Toxic Oil Syndrome

Environmental MONDO:0016421 Pathograph 30 Show in embeddings browser

Toxic oil syndrome is a multisystem disease caused by eating rapeseed oil that had been denatured with aniline for industrial use, then illegally re-refined to strip the denaturant and sold in unlabelled containers as cooking oil. It appeared as a point-source epidemic in central and north-western Spain in May 1981 and affected roughly 20,000 people. The unifying lesion is a non-necrotising endovasculitis of vessels of every calibre in essentially every organ. Endothelial injury is followed by inflammatory infiltration of the intima, then by myointimal and fibroblastic proliferation that narrows or obliterates the lumen, with in situ thrombosis compounding the ischaemia. The same inflammatory process outside the vessel wall produces a lymphocytic perineuritis that ends in axonal degeneration, an interstitial myopathy, and the dermal and fascial fibrosis that makes the chronic phase look like scleroderma. The disease runs in phases. An acute respiratory illness with non-cardiogenic pulmonary oedema, fever, rash, eosinophilia and myalgia gives way over months to intense myalgia, hepatic dysfunction and pulmonary hypertension, and then to a chronic sclerodermiform stage with joint contractures, peripheral neuropathy, sicca syndrome and Raynaud phenomenon. Pulmonary arterial hypertension is still being diagnosed in exposed survivors decades later. The agent has never been definitively identified. Fatty acid esters of 3-(N-phenylamino)-1,2-propanediol carry the strongest epidemiological association with case-related oils, and aniline itself does not cause the illness, but no compound has been shown to reproduce the disease.

Ask OpenScientist

Ask a research question about Toxic Oil Syndrome. OpenScientist will conduct autonomous deep research using the Disorder Mechanisms Knowledge Base and PubMed literature (typically 10-30 minutes).

Submitting...

Do not include personal health information in your question. Questions and results are cached in your browser's local storage.

17
Pathophys.
22
Phenotypes
4
Hypotheses
4
Gaps
30
Pathograph
2
Genes
3
Medical Actions
3
Models
3
References
1
Deep Research
◈

Mechanistic Hypotheses

4
Direct toxic injury to vascular endothelium
direct_toxic_endothelial_injury CANONICAL
Evidence balance 1 support
The initiating event is a direct chemical effect of the ingested toxin, or of a metabolite, on the vascular endothelium - the pathology series that defines the lesion proposes free radicals as the mediator. On this account the immune response follows the endothelial lesion and perpetuates it rather than starting it.
Pathograph links
Systemic Exposure to PAP Fatty Acid Esters -> Vascular Endothelial Injury INDIRECT UNKNOWN INTERMEDIATES The central unproven step of the whole entry. No cited source traces a route from an absorbed PAP ester to an injured endothelial cell; the pathology series proposes free radicals and says so hypothetically. The intermediates are recorded as unknown rather than filled in.
Show evidence (1 reference)
PMID:1654352 SUPPORT PRIMARY RESULT Human Clinical
"A direct effect of unidentified toxic substances, possibly free radicals, may cause the endothelial lesion."
The pathologists' own proposal, and note its hedging - "may cause", with the substances unidentified. This is the hypothesis stated as a hypothesis by the people who described the lesion.
Immunological mechanism as the initiating event
immune_mediated_initiation ALTERNATIVE
Evidence balance 2 support
An HLA-restricted immune response to the ingested xenobiotic is the primary pathogenetic mechanism rather than a secondary amplifier. The strongest evidence is the demonstration of T-cell activation and a cytokine shift in lung tissue from affected patients, alongside HLA class I and class II associations and the transgenic-mouse work built on them. The two accounts are not mutually exclusive and differ mainly on what comes first.
Pathograph links
T Cell Activation with Th2 Skewing in Lung Tissue -> Perivascular and Interstitial Inflammatory Infiltration DIRECT
Show evidence (2 references)
PMID:11501225 SUPPORT REVIEW SYNTHESIS Human Clinical
"There is good evidence that the initial pathogenetic mechanism is immunological."
States the claim this group makes, and states it about the *initial* mechanism, which is precisely where it competes with the direct-toxicity account.
PMID:9074654 SUPPORT PRIMARY RESULT Human Clinical
"Data presented in this paper are the first clear evidence that an immunological mechanism is directly implicated in this illness."
The primary result the group rests on, in the authors' own summary of what their lung-tissue cytokine measurements establish.
PAP esters acting as platelet-activating factor analogues
paf_analogue_signalling EMERGING
Evidence balance 1 support
The PAP esters resemble platelet-activating factor structurally, and some of them perturb PAF synthesis. This would supply a route by which a compound absorbed from the gut could produce a systemic vascular and eosinophilic disease. It is an inference from structure plus a measured effect on PAF synthesis; no step of it has been shown to operate in a patient, and the abstract does not name the system the PAF-synthesis measurement was made in, so this entry does not claim one.
Pathograph links
Systemic Exposure to PAP Fatty Acid Esters -> Vascular Endothelial Injury INDIRECT UNKNOWN INTERMEDIATES The central unproven step of the whole entry. No cited source traces a route from an absorbed PAP ester to an injured endothelial cell; the pathology series proposes free radicals and says so hypothetically. The intermediates are recorded as unknown rather than filled in.
Show evidence (1 reference)
PMID:11768157 SUPPORT PRIMARY RESULT Other
"Some of these PAP esters, when a long acyl chain was present in the sn-1 position of the molecule, showed an inhibitory effect on the PAF synthesis."
The measured effect the hypothesis is built on. Note it is an *inhibitory* effect on PAF synthesis, not PAF-like agonism, so the structural analogy and the measured activity do not point the same way. Graded OTHER because the abstract does not say whether this measurement was made in animals or in vitro; the companion item's animal grading covers the absorption work rather than this one.
The rapeseed oil rather than its contaminants
rapeseed_oil_itself ALTERNATIVE
Evidence balance 1 support
A minority account from a cardiac pathology series: rapeseed oil itself remains a suspect, with the aniline reaction products acting only as a facilitator. It is recorded because it is a published dissent from the contaminant consensus, not because the entry finds it likely - the epidemiology ties illness to specific refined batches rather than to rapeseed oil consumption in general.
Show evidence (1 reference)
PMID:8001309 SUPPORT REVIEW SYNTHESIS Model Organism
"it is suggested that this oil must remain a major suspected cause of the toxic oil syndrome, particularly in conjunction with some as yet unexplained facilitative influence by oleoanilids"
The dissenting proposal in the authors' words. The sentence it continues begins "Based upon observations by others with experimental feeding of rapeseed oil", so the evidence behind the suggestion is other groups' animal feeding studies, not this paper's own human autopsy series - hence MODEL_ORGANISM for the evidence and REVIEW_SYNTHESIS for a publication MEDLINE types as a Review. "As yet unexplained facilitative influence" is the weak point the authors themselves flag.
?

Discussions and Knowledge Gaps

4
Which compound in the re-refined oil caused toxic oil syndrome, and by what molecular mechanism does it injure vascular endothelium?
KNOWLEDGE GAP unidentified_etiologic_agent
Attached to
pathophysiology#Systemic Exposure to PAP Fatty Acid Esters pathophysiology#Vascular Endothelial Injury
Forty years on, the agent is unidentified. The PAP esters carry the strongest association, their absorption and hepatic biotransformation are measured, and a metabolite is shared with eosinophilia-myalgia syndrome - but no compound has been shown to produce the disease, and the step from an absorbed ester to an injured endothelial cell is entirely unfilled. One review argues on metabolic grounds that there was never a single agent to find.
Show evidence (2 references)
PMID:12192735 SUPPORT REVIEW SYNTHESIS Human Clinical
"Although the exact identity of the etiologic agent in toxic oil syndrome remains unknown, work on toxic oil syndrome continues."
States the gap directly, twenty years after the epidemic.
PMID:11501225 SUPPORT REVIEW SYNTHESIS Human Clinical
"On metabolic evidence, it is suggested that not one, but a group of, toxic agents was responsible for TOS."
Raises the possibility that the search has been for the wrong kind of answer - a group of agents rather than one - which would explain why no single compound has been convicted.
Why has no animal species reproduced the pathology of toxic oil syndrome, and does that failure reflect a missing host factor rather than the wrong compound?
HUMAN MODEL MISMATCH no_animal_model
Attached to
pathophysiology#Vascular Endothelial Injury animal_models#Mouse
This is the load-bearing gap in the whole disease, and it is not a KNOWLEDGE_GAP: evidence in model systems exists, it simply does not reproduce the human lesion. Because no animal develops the endovasculitis, candidate compounds cannot be tested by administering them, which is why the etiology has rested on epidemiological association with oil batches for four decades. A review of the problem proposed screening species and strains predisposed to vasculitis, eosinophilia and raised IgE - an approach that assumes the missing ingredient is host susceptibility rather than the wrong compound. The human HLA associations make that assumption reasonable and do not establish it.
Show evidence (2 references)
PMID:12176082 SUPPORT REVIEW SYNTHESIS Other
"To date, pathology characteristics of toxic oil syndrome (TOS), a disease associated with consumption of a contaminated cooking oil in Spain in 1981, have not been reproduced in an animal model."
States the mismatch. MEDLINE types this publication as a Review and it performed no animal experiment of its own, so the quote is the field's state as this review reports it - REVIEW_SYNTHESIS, and OTHER rather than MODEL_ORGANISM, which would assert an animal study that does not exist here.
PMID:12176082 SUPPORT REVIEW SYNTHESIS Other
"The intent was to determine predisposed strains or species that potentially might be effective in testing the toxic oils and thus defining the precise identity of the toxic contaminant(s)."
Makes the dependency explicit - identifying the agent is blocked on having a model, which is why this gap and the etiology gap are the same problem from two sides. Graded as the same review's synthesis, for the reason given on the item above.
Is the cognitive impairment measured in toxic oil syndrome survivors four decades after exposure a direct neurotoxic sequela, or a downstream effect of chronic fatigue, mood symptoms and medication?
KNOWLEDGE GAP cognitive_deficit_mechanism
Attached to
phenotypes#Cognitive impairment phenotypes#Fatigue
Two 2025 case-control studies, each 50 survivors against 50 matched controls, pull in opposite directions and neither settles it. Whether they report the same 50 people is not stated in either abstract and is not assumed here. Survivors perform worse on attention, executive function, processing speed and global cognition, but the difference disappears once fatigue, depression, anxiety and CNS-acting medication are adjusted for, and fatigue mediates most of it. In parallel, neurofilament light chain was slightly higher in survivors on the raw group comparison (p = 0.025) while GFAP and pTau217 were not, and clinical status predicted none of the three once age, sex and education entered the models. So there is a measurable deficit with no biomarker correlate and a plausible non-neurotoxic explanation, which is three different things being unresolved at once rather than one.
Show evidence (2 references)
PMID:40507935 SUPPORT PRIMARY RESULT Human Clinical
"Clinical status (TOS vs. control) did not significantly predict biomarker concentrations in any model."
The negative biomarker result stated at its strongest and narrowest - no predictive effect of disease status in any model.
PMID:40507935 SUPPORT PRIMARY RESULT Human Clinical
"However, the possibility of subtle, compartmentalized, or slowly evolving neurotoxic processes cannot be excluded."
The authors' own limit on their negative result, quoted so the gap is recorded as open rather than closed against neurotoxicity.
What drives the dermal and fascial fibrosis of the chronic phase, given that the growth factors responsible in eosinophilia-myalgia syndrome are absent from toxic oil syndrome skin?
KNOWLEDGE GAP fibrosis_driver_unknown
Attached to
pathophysiology#Dermal and Fascial Fibrosis
The two epidemics are routinely discussed as near-identical, and the temptation is to carry the eosinophilia-myalgia syndrome mechanism across. An immunohistochemical comparison blocks that: TGF-beta and PDGF-AA were present in the periappendageal dermis of 57% of EMS specimens and 0% of toxic oil syndrome specimens. Whatever drives the fibrosis here is not the driver there, and it has not been identified. The specimen count is small, so this is a finding to explain rather than a settled negative.
Show evidence (1 reference)
PMID:8285738 SUPPORT PRIMARY RESULT Human Clinical
"Seven skin biopsy specimens from EMS, six skin biopsy specimens from toxic oil syndrome, nine muscle biopsy specimens from EMS, and one sural nerve biopsy specimen from EMS were studied."
The specimen counts, quoted because the gap's force depends on them: six TOS skin biopsies is a small denominator for a 0% finding, and the rationale says so rather than leaving the reader to assume otherwise.
⚙

Pathophysiology

17
Ingestion of Aniline-Denatured Rapeseed Oil
Rapeseed oil imported for industrial use was denatured with 2% aniline, then illegally re-refined to remove the denaturant and sold door to door as cooking oil. Aniline itself is not the toxin; the refining step is what generated the compounds that track with disease.
Downstream
Systemic Exposure to PAP Fatty Acid Esters The esters are products of the refining process, not of the denaturing or of storage, which is what ties them to the one refinery whose oil caused illness.
Show evidence (2 references)
PMID:12192735 SUPPORT REVIEW SYNTHESIS Human Clinical
"The vehicle of the causative toxic agent was identified as an illicit oil that had been diverted from industrial use and refined in order to remove the aniline denaturant, and that was sold in unlabeled 5-liter containers by itinerant salesmen."
The vehicle and the route by which it reached households, which is what this node claims.
PMID:1869734 SUPPORT REVIEW SYNTHESIS Human Clinical
"Aniline itself did not cause the illness, but the causal agent may be a reaction product of aniline with some oil component."
The reason this node is named for the denatured oil rather than for aniline exposure, and the reason no ECTO aniline-exposure term is bound anywhere in this entry.
Systemic Exposure to PAP Fatty Acid Esters
The di- and mono-oleyl esters of 3-(N-phenylamino)-1,2-propanediol are the compounds most strongly associated with case-related oil. They are hydrolysed by pancreatic lipase like a triglyceride, absorbed from the gut, and distributed to organs - which is what makes a systemic disease from an ingested compound.
Direct toxic injury to vascular endothelium PAP esters acting as platelet-activating factor analogues
response to xenobiotic stimulus GO:0009410 Gene Ontology (GO) Relation: this pathophysiological event involves this biological process This pathophysiological event involves response to xenobiotic stimulus (GO:0009410). GO:0009410 is a biological process from the Gene Ontology.
Downstream
Hepatic Biotransformation of PAP to 3-(Phenylamino)alanine
Vascular Endothelial Injury The central unproven step of the whole entry. No cited source traces a route from an absorbed PAP ester to an injured endothelial cell; the pathology series proposes free radicals and says so hypothetically. The intermediates are recorded as unknown rather than filled in. Direct toxic injury to vascular endothelium PAP esters acting as platelet-activating factor analogues
Show evidence (3 references)
PMID:10069247 SUPPORT PRIMARY RESULT Human Clinical
"We found the odds ratio for exposure to DPAP (OR = 26.4, 95% CI = 6.4-76.3) is much higher than the odds ratio for exposure to oleyl anilide (OR = 4.1, 95% CI = 2.2-7.8), implying that exposure to DPAP was a more relevant risk factor for development of toxic oil syndrome than exposure to oleyl anilide."
Quantifies why this node names the PAP esters and not the fatty acid anilides that were the earlier suspect, and gives both odds ratios so the comparison is visible rather than asserted.
PMID:11768157 SUPPORT PRIMARY RESULT Model Organism
"Results indicate that PAP esters are absorbed in the gastrointestinal tract and are distributed and stored in different organs, particularly in the liver and brown adipose tissue."
Supplies the absorption and distribution step this node asserts. Graded MODEL_ORGANISM because the absorption and tissue-distribution measurements were made in animals.
PMID:15257613 SUPPORT PRIMARY RESULT In Vitro
"Taken together, these results showed that PAP esters are substrates of hPL and that the two hydrolytic steps exhibit kinetic resolution in favor of the (S)-enantiomers."
Shows the esters enter normal lipid handling at the first step, which is how a compound eaten in trace amounts becomes systemically available.
Hepatic Biotransformation of PAP to 3-(Phenylamino)alanine
Liver tissue converts PAP to 3-(phenylamino)alanine, the aniline derivative independently isolated from the L-tryptophan implicated in eosinophilia-myalgia syndrome. This is the chemical link between the two epidemics, and it is a shared metabolite rather than a shared exposure.
response to xenobiotic stimulus GO:0009410 Gene Ontology (GO) Relation: this pathophysiological event involves this biological process This pathophysiological event involves response to xenobiotic stimulus (GO:0009410). GO:0009410 is a biological process from the Gene Ontology.
Show evidence (2 references)
PMID:8555405 SUPPORT PRIMARY RESULT In Vitro
"Here, we demonstrate the biotransformation of PAP into PAA by both rat hepatocytes and human liver tissue."
The reaction this node names, shown in human liver tissue as well as rat hepatocytes. IN_VITRO because both preparations are tissue and cells outside an organism.
PMID:8555405 SUPPORT INDIRECT PRIMARY RESULT In Vitro
"This finding is the first reported chemical link between TOS and EMS and suggests that these two related diseases share a common etiology, namely, PAA."
Why the node is worth having as its own step rather than folded into the exposure node. INDIRECT because a shared metabolite makes a common etiology plausible without establishing it - the authors write "suggests".
Vascular Endothelial Injury
The first lesion in the vessel wall, graded in the pathology series from endothelial cell swelling through to cell necrosis. Every later vascular step in this entry follows from it.
vascular endothelial cell CL:0002139 Cell Ontology (CL) Relation: this pathophysiological event involves this cell type This pathophysiological event involves vascular endothelial cell, annotated with endothelial cell of vascular tree (CL:0002139). CL:0002139 is a cell type from the Cell Ontology.
blood vessel UBERON:0001981 Uberon multi-species anatomy ontology (UBERON) Relation: this pathophysiological event occurs in this anatomical location This pathophysiological event occurs in blood vessel (UBERON:0001981). UBERON:0001981 is an anatomical location from the Uberon multi-species anatomy ontology.
Downstream
Perivascular and Interstitial Inflammatory Infiltration
Show evidence (2 references)
PMID:1654352 SUPPORT PRIMARY RESULT Human Clinical
"The vascular lesions begins with endothelial damage that varies from cellular swelling to cellular necrosis."
Names this as the first lesion and gives its range. Quoted with the source's own grammatical slip intact, as a snippet must be.
PMID:9074654 SUPPORT BACKGROUND Human Clinical
"The pathologic findings in TOS showed primary endothelial injury, with cell proliferation and perivascular inflammatory infiltrates."
An independent statement that the endothelial injury is primary. BACKGROUND because it is this cytokine paper's framing of established pathology rather than its own measurement.
T Cell Activation with Th2 Skewing in Lung Tissue
Lung tissue from affected patients shows T-cell activation with both Th1 and Th2 cytokine profiles raised, and the Th2 increment over Th1 is itself significant. The response is not purely Th2 - both arms are up - and the entry says so rather than simplifying to a Th2 disease.
Immunological mechanism as the initiating event
T cell CL:0000084 Cell Ontology (CL) Relation: this pathophysiological event involves this cell type This pathophysiological event involves T cell (CL:0000084). CL:0000084 is a cell type from the Cell Ontology. T-helper 2 cell CL:0000546 Cell Ontology (CL) Relation: this pathophysiological event involves this cell type This pathophysiological event involves T-helper 2 cell (CL:0000546). CL:0000546 is a cell type from the Cell Ontology.
T cell activation GO:0042110 Gene Ontology (GO) Relation: this pathophysiological event involves this biological process This pathophysiological event involves increased T cell activation (GO:0042110). GO:0042110 is a biological process from the Gene Ontology. ↑ INCREASED type 2 immune response GO:0042092 Gene Ontology (GO) Relation: this pathophysiological event involves this biological process This pathophysiological event involves increased type 2 immune response (GO:0042092). GO:0042092 is a biological process from the Gene Ontology. ↑ INCREASED
lung UBERON:0002048 Uberon multi-species anatomy ontology (UBERON) Relation: this pathophysiological event occurs in this anatomical location This pathophysiological event occurs in lung (UBERON:0002048). UBERON:0002048 is an anatomical location from the Uberon multi-species anatomy ontology.
Animal models
HLA-DR2/DQ6 transgenic mouse fed toxic oil Mouse partially recapitulates Reproduces the HLA-dependent immunological readouts, and nothing else about the disease.
Downstream
Peripheral Eosinophilia
Perivascular and Interstitial Inflammatory Infiltration Immunological mechanism as the initiating event
Show evidence (2 references)
PMID:9074654 SUPPORT PRIMARY RESULT Human Clinical
"We found a significant increase in Th1 (P = 0.006) and Th2 (P = 0.003) cytokine profile in TOS patients with respect to controls."
Both arms rose, and the quote is chosen to carry that rather than the Th2 figure alone - the node's description depends on it.
PMID:9074654 SUPPORT PRIMARY RESULT Human Clinical
"The increment in TH2 response with respect to TH1 is significant (P = 0.03) in TOS lung specimens."
The skew itself, which is the part of the claim the previous quote does not establish.
Peripheral Eosinophilia
Marked peripheral eosinophilia is one of the two features that made the epidemic recognisable as a single disease, alongside incapacitating myalgia. Eosinophils are also found in the vessel wall infiltrate in some cases.
eosinophil CL:0000771 Cell Ontology (CL) Relation: this pathophysiological event involves this cell type This pathophysiological event involves eosinophil (CL:0000771). CL:0000771 is a cell type from the Cell Ontology.
eosinophil chemotaxis GO:0048245 Gene Ontology (GO) Relation: this pathophysiological event involves this biological process This pathophysiological event involves increased eosinophil chemotaxis (GO:0048245). GO:0048245 is a biological process from the Gene Ontology. ↑ INCREASED
Animal models
Mouse given fatty acid anilides and the linoleic PAP diester Mouse partially recapitulates Reproduces the blood eosinophilia and lung pathology of the acute phase from the candidate compounds, without the vasculitis that defines the disease.
Downstream
Increased total eosinophil count
Perivascular and Interstitial Inflammatory Infiltration The eosinophil-rich subset of the infiltrate. The pathology series is careful that this is "some cases" rather than the rule, and the edge claims no more than that.
Show evidence (1 reference)
PMID:8266108 SUPPORT REVIEW SYNTHESIS Human Clinical
"Both illnesses affect patients clinically by causing intense, incapacitating myalgias and a marked peripheral eosinophilia."
The WHO workshop's statement of the two defining clinical features. It is a comparative sentence covering TOS and EMS together, which is how the source makes the claim.
Perivascular and Interstitial Inflammatory Infiltration
The mixed cellular infiltrate that follows endothelial injury, in the intima and sometimes the media and adventitia, and more widely in the interstitium of affected organs. It is the branch point of the entry: the vascular, neural, muscular and cutaneous arms all descend from it.
T cell CL:0000084 Cell Ontology (CL) Relation: this pathophysiological event involves this cell type This pathophysiological event involves T cell (CL:0000084). CL:0000084 is a cell type from the Cell Ontology. eosinophil CL:0000771 Cell Ontology (CL) Relation: this pathophysiological event involves this cell type This pathophysiological event involves eosinophil (CL:0000771). CL:0000771 is a cell type from the Cell Ontology.
leukocyte migration GO:0050900 Gene Ontology (GO) Relation: this pathophysiological event involves this biological process This pathophysiological event involves increased leukocyte migration (GO:0050900). GO:0050900 is a biological process from the Gene Ontology. ↑ INCREASED
tunica intima UBERON:0002523 Uberon multi-species anatomy ontology (UBERON) Relation: this pathophysiological event occurs in this anatomical location This pathophysiological event occurs in tunica intima (UBERON:0002523). UBERON:0002523 is an anatomical location from the Uberon multi-species anatomy ontology.
Downstream
Myointimal and Fibroblastic Proliferation
Lymphocytic Perineuritis
Interstitial Inflammatory Myopathy
Dermal and Fascial Fibrosis Drawn with unknown intermediates deliberately. The fibrogenic growth factors that would supply the intermediates in the comparable disease were looked for in TOS skin and were not found - see the REFUTE item on the target node.
Pulmonary infiltrates The radiographic infiltrate of the acute phase is this same interstitial inflammatory process seen in the lung.
Show evidence (2 references)
PMID:8412642 SUPPORT PRIMARY RESULT Human Clinical
"The most important pathologic features of TOS were widespread interstitial infiltrates, non-necrotizing angiitis, endothelial proliferation, and tissue fibrosis."
Names the infiltrate as widespread and interstitial, not only intimal, which is what lets this one node serve as the branch point for the non-vascular arms.
PMID:1654352 SUPPORT PRIMARY RESULT Human Clinical
"Vessels of every type and size are involved, affecting practically every organ."
The distribution of the lesion, which is why this entry has arms reaching lung, nerve, muscle and skin rather than one target organ.
Myointimal and Fibroblastic Proliferation
Proliferation of myointimal cells, joined in advanced lesions by fibroblasts. This is what converts an inflamed vessel into a narrowed one.
myointimal smooth muscle cell CL:0000192 Cell Ontology (CL) Relation: this pathophysiological event involves this cell type This pathophysiological event involves myointimal smooth muscle cell, annotated with smooth muscle cell (CL:0000192). CL:0000192 is a cell type from the Cell Ontology. fibroblast CL:0000057 Cell Ontology (CL) Relation: this pathophysiological event involves this cell type This pathophysiological event involves fibroblast (CL:0000057). CL:0000057 is a cell type from the Cell Ontology.
smooth muscle cell proliferation GO:0048659 Gene Ontology (GO) Relation: this pathophysiological event involves this biological process This pathophysiological event involves increased smooth muscle cell proliferation (GO:0048659). GO:0048659 is a biological process from the Gene Ontology. ↑ INCREASED
Downstream
Vascular Luminal Narrowing and Obliteration
Show evidence (1 reference)
PMID:1654352 SUPPORT PRIMARY RESULT Human Clinical
"Proliferation of myointimal cells and in advanced stages fibroblastic proliferation causes narrowing or obliteration of the vascular lumen."
Names both cell populations, their sequence, and the consequence, which is this node and its single downstream edge in one sentence.
Vascular Luminal Narrowing and Obliteration
The narrowed or obliterated lumen that the proliferative response leaves behind, and the substrate on which thrombosis forms.
blood vessel UBERON:0001981 Uberon multi-species anatomy ontology (UBERON) Relation: this pathophysiological event occurs in this anatomical location This pathophysiological event occurs in blood vessel (UBERON:0001981). UBERON:0001981 is an anatomical location from the Uberon multi-species anatomy ontology.
Downstream
In Situ Thrombosis on the Damaged Vessel Wall
Pulmonary Vascular Remodelling and Plexiform Lesion Formation
Show evidence (1 reference)
PMID:3961509 SUPPORT PRIMARY RESULT Human Clinical
"Biopsy studies during this stage showed fibrosis and obliterating arteriopathy."
An independent biopsy series reaching the same obliterative endpoint, and dating it to the sclerodermiform stage.
In Situ Thrombosis on the Damaged Vessel Wall
Thrombosis on the damaged and narrowed vessel wall. The pathology series is explicit that this feeds back on the vascular lesion rather than simply following it.
blood coagulation GO:0007596 Gene Ontology (GO) Relation: this pathophysiological event involves this biological process This pathophysiological event involves increased blood coagulation (GO:0007596). GO:0007596 is a biological process from the Gene Ontology. ↑ INCREASED
Animal models
Rat given 3-phenylamino-1,2-propanediol Rat fails to recapitulate Intraperitoneal PAP caused massive pulmonary thromboembolism in rats, but by thrombosis in mesenteric vessels that then embolised - not by the endovasculitis of the human disease.
Downstream
Tissue Ischemia and Parenchymal Atrophy
Thromboembolism
Show evidence (1 reference)
PMID:1654352 SUPPORT PRIMARY RESULT Human Clinical
"Thromboembolic complications perpetuate the vascular lesion and compound the ischemia and parenchymal atrophy of several organs."
Carries this node and both of its downstream targets, and states the feedback onto the vascular lesion that the description notes.
Tissue Ischemia and Parenchymal Atrophy
The downstream organ consequence of a progressively obliterated vascular bed, reported across several organs rather than confined to one.
Pulmonary Vascular Remodelling and Plexiform Lesion Formation
In the lung the same process produces plexiform lesions, thromboses and venous lesions - a pulmonary vasculopathy that autopsy studies found indistinguishable from primary pulmonary hypertension. Plexiform lesions appear late.
blood vessel remodeling GO:0001974 Gene Ontology (GO) Relation: this pathophysiological event involves this biological process This pathophysiological event involves increased blood vessel remodeling (GO:0001974). GO:0001974 is a biological process from the Gene Ontology. ↑ INCREASED
lung UBERON:0002048 Uberon multi-species anatomy ontology (UBERON) Relation: this pathophysiological event occurs in this anatomical location This pathophysiological event occurs in lung (UBERON:0002048). UBERON:0002048 is an anatomical location from the Uberon multi-species anatomy ontology.
Downstream
Pulmonary arterial hypertension
Show evidence (3 references)
PMID:2914483 SUPPORT PRIMARY RESULT Human Clinical
"The main pathologic pulmonary vascular findings consisted of plexiform lesions, thromboses, and venous lesions."
The three lesions this node names, from the autopsy series of patients who died of the pulmonary hypertension.
PMID:2914483 SUPPORT PRIMARY RESULT Human Clinical
"Endothelial damage induced by the toxic agents is suggested as an initial causative mechanism, perpetuated by intimal proliferation and in situ thrombosis."
Reaches the same ordering as the extracardiac pathology series in a different organ and a different cohort, which is why the entry models one vascular chain rather than an organ-specific one.
PMID:6648850 SUPPORT PRIMARY RESULT Human Clinical
"In muscular pulmonary arteries there was pronounced medial hypertrophy and intimal proliferation, which was so severe in one case that it completely occluded the arterial lumen."
The earliest necropsy description of the pulmonary arterial lesion, reporting the same intimal proliferation the systemic vessels show and carrying it through to complete occlusion.
Lymphocytic Perineuritis
The nerve lesion begins as an inflammatory neuropathy with lymphocytic infiltration of the perineurium, and progresses to perineural fibrosis.
T cell CL:0000084 Cell Ontology (CL) Relation: this pathophysiological event involves this cell type This pathophysiological event involves T cell (CL:0000084). CL:0000084 is a cell type from the Cell Ontology.
Downstream
Secondary Axonal Degeneration
Show evidence (1 reference)
PMID:1654352 SUPPORT PRIMARY RESULT Human Clinical
"The peripheral nerve lesions begin with an inflammatory neuropathy with lymphocytic perineuritis and progress to perineural fibrosis with secondary axonal degeneration."
Carries the node, its progression, and the edge to axonal degeneration - and the word "secondary" is what makes the ordering a claim rather than an inference.
Secondary Axonal Degeneration
Axonal loss following the perineural lesion rather than arising independently, which is what the pathology series means by calling it secondary.
Downstream
Peripheral neuropathy
Neurogenic Muscular Atrophy
Interstitial Inflammatory Myopathy
The first muscle lesion, an interstitial inflammatory myopathy, which is later replaced by denervation atrophy rather than progressing as a primary myopathy.
Show evidence (1 reference)
PMID:1654352 SUPPORT PRIMARY RESULT Human Clinical
"Skeletal muscle lesions exhibit an interstitial inflammatory myopathy at first, followed by a neurogenic muscular atrophy."
Both muscle nodes and their order in one sentence, including that the later atrophy is neurogenic - which is why the entry routes it from the nerve arm rather than from this node.
Neurogenic Muscular Atrophy
Denervation atrophy following axonal degeneration. It is placed downstream of the nerve lesion, not of the myopathy, because the pathology series calls it neurogenic.
Downstream
Muscle weakness
Dermal and Fascial Fibrosis
The fibrosis that gives the chronic phase its sclerodermiform appearance. What drives it is not established: the growth factors that drive the comparable lesion in eosinophilia-myalgia syndrome were absent from every TOS skin specimen examined.
fibroblast CL:0000057 Cell Ontology (CL) Relation: this pathophysiological event involves this cell type This pathophysiological event involves fibroblast (CL:0000057). CL:0000057 is a cell type from the Cell Ontology.
extracellular matrix organization GO:0030198 Gene Ontology (GO) Relation: this pathophysiological event involves this biological process This pathophysiological event involves increased extracellular matrix organization (GO:0030198). GO:0030198 is a biological process from the Gene Ontology. ↑ INCREASED
Downstream
Scleroderma
Joint contracture
Show evidence (3 references)
PMID:3961509 SUPPORT PRIMARY RESULT Human Clinical
"TOS is a new chemically induced scleroderma-like syndrome with features overlapping those of eosinophilic fasciitis, systemic sclerosis, and forms of localized scleroderma."
Places the fibrotic phenotype among the sclerodermas without collapsing it into any one of them.
PMID:8285738 REFUTE PRIMARY RESULT Human Clinical
"The presence of TGF-beta and platelet-derived growth factorAA in the periappendageal dermis was significantly more prevalent in EMS than toxic oil syndrome (57% vs 0%)."
REFUTE against the obvious mechanistic account of this node. TGF-beta and PDGF-AA are the standard fibrogenic drivers and were found in zero of the TOS skin specimens. The deep-research report cited this same paper as evidence that those growth factors drive TOS fibrosis, which inverts what it found.
PMID:8285738 SUPPORT PRIMARY RESULT Human Clinical
"suggest that the pathogenesis of tissue fibrosis in EMS and toxic oil syndrome may be dependent on different growth factors"
The authors' own reading of that negative result, and the reason this node's upstream edge carries unknown intermediates rather than borrowing the EMS mechanism.
⬡

