Stevens-Johnson syndrome and toxic epidermal necrolysis
Stevens-Johnson syndrome and TEN: recognising a drug emergency
Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are severe drug reactions causing full-thickness epidermal detachment and painful mucosal erosions. Mortality is 5–10% for SJS and 30–50% for TEN.
Top causative drugs in Australia: allopurinol (HLA-B*58:01 carriers), aromatic anticonvulsants including carbamazepine and lamotrigine (HLA-B*15:02 carriers), sulfonamide antibiotics, and oxicam NSAIDs. Onset is typically 4–28 days after starting the drug.
Immediate GP actions: stop all suspected culprit drugs; call 000 for Category-1 transfer to a burns unit or ICU; never restart the culprit. Report to TGA and flag in My Health Record.
A drug reaction that kills: why speed matters
Stevens-Johnson syndrome and toxic epidermal necrolysis (SJS/TEN) are rare — approximately 1–10 cases per million Australians per year — but when they occur, mortality is unacceptably high if the diagnosis is delayed or the culprit drug is not immediately stopped. TEN carries a mortality of 30–50% and exceeds 90% in patients with the highest severity scores.
The single most important determinant of outcome is how quickly the culprit drug is stopped and how quickly the patient reaches a burns unit. Both of these depend on the GP making the diagnosis.
A. Core clinical — the AU general-practice framework
Classification
SJS/TEN exists as a spectrum defined by the percentage of body surface area (BSA) affected by epidermal detachment:
| Diagnosis | BSA detachment | Mortality |
|---|---|---|
| Stevens-Johnson syndrome (SJS) | <10% | 5–10% |
| SJS/TEN overlap | 10–30% | 20–30% |
| Toxic epidermal necrolysis (TEN) | >30% | 30–50%; >90% if SCORTEN ≥5 |
Erythema multiforme major (EM major) — post-herpetic, with typical 3-zone target lesions, acral distribution — is a separate entity and not the same condition.
Risk factors and high-risk populations
The reaction is T-cell-mediated and idiosyncratic — it is dose-independent. A single tablet can trigger the reaction; risk is highest in the first 4–8 weeks of a new drug.
High-risk populations in Australia:
- HIV/AIDS — approximately 100× background rate, particularly with cotrimoxazole and nevirapine
- Han Chinese, Thai, Vietnamese, Filipino, Malay ancestry — HLA-B15:02 (aromatic anticonvulsants); HLA-B58:01 (allopurinol)
- Aboriginal and Torres Strait Islander Australians — elevated HLA-B*58:01 carrier frequency in some communities; coordinate allopurinol prescribing accordingly
- Active malignancy on immunotherapy or radiation
- CKD on dialysis
The prodrome and early recognition
The prodrome resembles viral illness: fever, malaise, sore throat, and conjunctivitis — 1–3 days before the rash. This prodrome is frequently managed as an URTI, delaying recognition.
The rash:
- Starts as dusky red or purple macules on the trunk, spreading centrifugally to face and limbs
- Atypical target lesions — purpuric centre with less defined zones than EM major
- Flaccid bullae form on the dusky macules
- Positive Nikolsky sign — skin shears with light tangential pressure
- Skin pain — the single most important discriminating feature; painful skin in any drug rash is SJS/TEN until proved otherwise
Mucosal involvement (≥2 sites in ~90% of cases):
- Oral: painful erosions, haemorrhagic crusts on lips
- Ocular: conjunctivitis, photophobia, mucopurulent discharge, pseudomembrane
- Genital, urethral, anal: dysuria, urinary retention, dyschezia
- Respiratory mucosal involvement (dysphonia, dyspnoea) — poor prognostic sign
The drug history
The diagnosis almost always lives in the drug history. Ask about every prescribed, OTC, complementary, and herbal product taken in the past 8 weeks, with exact start dates.
Top culprits in Australian general practice:
Allopurinol — single most common cause; risk highest in first 8 weeks; start at 50 mg/day and titrate slowly; HLA-B*58:01 testing recommended before prescribing in high-prevalence populations (PBS subsidy applies — confirm current criteria).
Aromatic anticonvulsants — carbamazepine, lamotrigine, phenytoin, oxcarbazepine, phenobarbital; HLA-B*15:02 testing before carbamazepine in high-risk ancestries per TGA guidance (Chen NEJM 2011); cross-reactivity exists within the class — lifelong avoidance applies to all aromatic anticonvulsants.
