Carmeliet G. Vitamin D: metabolism, molecular mechanism of action, and pleiotropic effects.10.1152/physrev.00014.2015 · PMID 26681795
Physiological Reviews, and it settles the 'vitamin D is a hormone' point on the mechanism rather than on rhetoric. 1,25(OH)2D3 is described as the HORMONALLY ACTIVE form; it works by binding a nuclear receptor (VDR) that heterodimerises with RXR and binds DNA directly, regulating genes many kilobases from the transcription start site. That is a steroid hormone's mode of action, not a cofactor's. Also names CYP2R1 as the principal 25-hydroxylase and notes that inactivating CYP24A1 mutations probably cause idiopathic infantile hypercalcaemia — evidence that this system can be dangerous when dysregulated.
Vitamin D as a hormone
Vitamin D is not a vitamin. It is a secosteroid hormone acting through a nuclear receptor — which is why 'just take more' is the wrong frame.
- Christakos S, Dhawan P, Verstuyf A, Verlinden L, Carmeliet G. Vitamin D: metabolism, molecular mechanism of action, and pleiotropic effects.10.1152/physrev.00014.2015 · PMID 26681795
Calcitriol is 1,25(OH)2D3, the end of the activation chain and the molecule that actually does the work through VDR/RXR. Extraskeletal effects reported in knockout and transgenic mice — cancer progression, cardiovascular, immunomodulation in autoimmune disease — with the authors noting only that SOME have also been observed in humans. The gap between mouse mechanism and human outcome is the whole vitamin D supplement debate.
Vitamin K2
Matrix Gla protein needs it for carboxylation; trials show no slower calcification.
Blueprint
- Open with the theory, which is elegant and easy to believe: matrix Gla protein inhibits arterial calcification, and it only works once vitamin K carboxylates it. Therefore K2 should protect arteries.
- The trial: 365 men with established aortic valve calcification, 720 micrograms of MK-7 plus vitamin D or placebo, two years, double-blind, in Circulation.
- The result: calcification progressed 275 arbitrary units on treatment, 292 on placebo. No difference. None in valve area, jet velocity, aortic or coronary calcification, surgery, death or cardiovascular events.
- Now the part that makes this a teaching episode rather than a null result: the biomarker worked perfectly. dp-ucMGP fell sharply on MK-7, exactly as the theory predicted.
- So the mechanism was confirmed and the disease was unmoved. Hold those two facts together and do not let either cancel the other.
- Close on what this should do to how anyone reads a supplement claim built on a plausible pathway plus a moving marker.
Define on air: matrix Gla protein · carboxylation · MK-7 / menaquinone · dp-ucMGP · Agatston / arbitrary units · surrogate vs clinical endpoint
Where it lands: The cleanest demonstration on this map that a correct mechanism and a responsive biomarker can coexist with no clinical benefit at all.
Still needed: Whether earlier intervention, before established calcification, would behave differently. This trial cannot answer that.
- Diederichsen ACP, Lindholt JS, Möller S, et al. Vitamin K2 and D in patients with aortic valve calcification: a randomized double-blinded clinical trial.10.1161/CIRCULATIONAHA.121.057008 · PMID 35465686
A clean NULL in Circulation, and the most instructive result in this batch. 365 men, 720 µg MK-7 plus vitamin D or placebo, two years. Aortic valve calcification progressed by 275 AU on treatment versus 292 on placebo — no difference. No difference in valve area, jet velocity, aortic or coronary calcification, surgery, death or cardiovascular events. AND YET the biomarker worked exactly as the theory predicted: dp-ucMGP fell sharply on MK-7. The carboxylation mechanism is real; the disease did not care. Anyone selling K2 for arterial calcification has to answer this trial.
Lipids
Statins and dementia
Thirty-six cohorts point the other way from the rumour — less dementia on statins, not more. Observational, so read it carefully.
Blueprint
- Open with the rumour, stated fairly: people say statins fog the mind, and some of them are describing something real that happened to them.
- Then the biggest number we have, pointed the other way. Thirty-six cohorts: less dementia on statins (OR 0.80), less Alzheimer's (OR 0.68). No sex difference. No difference between lipophilic and hydrophilic.
- Immediately undercut it. These are observational. Healthy-user bias is not a technicality here — people well enough to be prescribed a preventive drug, and organised enough to keep taking it for a decade, are not the same people as those who are not.
- Now hold both. Case reports and one RCT (simvastatin) do show cognitive impairment. It is rare against the totality of the literature, but rare is not never, and the person it happened to is not a rounding error.
- The actionable part, which is unusually concrete: if you suspect it, a supervised trial off the drug reveals a temporal relationship, and switching from a lipophilic statin to a hydrophilic one (pravastatin, rosuvastatin) may resolve it, because they cross the blood-brain barrier less.
- Close on the asymmetry of risk: the vascular benefit is measured in hard outcomes; the cognitive risk is measured in case reports. That is not a reason to dismiss the case reports. It is a reason not to stop a statin without a plan.
Define on air: cohort study vs randomised trial · odds ratio · healthy-user bias · confounding by indication · lipophilic vs hydrophilic · blood-brain barrier
Where it lands: The evidence does not support the claim that statins cause dementia, and mildly suggests the opposite — but the study design cannot carry that second claim either. What survives is: a rare, reversible, individual effect worth taking seriously in the person in front of you, and no population-level signal of harm.
Still needed: An RCT with cognition as a pre-specified primary endpoint. The authors of the meta-analysis ask for exactly this.
- Olmastroni E, Molari G, De Beni N, et al. Statin use and risk of dementia or Alzheimer's disease: a systematic review and meta-analysis of observational studies.10.1093/eurjpc/zwab208 · PMID 34871380
This is the opposite of what the node was expected to show, and it has to be said that way round. 36 studies: statin users had LOWER dementia risk (OR 0.80, 0.75–0.86) and lower Alzheimer's risk (OR 0.68, 0.56–0.81), with no difference between men and women and none between lipophilic and hydrophilic statins. The authors' own conclusion is 'the absence of a neurocognitive risk'. The caveat is the whole design: these are OBSERVATIONAL, so healthy-user bias and confounding by indication are live — people well enough to be prescribed and to keep taking a preventive drug are not comparable to people who are not. The authors call for purpose-built RCTs. No episode should present this as proof of neuroprotection.
- Rojas-Fernandez CH, Cameron JCF. Is statin-associated cognitive impairment clinically relevant? A narrative review and clinical recommendations.10.1345/aph.1Q620 · PMID 22474137
The other half, and the reason the rumour exists. Reports of statin-associated cognitive impairment come mainly from case reports and series; ONE RCT found simvastatin impaired some measures against placebo, while most RCTs and observational studies found a neutral or beneficial effect. Practical and testable: a trial discontinuation can reveal a temporal relationship, and switching from a lipophilic statin to a hydrophilic one (pravastatin, rosuvastatin, which cross the blood-brain barrier less) may resolve it. Their bottom line is that the effect is rare and does not justify changing practice.
Longevity & NAD
Glycine
Made in the body but not always in enough: shortfalls show up in glutathione, collagen — and in homocysteine.
Blueprint
- Open on the word 'non-essential' and why it is misleading. Glycine is CONDITIONALLY essential — the body makes it, and under ordinary strain it does not make enough.
- List the strains, because they are not exotic: low protein intake, malnutrition, late pregnancy, diabetes, insulin resistance, a heavy load of xenobiotics.
- What runs short when glycine does: glutathione, collagen, nucleotides, one-carbon units. Antioxidant defence, conjugation, neurotransmission.
- The connection this map cares about — glycine deficiency RAISES HOMOCYSTEINE, which puts it on the same page as betaine, folate and B12.
- The twist that stops this being a simple supplement story: glycine and L-serine are interconvertible through SHMT, so giving glycine alone can pull flux the wrong way, losing methylenetetrahydrofolate and generating ammonia. The reviewer argues for co-administration with serine.
- And the honest ending: no study has compared glycine alone against glycine plus serine, and controlled studies in people without deficiency have not been done.
Define on air: conditionally essential · glutathione · one-carbon metabolism · SHMT · methylenetetrahydrofolate · glycine cleavage system
Where it lands: Genuinely undersupplied in common metabolic states, and mechanistically tied to homocysteine and glutathione. The supplement question is more complicated than the deficiency question, and the review says so.
Still needed: Glycine alone vs glycine plus serine, in humans. It does not exist.
- Holeček M. Glycine as a conditionally essential amino acid and its relationship to l-serine.10.1016/j.metabol.2025.156330 · PMID 40527450
Recent review, and it connects three of this map's nodes in one sentence: glycine deficiency raises HOMOCYSTEINE, alongside impaired synthesis of glutathione, collagen, nucleotides and one-carbon units. Deficiency states listed are ordinary ones — low protein intake, malnutrition, late pregnancy, diabetes, insulin resistance, xenobiotic load. The practical argument is the interesting part: because serine hydroxymethyltransferase links the two, the author argues glycine should be co-administered with L-SERINE rather than given alone, to avoid losing methylenetetrahydrofolate to the SHMT flux and to avoid ammonia from the glycine cleavage system. He also states plainly that no study has compared glycine alone against glycine plus serine, and that controlled studies in people WITHOUT deficiency are still needed.
Glycine:methionine ratio
Muscle meat is rich in methionine and poor in glycine; skin, bone and connective tissue carry the balance. The ratio, not the protein, may be what matters.
Blueprint
- Frame it as a question about what we stopped eating, not what we started. Muscle meat is methionine-rich and glycine-poor; skin, bone, tendon and connective tissue carry the glycine.
- So an all-muscle diet loads one side of a balance that whole-animal eating used to hold level.
- Where the two amino acids meet: methionine is spent into homocysteine; getting it back needs methyl donors, and glycine sits in the one-carbon pool that supplies them.
- The known half, from the glycine literature: glycine deficiency raises homocysteine, and occurs in ordinary metabolic states.
- The speculative half, named as speculative: methionine restriction extends lifespan in model organisms, and glycine supplementation has been proposed to mimic part of it. That is animal work.
- Close by refusing the easy conclusion. This is a hypothesis with a real mechanism and no human outcome data, and an hour that pretends otherwise would be the kind of episode this show exists to correct.
Define on air: methionine · one-carbon pool · methyl donor · methionine restriction · model organism
Where it lands: A genuinely interesting hypothesis about dietary composition rather than dietary quantity — currently supported by mechanism and animal data, not by human outcomes.
Still needed: The methionine-restriction lifespan literature needs citing directly, and any human trial that shifted the ratio and measured something. Neither is sourced on this node yet.
- Holeček M. Glycine as a conditionally essential amino acid and its relationship to l-serine.10.1016/j.metabol.2025.156330 · PMID 40527450
Placeholder while the ratio itself is sourced. This review supplies the glycine half — that glycine is only CONDITIONALLY essential, that deficiency is common in ordinary metabolic states, and that one of its consequences is raised homocysteine, which is where methionine metabolism ends up when the recycling routes are short. Still needed before an episode: the methionine-restriction lifespan literature, and human data on whether shifting the dietary ratio changes anything measurable.
Hydrogen water
A gas dissolved in water at millimolar concentrations, with small positive trials and an unresolved question about how something that inert does anything at all.
Blueprint
- The proposed mechanism is selective scavenging of the hydroxyl radical, leaving the signalling species alone — which would be an attractive property if it were established.
- The pharmacological oddity: H2 is close to chemically inert at body temperature, which is the central objection.
- The trials are real but small, short and heavily concentrated in a few research groups.
- The delivery problem is physical, not commercial: an open bottle is degassing while you drink it.
Define on air: W · a · t · e · r · · w · i · t · h · · m · o · l · e · c · u · l · a · r · · h · y · d · r · o · g · e · n · · ( · H · 2 · ) · · d · i · s · s · o · l · v · e · d · · i · n · · i · t · , · · u · s · u · a · l · l · y · · b · y · · e · l · e · c · t · r · o · l · y · s · i · s · · o · r · · a · · t · a · b · l · e · t · . · · C · o · n · c · e · n · t · r · a · t · i · o · n · · i · s · · t · h · e · · w · h · o · l · e · · p · r · o · d · u · c · t · · — · · m · o · s · t · · o · f · · t · h · e · · g · a · s · · i · s · · g · o · n · e · · w · i · t · h · i · n · · m · i · n · u · t · e · s · · o · f · · o · p · e · n · i · n · g · .
Where it lands: Not obviously nonsense, not remotely established. The most honest thing on this map to describe as genuinely unresolved.
Still needed: A large, multi-centre, independently funded trial with a verified dissolved-H2 concentration at the point of drinking and a clinical rather than biochemical endpoint.
- LeBaron TW, Singh RB, Fatima G, Kartikey K, Sharma JP, Ostojic SM, Gvozdjakova A, Kura B, Noda M, Mojto V, Niaz MA, Slezak J. The effects of 24-week, high-concentration hydrogen-rich water on body composition, blood lipid profiles and inflammation biomarkers in men and women with metabolic syndrome: a randomized controlled trial.10.2147/DMSO.S240122 · PMID 32273740
Sixty people, randomised, double-blinded, placebo-controlled, 24 weeks, >5.5 mmol of H2 per day. Reductions in cholesterol, glucose and HbA1c against placebo, with improved inflammation and redox markers. Longer and better-designed than most of this literature — and still sixty people at one site, with several authors affiliated to a molecular hydrogen institute, which is a declared interest rather than a hidden one but belongs in the reading.
Botanicals & extracts
Crocin
The carotenoid that makes saffron red and most of what a standardised extract is standardised to.
