Allopurinol-Induced Stevens-Johnson Syndrome/Toxic Epidermal Necrolysis

1. Disease Information

1.1 Overview

Allopurinol-induced SJS/TEN is a delayed (Type IVc), HLA class I–restricted, CD8+ T-cell–mediated severe cutaneous adverse reaction in which allopurinol (or, more precisely, its active metabolite oxypurinol) triggers widespread keratinocyte death, epidermal detachment, and mucosal erosion. It sits on a severity spectrum defined by body-surface-area (BSA) detachment:

This BSA-based consensus classification is the Bastuji-Garin scheme (PMID:8420497).

Allopurinol also causes two sibling SCAR phenotypes that share the same HLA restriction and the same culprit metabolite but differ in clinical expression: DRESS/DIHS (drug reaction with eosinophilia and systemic symptoms) and the older umbrella term allopurinol hypersensitivity syndrome (AHS). Most genetic-association literature reports the pooled endpoint "allopurinol-SCAR" (SJS/TEN + DRESS), which is a real limitation when curating SJS/TEN specifically — flag it in every evidence item drawn from those studies.

Why allopurinol matters disproportionately: it is the single most common culprit drug for SJS/TEN in Europe and Israel.

"Allopurinol was the drug most frequently associated with SJS or TEN, with 66 exposed patients (17.4%) and 28 exposed control subjects (1.9%) (adjusted odds ratio = 18, 95% confidence interval: 11-32)." — Halevy et al., EuroSCAR Study Group, J Am Acad Dermatol 2008 (PMID:17919772)

1.2 Key identifiers

⚠️ Curation decision flagged: there is no MONDO term for "allopurinol-induced SJS/TEN." No MONDO:* concept encodes the drug-specific entity. Options: (a) curate as a drug-toxicity Disease entry anchored to MONDO:0018229 and/or MONDO:0019810 via mappings.mondo_mappings with mapping_predicate: skos:narrowMatch; (b) curate the general SJS/TEN entity and carry allopurinol as an environmental[] / exposure-linked mechanism. Option (a) is more consistent with how dismech treats drug-toxicity entries (the "side effect as mechanism" module family), but the call belongs in the design register.

1.3 Synonyms and alternative names

• Allopurinol hypersensitivity syndrome (AHS) — broader; includes DRESS and non-SCAR rashes
• Allopurinol-induced SCAR / allopurinol-SCAR — broader; pooled SJS/TEN + DRESS endpoint used in most genetics papers
• Allopurinol-induced epidermal necrolysis (EN) — the RegiSCAR-preferred umbrella for SJS + overlap + TEN
• Allopurinol-induced Lyell syndrome (chiefly for the TEN pole)
• Drug-induced Stevens-Johnson syndrome (MeSH entry term)

1.4 Data provenance — patient-level vs aggregated

Both, and the distinction matters for evidence grading:

• Individual-patient (EHR/registry-derived): the RegiSCAR and EuroSCAR case–control networks; the Taiwan Allopurinol-SCAR Consortium prospective cohort (PMID:26399967); Vanderbilt BioVU–linked SJS/TEN Survivor Study (PMID:41160012); Ontario ICES administrative-claims cohort (PMID:35644439); Taiwan NHIRD (PMID:39969876); Danish national registries (PMID:42467870).
• Aggregated disease-level: CPIC/PharmGKB guideline annotations, Orphanet, MeSH/ICD, and pooled meta-analyses of HLA association.
• Mechanistic: ex vivo patient PBMC and blister-cell work — individual-patient biospecimens, small n (typically 5–25 cases per study).

2. Etiology

2.1 Primary causal factors

The disease is obligately two-hit: it requires (i) exposure to allopurinol and accumulation of oxypurinol, and (ii) a permissive HLA class I background — plus, almost always, (iii) a naive T-cell repertoire containing clonotypes capable of reading the drug–HLA complex.

Necessary exposure: allopurinol (CHEBI:40279), a xanthine oxidase inhibitor used for gout, hyperuricemia, tumor lysis prophylaxis, and (often inappropriately) asymptomatic hyperuricemia. Allopurinol is rapidly converted in vivo to oxypurinol (CHEBI:748196; also indexed as alloxanthine, CHEBI:28315), which has a long half-life and is renally cleared — the pivot on which most of the risk architecture turns.

Dose- and time-dependence (unusual for a Type B idiosyncratic reaction, and one of the most curation-relevant facts here):

"Daily doses equal to or greater than 200 mg were associated with a higher risk (adjusted odds ratio = 36, 95% confidence interval: 17-76) than lower doses (adjusted odds ratio = 3.0, 95% confidence interval: 1.1-8.4). The risk was restricted to short-term use (<or=8 weeks)." — Halevy et al. 2008 (PMID:17919772)

"Despite the prevailing dogma that Type B adverse drug reactions are dose independent, allopurinol hypersensitivity is primarily driven by oxypurinol-specific T cell response in a dose-dependent manner, particular in the presence of HLA-B58:01 allele." — Yun et al., Clin Exp Allergy 2013 (PMID:24152157*)

2.2 Genetic risk factors

HLA-B*58:01 — the dominant locus

Discovered in Han Chinese in a candidate-SNP screen that landed squarely in the MHC:

"The HLA-B5801 allele was present in all (100%) 51 patients with allopurinol-SCAR, but only in 20 (15%) of 135 tolerant patients [odds ratio 580.3 (95% confidence interval, 34.4-9780.9); corrected P value = 4.7 x 10(-24)]" — Hung et al., PNAS 2005 (PMID:15743917*)

The same paper notes the risk allele sits on an extended haplotype: "HLA alleles A3303, Cw0302, and DRB10301 were in linkage disequilibrium and formed an extended haplotype with HLA-B5801" — relevant because tag-SNP proxies (e.g. rs3131003, rs9263726, PSORS1C1) are used in some screening programs and are LD-based, not causal (PMID:41846573).

Effect size is strongly ancestry-dependent and weaker outside East Asia. A second, independent, Taiwanese cohort estimate:

"HLA-B58:01 was strongly associated with allopurinol-SCAR (p<0.001, OR (95% CI) 109 (25 to 481))" — Chung et al., Ann Rheum Dis 2015 (PMID:25115449*)

And a modern range across phenotype severity:

"Individuals who carry the human leukocyte antigen (HLA)-B58:01 allotype are at higher risk of experiencing a hypersensitivity reaction (odds ratios ranging from 5.62 to 580.3 for mild to severe reactions, respectively)." — Mifsud et al., Allergy 2023 (PMID:37452515*)

HLA-A*34:02 — a newly identified second independent locus (2025)

This is the most important recent genetic development, and it directly undercuts single-allele screening in admixed populations:

"Two HLA class I alleles were found to be independently associated with increased risk of allopurinol-induced SCAR: HLA-B58:01 (OR, 28.0 [95% CI, 8.6-100.6]) and HLA-A34:02 (OR, 20.6 [95% CI, 3.3-131.1]). No HLA class II alleles meeting the Bonferroni-corrected P < .05 level of significance were identified." — Campbell et al., JAMA Dermatol 2025 (PMID:41160012; preprint PMID:40475155)

"the allele was absent in more than one-third of the patient cohort and is therefore an incomplete indicator of risk" — same source

Caveats to record with this finding: n = 16 cases, primarily self-identified Black race, single-center, HLA imputed from genotyping array in controls. Treat HLA-A*34:02 as SUSCEPTIBILITY, not CAUSATIVE, and mark the evidence directness: DIRECT but note the small cohort.

Other reported HLA associations (weaker / not replicated)

• HLA-DR9, HLA-DR14 in allopurinol hypersensitivity arising in hematologic malignancy (Korean cohort, PMID:24858023) — class II, unreplicated.
• In a Turkish DRESS cohort, HLA-B*58:01 co-occurred with other risk alleles in multiple-drug-hypersensitivity syndrome (PMID:42014364).

Penetrance is low — this is the single most misused number in the field

Carrying HLA-B58:01 is necessary-ish but nowhere near sufficient*. Population frequencies (see §9) put carriers at 10–20% of Han Chinese, while allopurinol-SCAR incidence is ~0.3%/year of treated patients. The allele is a permissive background, not a determinant.