Cotrimoxazole and sulfasalazine — particularly at high risk in HIV/AIDS.
Oxicam NSAIDs — piroxicam, meloxicam, oxaprozin.
Nevirapine — class effect for non-nucleoside reverse transcriptase inhibitors.
Severity scoring: SCORTEN
Calculate SCORTEN within 24 hours of presentation (Bastuji-Garin J Invest Dermatol 2000):
| Factor | Score |
|---|---|
| Age >40 | 1 |
| Heart rate >120 | 1 |
| Active malignancy | 1 |
| BSA detachment >10% | 1 |
| Serum urea >10 mmol/L | 1 |
| Serum bicarbonate <20 mmol/L | 1 |
| Serum glucose >14 mmol/L | 1 |
Predicted mortality: 0–1 = 3%, 2 = 12%, 3 = 35%, 4 = 58%, ≥5 = >90%.
Immediate GP actions
Step 1: Stop the culprit. Stop all drugs started in the past 4–28 days. Stop all non-essential medications. Document precisely: drug name, dose, start date, lot number if available.
Step 2: Transfer immediately. Call 000. Pre-notify the ED with an ISBAR: “Category-1 transfer, suspected SJS/TEN, BSA detachment approximately [%], mucosal sites [list], suspected culprit [drug name].” Contact the regional burns unit directly. For regional and remote AU, activate RFDS retrieval early — do not wait for local management to fail.
Photograph the rash (with patient consent) — digital photographs are invaluable for specialist assessment on arrival, particularly for patients who deteriorate rapidly during transfer.
Pre-transfer supportive care:
- Fluid resuscitation (like a major burn; titrate to urine output 0.5–1 mL/kg/h)
- Analgesia — paracetamol first-line; opioid if required; note: if paracetamol is the suspected culprit, use an alternative
- Non-adherent wound dressings; avoid silver-sulphadiazine (contains sulfa) and chlorhexidine on raw mucosa
- Warm room (epidermal barrier failure causes heat loss)
- Ophthalmology consult on same day if possible — ocular synechiae develop rapidly
Never restart the culprit drug. Rechallenge is associated with an accelerated and more severe reaction. This is a lifelong contraindication.
Hospital management (specialist)
Specific immunomodulatory therapy remains controversial — no randomised trial demonstrates clear mortality benefit for any single agent. Supportive care (wound care, nutrition, eye care, fluid balance, sepsis prevention) is the cornerstone.
Cyclosporine (3–5 mg/kg/day for 7–14 days) has the most observational support; meta-analyses suggest mortality reduction of approximately 30–40%; use is specialist-led given nephrotoxicity and infection risk.
IVIG — mixed evidence; no consistent mortality benefit.
Corticosteroids — observational data suggests increased infection and mortality in extensive TEN; may be used for early SJS at some centres.
Etanercept — single-centre RCT (Wang Lancet 2018) suggested faster healing with a single 50 mg SC dose versus corticosteroid; emerging evidence only, not yet AU standard.
Prophylactic antibiotics are not recommended — they select resistant organisms without mortality benefit. Empirical broad-spectrum antibiotics are appropriate at the first sign of infection.
B. Evidence and controversies
HLA screening before high-risk drugs
Chen NEJM 2011 — universal HLA-B*15:02 screening before carbamazepine in Han Chinese reduced CBZ-induced SJS to near zero in Taiwan, demonstrating the principle of pharmacogenomic prevention. Cost-effectiveness is strongest in high-prevalence populations.
Hung PNAS 2005 established the HLA-B*58:01 association with allopurinol-induced SJS/TEN, now widely validated across Han Chinese, Thai, Vietnamese, Korean, and Aboriginal populations. The PBS subsidises pre-prescription testing in indicated groups — confirm current criteria.
Cyclosporine
Multiple retrospective and prospective cohort studies, synthesised in the Cochrane review (Mohanty 2022), support cyclosporine as the specific therapy with the strongest observational backing currently available. The effect size is uncertain due to confounded observational data, but the signal is consistent. It remains the first choice at most Australian burns and dermatology centres managing SJS/TEN.
SCORTEN validation
SCORTEN has been externally validated across multiple cohorts and predicts short-term mortality reliably within 24 hours of presentation. Calculate and document it — it informs ICU-level decision-making and family communication about prognosis.
