Stevens-Johnson syndrome and toxic epidermal necrolysis — branching viva

Opening question

Examiner: Take me through this child. What is the diagnosis, and what is your immediate frame? [6]

Candidate: This is Stevens-Johnson syndrome-toxic epidermal necrolysis overlap, a severe drug-induced mucocutaneous reaction. The keys are the carbamazepine started six weeks ago, the atypical target lesions, the sheet-like detachment of about 12 percent of body surface area, the positive Nikolsky sign, and the mucositis at two or more sites — the eyes and the mouth. My immediate frame is three things: stop the culprit drug at once, grade the severity with SCORTEN and admit to a burn-capable setting, and involve ophthalmology within hours to protect the eye. [6] [1]

Examiner: How do you classify this spectrum? [1]

Candidate: By the percentage of body-surface-area with epidermal detachment. Stevens-Johnson syndrome is under 10 percent, SJS-TEN overlap is 10 to 30 percent, and toxic epidermal necrolysis is more than 30 percent. All three share atypical target lesions, a positive Nikolsky sign and mucositis at two or more sites. This child, at 12 percent, is in the overlap band. [1]

Branch 1 — pathophysiology and pharmacogenetics

Examiner: Explain the mechanism by which a drug taken for weeks suddenly begins to kill the epidermis. [12]

Candidate: Stevens-Johnson syndrome and TEN are a type IV delayed drug-hypersensitivity reaction. Drug-specific CD8-positive cytotoxic T lymphocytes infiltrate the epidermis and release effector molecules against keratinocytes presenting the offending drug. The key mediator is granulysin — the most abundant and most cytotoxic molecule in early blister fluid — with the Fas-FasL death-receptor and perforin-granzyme pathways also contributing. The result is full-thickness epidermal necrosis at the dermo-epidermal junction. [12]

Examiner: Why is this child's carbamazepine particularly relevant pharmacogenetically? [4]

Candidate: Because HLA-B15:02 confers a strong risk of carbamazepine-induced SJS and TEN, and is common in Han Chinese, Thai, Malay, Filipino and Indonesian populations. The landmark discovery of this association introduced HLA screening into prescribing, and the European five-drug study confirmed the pharmacogenetic map. In at-risk populations, HLA-B15:02 screening before carbamazepine is recommended. HLA-B*58:01 is the analogous association for allopurinol. [4] [3]

Branch 2 — severity and the differential

Examiner: How will you grade her severity, and how does that change disposition? [1]

Candidate: I calculate SCORTEN on day one and re-check on day three. One point each for age over 40, active malignancy, heart rate over 120, detachment over 10 percent body surface area, urea over 10 mmol per litre, bicarbonate under 20, and glucose over 14. Her tachycardia, her detachment over 10 percent, and likely a rising urea or falling bicarbonate will score her highly, which means admission to a burn unit or PICU. The mortality bands run from about 3 percent at a score of 0 to 1 up to 90 percent at 5 or more. [1]

Examiner: How do you distinguish this from staphylococcal scalded skin syndrome? [12]

Candidate: The decisive feature is the mucosae. In SJS and TEN the mucosae are involved — the eyes and mouth are affected here. In staphylococcal scalded skin syndrome the mucosae are spared, because the exfoliative toxin cleaves desmoglein-1 in the superficial granular layer and the mucosa co-expresses desmoglein-3 which compensates. The split in SSSS is also superficial, not full-thickness. SSSS affects neonates and young infants and needs intravenous anti-staphylococcal therapy, whereas this child needs drug withdrawal and burns-style care. [12] [5]

Branch 3 — management and the supportive bundle

Examiner: Walk me through your stepwise management. [9]

Candidate: First, stop the carbamazepine immediately and withdraw all non-essential recent medicines, because earlier withdrawal improves survival. Second, calculate SCORTEN and admit to a burn unit or PICU given the overlap extent and her score. Third, deliver the supportive bundle: cautious intravenous fluids at about two-thirds of a thermal-burn formula to avoid fluid overload, early enteral nutrition, non-adherent wound care, analgesia, venous thromboprophylaxis, and targeted antibiotics only for proven infection. Fourth, involve ophthalmology within hours with daily review and amniotic membrane transplantation if the ocular involvement is severe. [9] [10]

Examiner: What about IVIG and cyclosporine? [9]

Candidate: I would not regard them as standard of care. The systematic review and meta-analysis of systemic immunomodulating therapies found no convincing survival benefit for IVIG, cyclosporine or corticosteroids, with heterogeneous and conflicting results and an association between corticosteroids and increased sepsis in some analyses. Supportive care is the standard; if the burns or intensive-care team considers an adjunct, it is discussed on its merits and never allowed to delay or replace the supportive bundle. [9]

Examiner: Name two interventions you would explicitly avoid. [10]

Candidate: I would not give prophylactic systemic antibiotics, because they do not prevent infection and drive resistance; antibiotics are reserved for proven or strongly suspected secondary infection. And I would not surgically debride the skin, because the necrosis is full-thickness but the dermis and its appendages survive, so the skin re-epithelialises from follicular stems and debridement destroys that regenerative capacity. [10]

Branch 4 — long-term and family care

Examiner: What will you tell the family about the future? [4]

Candidate: That she must never receive carbamazepine again and should avoid cross-reactive aromatic anticonvulsants such as phenytoin, lamotrigine and oxcarbazepine, with levetiracetam or valproate chosen for her ongoing seizure control. I will arrange MedicAlert identification, document the reaction and the culprit in her chart and the patient record, and offer HLA-B*15:02 screening of at-risk relatives before any aromatic anticonvulsant is prescribed. I will arrange ophthalmology follow-up for months regardless of how her eyes look at discharge, because late ocular complications develop silently. [4] [10]

Branch 5 — the Mycoplasma variant

Examiner: If instead this were a 12-year-old with a week of cough and fever, striking mouth, eye and urogenital ulceration, and almost no skin lesions, what would change? [5]

Candidate: I would consider Mycoplasma-induced rash and mucositis, which is triggered by Mycoplasma pneumoniae rather than a drug. The respiratory prodrome, the prominent mucositis at two or more sites and the sparse skin lesions are the clues, and there is no culprit drug to stop. Management centres on treating the Mycoplasma where indicated and on supportive mucosal and eye care. The prognosis is markedly better than drug-induced SJS, though ocular sequelae still occur and ophthalmology follow-up is essential. The key teaching point is not to over-attribute it to a medicine the child actually needs. [5]

Wrap

Examiner: Summarise the Stevens-Johnson syndrome and toxic epidermal necrolysis stance in one sentence. [1]

Candidate: Recognise the drug-induced blistering mucocutaneous reaction with mucositis at two or more sites, stop the culprit drug immediately, grade with SCORTEN and admit to a burn-capable setting, deliver burns-grade supportive care with ophthalmology at its centre, treat immunomodulation as unproven, and plan lifelong drug avoidance and HLA pharmacogenetic screening — because the child who is recognised early, whose drug is withdrawn, and whose eye is protected from the first hours is the child who survives and keeps their sight. [1] [10]