Stevens-Johnson Syndrome and Toxic Epidermal Necrolysis (SJS/TEN)

Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are not 2 separate diseases. They are one severe, drug-induced mucocutaneous reaction — a form of epidermal necrolysis driven by widespread, cytotoxic T-cell and natural-killer-cell-mediated keratinocyte death — that differs from case to case only in how much body-surface area (BSA) the epidermis sloughs from. Both share the same high-risk culprit drugs, the same prodrome and mucosal findings, and the same management approach, which is why this chapter treats them as one condition classified by severity rather than as 2 unrelated entries.^[1]

Classification by body-surface-area detachment

Classify using the maximum detachment reached at any point during the illness, not only the detachment present at the first assessment — SJS can and often does progress to overlap or TEN over the following days. This is exactly why early recognition, immediate causative-drug withdrawal, and early transfer to a unit capable of burn-level care matter so much.^[2]

This is rare but genuinely life-threatening. A drug cause is identified in most adult cases; in children, infection — particularly Mycoplasma pneumoniae — accounts for a larger share (see Special Populations). Mortality rises sharply with the amount of skin involved and is estimated using the SCORTEN severity score (see Assessment and Diagnosis). The single most important intervention in this entire chapter is immediate identification and withdrawal of the causative drug — every other intervention is supportive or adjunctive.^[3]

Prodrome

A nonspecific prodrome of fever, malaise, sore throat, cough, and conjunctival burning or itching typically precedes skin findings by 1 to 3 days, occasionally longer. This is easy to mistake for a routine viral illness, and the diagnosis is often missed at this stage — which matters, because the causative drug is usually still being taken during it.

Cutaneous findings

• Ill-defined, dusky-red or purpuric macules and atypical (flat, 2-zone) target-like lesions, often starting on the trunk and face and spreading, sometimes rapidly, over hours to days.
• Lesions coalesce and progress to flaccid blisters and sheet-like epidermal detachment, exposing a raw, red, weeping dermis that resembles a partial-thickness burn.
• Positive Nikolsky sign — gentle lateral pressure or rubbing on erythematous (but not yet blistered) skin causes the epidermis to shear off. A positive Nikolsky sign strongly supports the diagnosis, but its absence does not rule it out.
• Skin pain, often described as burning, that is out of proportion to how the skin looks early on — pain disproportionate to visible findings is itself a warning sign, not a reassurance.

Mucosal findings

Mucosal involvement — painful erosions, blistering, and crusting — affects 2 or more mucosal surfaces in most cases, and is one of the features that most reliably distinguishes SJS/TEN from a routine morbilliform drug rash:

• Oral/oropharyngeal — painful erosions of the lips, buccal mucosa, and oropharynx; hemorrhagic crusting of the lips is a classic finding.
• Ocular — conjunctival injection, photophobia, pain, discharge, and erosions. This is the mucosal site with the greatest risk of severe, permanent long-term sequelae if not managed correctly from day 1 (see Non-Pharmacologic Treatment and Monitoring).
• Genital/urogenital — erosions of the genital mucosa are common and frequently under-examined; ask about and examine this site specifically rather than relying on the patient to raise it.

Oral mucosal desquamation and hemorrhagic lip crusting in Stevens-Johnson syndrome. Involvement of 2 or more mucosal sites is one of the most reliable clues distinguishing this from a routine drug rash. (James Heilman, MD, CC BY-SA 3.0, Wikimedia Commons)

Drugs — the dominant cause in adults

High-risk drug classes — covered with full cross-reactivity and avoidance guidance in Drug Tables — include the aromatic anticonvulsants (carbamazepine, oxcarbazepine, phenytoin, phenobarbital), lamotrigine, allopurinol, sulfonamide antibiotics (particularly sulfamethoxazole-trimethoprim), oxicam-class NSAIDs (piroxicam, meloxicam), nevirapine, and sulfasalazine.^[4]

Timing

For a first exposure to a new drug, onset is classically 4 to 28 days after starting it, most often within the first several weeks. On re-exposure to a drug that previously caused a reaction, onset can be far more rapid — sometimes within 48 hours. This timing window is central to identifying which of several recently started drugs is the likely culprit (see Assessment and Diagnosis, and the ALDEN causality method).^[5]