Pathograph

Use the checkboxes to hide or show graph categories. Hover nodes for evidence and cross-linked metadata.
Pathograph: causal mechanism network for Toxic Oil Syndrome Interactive directed graph showing how pathophysiology mechanisms, phenotypes, genetic factors and variants, experimental models, environmental triggers, and treatments relate through causal and linked edges.
●

Phenotypes

22
Blood 2
Increased total eosinophil count HP:0001880 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Peripheral eosinophilia, annotated with Increased total eosinophil count (HP:0001880). HP:0001880 is a phenotype from the Human Phenotype Ontology.
Show evidence (1 reference)
PMID:8412642 SUPPORT PRIMARY RESULT Human Clinical
"The acute phase lasted 2 months, and was characterized by pulmonary edema, rash, eosinophilia, and myalgia."
Names eosinophilia as an acute-phase feature in the follow-up cohort.
Thromboembolism HP:0001907 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Thromboembolism (HP:0001907). HP:0001907 is a phenotype from the Human Phenotype Ontology.
Show evidence (1 reference)
PMID:1654352 SUPPORT PRIMARY RESULT Human Clinical
"Thromboembolic complications perpetuate the vascular lesion and compound the ischemia and parenchymal atrophy of several organs."
Names thromboembolic complications as a feature of the vascular disease.
Cardiovascular 2
Pulmonary arterial hypertension HP:0002092 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Pulmonary arterial hypertension (HP:0002092), qualified as course progressive. HP:0002092 is a phenotype from the Human Phenotype Ontology.
Course: PROGRESSIVE
Sequelae: Dyspnea
Show evidence (2 references)
PMID:8412642 SUPPORT PRIMARY RESULT Human Clinical
"During the intermediate phase (second to fourth months), severe myalgia, skin tenderness, subcutaneous edema, altered liver function, and pulmonary hypertension developed."
Dates the onset of pulmonary hypertension to the intermediate phase.
PMID:41473551 SUPPORT PRIMARY RESULT Human Clinical
"TOS-PAH remains a distinct clinical entity, with new cases diagnosed decades after toxic exposure."
A registry study's conclusion that the pulmonary vascular disease is still declaring itself in this cohort, which is why the phenotype carries a progressive course rather than being scoped to the 1980s.
Raynaud phenomenon HP:0030880 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Raynaud phenomenon (HP:0030880), qualified as temporality chronic. HP:0030880 is a phenotype from the Human Phenotype Ontology.
Temporal: CHRONIC
Show evidence (1 reference)
PMID:8412642 SUPPORT PRIMARY RESULT Human Clinical
"Its most prominent clinical features were muscle cramps, chronic musculoskeletal pain, chronic lung disease, Raynaud phenomenon, carpal tunnel syndrome, and psychologic disturbances."
Names Raynaud phenomenon among the most prominent features of the late chronic phase.
Digestive 1
Abnormal liver physiology HP:0031865 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Altered liver function, annotated with Abnormal liver physiology (HP:0031865). HP:0031865 is a phenotype from the Human Phenotype Ontology.
Show evidence (2 references)
PMID:8412642 SUPPORT PRIMARY RESULT Human Clinical
"During the intermediate phase (second to fourth months), severe myalgia, skin tenderness, subcutaneous edema, altered liver function, and pulmonary hypertension developed."
Names altered liver function among the intermediate-phase features.
PMID:3961509 SUPPORT PRIMARY RESULT Human Clinical
"They subsequently developed peripheral neuropathy, joint contractures, scleroderma-like changes, Raynaud phenomenon, pulmonary hypertension, sicca syndrome, and liver disease."
An independent series carrying liver disease into the chronic phase alongside the sclerodermiform features.
Head and Neck 1
Xerostomia HP:0000217 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Dry mouth, annotated with Xerostomia (HP:0000217), qualified as temporality chronic. HP:0000217 is a phenotype from the Human Phenotype Ontology.
Temporal: CHRONIC
Show evidence (1 reference)
PMID:8412642 SUPPORT PRIMARY RESULT Human Clinical
"It was marked by scleroderma, sicca syndrome, polyneuropathy, joint contractures, weight loss, and functional limitations."
Names sicca syndrome in the early chronic phase. The HP binding is Xerostomia because HPO has no sicca-syndrome term and dry mouth is the component this entry can name.
Immune 1
Skin rash HP:0000988 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Skin rash (HP:0000988), qualified as temporality acute. HP:0000988 is a phenotype from the Human Phenotype Ontology.
Temporal: ACUTE
Show evidence (1 reference)
PMID:8412642 SUPPORT PRIMARY RESULT Human Clinical
"The acute phase lasted 2 months, and was characterized by pulmonary edema, rash, eosinophilia, and myalgia."
Names rash as one of the four defining acute-phase features.
Integument 2
Scleroderma HP:0100324 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Scleroderma-like skin change, annotated with Scleroderma (HP:0100324), qualified as temporality chronic. HP:0100324 is a phenotype from the Human Phenotype Ontology.
Temporal: CHRONIC
Show evidence (1 reference)
PMID:3961509 SUPPORT PRIMARY RESULT Human Clinical
"They subsequently developed peripheral neuropathy, joint contractures, scleroderma-like changes, Raynaud phenomenon, pulmonary hypertension, sicca syndrome, and liver disease."
The chronic-phase sequence in a series selected for scleroderma-like change. The binding was narrowed to match: an earlier draft bound HP:0011838 Sclerodactyly and defended the digital specificity by pointing at the same abstract's "digital tuft changes", which is acro-osteolysis rather than skin sclerosis of the digits and does not support it. HP:0100324 Scleroderma is what the quoted sentence actually says.
Alopecia HP:0001596 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Alopecia (HP:0001596). HP:0001596 is a phenotype from the Human Phenotype Ontology.
Show evidence (1 reference)
PMID:1869734 SUPPORT REVIEW SYNTHESIS Human Clinical
"the remaining patients developed an intermediate or chronic phase, or both, of illness involving severe myalgia, eosinophilia, peripheral nerve damage, sclerodermiform skin lesions, sicca syndrome, alopecia and joint contractures, among other findings"
Names alopecia among the intermediate and chronic phase features.
Metabolism 3
Fever HP:0001945 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Fever (HP:0001945), qualified as temporality acute. HP:0001945 is a phenotype from the Human Phenotype Ontology.
Temporal: ACUTE
Show evidence (1 reference)
PMID:1869734 SUPPORT REVIEW SYNTHESIS Human Clinical
"patients presented primarily with a respiratory syndrome involving cough, fever, dyspnea, hypoxemia, pulmonary infiltrates and pleural effusions"
Names fever among the acute presenting features.
Edema HP:0000969 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Subcutaneous edema, annotated with Edema (HP:0000969). HP:0000969 is a phenotype from the Human Phenotype Ontology.
Show evidence (1 reference)
PMID:8412642 SUPPORT PRIMARY RESULT Human Clinical
"During the intermediate phase (second to fourth months), severe myalgia, skin tenderness, subcutaneous edema, altered liver function, and pulmonary hypertension developed."
Names subcutaneous oedema among the intermediate-phase features. Bound to the general HP:0000969 Edema rather than HP:0012398 Peripheral edema: the source says subcutaneous, which is a tissue plane, not a distribution, and narrowing to peripheral would assert a limb predominance the sentence does not carry.
Pleural effusion HP:0002202 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Pleural effusion (HP:0002202), qualified as temporality acute. HP:0002202 is a phenotype from the Human Phenotype Ontology.
Temporal: ACUTE
Show evidence (1 reference)
PMID:1869734 SUPPORT REVIEW SYNTHESIS Human Clinical
"patients presented primarily with a respiratory syndrome involving cough, fever, dyspnea, hypoxemia, pulmonary infiltrates and pleural effusions"
Names pleural effusions among the acute presenting features.
Musculoskeletal 2
Joint contracture HP:0034392 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Joint contracture (HP:0034392), qualified as temporality chronic. HP:0034392 is a phenotype from the Human Phenotype Ontology.
Temporal: CHRONIC
Show evidence (1 reference)
PMID:8266108 SUPPORT REVIEW SYNTHESIS Human Clinical
"Long-term complications include pulmonary hypertension, peripheral neuropathies, and joint contractures."
Names joint contractures among the three long-term complications.
Muscle weakness HP:0001324 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Muscle weakness (HP:0001324). HP:0001324 is a phenotype from the Human Phenotype Ontology.
Show evidence (1 reference)
PMID:1654352 SUPPORT INDIRECT PRIMARY RESULT Human Clinical
"Skeletal muscle lesions exhibit an interstitial inflammatory myopathy at first, followed by a neurogenic muscular atrophy."
INDIRECT because the source reports the muscle pathology rather than the clinical weakness; the weakness follows from neurogenic atrophy by an inference step this entry makes explicit.
Nervous System 3
Peripheral neuropathy HP:0009830 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Peripheral neuropathy (HP:0009830), qualified as temporality chronic. HP:0009830 is a phenotype from the Human Phenotype Ontology.
Temporal: CHRONIC
Show evidence (1 reference)
PMID:8266108 SUPPORT REVIEW SYNTHESIS Human Clinical
"Long-term complications include pulmonary hypertension, peripheral neuropathies, and joint contractures."
Names peripheral neuropathy among the three long-term complications.
Cognitive impairment HP:0100543 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Cognitive impairment (HP:0100543), qualified as temporality chronic. HP:0100543 is a phenotype from the Human Phenotype Ontology.
Temporal: CHRONIC
Show evidence (2 references)
PMID:40507507 SUPPORT PRIMARY RESULT Human Clinical
"TOS survivors showed significantly poorer performance than controls in attention, executive function, processing speed, and global cognition after adjusting for demographic and vascular risk factors."
The measured deficit across four domains in a matched case-control design, adjusted for demographic and vascular confounders.
PMID:40507507 REFUTE PRIMARY RESULT Human Clinical
"However, these differences were no longer statistically significant after additional adjustment for fatigue, depression, anxiety, and central nervous system-acting medications."
REFUTE against reading the deficit as an independent neurocognitive sequela. The same study reports both results, so both are carried; taking only the first sentence would state the finding as the authors specifically declined to state it.
Constrictive median neuropathy HP:0012185 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Carpal tunnel syndrome, annotated with Constrictive median neuropathy (HP:0012185), qualified as temporality chronic. HP:0012185 is a phenotype from the Human Phenotype Ontology.
Temporal: CHRONIC
Show evidence (1 reference)
PMID:8412642 SUPPORT PRIMARY RESULT Human Clinical
"Its most prominent clinical features were muscle cramps, chronic musculoskeletal pain, chronic lung disease, Raynaud phenomenon, carpal tunnel syndrome, and psychologic disturbances."
Names carpal tunnel syndrome among the late-phase features. HPO files this as Constrictive median neuropathy, which carries "Carpal tunnel syndrome" as a synonym.
Respiratory 2
Pulmonary infiltrates HP:0002113 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Pulmonary infiltrates (HP:0002113), qualified as temporality acute. HP:0002113 is a phenotype from the Human Phenotype Ontology.
Temporal: ACUTE
Show evidence (1 reference)
PMID:1869734 SUPPORT REVIEW SYNTHESIS Human Clinical
"patients presented primarily with a respiratory syndrome involving cough, fever, dyspnea, hypoxemia, pulmonary infiltrates and pleural effusions"
The acute presenting syndrome, naming the infiltrates directly.
Dyspnea HP:0002094 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Dyspnea (HP:0002094). HP:0002094 is a phenotype from the Human Phenotype Ontology.
Show evidence (1 reference)
PMID:1869734 SUPPORT REVIEW SYNTHESIS Human Clinical
"patients presented primarily with a respiratory syndrome involving cough, fever, dyspnea, hypoxemia, pulmonary infiltrates and pleural effusions"
Names dyspnoea among the acute presenting features.
Constitutional 2
Myalgia HP:0003326 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Myalgia (HP:0003326), qualified as severity severe. HP:0003326 is a phenotype from the Human Phenotype Ontology.
Severity: SEVERE
Show evidence (1 reference)
PMID:8412642 SUPPORT PRIMARY RESULT Human Clinical
"During the intermediate phase (second to fourth months), severe myalgia, skin tenderness, subcutaneous edema, altered liver function, and pulmonary hypertension developed."
Places severe myalgia in the intermediate phase and names the other features that accompany it there.
Fatigue HP:0012378 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Fatigue (HP:0012378), qualified as temporality chronic. HP:0012378 is a phenotype from the Human Phenotype Ontology.
Temporal: CHRONIC
Show evidence (1 reference)
PMID:40507507 SUPPORT PRIMARY RESULT Human Clinical
"Structural equation modeling analyses revealed that affective symptoms-particularly fatigue-substantially mediated the relationship between TOS and cognitive performance."
Establishes both that fatigue is present in the cohort and the mediating role described above. No graph edge records that mediation - see this phenotype's description for why.
Growth 1
Weight loss HP:0001824 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Weight loss (HP:0001824), qualified as temporality chronic. HP:0001824 is a phenotype from the Human Phenotype Ontology.
Temporal: CHRONIC
Show evidence (1 reference)
PMID:8412642 SUPPORT PRIMARY RESULT Human Clinical
"It was marked by scleroderma, sicca syndrome, polyneuropathy, joint contractures, weight loss, and functional limitations."
Names weight loss in the early chronic phase.
🧬

