Stevens–Johnson syndrome after three weeks of carbamazepine: causality, referral, and reporting
A 41-year-old woman from Darkhan was started on carbamazepine 200 mg twice daily three weeks ago for trigeminal neuralgia. She presents to the aimag hospital with four days of fever up to 39.2 °C, painful erosions of the mouth and lips, and a spreading dusky maculopapular rash; today about 12% of her body surface has begun to detach and the Nikolsky sign is positive. She also takes a combined oral contraceptive and denies any other medicine or herbal product. She had bought paracetamol and an antihistamine from a local pharmacy. Her 19-year-old nephew had a similar rash years ago after an antiepileptic. The hospital has no dermatology unit; burns intensive care is in Ulaanbaatar, seven hours away by road, and the outdoor temperature is -25 °C. Her daughter asks whether the rash came from the 'nerve tablets'.
1. State the most likely diagnosis and the single triggering medicine, and justify both from the timing, the pattern, and her history.
Read the worked answer
Diagnosis: Stevens–Johnson syndrome/toxic epidermal necrolysis overlap - fever plus painful mucosal erosions plus dusky, confluent rash with a positive Nikolsky sign and about 12% body-surface detachment (SJS is under 10%, overlap 10–30%, TEN above 30%).
Trigger: carbamazepine. The reasoning is not recall, it is pattern-matching:
- Latency: aromatic anticonvulsants characteristically cause severe cutaneous reactions 2–8 weeks after starting. Three weeks fits exactly; paracetamol, taken for a few days during the fever, does not.
- Drug class: carbamazepine is one of the commonest causes of SJS/TEN worldwide, and risk is strongly linked to HLA-B*15:02, which is frequent in East and Southeast Asian populations. A nephew's severe reaction to an antiepileptic is a genuine clue to familial susceptibility, not a coincidence.
- Exclusion: antihistamines and combined oral contraceptives are rare causes. There is no other new medicine and no herbal product, so there is no competing explanation to dilute causality.
- Severity predictors: age over 40, tachycardia, fever, and 12% detachment give an SCORTEN that predicts substantial mortality; this is a medical emergency, not a dermatology outpatient problem.
2. Prioritise management for the first six hours, including whether and how she should be transferred.
Read the worked answer
First, stop the drug - carbamazepine is withdrawn immediately and permanently; this is the single most important action and it must be written in the chart.
Supportive care in the first hours
- Airway and mucosal assessment: mouth and lip erosions can progress to pharyngeal involvement; have suction and an airway plan, and involve anaesthesia early before intubation becomes difficult.
- Fluids and electrolytes: large insensible losses through denuded skin; start isotonic crystalloid and monitor urine output.
- Eyes: daily ophthalmology review, lubricant drops and ointment, symblepharon prevention - ocular damage is permanent if neglected.
- Skin: non-adherent dressings, no adhesive tapes, analgesia (paracetamol or opioid; avoid NSAIDs), temperature management, and sepsis surveillance without routine prophylactic antibiotics.
- Do not start corticosteroids routinely; immunomodulation (e.g. cyclosporine or IVIG) is a referral-centre decision.
Transfer She needs burns-unit or intensive-care level nursing, which this hospital cannot provide, so transfer to Ulaanbaatar is indicated - but she must not travel cold and exposed. Arrange: ambulance with a warmed cabin, warmed blankets and heat packs, fluids running, airway equipment and a staff member able to manage the airway, and pre-notification of the receiving unit so a bed is ready. Discuss with the receiving intensivist by phone before departure; if the airway is already threatened, stabilise and secure it locally first.
3. How would you assess causality and report this case nationally, and what advice prevents a recurrence in this patient and her family?
Read the worked answer
Causality Using WHO-UMC criteria: the reaction appeared within a plausible interval after a drug with a well-documented association (2–8 weeks for aromatic anticonvulsants), the clinical picture is a recognised hypersensitivity syndrome, no other cause was identified, and she is improving after withdrawal (dechallenge). That supports probable to certain; rechallenge is absolutely contraindicated and must never be attempted.
National reporting Complete the national adverse drug reaction report to the National Agency for Medicines and Medical Devices using the expedited route for serious reactions - do not wait for the outcome, and mention the nephew's reaction in the narrative as a signal for a possible familial/pharmacogenetic cluster. Attach the drug batch and the dates of starting and stopping. In the hospital, enter the reaction in the allergy field of the medical record, not only in free-text notes.
Prevention advice
- Give her a written card: 'Carbamazepine - Stevens–Johnson syndrome. Avoid carbamazepine, oxcarbazepine, phenytoin, phenobarbital, and caution with lamotrigine.' Cross-reactivity between aromatic anticonvulsants is substantial.
- Tell every future prescriber and dentist, and tell the daughter that any relative needing carbamazepine should be tested for HLA-B*15:02 first.
- For the trigeminal neuralgia, refer for a non-aromatic option (e.g. gabapentin, baclofen, or a surgical/neuromodulation referral) rather than simply restarting another anticonvulsant at home.