Genetic risk (HLA) — real, actionable pharmacogenomics

• HLA-B*15:02 and carbamazepine or oxcarbazepine — strongly associated with SJS/TEN specifically (not milder rashes) in patients of Han Chinese and broader Southeast Asian ancestry, where the allele is comparatively common. Current pharmacogenomic guidance recommends genotyping before starting either drug in a patient from a population where this allele is common, and avoiding both drugs if the patient tests positive and a suitable alternative exists.^[6]
• HLA-B*58:01 and allopurinol — associated with allopurinol-induced SJS/TEN, again with a markedly higher allele frequency in Han Chinese, Thai, and Korean populations; risk is compounded by higher allopurinol starting doses and renal impairment.^[1]

Other risk factors

• HIV infection increases the risk of SJS/TEN markedly, with the sharpest increase seen for sulfonamide antibiotics and nevirapine.
• Active malignancy, and radiotherapy given around the same time as an anticonvulsant.
• Active or recent infection — in children specifically, Mycoplasma pneumoniae can trigger a related but distinct entity, MIRM (Mycoplasma-induced rash and mucositis), with different management emphasis and a better prognosis than drug-induced disease (see Special Populations).
• A personal history of SJS/TEN — re-exposure to the same or a cross-reactive drug carries a high risk of a faster, often more severe recurrence.
• Connective tissue disease (e.g., systemic lupus erythematosus) has been proposed as a contributing factor in some patients.

A clinical diagnosis, supported by biopsy

Diagnosis is made on the clinical picture described above — prodrome, characteristic skin lesions, a positive Nikolsky sign, and mucosal involvement. Skin biopsy (with a portion sent for urgent frozen section where available) supports the diagnosis, typically showing full-thickness epidermal necrosis with minimal dermal inflammation, and helps exclude mimics — but do not wait for biopsy results before stopping the suspected causative drug.

Estimating body-surface-area detachment

Use the rule-of-nines, as in burn assessment, or the palm method — the patient's own palm and fingers, not the examiner's, represents approximately 1% BSA — to estimate the percentage of detached and detachable (Nikolsky-positive) skin. Reassess BSA at least daily during the acute phase, since detachment typically continues to progress for several days even after the causative drug is stopped.

Distinguishing SJS/TEN from its mimics

SCORTEN severity and mortality-prediction score

SCORTEN is calculated from 7 variables, each scored 1 point if present, assessed within the first 24 hours of admission and ideally recalculated on day 3, when it becomes more accurate. Use it to guide the level of care and to set realistic expectations with the patient and family — not as a gate for whether supportive care is offered.^[3]

Causality assessment — the pharmacist's central role

There is no laboratory test that confirms which drug caused an episode of SJS/TEN. Causality is assessed through structured clinical reasoning, and a pharmacist is often best positioned to do this well: build a complete, dated timeline of every prescription, over-the-counter product, herbal or natural health product, and recently restarted drug in the weeks before onset, including exact start dates. A structured algorithm such as ALDEN (Algorithm of Drug Causality for Epidermal Necrolysis) formalizes this using time-to-onset, whether the drug was still being taken at onset, response to withdrawal, prior reaction history, and how strongly that drug is already known to cause SJS/TEN — and performs meaningfully better than unstructured clinical judgment alone.^[5]

• Identify and immediately withdraw every drug that cannot be excluded as the cause — the single highest-yield intervention, and the only one with a clear, consistent mortality benefit.
• Prevent progression and further epidermal detachment through burn-level supportive and wound care.
• Prevent and promptly treat secondary infection and sepsis, the leading cause of death in TEN.
• Maintain fluid, electrolyte, thermal, and nutritional balance during the acute hypermetabolic phase.
• Prevent long-term mucosal and ocular sequelae through early, aggressive site-specific care.
• Control pain, which is often severe and persistent.
• Establish a confirmed causality assessment and a durable, documented future-avoidance plan before discharge.

Wound care

Manage detached and denuded skin as a partial-thickness burn: a warm, temperature-controlled environment (thermoregulation is impaired once a large surface area is denuded), minimal unnecessary handling, and nonadherent dressings. Current guidance favours a conservative approach to the necrotic epidermis itself — leaving it in place as a biologic dressing over the wound rather than aggressive surgical debridement, since it appears to protect the underlying dermis and reduce infection risk. This differs from traditional burn-unit teaching for thermal burns and should be discussed explicitly with the unit managing the patient.^[2]

Fluid and nutrition

Patients need aggressive fluid resuscitation and nutritional support, but fluid requirements in SJS/TEN are generally lower than for a thermal burn of the same body-surface area, because dermal capillary leak is less severe — applying standard burn-resuscitation formulas unmodified can lead to over-resuscitation. Start enteral nutrition early; the hypermetabolic state and substantial protein loss through denuded skin make adequate caloric and protein intake important for wound healing.^[2]