Genetic Associations

2
HLA class II association with fatal disease
Gene: HLA-DRB1 hgnc:4948 HUGO Gene Nomenclature Committee (hgnc) Relation: this disease-associated gene is this gene This disease-associated gene is HLA-DRB1 (hgnc:4948). hgnc:4948 is a gene from the HUGO Gene Nomenclature Committee. relationship_type: MODIFIER
Show evidence (2 references)
PMID:10746782 SUPPORT PRIMARY RESULT Human Clinical
"In contrast, an increase in phenotypic frequency of DR2 antigen, was found in patients who had died from TOS (73.5%)"
The positive finding, and it is specifically about patients who died rather than about susceptibility to the disease.
PMID:10746782 SUPPORT PRIMARY RESULT Human Clinical
"Regarding surviving patients no significant association was found between HLA and disease."
SUPPORT, not REFUTE, and the distinction is the whole point of this record. The absence of any association among survivors is what confines the DR2 finding to fatal disease, which is exactly what this record claims and why it is typed MODIFIER. The same sentence is carried as REFUTE on the susceptibility record below, where it does cut against the claim being made.
HLA class I and DR4-DQ8 association with susceptibility
Gene: HLA-A hgnc:4931 HUGO Gene Nomenclature Committee (hgnc) Relation: this disease-associated gene is this gene This disease-associated gene is HLA-A (hgnc:4931). hgnc:4931 is a gene from the HUGO Gene Nomenclature Committee. relationship_type: RISK_FACTOR
Show evidence (2 references)
PMID:8803534 SUPPORT PRIMARY RESULT Human Clinical
"In this paper it is also established that a human class I antigen (HLA-A24) and, independently, an HLA class II haplotype (DR4-DQ8, Pcorrected = 0.04) and arginine 52 in the alpha-DQ chains (Pcorrected = 0.03) are associated with TOS susceptibility, similarly to insulin-dependent diabetes."
The reported associations with their corrected p-values. The gene binding is HLA-A because A24 is a serotype of that locus; allele-level detail in a gene field follows the HLA-B27 precedent in this repository.
PMID:10746782 REFUTE PRIMARY RESULT Human Clinical
"Regarding surviving patients no significant association was found between HLA and disease."
REFUTE against this record's susceptibility claim. A later, larger case-control study with family and unrelated controls found no HLA association among surviving patients at all, which is the population a susceptibility claim is about.
💊

Medical Actions

3
Corticosteroid Therapy
Action: PharmacotherapyNCI Thesaurus (NCIT) Relation: this treatment is this clinical intervention This treatment is Pharmacotherapy (NCIT:C15986). NCIT:C15986 is a clinical intervention from the NCI Thesaurus. NCIT:C15986
Agent: corticosteroid CHEBI:50858 Chemical Entities of Biological Interest (CHEBI) Relation: this treatment uses this therapeutic agent This treatment uses corticosteroid (CHEBI:50858). CHEBI:50858 is a therapeutic agent from Chemical Entities of Biological Interest.
Platform: Small molecule
Used in the acute and subacute phases. It relieves symptoms without changing the course of the disease, which is the conclusion the WHO workshop reached for both TOS and eosinophilia-myalgia syndrome.
Show evidence (2 references)
PMID:8266108 SUPPORT REVIEW SYNTHESIS Human Clinical
"Although treatment with corticosteroids has resulted in significant symptomatic relief in persons with either disorder, it does not alter the clinical course or long-term outcome."
Supports the symptomatic benefit this treatment claims. The same sentence is carried again below as REFUTE against disease modification, because it makes both claims at once.
PMID:8266108 REFUTE REVIEW SYNTHESIS Human Clinical
"Although treatment with corticosteroids has resulted in significant symptomatic relief in persons with either disorder, it does not alter the clinical course or long-term outcome."
REFUTE against corticosteroids as disease-modifying therapy. There is deliberately no `target_mechanisms` edge: on this evidence the drug does not act on the pathograph, and recording one would assert an effect the only cited source denies.
Supportive Care
Action: Supportive CareNCI Thesaurus (NCIT) Relation: this treatment is this clinical intervention This treatment is Supportive Care (NCIT:C15747). NCIT:C15747 is a clinical intervention from the NCI Thesaurus. NCIT:C15747
The mainstay throughout: respiratory support in the acute pulmonary oedema, and management of the chronic complications. An earlier draft said no intervention trialled during or after the epidemic altered the course of the disease. Only corticosteroids are cache-backed for that claim here, so it is narrowed - the other agents the literature reports as trialled without benefit (azathioprine, penicillamine, plasmapheresis, vitamin E, superoxide dismutase) are not cited in this entry and are not asserted by it.
Show evidence (1 reference)
PMID:8266108 SUPPORT INDIRECT REVIEW SYNTHESIS Human Clinical
"Research into the etiologic agents, preferred treatments, and ways to avoid similar problems in the future is needed."
INDIRECT, and quoted for what it concedes: a decade after the epidemic the WHO workshop still listed preferred treatment as an open research question, which is the situation that leaves supportive care as the mainstay. It is not a study of supportive care.
Rehabilitation
Action: RehabilitationNCI Thesaurus (NCIT) Relation: this treatment is this clinical intervention This treatment is Rehabilitation (NCIT:C15315). NCIT:C15315 is a clinical intervention from the NCI Thesaurus. NCIT:C15315
Platform: Behavioral / lifestyle
For the joint contractures and the neuropathic disability of the chronic phase, which are among the three complications the WHO workshop named as long-term.
Show evidence (1 reference)
PMID:8266108 SUPPORT INDIRECT REVIEW SYNTHESIS Human Clinical
"Long-term complications include pulmonary hypertension, peripheral neuropathies, and joint contractures."
INDIRECT: the source establishes the disabling complications that rehabilitation addresses, not the efficacy of rehabilitation in this disease, which no cited source reports.
🌍

Environmental Factors

1
Ingestion of aniline-denatured rapeseed oil sold as cooking oil
exposure to a food adulterated with an industrial denaturant ECTO:9002126 Environmental Conditions, Treatments and Exposures Ontology (ECTO) Relation: this environmental factor is this exposure This environmental factor is exposure to a food adulterated with an industrial denaturant, annotated with exposure to environmental food contaminant (ECTO:9002126). ECTO:9002126 is an exposure from the Environmental Conditions, Treatments and Exposures Ontology.
ECTO has an `exposure to aniline` term (`ECTO:9000980`, found with `runoak -i ols:ecto search "exposure to aniline"`), and it is deliberately not used: `PMID:1869734` states that aniline itself did not cause the illness. Binding it would assert the one thing the sources rule out. The bound term names the concept the epidemiology actually supports - a food carrying an industrial contaminant.
Rapeseed oil imported under a 2% aniline denaturant for industrial use, re-refined to strip the denaturant, and distributed in unlabelled 5-litre containers by itinerant salesmen in central and north-western Spain in 1981.
Show evidence (1 reference)
PMID:12192735 SUPPORT REVIEW SYNTHESIS Human Clinical
"The vehicle of the causative toxic agent was identified as an illicit oil that had been diverted from industrial use and refined in order to remove the aniline denaturant, and that was sold in unlabeled 5-liter containers by itinerant salesmen."
The exposure, its provenance and its distribution route in one sentence.
Mechanism Target:
TRIGGERS Ingestion of Aniline-Denatured Rapeseed Oil — The exposure is the disease's sole cause; withdrawal of the oil from sale ended the epidemic and there has been no ongoing incidence since.
Show evidence (1 reference)
PMID:10069247 SUPPORT BACKGROUND Human Clinical
"Epidemiologic studies have demonstrated that illness was caused by consumption of rapeseed oil that had been denatured with aniline."
States the exposure-disease link this edge asserts. BACKGROUND because it is this chemistry paper's framing of the established epidemiology rather than its own analysis.
📈

Progression

4
Acute phase
The first two months: non-cardiogenic pulmonary oedema, rash, eosinophilia and myalgia. Roughly half of patients across the epidemic recovered from this phase without apparent sequelae (PMID:1869734), against 9% achieving remission in the 332-patient cohort followed below. The two figures are not in conflict and are not averaged here: the first is the whole ~20,000-case epidemic, the second a referral cohort selected for longer follow-up.
Show evidence (1 reference)
PMID:8412642 SUPPORT PRIMARY RESULT Human Clinical
"The acute phase lasted 2 months, and was characterized by pulmonary edema, rash, eosinophilia, and myalgia."
The duration and the four defining features.
Intermediate phase
Months two to four: severe myalgia, skin tenderness, subcutaneous oedema, altered liver function and pulmonary hypertension.
Show evidence (1 reference)
PMID:8412642 SUPPORT PRIMARY RESULT Human Clinical
"During the intermediate phase (second to fourth months), severe myalgia, skin tenderness, subcutaneous edema, altered liver function, and pulmonary hypertension developed."
The window and its features.
Early chronic phase
Month four to the end of the second year: scleroderma, sicca syndrome, polyneuropathy, joint contractures, weight loss and functional limitation. Only 9% of patients remitted after the acute phase.
Show evidence (1 reference)
PMID:8412642 SUPPORT PRIMARY RESULT Human Clinical
"Only 9% of the patients achieved remission after the acute phase, the rest developing late clinical manifestations of the disease."
The proportion that did not progress, which is what makes the chronic phase the expected course rather than a complication.
Late chronic phase
Beyond two years, with partial improvement in many: muscle cramps, chronic musculoskeletal pain, chronic lung disease, Raynaud phenomenon, carpal tunnel syndrome and psychological disturbance.
Show evidence (1 reference)
PMID:8412642 SUPPORT PRIMARY RESULT Human Clinical
"Its most prominent clinical features were muscle cramps, chronic musculoskeletal pain, chronic lung disease, Raynaud phenomenon, carpal tunnel syndrome, and psychologic disturbances."
The late-phase feature set.
📊

Prevalence

1
Spain, 1981 epidemic cohort
Cases In Literature Not yet documented
A closed point-source epidemic with no ongoing incidence, so a population rate would be misleading and no numeric slot is populated. prevalence_class is NOT_YET_DOCUMENTED because no rate has been documented, not because the epidemic was uncounted - the cohort size is documented, and appears in the evidence snippet below. Estimates differ slightly with the cohort-closure date used.
Show evidence (1 reference)
PMID:8412642 SUPPORT PRIMARY RESULT Human Clinical
"It affected 19,748 people, of whom 457 died."
One cohort count with its own death toll. See the entry `notes:` for why this is not reconciled with the other published figures.
⚖️

Clinical Burden

High
Around 20,000 people were affected in a single point-source epidemic, most of whom did not remit after the acute phase, and mortality in the cohort was 8.4% over the first thirteen years. Pulmonary hypertension is one of only two causes of death that were not decreased relative to the general Spanish population - the cohort's overall mortality was lower than expected, so this is one cause spared a general decline rather than a demonstrated excess - and new cases of it were still being diagnosed in exposed survivors decades on.
Show evidence (2 references)
PMID:10024203 SUPPORT PRIMARY RESULT Human Clinical
"We identified 1663 deaths between 1 May 1981 and 31 December 1994 among 19 754 TOS cohort members, for a crude mortality rate of 8.4%."
The cohort size and its thirteen-year crude mortality.
PMID:10024203 SUPPORT PRIMARY RESULT Human Clinical
"except for deaths attributed to external causes including TOS and deaths due to pulmonary hypertension, all causes of death were decreased in TOS patients compared to the Spanish population"
Identifies pulmonary hypertension as one of only two causes of death not decreased relative to the Spanish population. Note what the sentence does and does not say: "not decreased" is weaker than "raised", and the same abstract reports the cohort's overall mortality as less than expected. It still makes this the burden-defining phenotype, because every other cause moved the other way.
🐁

Animal Models

3
HLA-DR2/DQ6 transgenic mouse fed toxic oil
Built to test the human HLA restriction rather than to reproduce the disease. DR2 and DQ6 transgenic mice given toxic oil showed higher eosinophil percentages and IgE than DR3 or DR4 transgenics, which matches the direction of the human HLA finding.
Species
Mouse
Genotype
HLA-DR2 and HLA-DQ6 transgenic
Publication
PMID:15979827
Pathograph links
T Cell Activation with Th2 Skewing in Lung Tissue partially recapitulates fidelity: low observes: cellular Reproduces the HLA-dependent immunological readouts, and nothing else about the disease.
Limitations: The model reports eosinophil percentage, IgE and a cytokine shift. It does not produce the endovasculitis that defines toxic oil syndrome, and no animal model does - see the `no_animal_model` discussion. Reading these mice as a model of the disease rather than of one immunological axis is the error the entry is guarding against.
Show evidence (1 reference)
PMID:15979827 SUPPORT PRIMARY RESULT Model Organism
"Results show that mice expressing human DR2 and DQ6 (both in linkage disequilibrium), had higher percentage of eosinophils (DQ6) and IgE (DR2) than other transgenic mice tested (DR3 and DR4)."
The measured result, and the reason the link is PARTIALLY rather than fully recapitulating: two immunological readouts, no disease.
Mouse given fatty acid anilides and the linoleic PAP diester
Intraperitoneal administration of the two suspect compound classes. The closest any animal has come to the acute human illness, and the authors are careful about how close that is.
Species
Mouse
Genotype
wild type
Publication
PMID:10650923
Pathograph links
Peripheral Eosinophilia partially recapitulates fidelity: low observes: organism Reproduces the blood eosinophilia and lung pathology of the acute phase from the candidate compounds, without the vasculitis that defines the disease.
Limitations: Intraperitoneal rather than oral, which is the route the disease took, and the companion rat study found intragastric PAP produced no toxicity at all. The endovasculitis is absent. The authors say the effects "resemble" the acute human disease and explicitly note the full spectrum was not produced.
Show evidence (2 references)
PMID:10650923 SUPPORT PRIMARY RESULT Model Organism
"Linoleic diester of PAP led to weight loss, haemorrhage, congestion and emphysema in the lungs and an increase in blood eosinophilia."
The measured findings, including the eosinophilia this link targets.
+ 1 more reference
Rat given 3-phenylamino-1,2-propanediol
A direct attempt to produce the disease with the leading candidate compound. It failed, and the way it failed is informative: the pathology it did produce was judged unrepresentative, and the oral route - the route of the actual epidemic - produced nothing at all.
Species
Rat
Genotype
wild type
Publication
PMID:7779449
Pathograph links
In Situ Thrombosis on the Damaged Vessel Wall fails to recapitulate fidelity: low observes: organism Intraperitoneal PAP caused massive pulmonary thromboembolism in rats, but by thrombosis in mesenteric vessels that then embolised - not by the endovasculitis of the human disease.
Limitations: The authors state the pathology is not representative of human toxic oil syndrome. The mono-oleoyl ester caused no toxicity at all, and intragastric PAP caused none either, which is the finding that matters most: the human disease was caused by eating the compound.
Show evidence (2 references)
PMID:7779449 REFUTE PRIMARY RESULT Model Organism
"Comparatively, the pathology seen after intraperitoneal administration of PAP was not thought to be representative of the pathology of the toxic oil syndrome in man."
The authors' own verdict on their model, which is what makes this a FAILS_TO_RECAPITULATE link rather than a partial one.
+ 1 more reference
{ }

Source YAML

click to show 
name: Toxic Oil Syndrome
creation_date: "2026-09-19T01:45:00Z"
category: Environmental
categories:
- Toxic Exposure Disorder
- Food Adulteration Disorder
- Acquired Sclerodermiform Syndrome
synonyms:
- toxic oil syndrome
- Spanish toxic oil syndrome
- TOS
- toxic allergic syndrome
- denatured rapeseed oil syndrome
description: >-
  Toxic oil syndrome is a multisystem disease caused by eating rapeseed oil that
  had been denatured with aniline for industrial use, then illegally re-refined
  to strip the denaturant and sold in unlabelled containers as cooking oil. It
  appeared as a point-source epidemic in central and north-western Spain in May
  1981 and affected roughly 20,000 people.

  The unifying lesion is a non-necrotising endovasculitis of vessels of every
  calibre in essentially every organ. Endothelial injury is followed by
  inflammatory infiltration of the intima, then by myointimal and fibroblastic
  proliferation that narrows or obliterates the lumen, with in situ thrombosis
  compounding the ischaemia. The same inflammatory process outside the vessel
  wall produces a lymphocytic perineuritis that ends in axonal degeneration, an
  interstitial myopathy, and the dermal and fascial fibrosis that makes the
  chronic phase look like scleroderma.

  The disease runs in phases. An acute respiratory illness with non-cardiogenic
  pulmonary oedema, fever, rash, eosinophilia and myalgia gives way over months
  to intense myalgia, hepatic dysfunction and pulmonary hypertension, and then
  to a chronic sclerodermiform stage with joint contractures, peripheral
  neuropathy, sicca syndrome and Raynaud phenomenon. Pulmonary arterial
  hypertension is still being diagnosed in exposed survivors decades later.

  The agent has never been definitively identified. Fatty acid esters of
  3-(N-phenylamino)-1,2-propanediol carry the strongest epidemiological
  association with case-related oils, and aniline itself does not cause the
  illness, but no compound has been shown to reproduce the disease.
disease_term:
  preferred_term: toxic oil syndrome
  term:
    id: MONDO:0016421
    label: toxic oil syndrome
notes: >-
  **The etiologic agent is unidentified and this entry does not assert one.**
  The pathograph names the PAP esters because they carry the strongest
  epidemiological association with case-related oil and because their
  absorption, lipase hydrolysis and hepatic biotransformation have each been
  measured. That is an association plus a plausible route, not a demonstrated
  cause, and the `unidentified_etiologic_agent` discussion records what is
  missing. No animal has developed the disease, though two rodent studies cited here got
  partway and are curated under animal_models: anilides and a PAP diester
  reproduced weight loss, lung pathology and blood eosinophilia in mice, and
  PAP produced pulmonary thromboembolism in rats that the authors judged
  unrepresentative of the human pathology.

  **Mortality figures in this entry are not interchangeable and are
  deliberately reported with their windows attached.** The sources count
  different things over different periods: 457 deaths in a cohort description
  (`PMID:8412642`), over 300 in the first year (`PMID:1869734`), 1,663 deaths
  from all causes among 19,754 cohort members by the end of 1994
  (`PMID:10024203`), and over 1,200 all-cause deaths in the 20-year review
  (`PMID:12192735`). These are not four estimates of one quantity, and the
  entry does not average or reconcile them.

  **Most phenotypes here are deliberately left unwired, and the rule is the
  same for all of them: an edge goes in only where a cited source makes the
  causal link.** Run `just list-disconnected-phenotypes
  kb/disorders/Toxic_Oil_Syndrome.yaml` for the current set rather than
  trusting a list in this note - an earlier version of this paragraph
  enumerated them and was wrong within the hour, because adding a phenotype
  silently falsifies a count. Three cases are worth naming. The constitutional
  features (fever, rash, weight loss) have no named lesion in any cited
  source. Cognitive impairment and fatigue sit on the open question the
  `cognitive_deficit_mechanism` discussion states, and the one candidate route
  in the literature, fatigue mediating the cognitive score, is a
  structural equation result rather than a mechanism. Myalgia is the pointed
  one: it is among the two features that defined the epidemic, the muscle
  shows an interstitial inflammatory myopathy, and it would be easy to draw an
  edge between them - but the myalgia of this disease and of
  eosinophilia-myalgia syndrome is conspicuously not explained by the muscle
  pathology, and no source here asserts that it is. Connectivity in this entry
  is correspondingly low. An edge drawn to raise that figure would be worse
  than the gap.

  **Two scope exclusions.** Eosinophilia-myalgia syndrome is not curated here.
  It is a separate epidemic with a separate vehicle (L-tryptophan) that
  resembles TOS closely enough that the two are routinely discussed together,
  and several references cited here are explicitly comparative; the comparison
  is used where a source makes it, and no EMS finding is imported as a TOS
  finding. The 2025 long-term survivor studies *are* curated, as
  `Cognitive impairment` and `Fatigue` and the `cognitive_deficit_mechanism`
  discussion. The deep-research report cites them only by PMC identifier and
  web link; their PubMed records (`PMID:40507507`, `PMID:40507935`) were
  recovered when the report's citations were resolved, which is why they are
  quotable here.