2.3 Environmental / clinical (non-genetic) risk factors

The Ontario CKD result is the cleanest actionable statement in the whole risk literature:

"Older patients with CKD who started allopurinol at >100 mg/d versus ≤100 mg/d were twice as likely to visit a hospital with a severe cutaneous reaction in the next 180 days." — Bathini et al., Am J Kidney Dis 2022 (PMID:35644439)

2.4 Protective factors

• Non-carriage of HLA-B*58:01 — the dominant protective state, but explicitly incomplete (PMID:41160012).
• Low starting dose with slow titration (≤100 mg/day, lower in CKD), strongly recommended by ACR 2020 (PMID:32390306). Mechanistically coherent: the T-cell response is oxypurinol-concentration-dependent (PMID:24152157).
• Substituting a structurally unrelated xanthine oxidase inhibitor. Febuxostat does not cross-react with oxypurinol-specific T cells:

  "Oxypurinol induced T-cell response in a concentration- and time-dependent manner, whereas allopurinol or febuxostat did not. T cells from patients with allopurinol-SCAR showed no crossreactivity with febuxostat." — Chung et al., J Invest Dermatol 2015 (PMID:25946710)

• Prospective HLA-B*58:01 screening — demonstrated preventive effect (see §13).
• No genetic protective variant has been characterized. Not applicable / no data.

2.5 Gene–environment interaction

This disease is a gene–environment interaction; the two axes are not separable. Three interacting dimensions are documented:

1. HLA genotype × drug concentration. "TCL induction data show that both the presence of HLA-B58:01 allele and high concentration of drug are important for the generation of drug-specific T cells… functional avidity of ALP/OXP-TCL is dependent on both the induction dose and HLA-B58:01 status." (PMID:24152157)
2. HLA genotype × renal function. Impaired eGFR → oxypurinol accumulation → higher effective drug concentration at the HLA groove. "Poor renal function was significantly associated with the delayed clearance of plasma oxypurinol, and increased the risk of allopurinol-SCAR" (PMID:25115449). This is a pharmacokinetic GxE, not a pharmacodynamic one.
3. HLA genotype × TCR repertoire. Even among carriers, the responding clonotypes are private — see §6.

3. Phenotypes

3.1 Prodromal / constitutional

3.2 Cutaneous — the defining phenotypes

⚠️ Ontology gap. OLS4 returned no HP term for "Nikolsky sign." Per the dismech term contract, leave term: absent rather than binding a near-miss; no term beats a bad one.

3.3 Mucosal

At least two mucosal sites are required by the consensus definition (PMID:8420497).

3.4 Systemic / laboratory

3.5 Onset timing, severity, and course

• Latency: typically 2–8 weeks after allopurinol initiation; EuroSCAR restricts the risk window to ≤8 weeks (PMID:17919772). This is longer than the ~1-week latency typical of antiseizure-drug SJS/TEN and is a useful discriminator for ALDEN causality scoring.
• Onset pattern: acute, but with a subacute prodrome.
• Severity: severe by definition; graded by SCORTEN (§10) and BSA.
• Progression: rapidly progressive over days, then plateau, then re-epithelialization over ~2–3 weeks. Self-limited if the drug is withdrawn — but re-exposure can be worse and is potentially fatal.
• Duration: the acute episode is self-limited; the sequelae are lifelong (§11).

3.6 Quality-of-life impact

Acute phase: intensive/burn-unit-level care, severe pain, inability to eat or open eyes. Survivor phase carries substantial and durable burden — chronic ocular surface disease (dry eye, symblepharon, corneal opacification, blindness), cutaneous dyspigmentation and scarring, nail dystrophy, genital adhesions, chronic pruritus, and post-traumatic stress. Disease-specific QoL instrument data (EQ-5D / SF-36 / DLQI) for allopurinol-specific SJS/TEN was not identified in this search; general SJS/TEN survivor QoL literature exists but was not retrieved in a form suitable for quoting. Gap — flag as KNOWLEDGE_GAP.

4. Genetic / Molecular Information

4.1 There are no causal genes in the Mendelian sense

This is a pharmacogenomic susceptibility trait, not a monogenic disease. No pathogenic variants, no ACMG/AMP classification, no gnomAD pathogenic allele frequency, no somatic/germline distinction, no chromosomal abnormality. All of these template sub-items are not applicable.

What exists instead is an HLA allelic risk architecture: common, functionally normal MHC alleles that happen to have a peptide-binding groove chemistry permissive to oxypurinol.

4.2 Susceptibility loci

Curation note on relationship types: use relationship_type: SUSCEPTIBILITY for all of these. None is CAUSATIVE. Inheritance: this is not an inherited disease — but the susceptibility is inherited codominantly as an HLA allele. If an inheritance: block is added at all, HP:0010982 (Polygenic inheritance) is a poor fit; a free-text description of HLA-allelic susceptibility with no term: binding is the honest choice.

4.3 Genes central to the effector mechanism (not risk genes)

These belong in pathophysiology nodes, not in genetic::

That last row deserves emphasis for the causal graph: the toxicity has nothing to do with the drug's intended pharmacology. Xanthine oxidase inhibition and HLA-B*58:01 binding are two entirely separate molecular events that happen to involve the same molecule. This is why febuxostat — same target, different scaffold — is safe in these patients.

4.4 Modifier genes, epigenetics, chromosomal abnormalities

• Modifier genes: none characterized. The functional modifier is the TCR repertoire, which is somatically generated rather than germline-encoded (§6).
• Epigenetics: no DNA-methylation or histone-modification data specific to allopurinol-SJS/TEN was identified. Not available.
• Chromosomal abnormalities: not applicable.

5. Environmental Information

• The obligate environmental factor is a drug exposure. Model as an environmental[] entry with influences_mechanisms and environmental_effect: TRIGGERS. ECTO has exposure-to-chemical patterns; a specific "exposure to allopurinol" term was not confirmed in this search — check cache/ecto/terms.csv and run just environmental-term-audit before assuming a gap. If unbound, record the search in review_notes: per the check-environmental-evidence waiver convention.
• Modifying environmental/clinical exposures: renal impairment (from any cause), high initiation dose, concurrent illness. Diuretic co-medication was tested and not supported (PMID:17919772).
• Lifestyle factors: no independent effect established. Gout's own lifestyle determinants (alcohol, purine-rich diet, obesity) act only by increasing the probability of allopurinol prescription — an indirect, non-mechanistic path. Do not draw a causal edge.
• Infectious agents: not applicable to the allopurinol-induced form. (Contrast: Mycoplasma pneumoniae causes a distinct mucocutaneous entity, MIRM, which is a differential — see §10.) Herpesvirus reactivation (HHV-6, EBV, CMV) is a feature of DRESS, not of SJS/TEN, and should not be imported into this entry.

6. Mechanism / Pathophysiology

6.1 The causal chain, step by step

Read this as a relay race where each runner hands off to the next. Steps 1–5 are the sensitization/recognition leg; steps 6–10 are the execution leg; step 11 is what you see at the bedside.

1. Allopurinol is ingested and rapidly metabolized to oxypurinol by xanthine oxidase. Oxypurinol has a long half-life and is renally cleared. → leads to a systemic oxypurinol burden proportional to dose and inversely proportional to eGFR. (Demonstrated: PMID:24152157, PMID:25115449.)

2. Renal impairment or high dosing raises the steady-state oxypurinol concentration. → results in sufficient local drug concentration at the antigen-presenting cell surface to occupy HLA grooves. (Demonstrated in humans: "Poor renal function was significantly associated with the delayed clearance of plasma oxypurinol, and increased the risk of allopurinol-SCAR (p<0.001, OR (95% CI) 8.0 (3.9 to 17))" — PMID:25115449.)

3. Oxypurinol binds non-covalently and reversibly into the peptide-binding groove of HLA-B*58:01. This is the p-i (pharmacological interaction with immune receptors) mechanism — no hapten, no covalent adduct, no antigen processing required. → leads to an altered peptide/HLA surface.