- Lopresti AL, Drummond PD. Saffron (Crocus sativus) for depression: a systematic review of clinical studies and examination of underlying antidepressant mechanisms of action.10.1002/hup.2434 · PMID 25384672
Crocin and safranal are the two constituents the antidepressant work keeps pointing at, and the constituents commercial extracts are standardised against. Same structural problem as curcumin: a brightly coloured, poorly absorbed plant pigment with a large in-vitro literature and a much smaller human one.
Lemon balm
It inhibits the enzyme that clears GABA — the same target as an epilepsy drug. Whether that survives a cup of tea is the question.
- Awad R, Muhammad A, Durst T, Trudeau VL, Arnason JT. Bioassay-guided fractionation of lemon balm (Melissa officinalis L.) using an in vitro measure of GABA transaminase activity.10.1002/ptr.2712 · PMID 19165747
A real, named mechanism rather than 'calming': the methanol extract inhibits GABA transaminase, the enzyme that breaks GABA down. Rosmarinic acid is the main active — 40% inhibition at 100 µg/mL, and about 1.5% of the dry leaf. GABA-T is the target of vigabatrin, a licensed anticonvulsant, so this is a herb and a drug aiming at the same enzyme. The gap is everything between an in-vitro rat-brain assay and a human who drank tea.
Thujone
Blamed for a century of absinthe madness. Measured in preserved pre-ban bottles, there was never enough of it to matter.
- Lachenmeier DW. [Thujone-attributable effects of absinthe are only an urban legend — toxicology uncovers alcohol as real cause of absinthism].PMID 18429531
The dose-response argument, stated plainly: thujone is present in both modern and historic pre-ban absinthe at concentrations so low that a pharmacological effect can be excluded outright. Absinthism, as described in 1890s France, is explained by chronic alcohol misuse alone. The author's warning is worth keeping — blaming thujone trivialises the alcohol.
- Lachenmeier DW, Nathan-Maister D, Breaux TA, Kuballa T. Long-term stability of thujone, fenchone, and pinocamphone in vintage preban absinthe.10.1021/jf803975m · PMID 19256492
The obvious objection to measuring century-old bottles is that the thujone might have degraded. It did not: green glass blocked the light, samples reanalysed years apart were unchanged, and only clear-glass storage under UV lost any (18%). The historical measurement stands.
Wormwood
A herb famous for a psychosis it does not cause, with a small steroid-sparing trial in Crohn's that nobody has replicated.
Blueprint
- The reputation: absinthe, hallucinations, Van Gogh's ear.
- The toxicology: thujone concentrations in both historic and modern absinthe are far too low to do anything, and absinthism is alcoholism.
- The actual evidence: one 40-patient double-blind trial showing a steroid-sparing effect in Crohn's, unreplicated for nearly two decades.
- The family resemblance: its cousin gave us the best antimalarial ever found.
Define on air: A · r · t · e · m · i · s · i · a · · a · b · s · i · n · t · h · i · u · m · · — · · t · h · e · · b · i · t · t · e · r · · h · e · r · b · · i · n · · v · e · r · m · o · u · t · h · · a · n · d · · a · b · s · i · n · t · h · e · . · · N · o · t · · t · h · e · · s · a · m · e · · p · l · a · n · t · · a · s · · A · r · t · e · m · i · s · i · a · · a · n · n · u · a · , · · w · h · i · c · h · · i · s · · w · h · e · r · e · · a · r · t · e · m · i · s · i · n · i · n · · c · o · m · e · s · · f · r · o · m · .
Where it lands: Famous for the wrong reason and unstudied for the right one.
Still needed: A properly powered multi-centre replication of the Crohn's trial with objective endpoints — endoscopic healing and calprotectin, not a symptom index.
- Omer B, Krebs S, Omer H, Noor TO. Steroid-sparing effect of wormwood (Artemisia absinthium) in Crohn's disease: a double-blind placebo-controlled study.10.1016/j.phymed.2007.01.001 · PMID 17240130
Forty patients across five German sites, tapered off steroids over ten weeks. 13 of 20 on wormwood reached near-complete symptom remission and held it to week 20; 16 of 20 on placebo needed steroids restarted. That is a striking result in a disease with few steroid-sparing options — and it is forty people, in one trial, from 2007, never replicated. Treat it as the reason to run a bigger trial, not as a reason to take it.
Minerals & iron
Magnesium glycinate
The glycine half is the interesting half — it is the part with a route to homocysteine. The autism claim is weaker than it sounds.
- Nye C, Brice A. Combined vitamin B6-magnesium treatment in autism spectrum disorder.10.1002/14651858.CD003497.pub2 · PMID 16235322
THE AUTISM PREMISE DOES NOT SURVIVE THIS. Cochrane found three randomised studies, 33 people in TOTAL. One gave no usable data. One found no difference from placebo on social interaction, communication, compulsivity, impulsivity or hyperactivity. The third, eight children selected for features resembling pyridoxine-dependent epilepsy, found an IQ gain of 5.2 points (CI 0.2–10.3) — a confidence interval that nearly touches zero, in eight children. Cochrane's verdict: 'no recommendation can be advanced'. Note also this is B6 PLUS magnesium, where magnesium was added to blunt B6's side effects — not magnesium glycinate, and not magnesium alone. The honest episode says the idea has been tested and the tests were too small to conclude anything, which is different from saying it failed. The homocysteine half of this node is the stronger half: glycine deficiency raises homocysteine (see the glycine node), and glycinate is a glycine salt.
Sickle cell disease
The mutation that protects against malaria does not protect against its cousin — and the spleen that would clear it has usually stopped working by age five.
Blueprint
- Teach balanced polymorphism properly, because it is the most elegant idea in human genetics and it is usually taught badly. One letter changes in the haemoglobin gene. One copy of that change makes your red cells a worse home for the malaria parasite, so in malarial regions the gene spreads. Two copies makes haemoglobin that polymerises when oxygen is low, deforming the cell into a crescent that jams capillaries. The same mutation is protection and disease depending on the dose.
- Explain the spleen, since it carries the rest of the story. The spleen filters blood and clears encapsulated organisms and blood parasites. In sickle cell disease repeated infarction destroys it — most children are functionally asplenic by age five. The organ that would clear Babesia is gone before school.
- Then the transfusion burden: chronic transfusion is standard therapy for stroke prevention and acute chest syndrome. So the exposure route is medical.
- Pivot to the Babesia work and let the two facts collide: the parasite is transfusion-transmissible, the patients are transfused, the spleen that would clear it is gone, the red cell it invades is already failing, and the resulting haemolysis is indistinguishable from the transfusion reaction that will be diagnosed instead.
- Close on what a physician should actually do with this — a thick smear and a Babesia PCR before reaching for steroids in a transfused sickle cell patient with unexplained haemolysis in an endemic region.
Define on air: balanced polymorphism · haemoglobin S · polymerisation · vaso-occlusion · functional asplenia · acute chest syndrome · thick smear
Where it lands: The mutation that protects against one parasite leaves its carrier maximally exposed to another — no spleen, a fragile cell, and a treatment that is also a transmission route. It is a genuinely tragic piece of biology and it is also completely actionable at the bedside.
Still needed: No prospective study has screened a transfused sickle cell cohort in an endemic region for Babesia. The denominator is unknown.
- Beri D, Rodriguez M, Singh M, McLaughlin D, Liu Y, Zhong H, Mendelson A, An X, Manwani D, Yazdanbakhsh K, Lobo CA. Babesiosis and sickle red blood cells: loss of deformability, altered osmotic fragility, and hypervesiculation.10.1182/blood.2024027602 · PMID 39869831
Read from the sickle cell side this is a paper about membranes. The SS red cell is already rigid, already fragile, already shedding vesicles; a parasite that remodels the membrane to get in and out lands on a cell with no reserve. Pair it with the transfusion cases: both the route in and the tissue it lands on are worse in sickle cell disease than anywhere else.
- Cursino-Santos JR, Singh M, Senaldi E, Manwani D, Yazdanbakhsh K, Lobo CA. Altered parasite life-cycle processes characterize Babesia infection in human sickle cell anemia.10.3324/haematol.2018.214304 · PMID 30923098
The evolutionary frame for the sickle cell node: haemoglobin S is the textbook case of balanced polymorphism — one copy protects against malaria, two copies cause disease. This paper tests whether that protection generalises to Babesia and finds it does not transfer to the trait. Explain balanced polymorphism plainly when this runs; it is the single most elegant idea in human genetics and it is usually taught badly.
Environment & exposure
Fluoride
The most-swallowed drug never prescribed: added to water at 0.7 mg/L, and the contemporary caries benefit measures about a quarter of a tooth.
Blueprint
- Open on the arithmetic nobody quotes. Cochrane 2024, the most thorough review ever done of community water fluoridation: 157 studies, and in the contemporary evidence — after 1975, once fluoride toothpaste was universal — starting fluoridation changes decay by 0.24 teeth. A quarter of one tooth. Confidence interval -0.03 to 0.52, which includes no benefit at all.
- Then the other side of that ledger from the same review: at 0.7 parts per million, about 12% of people have dental fluorosis of aesthetic concern and about 40% have fluorosis of some degree.
- Establish what fluoride IS before anything else. Fluorine is the most electronegative element on the periodic table — it grabs electrons harder than anything else in nature. That is why it is never found free; it is always bound to something. Fluoride is fluorine that has finished grabbing: it has taken its electron and settled. Explain that this is why it slots into the crystal of enamel so readily, and why it also slots into bone and stays there for decades.
- The mechanism, correctly: fluoride works TOPICALLY. It converts hydroxyapatite in enamel to fluorapatite, which dissolves at a lower pH. That is a surface chemistry event in the mouth. Grandjean's question follows from the mechanism and the dental literature has never answered it well — if the action is on the surface of the tooth, why is the delivery system the bloodstream?
- Cochrane again, the sentence that should end the 'settled science' framing: they found no eligible study reporting caries outcomes in ADULTS. Seventy years, and the outcome literature is children.
- Note that ALL 157 studies were non-randomised and every one was downgraded for risk of bias. This is not a conspiracy; you cannot randomise a water supply. But it means the strongest claim available is association, in both directions.
Define on air: fluoride · electronegativity · hydroxyapatite · fluorapatite · dmft and DMFT · parts per million · dental fluorosis · confidence interval
Where it lands: The benefit is real, small, and smaller than it was before toothpaste existed. The cosmetic harm is measurable in two in five people. That is a defensible trade for some populations and an indefensible one for others, and the honest position is that it is a POLICY question wearing a science costume — because the science, by Cochrane's own grading, is low certainty in both columns.
Still needed: There is no adult outcome data at all, and no randomised evidence of any kind. The cessation studies — what happens when a city stops — are very low certainty and there is essentially one of them.
- Iheozor-Ejiofor Z, Walsh T, Lewis SR, Riley P, Boyers D, Clarkson JE, Worthington HV, Glenny AM, O'Malley L. Water fluoridation for the prevention of dental caries.10.1002/14651858.CD010856.pub3 · PMID 39362658
THE central document and the one almost nobody quotes accurately. 157 studies, ALL non-randomised, every one downgraded for risk of bias. Contemporary evidence (post-1975, i.e. after fluoride toothpaste became universal): starting fluoridation changes dmft by 0.24 teeth (95% CI -0.03 to 0.52) — roughly a QUARTER OF ONE TOOTH, low certainty, and the confidence interval includes no benefit at all. Caries-free children: 4 percentage points (primary), 3 points (permanent), both CIs crossing zero. Cessation: very low certainty, cannot say. No eligible study reports caries outcomes in ADULTS at all. Harms, from the 2015 arm: at 0.7 ppm about 12% have fluorosis of aesthetic concern (95% CI 8-17%, 40 studies, 59,630 people) and about 40% have fluorosis of any level (90 studies, 180,530 people). That is the honest trade — a quarter of a tooth against a two-in-five chance of marking the enamel.
- Yeh CH, Wang YL, Vo TTT, Lee YC, Lee IT. Fluoride in Dental Caries Prevention and Treatment: Mechanisms, Clinical Evidence, and Public Health Perspectives.10.3390/healthcare13172246 · PMID 40941599
A pro-fluoride narrative review, useful precisely because it concedes the shape of the problem: study heterogeneity, protocol variation, fluorosis, and 'possible neurodevelopmental effects' named as persistent uncertainties. Also the cleanest statement that emerging alternatives — nano-hydroxyapatite, bioactive glass, probiotics — are 'promising but currently supported by limited clinical data'.
Fluoride & cognition
Seventy-four studies, a federal court, and one number that keeps moving: where does the dose-response actually start?
Blueprint
- Start with the number, because the number is not in dispute: across 74 studies the pooled effect of higher fluoride exposure on children's IQ is a standardised mean difference of -0.45. Harvard found -0.45 in 2012 with 27 studies. The US National Toxicology Program found -0.45 in 2025 with 74. Two independent teams, thirteen years apart, landed on the same figure.
- Then the number that IS in dispute, and be scrupulous about it. In the NTP meta-analysis, measured by DRINKING WATER alone, the inverse association held below 4 mg/L and below 2 mg/L — and was NULL below 1.5. Measured by URINARY fluoride, it stayed inverse below 1.5. US fluoridation is 0.7. Say all three of those numbers in the same breath, every time.
- The dose-response curve, from Veneri and Vinceti: substantially linear above 1 mg/L, about -3.05 IQ points per 1 mg/L up to 2 mg/L, steeper above. And their own caveat, read out loud: the association was strongest in high-risk-of-bias studies, and NO adverse effect appeared in the only study they judged low risk of bias.
- The shift that made this serious: from ecological village comparisons in China to three prospective cohorts in Mexico and Canada with MEASURED maternal urinary fluoride. Individual exposure, measured before the outcome. That is a different quality of evidence and it pointed the same way.