  No GeneReviews chapter exists and none is expected. This is an acquired
  point-source intoxication with no Mendelian basis; `just check-genereviews`
  returns NO_CHAPTER for both Bookshelf collections.
mechanistic_hypotheses:
- hypothesis_group_id: direct_toxic_endothelial_injury
  hypothesis_label: Direct toxic injury to vascular endothelium
  status: CANONICAL
  description: >-
    The initiating event is a direct chemical effect of the ingested toxin, or
    of a metabolite, on the vascular endothelium - the pathology series that
    defines the lesion proposes free radicals as the mediator. On this account
    the immune response follows the endothelial lesion and perpetuates it
    rather than starting it.
  evidence:
  - reference: PMID:1654352
    reference_title: Extracardiac vascular and neural lesions in the toxic oil syndrome.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "A direct effect of unidentified toxic substances, possibly free radicals, may cause the endothelial lesion."
    explanation: >-
      The pathologists' own proposal, and note its hedging - "may cause", with
      the substances unidentified. This is the hypothesis stated as a
      hypothesis by the people who described the lesion.
- hypothesis_group_id: immune_mediated_initiation
  hypothesis_label: Immunological mechanism as the initiating event
  status: ALTERNATIVE
  description: >-
    An HLA-restricted immune response to the ingested xenobiotic is the primary
    pathogenetic mechanism rather than a secondary amplifier. The strongest
    evidence is the demonstration of T-cell activation and a cytokine shift in
    lung tissue from affected patients, alongside HLA class I and class II
    associations and the transgenic-mouse work built on them. The two accounts
    are not mutually exclusive and differ mainly on what comes first.
  evidence:
  - reference: PMID:11501225
    reference_title: 'The toxic oil syndrome: 20 years on.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "There is good evidence that the initial pathogenetic mechanism is immunological."
    explanation: >-
      States the claim this group makes, and states it about the *initial*
      mechanism, which is precisely where it competes with the direct-toxicity
      account.
  - reference: PMID:9074654
    reference_title: 'Cytokine mRNA expression in lung tissue from toxic oil syndrome patients: a TH2 immunological mechanism.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Data presented in this paper are the first clear evidence that an immunological mechanism is directly implicated in this illness."
    explanation: >-
      The primary result the group rests on, in the authors' own summary of
      what their lung-tissue cytokine measurements establish.
- hypothesis_group_id: paf_analogue_signalling
  hypothesis_label: PAP esters acting as platelet-activating factor analogues
  status: EMERGING
  description: >-
    The PAP esters resemble platelet-activating factor structurally, and some of
    them perturb PAF synthesis. This would supply a route by which a compound
    absorbed from the gut could produce a systemic vascular and eosinophilic
    disease. It is an inference from structure plus a measured effect on PAF synthesis; no
    step of it has been shown to operate in a patient, and the abstract does
    not name the system the PAF-synthesis measurement was made in, so this
    entry does not claim one.
  evidence:
  - reference: PMID:11768157
    reference_title: Absorption and effects of 3-(N-phenylamino)-1,2-propanediol esters in relation to toxic oil syndrome.
    supports: SUPPORT
    evidence_source: OTHER
    quote_role: PRIMARY_RESULT
    snippet: "Some of these PAP esters, when a long acyl chain was present in the sn-1 position of the molecule, showed an inhibitory effect on the PAF synthesis."
    explanation: >-
      The measured effect the hypothesis is built on. Note it is an
      *inhibitory* effect on PAF synthesis, not PAF-like agonism, so the
      structural analogy and the measured activity do not point the same way. Graded
      OTHER because the abstract does not say whether this measurement was
      made in animals or in vitro; the companion item's animal grading covers
      the absorption work rather than this one.
- hypothesis_group_id: rapeseed_oil_itself
  hypothesis_label: The rapeseed oil rather than its contaminants
  status: ALTERNATIVE
  description: >-
    A minority account from a cardiac pathology series: rapeseed oil itself
    remains a suspect, with the aniline reaction products acting only as a
    facilitator. It is recorded because it is a published dissent from the
    contaminant consensus, not because the entry finds it likely - the
    epidemiology ties illness to specific refined batches rather than to
    rapeseed oil consumption in general.
  evidence:
  - reference: PMID:8001309
    reference_title: The toxic oil syndrome.
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    quote_role: REVIEW_SYNTHESIS
    snippet: "it is suggested that this oil must remain a major suspected cause of the toxic oil syndrome, particularly in conjunction with some as yet unexplained facilitative influence by oleoanilids"
    explanation: >-
      The dissenting proposal in the authors' words. The sentence it continues
      begins "Based upon observations by others with experimental feeding of
      rapeseed oil", so the evidence behind the suggestion is other groups'
      animal feeding studies, not this paper's own human autopsy series -
      hence MODEL_ORGANISM for the evidence and REVIEW_SYNTHESIS for a
      publication MEDLINE types as a Review. "As yet unexplained facilitative
      influence" is the weak point the authors themselves flag.
pathophysiology:
- name: Ingestion of Aniline-Denatured Rapeseed Oil
  biological_scale: ORGANISM
  description: >-
    Rapeseed oil imported for industrial use was denatured with 2% aniline,
    then illegally re-refined to remove the denaturant and sold door to door as
    cooking oil. Aniline itself is not the toxin; the refining step is what
    generated the compounds that track with disease.
  chemical_entities:
  - preferred_term: aniline
    term:
      id: CHEBI:17296
      label: aniline
  downstream:
  - target: Systemic Exposure to PAP Fatty Acid Esters
    causal_link_type: DIRECT
    description: >-
      The esters are products of the refining process, not of the denaturing or
      of storage, which is what ties them to the one refinery whose oil caused
      illness.
    evidence:
    - reference: PMID:7710294
      reference_title: 'Possible etiologic agents for toxic oil syndrome: fatty acid esters of 3-(N-phenylamino)-1,2-propanediol.'
      supports: SUPPORT
      evidence_source: OTHER
      quote_role: PRIMARY_RESULT
      snippet: "These results show that the esters of PAP were products of the ITH refining process and were not formed spontaneously during storage."
      explanation: >-
        Establishes this edge specifically - that the ingested oil carried the
        esters because of what was done to it, which is the step this link
        asserts. Graded OTHER because it is an analytical chemistry result on
        oil samples rather than a study of people, animals or cells.
  evidence:
  - reference: PMID:12192735
    reference_title: 'Toxic oil syndrome: the perspective after 20 years.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "The vehicle of the causative toxic agent was identified as an illicit oil that had been diverted from industrial use and refined in order to remove the aniline denaturant, and that was sold in unlabeled 5-liter containers by itinerant salesmen."
    explanation: >-
      The vehicle and the route by which it reached households, which is what
      this node claims.
  - reference: PMID:1869734
    reference_title: 'Toxic oil syndrome: a current clinical and epidemiologic summary, including comparisons with the eosinophilia-myalgia syndrome.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "Aniline itself did not cause the illness, but the causal agent may be a reaction product of aniline with some oil component."
    explanation: >-
      The reason this node is named for the denatured oil rather than for
      aniline exposure, and the reason no ECTO aniline-exposure term is bound
      anywhere in this entry.
- name: Systemic Exposure to PAP Fatty Acid Esters
  biological_scale: ORGANISM
  description: >-
    The di- and mono-oleyl esters of 3-(N-phenylamino)-1,2-propanediol are the
    compounds most strongly associated with case-related oil. They are
    hydrolysed by pancreatic lipase like a triglyceride, absorbed from the gut,
    and distributed to organs - which is what makes a systemic disease from an
    ingested compound.
  biological_processes:
  - preferred_term: response to xenobiotic stimulus
    term:
      id: GO:0009410
      label: response to xenobiotic stimulus
  downstream:
  - target: Hepatic Biotransformation of PAP to 3-(Phenylamino)alanine
    causal_link_type: DIRECT
  - target: Vascular Endothelial Injury
    causal_link_type: INDIRECT_UNKNOWN_INTERMEDIATES
    hypothesis_groups:
    - direct_toxic_endothelial_injury
    - paf_analogue_signalling
    description: >-
      The central unproven step of the whole entry. No cited source traces a
      route from an absorbed PAP ester to an injured endothelial cell; the
      pathology series proposes free radicals and says so hypothetically. The
      intermediates are recorded as unknown rather than filled in.
  evidence:
  - reference: PMID:10069247
    reference_title: Epidemiologic evidence for a new class of compounds associated with toxic oil syndrome.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "We found the odds ratio for exposure to DPAP (OR = 26.4, 95% CI = 6.4-76.3) is much higher than the odds ratio for exposure to oleyl anilide (OR = 4.1, 95% CI = 2.2-7.8), implying that exposure to DPAP was a more relevant risk factor for development of toxic oil syndrome than exposure to oleyl anilide."
    explanation: >-
      Quantifies why this node names the PAP esters and not the fatty acid
      anilides that were the earlier suspect, and gives both odds ratios so the
      comparison is visible rather than asserted.
  - reference: PMID:11768157
    reference_title: Absorption and effects of 3-(N-phenylamino)-1,2-propanediol esters in relation to toxic oil syndrome.
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    quote_role: PRIMARY_RESULT
    snippet: "Results indicate that PAP esters are absorbed in the gastrointestinal tract and are distributed and stored in different organs, particularly in the liver and brown adipose tissue."
    explanation: >-
      Supplies the absorption and distribution step this node asserts. Graded
      MODEL_ORGANISM because the absorption and tissue-distribution
      measurements were made in animals.
  - reference: PMID:15257613
    reference_title: 'Studies on toxic oil syndrome: stereoselective hydrolysis of 3-(phenylamino)propane-1,2-diol esters by human pancreatic lipase.'
    supports: SUPPORT
    evidence_source: IN_VITRO
    quote_role: PRIMARY_RESULT
    snippet: "Taken together, these results showed that PAP esters are substrates of hPL and that the two hydrolytic steps exhibit kinetic resolution in favor of the (S)-enantiomers."
    explanation: >-
      Shows the esters enter normal lipid handling at the first step, which is
      how a compound eaten in trace amounts becomes systemically available.
- name: Hepatic Biotransformation of PAP to 3-(Phenylamino)alanine
  biological_scale: MOLECULAR
  description: >-
    Liver tissue converts PAP to 3-(phenylamino)alanine, the aniline derivative
    independently isolated from the L-tryptophan implicated in
    eosinophilia-myalgia syndrome. This is the chemical link between the two
    epidemics, and it is a shared metabolite rather than a shared exposure.
  biological_processes:
  - preferred_term: response to xenobiotic stimulus
    term:
      id: GO:0009410
      label: response to xenobiotic stimulus
  evidence:
  - reference: PMID:8555405
    reference_title: 'Biotransformation of 3-(phenylamino)-1,2-propanediol to 3-(phenylamino)alanine: a chemical link between toxic oil syndrome and eosinophilia-myalgia syndrome.'
    supports: SUPPORT
    evidence_source: IN_VITRO
    quote_role: PRIMARY_RESULT
    snippet: "Here, we demonstrate the biotransformation of PAP into PAA by both rat hepatocytes and human liver tissue."
    explanation: >-
      The reaction this node names, shown in human liver tissue as well as rat
      hepatocytes. IN_VITRO because both preparations are tissue and cells
      outside an organism.
  - reference: PMID:8555405
    reference_title: 'Biotransformation of 3-(phenylamino)-1,2-propanediol to 3-(phenylamino)alanine: a chemical link between toxic oil syndrome and eosinophilia-myalgia syndrome.'
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: IN_VITRO
    quote_role: PRIMARY_RESULT
    snippet: "This finding is the first reported chemical link between TOS and EMS and suggests that these two related diseases share a common etiology, namely, PAA."
    explanation: >-
      Why the node is worth having as its own step rather than folded into the
      exposure node. INDIRECT because a shared metabolite makes a common
      etiology plausible without establishing it - the authors write
      "suggests".
- name: Vascular Endothelial Injury
  biological_scale: CELLULAR
  description: >-
    The first lesion in the vessel wall, graded in the pathology series from
    endothelial cell swelling through to cell necrosis. Every later vascular
    step in this entry follows from it.
  cell_types:
  - preferred_term: vascular endothelial cell
    term:
      id: CL:0002139
      label: endothelial cell of vascular tree
  locations:
  - preferred_term: blood vessel
    term:
      id: UBERON:0001981
      label: blood vessel
  downstream:
  - target: Perivascular and Interstitial Inflammatory Infiltration
    causal_link_type: DIRECT
    evidence:
    - reference: PMID:1654352
      reference_title: Extracardiac vascular and neural lesions in the toxic oil syndrome.
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      quote_role: PRIMARY_RESULT
      snippet: "It then progresses by mixed cellular inflammatory infiltration of the intima and, in some cases, of the media and adventitia."
      explanation: >-
        States the ordering this edge asserts - infiltration follows the
        endothelial lesion - rather than merely that both are present.
  evidence:
  - reference: PMID:1654352
    reference_title: Extracardiac vascular and neural lesions in the toxic oil syndrome.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "The vascular lesions begins with endothelial damage that varies from cellular swelling to cellular necrosis."
    explanation: >-
      Names this as the first lesion and gives its range. Quoted with the
      source's own grammatical slip intact, as a snippet must be.
  - reference: PMID:9074654
    reference_title: 'Cytokine mRNA expression in lung tissue from toxic oil syndrome patients: a TH2 immunological mechanism.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: BACKGROUND
    snippet: "The pathologic findings in TOS showed primary endothelial injury, with cell proliferation and perivascular inflammatory infiltrates."
    explanation: >-
      An independent statement that the endothelial injury is primary. BACKGROUND
      because it is this cytokine paper's framing of established pathology
      rather than its own measurement.
- name: T Cell Activation with Th2 Skewing in Lung Tissue
  biological_scale: CELLULAR
  description: >-
    Lung tissue from affected patients shows T-cell activation with both Th1 and
    Th2 cytokine profiles raised, and the Th2 increment over Th1 is itself
    significant. The response is not purely Th2 - both arms are up - and the
    entry says so rather than simplifying to a Th2 disease.
  cell_types:
  - preferred_term: T cell
    term:
      id: CL:0000084
      label: T cell
  - preferred_term: T-helper 2 cell
    term:
      id: CL:0000546
      label: T-helper 2 cell
  biological_processes:
  - preferred_term: T cell activation
    term:
      id: GO:0042110
      label: T cell activation
    modifier: INCREASED
  - preferred_term: type 2 immune response
    term:
      id: GO:0042092
      label: type 2 immune response
    modifier: INCREASED
  locations:
  - preferred_term: lung
    term:
      id: UBERON:0002048
      label: lung
  downstream:
  - target: Peripheral Eosinophilia
    causal_link_type: DIRECT
  - target: Perivascular and Interstitial Inflammatory Infiltration
    causal_link_type: DIRECT
    hypothesis_groups:
    - immune_mediated_initiation
  evidence:
  - reference: PMID:9074654
    reference_title: 'Cytokine mRNA expression in lung tissue from toxic oil syndrome patients: a TH2 immunological mechanism.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "We found a significant increase in Th1 (P = 0.006) and Th2 (P = 0.003) cytokine profile in TOS patients with respect to controls."
    explanation: >-
      Both arms rose, and the quote is chosen to carry that rather than the
      Th2 figure alone - the node's description depends on it.
  - reference: PMID:9074654
    reference_title: 'Cytokine mRNA expression in lung tissue from toxic oil syndrome patients: a TH2 immunological mechanism.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "The increment in TH2 response with respect to TH1 is significant (P = 0.03) in TOS lung specimens."
    explanation: >-
      The skew itself, which is the part of the claim the previous quote does
      not establish.
- name: Peripheral Eosinophilia
  biological_scale: ORGANISM
  description: >-
    Marked peripheral eosinophilia is one of the two features that made the
    epidemic recognisable as a single disease, alongside incapacitating myalgia.
    Eosinophils are also found in the vessel wall infiltrate in some cases.
  cell_types:
  - preferred_term: eosinophil
    term:
      id: CL:0000771
      label: eosinophil
  biological_processes:
  - preferred_term: eosinophil chemotaxis
    term:
      id: GO:0048245
      label: eosinophil chemotaxis
    modifier: INCREASED
  downstream:
  - target: Increased total eosinophil count
    causal_link_type: DIRECT
  - target: Perivascular and Interstitial Inflammatory Infiltration
    causal_link_type: DIRECT
    description: >-
      The eosinophil-rich subset of the infiltrate. The pathology series is
      careful that this is "some cases" rather than the rule, and the edge
      claims no more than that.
    evidence:
    - reference: PMID:1654352
      reference_title: Extracardiac vascular and neural lesions in the toxic oil syndrome.
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      quote_role: PRIMARY_RESULT
      snippet: "In some cases the infiltrate is rich in eosinophils and a few show foamy histiocytes."
      explanation: >-
        Supports the edge and bounds it - eosinophil-rich infiltration is a
        subset finding, not the universal picture.
  evidence:
  - reference: PMID:8266108
    reference_title: 'Toxic oil syndrome and eosinophilia-myalgia syndrome: May 8-10, 1991, World Health Organization meeting report.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "Both illnesses affect patients clinically by causing intense, incapacitating myalgias and a marked peripheral eosinophilia."
    explanation: >-
      The WHO workshop's statement of the two defining clinical features. It is
      a comparative sentence covering TOS and EMS together, which is how the
      source makes the claim.
- name: Perivascular and Interstitial Inflammatory Infiltration
  biological_scale: TISSUE
  description: >-
    The mixed cellular infiltrate that follows endothelial injury, in the intima
    and sometimes the media and adventitia, and more widely in the interstitium
    of affected organs. It is the branch point of the entry: the vascular,
    neural, muscular and cutaneous arms all descend from it.
  cell_types:
  - preferred_term: T cell
    term:
      id: CL:0000084
      label: T cell
  - preferred_term: eosinophil
    term:
      id: CL:0000771
      label: eosinophil
  biological_processes:
  - preferred_term: leukocyte migration
    term:
      id: GO:0050900
      label: leukocyte migration
    modifier: INCREASED
  locations:
  - preferred_term: tunica intima
    term:
      id: UBERON:0002523
      label: tunica intima
  downstream:
  - target: Myointimal and Fibroblastic Proliferation
    causal_link_type: DIRECT
  - target: Lymphocytic Perineuritis
    causal_link_type: DIRECT
  - target: Interstitial Inflammatory Myopathy
    causal_link_type: DIRECT
  - target: Dermal and Fascial Fibrosis
    causal_link_type: INDIRECT_UNKNOWN_INTERMEDIATES
    description: >-
      Drawn with unknown intermediates deliberately. The fibrogenic growth
      factors that would supply the intermediates in the comparable disease
      were looked for in TOS skin and were not found - see the REFUTE item on
      the target node.
  - target: Pulmonary infiltrates
    causal_link_type: DIRECT
    description: >-
      The radiographic infiltrate of the acute phase is this same interstitial
      inflammatory process seen in the lung.
    evidence:
    - reference: PMID:3961509
      reference_title: 'Toxic oil syndrome: a syndrome with features overlapping those of various forms of scleroderma.'
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      quote_role: PRIMARY_RESULT
      snippet: "Histologic investigations showed a widespread chronic interstitial infiltrate with lymphocytic vasculitis."
      explanation: >-
        Supports this edge rather than either node alone: the histology of
        patients presenting with the acute pulmonary picture is a widespread
        interstitial infiltrate, which is the lesion behind the radiographic
        finding.
  evidence:
  - reference: PMID:8412642
    reference_title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "The most important pathologic features of TOS were widespread interstitial infiltrates, non-necrotizing angiitis, endothelial proliferation, and tissue fibrosis."
    explanation: >-
      Names the infiltrate as widespread and interstitial, not only intimal,
      which is what lets this one node serve as the branch point for the
      non-vascular arms.
  - reference: PMID:1654352
    reference_title: Extracardiac vascular and neural lesions in the toxic oil syndrome.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Vessels of every type and size are involved, affecting practically every organ."
    explanation: >-
      The distribution of the lesion, which is why this entry has arms reaching
      lung, nerve, muscle and skin rather than one target organ.
- name: Myointimal and Fibroblastic Proliferation
  biological_scale: TISSUE
  description: >-
    Proliferation of myointimal cells, joined in advanced lesions by
    fibroblasts. This is what converts an inflamed vessel into a narrowed one.
  cell_types:
  - preferred_term: myointimal smooth muscle cell
    term:
      id: CL:0000192
      label: smooth muscle cell
  - preferred_term: fibroblast
    term:
      id: CL:0000057
      label: fibroblast
  biological_processes:
  - preferred_term: smooth muscle cell proliferation
    term:
      id: GO:0048659
      label: smooth muscle cell proliferation
    modifier: INCREASED
  downstream:
  - target: Vascular Luminal Narrowing and Obliteration
    causal_link_type: DIRECT
  evidence:
  - reference: PMID:1654352
    reference_title: Extracardiac vascular and neural lesions in the toxic oil syndrome.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Proliferation of myointimal cells and in advanced stages fibroblastic proliferation causes narrowing or obliteration of the vascular lumen."
    explanation: >-
      Names both cell populations, their sequence, and the consequence, which
      is this node and its single downstream edge in one sentence.
- name: Vascular Luminal Narrowing and Obliteration
  biological_scale: TISSUE
  description: >-
    The narrowed or obliterated lumen that the proliferative response leaves
    behind, and the substrate on which thrombosis forms.
  locations:
  - preferred_term: blood vessel
    term:
      id: UBERON:0001981
      label: blood vessel
  downstream:
  - target: In Situ Thrombosis on the Damaged Vessel Wall
    causal_link_type: DIRECT
  - target: Pulmonary Vascular Remodelling and Plexiform Lesion Formation
    causal_link_type: DIRECT
  evidence:
  - reference: PMID:3961509
    reference_title: 'Toxic oil syndrome: a syndrome with features overlapping those of various forms of scleroderma.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Biopsy studies during this stage showed fibrosis and obliterating arteriopathy."
    explanation: >-
      An independent biopsy series reaching the same obliterative endpoint, and
      dating it to the sclerodermiform stage.
- name: In Situ Thrombosis on the Damaged Vessel Wall
  biological_scale: TISSUE
  description: >-
    Thrombosis on the damaged and narrowed vessel wall. The pathology series is
    explicit that this feeds back on the vascular lesion rather than simply
    following it.
  biological_processes:
  - preferred_term: blood coagulation
    term:
      id: GO:0007596
      label: blood coagulation
    modifier: INCREASED
  downstream:
  - target: Tissue Ischemia and Parenchymal Atrophy
    causal_link_type: DIRECT
  - target: Thromboembolism
    causal_link_type: DIRECT
  evidence:
  - reference: PMID:1654352
    reference_title: Extracardiac vascular and neural lesions in the toxic oil syndrome.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Thromboembolic complications perpetuate the vascular lesion and compound the ischemia and parenchymal atrophy of several organs."
    explanation: >-
      Carries this node and both of its downstream targets, and states the
      feedback onto the vascular lesion that the description notes.
- name: Tissue Ischemia and Parenchymal Atrophy
  biological_scale: TISSUE
  description: >-
    The downstream organ consequence of a progressively obliterated vascular
    bed, reported across several organs rather than confined to one.
- name: Pulmonary Vascular Remodelling and Plexiform Lesion Formation
  biological_scale: TISSUE
  description: >-
    In the lung the same process produces plexiform lesions, thromboses and
    venous lesions - a pulmonary vasculopathy that autopsy studies found
    indistinguishable from primary pulmonary hypertension. Plexiform lesions
    appear late.
  biological_processes:
  - preferred_term: blood vessel remodeling
    term:
      id: GO:0001974
      label: blood vessel remodeling
    modifier: INCREASED
  locations:
  - preferred_term: lung
    term:
      id: UBERON:0002048
      label: lung
  downstream:
  - target: Pulmonary arterial hypertension
    causal_link_type: DIRECT
  evidence:
  - reference: PMID:2914483
    reference_title: Pulmonary hypertension due to toxic oil syndrome. A clinicopathologic study.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "The main pathologic pulmonary vascular findings consisted of plexiform lesions, thromboses, and venous lesions."
    explanation: >-
      The three lesions this node names, from the autopsy series of patients
      who died of the pulmonary hypertension.
  - reference: PMID:2914483
    reference_title: Pulmonary hypertension due to toxic oil syndrome. A clinicopathologic study.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Endothelial damage induced by the toxic agents is suggested as an initial causative mechanism, perpetuated by intimal proliferation and in situ thrombosis."