"ALP/OXP-specific T cells reacted immediately to the addition of the drugs and bypassed intracellular Ag processing, which is consistent with the 'pharmacological interaction with immune receptors' (p-i) concept." — Yun et al., J Immunol 2014 (PMID:24591375)

"this response supported the pharmacological interaction with immune receptors (p-i) concept by showcasing (i) the labile metabolite interaction with peptide/HLA complexes, (ii) immunogenic complex formation at the cell surface, and (iii) lack of requirement for antigen processing to elicit drug-induced T cell responsiveness." — Mifsud et al., Allergy 2023 (PMID:37452515)

1. BRANCH — a partly complementary "altered peptide repertoire" mechanism. Structural work shows the drug can also enable self-peptides that would otherwise not bind at all:

"a Lamin A/C peptide KAGQVVTI which is unable to bind HLA-B58:01 on its own, is enabled to form a stable peptide-HLA complex only in the presence of allopurinol. Crystal structure analysis reveal that allopurinol non-covalently facilitated KAGQVVTI to adopt an unusual binding conformation, whereby the C-terminal isoleucine does not engage as a PΩ that typically fit deeply in the binding F-pocket." — Huan et al., Sci Rep 2023 (PMID:37296297*)

The authors extend this to viral peptide (EBNA3B) presentation and propose the result is anti-self reactivity: "aberrant loading of unconventional peptides in the presence of allopurinol or oxypurinol may be able to trigger anti-self reactions that can lead to Stevens-Johnson syndrome/toxic epidermal necrolysis." (Structural/in vitro — mark evidence_source: IN_VITRO, and note the two branches are not mutually exclusive; the field has not settled which dominates in vivo. Good candidate for a mechanistic_hypotheses group with status: EMERGING.)

1. Drug-modified pHLA complexes are read by CD8+ T cells bearing specific, private, oligoclonal αβTCRs. → results in immediate T-cell activation without a processing delay.

"Preferential TCR-V-β usage and clonal expansion of specific CDR3 (third complementarity-determining region) were found in the blister cells from skin lesions (n=8) and oxypurinol-activated T-cell cultures (n=4) from patients with allopurinol-SCAR." — Chung et al., J Invest Dermatol 2015 (PMID:25946710)

"Examination of paired OXP-induced αβTCR repertoires highlighted an oligoclonal and private clonotypic profile in both resolved ALP-induced SJS/TEN cases and drug-naïve healthy donors." — Mifsud et al. 2023 (PMID:37452515)

Note the striking implication of that last clause: drug-naive healthy donors already carry the responsive clonotypes. The repertoire is not the bottleneck; the drug concentration and the HLA are.

1. Activated drug-specific CD8+ CTLs (and NK/NKT cells) traffic to skin and mucosa and accumulate in the subepidermal blister space. → leads to a local cytotoxic milieu.

2. CTL/NK degranulation releases granulysin as the dominant soluble killer. → results in keratinocyte death without requiring cell–cell contact, which is what makes the damage disseminated rather than focal.

"Granulysin concentrations in the blister fluids were two to four orders of magnitude higher than perforin, granzyme B or soluble Fas ligand concentrations, and depleting granulysin reduced the cytotoxicity. Granulysin in the blister fluids was a 15-kDa secretory form, and injection of it into mouse skin resulted in features mimicking SJS-TEN." — Chung et al., Nat Med 2008 (PMID:19029983)

1. BRANCH — a parallel monocyte-driven necroptosis arm. Drug-exposed monocytes secrete annexin A1, which engages FPR1 on keratinocytes:

"Mass spectrometric analysis identified annexin A1 as a key mediator of keratinocyte death; depletion of annexin A1 by a specific antibody diminished supernatant cytotoxicity. The necroptosis-mediating complex of RIP1 and RIP3 was indispensable for SJS/TEN supernatant-induced keratinocyte death, and SJS/TEN keratinocytes expressed abundant formyl peptide receptor 1 (FPR1), the receptor for annexin A1, whereas control keratinocytes did not." — Saito et al., Sci Transl Med 2014 (PMID:25031270)

→ leads to RIPK1/RIPK3 necrosome assembly, MLKL phosphorylation, membrane rupture. Note: this work used SJS/TEN broadly, not allopurinol specifically — grade accordingly and say so in the explanation.

1. Auxiliary arms: Fas/FasL-mediated apoptosis, perforin/granzyme B, and TNF-α. TNF-α is the arm with the strongest therapeutic validation (§12). Macrophage-derived CXCL10 recruiting CXCR3+ CTLs has been demonstrated in immune-checkpoint-inhibitor-induced SJS/TEN (PMID:39737932) — mechanistically attractive, but not yet shown for allopurinol. Curate as KNOWLEDGE_GAP or mechanistic_hypotheses, not as established.

2. Massive keratinocyte apoptosis + necroptosis → full-thickness epidermal necrosis → results in loss of dermal–epidermal adhesion, sub-epidermal cleavage, Nikolsky-positive sloughing.

3. Loss of the epidermal barrier → the clinical syndrome: fluid and electrolyte loss, thermoregulatory failure, protein catabolism, and a wide-open portal for sepsis — which is what actually kills most patients.

4. Feedback loop worth modeling explicitly: sustained oxypurinol levels after drug withdrawal keep feeding step 3. "Sustained high levels of oxypurinol after allopurinol withdrawal correlated with the poor prognosis of allopurinol-SCAR." (PMID:25115449). This is the mechanistic rationale for the (unproven) proposal to actively enhance oxypurinol clearance — the same paper: "An early intervention to increase the clearance of plasma oxypurinol may improve the prognosis of allopurinol-SCAR."

6.2 Suggested GO / CL / UBERON terms for pathophysiology nodes

6.3 Molecular profiling

• Transcriptomics: the granulysin discovery came from gene-expression profiling of blister cells (PMID:19029983). scRNA-seq of SJS/TEN lesional skin exists for ICI-induced (PMID:39737932) and PD-1-inhibitor-induced (PMID:41418419) disease. No published single-cell atlas of allopurinol-specific SJS/TEN lesions was identified. Genuine, curatable gap.
• Proteomics: mass spectrometry of PBMC supernatant identified annexin A1 (PMID:25031270).
• Immune repertoire sequencing: next-generation TCR sequencing of blister cells and drug-stimulated cultures (PMID:25946710); paired single-cell αβTCR sequencing (PMID:37452515).
• Structural biology: X-ray crystallography of allopurinol/oxypurinol–HLA-B58:01–peptide ternary complexes (PMID:37296297). Also in silico docking showing oxypurinol binds the HLA-B58:01 groove with higher affinity than allopurinol (PMID:24591375, evidence_source: COMPUTATIONAL).
• Metabolomics / lipidomics: none identified. Not available.
• Functional genomics screens (CRISPR/RNAi): none identified for this disease. Not available.

7. Anatomical Structures Affected

Organ level

• Primary: skin (UBERON:0002097 skin of body) and stratified squamous mucosae — oral (UBERON:0000165 mouth), ocular surface / conjunctiva (UBERON:0001811 conjunctiva), genital, anal, and in severe cases tracheobronchial and esophageal epithelium.
• Secondary: kidney (AKI, both risk factor and complication), liver (transaminitis), lung (bronchiolitis obliterans, ARDS), bone marrow (cytopenias), cardiovascular system (long-term — see §11).
• Body systems: integumentary (primary), immune (driver), ocular/visual, gastrointestinal, respiratory, renal, cardiovascular.

Tissue and cell level

• Tissue: stratified squamous epithelium — full-thickness epidermal necrosis with a sparse dermal inflammatory infiltrate. The mismatch between the tiny number of infiltrating cells and the enormous amount of dead epidermis is precisely what the soluble-mediator model (granulysin, annexin A1) was invented to explain.
• Cells targeted: keratinocytes (CL:0000312; CL:4052061 epidermal keratinocyte; CL:0002187 basal cell of epidermis).
• Cells driving: CD8+ cytotoxic T cells (CL:0000794), NK cells, NKT cells, monocytes/macrophages.

Subcellular level

• Plasma membrane (GO:0005886) — where the drug–pHLA complex forms, where FPR1 sits, and where phospho-MLKL executes.
• Endoplasmic reticulum (GO:0005783) — conventional peptide loading; notably bypassed in the p-i model.
• Cytosol (GO:0005829) — necrosome assembly.
• Mitochondrion (GO:0005739) — intrinsic apoptotic arm.

Localization and laterality

• Bilateral and symmetric, characteristically starting on the face and upper trunk and spreading centrifugally. Palms and soles are frequently involved; the scalp is typically spared. Lesions are not dermatomal or unilateral — asymmetry should prompt reconsidering the diagnosis.