- The legal history in order. 2016: petition to EPA under section 21 of the Toxic Substances Control Act. 2017: EPA denies it. September 24, 2024: Judge Edward Chen of the Northern District of California finds fluoridation at 0.7 mg/L 'poses an unreasonable risk of reduced IQ in children' and orders EPA to begin rulemaking — while stating explicitly that this 'does not conclude with certainty that fluoridated water is injurious'. January 2025: EPA appeals. May 21, 2026: the Ninth Circuit VACATES and remands, on the procedural ground that the district court relied on the NTP monograph after the parties had agreed not to, not on the science.
- Then the policy consequences that ran ahead of the courts: Utah in March 2025, in force May 7; Florida the same day. Dozens of communities. And the honest counterpoint — as of 2026 no further state has followed.
- Give the opposition its best case, properly. The 2026 JADA review concludes there is no association at fluoridation-relevant levels and 'no established biological mechanism'. The methodological critique of the Chinese cross-sectional literature is largely CORRECT. Then say where it is published.
Define on air: IQ and standardised mean difference · meta-analysis · risk of bias · cross-sectional versus prospective cohort · urinary biomarker · confounding · milligrams per litre
Where it lands: Above 1.5 mg/L the association is strong, consistent across three decades and two continents, and dose-dependent. At 0.7 mg/L the drinking-water evidence does not resolve it and the urinary evidence suggests the curve does not stop. A federal judge found that unresolved state sufficient to require regulatory action; an appeals court vacated that on procedure without touching the science. Anyone telling you this is settled, in either direction, is telling you about themselves.
Still needed: The study that would end it does not exist: a prospective cohort in a fluoridated Western country with individual measured exposure from pregnancy, iodine status measured alongside, and cognitive outcomes to school age.
- Taylor KW, Eftim SE, Sibrizzi CA, Blain RB, Magnuson K, Hartman PA, Rooney AA, Bucher JR. Fluoride Exposure and Children's IQ Scores: A Systematic Review and Meta-Analysis.10.1001/jamapediatrics.2024.5542 · PMID 39761023
The National Toxicology Program's own meta-analysis, and the single most important number in the fluoride debate. 74 studies, 10 of them cohorts. Pooled SMD -0.45 (95% CI -0.57 to -0.33). Individual-level data: IQ falls 1.63 points per 1 mg/L of URINARY fluoride (1.14 points among low-risk-of-bias studies only). THE CRITICAL NUANCE, which both sides misquote: for DRINKING WATER measured alone the inverse association held below 4 and below 2 mg/L but was NULL below 1.5 mg/L. For URINARY fluoride it stayed inverse below 1.5. The authors' own conclusion names 'limited data and uncertainty in the dose-response association ... when fluoride exposure was estimated by drinking water alone at concentrations less than 1.5 mg/L'. US fluoridation is 0.7 mg/L. So the honest sentence is: the effect is real and dose-dependent above 1.5, and at 0.7 the drinking-water evidence does not resolve it while the urinary evidence suggests it does not stop.
- Taylor KW, Eftim SE, Sibrizzi CA, Blain RB, Magnuson K, Hartman PA, Bucher JR, Rooney AA. Addressing Critiques of the Evidence Linking Fluoride and Children's IQ.10.5334/aogh.4853 · PMID 41393310
The NTP authors answering their critics in December 2025. Their claim here is stronger than in the JAMA paper: 'among the high-quality evidence, inverse associations were still observed at fluoride exposure levels below 1.5 mg/L, based on both urinary AND drinking-water measurements'. Quote it as their position, not as settled fact, and put it next to the JAMA abstract, which says the water-only association was null below 1.5. That gap between a team's meta-analysis and the same team's defence of it is itself worth airtime.
- Veneri F, Vinceti M, Generali L, Giannone ME, Mazzoleni E, Birnbaum LS, Consolo U, Filippini T. Fluoride exposure and cognitive neurodevelopment: Systematic review and dose-response meta-analysis.10.1016/j.envres.2023.115239 · PMID 36639015
The best dose-response modelling anyone has done: a cubic-spline one-stage meta-analysis of 30 studies. Highest vs lowest exposure differs by -4.68 IQ points (95% CI -6.45 to -2.92); for drinking water specifically, -5.60. The curve is substantially LINEAR above 1 mg/L — about -3.05 IQ points per 1 mg/L up to 2 mg/L, and steeper above that. Urinary fluoride shows a weaker slope starting around 0.28 mg/L, which they note corresponds to roughly 0.7 mg/L in water. AND THE LINE THAT MUST BE READ ON AIR: 'the inverse association ... was particularly strong in the studies at high risk of bias, while NO ADVERSE EFFECT emerged in the ONLY study judged at low risk of bias.' They could not rule out residual confounding.
- Choi AL, Sun G, Zhang Y, Grandjean P. Developmental fluoride neurotoxicity: a systematic review and meta-analysis.10.1289/ehp.1104912 · PMID 22820538
The Harvard paper that started the modern argument. 27 studies, standardised mean difference -0.45 — the identical figure the NTP would land on thirteen years later with 74 studies. Its own limitation is the one everyone forgot to mention when it went viral: these were high-fluoride endemic areas in China, comparing badly-affected villages with less-affected villages, not fluoridated cities with unfluoridated ones.
- Grandjean P. Developmental fluoride neurotoxicity: an updated review.10.1186/s12940-019-0551-x · PMID 31856837
Grandjean's update, and the paper that moved the argument from ecological studies to individual exposure: three prospective cohorts from Mexico and Canada with measured maternal urinary fluoride, all negative for cognition. His benchmark-dose calculation concludes safe exposures are 'likely to be below currently accepted or recommended fluoride concentrations in drinking water'. Also makes the argument the dental literature has never answered well: topical fluoride works in the mouth, so why swallow it?
- Kumar JV, Levy SM, Warren JJ. An update on community water fluoridation, part 2: Fluoride exposure and children's intelligence (IQ) scores.10.1016/j.adaj.2026.02.010 · PMID 41941356
The 2026 counter-case, in the Journal of the American Dental Association. Conclusion: 'no association between F concentrations relevant to CWF and IQ', the endemic-area studies are methodologically weak, and 'there is no established biological mechanism to explain F's effect on IQ'. Two things to say about it on air, both true: the methodological critique of the Chinese cross-sectional literature is largely correct, and this is the dental profession reviewing the safety of the dental profession's flagship public health measure. Give it its argument and give it its address.
- Kumar JV, Moss ME, Liu H, Fisher-Owens S. Association between low fluoride exposure and children's intelligence: a meta-analysis relevant to community water fluoridation.10.1016/j.puhe.2023.03.011 · PMID 37120936
The other side's primary evidence, and it is a real analysis, not a press release. Restricted to eight studies from NON-endemic areas: SMD 0.07 (95% CI -0.02 to 0.17), I-squared 0%, no significant fluctuation across the fluoride range on spline modelling. Their honest closing sentence matters: 'the reported association observed at higher fluoride levels in endemic areas requires further investigation.' Eight studies is a small evidence base to clear a public health measure with, and they say so.
- Gunasekaran S, Sakthivel S, Rajan RE, Balasubramanian M, Latkar YS, Mathew BM. Fluoride-induced effects on cognitive development in Indian children: A systematic review and meta-analysis.10.4103/jisppd.jisppd_259_25 · PMID 41026552
Eleven studies, 6,000-plus children, September 2025. Above 2.0 ppm the pooled effect is -6.5 IQ points (95% CI -7.3 to -5.7), worst in 6-to-10-year-olds. Useful because it is recent and specific about the exposure level; limited by the same problems as the rest — mostly cross-sectional, substantial heterogeneity, and the authors report some publication bias themselves.
- Tang QQ, Du J, Ma HH, Jiang SJ, Zhou XJ. Fluoride and children's intelligence: a meta-analysis.10.1007/s12011-008-8204-x · PMID 18695947
The 2008 Chinese meta-analysis, sixteen case-control studies, weighted mean difference about -5 IQ points. Mostly of historical interest now, but it establishes that the signal was in the literature a full four years before Harvard found it and sixteen years before a US court did.
Fluoride & iodine
Two halogens, one transporter. The displacement story is popular, mechanistically plausible, and the best direct test of it came back negative.
Blueprint
- Open with the chemistry, because it is genuinely elegant and it is why the hypothesis will not die. Fluorine and iodine are both HALOGENS — column 17 of the periodic table, the elements one electron short of full. Fluorine is the smallest and most aggressive; iodine is the largest and laziest. The thyroid runs on iodine. So the question writes itself: does the pushy halogen get in the way of the quiet one?
- Correct the popular version immediately. Fluoride does not displace iodine from the thyroid hormone molecule — that is not a reaction that happens in the body. The serious hypothesis is about the SODIUM-IODIDE SYMPORTER, the pump in the thyroid cell membrane that drags iodide in from the blood against its concentration gradient. Waugh's proposal is that fluoride suppresses the pump indirectly, by raising inflammatory signals — TNF-alpha, IL-6, IL-1-beta, interferon-gamma — each of which independently turns the pump down.
- Then the direct test, and do not soften it. In 2026, sodium fluoride was run against the symporter itself. Up to 300 micromolar: no change in iodide uptake. No thyroid peroxidase inhibition. No binding at thyroid hormone receptors. Also nothing at oestrogen, androgen, progesterone or any of eight other receptors.
- And then who ran it. All three authors work for Procter & Gamble, which sells fluoride toothpaste. Say the result. Say the address. Then say why neither cancels the other: an in-vitro assay at fixed concentrations for a fixed time is not thirty years of drinking water, and a negative result from an interested party is still the best direct measurement anyone has made.
- The population evidence, with its own correction attached. Peckham 2015: GP practices in the fully fluoridated West Midlands were nearly twice as likely to report high hypothyroidism prevalence as practices in unfluoridated Greater Manchester. Public Health England's reply in the same issue: the study never measured iodine status, which is the actual driver of hypothyroidism, and Birmingham differs from Manchester in a hundred ways.
- Close on the finding that should embarrass everyone. A 2023 scoping review went looking for studies measuring BOTH iodine and fluoride in pregnancy — the only design that could test whether fluoride matters more when iodine is scarce. They found one. One study, after seventy years of fluoridation.
Define on air: halogen · electronegativity · sodium-iodide symporter · iodide versus iodine · thyroid peroxidase · micromolar · in vitro versus in vivo · ecological study
Where it lands: The mechanism is chemically plausible and the direct experimental test of it is negative. The population signal exists and cannot be separated from everything else that differs between the places compared. The interesting question is not whether fluoride displaces iodine — it does not, in the literal sense — but whether fluoride exposure matters more in someone whose iodine is already low. That question has been asked once.
Still needed: One study has ever measured both. Until a pregnancy cohort measures iodine status and fluoride exposure together, the honest answer is that the most important version of this question is unasked, not unanswered.
- Waugh DT. Fluoride Exposure Induces Inhibition of Sodium/Iodide Symporter (NIS) Contributing to Impaired Iodine Absorption and Iodine Deficiency.10.3390/ijerph16061086 · PMID 30917615
The fullest statement of the displacement hypothesis, and the source most of the internet is downstream of. Proposed route: fluoride does not swap for iodine atom-for-atom, it suppresses the SODIUM-IODIDE SYMPORTER — the pump that drags iodide into the thyroid — partly by driving up TNF-alpha, IL-6, IL-1-beta, IFN-gamma and TGF-beta-1, all of which independently downregulate NIS, and partly via prolactin and megalin. IMPORTANT FOR HONESTY: this is a single-author narrative review by an environmental consultant, not primary data, and it hypothesises the mechanism rather than measuring it. It is the right place to get the hypothesis and the wrong place to stop.
- Mudd AM, Ovando BJ, Daston GP. Probing the Biological Plausibility of Fluoride as an Endocrine Disruptor.10.1002/bdr2.70046 · PMID 42138003
The direct experimental test of the displacement hypothesis, and it came back NEGATIVE. Sodium fluoride up to 300 micromolar produced NO change in iodide uptake through the sodium-iodide symporter; up to 10 micromolar it did not bind thyroid hormone receptors, did not inhibit thyroid peroxidase, and did not touch oestrogen, androgen, progesterone, PPAR, AhR, CAR, PXR, RAR or glucocorticoid receptors; up to 316 micromolar it did not alter oestrogen or testosterone synthesis. TWO THINGS MUST BE SAID TOGETHER. First: all three authors work for Procter & Gamble, which sells fluoride toothpaste, and that belongs on air in the same breath as the result. Second: the result is still the best direct measurement anyone has of this mechanism, and an in-vitro assay at fixed concentrations is not the same experiment as thirty years of drinking the stuff. Report the finding, report the funder, and do not pretend either one cancels the other.
- Patial B, Thakur R. Exploring the Impacts of Fluoride-Induced Thyroid Toxicity: A Comprehensive Review.10.1007/s12011-025-04840-6 · PMID 41073679
Reviews everything published on fluoride and the thyroid from 1976 to 2025. Reported effects: interference with iodine metabolism, lower T3 and T4, higher TSH, oxidative stress, downregulated antioxidant and thyroid genes, and histology showing follicular-cell apoptosis, necrosis and hyperplasia. Their framing — 'the thyroid gland in particular is highly susceptible to fluoride accumulation' — is worth quoting, but note that the weight of what they review is ANIMAL work at exposures well above 0.7 mg/L.
- Peckham S, Lowery D, Spencer S. Are fluoride levels in drinking water associated with hypothyroidism prevalence in England? A large observational study of GP practice data and fluoride levels in drinking water.10.1136/jech-2014-204971 · PMID 25714098
The population study everyone cites: GP practices in the fully fluoridated West Midlands were nearly twice as likely to report high hypothyroidism prevalence as practices in unfluoridated Greater Manchester. It is ECOLOGICAL — practice-level, not person-level — and adjusted only for deprivation, so it cannot separate fluoride from everything else that differs between Birmingham and Manchester. Always pair it with the reply below; running it alone would be exactly the kind of one-sided citation this show exists to refuse.