    explanation: >-
      Reaches the same ordering as the extracardiac pathology series in a
      different organ and a different cohort, which is why the entry models one
      vascular chain rather than an organ-specific one.
  - reference: PMID:6648850
    reference_title: Pulmonary vascular lesions in the toxic oil syndrome in Spain.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "In muscular pulmonary arteries there was pronounced medial hypertrophy and intimal proliferation, which was so severe in one case that it completely occluded the arterial lumen."
    explanation: >-
      The earliest necropsy description of the pulmonary arterial lesion,
      reporting the same intimal proliferation the systemic vessels show and
      carrying it through to complete occlusion.
- name: Lymphocytic Perineuritis
  biological_scale: TISSUE
  description: >-
    The nerve lesion begins as an inflammatory neuropathy with lymphocytic
    infiltration of the perineurium, and progresses to perineural fibrosis.
  cell_types:
  - preferred_term: T cell
    term:
      id: CL:0000084
      label: T cell
  downstream:
  - target: Secondary Axonal Degeneration
    causal_link_type: DIRECT
  evidence:
  - reference: PMID:1654352
    reference_title: Extracardiac vascular and neural lesions in the toxic oil syndrome.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "The peripheral nerve lesions begin with an inflammatory neuropathy with lymphocytic perineuritis and progress to perineural fibrosis with secondary axonal degeneration."
    explanation: >-
      Carries the node, its progression, and the edge to axonal degeneration -
      and the word "secondary" is what makes the ordering a claim rather than
      an inference.
- name: Secondary Axonal Degeneration
  biological_scale: CELLULAR
  description: >-
    Axonal loss following the perineural lesion rather than arising
    independently, which is what the pathology series means by calling it
    secondary.
  downstream:
  - target: Peripheral neuropathy
    causal_link_type: DIRECT
  - target: Neurogenic Muscular Atrophy
    causal_link_type: DIRECT
- name: Interstitial Inflammatory Myopathy
  biological_scale: TISSUE
  description: >-
    The first muscle lesion, an interstitial inflammatory myopathy, which is
    later replaced by denervation atrophy rather than progressing as a primary
    myopathy.
  evidence:
  - reference: PMID:1654352
    reference_title: Extracardiac vascular and neural lesions in the toxic oil syndrome.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Skeletal muscle lesions exhibit an interstitial inflammatory myopathy at first, followed by a neurogenic muscular atrophy."
    explanation: >-
      Both muscle nodes and their order in one sentence, including that the
      later atrophy is neurogenic - which is why the entry routes it from the
      nerve arm rather than from this node.
- name: Neurogenic Muscular Atrophy
  biological_scale: TISSUE
  description: >-
    Denervation atrophy following axonal degeneration. It is placed downstream
    of the nerve lesion, not of the myopathy, because the pathology series calls
    it neurogenic.
  downstream:
  - target: Muscle weakness
    causal_link_type: DIRECT
- name: Dermal and Fascial Fibrosis
  biological_scale: TISSUE
  description: >-
    The fibrosis that gives the chronic phase its sclerodermiform appearance.
    What drives it is not established: the growth factors that drive the
    comparable lesion in eosinophilia-myalgia syndrome were absent from every
    TOS skin specimen examined.
  cell_types:
  - preferred_term: fibroblast
    term:
      id: CL:0000057
      label: fibroblast
  biological_processes:
  - preferred_term: extracellular matrix organization
    term:
      id: GO:0030198
      label: extracellular matrix organization
    modifier: INCREASED
  downstream:
  - target: Scleroderma
    causal_link_type: DIRECT
  - target: Joint contracture
    causal_link_type: DIRECT
  evidence:
  - reference: PMID:3961509
    reference_title: 'Toxic oil syndrome: a syndrome with features overlapping those of various forms of scleroderma.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "TOS is a new chemically induced scleroderma-like syndrome with features overlapping those of eosinophilic fasciitis, systemic sclerosis, and forms of localized scleroderma."
    explanation: >-
      Places the fibrotic phenotype among the sclerodermas without collapsing
      it into any one of them.
  - reference: PMID:8285738
    reference_title: Fibrogenic growth factors in the eosinophilia-myalgia syndrome and the toxic oil syndrome.
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "The presence of TGF-beta and platelet-derived growth factorAA in the periappendageal dermis was significantly more prevalent in EMS than toxic oil syndrome (57% vs 0%)."
    explanation: >-
      REFUTE against the obvious mechanistic account of this node. TGF-beta and
      PDGF-AA are the standard fibrogenic drivers and were found in zero of the
      TOS skin specimens. The deep-research report cited this same paper as
      evidence that those growth factors drive TOS fibrosis, which inverts what
      it found.
  - reference: PMID:8285738
    reference_title: Fibrogenic growth factors in the eosinophilia-myalgia syndrome and the toxic oil syndrome.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "suggest that the pathogenesis of tissue fibrosis in EMS and toxic oil syndrome may be dependent on different growth factors"
    explanation: >-
      The authors' own reading of that negative result, and the reason this
      node's upstream edge carries unknown intermediates rather than borrowing
      the EMS mechanism.
phenotypes:
- category: Respiratory
  name: Pulmonary infiltrates
  description: >-
    Bilateral infiltrates with non-cardiogenic pulmonary oedema were the
    presenting abnormality of the acute phase.
  phenotype_term:
    preferred_term: Pulmonary infiltrates
    term:
      id: HP:0002113
      label: Pulmonary infiltrates
    temporality: ACUTE
  evidence:
  - reference: PMID:1869734
    reference_title: 'Toxic oil syndrome: a current clinical and epidemiologic summary, including comparisons with the eosinophilia-myalgia syndrome.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "patients presented primarily with a respiratory syndrome involving cough, fever, dyspnea, hypoxemia, pulmonary infiltrates and pleural effusions"
    explanation: The acute presenting syndrome, naming the infiltrates directly.
- category: Respiratory
  name: Dyspnea
  phenotype_term:
    preferred_term: Dyspnea
    term:
      id: HP:0002094
      label: Dyspnea
  evidence:
  - reference: PMID:1869734
    reference_title: 'Toxic oil syndrome: a current clinical and epidemiologic summary, including comparisons with the eosinophilia-myalgia syndrome.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "patients presented primarily with a respiratory syndrome involving cough, fever, dyspnea, hypoxemia, pulmonary infiltrates and pleural effusions"
    explanation: Names dyspnoea among the acute presenting features.
- category: Constitutional
  name: Fever
  phenotype_term:
    preferred_term: Fever
    term:
      id: HP:0001945
      label: Fever
    temporality: ACUTE
  evidence:
  - reference: PMID:1869734
    reference_title: 'Toxic oil syndrome: a current clinical and epidemiologic summary, including comparisons with the eosinophilia-myalgia syndrome.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "patients presented primarily with a respiratory syndrome involving cough, fever, dyspnea, hypoxemia, pulmonary infiltrates and pleural effusions"
    explanation: Names fever among the acute presenting features.
- category: Dermatologic
  name: Skin rash
  phenotype_term:
    preferred_term: Skin rash
    term:
      id: HP:0000988
      label: Skin rash
    temporality: ACUTE
  evidence:
  - reference: PMID:8412642
    reference_title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "The acute phase lasted 2 months, and was characterized by pulmonary edema, rash, eosinophilia, and myalgia."
    explanation: Names rash as one of the four defining acute-phase features.
- category: Hematologic
  name: Increased total eosinophil count
  description: >-
    Marked peripheral eosinophilia, one of the two features that identified the
    epidemic as a single disease.
  phenotype_term:
    preferred_term: Peripheral eosinophilia
    term:
      id: HP:0001880
      label: Increased total eosinophil count
  evidence:
  - reference: PMID:8412642
    reference_title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "The acute phase lasted 2 months, and was characterized by pulmonary edema, rash, eosinophilia, and myalgia."
    explanation: Names eosinophilia as an acute-phase feature in the follow-up cohort.
- category: Musculoskeletal
  name: Myalgia
  description: >-
    Intense and incapacitating, worst in the intermediate phase, and persisting
    as chronic musculoskeletal pain in the late chronic phase.
  phenotype_term:
    preferred_term: Myalgia
    term:
      id: HP:0003326
      label: Myalgia
    severity: SEVERE
  evidence:
  - reference: PMID:8412642
    reference_title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "During the intermediate phase (second to fourth months), severe myalgia, skin tenderness, subcutaneous edema, altered liver function, and pulmonary hypertension developed."
    explanation: >-
      Places severe myalgia in the intermediate phase and names the other
      features that accompany it there.
- category: Cardiovascular
  name: Pulmonary arterial hypertension
  description: >-
    Present in the intermediate phase and persisting chronically. It is one of
    only two causes of death that were not decreased in the cohort relative to the
    Spanish population, and new cases were still being diagnosed in exposed
    survivors decades later.
  phenotype_term:
    preferred_term: Pulmonary arterial hypertension
    term:
      id: HP:0002092
      label: Pulmonary arterial hypertension
    clinical_course: PROGRESSIVE
  evidence:
  - reference: PMID:8412642
    reference_title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "During the intermediate phase (second to fourth months), severe myalgia, skin tenderness, subcutaneous edema, altered liver function, and pulmonary hypertension developed."
    explanation: Dates the onset of pulmonary hypertension to the intermediate phase.
  - reference: PMID:41473551
    reference_title: Pulmonary arterial hypertension associated with exposure to toxic rapeseed oil.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "TOS-PAH remains a distinct clinical entity, with new cases diagnosed decades after toxic exposure."
    explanation: >-
      A registry study's conclusion that the pulmonary vascular disease is
      still declaring itself in this cohort, which is why the phenotype carries
      a progressive course rather than being scoped to the 1980s.
  sequelae:
  - target: Dyspnea
    causal_link_type: DIRECT
    description: >-
      Increasing dyspnoea is the leading clinical feature of the late
      pulmonary-hypertensive deterioration, alongside chest pain and syncope.
    evidence:
    - reference: PMID:2914483
      reference_title: Pulmonary hypertension due to toxic oil syndrome. A clinicopathologic study.
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      quote_role: PRIMARY_RESULT
      snippet: "These cases correspond to a late stage of evolution of the disease characterized by progressive deterioration in clinical features--increasing dyspnea, chest pain, syncope, and death (in low-output heart failure)."
      explanation: >-
        Names dyspnoea as a feature of the pulmonary-hypertensive stage in the
        autopsy series of patients who died of it, which is the link this edge
        asserts.
- category: Dermatologic
  name: Scleroderma
  description: >-
    Sclerodermiform skin change of the digits, part of the chronic-phase
    picture that led to TOS being classified among the chemically induced
    sclerodermas.
  phenotype_term:
    preferred_term: Scleroderma-like skin change
    term:
      id: HP:0100324
      label: Scleroderma
    temporality: CHRONIC
  evidence:
  - reference: PMID:3961509
    reference_title: 'Toxic oil syndrome: a syndrome with features overlapping those of various forms of scleroderma.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "They subsequently developed peripheral neuropathy, joint contractures, scleroderma-like changes, Raynaud phenomenon, pulmonary hypertension, sicca syndrome, and liver disease."
    explanation: >-
      The chronic-phase sequence in a series selected for scleroderma-like change.
      The binding was narrowed to match: an earlier draft bound HP:0011838
      Sclerodactyly and defended the digital specificity by pointing at the
      same abstract's "digital tuft changes", which is acro-osteolysis rather
      than skin sclerosis of the digits and does not support it. HP:0100324
      Scleroderma is what the quoted sentence actually says.
- category: Musculoskeletal
  name: Joint contracture
  phenotype_term:
    preferred_term: Joint contracture
    term:
      id: HP:0034392
      label: Joint contracture
    temporality: CHRONIC
  evidence:
  - reference: PMID:8266108
    reference_title: 'Toxic oil syndrome and eosinophilia-myalgia syndrome: May 8-10, 1991, World Health Organization meeting report.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "Long-term complications include pulmonary hypertension, peripheral neuropathies, and joint contractures."
    explanation: Names joint contractures among the three long-term complications.
- category: Neurologic
  name: Peripheral neuropathy
  description: >-
    Sensorimotor neuropathy following the perineural lesion, and one of the
    three named long-term complications.
  phenotype_term:
    preferred_term: Peripheral neuropathy
    term:
      id: HP:0009830
      label: Peripheral neuropathy
    temporality: CHRONIC
  evidence:
  - reference: PMID:8266108
    reference_title: 'Toxic oil syndrome and eosinophilia-myalgia syndrome: May 8-10, 1991, World Health Organization meeting report.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "Long-term complications include pulmonary hypertension, peripheral neuropathies, and joint contractures."
    explanation: Names peripheral neuropathy among the three long-term complications.
- category: Musculoskeletal
  name: Muscle weakness
  description: >-
    Weakness attributable to denervation atrophy rather than to a persisting
    primary myopathy.
  phenotype_term:
    preferred_term: Muscle weakness
    term:
      id: HP:0001324
      label: Muscle weakness
  evidence:
  - reference: PMID:1654352
    reference_title: Extracardiac vascular and neural lesions in the toxic oil syndrome.
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Skeletal muscle lesions exhibit an interstitial inflammatory myopathy at first, followed by a neurogenic muscular atrophy."
    explanation: >-
      INDIRECT because the source reports the muscle pathology rather than the
      clinical weakness; the weakness follows from neurogenic atrophy by an
      inference step this entry makes explicit.
- category: Rheumatologic
  name: Xerostomia
  description: Dry mouth as part of the sicca syndrome of the chronic phase.
  phenotype_term:
    preferred_term: Dry mouth
    term:
      id: HP:0000217
      label: Xerostomia
    temporality: CHRONIC
  evidence:
  - reference: PMID:8412642
    reference_title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "It was marked by scleroderma, sicca syndrome, polyneuropathy, joint contractures, weight loss, and functional limitations."
    explanation: >-
      Names sicca syndrome in the early chronic phase. The HP binding is
      Xerostomia because HPO has no sicca-syndrome term and dry mouth is the
      component this entry can name.
- category: Constitutional
  name: Weight loss
  phenotype_term:
    preferred_term: Weight loss
    term:
      id: HP:0001824
      label: Weight loss
    temporality: CHRONIC
  evidence:
  - reference: PMID:8412642
    reference_title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "It was marked by scleroderma, sicca syndrome, polyneuropathy, joint contractures, weight loss, and functional limitations."
    explanation: Names weight loss in the early chronic phase.
- category: Constitutional
  name: Edema
  description: >-
    Subcutaneous oedema of the intermediate phase, alongside the skin
    tenderness and severe myalgia of that window.
  phenotype_term:
    preferred_term: Subcutaneous edema
    term:
      id: HP:0000969
      label: Edema
  evidence:
  - reference: PMID:8412642
    reference_title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "During the intermediate phase (second to fourth months), severe myalgia, skin tenderness, subcutaneous edema, altered liver function, and pulmonary hypertension developed."
    explanation: >-
      Names subcutaneous oedema among the intermediate-phase features. Bound to
      the general HP:0000969 Edema rather than HP:0012398 Peripheral edema: the
      source says subcutaneous, which is a tissue plane, not a distribution,
      and narrowing to peripheral would assert a limb predominance the sentence
      does not carry.
- category: Cardiovascular
  name: Raynaud phenomenon
  phenotype_term:
    preferred_term: Raynaud phenomenon
    term:
      id: HP:0030880
      label: Raynaud phenomenon
    temporality: CHRONIC
  evidence:
  - reference: PMID:8412642
    reference_title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Its most prominent clinical features were muscle cramps, chronic musculoskeletal pain, chronic lung disease, Raynaud phenomenon, carpal tunnel syndrome, and psychologic disturbances."
    explanation: >-
      Names Raynaud phenomenon among the most prominent features of the late
      chronic phase.
- category: Neurologic
  name: Cognitive impairment
  description: >-
    Survivors assessed more than four decades after exposure performed worse
    than matched controls on attention, executive function, processing speed
    and global cognition. The deficit did not survive adjustment for fatigue,
    depression, anxiety and CNS-acting medication, and fatigue substantially
    mediated it - so whether this is a direct neurotoxic sequela or a
    downstream effect of the chronic illness is unsettled.
  phenotype_term:
    preferred_term: Cognitive impairment
    term:
      id: HP:0100543
      label: Cognitive impairment
    temporality: CHRONIC
  evidence:
  - reference: PMID:40507507
    reference_title: 'Cognitive Functioning in Toxic Oil Syndrome Survivors: A Case-Control Study Four Decades After the Epidemic.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "TOS survivors showed significantly poorer performance than controls in attention, executive function, processing speed, and global cognition after adjusting for demographic and vascular risk factors."
    explanation: >-
      The measured deficit across four domains in a matched case-control
      design, adjusted for demographic and vascular confounders.
  - reference: PMID:40507507
    reference_title: 'Cognitive Functioning in Toxic Oil Syndrome Survivors: A Case-Control Study Four Decades After the Epidemic.'
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "However, these differences were no longer statistically significant after additional adjustment for fatigue, depression, anxiety, and central nervous system-acting medications."
    explanation: >-
      REFUTE against reading the deficit as an independent neurocognitive
      sequela. The same study reports both results, so both are carried; taking
      only the first sentence would state the finding as the authors
      specifically declined to state it.
- category: Constitutional
  name: Fatigue
  description: >-
    Persistent fatigue in long-term survivors, and the variable that mediates
    most of their measured cognitive deficit. Deliberately not linked to
    `Cognitive impairment` in the graph: the mediation is a structural-equation
    result in one cohort, and the schema's `reports_on` slot is for a test
    result reporting on a mechanism, which this is not. No edge type here would
    say what the finding says.
  phenotype_term:
    preferred_term: Fatigue
    term:
      id: HP:0012378
      label: Fatigue
    temporality: CHRONIC
  evidence:
  - reference: PMID:40507507
    reference_title: 'Cognitive Functioning in Toxic Oil Syndrome Survivors: A Case-Control Study Four Decades After the Epidemic.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Structural equation modeling analyses revealed that affective symptoms-particularly fatigue-substantially mediated the relationship between TOS and cognitive performance."
    explanation: >-
      Establishes both that fatigue is present in the cohort and the mediating role
      described above. No graph edge records that mediation - see this
      phenotype's description for why.
- category: Hepatic
  name: Abnormal liver physiology
  description: >-
    Altered liver function appears in the intermediate phase and liver disease
    persists into the chronic phase. The entry binds the broad HPO physiology
    term because the sources say "altered liver function" and "liver disease"
    without naming the abnormality.
  phenotype_term:
    preferred_term: Altered liver function
    term:
      id: HP:0031865
      label: Abnormal liver physiology
  evidence:
  - reference: PMID:8412642
    reference_title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "During the intermediate phase (second to fourth months), severe myalgia, skin tenderness, subcutaneous edema, altered liver function, and pulmonary hypertension developed."
    explanation: Names altered liver function among the intermediate-phase features.
  - reference: PMID:3961509
    reference_title: "Toxic oil syndrome: a syndrome with features overlapping those of various forms of scleroderma."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "They subsequently developed peripheral neuropathy, joint contractures, scleroderma-like changes, Raynaud phenomenon, pulmonary hypertension, sicca syndrome, and liver disease."
    explanation: >-
      An independent series carrying liver disease into the chronic phase
      alongside the sclerodermiform features.
- category: Respiratory
  name: Pleural effusion
  phenotype_term:
    preferred_term: Pleural effusion
    term:
      id: HP:0002202
      label: Pleural effusion
    temporality: ACUTE
  evidence:
  - reference: PMID:1869734
    reference_title: "Toxic oil syndrome: a current clinical and epidemiologic summary, including comparisons with the eosinophilia-myalgia syndrome."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "patients presented primarily with a respiratory syndrome involving cough, fever, dyspnea, hypoxemia, pulmonary infiltrates and pleural effusions"
    explanation: Names pleural effusions among the acute presenting features.
- category: Neurologic
  name: Constrictive median neuropathy
  description: >-
    Carpal tunnel syndrome, one of the most prominent features of the late
    chronic phase and shared with eosinophilia-myalgia syndrome.
  phenotype_term:
    preferred_term: Carpal tunnel syndrome
    term:
      id: HP:0012185
      label: Constrictive median neuropathy
    temporality: CHRONIC
  evidence:
  - reference: PMID:8412642
    reference_title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Its most prominent clinical features were muscle cramps, chronic musculoskeletal pain, chronic lung disease, Raynaud phenomenon, carpal tunnel syndrome, and psychologic disturbances."
    explanation: >-
      Names carpal tunnel syndrome among the late-phase features. HPO files
      this as Constrictive median neuropathy, which carries "Carpal tunnel
      syndrome" as a synonym.
- category: Dermatologic
  name: Alopecia
  phenotype_term:
    preferred_term: Alopecia
    term:
      id: HP:0001596
      label: Alopecia
  evidence:
  - reference: PMID:1869734
    reference_title: "Toxic oil syndrome: a current clinical and epidemiologic summary, including comparisons with the eosinophilia-myalgia syndrome."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "the remaining patients developed an intermediate or chronic phase, or both, of illness involving severe myalgia, eosinophilia, peripheral nerve damage, sclerodermiform skin lesions, sicca syndrome, alopecia and joint contractures, among other findings"
    explanation: Names alopecia among the intermediate and chronic phase features.
- category: Cardiovascular
  name: Thromboembolism
  phenotype_term:
    preferred_term: Thromboembolism
    term:
      id: HP:0001907
      label: Thromboembolism
  evidence:
  - reference: PMID:1654352
    reference_title: Extracardiac vascular and neural lesions in the toxic oil syndrome.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Thromboembolic complications perpetuate the vascular lesion and compound the ischemia and parenchymal atrophy of several organs."
    explanation: Names thromboembolic complications as a feature of the vascular disease.
environmental:
- name: Ingestion of aniline-denatured rapeseed oil sold as cooking oil
  description: >-
    Rapeseed oil imported under a 2% aniline denaturant for industrial use,
    re-refined to strip the denaturant, and distributed in unlabelled 5-litre
    containers by itinerant salesmen in central and north-western Spain in 1981.
  exposure_term:
    preferred_term: exposure to a food adulterated with an industrial denaturant
    term:
      id: ECTO:9002126
      label: exposure to environmental food contaminant
  influences_mechanisms:
  - target: Ingestion of Aniline-Denatured Rapeseed Oil
    environmental_effect: TRIGGERS
    causal_link_type: DIRECT
    description: >-
      The exposure is the disease's sole cause; withdrawal of the oil from sale
      ended the epidemic and there has been no ongoing incidence since.
    evidence:
    - reference: PMID:10069247
      reference_title: Epidemiologic evidence for a new class of compounds associated with toxic oil syndrome.
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      quote_role: BACKGROUND
      snippet: "Epidemiologic studies have demonstrated that illness was caused by consumption of rapeseed oil that had been denatured with aniline."
      explanation: >-
        States the exposure-disease link this edge asserts. BACKGROUND because
        it is this chemistry paper's framing of the established epidemiology
        rather than its own analysis.
  evidence:
  - reference: PMID:12192735
    reference_title: 'Toxic oil syndrome: the perspective after 20 years.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "The vehicle of the causative toxic agent was identified as an illicit oil that had been diverted from industrial use and refined in order to remove the aniline denaturant, and that was sold in unlabeled 5-liter containers by itinerant salesmen."
    explanation: The exposure, its provenance and its distribution route in one sentence.
  notes: >-
    ECTO has an `exposure to aniline` term (`ECTO:9000980`, found with
    `runoak -i ols:ecto search "exposure to aniline"`), and it is deliberately
    not used: `PMID:1869734` states that aniline itself did not cause the
    illness. Binding it would assert the one thing the sources rule out. The
    bound term names the concept the epidemiology actually supports - a food
