One paragraph each, with the primary source behind it. Written to be quoted.
They are different substances and the difference is chemistry, not semantics. Household bleach is sodium hypochlorite, NaClO, an alkaline, caustic compound. CDS is chlorine dioxide, ClO2, a gas dissolved in water giving a neutral solution with no sodium and no hydroxide. A third substance is routinely folded into both: sodium chlorite, NaClO2, the base of MMS once acidified. The three have different CAS numbers and different behaviour. Collapsing them into one word is not a nuance: it is the foundation of almost every allegation in circulation.
→ #C01 · full entry, verdict and source
They share the source reaction, not the final product. MMS is the reacting mixture: acidified sodium chlorite drunk as such, with the acid and residual chlorite still in it. CDS uses the same reaction, but the chlorine dioxide generated is drawn off as gas and dissolved in separate water; the acid and chlorite stay in the generator. The difference is what is ingested. It can be checked with two glasses and a pH meter — and it should be, before attributing to CDS what was observed with the acid mixture.
→ #C02 · full entry, verdict and source
The published figure points the other way. The US EPA assessment (Case 4023) records an oral LD50 for chlorine dioxide of 292 mg/kg in male rats and 340 mg/kg in females. Published values for sodium hypochlorite sit in the hundreds to thousands of mg/kg depending on concentration. Orally, chlorine dioxide is the less acutely toxic of the two, not thirty times more. And the arithmetic matters: 292 mg/kg in a 70 kg adult is 20.4 grams — about 6.8 litres of the 0.3% stock solution.
→ #C05 · full entry, verdict and source
I went through the FDA public file (openFDA, FAERS) report by report. Under “chlorine dioxide” there are eight: none records a dose, in none was the substance verified, none is a death, none describes methemoglobinemia. Under “sodium chlorite” there are sixteen, eleven unrelated to the question. That does not show the product is harmless — a voluntary system without doses proves nothing in either direction — but it shows the file does not support the claim hung on it.
→ #C06 · full entry, verdict and source
The Plottier case, Neuquén, 2020. The forensic board of the Public Prosecutor's Office wrote on 27 August 2020 that the findings were “compatible” — the word appears three times — and left histopathology and toxicology pending. “Compatible” does not mean “caused”. The case closed without conviction: ingestion was not held proven, because the evidence was the parents' statements and the court found them inadmissible. The headlines said “confirmed”. The document does not.
→ #C08 · full entry, verdict and source
It was tested, and under control. Lubbers, Chauhan, Miller and Bianchine published in 1981 a study in 334 healthy adult males: rising single-dose tolerance to chlorine dioxide (0.1 to 24 ppm) and chlorite, a twelve-week chronic phase at 5 mg/L, and a phase in G6PD-deficient subjects — the population most exposed to oxidative haemolysis. The EPA review derives from that work a human NOAEL of 0.036 mg/kg/day. What does not exist is a randomised trial with clinical endpoints: those are two different statements.
→ #C12 · full entry, verdict and source
A scoping review using PRISMA methodology identified 34 studies, including fifteen in animals and nineteen in humans, among them six clinical studies. Their quality can be argued — and should be: none is a randomised trial with clinical endpoints — but “no evidence at all” is a different statement and it is false. The distinction matters because whoever uses it loses credibility for the rest of the argument, even where the rest is right.
→ #C14 · full entry, verdict and source
There is a randomised trial, the Osaka one from 2024, and it did not prove that. It compared a 0.01% solution used as a mouthrinse against the salivary SARS-CoV-2 titre. That is a topical intervention, thirty times more dilute, on a virological outcome: it says nothing about a 0.3% solution ingested and its clinical endpoints. Citing it as proof of systemic inefficacy is a leap the paper itself does not make.
→ #C15 · full entry, verdict and source
No conviction has been located in Spain, Germany or Switzerland. The Swiss federal criminal record extract, with a qualified electronic signature, dated 5 November 2021, states: “ist im Strafregister nicht verzeichnet” — not registered. In Spain there is a record of a private prosecution being admitted in Alicante in 2019 and later dismissed. Admission is not a conviction, and the Granada acquittal is published with its document. Anyone asserting a conviction can produce the case number; so far, nobody has.
→ #C19 · full entry, verdict and source
He has one, published in 2026: Kalcker AL, “The Mechanism of Action of Chlorine Dioxide (CDS): Electron-Transfer Chemistry, Antimicrobial Selectivity and Redox-Regulatory Effects: A Narrative Review”, Biophysica 2026;6(5):91, doi:10.3390/biophysica6050091. Received 1 August, accepted 4 September, published 21 September 2026, after peer review. It is a narrative review, not primary clinical data — a different and legitimate criticism — but the claim that no peer-reviewed publication exists is false and checkable in a minute.