8. Temporal Development

• Age of onset: adult to geriatric, tracking the age distribution of gout and CKD. Mean 61.1 y (SD 12.6) in the US SCAR cohort (PMID:41160012); median 76 y in the Ontario CKD allopurinol-initiator cohort (PMID:35644439). Pediatric cases are exceptional.
• Onset pattern: subacute prodrome (1–3 d) → acute explosive cutaneous phase.
• Latency from drug start: typically 2–8 weeks; EuroSCAR found risk "restricted to short-term use (≤8 weeks)" (PMID:17919772). Longer than for antiseizure drugs.
• Stages:
• Prodromal — fever, malaise, skin pain, ocular grittiness.
• Acute/progressive — 2–5 days of expanding erythema, blistering, detachment. Maximal BSA usually by day 5–7.
• Plateau/nadir — sepsis and multi-organ risk peak.
• Re-epithelialization — median 14–19 days depending on treatment (PMID:29400697).
• Chronic sequelae — lifelong (§11).
• Progression rate: rapid (days).
• Course: monophasic and self-limited if the drug is stopped; not relapsing-remitting. Recurrence requires re-exposure — and is typically more severe.
• Remission: treatment-independent resolution follows drug withdrawal; no spontaneous remission while the drug continues. Because oxypurinol clearance is slow in CKD, "withdrawal" is not instantaneous at the tissue level (PMID:25115449).
• Critical intervention window: the earliest possible drug withdrawal. Every hour of continued exposure is more antigen. Beyond that, the therapeutic window for immunomodulation is the first few days of the acute phase.

9. Inheritance and Population

9.1 Epidemiology

Denominators worth keeping straight — SJS/TEN overall vs. allopurinol-attributable vs. per-initiator risk:

Note the two-orders-of-magnitude gap between the general-population SJS/TEN rate and the rate among allopurinol initiators. That gap is the drug effect.

9.2 Inheritance

• Inheritance pattern: not a heritable disease. The susceptibility allele HLA-B58:01 is inherited codominantly at the HLA-B locus. Most reported cases are heterozygous — in Thai patients, "96.7% of patients with allopurinol-induced cutaneous adverse drug reactions were found to be heterozygous for HLA-B58:01." Do not bind an HPO mode-of-inheritance term to this entry.
• Penetrance: very low and jointly conditioned on drug exposure, dose, and renal function. Carriers who never take allopurinol never manifest.
• Expressivity: highly variable — the same allele and the same drug yield mild rash, DRESS, SJS, or TEN in different people. What determines which is not known. Genuine KNOWLEDGE_GAP.
• Anticipation, germline mosaicism, consanguinity: not applicable.
• Carrier frequency: see allele frequencies below — but "carrier" is a misnomer here; these are common functional HLA alleles, not recessive disease alleles.

9.3 Population demographics and geography

HLA-B*58:01 allele/carrier frequency by ancestry (figures are a synthesis across sources; the authoritative consolidated dataset is Zhou et al. 2021, which pooled HLA genotypes from 3.5–6.4 million individuals across up to 74 countries, PMID:32535895):

"We find major ethnogeographic differences in risk allele prevalence, which translated into pronounced differences in the number of patients needed to test to prevent one case of severe hypersensitivity reactions between countries and populations… Testing of HLA-B58:01 is more likely to be cost-effective throughout Africa and Asia compared with Europe and the Americas." — Zhou et al., Clin Pharmacol Ther 2021 (PMID:32535895*)

Ancestry-dependence of the association strength is separate from allele frequency and is the more important curation point. HLA-B58:01 accounts for essentially 100% of cases in Han Chinese but only ~55–64% in Northern and Southern European cohorts, and was absent in more than one third* of a US cohort of predominantly self-identified Black patients (PMID:41160012). The Portuguese cohort study (PMID:23600531) confirmed a real but weaker association in a European population.

Sex ratio: roughly balanced. US SCAR cohort: 9 female (56.25%) / 7 male (43.75%) (PMID:41160012). Taiwan SJS/TEN survivor cohort: ~50.7% female (PMID:39969876). No convincing sex effect.

Age distribution: concentrated in the 6th–8th decades, following gout/CKD prescribing.

10. Diagnostics

10.1 Clinical diagnosis is primary

There is no confirmatory acute-phase test. Diagnosis rests on: recent drug exposure with compatible latency + prodrome + painful dusky erythema evolving to blistering and detachment + ≥2 mucosal sites + positive Nikolsky sign + compatible histology.

10.2 Histopathology / biopsy

• Frozen section or punch biopsy of lesional skin: full-thickness epidermal necrosis, subepidermal detachment, sparse lymphocytic infiltrate, apoptotic/necrotic keratinocytes at all epidermal levels. This is the key rule-out for staphylococcal scalded skin syndrome (which cleaves at the granular layer, not the DEJ).
• Direct immunofluorescence: negative — used to exclude autoimmune blistering disease (pemphigus, bullous pemphigoid, linear IgA).
• Full-thickness epidermal necrosis on biopsy was the entry criterion in the PD-1-inhibitor SJS/TEN series (PMID:41418419).

10.3 Laboratory tests and biomarkers

"the increased plasma levels of oxypurinol and granulysin linked to the high mortality of allopurinol-SJS/TEN (p<0.01), and strongly associated with prolonged cutaneous reactions in allopurinol-DRESS (p<0.05)" — Chung et al. 2015 (PMID:25115449)

These two are the closest thing this disease has to a disease-specific prognostic biomarker pair, and they are attractive biochemical: entries with BiomarkerReadout links to the oxypurinol-accumulation and granulysin-cytotoxicity pathophysiology nodes.

10.4 Genetic testing

• Single-allele targeted testing is the standard: HLA-B*58:01 genotyping, pre-prescription. Widely available (real-time PCR, LAMP — PMID:23066948; single-tube duplex real-time PCR — PMID:26652271; SSO/SSP typing).
• HLA imputation from genotyping arrays is used in biobank research (PMID:41160012) but is not a clinical test.
• WGS/WES, gene panels, CMA, karyotype, FISH, mtDNA, repeat-expansion testing: not applicable.
• After the fact, HLA typing has diagnostic value too: in the Turkish DRESS cohort, HLA genotyping was used alongside patch testing and the lymphocyte transformation test to attribute causality among multiple candidate drugs (PMID:42014364).
• Emerging caveat: because HLA-B58:01 is absent in > ⅓ of US allopurinol-SCAR cases and HLA-A34:02 is independently associated, the authors argue for broader panels — "These findings underscore the need to conduct population-based studies that both reproduce known and uncover novel HLA associations to reduce harm through contributions to screening, risk stratification, and diagnosis" (PMID:41160012).

10.5 Ex vivo / delayed drug-allergy testing

• Lymphocyte transformation test (LTT) with oxypurinol — critically, LTT with the parent drug is far less sensitive. "Allopurinol allergic patients are primarily sensitized to oxypurinol in a dose-dependent manner… OXP-TCLs do not recognize allopurinol and vice versa." (PMID:24152157). Practical implication: an LTT run with allopurinol alone can produce a false negative. This is an excellent, concrete, curatable diagnostic pitfall.
• Patch testing — used in the DRESS setting; low sensitivity, and contraindicated as provocation in SJS/TEN survivors.
• Drug provocation/rechallenge is absolutely contraindicated in SJS/TEN.

10.6 Severity scoring and causality assessment

• SCORTEN — seven independent mortality predictors, developed and validated in 165 + 75 patients:

  "We identified seven independent risk factors for death and constituted the toxic epidermal necrolysis-specific severity-of-illness score: age above 40 y, malignancy, tachycardia above 120 per min, initial percentage of epidermal detachment above 10%, serum urea above 10 mmol per liter, serum glucose above 14 mmol per liter, and bicarbonate below 20 mmol per liter. For each toxic epidermal necrolysis-specific severity-of-illness score point the odds ratio was 3.45 (confidence interval 2.26-5.25)." — Bastuji-Garin et al., J Invest Dermatol 2000 (PMID:10951229) Calibration was excellent in the derivation setting: "excellent agreement between expected (19.6%) and actual (20%) mortality; discrimination was also excellent with a receiver operating characteristic area of 82%."