    carrying an industrial contaminant.
genetic:
- name: HLA class II association with fatal disease
  notes: >-
    Susceptibility and severity have repeatedly been linked to HLA, but the
    findings do not all agree, and the best-designed study found the association
    only in those who died. No causal gene is established, and TOS is not
    heritable - these are host susceptibility factors for an environmental
    exposure.
  relationship_type: MODIFIER
  gene_term:
    preferred_term: HLA-DRB1
    term:
      id: hgnc:4948
      label: HLA-DRB1
  evidence:
  - reference: PMID:10746782
    reference_title: DR2 antigens are associated with severity of disease in toxic oil syndrome (TOS).
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "In contrast, an increase in phenotypic frequency of DR2 antigen, was found in patients who had died from TOS (73.5%)"
    explanation: >-
      The positive finding, and it is specifically about patients who died
      rather than about susceptibility to the disease.
  - reference: PMID:10746782
    reference_title: DR2 antigens are associated with severity of disease in toxic oil syndrome (TOS).
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Regarding surviving patients no significant association was found between HLA and disease."
    explanation: >-
      SUPPORT, not REFUTE, and the distinction is the whole point of this
      record. The absence of any association among survivors is what confines
      the DR2 finding to fatal disease, which is exactly what this record
      claims and why it is typed MODIFIER. The same sentence is carried as
      REFUTE on the susceptibility record below, where it does cut against the
      claim being made.
- name: HLA class I and DR4-DQ8 association with susceptibility
  notes: >-
    An earlier case series reported HLA-A24, the DR4-DQ8 haplotype and DQ-alpha
    arginine 52 as susceptibility markers, and read the pattern as supporting an
    autoimmune classification.
  relationship_type: RISK_FACTOR
  gene_term:
    preferred_term: HLA-A
    term:
      id: hgnc:4931
      label: HLA-A
  evidence:
  - reference: PMID:8803534
    reference_title: Frequencies of HLA-A24 and HLA-DR4-DQ8 are increased and that of HLA-B blank is decreased in chronic toxic oil syndrome.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "In this paper it is also established that a human class I antigen (HLA-A24) and, independently, an HLA class II haplotype (DR4-DQ8, Pcorrected = 0.04) and arginine 52 in the alpha-DQ chains (Pcorrected = 0.03) are associated with TOS susceptibility, similarly to insulin-dependent diabetes."
    explanation: >-
      The reported associations with their corrected p-values. The gene binding
      is HLA-A because A24 is a serotype of that locus; allele-level detail in
      a gene field follows the HLA-B27 precedent in this repository.
  - reference: PMID:10746782
    reference_title: DR2 antigens are associated with severity of disease in toxic oil syndrome (TOS).
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Regarding surviving patients no significant association was found between HLA and disease."
    explanation: >-
      REFUTE against this record's susceptibility claim. A later, larger
      case-control study with family and unrelated controls found no HLA
      association among surviving patients at all, which is the population
      a susceptibility claim is about.
treatments:
- name: Corticosteroid Therapy
  description: >-
    Used in the acute and subacute phases. It relieves symptoms without
    changing the course of the disease, which is the conclusion the WHO
    workshop reached for both TOS and eosinophilia-myalgia syndrome.
  therapeutic_modality: SMALL_MOLECULE
  treatment_term:
    preferred_term: Pharmacotherapy
    term:
      id: NCIT:C15986
      label: Pharmacotherapy
    therapeutic_agent:
    - preferred_term: corticosteroid
      term:
        id: CHEBI:50858
        label: corticosteroid
  evidence:
  - reference: PMID:8266108
    reference_title: 'Toxic oil syndrome and eosinophilia-myalgia syndrome: May 8-10, 1991, World Health Organization meeting report.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "Although treatment with corticosteroids has resulted in significant symptomatic relief in persons with either disorder, it does not alter the clinical course or long-term outcome."
    explanation: >-
      Supports the symptomatic benefit this treatment claims. The same sentence
      is carried again below as REFUTE against disease modification, because it
      makes both claims at once.
  - reference: PMID:8266108
    reference_title: 'Toxic oil syndrome and eosinophilia-myalgia syndrome: May 8-10, 1991, World Health Organization meeting report.'
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "Although treatment with corticosteroids has resulted in significant symptomatic relief in persons with either disorder, it does not alter the clinical course or long-term outcome."
    explanation: >-
      REFUTE against corticosteroids as disease-modifying therapy. There is
      deliberately no `target_mechanisms` edge: on this evidence the drug does
      not act on the pathograph, and recording one would assert an effect the
      only cited source denies.
  notes: >-
    The same quoted sentence appears twice with opposite `supports` values.
    That is the claim-relative behaviour `supports` is meant to have - the
    sentence supports symptomatic relief and refutes disease modification -
    and `evidence_source` is identical in both, so the grading gate is
    satisfied.
- name: Supportive Care
  description: >-
    The mainstay throughout: respiratory support in the acute pulmonary oedema,
    and management of the chronic complications. An earlier draft said no
    intervention trialled during or after the epidemic altered the course of
    the disease. Only corticosteroids are cache-backed for that claim here, so
    it is narrowed - the other agents the literature reports as trialled
    without benefit (azathioprine, penicillamine, plasmapheresis, vitamin E,
    superoxide dismutase) are not cited in this entry and are not asserted by
    it.
  treatment_term:
    preferred_term: Supportive Care
    term:
      id: NCIT:C15747
      label: Supportive Care
  evidence:
  - reference: PMID:8266108
    reference_title: 'Toxic oil syndrome and eosinophilia-myalgia syndrome: May 8-10, 1991, World Health Organization meeting report.'
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "Research into the etiologic agents, preferred treatments, and ways to avoid similar problems in the future is needed."
    explanation: >-
      INDIRECT, and quoted for what it concedes: a decade after the epidemic
      the WHO workshop still listed preferred treatment as an open research
      question, which is the situation that leaves supportive care as the
      mainstay. It is not a study of supportive care.
- name: Rehabilitation
  description: >-
    For the joint contractures and the neuropathic disability of the chronic
    phase, which are among the three complications the WHO workshop named as
    long-term.
  therapeutic_modality: BEHAVIORAL
  treatment_term:
    preferred_term: Rehabilitation
    term:
      id: NCIT:C15315
      label: Rehabilitation
  evidence:
  - reference: PMID:8266108
    reference_title: 'Toxic oil syndrome and eosinophilia-myalgia syndrome: May 8-10, 1991, World Health Organization meeting report.'
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "Long-term complications include pulmonary hypertension, peripheral neuropathies, and joint contractures."
    explanation: >-
      INDIRECT: the source establishes the disabling complications that
      rehabilitation addresses, not the efficacy of rehabilitation in this
      disease, which no cited source reports.
prevalence:
- population: Spain, 1981 epidemic cohort
  measure_type: CASES_IN_LITERATURE
  prevalence_class: NOT_YET_DOCUMENTED
  notes: >-
    A closed point-source epidemic with no ongoing incidence, so a population rate
    would be misleading and no numeric slot is populated. prevalence_class is
    NOT_YET_DOCUMENTED because no rate has been documented, not because the
    epidemic was uncounted - the cohort size is documented, and appears in the
    evidence snippet below. Estimates differ slightly with the cohort-closure
    date used.
  evidence:
  - reference: PMID:8412642
    reference_title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "It affected 19,748 people, of whom 457 died."
    explanation: >-
      One cohort count with its own death toll. See the entry `notes:` for why
      this is not reconciled with the other published figures.
progression:
- phase: Acute phase
  notes: >-
    The first two months: non-cardiogenic pulmonary oedema, rash, eosinophilia and
    myalgia. Roughly half of patients across the epidemic recovered from this
    phase without apparent sequelae (PMID:1869734), against 9% achieving
    remission in the 332-patient cohort followed below. The two figures are
    not in conflict and are not averaged here: the first is the whole
    ~20,000-case epidemic, the second a referral cohort selected for longer
    follow-up.
  evidence:
  - reference: PMID:8412642
    reference_title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "The acute phase lasted 2 months, and was characterized by pulmonary edema, rash, eosinophilia, and myalgia."
    explanation: The duration and the four defining features.
- phase: Intermediate phase
  notes: >-
    Months two to four: severe myalgia, skin tenderness, subcutaneous oedema,
    altered liver function and pulmonary hypertension.
  evidence:
  - reference: PMID:8412642
    reference_title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "During the intermediate phase (second to fourth months), severe myalgia, skin tenderness, subcutaneous edema, altered liver function, and pulmonary hypertension developed."
    explanation: The window and its features.
- phase: Early chronic phase
  notes: >-
    Month four to the end of the second year: scleroderma, sicca syndrome,
    polyneuropathy, joint contractures, weight loss and functional limitation.
    Only 9% of patients remitted after the acute phase.
  evidence:
  - reference: PMID:8412642
    reference_title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Only 9% of the patients achieved remission after the acute phase, the rest developing late clinical manifestations of the disease."
    explanation: >-
      The proportion that did not progress, which is what makes the chronic
      phase the expected course rather than a complication.
- phase: Late chronic phase
  notes: >-
    Beyond two years, with partial improvement in many: muscle cramps, chronic
    musculoskeletal pain, chronic lung disease, Raynaud phenomenon, carpal
    tunnel syndrome and psychological disturbance.
  evidence:
  - reference: PMID:8412642
    reference_title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Its most prominent clinical features were muscle cramps, chronic musculoskeletal pain, chronic lung disease, Raynaud phenomenon, carpal tunnel syndrome, and psychologic disturbances."
    explanation: The late-phase feature set.
animal_models:
- name: HLA-DR2/DQ6 transgenic mouse fed toxic oil
  species: Mouse
  genotype: HLA-DR2 and HLA-DQ6 transgenic
  publication: PMID:15979827
  description: >-
    Built to test the human HLA restriction rather than to reproduce the
    disease. DR2 and DQ6 transgenic mice given toxic oil showed higher
    eosinophil percentages and IgE than DR3 or DR4 transgenics, which matches
    the direction of the human HLA finding.
  modeled_mechanisms:
  - target: T Cell Activation with Th2 Skewing in Lung Tissue
    relationship: PARTIALLY_RECAPITULATES
    fidelity: LOW
    model_scale: CELLULAR
    description: >-
      Reproduces the HLA-dependent immunological readouts, and nothing else
      about the disease.
    limitations: >-
      The model reports eosinophil percentage, IgE and a cytokine shift. It
      does not produce the endovasculitis that defines toxic oil syndrome, and
      no animal model does - see the `no_animal_model` discussion. Reading
      these mice as a model of the disease rather than of one immunological
      axis is the error the entry is guarding against.
    evidence:
    - reference: PMID:15979827
      reference_title: 'Toxic oil syndrome: genetic restriction and immunomodulatory effects due to adulterated oils in a model of HLA transgenic mice.'
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      quote_role: PRIMARY_RESULT
      snippet: "Results show that mice expressing human DR2 and DQ6 (both in linkage disequilibrium), had higher percentage of eosinophils (DQ6) and IgE (DR2) than other transgenic mice tested (DR3 and DR4)."
      explanation: >-
        The measured result, and the reason the link is PARTIALLY rather than
        fully recapitulating: two immunological readouts, no disease.
- name: Mouse given fatty acid anilides and the linoleic PAP diester
  species: Mouse
  genotype: wild type
  publication: PMID:10650923
  description: >-
    Intraperitoneal administration of the two suspect compound classes. The
    closest any animal has come to the acute human illness, and the authors
    are careful about how close that is.
  modeled_mechanisms:
  - target: Peripheral Eosinophilia
    relationship: PARTIALLY_RECAPITULATES
    fidelity: LOW
    model_scale: ORGANISM
    description: >-
      Reproduces the blood eosinophilia and lung pathology of the acute phase
      from the candidate compounds, without the vasculitis that defines the
      disease.
    limitations: >-
      Intraperitoneal rather than oral, which is the route the disease took,
      and the companion rat study found intragastric PAP produced no toxicity
      at all. The endovasculitis is absent. The authors say the effects
      "resemble" the acute human disease and explicitly note the full spectrum
      was not produced.
    evidence:
    - reference: PMID:10650923
      reference_title: 'The acute pathology of fatty acid anilides and linoleic diester of 3-phenylamino-1,2-propanediol in mice: possible implication as aetiologic agents for the toxic oil syndrome.'
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      quote_role: PRIMARY_RESULT
      snippet: "Linoleic diester of PAP led to weight loss, haemorrhage, congestion and emphysema in the lungs and an increase in blood eosinophilia."
      explanation: >-
        The measured findings, including the eosinophilia this link targets.
    - reference: PMID:10650923
      reference_title: 'The acute pathology of fatty acid anilides and linoleic diester of 3-phenylamino-1,2-propanediol in mice: possible implication as aetiologic agents for the toxic oil syndrome.'
      supports: REFUTE
      evidence_source: MODEL_ORGANISM
      quote_role: PRIMARY_RESULT
      snippet: "Although not producing the full spectrum of symptoms the effects of the substances resemble the acute human disease."
      explanation: >-
        REFUTE against reading this as a model of toxic oil syndrome. The
        authors' own qualification, and the reason the link is PARTIALLY at
        LOW fidelity rather than recapitulating.
- name: Rat given 3-phenylamino-1,2-propanediol
  species: Rat
  genotype: wild type
  publication: PMID:7779449
  description: >-
    A direct attempt to produce the disease with the leading candidate
    compound. It failed, and the way it failed is informative: the pathology
    it did produce was judged unrepresentative, and the oral route - the route
    of the actual epidemic - produced nothing at all.
  modeled_mechanisms:
  - target: In Situ Thrombosis on the Damaged Vessel Wall
    relationship: FAILS_TO_RECAPITULATE
    fidelity: LOW
    model_scale: ORGANISM
    description: >-
      Intraperitoneal PAP caused massive pulmonary thromboembolism in rats,
      but by thrombosis in mesenteric vessels that then embolised - not by the
      endovasculitis of the human disease.
    limitations: >-
      The authors state the pathology is not representative of human toxic oil
      syndrome. The mono-oleoyl ester caused no toxicity at all, and
      intragastric PAP caused none either, which is the finding that matters
      most: the human disease was caused by eating the compound.
    evidence:
    - reference: PMID:7779449
      reference_title: A comparison of the acute pathology induced by 3-phenylamino-1,2-propanediol (PAP) and its mono-oleoyl ester in rodents with the toxic oil syndrome in man.
      supports: REFUTE
      evidence_source: MODEL_ORGANISM
      quote_role: PRIMARY_RESULT
      snippet: "Comparatively, the pathology seen after intraperitoneal administration of PAP was not thought to be representative of the pathology of the toxic oil syndrome in man."
      explanation: >-
        The authors' own verdict on their model, which is what makes this a
        FAILS_TO_RECAPITULATE link rather than a partial one.
    - reference: PMID:7779449
      reference_title: A comparison of the acute pathology induced by 3-phenylamino-1,2-propanediol (PAP) and its mono-oleoyl ester in rodents with the toxic oil syndrome in man.
      supports: REFUTE
      evidence_source: MODEL_ORGANISM
      quote_role: PRIMARY_RESULT
      snippet: "Intra-gastric administration of PAP caused no toxicity in rats."
      explanation: >-
        The sharpest negative in the entry. The human disease came from eating
        the oil; giving rats the leading candidate compound by mouth did
        nothing. Any account of PAP as the agent has to explain this.
discussions:
- kind: KNOWLEDGE_GAP
  discussion_id: unidentified_etiologic_agent
  prompt: >-
    Which compound in the re-refined oil caused toxic oil syndrome, and by what
    molecular mechanism does it injure vascular endothelium?
  attaches_to:
  - pathophysiology#Systemic Exposure to PAP Fatty Acid Esters
  - pathophysiology#Vascular Endothelial Injury
  rationale: >-
    Forty years on, the agent is unidentified. The PAP esters carry the
    strongest association, their absorption and hepatic biotransformation are
    measured, and a metabolite is shared with eosinophilia-myalgia syndrome -
    but no compound has been shown to produce the disease, and the step from an
    absorbed ester to an injured endothelial cell is entirely unfilled. One
    review argues on metabolic grounds that there was never a single agent to
    find.
  evidence:
  - reference: PMID:12192735
    reference_title: 'Toxic oil syndrome: the perspective after 20 years.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "Although the exact identity of the etiologic agent in toxic oil syndrome remains unknown, work on toxic oil syndrome continues."
    explanation: States the gap directly, twenty years after the epidemic.
  - reference: PMID:11501225
    reference_title: 'The toxic oil syndrome: 20 years on.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "On metabolic evidence, it is suggested that not one, but a group of, toxic agents was responsible for TOS."
    explanation: >-
      Raises the possibility that the search has been for the wrong kind of
      answer - a group of agents rather than one - which would explain why no
      single compound has been convicted.
- kind: HUMAN_MODEL_MISMATCH
  discussion_id: no_animal_model
  prompt: >-
    Why has no animal species reproduced the pathology of toxic oil syndrome,
    and does that failure reflect a missing host factor rather than the wrong
    compound?
  attaches_to:
  - pathophysiology#Vascular Endothelial Injury
  - animal_models#Mouse
  rationale: >-
    This is the load-bearing gap in the whole disease, and it is not a
    KNOWLEDGE_GAP: evidence in model systems exists, it simply does not
    reproduce the human lesion. Because no animal develops the endovasculitis,
    candidate compounds cannot be tested by administering them, which is why
    the etiology has rested on epidemiological association with oil batches for
    four decades. A review of the problem proposed screening species and
    strains predisposed to vasculitis, eosinophilia and raised IgE - an
    approach that assumes the missing ingredient is host susceptibility rather
    than the wrong compound. The human HLA associations make that assumption
    reasonable and do not establish it.
  evidence:
  - reference: PMID:12176082
    reference_title: A search for an animal model of the Spanish toxic oil syndrome.
    supports: SUPPORT
    evidence_source: OTHER
    quote_role: REVIEW_SYNTHESIS
    snippet: "To date, pathology characteristics of toxic oil syndrome (TOS), a disease associated with consumption of a contaminated cooking oil in Spain in 1981, have not been reproduced in an animal model."
    explanation: >-
      States the mismatch. MEDLINE types this publication as a Review and it
      performed no animal experiment of its own, so the quote is the field's
      state as this review reports it - REVIEW_SYNTHESIS, and OTHER rather
      than MODEL_ORGANISM, which would assert an animal study that does not
      exist here.
  - reference: PMID:12176082
    reference_title: A search for an animal model of the Spanish toxic oil syndrome.
    supports: SUPPORT
    evidence_source: OTHER
    quote_role: REVIEW_SYNTHESIS
    snippet: "The intent was to determine predisposed strains or species that potentially might be effective in testing the toxic oils and thus defining the precise identity of the toxic contaminant(s)."
    explanation: >-
      Makes the dependency explicit - identifying the agent is blocked on
      having a model, which is why this gap and the etiology gap are the same
      problem from two sides. Graded as the same review's synthesis, for the
      reason given on the item above.
- kind: KNOWLEDGE_GAP
  discussion_id: cognitive_deficit_mechanism
  prompt: >-
    Is the cognitive impairment measured in toxic oil syndrome survivors four
    decades after exposure a direct neurotoxic sequela, or a downstream effect
    of chronic fatigue, mood symptoms and medication?
  attaches_to:
  - phenotypes#Cognitive impairment
  - phenotypes#Fatigue
  rationale: >-
    Two 2025 case-control studies, each 50 survivors against 50 matched controls,
    pull in opposite directions and neither settles it. Whether they report
    the same 50 people is not stated in either abstract and is not assumed
    here. Survivors perform worse on
    attention, executive function, processing speed and global cognition, but
    the difference disappears once fatigue, depression, anxiety and CNS-acting
    medication are adjusted for, and fatigue mediates most of it. In parallel, neurofilament light chain was slightly higher in survivors on the
    raw group comparison (p = 0.025) while GFAP and pTau217 were not, and
    clinical status predicted none of the three once age, sex and education
    entered the models. So there is a measurable deficit with no biomarker
    correlate and a plausible non-neurotoxic explanation, which is three
    different things being unresolved at once rather than one.
  evidence:
  - reference: PMID:40507935
    reference_title: 'Blood Biomarkers of Neurodegeneration over Four Decades After Toxic Oil Syndrome: A Case-Control Study.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Clinical status (TOS vs. control) did not significantly predict biomarker concentrations in any model."
    explanation: >-
      The negative biomarker result stated at its strongest and narrowest -
      no predictive effect of disease status in any model.
  - reference: PMID:40507935
    reference_title: 'Blood Biomarkers of Neurodegeneration over Four Decades After Toxic Oil Syndrome: A Case-Control Study.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "However, the possibility of subtle, compartmentalized, or slowly evolving neurotoxic processes cannot be excluded."
    explanation: >-
      The authors' own limit on their negative result, quoted so the gap is
      recorded as open rather than closed against neurotoxicity.
- kind: KNOWLEDGE_GAP
  discussion_id: fibrosis_driver_unknown
  prompt: >-
    What drives the dermal and fascial fibrosis of the chronic phase, given
    that the growth factors responsible in eosinophilia-myalgia syndrome are
    absent from toxic oil syndrome skin?
  attaches_to:
  - pathophysiology#Dermal and Fascial Fibrosis
  rationale: >-
    The two epidemics are routinely discussed as near-identical, and the
    temptation is to carry the eosinophilia-myalgia syndrome mechanism across.
    An immunohistochemical comparison blocks that: TGF-beta and PDGF-AA were
    present in the periappendageal dermis of 57% of EMS specimens and 0% of
    toxic oil syndrome specimens. Whatever drives the fibrosis here is not the
    driver there, and it has not been identified. The specimen count is small,
    so this is a finding to explain rather than a settled negative.
  evidence:
  - reference: PMID:8285738
    reference_title: Fibrogenic growth factors in the eosinophilia-myalgia syndrome and the toxic oil syndrome.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "Seven skin biopsy specimens from EMS, six skin biopsy specimens from toxic oil syndrome, nine muscle biopsy specimens from EMS, and one sural nerve biopsy specimen from EMS were studied."
    explanation: >-
      The specimen counts, quoted because the gap's force depends on them: six
      TOS skin biopsies is a small denominator for a 0% finding, and the
      rationale says so rather than leaving the reader to assume otherwise.
clinical_burden:
  burden_level: HIGH
  rationale: >-
    Around 20,000 people were affected in a single point-source epidemic, most
    of whom did not remit after the acute phase, and mortality in the cohort was
    8.4% over the first thirteen years. Pulmonary hypertension is one of only
    two causes of death that were not decreased relative to the general Spanish
    population - the cohort's overall mortality was lower than expected, so
    this is one cause spared a general decline rather than a demonstrated
    excess - and new cases of it were still being diagnosed in exposed
    survivors decades on.
  evidence:
  - reference: PMID:10024203
    reference_title: 'Toxic oil syndrome mortality: the first 13 years.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "We identified 1663 deaths between 1 May 1981 and 31 December 1994 among 19 754 TOS cohort members, for a crude mortality rate of 8.4%."
    explanation: The cohort size and its thirteen-year crude mortality.
  - reference: PMID:10024203
    reference_title: 'Toxic oil syndrome mortality: the first 13 years.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "except for deaths attributed to external causes including TOS and deaths due to pulmonary hypertension, all causes of death were decreased in TOS patients compared to the Spanish population"
    explanation: >-
      Identifies pulmonary hypertension as one of only two causes of death not
      decreased relative to the Spanish population. Note what the sentence
      does and does not say: "not decreased" is weaker than "raised", and the
      same abstract reports the cohort's overall mortality as less than
      expected. It still makes this the burden-defining phenotype, because
      every other cause moved the other way.
references:
- reference: PMID:12192735
  title: "Toxic oil syndrome: the perspective after 20 years."
- reference: PMID:8412642
  title: Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
- reference: PMID:1654352
  title: Extracardiac vascular and neural lesions in the toxic oil syndrome.
📚