Mucosal and site-specific care

Infection prevention

Do not give prophylactic systemic antibiotics in the absence of documented infection — this does not reduce infection risk and contributes to resistance and adverse effects. Instead, monitor closely (see Monitoring and Follow-Up) and treat promptly and specifically once infection is actually identified.^[2]

Line and catheter minimization

Minimize invasive lines, catheters, and adhesive tape on compromised skin wherever possible — each is a potential entry point for infection and a source of further mechanical skin trauma.

Supportive care and immediate causative-drug withdrawal are the only interventions in SJS/TEN with consistent evidence behind them. Several systemic immunomodulating therapies are used adjunctively at many centres, but the evidence for each remains genuinely mixed, and there is no consensus recommendation to use any one of them routinely. Present this honestly rather than overstating a certainty that does not exist in the current literature — the decision, where one is made, belongs to the specialized team managing the patient, generally case by case.^[1]

Genetic counselling for blood relatives

Because HLA-B*15:02 and HLA-B*58:01 are inherited, a confirmed or strongly suspected HLA-linked reaction has implications beyond the patient in front of you: first-degree relatives of the same ancestry share a meaningfully elevated chance of carrying the same allele. Counsel the patient to inform blood relatives of the reaction and the specific drug involved, so this can be discussed with their own prescriber and factored into HLA testing decisions before that relative is ever started on the same drug — a genuine, practical extension of allergy counselling, not routine genetic-counselling overreach.

Arrange structured multidisciplinary follow-up after discharge — dermatology, ophthalmology, and, where genital or urethral scarring occurred, gynecology or urology — and monitor for longer-term skin, nail, and pigmentary changes. Before discharge, confirm the causality assessment was actually completed and that the resulting allergy documentation was communicated to the patient, their family physician, and their community pharmacy record, not only recorded in the hospital chart.

1. Confirm every possible causative drug was actually identified and stopped. A partial causality assessment that stopped only the "obvious" drug while a genuine culprit continued is a common, preventable cause of ongoing progression.
2. Confirm the patient has not been inadvertently re-exposed to a structurally related, cross-reactive drug (see Drug Tables) through a substitution made without this specific risk in mind.
3. Reassess for an undiagnosed or new secondary infection driving apparent deterioration, rather than assuming the underlying SJS/TEN is simply progressing on its own.
4. Confirm the patient has been transferred to, or is already at, a unit with genuine burn-level wound-care capability — ongoing deterioration in a unit without this capability is itself an indication to escalate the level of care.
5. Revisit the adjunctive-therapy discussion with the specialized team managing the patient (see Pharmacologic Treatment Options). This is a specialist, case-by-case decision, not a default step.

• Waiting for biopsy or specialist confirmation before stopping the suspected causative drug — every day of continued exposure worsens the prognosis.
• Stopping only the "most obvious" recently started drug while continuing others that cannot actually be excluded.
• Substituting one aromatic anticonvulsant for another (for example, carbamazepine to phenytoin) without accounting for real cross-reactivity risk within this drug group.
• Giving prophylactic systemic antibiotics without a documented infection.
• Aggressive surgical debridement of intact-but-detached epidermis rather than managing it as a biologic dressing.
• Underestimating severity because skin findings look mild while significant mucosal, especially ocular, involvement is present.
• Deferring ophthalmology consultation because eye symptoms seem minor at first assessment.
• Vague allergy documentation ("rash with antibiotic") instead of the specific drug, reaction, date, and severity — this actively endangers the patient at the next prescribing decision.
• Failing to counsel the patient, and blood relatives where an HLA-linked drug is involved, on the specific drug and drug-class avoidance needed going forward.

How to use this table: doses are adult doses and require patient-specific confirmation. Select any drug name to open its complete Pharmasist Drug Database page.

High-risk culprit drug classes

Evidence note: culprit-drug risk data reflect the RegiSCAR/EuroSCAR case-control medication-risk literature.^[4] HLA-genotyping guidance reflects the current Clinical Pharmacogenetics Implementation Consortium recommendation.^[6]

Adjunctive immunomodulating therapies — formulation detail

None of these has unambiguous, consensus-level evidence of benefit (see Pharmacologic Treatment Options); use is a specialized-team, case-by-case decision.