- Newton JN, Young N, Verne J, Morris J. Water fluoridation and hypothyroidism: results of this study need much more cautious interpretation.10.1136/jech-2015-205917 · PMID 26068199
Public Health England's response, published in the same issue. The core objection is a good one: iodine status, the actual driver of hypothyroidism, was not measured at all, and the two comparison regions differ in far more than their water. Worth reading alongside the 2017 follow-up exchange (PMID 28093451) in which neither side moves.
- Griebel-Thompson AK, Sands S, Chollet-Hinton L, Christifano D, Sullivan DK, Hull H, Carlson SE. A Scoping Review of Iodine and Fluoride in Pregnancy in Relation to Maternal Thyroid Function and Offspring Neurodevelopment.10.1016/j.advnut.2023.01.003 · PMID 36796438
The most useful single sentence in this whole literature. These authors went looking for studies that measured BOTH iodine and fluoride in pregnancy — the only design that could actually test whether fluoride matters more in iodine-deficient women — and found exactly ONE. After seventy years of fluoridation and thirty years of iodine-deficiency research, one study has asked the obvious question. Their conclusion is four words long: more studies are needed.
Nano-hydroxyapatite
The fluoride-free alternative is the mineral your enamel is already made of — which is both the case for it and the reason to ask what a nanoparticle of it does elsewhere.
Blueprint
- Start with what it is, because the name does all the frightening and none of the explaining. Hydroxyapatite is the mineral your enamel and bones are already made of — calcium, phosphate and a hydroxide group in a repeating crystal. 'Nano' just means the particles are small enough to fit into the microscopic pores acid has opened in enamel. You are putting tooth back on tooth.
- Its origin story is worth the airtime: NASA developed it for astronauts, whose teeth and bones demineralise in microgravity. Japan approved it as an anticaries agent in 1993. It has been mainstream in Japanese toothpaste for thirty years.
- The safety question, answered properly. The definitive review finds no inherent toxicity; cells that take the particles in digest them and clear the calcium within hours. Cytotoxicity in culture is mostly an artefact of particles clumping and settling onto the cell layer. And the sentence that settles the toothpaste question: there is NO RISK from oral uptake, because the particles DISSOLVE IN STOMACH ACID. A nanoparticle that does not survive your stomach cannot be a nanoparticle anywhere downstream.
- So where did the scare come from? Two things in 2016 — an EU scientific committee statement about NON-SPHERICAL nano-hydroxyapatite in cosmetics, and a US argument about nano-calcium-phosphate in baby food. Shape and route, not substance. The one real exposure concern the review names is INHALATION, which is a question about aerosols and factories, not about brushing.
- The efficacy evidence, honestly graded. The largest trial — 610 children, 24 months, triple-blind — favoured hydroxyapatite-plus-fluoride over conventional fluoride paste on enamel lesions. But BOTH ARMS CONTAINED FLUORIDE. It tested hydroxyapatite as an addition, not as a replacement, which is not the question most people are asking.
- Close on the honest state of play: a 2025 review calls it 'promising but currently supported by limited clinical data'. That is the correct verdict and it is not a criticism.
Define on air: hydroxyapatite · nanoparticle · remineralisation · demineralisation · triple-blind · in situ study
Where it lands: The harms are the weakest part of the case against it: the particles dissolve in stomach acid, the dermal and mucosal risk is very low, and the 2016 EU concern was about a different particle shape in a different product. Efficacy is where the real gap is — the evidence is in-situ enamel models, sensitivity outcomes, and one large trial in which both arms also contained fluoride. Nobody has run the trial that matters: hydroxyapatite alone against fluoride alone, caries as the endpoint, in a population.
Still needed: A head-to-head, fluoride-free-arm, caries-outcome randomised trial. It has not been done, and until it is, choosing hydroxyapatite over fluoride is a reasonable preference and not an evidence-based decision.
- Epple M. Review of potential health risks associated with nanoscopic calcium phosphate.10.1016/j.actbio.2018.07.036 · PMID 30031162
The safety review the whole nano-hydroxyapatite question turns on, and it does not say what either camp wants. Findings: no inherent toxicity; particles taken into the cell are digested in the lysosome and raise intracellular calcium, which cells clear within hours unless the dose is very high; observed cytotoxicity in culture is mostly agglomeration and sedimentation producing a huge LOCAL dose. Crucially for toothpaste: 'there is NO RISK from an oral uptake of calcium phosphate nanoparticles due to their RAPID DISSOLUTION IN THE STOMACH' — stomach acid takes the nanoparticle apart before it can be a nanoparticle. Dermal and mucosal risk very low. The one real exposure route is INHALATION, which is an aerosol and occupational question, not a brushing question. He also names the origin of the scare: a 2016 EU SCCS statement about non-spherical nano-hydroxyapatite in cosmetics, and a 2016 US argument about nano-calcium-phosphate in baby food.
- Cocco F, Salerno C, Wierichs RJ, Wolf TG, Arghittu A, Cagetti MG, Campus G. Hydroxyapatite-Fluoride Toothpastes on Caries Activity: A Triple-Blind Randomized Clinical Trial.10.1016/j.identj.2024.09.037 · PMID 39971658
The largest clinical trial in this space: 610 children aged 4-7, 518 completing, 24 months, triple-blind. Hydroxyapatite-plus-fluoride pastes beat conventional monofluorophosphate pastes on enamel lesions (p<0.01); of 78 active lesions at baseline in the hydroxyapatite arm, 58 were inactive at two years. NOTE WHAT IT IS AND IS NOT: both arms contained fluoride. This trial tests hydroxyapatite as an ADDITION, not as a replacement — which is the question most people asking about it actually want answered.
- Amaechi BT, Alshareif DO, Azees PAA, Shehata MA, Lima PP, Abdollahi A, Kalkhorani PS, Evans V, Bagheri A, Okoye LO. Anti-caries evaluation of a nano-hydroxyapatite dental lotion for use after toothbrushing: An in situ study.10.1016/j.jdent.2021.103863 · PMID 34743963
Randomised, double-blind, crossover, 30 adults wearing intra-oral appliances with real human enamel blocks. 5% nano-hydroxyapatite toothpaste remineralised early lesions and completely prevented demineralisation of sound enamel; adding a 5% lotion afterwards raised remineralisation from 37.7% to 58.4%. Real mechanism, real measurement — and an in-situ appliance study on 64 enamel blocks, which is a long way from a caries outcome in a population.
- Butera A, Gallo S, Pascadopoli M, Montasser MA, Abd El Latief MH, Modica GG, Scribante A. Home Oral Care with Biomimetic Hydroxyapatite vs. Conventional Fluoridated Toothpaste for the Remineralization and Desensitizing of White Spot Lesions.10.3390/ijerph19148676 · PMID 35886524
40 patients, three months, hydroxyapatite against 1450 ppm fluoride. Hydroxyapatite beat fluoride on sensitivity and pain; erosive wear scores did not move in either arm. Forty patients and one outcome that is not caries — this is a signal, not a verdict, and should be labelled that way on air.
Cancer, found and missed
Amygdalin (Laetrile)
178 patients, no benefit on any endpoint, and blood cyanide approaching the lethal range. Cochrane calls the risk-benefit balance unambiguously negative.
Blueprint
- The mechanism claimed: cancer cells contain more beta-glucosidase, so they liberate cyanide selectively.
- The mechanism observed: gut bacteria liberate the cyanide, which is why oral dosing is more dangerous than intravenous.
- The trial: 178 patients, 1982, no benefit, measurable cyanide toxicity.
- The persistence: still sold, still as 'B17', four decades after it was tested.
Define on air: A · · c · y · a · n · o · g · e · n · i · c · · g · l · y · c · o · s · i · d · e · · f · r · o · m · · t · h · e · · k · e · r · n · e · l · s · · o · f · · a · p · r · i · c · o · t · s · · a · n · d · · o · t · h · e · r · · P · r · u · n · u · s · · s · p · e · c · i · e · s · . · · ' · L · a · e · t · r · i · l · e · ' · · i · s · · a · · s · e · m · i · - · s · y · n · t · h · e · t · i · c · · r · e · l · a · t · i · v · e · , · · s · o · l · d · · s · i · n · c · e · · t · h · e · · 1 · 9 · 5 · 0 · s · · a · n · d · · m · a · r · k · e · t · e · d · · a · s · · ' · v · i · t · a · m · i · n · · B · 1 · 7 · ' · , · · w · h · i · c · h · · i · t · · i · s · · n · o · t · · — · · i · t · · i · s · · n · o · t · · a · · v · i · t · a · m · i · n · · a · n · d · · t · h · e · r · e · · i · s · · n · o · · d · e · f · i · c · i · e · n · c · y · · s · t · a · t · e · .
Where it lands: One of the very few things on this map where the evidence is not thin or mixed. It was tested, it failed, and it poisons people.
Still needed: None on efficacy. The open question is regulatory: why a compound with a negative NEJM trial and a negative Cochrane review is still purchasable.
- Moertel CG, Fleming TR, Rubin J, Kvols LK, Sarna G, Koch R, Currie VE, Young CW, Jones SE, Davignon JP. A clinical trial of amygdalin (Laetrile) in the treatment of human cancer.10.1056/NEJM198201283060403 · PMID 7033783
The trial the advocates demanded and then rejected. 178 patients, most in good general condition, a third chemotherapy-naive, given amygdalin at doses and schedules representative of actual Laetrile practice plus the full 'metabolic therapy' diet-and-enzyme programme. No benefit in cure, improvement, stabilisation, symptoms or survival. Several patients showed cyanide toxicity or blood cyanide approaching lethal. The conclusion is one sentence: a toxic drug that is not effective as a cancer treatment.
- Milazzo S, Horneber M. Laetrile treatment for cancer.10.1002/14651858.CD005476.pub4 · PMID 25918920
Eight databases, two registers, over 200 references screened across three updates — and not one study met the inclusion criteria. Zero randomised trials. The review's own verdict: the risk-benefit balance is 'unambiguously negative'. Worth noting how rare that phrasing is in a Cochrane conclusion.
Apricot seeds
Sold as a cancer cure. The active ingredient is metabolised to cyanide, and enough of it will make a pulse oximeter read a healthy man as hypoxic.
- Konstantatos A, Shiv Kumar M, Burrell A, Smith J. An unusual presentation of chronic cyanide toxicity from self-prescribed apricot kernel extract.10.1136/bcr-2017-220814 · PMID 28893740
A 67-year-old man looked hypoxic under routine anaesthesia. Cardiac and respiratory causes excluded, haemoglobinopathies excluded. He had been taking apricot kernel extract daily for FIVE YEARS; blood analysis confirmed cyanide toxicity, and the free cyanide was interfering with the pulse oximeter itself. Saturations normalised when he stopped. A chronic toxicity from a complementary medicine that presented as a monitoring artefact — which is exactly how these are missed.
Formalin fixation
Formaldehyde stops biology mid-sentence by cross-linking protein. It preserves shape beautifully and shreds DNA doing it.
Blueprint
- Open on the trade, because it is the cleanest example in medicine of buying one thing with another. Formaldehyde cross-links proteins. That is what preserves the architecture — and it is also what shreds the DNA.
- Quote the forensic review directly: formalin 'is known to damage DNA through crosslinking activity', producing 'severely fragmented DNA of variable yields', which 'reduces the ability to perform downstream molecular analyses'.
- So the step that makes tissue readable by eye is the same step that degrades the molecules you would need to identify an organism nobody suspected. That is not a conspiracy; it is a hundred-and-thirty-year-old engineering decision nobody has revisited.
- Tell the history, because it explains why. F. Blum introduced formalin in 1893. It won because it was cheaper than the ethyl alcohol Dutch scientists had used since the eighteenth century. Alcohol-based coagulant fixatives do NOT cross-link — they do not compromise immunostaining and they do not damage DNA or RNA. One Leiden laboratory has processed over 40,000 skin biopsies and 100,000 cervical samples that way.
- Then the prediction made in 2008 and still unfulfilled: formalin will be legislated out of diagnostic pathology once health authorities realise 'that formalin invalidates expensive tests'.
- And the counterweight, which must be given equal weight. Modern methods recover usable sequence from FFPE anyway — roughly 7,000 genes from a 100-micrometre punch of six-year-old lung cancer blocks. The archive is damaged. It is not unreadable.
- Close on what that means: every hospital on earth is sitting on decades of paraffin blocks at room temperature, and the technique to interrogate them for non-human sequence exists and is routine. It has not been pointed at this.
Define on air: formaldehyde · cross-linking · coagulant fixative · FFPE · fragmentation · immunostaining
Where it lands: Formalin degrades nucleic acid — that is established and quantified. It does not erase morphology, which is what a parasite would be recognised by. The molecular barrier is real; the visual one is not.
Still needed: No study has asked how formalin fixation specifically affects the recoverability of PARASITE DNA from human tumour tissue. The question has not been posed.
- Reid KM, Maistry S, Ramesar R, Heathfield LJ. A review of the optimisation of the use of formalin fixed paraffin embedded tissue for molecular analysis in a forensic post-mortem setting.10.1016/j.forsciint.2017.09.020 · PMID 29078160
The cost of fixation, stated without euphemism: 'The preservation of tissue in formalin is KNOWN TO DAMAGE DNA through crosslinking activity. This results in the extraction of SEVERELY FRAGMENTED DNA of variable yields, which subsequently REDUCES THE ABILITY TO PERFORM DOWNSTREAM MOLECULAR ANALYSES.' 111 studies reviewed on how to work around it. This is the single most important fact for the parasite question: the step that makes the tissue readable by eye is the same step that degrades the molecules a PCR would need to name an organism nobody suspected.