• ALDEN (ALgorithm of Drug causality for Epidermal Necrolysis) — Sassolas et al. 2010 (PMID:20375998). The standard causality instrument; used to nominate allopurinol as the culprit when several drugs are in play.
• RegiSCAR score — for DRESS specifically, not SJS/TEN (PMID:42499701, PMID:42014364).

10.7 Differential diagnosis

10.8 Screening in asymptomatic individuals

The screening question here is pharmacogenomic pre-prescription genotyping, not disease screening. See §13.

11. Outcome / Prognosis

11.1 Mortality

• General SJS/TEN: mortality historically ~30% for TEN (PMID:10951229); modern SJS/TEN cohorts report lower figures with specialist care — 3.8% observed vs 8.6% SCORTEN-predicted in a 213-patient Chinese series (PMID:31898979). One-year mortality of 17% among Danish antiseizure-drug-associated SJS/TEN cases (PMID:42467870).
• Allopurinol-specific: allopurinol-SCAR carries a worse prognosis than SCAR from many other culprits, and the driver is renal.

  "Impaired renal function and increased plasma levels of oxypurinol and granulysin correlated with the poor prognosis of allopurinol-SCAR." — Chung et al. 2015 (PMID:25115449) "Patients with allopurinol-induced SCARs with renal impairment have significantly higher risk of mortality." — Wang, Dao & Chung, Curr Opin Allergy Clin Immunol 2016 (PMID:27362322)

• A precise allopurinol-SJS/TEN-specific case-fatality rate from a large modern cohort was not identified in this search. Curate the general SJS/TEN figures and the allopurinol-specific direction of effect, and flag the exact number as a gap.

11.2 Morbidity, complications, and disability

Acute complications (from a 213-case SJS/TEN series, PMID:31898979):

Sepsis from the denuded skin is the leading proximate cause of death. Other acute complications: hypovolemia, AKI, ARDS/bronchiolitis obliterans, GI hemorrhage (18.2% under corticosteroids vs 2.6% under etanercept, PMID:29400697), thermoregulatory failure, protein-losing catabolic state.

Chronic sequelae — this is a disease that does not stop when the skin heals: - Ocular: chronic dry eye, symblepharon, trichiasis, limbal stem cell deficiency, corneal neovascularization and opacification, blindness. The most common and most disabling long-term sequela. - Cutaneous: dyspigmentation, scarring, nail dystrophy/anonychia, chronic pruritus, alopecia. - Mucosal: oral synechiae, esophageal strictures, vaginal/urethral stenosis, phimosis. - Pulmonary: obstructive disease, bronchiolitis obliterans. - Cardiovascular — recently quantified and previously underappreciated:

"compared with non-SJS/TEN participants, patients with SJS/TEN had higher risks of cardiovascular morbidity (CVA: HR, 1.65 [95% CI, 1.57-1.72] … IHD: HR, 1.58 [95% CI, 1.51-1.65] …) and death due to cardiovascular disease (CVA: HR, 1.69; 95% CI, 1.46-1.96; IHD: HR, 1.55; 95% CI, 1.32-1.82). The increased cardiovascular mortality risks peaked at 1 year after SJS/TEN and persisted for 4 to 7 years." — Chiu & Chiu, JAMA Dermatol 2025 (PMID:39969876) — n = 10,571 (CVA cohort) / 11,084 (IHD cohort) SJS/TEN survivors, Taiwan NHIRD 1998–2021 - Psychological: PTSD, depression, anxiety — well described in survivor cohorts. - Iatrogenic: lifelong avoidance of allopurinol and uncertainty about related urate-lowering options; loss of a first-line gout therapy in someone who often has CKD and few alternatives.

11.3 Prognostic factors

• SCORTEN (age, malignancy, tachycardia, BSA, urea, glucose, bicarbonate) — PMID:10951229.
• Renal function / eGFR — allopurinol-specific, and the strongest disease-specific modifier (PMID:25115449).
• Plasma oxypurinol level and its trajectory after withdrawal (PMID:25115449).
• Plasma granulysin level (PMID:25115449).
• Speed of culprit-drug withdrawal.
• ICU admission at diagnosis predicts higher long-term cardiovascular mortality (PMID:39969876).

12. Treatment

12.1 The one intervention everyone agrees on

Immediate and permanent withdrawal of allopurinol, plus lifelong documented avoidance. Everything else is contested.

• NCIT: NCIT:C15747 Supportive Care; NCIT:C49236 Therapeutic Procedure
• Therapeutic modality: BEHAVIORAL is a poor fit; this is really "drug discontinuation," for which no clean NCIT clinical-action term was identified. Consider a free-text preferred_term with no term: binding.

12.2 Supportive care — the backbone

Wound care, fluid/electrolyte resuscitation, nutritional support, analgesia, temperature control, infection surveillance (without prophylactic systemic antibiotics), airway assessment, early ophthalmology involvement (amniotic membrane transplantation for the ocular surface), urology/gynecology for genital care. Management in a burn unit or specialized SCAR center where available.

"There is currently no evidence-based treatment outside of aggressive supportive care" — Justice, Mukherjee, Martin-Pozo & Phillips, Allergol Int 2025 (PMID:40473510)

That sentence, from a 2025 review by the group running the definitive trial, is the honest summary of the field and should be quoted directly in the entry.

12.3 TNF-α blockade (etanercept) — the strongest positive RCT evidence

Randomized trial, n = 96 SJS/TEN patients, etanercept vs corticosteroids (NCT01276314):

"Etanercept improved clinical outcomes in patients with SJS-TEN. Etanercept decreased the SCORTEN-based predicted mortality rate (predicted and observed rates, 17.7% and 8.3%, respectively). Compared with corticosteroids, etanercept further reduced the skin-healing time in moderate-to-severe SJS-TEN patients (median time for skin healing was 14 and 19 days for etanercept and corticosteroids, respectively; P = 0.010), with a lower incidence of gastrointestinal hemorrhage in all SJS-TEN patients (2.6% for etanercept and 18.2% for corticosteroids; P = 0.03)." — Wang et al., J Clin Invest 2018 (PMID:29400697)

Mechanistic corroboration in the same trial: "etanercept decreased the TNF-α and granulysin secretions in blister fluids and plasma (45.7%-62.5% decrease after treatment; all P < 0.05) and increased the Treg population (2-fold percentage increase after treatment; P = 0.002)." That is a rare instance of a trial validating the mechanism it targets — good material for a treatments.target_mechanisms link with its own evidence.

• Annotation: treatment_term NCIT:C15986 Pharmacotherapy; therapeutic_agent NCIT:C2381 Etanercept; therapeutic_modality: MONOCLONAL_ANTIBODY — strictly, etanercept is a soluble receptor–Fc fusion protein, not a monoclonal antibody. PROTEIN_REPLACEMENT is also wrong. OTHER is the honest value; flag this as a modality-enum gap.
• Supporting data in adjacent settings: TNF blockade in ICI-induced SJS/TEN "showed a significantly rapid recovery and no recurrence of SCAR with continuous ICI therapy" (PMID:39737932); etanercept + corticosteroid combination in PD-1-inhibitor SJS/TEN gave re-epithelialization in 5.75 ± 1.48 days (n = 5, PMID:41418419). Both are non-allopurinol contexts — grade accordingly.

12.4 Ciclosporin

Widely used, mechanistically plausible (calcineurin inhibition blocks CTL activation), supported by cohort studies and meta-analyses suggesting a mortality benefit — but no adequately powered RCT. therapeutic_agent: CHEBI:4031 cyclosporin A / CHEBI:748220 cyclosporine; therapeutic_modality: SMALL_MOLECULE; treatment_term NCIT:C15986.

12.5 Systemic corticosteroids

The historical default, and still the comparator arm in the etanercept trial. Associated there with slower healing and markedly more GI hemorrhage (18.2% vs 2.6%; PMID:29400697). Evidence remains "a subject of ongoing debate, with inconsistent evidence regarding their efficacy" (PMID:40909037). NCIT:C2322 Corticosteroid as therapeutic_agent.

12.6 IVIG and plasmapheresis

A large Japanese administrative-database comparison of second-line therapy after corticosteroid failure (1,215 screened; 53 plasmapheresis-first vs 213 IVIG-first) found no significant difference in inpatient mortality between the two (PMID:36884227). IVIG's Fas-blockade rationale has not translated into consistent survival benefit. - NCIT:C121331 Intravenous Immunoglobulin Therapy; NCIT:C15304 Plasmapheresis.