C. SJS and TEN in Australian practice
MBS items
- Standard consultation (items 23, 36, 44) for recognition and transfer
- Urgent and after-hours (items 597–600)
- HLA typing — Medicare-funded in indicated pharmacogenomic indications; confirm current PBS criteria for pre-allopurinol or pre-carbamazepine testing
- Chronic Disease Management Plan (items 965/967) — post-SJS/TEN chronic sequelae (eye, skin, psychological)
- Mental Health Care Plan (items 2715/2717) — 10 Better Access psychology sessions per year
- Telehealth post-discharge (items 91890/91891) — for existing-patient follow-up
Mandatory reporting and drug allergy flagging
TGA Adverse Drug Reaction reporting — all suspected SJS/TEN cases should be reported at tga.gov.au/safety/adverse-events using the Blue Card. This is how signals of new culprit drugs are detected.
My Health Record — flag the drug allergy immediately. Cross-reactive drugs in the same class (all aromatic anticonvulsants if carbamazepine is the culprit; febuxostat if allopurinol is the culprit) should also be flagged.
MedicAlert bracelet — strongly recommended for all SJS/TEN survivors.
Coronial referral if the patient dies — any death from drug-induced severe cutaneous adverse reaction (SCAR) requires referral.
Aboriginal and Torres Strait Islander considerations
The higher HLA-B*58:01 carrier frequency in some Aboriginal communities makes pre-prescription testing before allopurinol a particularly important practice. Engage with the local ACCHO when prescribing allopurinol to Aboriginal patients and make HLA testing part of the consent conversation. Closing the Gap PBS Co-payment covers ongoing topical and systemic medications post-discharge.
D. Special populations
Children and adolescents. Paediatric SJS/TEN is more often triggered by Mycoplasma pneumoniae than in adults — an URTI with target-like lesions and mucosal involvement in a child warrants Mycoplasma serology. Transfer to a state children’s hospital burns unit; management principles are similar to adults.
HIV/AIDS. Risk of SJS/TEN is approximately 100 times the background rate, primarily driven by cotrimoxazole and nevirapine. Counsel patients and their prescribers about the warning signs before initiating these agents.
Older adults. SCORTEN scores are higher in older patients due to the age criterion and comorbidities. Higher mortality risk should inform early transfer decisions and advance care planning conversations.
Patients on anticonvulsant therapy. Abrupt withdrawal of anticonvulsants creates seizure risk. Hospital management of the SJS/TEN reaction includes safe anticonvulsant substitution — this is a specialist decision, not a reason to delay stopping the culprit.
When to escalate
Immediate (000 / Cat-1 transfer to burns unit or ICU):
- Any painful drug rash
- Mucosal involvement (oral, ocular, genital) with a new or recent drug
- Skin detachment or positive Nikolsky sign
- Fever with conjunctivitis and a rash on a new medication
Post-discharge — refer urgently if:
- Worsening ocular symptoms — corneal erosion, visual change
- Signs of infection in healing skin
- Urine retention from genital or urethral involvement
Post-discharge routine:
- 1-week GP review: drug reconciliation, wound healing, mental health screen
- 1 month and 3 months: ophthalmology plus dermatology plus GP
- Annually: ophthalmology review for all survivors
What this article is and is not
This is general health information drawn from Australasian College of Dermatologists, ANZBA, Therapeutic Guidelines (eTG), the 2024 European S3 guideline, and published trials. It is not personal medical advice and does not replace urgent clinical assessment and specialist management.
For AU consumer information: HealthDirect, Better Health Channel, TGA adverse event reporting.
Sources cited
- Australasian College of Dermatologists
- ANZBA
- Therapeutic Guidelines (eTG)
- S3 European guideline for epidermal necrolysis (Paulmann 2024)
- Society of Dermatology Hospitalists 2020
- Bastuji-Garin — SCORTEN (J Invest Dermatol 2000)
- Bastuji-Garin — SJS/TEN classification (Arch Dermatol 1993)
- Chen — HLA-B*15:02 screening (NEJM 2011)
- Hung — HLA-B*58:01 and allopurinol (PNAS 2005)
- Wang — etanercept for SJS/TEN (Lancet 2018)
- TGA — Adverse Drug Reaction Reporting
- HealthDirect — Stevens-Johnson syndrome
- Better Health Channel
Frequently asked questions
-
How do I recognise SJS or TEN early?