- Matsunaga H, Arikawa K, Yamazaki M, Wagatsuma R, Ide K, Samuel AZ, Takamochi K, Suzuki K, Hayashi T, Hosokawa M, Kambara H, Takeyama H. Reproducible and sensitive micro-tissue RNA sequencing from formalin-fixed paraffin-embedded tissues for spatial gene expression analysis.10.1038/s41598-022-23651-6 · PMID 36376423
The counterweight, and the reason the episode should not end in despair. Yes, FFPE tissue 'may be severely damaged by methylene crosslinking'. But this group recovered roughly 7,000 genes from a 100-micrometre punch, 10 micrometres thick, from mouse liver fixed TWO YEARS earlier, and characterised the tumour microenvironment in human lung cancer blocks fixed SIX YEARS earlier. Every hospital in the developed world holds decades of these blocks at room temperature. The archive is readable. Nobody has read it for this.
- Boon ME, Kok LP. Theory and practice of combining coagulant fixation and microwave histoprocessing.10.1080/10520290802553476 · PMID 19031284
The history, and the reason it is formalin at all. F. Blum introduced formaldehyde as a fixative in 1893; it became standard because it hardened and preserved gross specimens, and because it was CHEAPER than the ethyl alcohol Dutch scientists had used since the eighteenth century. Alcohol-based coagulant fixatives do not cross-link, so they do not compromise immunostaining and do not damage DNA or RNA. Over 40,000 skin biopsies and 100,000 cervical samples have been processed this way in one Leiden laboratory. The authors' prediction, made in 2008 and still not realised: formalin will eventually be legislated out once health authorities 'realize that formalin invalidates expensive tests'.
H&E
Two dyes, pink and blue, unchanged since the 1870s. Almost every cancer diagnosis on earth is made on them.
Blueprint
- Two dyes. Haematoxylin, from the heartwood of a Central American logwood tree, binds acidic things — DNA and RNA — and goes blue-purple. Eosin, a synthetic dye named for the Greek goddess of dawn, binds basic things — most proteins — and goes pink.
- That is it. That is the test almost every cancer diagnosis on earth is made on, essentially unchanged since the 1870s.
- And it works — that is the part that has to be said before any criticism. Pattern recognition on H&E has been refined over a century and a half and it is astonishingly good at answering the question it asks.
- Now the question it asks, from a paper defining adequacy criteria for a slide: cellular detail, tissue arrangement, tissue integrity, stain uptake, and visual distinction of tissue STRUCTURE. Every criterion is about human tissue architecture. None of them asks whether something non-human is present.
- Then the geometry. The section is four micrometres thick. A schistosome egg is roughly 150 micrometres long; a tapeworm proglottid is millimetres. An organism crossing a four-micrometre plane appears as a slice through something — a fragment, a ring, a smear of unfamiliar texture. Recognising it requires having seen it before.
- Which is where training comes in, and it is the real barrier. A pathologist in Chicago may complete an entire residency without seeing a schistosome egg in tissue. Not because it was destroyed. Because it was never in the teaching set.
- Close on the contrast that makes the point: in parasitology, IODINE MOUNT microscopy is a standard method for FINDING parasites, listed alongside direct wet mount. In the laboratory that goes looking, the finding rate is over 90%. The difference between those two laboratories is not the reagents. It is the question.
Define on air: haematoxylin · eosin · basophilic and eosinophilic · micrometre · proglottid · wet mount
Where it lands: H&E is not blind to parasites; pathologists diagnose them on it routinely when they are looking. It is blind to parasites in tumours for the same reason a radiologist reading a chest film for pneumonia can miss a rib fracture — the eye finds what the mind has queued up.
Still needed: Nobody has measured how often a trained parasitologist, re-reading archived tumour H&E slides blind, finds organisms the original pathologist did not. That is a cheap study and it has not been done.
- Haghbin N, Oveisi B, Banitaba AP. Automated variable power cold microwave tissue processing: A novel universal tissue processing protocol without using formaldehyde and xylene.10.1016/j.acthis.2022.151880 · PMID 35344896
Used here for their MAS criteria — morphology, artifacts, staining — which is the honest list of what a pathologist is actually judging when they say a slide is adequate: cellular detail, tissue arrangement, tissue integrity, stain uptake, and visual distinction under LIGHT microscopy. Every one of those criteria is about human tissue architecture. None of them asks whether something non-human is present.
- Mewara A, Khurana S, Gupta S, Munda VS, Singh S, Sehgal R. Diagnostic performance of mini parasep solvent-free faecal parasite concentrator for the diagnosis of intestinal parasitic infections.10.4103/ijmm.IJMM_19_44 · PMID 32003337
The fact that breaks the simple version of the iodine story. In parasitology, IODINE MOUNT MICROSCOPY is a routine method for FINDING parasites — it is listed here alongside direct wet mount as standard practice, and the concentration methods improved the yield of Hymenolepis nana, Trichuris trichiura, Entamoeba coli and Giardia lamblia. Iodine kills protozoa and stains their nuclei so a human can identify them. In the laboratory that goes looking for parasites, iodine is the tool that reveals them. Whatever iodine does in a cancer specimen, 'makes the organism invisible' is not it.
How cancer is tested
Six steps between the tumour and the diagnosis, and every one of them was designed to answer a different question than the one a parasite would raise.
Blueprint
- Walk the pipeline in order, slowly, because almost nobody outside medicine has been told it and it is the spine of the whole episode. Step one: something is taken out of a person — a needle core, an endoscopic pinch, a whole organ.
- Step two, FIXATION. The tissue goes into 10% neutral buffered formalin, usually within minutes, and sits for hours. Formaldehyde cross-links protein, locking every molecule to its neighbour, which freezes the architecture exactly as it was. Without this the tissue digests itself within hours.
- Step three, PROCESSING. Water is pulled out through graded alcohols, the alcohol is replaced by a clearing agent — usually xylene — and the xylene is replaced by molten paraffin wax that infiltrates every space water used to occupy. Hours, usually overnight, usually automated.
- Step four, EMBEDDING and SECTIONING. The wax block is chilled and shaved on a microtome into ribbons FOUR MICROMETRES thick. A human cell is ten to thirty micrometres across, so a section is a fraction of one cell's depth.
- Step five, STAINING. The wax comes off, the water goes back in, and the section is dipped in haematoxylin and eosin — blue for nucleic acid, pink for protein. Two dyes, unchanged in principle since the 1870s. Almost every cancer diagnosis on earth is made on this.
- Step six, READING. A human being looks down a microscope and compares what they see with a catalogue in their head.
- Then, only if the H&E raises a specific question: immunohistochemistry (antibodies against named human proteins) and molecular testing (PCR or sequencing for named human genes).
- Now make the point the episode exists for, and make it precisely. Every one of those steps is superbly designed to answer ONE question: what kind of human cell is this and is it behaving badly? Steps 7 and 8 are literally targeted — antibodies raised against human antigens, primers designed for human sequence. A parasite is not a negative result on these tests. It is not a question they were asked.
Define on air: biopsy versus resection · formalin · fixation · cross-linking · clearing agent · microtome · micrometre · haematoxylin and eosin · immunohistochemistry · antibody · PCR primer
Where it lands: The pipeline is not broken and it is not hiding anything. It is exquisitely optimised for the question it was built to answer, and a parasite falls outside that question at every single step — not because anything destroys the evidence, but because nothing in the workflow ever asks.
Still needed: There is no routine step at which a specimen is examined for non-human organisms unless a clinician specifically requests it, and clinicians request it when a parasite is already on the differential.
- Haghbin N, Oveisi B, Banitaba AP. Automated variable power cold microwave tissue processing: A novel universal tissue processing protocol without using formaldehyde and xylene.10.1016/j.acthis.2022.151880 · PMID 35344896
The cleanest plain statement of the pipeline every cancer diagnosis passes through: fixation, dehydration, clearing, paraffin infiltration, embedding, microtome slicing, staining, slide reading. Use their framing for the episode's spine, and note the two reagents they built the whole machine to avoid — formaldehyde and xylene — 'cause severe health and safety issues for humans and the environment'. The conventional protocol takes many hours; theirs takes 97 minutes. Tissue up to 4 mm thick.
- Roh CK, Jung MJ. Laparoscopic excision for ectopic peritoneal paragonimiasis mimicking a gastric duplication cyst: A case report.10.1016/j.amsu.2021.102754 · PMID 34484726
A 47-year-old man, no symptoms, an incidental peritoneal mass on a screening CT. Radiology called it a gastric duplication cyst. It was a lung fluke that had migrated out of the chest. The line to quote is the authors' own: 'the detection of the ova of Paragonimus parasites in sputum and BIOPSY SPECIMENS may be difficult due to an INSUFFICIENT AMOUNT.' Not because iodine destroyed it. Because there was not enough of it in the piece that was taken.
- Al Laham O, Abdul Khalek G, Alboushi H, Abazid E, Darwish A, Hamza A. An incidentally diagnosed primary pancreatic body hydatid cyst: A case report and literature review.10.1016/j.ijscr.2024.109392 · PMID 38367420
The mirror image of the episode's whole question. A 46-year-old woman with epigastric pain and a 'pancreatic body mass' on imaging went to theatre for what everybody expected to be pancreatic cancer — and lost her pancreatic body, tail and spleen. Histopathology found a hydatid cyst: a tapeworm larva. The parasite was only found BECAUSE the specimen was examined. That is the case for the pipeline, not against it, and an honest episode has to carry it.
Iodine in the specimen
Iodine reaches tissue three ways before a pathologist ever sees it, and it kills organisms on contact. Whether that is why parasites go unreported is a different question.
Blueprint
- State the hypothesis first, in its strongest form, because it deserves that: iodine touches cancer specimens, iodine kills parasites, therefore iodine could be erasing the evidence.
- Establish the first half. Iodine reaches tissue on three documented routes before a pathologist ever sees it. Povidone-iodine surgical skin preparation and cavity lavage, at 7.5 to 10%. Lugol's iodine chromoendoscopy, sprayed directly onto oesophageal lining to reveal dysplasia before the biopsy is taken. And Lugol's iodine in the histology laboratory itself, historically used to strip mercury pigment from tissue fixed in mercury-containing fixatives.
- Establish the second half, and it is stronger than most people expect. Of eight anti-amoebic drugs tested against Acanthamoeba cysts, povidone iodine at 1% was the ONLY one that completely suppressed trophozoite formation across every isolate and every replicate. Chlorhexidine and miltefosine mostly failed. Surgical prep is roughly ten times that concentration.
- Then the finding that makes this a real question rather than a speculation — and read it verbatim. From an ophthalmology journal in 2019: 'Ophthalmologists should be aware that certain topical anesthetics and ophthalmic preparations containing BAC prior to specimen sampling may affect the viability of Acanthamoeba spp. in vivo, resulting in FALSE-NEGATIVE RESULTS IN DIAGNOSTIC TESTS.' A specialty looked at exactly this mechanism and found it, in print, seven years ago.
- NOW THE TURN, and it is the honest heart of the whole node. Read what test they meant. A VIABILITY test — culture. A dead organism does not grow. But a dead organism is still THERE, and histopathology does not culture anything. Formalin kills everything in the specimen within minutes and preserves it perfectly; that is the entire purpose of formalin. A dead fluke in a paraffin section still looks exactly like a fluke.
- So the accurate sentence is narrower and more interesting than the popular one: iodine would defeat a CULTURE and would not defeat a MICROSCOPE. Where it could genuinely matter is nucleic acid — iodine is a strong oxidiser, and nobody has measured what surgical-concentration povidone-iodine does to parasite DNA recoverable from a specimen. That study does not exist.
- Finish by naming the real obstacles, in order of size: nobody orders the test; the section is four micrometres thick; the needle sampled a thousandth of the tumour; the antibodies are raised against human antigens; the primers are designed for human sequence; and formalin has fragmented the DNA. Iodine is on that list. It is not at the top of it.
Define on air: povidone-iodine · Lugol's iodine · chromoendoscopy · cysticidal · trophozoite · viability testing versus morphology · oxidiser · benzalkonium chloride
Where it lands: The mechanism is real, documented and in the literature — for culture-based diagnosis, in ophthalmology. Extending it to histopathology requires a step nobody has taken, and the step points the wrong way, because fixation kills everything anyway and morphology survives death. Iodine is a genuine contributor to under-detection of LIVE organisms and a poor explanation for under-detection on a slide.
Still needed: The measurable question nobody has measured: does povidone-iodine at surgical concentration degrade parasite DNA to the point of defeating a broad eukaryotic PCR on the specimen? That is one bench experiment and it would settle it.
- Heaselgrave W, Hamad A, Coles S, Hau S. In Vitro Evaluation of the Inhibitory Effect of Topical Ophthalmic Agents on Acanthamoeba Viability.10.1167/tvst.8.5.17 · PMID 31588380
THE central citation for this node, and it says the thing out loud. Testing topical agents against Acanthamoeba cysts and trophozoites, povidone iodine's antiamoebic effect was 'superior to the current diamidines' — and then the clinical warning, verbatim: 'Ophthalmologists should be aware that certain topical anesthetics and ophthalmic preparations containing BAC PRIOR TO SPECIMEN SAMPLING may affect the viability of Acanthamoeba spp. IN VIVO, resulting in FALSE-NEGATIVE RESULTS IN DIAGNOSTIC TESTS.' So the mechanism is documented, in print, in a specialty that looked. NOW THE PART THAT MUST BE SAID IN THE SAME BREATH: the false negative here is in a VIABILITY test — culture. A dead organism does not grow. It is still visible under a microscope. Applying this to histopathology requires an extra step nobody has taken, which is exactly why it is worth naming as an open question rather than a finding.