12.7 The definitive trial: NATIENS

⚠️ Read this one carefully before curating. The ClinicalTrials.gov record returns an enrollment count of 2 alongside status COMPLETED. That is almost certainly a study that closed with minimal accrual rather than a completed 100+ patient trial, but I did not find a published results paper to confirm either reading. Do not curate NATIENS as having answered the question. Record status: COMPLETED, phase: PHASE_III, and put the accrual caveat in notes:.

12.8 Experimental / emerging

• JAK inhibitors (tofacitinib and others) — case reports and small series; rationale is IFN-γ/JAK-STAT blockade. No controlled data identified.
• Necroptosis inhibition (RIPK1/RIPK3/MLKL) — "Inhibition of necroptosis completely prevented SJS/TEN-like responses in a mouse model of SJS/TEN" (PMID:25031270). Preclinical only; evidence_source: MODEL_ORGANISM.
• Granulysin neutralization — direct target validation exists (depletion reduced cytotoxicity, PMID:19029983); no clinical agent.
• Accelerated oxypurinol clearance (hemodialysis/hemoperfusion) — proposed but untested: "An early intervention to increase the clearance of plasma oxypurinol may improve the prognosis of allopurinol-SCAR" (PMID:25115449). This is an allopurinol-specific therapeutic hypothesis with no other analogue in SCAR, and worth curating as such.

12.9 Pharmacogenomics — replacing the drug

• HLA-B*58:01–positive patients should not receive allopurinol. CPIC assigns a Level 1A recommendation; the 2015 update confirmed the 2013 recommendations stand:

  "We reviewed the recent literature and concluded that none of the evidence would change the therapeutic recommendations in the original guideline; therefore, the original publication remains clinically current." — Saito et al., Clin Pharmacol Ther 2016 (PMID:26094938)

• Febuxostat (CHEBI:31596) is the principal substitute — structurally unrelated, and immunologically non-cross-reactive (PMID:25946710). Its own cardiovascular safety signal (CARES trial) is a separate consideration.
• Probenecid (uricosuric) — the Thai cost-effectiveness model found HLA screening cost-effective only if probenecid was the alternative (PMID:40829930).
• Desensitization protocols exist for mild allopurinol rash but are contraindicated after SJS/TEN.

12.10 Treatment algorithm (synthesis)

1. Recognize; stop allopurinol immediately; stop all other non-essential drugs (ALDEN to assign causality).
2. Assess BSA detachment; calculate SCORTEN; check eGFR and, if available, plasma oxypurinol.
3. Transfer to burn unit / specialist center. Institute aggressive supportive care.
4. Early ophthalmology, urology/gynecology, ENT/pulmonology consultation.
5. Consider systemic immunomodulation — etanercept has the best randomized evidence (PMID:29400697); ciclosporin is a reasonable alternative; corticosteroids carry a GI bleeding penalty; IVIG/plasmapheresis show no clear benefit over one another.
6. On recovery: document allergy in every record system; HLA-B58:01 typing (for the patient and* consider family cascade testing given the shared haplotype); switch to febuxostat or a uricosuric; long-term ophthalmologic follow-up; recognize the elevated cardiovascular risk in survivorship care.

13. Prevention

13.1 Primary prevention — this is where the wins are

(a) Don't prescribe allopurinol when it isn't indicated. ACR 2020 conditionally recommends against urate-lowering therapy for asymptomatic hyperuricemia. Every avoided prescription is an avoided exposure.

(b) Start low, go slow. ACR 2020 strongly recommends:

"using a low starting dose of allopurinol (≤100 mg/day, and lower in CKD) or febuxostat (<40 mg/day); and a treat-to-target management strategy with ULT dose titration guided by serial serum urate (SU) measurements, with an SU target of <6 mg/dl." — FitzGerald et al., Arthritis Rheumatol 2020 (PMID:32390306)

Note that the same guideline still names allopurinol "the preferred first-line ULT, including for those with moderate-to-severe chronic kidney disease (CKD; stage >3)" — the low-dose-start recommendation is what makes that safe.

(c) Pre-prescription HLA-B*58:01 genotyping. The strongest prevention evidence in the entire report:

"Participants who tested positive for HLA-B58:01 (19.6%, n=571) were advised to avoid allopurinol… SCARs did not develop in any of the participants receiving allopurinol who screened negative for HLA-B58:01. By contrast, seven cases of SCARs were expected, based on the estimated historical incidence of allopurinol induced SCARs nationwide (0.30% per year, 95% confidence interval 0.28% to 0.31%; P=0.0026…)" — Ko et al., BMJ 2015 (PMID:26399967), 2,910 participants, 15 Taiwanese medical centres

NCIT:C92803 Genetic Screening; NCIT:C200724 Cascade Testing.

13.2 Who should be screened — and the economics

• CPIC: if HLA-B*58:01 genotype is known and positive, allopurinol is contraindicated (PMID:26094938). CPIC does not mandate universal testing.
• ACR 2020: conditionally recommends HLA-B58:01 testing before allopurinol in patients of Southeast Asian descent and African American* patients (PMID:32390306).
• Cost-effectiveness varies sharply by country and by what the alternative drug is:

  "Testing of HLA-B58:01 is more likely to be cost-effective throughout Africa and Asia compared with Europe and the Americas." — Zhou et al. 2021 (PMID:32535895) "HLA-B58:01 testing was not cost-effective before allopurinol initiation in Thai patients with gout at the current price of 1,000 THB per test. However, HLA-B58:01 testing would be cost-effective if only probenecid was the alternative treatment for patients with positive HLA-B58:01 results." — Dilokthornsakul et al., ACR Open Rheumatol 2025 (PMID:40829930), ICER 1,093,068 THB/QALY ≈ $31,404/QALY

This is a nice illustration that a screening program's value depends less on the test than on how good the second-line drug is.

13.3 The implementation gap

Recommended ≠ done. A VA study of clinical decision support:

"The percentage of Asian or African American/Black patients who had HLA-B58:01 testing before or during the month allopurinol was prescribed increased from 8.8% in October 2022 to 35.5% in December 2023 at the dashboard + BPA site and from 2.1% to 4.5% at the dashboard only site (P < .0001 for difference in difference)." — Fadairo-Azinge et al., Arthritis Care Res 2026 (PMID:41664544*)

Baseline testing rates of 2–9% in a guideline-eligible population, rising only to ~35% with an active alert. National implementation programs exist (e.g. Singapore — PMID:39405418).

13.4 Secondary and tertiary prevention

• Secondary (early detection): patient counseling to stop the drug and seek care at the first rash — especially in the first 8 weeks; clinician education on the skin-pain-out-of-proportion red flag.
• Tertiary (preventing complications in those affected): early transfer to specialist care; aggressive ocular surface management to prevent blindness; sepsis surveillance; permanent allergy documentation; cardiovascular risk-factor management in survivors, given HR ~1.6 for CVA and IHD persisting 4–7 years (PMID:39969876).

13.5 Genetic counseling, immunization, public health

• Genetic counseling: relevant in a narrow sense. HLA-B58:01 is inherited; first-degree relatives of a case have elevated carriage probability, and cascade testing before their* allopurinol prescription is defensible. This is pharmacogenomic counseling, not reproductive counseling.
• Immunization: not applicable.
• Public health / environmental interventions: regulatory labeling (HLA-B*58:01 appears in Korean drug labels and, less consistently, US labels — PMID:41846573); pharmacovigilance; national genotype-guided prescribing programs.

14. Other Species / Natural Disease

14.1 Is there a natural animal counterpart?

Not for allopurinol-induced SJS/TEN specifically. But TEN as a drug-hypersensitivity phenotype does occur naturally in companion animals:

"Cutaneous drug allergies in veterinary medicine can have a variety of clinical manifestations, ranging from pruritus to often fatal toxic epidermal necrolysis… There are multiple theories that attempt to explain how drug allergies occur… These include the (pro)-hapten hypothesis, the Danger Theory, the pi concept, and the viral reactivation theory." — Voie, Campbell & Lavergne, J Vet Intern Med 2012 (PMID:22519673)

Note that the p-i concept — the exact mechanism at work in allopurinol-SJS/TEN — is already part of the veterinary conceptual framework. But no HLA-B*58:01 orthologue exists: MHC class I alleles are not conserved across species at allelic resolution, and the dog leukocyte antigen (DLA) and feline leukocyte antigen (FLA) systems have their own architectures. There is no reported drug–MHC association in animals analogous to HLA-B*58:01/allopurinol.