The prodrome resembles a viral illness: fever 38–40°C, malaise, sore throat, and conjunctivitis, appearing 1–3 days before the rash. The rash begins as dusky red or purple macules on the trunk, spreading centrifugally to the face and limbs. Atypical targets (purpuric centre, less than 3 zones) or dusky macules, combined with painful skin and mucosal involvement, distinguish SJS/TEN from morbilliform (non-blistering) drug rashes. The Nikolsky sign — skin shearing with light tangential pressure — confirms epidermal detachment. The critical discriminating feature is skin pain: a painful drug rash should be treated as SJS/TEN until proved otherwise. Involvement of ≥2 mucosal sites (oral, ocular, genital) is present in approximately 90% of cases.
-
Which drugs most commonly cause SJS and TEN in Australia?
Allopurinol is the single most common cause in Australia, particularly in patients who are HLA-B*58:01 carriers — a genotype prevalent in Han Chinese, Thai, Vietnamese, Korean, and some Aboriginal and Torres Strait Islander populations. Aromatic anticonvulsants (carbamazepine, lamotrigine, phenytoin, oxcarbazepine, phenobarbital) are the second major group, with strong HLA-B*15:02 association in Han Chinese, Thai, Vietnamese, Filipino, and Malay patients. Other important culprits: cotrimoxazole and sulfasalazine; oxicam NSAIDs (piroxicam, meloxicam); nevirapine (HIV antiretroviral); and, rarely, paracetamol. The reaction onset window is typically 4–28 days after first starting the drug — outside this window, existing drugs are unlikely culprits.
-
What is HLA-B*58:01 and B*15:02 testing and who should have it?
HLA-B*58:01 is a human leukocyte antigen allele strongly associated with allopurinol-induced SJS/TEN, particularly in East and South-East Asian and some Aboriginal populations. The PBS subsidises HLA-B*58:01 testing before initiating allopurinol in high-prevalence populations — confirm current PBS criteria at the time of prescribing. Testing before prescribing in these populations effectively eliminates the pharmacogenomic risk. HLA-B*15:02 testing is recommended before starting carbamazepine in patients of Han Chinese, Thai, Vietnamese, Filipino, or Malay ancestry per TGA precaution, as carbamazepine-induced SJS rates are dramatically reduced by pre-prescription screening (Chen NEJM 2011). Lamotrigine, oxcarbazepine, and phenobarbital cross-react with B*15:02 — the same populations require the same precaution for these agents.
-
What burns units in Australia manage SJS and TEN?
Adult burns units that manage SJS/TEN include: Royal Adelaide Hospital; Royal Brisbane and Women's Hospital; Concord Hospital (Sydney); Royal North Shore Hospital; Royal Melbourne Hospital; Royal Perth Hospital; and Royal Hobart Hospital. The Australian and New Zealand Burn Association (ANZBA) maintains a referral directory. Paediatric cases should be directed to state children's hospitals (Royal Children's Hospital Melbourne, Sydney Children's Hospital, Queensland Children's Hospital, and equivalents). For patients in regional or remote Australia, Royal Flying Doctor Service (RFDS) retrieval should be activated early — do not wait for local management to fail before calling for transfer.
-
What are the long-term consequences of SJS or TEN?
Survivors of SJS/TEN carry significant long-term sequelae. Ocular complications are the most disabling: synechiae (adhesions between the eyelid and eyeball), corneal scarring, chronic dry eye, and — in severe cases — permanent visual impairment or blindness. Annual ophthalmology review is essential for all survivors. Other long-term effects include oral xerostomia and taste changes; pulmonary bronchiolitis obliterans from airway mucosal involvement; cutaneous dyspigmentation, scarring, and nail dystrophy; and psychological consequences including PTSD, depression, and anxiety. A Mental Health Care Plan with referral to a psychologist should be arranged at the first post-discharge GP visit. The culprit drug — and all cross-reactive drugs in the same class — must be flagged in the patient's My Health Record as a lifelong allergy and a MedicAlert bracelet is recommended.
Source quality
Sources grouped by evidence tier. AU primary tier first; international where AU is silent or lagging; named-author reconstruction where guidelines have not yet caught up. How tiers work.
-
T1 AU primary 6 sources -
T2 International primary 2 sources -
T3 Named-author reconstruction 5 sources