- Muthukumar V, Shi L, Chai N, Langenbucher A, Becker SL, Seitz B, Orosz E, Stachon T, Kiderlen AF, Bischoff M, Szentmáry N. Efficacy of Off-Label Anti-Amoebic Agents to Suppress Trophozoite Formation of Acanthamoeba spp. on Non-Nutrient Agar Escherichia coli Plates.10.3390/microorganisms10081642 · PMID 36014060
How good an antiparasitic iodine actually is. Eight off-label anti-amoebic drugs tested against four Acanthamoeba isolates in five biological replicates. Only ONE — povidone iodine at 1% — completely suppressed trophozoite formation from drug-challenged cysts, in every isolate, every replicate. Chlorhexidine, miltefosine and dibromopropamidine mostly failed. Iodine is not a mild antiseptic that happens to be near parasites; on this evidence it is the most reliably cysticidal agent on the shelf. Surgical skin prep is 7.5-10% povidone iodine — roughly ten times the concentration tested here.
- Padzik M, Baltaza W, Conn DB, Szaflik JP, Chomicz L. Effect of povidone iodine, chlorhexidine digluconate and toyocamycin on amphizoic amoebic strains, infectious agents of Acanthamoeba keratitis.10.26444/aaem/99683 · PMID 30586959
Corroborates the effect against both a clinical corneal isolate and a laboratory strain, with povidone iodine 'promising' against reference chlorhexidine. Note the co-author: David Bruce Conn of Harvard's Museum of Comparative Zoology, the same parasitologist who wrote the Hymenolepis correspondence in the New England Journal. A small field.
- Mewara A, Khurana S, Gupta S, Munda VS, Singh S, Sehgal R. Diagnostic performance of mini parasep solvent-free faecal parasite concentrator for the diagnosis of intestinal parasitic infections.10.4103/ijmm.IJMM_19_44 · PMID 32003337
The necessary correction to the iodine story, kept on this node so it cannot be quoted without it. Iodine mount is a ROUTINE parasitology stain. Killing an organism and hiding an organism are different operations, and iodine performs the first while assisting the second's opposite.
Sampling error
A core needle takes about a thousandth of a tumour and the slide shows four microns of that. The arithmetic of what is never looked at.
Blueprint
- Do the arithmetic out loud, because the arithmetic is the argument. A core needle biopsy takes a cylinder roughly one millimetre across and a centimetre or two long. A three-centimetre tumour is about fourteen cubic centimetres. The core is about fifteen thousandths of a cubic centimetre. That is a thousandth of the tumour, and then the slide shows four micrometres of that.
- So the honest statement of what a cancer diagnosis is: a confident conclusion about an organ, drawn from a vanishingly small and non-randomly chosen sample of it. That works for cancer because cancer cells are everywhere in the mass. It works badly for anything rare and focal.
- Give the two case reports, because they show both directions. A peritoneal mass called a gastric duplication cyst on CT turned out to be a migrated lung fluke — and the authors' explanation for why biopsy misses these is not chemistry, it is 'an INSUFFICIENT AMOUNT'. And a pancreatic 'mass' that cost a woman her pancreatic body, tail and spleen, which histology then showed to be a hydatid cyst — a tapeworm larva, found ONLY because the specimen was examined.
- Name the real barrier, in the authors' own words: 'clinicians should bear in mind that an intra-abdominal mass may be related to a parasitic infection.' The bottleneck is the differential diagnosis, not the laboratory.
- Then show what is already possible, so the episode ends on something actionable. PCR plus pyrosequencing of a 46-nucleotide window distinguishes five fish-borne trematodes and detects a SINGLE fluke egg in 100 mg of stool. Broad eukaryotic PCR on FFPE tissue is what solved the tapeworm case. Micro-tissue RNA sequencing recovers thousands of genes from a 100-micrometre punch of six-year-old blocks.
- Close on the actual study. Take a series of archived tumour blocks. Run a broad eukaryotic PCR — the one that asks 'what organism is this' rather than 'is this cancer'. Report what fraction return non-human sequence. It is cheap, the tissue is already in the building, the method is published, and nobody has done it.
Define on air: core needle biopsy · resection specimen · sampling error · focal lesion · broad-range PCR · pyrosequencing · FFPE archive
Where it lands: If parasites are being systematically missed in tumour pathology, the arithmetic of sampling is a far larger explanation than any reagent. A thousandth of the mass, four micrometres deep, read by someone who was asked a different question.
Still needed: The definitive experiment has not been run and would not be expensive: broad eukaryotic PCR across an archived tumour series, with the fraction of non-human sequence as the primary endpoint.
- Roh CK, Jung MJ. Laparoscopic excision for ectopic peritoneal paragonimiasis mimicking a gastric duplication cyst: A case report.10.1016/j.amsu.2021.102754 · PMID 34484726
Their stated reason a fluke is missed on biopsy is not chemistry, it is arithmetic: 'insufficient amount'. And their recommendation is not a new reagent, it is a new habit — 'clinicians should bear in mind that an intra-abdominal mass may be related to a parasitic infection'. The barrier they name is the differential diagnosis, not the laboratory.
- Tantrawatpan C, Intapan PM, Thanchomnang T, Sanpool O, Janwan P, Lulitanond V, Sadaow L, Maleewong W. Development of a PCR assay and pyrosequencing for identification of important human fish-borne trematodes and its potential use for detection in fecal specimens.10.1186/1756-3305-7-88 · PMID 24589167
What is possible when somebody builds the assay. PCR plus pyrosequencing of a 46-nucleotide window in the 28S ribosomal gene distinguishes five fish-borne trematodes, and detects a SINGLE Opisthorchis viverrini or Clonorchis sinensis egg in 100 mg of stool. The authors note the reason it was needed: the eggs are 'difficult to differentiate MORPHOLOGICALLY ... even for experienced technicians'. The technology to name an organism from a few dozen bases has existed for over a decade. It is aimed at stool. It has not been aimed at tumours.
- Matsunaga H, Arikawa K, Yamazaki M, Wagatsuma R, Ide K, Samuel AZ, Takamochi K, Suzuki K, Hayashi T, Hosokawa M, Kambara H, Takeyama H. Reproducible and sensitive micro-tissue RNA sequencing from formalin-fixed paraffin-embedded tissues for spatial gene expression analysis.10.1038/s41598-022-23651-6 · PMID 36376423
The arithmetic of the archive, and the change-my-mind study for the whole episode. A 100-micrometre punch, 10 micrometres thick — a few dozen cells — from six-year-old lung cancer blocks yielded usable sequence. The Hymenolepis case was solved by a broad eukaryotic PCR on FFPE tissue. Everything needed to ask 'is there non-human DNA in this tumour?' already exists and is already routine. Nobody has run it at scale.
Protozoa
Babesia
A malaria-like parasite that lives inside the red cell — and in sickle cell disease the cell it moves into is already failing.
Blueprint
- Start with what it is and why it is unfamiliar: Babesia is an apicomplexan — the same phylum as the malaria parasite — and like malaria it lives INSIDE the red blood cell. Unlike malaria it is carried by a tick, in the United States, right now.
- In a healthy adult it is often a flu-like illness or nothing. The severe disease belongs to people without a working spleen, and that is the bridge to sickle cell.
- The evolutionary hook, because it is genuinely surprising. Sickle cell trait exists at all because one copy of the haemoglobin S gene protects against falciparum malaria. So does it protect against malaria's cousin? Partly — in sickle cell ANAEMIA red cells, Babesia does badly: defective merozoite infectivity, impaired egress, much lower parasitaemia. But in sickle TRAIT cells, infection proceeds at rates comparable to normal red cells. The shield does not transfer.
- The damage, measured. Ektacytometry across normal, trait and sickle cells before and after infection: loss of deformability along a spectrum, worst in SS cells — higher point-of-sickling values, loss of surface area, altered osmotic fragility, and hypervesiculation, the cell shedding membrane as vesicles. That is why babesiosis in sickle cell disease produces catastrophic haemolysis rather than a fever.
- The clinical trap, and this is the part with a body count. Babesia is TRANSFUSION-TRANSMISSIBLE. People with sickle cell disease are among the most transfused patients in medicine. Post-transfusion haemolysis in a sickle cell patient looks exactly like a delayed haemolytic transfusion reaction — so the reflex is steroids, which is the wrong treatment and makes the parasite worse.
- Close on the policy lag: the FDA did not require region-based Babesia screening of the US blood supply until 2019.
Define on air: apicomplexan · merozoite · parasitaemia · haemolysis · ektacytometry · deformability · osmotic fragility · functional asplenia · delayed haemolytic transfusion reaction
Where it lands: A tick-borne parasite that is mild in most people and devastating in exactly the group most likely to receive the transfusions that carry it — and whose presentation mimics the transfusion reaction it will be mistaken for. This is not an exotic risk; it is an American one, and it is a diagnostic-reflex problem more than a biology problem.
Still needed: Nobody knows the true incidence of babesiosis in the sickle cell population, because the test is only run when somebody thinks of it.
- Beri D, Rodriguez M, Singh M, McLaughlin D, Liu Y, Zhong H, Mendelson A, An X, Manwani D, Yazdanbakhsh K, Lobo CA. Babesiosis and sickle red blood cells: loss of deformability, altered osmotic fragility, and hypervesiculation.10.1182/blood.2024027602 · PMID 39869831
The mechanism, measured. Using ektacytometry on AA, AS and SS red cells before and after Babesia infection, the damage lands on a spectrum: normal cells lose some deformability, sickle-trait cells more, and SS cells most of all — higher point-of-sickling values, loss of surface area, altered osmotic fragility, and hypervesiculation, shedding membrane as extracellular vesicles. That is a mechanistic explanation for why babesiosis in sickle cell disease produces hyperhaemolysis and extreme anaemia rather than the flu-like illness a healthy adult gets.
- Cursino-Santos JR, Singh M, Senaldi E, Manwani D, Yazdanbakhsh K, Lobo CA. Altered parasite life-cycle processes characterize Babesia infection in human sickle cell anemia.10.3324/haematol.2018.214304 · PMID 30923098
The counter-intuitive half, and the reason this topic is interesting rather than merely sad. Sickle cell TRAIT was selected because it protects against Plasmodium falciparum. So does it protect against Babesia, a close apicomplexan relative? Partly: in sickle-cell-ANAEMIA red cells the parasite does badly — defective merozoite infectivity, impaired egress, much lower parasitaemia. But sickle TRAIT cells support infection at rates comparable to normal red cells. So the trait that shields against malaria does not shield against its cousin, and the disease that limits the parasite is the same disease that cannot survive losing the cells.
- Karkoska K, Louie J, Appiah-Kubi AO, Wolfe L, Rubin L, Rajan S, Aygun B. Transfusion-transmitted babesiosis leading to severe hemolysis in two patients with sickle cell anemia.10.1002/pbc.26734 · PMID 28766838
The clinical trap, in two cases. Babesia microti and B. duncani are transmissible by TRANSFUSION — and people with sickle cell disease are among the most transfused patients in medicine. Post-transfusion haemolysis in a sickle cell patient looks exactly like a delayed haemolytic transfusion reaction, so the reflex is steroids or other immunosuppression, which is the wrong treatment and makes the parasite worse. The authors argue for universal blood screening in endemic areas. Note for the episode: the FDA did not require region-based Babesia screening of the US blood supply until 2019.
Blastocystis
The commonest parasite in human stool — and in a 1,098-person study it tracked with BETTER glucose after meals, not worse.
- Asnicar F, Berry SE, Valdes AM, et al. Microbiome connections with host metabolism and habitual diet from 1,098 deeply phenotyped individuals.10.1038/s41591-020-01183-8 · PMID 33432175
PREDICT 1. Deep metagenomics on 1,203 gut microbiomes against hundreds of cardiometabolic markers. Blastocystis spp. came out as an indicator of FAVOURABLE postprandial glucose metabolism, alongside Prevotella copri. A marker of a good diet is not the same as a cause — but it is the strongest single argument against treating every carrier, and it is the number to open the episode with.
Trichomonas vaginalis
Wet-mount microscopy found 26% of what PCR found in one 980-woman comparison — the commonest curable STI, missed by the test most often used.
- Kim TG, Young MR, Goggins ER, et al. Trichomonas vaginalis in pregnancy: patterns and predictors of testing, infection, and treatment.10.1097/AOG.0000000000003776 · PMID 32282605
3,265 pregnant women; 980 tested by BOTH wet mount and NAAT, so the comparison is within-patient. Microscopy sensitivity 26%, specificity 99% — it almost never calls a false positive and misses three quarters of the true ones. Prevalence 15%. Also worth saying on air: 29% of the retested women were positive again, which is a reinfection problem, not a drug-failure problem.
Trypanosoma cruzi
An estimated 240,000–350,000 people in the United States carry Chagas disease, and about 30% will develop heart or gut damage from it.
- Hochberg NS, Montgomery SP. Chagas disease.10.7326/AITC202302210 · PMID 36780647
In the Clinic review co-authored by the CDC's parasitic diseases branch. 240,000–350,000 infected in the US, mostly immigrants from endemic Latin America; transmission also by transfusion, transplant and vertically. ~30% develop cardiac or GI complications. The teaching point is that this is a US disease that is almost never screened for.
Roundworms
Strongyloides stercoralis
It completes its whole life cycle inside one host, so it persists for decades — and steroids turn that into hyperinfection with 60% ICU mortality.
- Geri G, Rabbat A, Mayaux J, et al. Strongyloides stercoralis hyperinfection syndrome: a case series and a review of the literature.10.1007/s15010-015-0799-1 · PMID 26008854
133 ICU patients. 83.5% were on long-term corticosteroids; in-ICU mortality 60.3%. Two details for the episode: shock was usually driven by a concomitant gram-negative infection carried by the migrating larvae, and eosinophilia was present in only about a third — so the test people rely on is least reliable exactly when the stakes are highest.
