Anticonvulsant Allergy: DRESS, SJS/TEN, and HLA-B*15:02 Pharmacogenomics

Anticonvulsant hypersensitivity syndrome, first characterized by Shear and Spielberg in the Journal of Clinical Investigation in 1988, is a severe systemic reaction to aromatic antiepileptic drugs presenting with fever, maculopapular rash, hepatitis, lymphadenopathy, and eosinophilia. It occurs in 1 per 1,000 to 1 per 10,000 treatment courses with carbamazepine, phenytoin, phenobarbital, or lamotrigine. Modern terminology recognizes AHS as essentially identical to DRESS syndrome. Mortality approaches 10–20%, primarily from hepatic necrosis, making it one of the most dangerous drug hypersensitivity reactions in clinical medicine. Immediate drug discontinuation is mandatory.

Aromatic AEDs — phenytoin, carbamazepine, phenobarbital, lamotrigine, primidone — all share the aromatic ring structural motif that CYP450 enzymes oxidize into reactive arene oxide metabolites. The shared mechanism of sensitization, rather than structural cross-reactivity per se, explains the 40–80% cross-reactivity figure established by Shear and Spielberg in 1988. Individuals sensitized to the arene oxide hapten from one aromatic AED will typically react to another because the same reactive intermediate is generated. This is why, after DRESS or SJS/TEN from any aromatic AED, the entire class must be permanently avoided in favor of non-aromatic alternatives like levetiracetam or valproate.

The FDA mandates HLA-B*15:02 genetic testing before starting carbamazepine in patients of Han Chinese, Thai, Malay, Filipino, Indonesian, Vietnamese, or other Southeast Asian ancestry. The allele occurs in 2–12% of Han Chinese and significantly less often in Europeans or Africans. Carriers face an odds ratio of 1,357 to 2,504 for carbamazepine-induced SJS/TEN (Chung et al., Nature 2004). The CPIC guideline recommends alternative AEDs for all HLA-B*15:02 carriers. HLA-B*15:02 also increases risk with phenytoin and oxcarbazepine, not only carbamazepine. Testing takes 1–2 weeks — urgent seizure situations may require bridging with a non-aromatic AED while awaiting results.

HLA-A*31:01 testing is increasingly recommended for European and Japanese patients before starting carbamazepine. McCormack et al. (NEJM 2011) identified HLA-A*31:01 as associated with a spectrum of carbamazepine hypersensitivity reactions in Europeans including DRESS, maculopapular eruption, and SJS/TEN — broader phenotype than HLA-B*15:02. The allele frequency is approximately 2–5% in Europeans and 9% in Japanese. CPIC guidelines recommend considering alternative AEDs for A*31:01 carriers. Unlike HLA-B*15:02, which is primarily a SJS/TEN risk allele, A*31:01 covers the full severity spectrum of carbamazepine hypersensitivity.

Lamotrigine-induced SJS/TEN risk is inversely related to the speed of dose escalation. Without slow titration, approximately 1% of patients develop SJS/TEN. The FDA-mandated slow titration protocol — starting at 25 mg/day for 2 weeks, then 50 mg/day for 2 more weeks, then increasing by 50 mg every 1–2 weeks — reduces incidence to 3.8 per 10,000. When lamotrigine is co-prescribed with valproate, valproate inhibits lamotrigine glucuronidation, raising lamotrigine blood levels substantially, requiring the dose to be cut in half. Exceeding the titration schedule, particularly in young patients or those on valproate, substantially increases SJS/TEN risk and is the most common preventable trigger.

Non-aromatic antiepileptic drugs are safe alternatives after DRESS, SJS, or TEN from an aromatic AED. Levetiracetam (Keppra) is first-line — broad-spectrum, no hepatic metabolism, minimal drug interactions, and no structural similarity to aromatic AEDs. Valproate (Depakote) is effective for many seizure types but carries hepatotoxicity risk in children under 2 and teratogenicity risk in women of childbearing age. Gabapentin and pregabalin are safe for neuropathic pain and partial seizures. Topiramate is effective for migraine prevention and multiple seizure types. Desensitization to aromatic AEDs is absolutely contraindicated after any SCAR reaction — avoidance of the entire aromatic AED class is mandatory.

DRESS can cause permanent organ damage depending on severity and the organ systems involved. Hepatic necrosis, the primary cause of DRESS-associated mortality, can leave fibrosis even in survivors who recover liver function. Interstitial nephritis from DRESS can progress to chronic kidney disease in a subset of patients, particularly if diagnosis and drug discontinuation are delayed. Thyroiditis developing as a DRESS complication sometimes evolves into permanent hypothyroidism requiring lifelong replacement. Ocular involvement from SJS/TEN — even without full TEN — can lead to conjunctival scarring, dry eye, and permanent visual impairment requiring ophthalmologic follow-up. The HHV-6 reactivation associated with DRESS can, rarely, cause encephalitis with lasting neurologic consequences.

HHV-6 (human herpesvirus 6) reactivation occurs in approximately 75% of DRESS cases and is thought to be driven by the profound immune activation triggered by the drug-specific T-cell response. The drug-induced immune dysregulation appears to disrupt normal viral latency control, allowing HHV-6 (and sometimes HHV-7, EBV, and CMV) to reactivate. This was characterized by Shiohara et al. in Allergology International 2006. The viral reactivation amplifies inflammation and contributes to the prolonged, fluctuating disease course that distinguishes DRESS from simpler drug rashes. It also means that DRESS resolves slowly — over 3–6 months on corticosteroids — rather than within days like a simple drug rash after drug discontinuation.

Fosphenytoin is a water-soluble phosphate ester prodrug of phenytoin used for IV loading — it is rapidly converted to phenytoin after administration. Patients with phenytoin allergy are also allergic to fosphenytoin because the active drug is phenytoin. Switching from phenytoin to fosphenytoin does not circumvent phenytoin allergy. Additionally, fosphenytoin contains phosphate ester and propylene glycol excipients that phenytoin does not, so in theory an excipient reaction to one formulation might not occur with the other — but the phenytoin-reactive patient should avoid both agents and switch to a non-aromatic AED. Phenytoin and fosphenytoin both cause anticonvulsant hypersensitivity syndrome via the arene oxide mechanism.

Anticonvulsant hypersensitivity syndrome shows familial clustering consistent with inherited genetic predisposition. First-degree relatives of AHS patients have approximately 25% risk of developing AHS if exposed to the same aromatic AED, compared to 1 in 1,000 to 1 in 10,000 in the general population. The inheritance is likely polygenic, involving epoxide hydrolase polymorphisms (determining arene oxide detoxification capacity) combined with HLA variants (determining T-cell recognition). First-degree relatives of patients who developed carbamazepine SJS/TEN should be HLA-B*15:02 tested before starting any aromatic AED. Family history of AHS should prompt consideration of non-aromatic AEDs as first-line therapy, particularly in Southeast Asian families where HLA-B*15:02 prevalence may be elevated.