References & Deep Research

References

3
PMID:12192735
Toxic oil syndrome: the perspective after 20 years.
No top-level findings curated for this source.
PMID:8412642
Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients.
No top-level findings curated for this source.
PMID:1654352
Extracardiac vascular and neural lesions in the toxic oil syndrome.
No top-level findings curated for this source.

Deep Research

1

Deep research results are used as seeds for research; they do not undergo the same validation as the main records and may contain errors. How we use deep research.

Evaluations and curation notes (2)

Record notes

**The etiologic agent is unidentified and this entry does not assert one.** The pathograph names the PAP esters because they carry the strongest epidemiological association with case-related oil and because their absorption, lipase hydrolysis and hepatic biotransformation have each been measured. That is an association plus a plausible route, not a demonstrated cause, and the `unidentified_etiologic_agent` discussion records what is missing. No animal has developed the disease, though two rodent studies cited here got partway and are curated under animal_models: anilides and a PAP diester reproduced weight loss, lung pathology and blood eosinophilia in mice, and PAP produced pulmonary thromboembolism in rats that the authors judged unrepresentative of the human pathology. **Mortality figures in this entry are not interchangeable and are deliberately reported with their windows attached.** The sources count different things over different periods: 457 deaths in a cohort description (`PMID:8412642`), over 300 in the first year (`PMID:1869734`), 1,663 deaths from all causes among 19,754 cohort members by the end of 1994 (`PMID:10024203`), and over 1,200 all-cause deaths in the 20-year review (`PMID:12192735`). These are not four estimates of one quantity, and the entry does not average or reconcile them. **Most phenotypes here are deliberately left unwired, and the rule is the same for all of them: an edge goes in only where a cited source makes the causal link.** Run `just list-disconnected-phenotypes kb/disorders/Toxic_Oil_Syndrome.yaml` for the current set rather than trusting a list in this note - an earlier version of this paragraph enumerated them and was wrong within the hour, because adding a phenotype silently falsifies a count. Three cases are worth naming. The constitutional features (fever, rash, weight loss) have no named lesion in any cited source. Cognitive impairment and fatigue sit on the open question the `cognitive_deficit_mechanism` discussion states, and the one candidate route in the literature, fatigue mediating the cognitive score, is a structural equation result rather than a mechanism. Myalgia is the pointed one: it is among the two features that defined the epidemic, the muscle shows an interstitial inflammatory myopathy, and it would be easy to draw an edge between them - but the myalgia of this disease and of eosinophilia-myalgia syndrome is conspicuously not explained by the muscle pathology, and no source here asserts that it is. Connectivity in this entry is correspondingly low. An edge drawn to raise that figure would be worse than the gap. **Two scope exclusions.** Eosinophilia-myalgia syndrome is not curated here. It is a separate epidemic with a separate vehicle (L-tryptophan) that resembles TOS closely enough that the two are routinely discussed together, and several references cited here are explicitly comparative; the comparison is used where a source makes it, and no EMS finding is imported as a TOS finding. The 2025 long-term survivor studies *are* curated, as `Cognitive impairment` and `Fatigue` and the `cognitive_deficit_mechanism` discussion. The deep-research report cites them only by PMC identifier and web link; their PubMed records (`PMID:40507507`, `PMID:40507935`) were recovered when the report's citations were resolved, which is why they are quotable here. No GeneReviews chapter exists and none is expected. This is an acquired point-source intoxication with no Mendelian basis; `just check-genereviews` returns NO_CHAPTER for both Bookshelf collections.

Create: Toxic Oil Syndrome (MONDO:0016421) · 2026-09-19T02:30:57Z · View source

New entry for the 1981 Spanish toxic oil syndrome epidemic. Deep research via the claude_code provider (the only one with credentials in this environment; falcon, asta, openscientist, perplexity and openai keys were all unset). Report validated retrospectively with just validate-research-reference (29 references checked, 28 resolved, 0 unresolved, 0 off topic) and just validate-research-terms (61 terms, 58 resolved, 0 unresolved, 2 obsolete and unused here). just preflight-dr returned SKIP because MONDO records no causal gene for a non-genetic disease; disease identity was checked manually against the report body instead. Every ontology CURIE was re-derived from a live OLS lookup at the point of writing and none taken from the report - which was the right call twice over: the report offered CL:0002548 as 'fibroblast of connective tissue' when it is fibroblast of cardiac tissue, and HP:0001880 is 'Increased total eosinophil count', not 'Eosinophilia' as memory suggested. Substantive curation decision: the report cites PMID:8285738 as evidence that TGF-beta and PDGF-AA drive the dermal fibrosis, and the paper found them in 0 percent of toxic oil syndrome skin specimens against 57 percent of eosinophilia-myalgia specimens, concluding the two diseases depend on different growth factors. That inversion is carried as a REFUTE item and the upstream edge to the fibrosis node uses INDIRECT_UNKNOWN_INTERMEDIATES because of it; the fibrosis_driver_unknown discussion records the gap. Four mechanistic_hypotheses are curated (direct toxic endothelial injury as CANONICAL, immune-mediated initiation as ALTERNATIVE, PAF-analogue signalling as EMERGING, and the minority rapeseed-oil-itself account), because the sources genuinely disagree on what initiates the lesion. Three discussions: the unidentified etiologic agent, the absence of any animal model (HUMAN_MODEL_MISMATCH, not KNOWLEDGE_GAP - evidence in animals exists and fails to reproduce the human lesion), and the unexplained cognitive deficit. A red-team review by a fresh-context subagent against dismech-pr-review returned 19 findings and all were applied. The ones worth recording: two self-contradictions I introduced mid-review (an evidence explanation referring to a reports_on link I had just deleted, and a notes paragraph counting six unwired phenotypes after I had added two more); five evidence-grading errors where the entry's own prose already stated the correct answer, including two items on MEDLINE-typed Review publications graded PRIMARY_RESULT; a mortality claim overstated in three places as pulmonary hypertension being 'raised' relative to the Spanish population when the cited sentence says only that it was not decreased, and the same abstract reports overall cohort mortality as lower than expected; both HLA records typed SUSCEPTIBILITY when the schema enum files HLA risk alleles under RISK_FACTOR and severity-only findings under MODIFIER; and a REFUTE item pointed at a claim the record it hung on did not make. The reviewer also found two rodent papers I had fetched and never read, which qualify the entry's absolute claim that no animal developed the disease - both are now curated as animal_models, and PMID:7779449 supplies the sharpest negative in the entry: intragastric PAP caused no toxicity in rats, though the human disease came from eating the oil. The phenotype-unwired note was rewritten to state the rule and point at just list-disconnected-phenotypes rather than enumerate, because the enumeration falsified itself within the hour. Every reference_title in the file is written programmatically from the cache frontmatter after I twice completed a truncated title from memory and the title gate caught it. Validation: just validate-disorders passes with 88/88 snippets verified; schema, terms, references, entity-refs, causal-targets, duplicate-keys, enum-values, qualifier-terms (offline and --resolve), folded-hyphens, not4curation and case-collisions all clean. just check-genereviews returns NO_CHAPTER for both Bookshelf collections, which is what the entry notes claim. Phenotype connectivity is deliberately low at 9/21.

Claude Code ▸
Toxic Oil Syndrome (TOS): Comprehensive Disease Characteristics Report
claude-haiku-4-5-20251001, claude-sonnet-5 47 citations 2026-09-19T01:45:58.600108

Toxic Oil Syndrome (TOS): Comprehensive Disease Characteristics Report

1. Disease Information

Overview. Toxic Oil Syndrome (TOS, Spanish: síndrome del aceite tóxico, colloquially "lo de la colza") is a multisystem toxic-immunologic disease caused by ingestion of illegally sold, industrial rapeseed (colza) oil that had been denatured with 2% aniline for non-food/industrial use and then illegally re-refined and sold door-to-door as cheap "olive oil" in Spain in spring 1981 (PubMed 6116011; PubMed 6633617). It is a point-source environmental/toxicologic epidemic rather than an infectious or classically genetic disease, and virtually all epidemiological, clinical, and mechanistic data derive from a single, extensively studied Spanish cohort (aggregated disease-level surveillance/registry data plus individual clinical follow-up studies), not from ongoing EHR-based case ascertainment.

The illness is a three-phase syndrome: an acute toxic-allergic pneumonopathy (respiratory distress, interstitial/alveolar infiltrates, fever, myalgia, rash, eosinophilia), an intermediate phase (peripheral edema, skin induration, hepatic dysfunction, pulmonary hypertension), and a chronic phase (scleroderma-like skin sclerosis, peripheral neuropathy, joint contractures, sicca syndrome) that can persist for decades (ScienceDirect S0190962288700468; PubMed 3961509).

Key identifiers: - MONDO: MONDO:0016421 - Orphanet: ORPHA:227972 ("Toxic oil syndrome") — described as "a rare intoxication, due to consumption of a rapeseed oil denatured with aniline 2%, characterized by generalized vascular lesions affecting all organs and vessels... and presenting with severe incapacitating myalgias, marked peripheral eosinophilia and pulmonary infiltrates" (Orphanet ORPHA:227972) - MeSH: "Toxic Oil Syndrome" (D019867) - ICD-10/ICD-11: No dedicated code identified in searched sources; historically coded under toxic-effect/adverse-event categories (e.g., ICD-9-CM 989.89-adjacent "toxic effect of other substances") rather than a named entity. Confirm exact ICD-10-CM/ICD-11 mapping via direct database lookup before curation. - OMIM: Not a Mendelian disorder; no OMIM phenotype number.

Synonyms: Spanish toxic oil syndrome; Spanish toxic-oil syndrome; toxic-allergic syndrome caused by ingestion of rapeseed oil denatured with aniline; "colza oil syndrome"; "lo de la colza."

2. Etiology

Primary causal factor — environmental/toxicologic, not genetic or infectious. The causal agent is the ingestion of illegally re-refined rapeseed oil that had been industrially denatured with aniline. The syndrome does not resemble classic aniline or acetanilide intoxication; instead it has been provisionally, and then more specifically, attributed to reaction products formed between aniline/acetanilide and the oil's fatty-acid components — termed collectively "oleoanilides" (PubMed 6116011). Subsequent case-control chemical epidemiology strongly implicated a specific chemical class: fatty-acid esters of 3-(N-phenylamino)-1,2-propanediol (PAP), byproducts of the aniline-oil reaction, as the most probable etiologic agents (PubMed 10069247; PubMed 7918809).

Risk factors: - Environmental/exposure: purchase of oil from itinerant door-to-door salesmen rather than licensed retail outlets was the dominant exposure risk — in one household study in the Orcasur district of Madrid, all affected households had purchased oil from traveling salesmen, versus only 34% of unaffected households (Grokipedia summary; "Lo de la colza" — Nursing Clio). - Geographic: concentrated in Madrid province (~71% of the ~14,292–19,828 reported cases) and 13 other central/northwestern Spanish provinces; incidence exceeded 300 cases/100,000 in Segovia and Palencia (NEJM 1983). - Sex/age: of the long-term surveillance cohort of 20,084 subjects, 60.6% were women and 39.4% men; TOS was the leading cause of death among subjects under 40 years of age, with the shortest post-onset survival among women and younger patients (ScienceDirect S0895435603001197). - Genetic susceptibility (host modifier, not causal): HLA class I/II alleles modulate severity and chronicity (see Section 4/9). - Storage/dose-persistence: "late cases" of TOS occurred among people who consumed oil that had been stored for up to a year, indicating the etiologic agent(s) persisted in stored oil over time (ScienceDirect 0278691589900471).

Protective factors: No genetic or dietary protective factor has been robustly established; certain HLA haplotypes were associated with lower likelihood of chronic/severe disease relative to DR2/DR4-DQ8/A24 carriers (see below), which functions as a relative rather than absolute protective association.

Gene–environment interaction: This is the central etiologic feature of TOS beyond the initial toxic exposure — identical oil exposure produced markedly different clinical trajectories (self-limited acute illness vs. progression to chronic scleroderma-like disease vs. death) that correlate with host HLA genotype, indicating a gene–environment interaction in which a xenobiotic (oleoanilide/PAP-ester) triggers an aberrant, HLA-restricted immune response in susceptible individuals (ScienceDirect S0378427405003103; PubMed 15979827).

3. Phenotypes

Phenotypes are drawn from Spanish Clinical Commission (Ministry of Health, August 1981) case criteria and subsequent cohort/case-series literature. Frequencies below are qualitative characterizations from the literature; a curated entry should mine the NEJM 1983 clinical-epidemiology paper and the 14-case immunopathologic series for precise percentages.

Phenotype Type Phase Suggested HP term
Fever Sign Acute HP:0001945 Fever
Cough / dyspnea Symptom Acute HP:0002090 Bronchitis; HP:0002094 Dyspnea
Pulmonary infiltrates (interstitial/alveolar) Radiographic sign Acute HP:0002088 Abnormal pulmonary interstitial morphology
Pleural effusion Sign Acute HP:0002202 Pleural effusion
Peripheral (blood) eosinophilia Lab abnormality Acute–chronic HP:0001880 Eosinophilia
Myalgia (often severe/incapacitating) Symptom Acute–chronic HP:0003326 Myalgia
Skin rash Sign Acute HP:0000988 Skin rash
Hepatosplenomegaly Sign Acute HP:0001433 Hepatosplenomegaly
Generalized lymphadenopathy Sign Acute HP:0002716 Lymphadenopathy
Peripheral edema Sign Intermediate HP:0000969 Edema
Skin induration / scleroderma-like sclerosis Sign Chronic HP:0100678 Skin plaque; HP:0100692 Skin nodule (or free text — no exact HP scleroderma-secondary term)
Hepatic dysfunction / liver disease Lab/clinical Intermediate–chronic HP:0001392 Abnormality of the liver
Pulmonary arterial hypertension Sign Intermediate–chronic HP:0002092 Pulmonary hypertension
Sicca syndrome (dry eyes/mouth) Symptom Chronic HP:0031000 Dry eye; HP:0031416 Dry mouth
Peripheral (sensorimotor) neuropathy Sign Chronic HP:0009830 Peripheral neuropathy
Joint contractures Sign Chronic HP:0034680 / HP:0001371 Flexion contracture
Raynaud phenomenon Symptom Chronic HP:0025595 Raynaud phenomenon
Livedo reticularis Sign Chronic HP:0100672 Livedo reticularis
Carpal tunnel syndrome Sign Chronic HP:0100628 Carpal tunnel syndrome
Dysphagia Symptom Chronic HP:0002015 Dysphagia
Alopecia Sign Chronic HP:0001596 Alopecia
Cachexia / weight loss Sign Intermediate–chronic HP:0004325 Decreased body weight
Fatigue Symptom Chronic (persistent) HP:0012378 Fatigue
Cognitive difficulties (frontal-subcortical) Symptom Chronic (decades later) HP:0100543 Cognitive impairment
Psychiatric complaints Symptom Chronic (context-dependent; no single HP term)

Characteristics: - Onset: adult-onset, epidemic point-source exposure (interval between ingestion of contaminated oil and symptom onset of 4–10 days) (ScienceDirect/CHEST summary). - Severity/frequency: roughly half of the ~20,000 affected individuals recovered from the acute phase without apparent sequelae; the remainder progressed to intermediate and/or chronic disease with severe myalgia, eosinophilia, peripheral nerve damage, sclerodermiform skin lesions, sicca syndrome, alopecia, and joint contractures (ScienceDirect overview). - Progression: staged and largely unidirectional (acute → intermediate → chronic), though partial recovery has been documented in chronic-phase patients beyond 2 years post-onset. - Quality of life (long-term): a 2022 SF-36 health-related quality-of-life study in survivors documented persistently reduced physical and mental health domain scores relative to the general population decades after exposure (IJE 2022). Among 91 individuals re-evaluated more than a decade after exposure, over half reported persistent fatigue, muscle cramps, arthralgias, subjective cognitive difficulties, and psychiatric complaints.

4. Genetic/Molecular Information

TOS has no causal single-gene etiology — it is a toxin-triggered disease — but host genetics substantially modifies severity and chronicity via HLA:

  • HLA-DR2: phenotypic frequency was markedly increased in patients who died of TOS (73.5%) compared with TOS-affected survivors (25.6%), unaffected family members (28.5%), unrelated controls (23.9%), and controls who died of other causes (38.4%) — i.e., DR2 tracks with fatal/severe disease (PubMed 10746782).
  • HLA-DR4-DQ8 and HLA-A24: independently associated with susceptibility to chronic TOS; HLA-B blank (homozygosity/no detected B antigen) frequency was decreased in chronic TOS (PubMed 8803534).
  • DQα-chain arginine-52: an amino-acid polymorphism in the DQ α-chain (Arg52) was independently associated with TOS susceptibility, analogous to a mechanism described in other HLA-linked autoimmune/toxin-triggered diseases (ScienceDirect S0378427405003103).
  • HLA-transgenic mouse functional validation: HLA-DR2/DQ6 transgenic mice exposed to TOS-implicated oils showed higher eosinophil percentages (DQ6-linked) and IgE levels (DR2-linked) than DR3- or DR4-transgenic mice, supporting a genetically restricted immunomodulatory mechanism (PubMed 15979827).

Molecular etiologic agent(s)/chemical entities (CHEBI candidates for curation): - Oleoanilides — the general class of fatty-acid–acetanilide reaction products originally implicated. - 3-(N-phenylamino)-1,2-propanediol (PAP) and its fatty-acid esters (mono- and di-oleoyl esters, linoleic diester) — the leading specific etiologic candidates, epidemiologically linked to TOS-implicated oils (PubMed 10069247; Lipids journal). - 3-(Phenylamino)alanine (PAA) — a biotransformation product of PAP (demonstrated in rat hepatocytes and human liver tissue), mechanistically linking TOS to the chemically distinct but clinically overlapping 1989 U.S. eosinophilia-myalgia syndrome (EMS), in which PAA was found as a contaminant of implicated L-tryptophan lots (PubMed 8555405; Mayo Clin Proc). - Incubation of PAP with human liver microsomes generates a reactive quinoneimine intermediate, implicating a reactive-metabolite bioactivation mechanism (search summary, per Chem Res Toxicol 8(7):911). - Structural analogy between PAP-diesters and platelet-activating factor (PAF) has been proposed as contributing to the eosinophilic/inflammatory response (Toxicology, S0378427496038623).

There is no described pathogenic germline variant, no described epigenetic signature specific to TOS, and no chromosomal abnormality associated with the disease — all molecular data concern the exogenous toxin and its host-immune interaction rather than a heritable lesion.

5. Environmental Information

  • Primary environmental factor: consumption of illegally distributed, industrially aniline-denatured (2%) rapeseed oil that was re-refined by unlicensed processors to remove color/odor and then fraudulently sold as edible/olive oil (PubMed 6116011). This is a discrete, non-recurring environmental/regulatory-failure exposure rather than an ongoing occupational or ambient toxin.
  • Persistence: the causal toxin(s) persisted in stored oil for up to a year, generating "late cases" after the initial 1981 outbreak (ScienceDirect 0278691589900471).
  • Lifestyle factor: reliance on door-to-door itinerant oil vendors (a socioeconomic/behavioral exposure route) rather than licensed retail supply chains was the operative lifestyle/consumer-behavior risk factor.
  • Infectious agents: none — TOS is not infectious and has no described microbial trigger or co-factor.

6. Mechanism / Pathophysiology

Ordered causal chain (as currently understood; several links are inferred rather than fully demonstrated):

  1. Ingestion of illegally re-refined rapeseed oil denatured with aniline leads to systemic exposure to reaction byproducts formed between aniline/acetanilide and oil fatty-acid components ("oleoanilides"), most specifically fatty-acid esters of 3-(N-phenylamino)-1,2-propanediol (PAP) — inferred from epidemiologic case-control chemistry, not directly demonstrated in exposed humans at the time of the epidemic (PubMed 10069247).
  2. Hepatic/microsomal metabolism of PAP results in bioactivation to a reactive quinoneimine intermediate and to the metabolite 3-(phenylamino)alanine (PAA) — demonstrated in vitro in rat hepatocytes and human liver tissue (PubMed 8555405).
  3. PAP/PAA and their esters trigger activation of circulating polymorphonuclear leukocytes, including generation of reactive oxygen metabolites in human PMNs — demonstrated in vitro (ScienceDirect S0378427496038623) — and structurally resemble platelet-activating factor (PAF), which may contribute to eosinophil recruitment and vascular permeability changes — inferred from structural analogy.
  4. In genetically susceptible individuals (particularly those carrying HLA-DR2, HLA-DR4-DQ8, or HLA-A24, or the DQα-Arg52 polymorphism), this toxic exposure drives an aberrant, HLA-restricted adaptive immune response with a Th2-skewed cytokine profile (elevated IL-4, IL-5 relative to IFN-γ in affected lung tissue) — demonstrated by cytokine mRNA expression analysis of TOS lung biopsies (PubMed 9074654).
  5. The Th2-skewed response leads to marked peripheral and tissue eosinophilia, elevated IgE, and T-cell activation — demonstrated in patient and transgenic-mouse studies (PubMed 15979827).
  6. Concurrently, the toxin(s) cause direct or immune-mediated endothelial injury in both small and large vessels across multiple organs, with pronounced intimal proliferation (edematous, with vacuolated cells in the media and loss of vascular smooth muscle in elastic pulmonary arteries) and medial hypertrophy with intimal proliferation in muscular pulmonary arteries, in one reported case severe enough to completely occlude the arterial lumen — demonstrated histopathologically (PMC459646; Virchows Arch, pathology of new toxic syndrome).
  7. Pulmonary vascular endothelial injury and remodeling result in pulmonary arterial hypertension, a major driver of intermediate- and chronic-phase morbidity and mortality — demonstrated clinically and by autopsy series.
  8. Perivascular and interstitial inflammatory infiltration and eosinophil-derived mediators drive release of fibrogenic growth factors, notably TGF-β and PDGF-AA, which lead to progressive tissue fibrosis — demonstrated in comparative EMS/TOS tissue studies (PubMed 8285738).
  9. Dermal and fascial fibrosis manifests as scleroderma-like skin sclerosis, joint contractures, and sicca syndrome in the chronic phase, overlapping clinically and histologically with idiopathic systemic sclerosis and related connective-tissue diseases — demonstrated clinically (PubMed 3961509).
  10. Concurrent nerve involvement (mechanism less well defined — possibly ischemic/microvascular and/or direct toxic/immune injury to peripheral nerve) leads to chronic peripheral (sensorimotor) neuropathy — largely inferred from clinical/electrophysiologic observation rather than a fully worked-out cellular mechanism.
  11. In a minority of long-term survivors, ongoing frontal-subcortical circuit dysfunction is detectable by eye-tracking and cognitive testing four decades after exposure, but blood biomarkers (NfL, GFAP, pTau217) do not show the elevation pattern typical of progressive neurodegenerative disease — suggesting a static or slowly evolving immune/vascular injury pattern rather than ongoing neurodegeneration — demonstrated in 2025 case-control biomarker and eye-tracking studies (PMC12155236; Frontiers 2025; PMC12155933).