• Species: NCBITaxon:9615 Canis lupus familiaris (dog); NCBITaxon:9685 Felis catus (cat) — for TEN generally, not for this drug.
• MONDO carries MONDO:1013225 toxic epidermal necrolysis, non-human animal.
• Breed (VBO): no breed predisposition to drug-induced TEN identified. Not applicable.

14.2 Allopurinol in veterinary medicine

Allopurinol is used long-term in dogs, chiefly for canine leishmaniosis and for urate urolithiasis in Dalmatians. Adverse effects reported are predominantly xanthine urolithiasis (an on-target pharmacological effect, not an immune one), not epidermal necrolysis (e.g. PMID:20178476, PMID:10622623). The immune-mediated SCAR phenotype does not appear to be a recognized entity in dogs on allopurinol — a striking negative that fits the HLA-restriction model, since dogs have no HLA-B*58:01.

14.3 Comparative biology and evolutionary conservation

• Conserved: the effector machinery — granulysin has no direct mouse orthologue (mice use NK-lysin), but perforin/granzyme, Fas/FasL, and the RIPK1/RIPK3/MLKL necroptosis axis are broadly conserved across mammals.
• Not conserved: the initiating recognition event. HLA allelic specificity is human-specific by definition. This is the central translational barrier for the disease and the reason animal modeling is so limited (§15).
• Zoonotic potential / cross-species transmission: not applicable.

15. Model Organisms

Bottom line first: this disease is exceptionally hard to model, because the initiating lesion is an interaction between a specific human MHC allele and a specific small molecule. You can model the dying (steps 6–11 of §6.1) reasonably well. You cannot easily model the recognizing (steps 3–5).

15.1 Human ex vivo systems — the workhorse, and effectively the primary model

These are not animal models and belong in experimental_models: as NAMs:

Notable design feature of the Mifsud study: it drew T-cell lines from drug-naive healthy donors as well as resolved cases (PMID:37452515) — a clean way to separate "the repertoire exists" from "the repertoire was expanded by disease."

15.2 Animal models

• Mouse granulysin injection model — recombinant 15-kDa secretory granulysin injected intradermally: "injection of it into mouse skin resulted in features mimicking SJS-TEN" (PMID:19029983). Models the effector step only. relationship: PARTIALLY_RECAPITULATES, fidelity: MODERATE, limitations: no drug, no HLA, no T cell — this is a pure downstream-mediator model.
• Mouse SJS/TEN necroptosis model — used to show that "Inhibition of necroptosis completely prevented SJS/TEN-like responses in a mouse model of SJS/TEN" (PMID:25031270). Same caveat: effector-arm only.
• HLA-B*58:01 transgenic mouse: no such model was identified in this search. Repeated targeted PubMed queries for HLA-B58:01 transgenic/humanized mice returned nothing. (For contrast, HLA-B57:01 transgenic mice exist for abacavir hypersensitivity, which is the obvious template.) This is the single largest model-system gap for this disease and a strong candidate for a HUMAN_MODEL_MISMATCH discussion entry, since it means every mechanistic claim about the initiating recognition event rests on human ex vivo work alone.

15.3 Model limitations, stated plainly

1. No animal expresses HLA-B*58:01 naturally. The initiating event is unmodellable without transgenesis.
2. Murine xanthine oxidase metabolism differs, so oxypurinol exposure is not equivalent.
3. Granulysin has no direct mouse orthologue, so the dominant human effector cannot be studied in a wild-type mouse.
4. Keratinocyte death models (granulysin injection, necroptosis inhibition) reproduce the lesion without reproducing its cause — useful for therapeutics aimed at steps 6–11, useless for prevention.
5. No organoid or skin-on-chip model of SJS/TEN was identified. Given that this disease is fundamentally an immune–epithelial interface problem, a co-culture immunocompetent skin construct with HLA-B*58:01-matched donor cells is the obvious missing NAM.

15.4 Resources

• MGI, IMPC, IMSR for the necroptosis-pathway alleles (Ripk1, Ripk3, Mlkl, Fasl) — all exist and are well characterized, but for necroptosis biology generally, not for this disease.
• Cellosaurus / ATCC for HLA-typed B-lymphoblastoid lines used as APCs.
• The NATIENS (NCT02987257) biorepository and the SJS/TEN Survivor Study (PMID:41160012) are the two most important human sample resources.

Appendix A — Curation-ready ontology suggestions

Module conformance candidates

Check just list-modules before creating anything new, but these look like natural conformance targets: - A treatment-toxicity / "side effect as mechanism" module — this entry is a textbook member of that family. - A T-cell-mediated cytotoxicity or adaptive immune effector module, if one exists. - A necroptosis / regulated cell death module for steps 8–10. - Not an infection or autoimmunity module — the trigger is exogenous and pharmacological, and it resolves on withdrawal.

Appendix B — Explicit gaps (candidate discussions with kind: KNOWLEDGE_GAP or HUMAN_MODEL_MISMATCH)

1. No HLA-B*58:01 transgenic animal model exists. Every claim about the initiating drug–HLA–TCR event rests on human ex vivo data. → HUMAN_MODEL_MISMATCH.
2. Why the same allele + same drug yields SJS/TEN in one carrier and DRESS in another is unknown. Variable expressivity with no explanation.
3. No allopurinol-specific case-fatality rate from a large modern cohort was identified; existing figures are for SJS/TEN pooled across culprits.
4. The p-i mechanism vs. the altered-peptide-repertoire mechanism are both supported and not reconciled. → mechanistic_hypotheses with two groups, status: EMERGING for the altered-repertoire model (PMID:37296297).
5. Macrophage-derived CXCL10 is established in ICI-induced SJS/TEN but untested in allopurinol-induced disease (PMID:39737932).
6. The annexin A1–FPR1 necroptosis arm (PMID:25031270) was demonstrated in SJS/TEN generally, not in allopurinol-specific cases.
7. Enhanced oxypurinol clearance as therapy is a stated hypothesis with no trial (PMID:25115449).
8. HLA-A*34:02 rests on 16 cases from a single center (PMID:41160012) and needs replication.
9. No QoL instrument data specific to allopurinol-induced SJS/TEN.
10. NATIENS results are unpublished and the registry enrollment figure suggests the trial may not have accrued.
11. No single-cell atlas of allopurinol-induced SJS/TEN lesional skin.

