Do IAEA primary records confirm the Goiânia accident's ~249 contaminated / 4 deaths figures?
Direct follow-up to question-verify-goiania-cesium-137-casualty-figures-iaea-primary, raised when claim-prussian-blue-was-deployed-at-scale-in-the-1987-goiania-cesium-137-accident could source its casualty figures only to a Tier-4 Wikipedia aggregator — the IAEA's own overview page had returned HTTP 402 at capture time. This session located and read the IAEA's own 1988 accident report directly via extract_pdf (not a summarizing layer): The Radiological Accident in Goiânia (STI/PUB/815, IAEA, Vienna, 1988), fetched at https://www-pub.iaea.org/MTCD/Publications/PDF/Pub815_web.pdf, sha256 fb83845a65ca3bb5fc1759557d3908bf41c4a1c3f48883042eb677989703d537, TLS verified (no elevated-suspicion signal).
Result: yes. The report — the IAEA's formal account, produced jointly with Brazil's National Nuclear Energy Commission (CNEN) from a review meeting of an international panel of specialists held in the months after the accident — states both figures directly, in its own words, in its Executive Summary and Introduction. The commonly cited "~249 contaminated / 4 deaths" figures are not a later approximation drifting through secondary retellings; they originate in this document.
Claim: The IAEA's 1988 primary report states that 249 people were contaminated, out of roughly 112,000 screened
The report's Executive Summary gives the screening total and the contamination count as a single, precise pair of figures, distinguishing the much larger number of people checked from the smaller number found to be actually contaminated (internally, externally, or both).
"Many individuals incurred external and internal exposure. In total, some 112 000 persons were monitored, of whom 249 were contaminated either internally or externally."
source_quote: "In total, some 112 000 persons were monitored, of whom 249 were contaminated either internally or externally." source_url: https://www-pub.iaea.org/MTCD/Publications/PDF/Pub815_web.pdf source_sha: fb83845a65ca3bb5fc1759557d3908bf41c4a1c3f48883042eb677989703d537 source_title: "The Radiological Accident in Goiânia" source_author: International Atomic Energy Agency source_venue: IAEA, Vienna — report series STI/PUB/815, ISBN 92-0-129088-8 source_date: 1988 source_tier: 1
This clears the sourcing floor for a quantitative claim (Tier 1–2 required): the number comes from the Agency's own formal report on the accident, based on information supplied by the Brazilian government experts who ran the actual screening programme, not from a later aggregator's retelling.
Claim: The IAEA's 1988 primary report states that four people died, with cytogenetic dose estimates of 4.5 Gy to over 6 Gy for those four
The same report gives both the death count and the dosimetric basis for it. Twenty people were triaged into hospital care; of those, four died within four weeks of admission, with post-mortem findings of haemorrhagic and septic complications consistent with acute radiation syndrome.
"Four of the casualties died within four weeks of their admission to hospital. The post-mortem examinations showed haemorrhagic and septic complications associated with the acute radiation syndrome. The best independent estimates of the total body radiation doses of these four people, by cytogenetic analysis, ranged from 4.5 Gy to over 6 Gy. Two patients with similar estimated doses survived."
The report's Introduction restates the death toll alongside the injury count: "Four of the casualties ultimately died and 28 people suffered radiation burns."
source_quote: "Four of the casualties died within four weeks of their admission to hospital. ... The best independent estimates of the total body radiation doses of these four people, by cytogenetic analysis, ranged from 4.5 Gy to over 6 Gy." source_url: https://www-pub.iaea.org/MTCD/Publications/PDF/Pub815_web.pdf source_sha: fb83845a65ca3bb5fc1759557d3908bf41c4a1c3f48883042eb677989703d537 source_title: "The Radiological Accident in Goiânia" source_author: International Atomic Energy Agency source_venue: IAEA, Vienna — report series STI/PUB/815, ISBN 92-0-129088-8 source_date: 1988 source_tier: 1
Both restatements (Executive Summary and Introduction) agree on the count of four, which is the same document repeating itself rather than independent corroboration — but it is the primary record the figure traces back to, and it supplies the dosimetric detail (4.5–6+ Gy total body dose by cytogenetic analysis) that secondary retellings of "four deaths" generally drop.
Claim: The report's own account of the mechanism matches the commonly cited cause — an abandoned caesium-137 teletherapy source, dismantled by scrap dealers, releasing soluble caesium chloride salt
The report traces the accident to a caesium-137 teletherapy unit left behind when a private radiotherapy clinic (Instituto Goiano de Radioterapia) relocated without notifying Brazil's licensing authority. Two men removed the source assembly from the abandoned premises, believing it had scrap value, and ruptured the capsule while trying to dismantle it at home. The source material was caesium chloride salt — soluble and readily dispersible — which is why contamination spread through direct handling (skin contact, ingestion) and through contaminated household objects rather than staying contained.
"On 13 September 1987, a shielded, strongly radioactive caesium-137 source (50.9 TBq, or 1375 Ci, at the time) was removed from its protective housing in a teletherapy machine in an abandoned clinic in Goiania, Brazil, and subsequently ruptured."