Flukes & tapeworms
Dwarf tapeworm
In 2015 a man's tumours turned out to be the tapeworm's cancer, not his. One case, and it changed what a tumour can be.
Blueprint
- Tell the case straight, in the order the doctors experienced it. A man with HIV, biopsied for enlarged lymph nodes and lung lesions. The pathologist sees nests of monomorphic undifferentiated cells, invading, crowded — cancer, by every criterion in the book.
- The detail that stopped them, in the authors' own words: the morphology and invasive behaviour were characteristic of cancer, 'but their SMALL SIZE suggested a nonhuman origin'.
- They ran a broad eukaryotic PCR — a test that asks not 'is this lymphoma' but 'what organism is this'. The answer was Hymenolepis nana, the dwarf tapeworm, the commonest tapeworm in humans and a parasite considered essentially harmless.
- Then the confirmation, which is what makes it publishable rather than a curiosity: immunohistochemistry and in-situ probes labelled the cells as tapeworm tissue, and deep sequencing found structural variants in the parasite genome 'compatible with mutations described in cancer'.
- State the conclusion precisely. The tapeworm's own cells had become malignant, and then colonised a human. This is NOT a demonstration that parasites cause human cancer. It is a demonstration that a competent pathologist can look at a slide, see unambiguous cancer, and be looking at another species — and that only an assay nobody orders would have told them.
- Explain why it could happen: the patient had untreated HIV and was profoundly immunosuppressed. In an intact immune system this almost certainly does not occur. Say that clearly.
- Close on the number that matters for the rest of the episode. Published in the New England Journal in 2015, it remains the only case of its kind ever described. Which is either because it is vanishingly rare, or because the test that found it is almost never ordered. Nobody knows which, and nobody has checked.
Define on air: Hymenolepis nana · monomorphic · undifferentiated · polymerase chain reaction · immunohistochemistry · in situ hybridisation · structural variant · immunosuppression
Where it lands: One case, one profoundly immunosuppressed host, no evidence this generalises. And a permanent change to what the word 'tumour' can mean. Both of those are true and neither is a hedge.
Still needed: The obvious follow-up — broad eukaryotic PCR on a series of undifferentiated tumours in immunosuppressed patients — has not been published. A decade on.
- Muehlenbachs A, Bhatnagar J, Agudelo CA, Hidron A, Eberhard ML, Mathison BA, Frace MA, Ito A, Metcalfe MG, Rollin DC, Visvesvara GS, Pham CD, Jones TL, Greer PW, Vélez Hoyos A, Olson PD, Diazgranados LR, Zaki SR. Malignant Transformation of Hymenolepis nana in a Human Host.10.1056/NEJMoa1505892 · PMID 26535513
The case that changed what a tumour can be. A Colombian man with HIV had nests of monomorphic undifferentiated cells in lymph node and lung — invasive, crowded, cancer by every morphological criterion — except the cells were far too small to be human. A broad eukaryotic PCR returned HYMENOLEPIS NANA, the dwarf tapeworm. Immunohistochemistry and in-situ probes lit the cells up as tapeworm. Deep sequencing found structural variants in the parasite's genome 'compatible with mutations described in cancer'. So: the tapeworm's cells had become malignant, and the man was carrying a cancer that was not his. Say clearly on air that this is ONE case, in a profoundly immunosuppressed host, and that it does not show parasites cause human cancer. What it shows is narrower and stranger — that a pathologist can look at human tissue, see unambiguous cancer, and be looking at another species.
- Conn DB. Malignant Transformation of Hymenolepis nana in a Human Host [letter].10.1056/NEJMc1600490 · PMID 27028926
The correspondence that followed, from a Harvard comparative-zoology parasitologist. Worth having because it is where the field argued about what the case actually demonstrated, and because it is a reminder that the finding survived expert scrutiny rather than being quietly dropped.
Liver flukes
Opisthorchis and Clonorchis cause bile-duct cancer and are also Group 1 carcinogens. Endemic to tens of millions of people.
Blueprint
- Two of the three Group 1 parasitic carcinogens are liver flukes: Opisthorchis viverrini, classified in 1994, and Clonorchis sinensis, classified in 2009.
- The route in is dinner. Raw or fermented freshwater fish carrying the larval stage. Adults reach the bile ducts and feed there for TEN TO THIRTY YEARS.
- The mechanism is three things at once, and they should be given in that order: mechanical damage to the bile-duct lining from an organism physically feeding on it; chronic inflammation and oxidative stress; and proteins the fluke SECRETES that drive host cell division and block host cell death. Plus a dietary co-factor — nitrosamines in the same fermented fish dishes.
- The experiment that turns association into causation, and it is beautiful. A team CRISPR-edited the granulin gene out of the fluke itself — the first programmed gene knockout in a parasitic flatworm — then infected hamsters with the edited parasites. The flukes still colonised the bile ducts and grew normally. But the bile-duct overgrowth and fibrosis that precede cancer were markedly reduced. Delete one parasite growth factor and the pre-cancer largely does not happen.
- The human scale: Khon Kaen province in Thailand has the highest cholangiocarcinoma incidence in the world. Five-year survival under 5%. Usually found late, usually unresectable.
- The part that should unsettle: mass deworming FAILED. Reinfection rates are high, and the authors raise the possibility that repeated cure-and-reinfection cycles may themselves PROMOTE carcinogenesis.
- Close on the fourth fluke. Opisthorchis felineus, endemic across Siberia, up to 80% infestation in endemic areas. In hamsters, infection plus a nitrosamine produced cholangiocarcinoma in every animal by week 18. It has no IARC category at all.
Define on air: cholangiocarcinoma · bile duct · granulin · CRISPR · gene knockout · fibrosis · nitrosamine · metacercaria
Where it lands: This is the strongest causal case in parasite oncology and it is very strong: a defined organism, a defined organ, a defined secreted molecule, and a gene-knockout experiment that removes the molecule and reduces the disease. If you want to know what proof looks like in this field, it looks like the granulin experiment.
Still needed: Opisthorchis felineus has a positive animal model, an 80%-infested population across Siberia, and no human epidemiology worth the name.
- Sivanand A, Talati D, Kalariya Y, Patel P, Gandhi SK. Associations of liver fluke infection and cholangiocarcinoma: a scoping review.10.7759/cureus.46400 · PMID 37927641
Scoping review. Mechanism is specific and worth the airtime: mechanical injury from the fluke feeding in the bile duct, plus a secreted protein (OvGRN-1) taken up by human cholangiocytes that shifts gene expression toward wound-healing and cancer pathways. Note the journal — Cureus has light peer review, so verify the primary OvGRN-1 work before quoting it.
- Kim TS, Pak JH, Kim JB, Bahk YY. Clonorchis sinensis, an oriental liver fluke, as a human biological agent of cholangiocarcinoma: a brief review.10.5483/bmbrep.2016.49.11.109 · PMID 27418285
Names the three organisms IARC lists as Group 1 human carcinogens: Clonorchis sinensis, Opisthorchis viverrini and Schistosoma haematobium. Also the scale nobody in the West registers — C. sinensis was the most prevalent parasite in the 2012 Korean national survey at 1.86% of the general population, and it is still actively transmitted.
- Buisson Y. [Control of Opisthorchis viverrini infection for cholangiocarcinoma prevention].10.1007/s13149-017-0544-8 · PMID 28105582
The clearest account of the mechanism, and it is three mechanisms braided: mechanical damage from adult flukes feeding on bile-duct lining for TEN TO THIRTY YEARS; chronic inflammation and oxidative stress; and mitogenic, anti-apoptotic proteins the fluke itself secretes. Plus a dietary co-factor — nitrosamines in the fermented raw fish that carries the parasite. Khon Kaen province in Thailand has the highest cholangiocarcinoma incidence on earth. Five-year survival under 5%. The detail that should end the episode's optimism: mass deworming FAILED, because people are reinfected, and repeated cure-and-reinfection cycles may themselves promote carcinogenesis.
- Arunsan P, Ittiprasert W, Smout MJ, Cochran CJ, Mann VH, Chaiyadet S, Karinshak SE, Sripa B, Young ND, Sotillo J, Loukas A, Brindley PJ, Laha T. Programmed knockout mutation of liver fluke granulin attenuates virulence of infection-induced hepatobiliary morbidity.10.7554/eLife.41463 · PMID 30644359
The experiment that turns the association into causation. The team CRISPR-edited the granulin gene out of Opisthorchis viverrini itself — the first programmed gene knockout in a parasitic flatworm — then let the edited flukes infect hamsters. The flukes colonised the bile ducts and grew to adults normally, but the bile-duct overgrowth and fibrosis that precede cholangiocarcinoma were markedly reduced. Delete one parasite growth factor, and the pre-cancer largely does not happen. This is as close to proof of a parasite-driven cancer mechanism as the field has.
- Feng M, Cheng X. Parasite-Associated Cancers (Blood Flukes/Liver Flukes).10.1007/978-981-10-5765-6_12 · PMID 29052139
The compact summary of all three Group 1 organisms in one place, and blunt about the hole: the cellular and molecular mechanisms linking fluke infection to cancer 'have yet to be defined'. Also the epidemiological point that makes this a diagnosis problem as much as a biology problem — infection is frequently ASYMPTOMATIC and therefore rarely diagnosed early.
- Maksimova GA, Zhukova NA, Kashina EV, Lvova MN, Katokhin AV, Tolstikova TG, Mordvinov VA. [Role of opisthorchis felineus on induction of bile duct cancer].PMID 26016330
The fourth fluke, and the one that is not on the IARC list. Opisthorchis felineus is endemic across Siberia and parts of Europe, with up to 80% infestation in endemic areas. In hamsters, infection alone produced hyperplastic and dysplastic bile-duct change; infection plus dimethylnitrosamine produced cholangiocarcinoma in EVERY animal by week 18. The authors argue their data should be used to assign it an IARC category. It still has none. That is not evidence of safety; it is evidence that nobody has run the human study.
Schistosoma haematobium
A parasite the IARC classifies as a Group 1 human carcinogen. Infection causing cancer is established science, not a fringe claim.
Blueprint
- Establish the category first, because the whole episode runs on it. IARC — the World Health Organization's cancer agency — sorts agents into groups. Group 1 means carcinogenic to humans, and the evidence is beyond argument: tobacco, asbestos, and three parasites.
- Schistosoma haematobium is one of them. The life cycle is worth telling properly: eggs in fresh water, larvae into a snail, out as cercariae that burrow through skin, adults pairing in the veins of the bladder wall, eggs pushed out through the bladder lining for decades.
- The cancer it causes is not the one Western urologists expect. Schistosoma-associated bladder cancer is typically SQUAMOUS CELL carcinoma, not the urothelial type that smoking causes. Different cell, different behaviour, same organ.
- Then the honest limit, from the review itself: several carcinogenic pathways have been proposed and the exact mechanisms are 'not defined yet'. Group 1 means the epidemiology is settled. It does not mean the biology is understood.
- The contrast that teaches the whole IARC system: S. mansoni, the close relative, is GROUP 3 — not classifiable. Case reports, animal models and cell culture all suggest it may predispose to liver and colorectal cancer, and there is not enough to classify. Group 3 is not a clean bill of health; it is a statement about the literature, not the organism.
- Land the migration point, which makes this a Chicago question and not only an African one: these infections travel with people, and clinicians in non-endemic countries have neither the experience nor the index of suspicion.
Define on air: IARC groups 1, 2A, 2B and 3 · squamous cell carcinoma · urothelial carcinoma · cercaria · endemic
Where it lands: One parasite in this genus is as established a human carcinogen as asbestos. Its nearest relative sits in the category that means nobody has looked hard enough. That asymmetry, in a single genus, is the shape of the entire parasite-cancer question.
Still needed: S. mansoni has no adequate human cohort. The animal and mechanistic work points somewhere; the epidemiology to test it has not been funded.
- Efared B, Bako ABA, Idrissa B, et al. Urinary bladder Schistosoma haematobium-related squamous cell carcinoma: a report of two fatal cases and literature review.10.1186/s40794-022-00161-x · PMID 35164874
Two fatal cases plus review, so weak as evidence of incidence — cite it for the mechanism and the histology, not for risk figures. States S. haematobium is a proven carcinogenic agent causing bladder SQUAMOUS cell carcinoma, which is the point: a different histology from the transitional cell cancer most bladder cancer is. Needs an IARC monograph alongside it before broadcast.
- Zaghloul MS, Zaghloul TM, Bishr MK, Baumann BC. Urinary schistosomiasis and the associated bladder cancer: update.10.1186/s43046-020-00055-z · PMID 33252773
The clinical review of the one parasite-cancer link nobody disputes. Schistosoma haematobium is an IARC GROUP 1 carcinogen — the same category as tobacco and asbestos — and the cancer it causes is usually SQUAMOUS CELL carcinoma of the bladder rather than the urothelial type the West sees. Note the honest gap they name: several carcinogenic pathways have been proposed and 'the exact mechanism(s) are not defined yet'. Group 1 does not mean the mechanism is known; it means the epidemiology is beyond argument.
- von Bülow V, Lichtenberger J, Grevelding CG, Falcone FH, Roeb E, Roderfeld M. Does Schistosoma mansoni Facilitate Carcinogenesis?10.3390/cells10081982 · PMID 34440754
The edge of the evidence, and the most useful paper for showing how IARC categories actually work. S. haematobium is Group 1 — definitely carcinogenic. S. mansoni, its close relative, is GROUP 3 — insufficient evidence to classify, which is not the same as evidence of safety. The authors argue from case reports, animal models and cell culture that mansoni may still predispose to liver and colorectal cancer. This is the honest shape of the parasite-cancer question outside the three settled organisms: a real hypothesis, a real mechanism, and a category that says nobody has done the study.