Cell types/processes for ontology binding (suggested): - GO biological processes: GO:0006954 inflammatory response; GO:0043304 regulation of mast cell degranulation (context-dependent); GO:0030101 natural killer cell activation (if relevant); GO:0030099 myeloid cell differentiation (eosinophilopoiesis); GO:0001525 angiogenesis / vascular remodeling; GO:0030198 extracellular matrix organization (fibrosis); response to xenobiotic stimulus GO:0009410. - CL cell types: CL:0000771 eosinophil; CL:0000542 lymphocyte / CL:0000909 CD4-positive, alpha-beta T cell (Th2 subset CL:0000546); CL:0000115 endothelial cell; CL:0000499 stromal cell / CL:0002548 fibroblast of connective tissue (fibrosis); CL:0002139 endothelial cell of vascular tree. - UBERON: UBERON:0002048 lung; UBERON:0001981 blood vessel; UBERON:0002037 cerebellum (n/a) — more relevantly UBERON:0002370 thymus is not central; primary organs: UBERON:0002048 lung, UBERON:0002107 liver, UBERON:0000178 skin (integument), UBERON:0001021 nerve, UBERON:0001981 blood vessel, UBERON:0000948 heart (secondary, cor pulmonale from PAH). - CHEBI: candidate entries for "3-(phenylamino)-1,2-propanediol", "3-phenylaminoalanine", "acetanilide", "aniline" — verify exact CHEBI CURIEs via OAK/ChEBI lookup before binding (not independently confirmed in this research pass).

7. Anatomical Structures Affected

  • Primary organs: lungs (pneumonitis, pulmonary vascular disease, pulmonary hypertension); skin/subcutaneous tissue and fascia (edema → induration → scleroderma-like sclerosis); peripheral nervous system (sensorimotor neuropathy).
  • Secondary/systemic involvement: liver (hepatic dysfunction, hepatosplenomegaly, elevated relative risk of liver disease — RR 3.83 in long-term follow-up, ScienceDirect S0895435603001197); joints (contractures); exocrine glands (sicca/Sjögren-like syndrome); vasculature systemically (large- and small-vessel involvement, "generalized vascular lesions affecting all organs and vessels" per Orphanet); reticuloendothelial system (lymphadenopathy, splenomegaly); central/peripheral nervous system in the long-term (frontal-subcortical cognitive dysfunction).
  • Body systems: respiratory, cardiovascular (pulmonary hypertension, thromboembolism of great arteries), integumentary, musculoskeletal, hepatic, immune, peripheral/central nervous system.
  • Tissue/cell level: vascular endothelium and media (elastic and muscular pulmonary arteries), dermal fibroblasts/fascia, hepatocytes (site of PAP bioactivation), circulating eosinophils and T lymphocytes.
  • Subcellular: hepatic microsomal/cytochrome P450-associated bioactivation of PAP to reactive quinoneimine (endoplasmic reticulum, GO Cellular Component: GO:0005783).
  • Laterality: bilateral/systemic — not a lateralized disease.

8. Temporal Development

  • Onset: adult-onset, epidemic exposure; symptom onset 4–10 days after ingestion of contaminated oil (CHEST summary).
  • Pattern: acute onset of respiratory/systemic illness, evolving in a subset of patients through a staged, largely progressive course.
  • Stages (well defined in this literature, unusually so for an environmental disease):
  • Acute (~first 2 months): pulmonary edema/infiltrates, fever, rash, eosinophilia, myalgia.
  • Intermediate (approximately months 2–4): peripheral edema, dermal induration, hepatic dysfunction, pulmonary hypertension, sicca syndrome, hypertriglyceridemia, cachexia — reached by roughly 60% of all TOS patients.
  • Chronic (beyond month 4, with partial recovery possible after ~2 years in some patients): scleroderma-like cutaneous sclerosis, peripheral neuropathy, joint contractures, Raynaud phenomenon, sicca/Sjögren syndrome, carpal tunnel syndrome, dysphagia, persistent pulmonary hypertension.
  • Progression rate: variable — roughly half of patients recovered from the acute phase without sequelae; the remainder progressed, and progression through intermediate to chronic phase correlates with HLA genotype and disease severity markers.
  • Duration: for those developing chronic disease, TOS is a lifelong condition — persistent fatigue, myalgia, arthralgia, cognitive complaints, and psychiatric symptoms are documented in survivors assessed more than four decades after the 1981 exposure (PMC12484009; PMC12155933).
  • Critical period: the initial acute/subacute phase (first weeks to ~4 months) appears to be the critical window in which the eventual severity/chronicity trajectory is largely determined, correlating with HLA genotype.

9. Inheritance and Population

  • Epidemiology: ~19,828–20,084 cases reported in Spain in 1981–1982 (estimates vary slightly by source/cohort-closure date), concentrated in Madrid province (71% of cases) and 13 other central/northwestern provinces, with incidence exceeding 300/100,000 in Segovia and Palencia (NEJM 1983). This is a historical point-source epidemic with zero ongoing incidence since the causal oil was withdrawn from sale in 1981 — it is not an endemic or recurring disease.
  • Mortality: 315 deaths reported by June 1982 in early surveillance; longer surveillance to 1995 of the full 20,084-subject cohort recorded 1,799 total deaths, of which 356 were TOS-related (overall epidemic mortality historically cited around 8.4%, with figures up to 839 deaths cited in some secondary EMS-comparison sources — cohort definitions and follow-up windows differ across sources and should be reconciled against the primary registry paper before final figures are curated) (ScienceDirect S0895435603001197; PubMed 8412642).
  • Inheritance pattern: not a Mendelian/heritable disease — inheritance pattern is not applicable; susceptibility is modulated by HLA genotype (a polygenic host-modifier effect on a toxin-induced disease), not by a single causal locus.
  • Sex ratio: cohort was 60.6% female / 39.4% male; TOS-related deaths occurred with shortest survival times in women and in those under 40.
  • Age distribution: all ages affected; TOS was the leading cause of death in the affected cohort among those under age 40.
  • Geographic distribution: strictly confined to the distribution network of the adulterated oil within Spain — no cases outside Spain, and no cases since the source oil was withdrawn, making TOS a geographically and temporally bounded historical epidemic.
  • Genetic population considerations: no founder effect, consanguinity role, or carrier-frequency concept applies (non-heritable disease); HLA allele frequencies (DR2, DR4-DQ8, A24) in the Spanish population are the relevant "population genetics" consideration, functioning purely as severity/susceptibility modifiers of an exogenous exposure.

10. Diagnostics

  • Clinical case definition: formal diagnostic criteria were established by the Spanish Clinical Commission (Ministry of Health and Consumer Affairs) in August 1981, combining epidemiologic exposure history (consumption of oil from the implicated distribution channels) with the characteristic clinical/laboratory triad of respiratory symptoms, marked peripheral eosinophilia, and myalgia, in the appropriate geographic/temporal context (CHEST summary).
  • Laboratory tests: peripheral blood eosinophil count (hallmark finding); liver function tests (elevated in intermediate/chronic phase); IgE levels (elevated in genetically susceptible subgroups); triglycerides (elevated, intermediate phase).
  • Imaging: chest radiography demonstrating interstitial or alveolar infiltrates with or without pleural effusion in the acute phase; later imaging for pulmonary hypertension assessment (echocardiography historically, right heart catheterization for confirmation).
  • Functional tests: pulmonary function testing (restrictive pattern in advanced pulmonary fibrosis/hypertension); right-heart catheterization for pulmonary hypertension confirmation.
  • Biopsy/histopathology: skin biopsy showing scleroderma-like dermal/fascial fibrosis in chronic phase; lung/vascular histopathology (in fatal cases) showing the characteristic intimal proliferation and medial hypertrophy of pulmonary arteries described above (PMC459646).
  • Genetic testing: not diagnostic for TOS itself, but HLA typing (DR2, DR4-DQ8, A24) has been used in research settings as a severity/prognostic biomarker rather than a diagnostic test.
  • Emerging/research biomarkers: recent (2025) blood biomarker panels for neurodegeneration — neurofilament light chain (NfL), glial fibrillary acidic protein (GFAP), phosphorylated tau 217 (pTau217) — have been used in long-term survivor case-control research to characterize the chronic neurocognitive phenotype, showing no elevation pattern typical of progressive neurodegenerative disease (PMC12155236); eye-tracking has been used as an objective research measure of frontal-subcortical dysfunction (Frontiers 2025).
  • Differential diagnosis: idiopathic eosinophilia-myalgia syndrome (chemically/clinically overlapping, distinguished by L-tryptophan exposure history rather than oil exposure); idiopathic systemic sclerosis/scleroderma; hypereosinophilic syndrome; eosinophilic pneumonia of other causes; other causes of pulmonary arterial hypertension.
  • Screening: not applicable — there is no ongoing population at risk; the relevant "screening" activity was 1981-era regulatory/epidemiologic case-finding, not a recurring clinical screening program.

11. Outcome/Prognosis

  • Mortality: acute-phase deaths were predominantly due to respiratory failure from noncardiogenic pulmonary edema; intermediate-phase deaths were dominated by thromboembolism of the great arteries and pulmonary hypertension; chronic-phase deaths were predominantly due to restrictive respiratory failure secondary to severe neurologic infection and pulmonary hypertension (ScienceDirect S0895435603001197).
  • Long-term relative risks (vs. general/unaffected comparison, from cohort follow-up): liver disease RR 3.83; pulmonary hypertension RR 3.19; motor neuropathy RR 2.24; pulmonary infection RR 1.54; eosinophilia RR 1.14.
  • Recovery: approximately half of all affected individuals recovered from the acute phase without apparent long-term sequelae. Among those progressing to intermediate/chronic disease, partial recovery of skin sclerosis has been documented beyond 2 years post-onset in some patients, but many carry persistent morbidity for life.
  • Quality of life: SF-36-based assessment in 2022 documented persistently reduced physical and mental health-related quality of life in survivors relative to the general Spanish population, decades after exposure (IJE 2022).
  • Prognostic factors: HLA genotype (DR2 with fatal disease; DR4-DQ8/A24 with chronicity), phase reached (acute-only vs. progression to intermediate/chronic), sex (female) and younger age (<40) associated with shorter survival among those who die of TOS-related causes.
  • Neurocognitive prognosis: four decades post-exposure, survivors show measurable frontal-subcortical cognitive dysfunction by objective eye-tracking testing, but blood neurodegeneration biomarkers argue against an ongoing progressive neurodegenerative process, suggesting a largely static injury pattern from the original toxic/immune insult rather than a worsening trajectory (PMC12484009; PMC12155236).

12. Treatment

No specific antidote exists. Management is supportive/symptomatic and multidisciplinary. Multiple immunomodulatory and antifibrotic agents were trialed without convincing benefit:

  • Corticosteroids (NCIT:C2942 / treatment_term NCIT:C15986 Pharmacotherapy) — used, particularly in acute/subacute phases, to control inflammation and eosinophilia; benefit was inconsistent and not clearly disease-modifying for chronic sequelae.
  • Azathioprine, penicillamine — immunomodulatory/antifibrotic agents trialed for chronic sclerodermiform disease; no convincing therapeutic effect demonstrated.
  • Plasmapheresis — trialed; no convincing effect.
  • Vitamin E, superoxide dismutase — antioxidant approaches trialed given the oxidative/PMN-activation component of pathogenesis; no convincing effect.
  • Vasodilators — used symptomatically, presumably for pulmonary hypertension and Raynaud phenomenon management (NCIT term candidates: NCIT:C29688-type vasodilator class terms — verify exact CURIE).
  • Supportive care (NCIT:C15747 Supportive Care) — mainstay of management: respiratory support in acute pulmonary edema, nutritional support for cachexia, management of secondary infection.
  • Rehabilitation (NCIT:C15315 Rehabilitation) / Physical therapy (NCIT:C15302) — for joint contractures and neuropathy-related disability.
  • Genetic counseling — not applicable (non-heritable disease).
  • Experimental/advanced therapeutics — no gene therapy, cell therapy, RNA-based therapy, targeted therapy, or immunotherapy has been developed or trialed specifically for TOS; this reflects both its status as a closed historical epidemic and the era (early 1980s) in which acute management decisions were made.
  • Treatment outcomes summary: per the WHO 1991 meeting report and subsequent reviews, none of the pharmacologic interventions trialed (corticosteroids, azathioprine, penicillamine, plasmapheresis, vitamin E, superoxide dismutase, vasodilators, anti-inflammatories) produced a convincing therapeutic effect on the underlying disease course (ScienceDirect S0049017205800174; syndrome.co.uk summary).
  • Clinical trials: no NCT-registered interventional trials specific to TOS were identified (consistent with its status as a closed 1981 epidemic predating ClinicalTrials.gov and lacking an ongoing at-risk population).

13. Prevention

  • Primary prevention: entirely regulatory/public-health — withdrawal of the adulterated oil from sale in 1981, and subsequent Spanish and EU regulatory tightening of edible-oil labeling, distribution licensing, and denaturant-tracking requirements to prevent industrial (non-food) oils from re-entering the food supply chain. There is no vaccine, chemoprophylaxis, or individual behavioral intervention relevant to this disease beyond avoiding the (now nonexistent) contaminated product.
  • Secondary prevention: 1981-era case-finding and epidemiologic surveillance (the Spanish Clinical Commission registry) to identify and treat affected individuals early in the acute phase, when corticosteroid/supportive intervention may reduce acute morbidity.
  • Tertiary prevention: multidisciplinary chronic-disease management (rehabilitation, symptomatic treatment of pulmonary hypertension, neuropathy, and sicca syndrome) to limit disability in those who progressed to chronic disease.
  • Genetic counseling / screening: not applicable (non-heritable).
  • Public health: the TOS epidemic directly motivated strengthened food-safety and adulteration-control legislation in Spain and contributed to broader European food-safety regulatory reform; it remains a canonical case study in toxicology and public-health/regulatory-failure literature (see the 2025 historical/political analysis, "Lo de la colza").
  • Environmental intervention: control of industrial denaturant (aniline) use and tracking, and enforcement against unlicensed oil re-refining/distribution — the specific regulatory intervention that ended the epidemic.

14. Other Species / Natural Disease

  • Taxonomy: TOS as a clinical entity is described only in humans (Homo sapiens, NCBITaxon:9606); it is not a naturally occurring veterinary disease.
  • Breed: not applicable.
  • Natural disease in other species: none reported — TOS is not a spontaneously occurring animal disease; all animal data derive from deliberate experimental exposure (see Section 15).
  • Comparative biology: the closest naturally/epidemically occurring comparator in another population is the U.S. eosinophilia-myalgia syndrome (EMS) of 1989, caused by contaminated L-tryptophan supplements, which shares clinical (eosinophilia, myalgia, fasciitis/fibrosis, in some cases scleroderma-like skin change) and chemical (shared PAA metabolite) features with TOS, suggesting partially convergent or shared pathogenic mechanisms across two independent toxin exposures (Mayo Clin Proc; NEJM 1990).
  • Zoonotic potential: not applicable — TOS is a toxin-induced disease, not a transmissible one.

15. Model Organisms

  • HLA-transgenic mice: mice expressing human HLA-DR2/DQ6, DR3, or DR4 haplotypes were exposed to TOS-implicated oils/oil components. DR2/DQ6-expressing mice showed higher eosinophilia and IgE than DR3/DR4-expressing mice, functionally validating the human HLA-association data and modeling genetic restriction of the immunomodulatory response (PubMed 15979827).
  • Mouse toxicologic models of PAP/fatty-acid anilides: administration of fatty acid anilides and the linoleic diester of PAP to mice produced weight loss, pulmonary hemorrhage/congestion/emphysema, and increased blood eosinophilia, supporting these compounds as candidate etiologic agents and providing an acute toxicologic model of the pulmonary/hematologic phenotype (Arch Toxicol, S002040050641; PubMed 7779449; PubMed 10650923).
  • Rodent (non-HLA) "search for an animal model": a dedicated study explicitly titled "A search for an animal model of the Spanish toxic oil syndrome" concluded that no conventional rodent model fully recapitulates the human multisystem chronic phenotype, particularly the scleroderma-like fibrotic and neuropathic components (ScienceDirect S027869150200114X).
  • Lewis rat EMS model (cross-referenced comparator): female Lewis rats given L-tryptophan implicated in human EMS developed fasciitis and perimyositis resembling human EMS pathology; while not a TOS model per se, this is the most relevant fibrotic-phenotype rodent model in the literature comparator disease and is cited as a partial model for the shared PAA-associated fibrotic mechanism (PubMed 2243145; JCI 114902).
  • In vitro human systems: human polymorphonuclear leukocyte cultures exposed to PAP and its mono-/di-oleyl esters, used to model reactive-oxygen-metabolite generation (ScienceDirect S0378427496038623); human liver microsome/hepatocyte incubations used to model PAP bioactivation to reactive quinoneimine and to PAA (PubMed 8555405).
  • Model limitations: no existing model — mouse or rat — reproduces the full chronic human phenotype (progressive scleroderma-like skin sclerosis, peripheral neuropathy, pulmonary arterial hypertension, and decades-long cognitive dysfunction). HLA-transgenic mice model the genetic-restriction/Th2-skewing immune mechanism; toxicologic mouse models capture acute pulmonary/hematologic effects of specific candidate compounds; neither models chronic fibrotic/vascular remodeling or the neurologic phenotype.
  • Applications: the HLA-transgenic mouse model is useful for studying gene–environment interaction and immune mechanism; the acute mouse toxicology models are useful for compound-identification/structure-activity work distinguishing candidate etiologic agents; no model exists for late-stage drug development, since TOS is a closed epidemic with no active patient-recruitment or therapeutic pipeline.

Summary of Key Ontology/Identifier Candidates for Curation

Category Suggested term (verify CURIE before binding)
Disease MONDO:0016421 (Toxic oil syndrome); ORPHA:227972
Genes/host modifiers HLA-DRB1 (DR2, DR4 alleles), HLA-DQB1 (DQ8), HLA-A (A24) — HGNC IDs for HLA loci should be resolved via HGNC before binding genetic: blocks
Chemical entities 3-(N-phenylamino)-1,2-propanediol (PAP); 3-(phenylamino)alanine (PAA); aniline; acetanilide; oleoanilide — resolve CHEBI CURIEs
Phenotypes See Section 3 table (HP terms)
Cell types CL:0000771 eosinophil; CL:0000115 endothelial cell; CL:0002548 fibroblast
Biological processes GO:0006954 inflammatory response; GO:0030198 extracellular matrix organization; angiogenesis/vascular remodeling GO terms
Anatomy UBERON:0002048 lung; UBERON:0000178 skin epidermis/integument; UBERON:0001021 nerve; UBERON:0002107 liver; UBERON:0001981 blood vessel
Treatment NCIT:C15986 Pharmacotherapy (corticosteroids); NCIT:C15747 Supportive Care; NCIT:C15315 Rehabilitation; NCIT:C15302 Physical Therapy

Note on evidence base: Nearly all primary literature on TOS dates from 1981–2005, reflecting its nature as a closed historical epidemic; a smaller but active cluster of 2022–2025 publications addresses long-term (four-decade) survivor outcomes (quality of life, cognition, neurodegeneration biomarkers). For dismech curation, prioritize: the 1983 NEJM clinical-epidemiology paper, the 2003 survival cohort paper, the PAP/PAA mechanistic series (1994–2001), the HLA association papers (1996–2005), and the 2025 long-term biomarker/cognitive papers — each should be fetched and quote-verified per standard evidence-curation practice before use.

Sources: - Toxic-allergic syndrome caused by ingestion of rapeseed oil denatured with aniline (PMID:6116011) - Clinical epidemiology of toxic-oil syndrome (PMID:6633617) - Factors associated with pathogenicity of oils related to TOS (PMID:7918809) - Epidemiologic evidence for a new class of compounds associated with TOS (PMID:10069247) - Pulmonary vascular lesions in the toxic oil syndrome in Spain (PMC459646) - The toxic oil syndrome (PMID:8001309) - Toxic Oil Syndrome: Review of Immune Aspects of the Disease - Toxic-Oil Syndrome — CHEST Journal - Biotransformation of PAP to PAA (PMID:8555405) - Absorption and effects of PAP esters — Lipids - Acute pathology of fatty acid anilides and PAP diester in mice (PMID:10650923) - Comparison of acute pathology induced by PAP and mono-oleoyl ester (PMID:7779449) - Effects of PAP esters on reactive oxygen metabolites in human PMNs - Orphanet: Toxic oil syndrome (ORPHA:227972) - Toxic oil syndrome — MalaCards / MONDO - Toxic oil syndrome: Survival in the whole cohort between 1981 and 1995 - Toxic oil syndrome: health-related quality-of-life assessment using SF-36 (IJE 2022) - Late cases of toxic oil syndrome: agent persisted in stored oil - Frontal-subcortical dysfunction in toxic oil syndrome (Frontiers 2025) - Toxic oil syndrome. A long-term follow-up of a cohort of 332 patients (PMID:8412642) - DR2 antigens associated with severity of disease in TOS (PMID:10746782) - Frequencies of HLA-A24 and HLA-DR4-DQ8 in chronic TOS (PMID:8803534) - Genetic approaches in the understanding of Toxic Oil Syndrome - A search for an animal model of the Spanish toxic oil syndrome - TOS: genetic restriction and immunomodulatory effects in HLA-transgenic mice (PMID:15979827) - The toxic oil syndrome: an exogenously induced autoimmune reaction (PMID:9112238) - Toxic oil syndrome: features overlapping various forms of scleroderma (PMID:3961509) - Clinical, pathologic, and immunopathologic manifestations of TOS: 14 cases - Cytokine mRNA expression in lung tissue from TOS patients: Th2 mechanism (PMID:9074654) - Fibrogenic growth factors in EMS and TOS (PMID:8285738) - L-tryptophan implicated in EMS causes fasciitis/perimyositis in Lewis rat (PMID:2243145) - 3-(Phenylamino)alanine — link between EMS and TOS (Mayo Clin Proc) - An investigation of the cause of EMS associated with tryptophan use (NEJM 1990) - Clinical Epidemiology of Toxic-Oil Syndrome (NEJM 1983) - "Lo de la colza": mass poisoning, state neglect, and corruption — Nursing Clio (2025) - Toxic Oil Syndrome: Current Knowledge and Future Perspectives (PMC1059810) - Blood Biomarkers of Neurodegeneration over Four Decades After TOS (PMC12155236) - Cognitive Functioning in TOS Survivors: Case-Control Study (PMC12155933)

Reference Validation

Checked with linkml-reference-validator 0.2.1.

Outcome Count
References checked 29
Resolved 28
Unresolved (possible confabulation) 0
Unverifiable 1
References weighed for topical relevance 28
On topic 15
Off topic 0

28 of 29 references resolved; the rest could not be looked up either way.

Term Validation

Checked with linkml-term-validator 0.4.5, through the ols: adapter.

Outcome Count
Terms checked 61
Resolved 58
Unresolved (possible confabulation) 0
Obsolete 2
Unverifiable 1
Terms whose name was checked 2
Terms named correctly 1
Terms named as a different term 1

Terms the report names something else

These identifiers resolve, so nothing about them looks wrong, and the ontology calls them something unrelated to what the report calls them. That usually means the identifier is not the one the sentence needs:

  • MONDO:0016421 (3 mentions) - the report calls it "if available"; MONDO calls it toxic oil syndrome

Obsolete terms

These terms are real but deprecated. Citing one is not a fabrication; it does mean the report is naming something the ontology has retired:

  • HP:0100672 (obsolete Vaginal hernia) (1 mention) - replaced by HP:0031607
  • NCIT:C29688 (GR6 Protein) (1 mention)

Prefixes with no resolver

Terms carrying these prefixes were not checked either way, because no configured ontology covers them. An unrecognised prefix may name an ontology this run could not reach as easily as one that does not exist, so nothing here is evidence of fabrication: ORPHA.