Sources

Primary literature (PubMed) - Hung SI et al. HLA-B5801 allele as a genetic marker for severe cutaneous adverse reactions caused by allopurinol. PNAS 2005;102(11):4134-9. PMID:15743917 · doi:10.1073/pnas.0409500102 - Halevy S et al. Allopurinol is the most common cause of Stevens-Johnson syndrome and toxic epidermal necrolysis in Europe and Israel. J Am Acad Dermatol 2008;58(1):25-32. PMID:17919772 - Chung WH et al. Granulysin is a key mediator for disseminated keratinocyte death in Stevens-Johnson syndrome and toxic epidermal necrolysis. Nat Med 2008;14(12):1343-50. PMID:19029983 - Bastuji-Garin S et al. SCORTEN: a severity-of-illness score for toxic epidermal necrolysis. J Invest Dermatol 2000;115(2):149-53. PMID:10951229 - Bastuji-Garin S et al. Clinical classification of cases of toxic epidermal necrolysis, Stevens-Johnson syndrome, and erythema multiforme. PMID:8420497 - Sassolas B et al. ALDEN, an algorithm for assessment of drug causality in Stevens-Johnson syndrome and toxic epidermal necrolysis. PMID:20375998 - Yun J et al. Allopurinol hypersensitivity is primarily mediated by dose-dependent oxypurinol-specific T cell response. Clin Exp Allergy 2013;43(11):1246-55. PMID:24152157 - Yun J et al. Oxypurinol directly and immediately activates the drug-specific T cells via the preferential use of HLA-B58:01. J Immunol 2014;192(7):2984-93. PMID:24591375 - Saito N et al. An annexin A1-FPR1 interaction contributes to necroptosis of keratinocytes in severe cutaneous adverse drug reactions. Sci Transl Med 2014;6(245):245ra95. PMID:25031270 - Chung WH et al. Insights into the poor prognosis of allopurinol-induced severe cutaneous adverse reactions. Ann Rheum Dis 2015;74(12):2157-64. PMID:25115449 - Chung WH et al. Oxypurinol-specific T cells possess preferential TCR clonotypes and express granulysin in allopurinol-induced severe cutaneous adverse reactions. J Invest Dermatol 2015;135(9):2237-48. PMID:25946710 - Lin CH et al. Immunologic basis for allopurinol-induced severe cutaneous adverse reactions. J Allergy Clin Immunol 2015;135(4):1063-1065.e5. PMID:25458913 - Ko TM et al. Use of HLA-B58:01 genotyping to prevent allopurinol induced severe cutaneous adverse reactions in Taiwan. BMJ 2015;351:h4848. PMID:26399967 - Saito Y et al. CPIC guidelines for HLA-B genotype and allopurinol dosing: 2015 update. Clin Pharmacol Ther 2016;99(1):36-7. PMID:26094938 - Wang CW, Dao RL, Chung WH. Immunopathogenesis and risk factors for allopurinol severe cutaneous adverse reactions. Curr Opin Allergy Clin Immunol 2016;16(4):339-45. PMID:27362322 - Wang CW et al. Randomized, controlled trial of TNF-α antagonist in CTL-mediated severe cutaneous adverse reactions. J Clin Invest 2018;128(3):985-996. PMID:29400697 - Yang L et al. Retrospective study of 213 cases of Stevens-Johnson syndrome and toxic epidermal necrolysis from China. Burns 2020;46(4):959-969. PMID:31898979 - FitzGerald JD et al. 2020 American College of Rheumatology Guideline for the Management of Gout. Arthritis Rheumatol 2020;72(6):879-895. PMID:32390306 - Zhou Y et al. Global frequencies of clinically important HLA alleles and their implications for the cost-effectiveness of preemptive pharmacogenetic testing. Clin Pharmacol Ther 2021;109(1):160-174. PMID:32535895 - Bathini L et al. Initiation dose of allopurinol and the risk of severe cutaneous reactions in older adults with CKD. Am J Kidney Dis 2022;80(6):730-739. PMID:35644439 - Miyamoto Y et al. Evaluation of plasmapheresis vs immunoglobulin as first treatment after ineffective systemic corticosteroid therapy for patients with SJS/TEN. JAMA Dermatol 2023;159(5):481-487. PMID:36884227 - Huan X et al. Allopurinol non-covalently facilitates binding of unconventional peptides to HLA-B58:01. Sci Rep 2023;13(1):9373. PMID:37296297 - Mifsud NA et al. The allopurinol metabolite, oxypurinol, drives oligoclonal expansions of drug-reactive T cells in resolved hypersensitivity cases and drug-naïve healthy donors. Allergy 2023;78(11):2980-2993. PMID:37452515 - Chen CB et al. Immune checkpoint inhibitor-induced severe epidermal necrolysis mediated by macrophage-derived CXCL10 and abated by TNF blockade. Nat Commun 2024;15(1):10733. PMID:39737932 - Chiu HY, Chiu YM. Risk of cardiovascular morbidity and mortality in Stevens-Johnson syndrome/toxic epidermal necrolysis survivors. JAMA Dermatol 2025;161(4):391-398. PMID:39969876 - Justice J, Mukherjee E, Martin-Pozo M, Phillips E. Updates in the pathogenesis of SJS/TEN. Allergol Int 2025;74(3):361-371. PMID:40473510 - Dilokthornsakul P et al. An updated economic evaluation of HLA-B58:01 genotype testing in gouty patients for preventing severe allopurinol hypersensitivity in Thailand. ACR Open Rheumatol 2025;7(8):e70093. PMID:40829930 - Campbell CN, Krantz MS, Yu A, Phillips EJ; SJS/TEN Survivor Study Collaborators. HLA-B58:01 and risk of allopurinol-induced severe cutaneous adverse reactions in the US. JAMA Dermatol 2025;161(12):1258-1263. PMID:41160012 (preprint PMID:40475155) - Xiong H, Shen Z. Increased TNF-α in SJS/TEN induced by PD-1 inhibitors supports the combination therapy of etanercept and systemic corticosteroids. Mol Immunol 2026;189:174-180. PMID:41418419 - Fadairo-Azinge A et al. Effect of implementing a dashboard with or without a best practice alert on HLA-B58:01 testing rates among allopurinol users at VA medical centers. Arthritis Care Res 2026;78(9):1265-1269. PMID:41664544 - Lee HK et al. Application of risk priority number of failure mode and effects analysis to drug-variant pairs for severe cutaneous adverse reactions in Korean and American populations. Pharmacogenet Genomics 2026;36(4):125-133. PMID:41846573 - Ünal D et al. Utility of patch testing, LTT, and HLA genotyping in identifying culprit drugs and diagnosing multiple drug hypersensitivity in definite DRESS. Clin Transl Allergy 2026;16(4):e70172. PMID:42014364 - Heerfordt IM et al. Risk of Stevens-Johnson syndrome and toxic epidermal necrolysis after initiation of antiseizure medication: a Danish nationwide cohort study. Neurology 2026;107(3):e218355. PMID:42467870 - Voie KL, Campbell KL, Lavergne SN. Drug hypersensitivity reactions targeting the skin in dogs and cats. J Vet Intern Med 2012;26(4):863-74. PMID:22519673 - Gonçalo M et al. HLA-B58:01 is a risk factor for allopurinol-induced DRESS and SJS/TEN in a Portuguese population. Br J Dermatol 2013. PMID:23600531 - Jung JW et al. HLA-DR9 and DR14 are associated with the allopurinol-induced hypersensitivity in hematologic malignancy. Tohoku J Exp Med 2014. PMID:24858023 - Chua HM et al. Implementation of HLA-related genotype-guided prescribing in Singapore. Am J Health Syst Pharm 2025;82(5):e285-e293. PMID:39405418

Databases and registries - ClinicalTrials.gov NCT02987257 — NATIENS · NCT01276314 — etanercept RCT - EBI Ontology Lookup Service (OLS4) — MONDO, HP, GO, CL, UBERON, CHEBI, NCIT term verification - HGNC REST API — gene identifier verification - NCBI MeSH — D013262 - PharmGKB / CPIC allopurinol guideline - ICD-11 MMS EB13 block - Orphanet — SJS/TEN spectrum (ORPHA:95455) — code from secondary sources; portal blocked automated retrieval, verify before binding - DermNet NZ — SJS/TEN - Mayo Clinic Laboratories — HLA-B*5801 genotype test · ARUP Consult — HLA-B*58:01 genotyping

Reference Validation

Checked with linkml-reference-validator 0.2.1.

Quotes not found in the cited source

Searched the abstract, any retrieved full text, and the title. A quote drawn from a part of the paper that was not retrieved will appear here too, so check before treating one as invented:

Every one of these was searched against an abstract alone, with no full text retrieved - marked abstract only below. Where full text can be fetched, re-running with it will settle them; where the source publishes only a summary to PubMed, as GeneReviews chapters do, it will not, and the quote has to be checked by hand against the chapter itself.

• PMID:24152157 (abstract only): "TCL induction data show that both the presence of HLA-B58:01 allele and high concentration of drug are important for the generation of drug-specific T cells… functional avidity of ALP/OXP-TCL is dependent on both the induction dose and HLA-B58:01 status."
• closest text in source: "TCL induction data show that both the presence of HLA-B*58:01 allele and high concentration of drug are important for the generation of drug-specific T cells"
• PMID:17919772 (abstract only): "restricted to short-term use (≤8 weeks)"
• closest text in source: "The risk was restricted to short-term use (<or=8 weeks)"
• PMID:24152157 (abstract only): "Allopurinol allergic patients are primarily sensitized to oxypurinol in a dose-dependent manner… OXP-TCLs do not recognize allopurinol and vice versa."
• closest text in source: "RESULTS: Allopurinol allergic patients are primarily sensitized to oxypurinol in a dose-dependent manner"