Antiparasitic drugs
Artemisinin
Pulled from a 1,600-year-old Chinese fever recipe, extracted cold because heat destroyed it, and now the backbone of malaria treatment.
Blueprint
- The false binary this dissolves: natural versus pharmaceutical. A Nobel Prize was awarded for reading a fourth-century Chinese text carefully.
- The mechanism is the peroxide bridge, cleaved by iron inside the parasite.
- The reason it is given as a combination is resistance — which has already emerged in the Mekong and is now reported in Africa.
- The thing sold as 'artemisinin' in supplement form is not artemisinin combination therapy and does not treat malaria.
Define on air: A · · s · e · s · q · u · i · t · e · r · p · e · n · e · · l · a · c · t · o · n · e · · w · i · t · h · · a · n · · e · n · d · o · p · e · r · o · x · i · d · e · · b · r · i · d · g · e · , · · f · r · o · m · · A · r · t · e · m · i · s · i · a · · a · n · n · u · a · · — · · s · w · e · e · t · · w · o · r · m · w · o · o · d · , · · n · o · t · · t · h · e · · a · b · s · i · n · t · h · e · · s · p · e · c · i · e · s · .
Where it lands: The strongest argument in existence that the plant-versus-pharma framing is a marketing invention — and a warning about what happens when you take the same molecule alone.
Still needed: Nothing about the antimalarial. The open question is whether the oncology work on artesunate is anything more than the concentration problem all over again.
- Tu Y. The discovery of artemisinin (qinghaosu) and gifts from Chinese medicine.10.1038/nm.2471 · PMID 21989013
Her own account, four years before the Nobel. The detail that matters for this show: the classical text said to steep the herb in cold water and wring it out, and that instruction is why the extraction worked — artemisinin degrades with heat, so the conventional hot extraction had been destroying it. The tradition did not contain the pharmacology. It contained the method, and somebody had to notice.
Ivermectin
Avermectins came out of one soil sample near a Japanese golf course and won a Nobel. The cancer claim is a real preclinical signal with no human trial behind it.
Blueprint
- Set the terms before anything else, because this molecule arrives pre-loaded: this episode is about oncology, and it is neither a defence nor a debunking of anything else.
- The preclinical signal is real and worth describing — several distinct cancer pathways, across multiple cell lines and animal models.
- That is a legitimate reason to run a trial. It is not a reason to take a drug.
- The finding that has to be stated plainly: a PubMed search for human trials of ivermectin in cancer returned ZERO results. The review itself says clinical validation remains limited.
- Explain why preclinical-to-human failure is the norm, not the exception — concentrations achievable in a dish are often unreachable in a person.
- End on what an honest integrative position looks like here: interested, unconvinced, and specific about what would change its mind.
Define on air: preclinical · cell line · xenograft · pharmacokinetics · clinical validation
Where it lands: A real preclinical signal with no human evidence. Both halves of that sentence are load-bearing.
Still needed: Any registered human trial. Until one reports, nothing changes.
- Robalino KN, Vivanco-Galván O, Romero-Benavides JC, Jiménez-Gaona Y. Ivermectin as an Alternative Anticancer Agent: A Review of Its Chemical Properties and Therapeutic Potential.10.3390/ph18101459 · PMID 41155573
Review. States plainly that despite preclinical evidence, clinical validation remains limited. A PubMed search for human ivermectin cancer trials returned zero results on 2026-09-16.
- Hu B, Tan H, Yu L, Liao Q, Guo W. Repurposing Ivermectin to augment chemotherapy's efficacy in osteosarcoma.10.1177/09603271221143693 · PMID 36503300
Cell lines and mouse xenograft. Synergy with doxorubicin. Preclinical only — this is the shape of the whole literature.
- Bhadra S, Xu YJ. TTT (Tel2-Tti1-Tti2) Complex, the Co-Chaperone of PIKKs and a Potential Target for Cancer Chemotherapy.10.3390/ijms24098268 · PMID 37175973
The best answer to 'but what would it even DO?' Ivermectin binds TEL2, part of the TTT complex, which is the chaperone that folds a family of very large kinases — ATR, ATM, DNA-PKcs, mTOR, SMG1, TRRAP — the machinery of DNA-damage response and growth signalling. A HUMAN TARGET, not a worm one. That is a plausible anticancer mechanism, and it is a long way from a plausible mechanism to a treatment.
- Hayashi A, Kamio K, Miyanaga A, Yoshida K, Noro R, Matsuda K, Tozuka T, Omori M, Hirao M, Fukuizumi A, Hisakane K, Takeuchi S, Matsumoto M, Kasahara K, Amano T, Honda K, Seike M. Ivermectin Enhances Paclitaxel Efficacy by Overcoming Resistance Through Modulation of ABCB1 in Non-small Cell Lung Cancer.10.21873/anticanres.17355 · PMID 39626921
The most interesting preclinical result, because it is not about killing cells. A549 lung cancer cells made resistant to paclitaxel did it by overexpressing P-glycoprotein, the pump that throws chemotherapy back out of the cell. Giving ivermectin alongside the escalating paclitaxel ABOLISHED that P-glycoprotein induction, so the drug stayed inside and the cells never became resistant. If ivermectin has a real place in oncology, resistance-prevention alongside chemotherapy is a better hypothesis than tumour-killing on its own.
- Shukla A, Sharma A, Gupta S, Mishra A, Singh A. Antitumor potential of ivermectin against T-cell lymphoma-bearing hosts.10.1007/s12032-025-02726-0 · PMID 40257544
Representative of the preclinical literature, and useful for the number that matters. Ivermectin killed lymphoma cells with an IC50 of 10.55 micrograms per millilitre; cisplatin managed 8.32 in the same assay. The catch is the one nobody quotes: ordinary human dosing produces plasma levels of roughly 0.02-0.05 micrograms per millilitre. The concentration that kills cancer cells in a dish is two to three ORDERS OF MAGNITUDE above what a person taking ivermectin achieves. That gap is the entire story of this drug in oncology.
- Sheir MM, El-Habashy SE, Sheta E, Nasra MMA, Abdallah OY. Biomimetic platelet-membrane camouflaged ivermectin nanocrystals for tumor homing and breast cancer management.10.1007/s13346-025-02032-2 · PMID 41495333
What serious pharmacologists do about that gap: they stop giving it as a tablet. Ivermectin nanocrystals wrapped in platelet membrane to home to tumour and evade immune clearance, active against triple-negative breast cancer in mice. Note what the authors say plainly in their own opening — repurposing ivermectin is 'hindered by poor solubility and HIGH TOXICITY, restricting its parenteral administration'. The people most invested in this drug's anticancer future are the ones building whole delivery systems to avoid giving it the way people are buying it.
- Expression of Concern [regarding Hulscher et al., Real-world Clinical Outcomes of Ivermectin and Mebendazole in Cancer Patients: Results from a Prospective Observational Cohort].10.21873/anticanres.18276 · PMID 42300708
The single most important document for this node, and it must be read on air. In 2026 Anticancer Research published a cohort reporting an 84.4% 'Clinical Benefit Ratio' for ivermectin and mebendazole in cancer patients. The journal then issued a formal Expression of Concern and opened a Post-Publication Data Integrity and Ethical Oversight Audit — examining whether IRB approval existed, whether the 197 participants' baseline cancer diagnoses can be source-verified, and whether the reported tumour regressions have objective medical documentation. The editorial board's sentence is the one to quote: 'Disclosing limitations does not exempt a clinical dataset from the foundational scientific requirements of empirical verifiability and independent ethical oversight.' This is what the strongest 'real-world evidence' for ivermectin in cancer currently amounts to.
Bacteria
H. pylori
The precedent everybody forgot: a bacterium causing ulcers and gastric cancer, dismissed for a decade, then a Nobel Prize.
Blueprint
- Open with the logic everyone accepts: a bacterium is a class-1 carcinogen for gastric cancer, so eradicating it should prevent gastric cancer.
- The trial that tested it at scale: 240,000 adults in Taiwan, randomised to an invitation to H. pylori stool-antigen testing plus bowel-cancer screening, or bowel screening alone.
- The primary outcome: NULL. No reduction in gastric cancer incidence. No reduction in gastric cancer mortality.
- Then the post-hoc analysis — adjusting for participation rates and follow-up — which found lower incidence (0.79) but still no mortality benefit.
- Teach the distinction, because it is the most useful thing in the episode: a pre-specified primary outcome is a test. A post-hoc analysis is a hypothesis. They are not interchangeable, and headlines routinely report the second as the first.
- Close on what the trial actually tested — an INVITATION to screening, not eradication itself — and why that distinction matters for what can be concluded.
Define on air: pragmatic trial · primary outcome · post-hoc analysis · intention-to-treat · stool antigen test · carcinogen classification
Where it lands: The best test we have of the screening strategy returned null on its own terms. The favourable number exists but comes from an analysis that cannot carry the claim.
Still needed: Longer follow-up; and whether targeting high-prevalence or high-risk groups performs differently.
- Lee YC, Chiang TH, Chiu HM, et al. Screening for Helicobacter pylori to prevent gastric cancer: a pragmatic randomized clinical trial.10.1001/jama.2024.14887 · PMID 39348147
240,000 adults in Taiwan. The PRIMARY outcome was null: inviting people to H. pylori screening did not reduce gastric cancer incidence or mortality versus colorectal screening alone. Only a POST HOC analysis adjusting for participation and follow-up found lower incidence (0.79, 0.63–0.98) — and still no mortality benefit. Anyone citing this as proof that eradication prevents gastric cancer is citing the post-hoc number, and the episode should say which number it is using.
What is actually in the bottle
What is in a melatonin gummy
A 2023 analysis of 25 products found doses from 74% to 347% of the label, and one carried 1.3 mg of CBD that was not declared at all.
- Cohen PA, Avula B, Wang YH, Katragunta K, Khan I. Quantity of Melatonin and CBD in Melatonin Gummies Sold in the US.10.1001/jama.2023.2296 · PMID 37097362
25 products analysed. Melatonin content ranged from 74% to 347% of the labelled amount; one product contained no detectable melatonin and 31.3 mg of CBD. Check the CBD figure against the paper before it is said aloud — reports of this study quote it both ways.
Skin
Niacinamide
It blocks pigment transfer to the keratinocyte and rebuilds ceramides. A 2015 trial also cut new skin cancers by 23% in high-risk patients — at 500 mg twice a day, oral.
- Chen AC, Martin AJ, Choy B, et al. A Phase 3 Randomized Trial of Nicotinamide for Skin-Cancer Chemoprevention.10.1056/NEJMoa1506197 · PMID 26488693
ONTRAC. 386 high-risk participants, 500 mg twice daily for 12 months: new non-melanoma skin cancers 23% lower (95% CI 4 to 38, P=0.02). Note for the episode: the confidence interval reaches almost to 4%, the trial was in an Australian high-risk population, and the benefit VANISHED once the tablets stopped. Not a general-population claim.
Oxygen & breathing
Pulse oximetry
It infers saturation from two wavelengths of light, and the dissociation curve is flat at the top — so a patient on far too much oxygen reads the same as one on exactly enough.
- Sjoding MW, Dickson RP, Iwashyna TJ, Gay SE, Valley TS. Racial bias in pulse oximetry measurement.10.1056/NEJMc2029240 · PMID 33326721
A research letter, not a full paper — worth saying when it is cited, because it changed practice at that weight. Occult hypoxaemia roughly three times more common in Black patients.
The oxygen target
Pooling 25 trials and 16,037 acutely ill adults, MORE oxygen raised in-hospital mortality by 21%. The harm appears to start around a saturation of 94 to 96 per cent.
- Chu DK, Kim LH-Y, Young PJ, et al. Mortality and morbidity in acutely ill adults treated with liberal versus conservative oxygen therapy (IOTA): a systematic review and meta-analysis.10.1016/S0140-6736(18)30479-3 · PMID 29726345
25 RCTs, 16,037 patients. In-hospital mortality RR 1.21 (1.03–1.43), 30-day 1.14 (1.01–1.29), longest follow-up 1.10 (1.00–1.20), I² = 0%, GRADE high. The trials span sepsis, stroke, trauma, MI and cardiac arrest — heterogeneous populations, homogeneous result, which is the striking part. The longest-follow-up CI touches 1.00.
Hydrogen & the proton
Hydrogen breath testing
Human cells make no hydrogen, so every molecule in your breath came from gut bacteria. Elegant physiology, and the weakest widely-ordered test in this atlas.
- Rezaie A, Buresi M, Lembo A, et al. Hydrogen and methane-based breath testing in gastrointestinal disorders: the North American Consensus.10.1038/ajg.2017.46 · PMID 28323273
Hydrogen rise ≥20 ppm by 90 minutes is positive for SIBO; methane ≥10 ppm is methane-positive. Doses: lactulose 10 g, glucose 75 g, fructose 25 g, lactose 25 g. It is a consensus statement voted by ten specialists — the thresholds are agreed, not validated against a reference standard, and the document says so.
Tin & the obesogens
Tributyltin
The most effective antifouling paint ever put on a hull, and a nanomolar agonist at the two nuclear receptors that decide whether a cell becomes a fat cell.
- Grün F, Blumberg B. Environmental obesogens: organotins and endocrine disruption via nuclear receptor signaling.10.1210/en.2005-1129 · PMID 16690801
Tributyltin and triphenyltin as nanomolar agonists at RXRα/β/γ and PPARγ. A review advancing the obesogen hypothesis — the receptor pharmacology is solid, the human obesity claim is the hypothesis, and the two should not be quoted as one thing.