Analysis of HLA-B Allelic Variation and IFN-γ ELISpot Responses in Patients with Severe Cutaneous Adverse Reactions Associated with Drugs.

BACKGROUND: The prevention and confirmation of drug-induced severe cutaneous adverse reactions (SCARs) are difficult.

OBJECTIVE: To determine the benefit of HLA-B allele prescreening and the measurement of drug-specific IFN-γ-releasing cells in the prevention and identification of the culprit drug in patients with SCARs.

METHODS: A total of 160 patients with SCARs were recruited from 6 university hospitals in Thailand over a 3-year period. HLA-B alleles were genotypically analyzed. The frequencies of drug-specific IFN-γ-releasing cells in patients with SCARs were also measured.

RESULTS: The drugs commonly responsible for SCARs were anticonvulsants, allopurinol, beta-lactams, antituberculosis agents, and sulfonamides. If culprit drugs had been withheld in patients carrying known HLA-B alleles at risk, it would have prevented 21.2% of SCAR cases, mainly allopurinol- and carbamazepine-related SCARs. Culprit drug-specific IFN-γ-releasing cells could be identified in 45.7% (53 of 116) of patients with SCARs caused by 5 major drug groups, particularly in patients diagnosed with drug reactions with eosinophilia and systemic symptoms (DRESS) (50.0%), followed by Stevens-Johnson syndrome/toxic epidermal necrolysis (46.0%), and acute generalized exanthematous pustulosis (31.3%). According to our study, high frequencies of drug-specific IFN-γ-releasing cells were significantly demonstrated in patients who suffered from DRESS phenotype, having anticonvulsants or the drugs belonging to the "probable" category based on the Naranjo algorithm scale, as the culprit drugs.

CONCLUSIONS: HLA-B prescreening would succeed in preventing only a minority of SCAR victims. Drug-specific IFN-γ-releasing cells are detectable in almost half of patients. Better strategies are required for better SCAR prevention and culprit drug confirmation.