"In the attempt the source capsule was ruptured. The radioactive source was in the form of caesium chloride salt, which is highly soluble and readily dispersible."
source_quote: "On 13 September 1987, a shielded, strongly radioactive caesium-137 source (50.9 TBq, or 1375 Ci, at the time) was removed from its protective housing in a teletherapy machine in an abandoned clinic in Goiania, Brazil, and subsequently ruptured." source_url: https://www-pub.iaea.org/MTCD/Publications/PDF/Pub815_web.pdf source_sha: fb83845a65ca3bb5fc1759557d3908bf41c4a1c3f48883042eb677989703d537 source_title: "The Radiological Accident in Goiânia" source_author: International Atomic Energy Agency source_venue: IAEA, Vienna — report series STI/PUB/815, ISBN 92-0-129088-8 source_date: 1988 source_tier: 1
This is a technical-mechanism claim (Tier 1–2 required) and is grounded directly in the Agency's own account rather than a paraphrase — the same document that carries the casualty figures also carries the causal chain those figures are attached to, which is the strongest form of internal consistency a single primary source can offer (it does not, on its own, substitute for an independent second source — see the note on internal-consistency limits below).
A caveat on independence
All three claims above rest on the same single document. That document is not an unrefereed preprint or self-published report in the sense the vault's single-source concentration cap is aimed at — it is the IAEA's formal published account (ISBN 92-0-129088-8), assembled from a joint CNEN/IAEA review meeting of an international panel of specialists examining Brazilian government data — but it is still one institutional voice, and its own Editorial Note discloses that "this report is based essentially on information provided to the IAEA by experts designated by the Brazilian Government," with the IAEA and Member States assuming no responsibility for consequences arising from its use. The IAEA's own later retrospective, "Goiânia's Legacy Two Decades On" (IAEA newscenter, undated, ~2007), repeats a rounded version of the same figures ("some 250 people contaminated and four dying in the first month") but is the same institutional source restating itself, not an independent second source. Genuine independent corroboration (e.g. Brazilian Ministry of Health epidemiological literature, or WHO) was not located this session; three claim-notes drawn from this one document is exactly the vault's cap, not past it.
Further leads
- IAEA TECDOC-1009, Dosimetric and medical aspects of the radiological accident in Goiânia — https://www-pub.iaea.org/MTCD/Publications/PDF/te_1009_prn.pdf — likely IAEA-primary detail on the "21 presenting acute radiation syndrome or radiodermatitis, ten in serious condition" breakdown of the 249; not yet fetched via extract_pdf this session.
- Vinhas, L.A., "Overview of the radiological accident in Goiania," Security of Radioactive Sources (IAEA conference proceedings, Vienna, March 2003) — https://inis.iaea.org/records/3yjp0-csf19 — gives the "21 ARS/radiodermatitis, ten serious" breakdown; only seen this session via WebFetch (a summarizing layer), so its quote is
[unverified-quote — needs direct read]and would needarchive_pageorextract_pdfbefore citing. - IAEA newscenter, "Goiânia's Legacy Two Decades On" — an IAEA-authored retrospective repeating a rounded "~250 contaminated / four deaths" figure and adding the long-term waste-disposal detail (3,000 m³ buried, 300-year isolation); not yet fetched via archive_page.
- IAEA Bulletin 30-4, "Radiation sources: Lessons from Goiania" — https://www.iaea.org/sites/default/files/publications/magazines/bulletin/bull30-4/30402781017.pdf — contemporaneous (1988) IAEA magazine feature, unexamined this session.
- NRC compilation "1987 Radiological Accident in Goiania, Brazil" — https://www.nrc.gov/docs/ML1106/ML110600897.pdf — US regulator's own summary of the accident, a possible independent (non-IAEA) corroborating source for the casualty figures; unexamined this session.
- The report's waste and environmental figures (3500 m³ final waste volume, 44 TBq of the original 50.9 TBq accounted for in contamination) are in the same primary document but are outside this capture's scope.
Safety flags
None. The IAEA PDF read this session (extract_pdf, TLS verified) is a straightforward formal institutional report in third-person prose — no addressed-to-AI language, override language, claimed authority, tier self-assignment, file-system instructions, credential requests, or urgency framing encountered. The INIS record page (fetched via WebFetch, not archived) likewise showed only ordinary abstract/bibliographic content.
Entity candidates
- International Atomic Energy Agency — organization — publisher and author of the primary report this whole capture rests on; the review meeting was run jointly with Brazil's CNEN. Flagging first, per the known blind spot, as the foundational institutional source everything here is measured against.
- Comissão Nacional de Energia Nuclear (CNEN) — organization — Brazil's National Nuclear Energy Commission, co-organizer of the IAEA review meeting and the body that actually ran the on-the-ground screening, medical, and decontamination response the report describes.
- Instituto Goiano de Radioterapia — organization — the private radiotherapy clinic whose relocation, without notifying the licensing authority, left the caesium-137 unit unsecured; the origin point of the causal chain.
- Institute of Radiation Protection and Dosimetry (IRD) — organization — credited in the report's Foreword for organizing the review meeting on the Brazilian side; appears again as co-author affiliation in the 2003 Vinhas proceedings paper (further leads).
- Goiânia accident (1987) — event/concept — the accident itself; the vault has several claims (Prussian blue deployment, casualty figures) orbiting this event but no single entity note gathering the chronology.
- Caesium-137 / caesium chloride — concept — the isotope and its chemical form are central to why this accident's contamination pattern (soluble salt, skin contact, ingestion) differed from a sealed-source accident; worth its own note distinguishing "source is radioactive" from "source is a soluble powder."