A study of HLA class I and class II 4-digit allele level in Stevens–Johnson syndrome and toxic epidermal necrolysis

Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are represented by rare but life-threatening cutaneous adverse reactions to different drugs. Previous studies have found that in a Han Chinese population from Taiwan and other Asian Countries, a strong genetic association between HLA-class I alleles (B*15:02, B*58:01) and SJS and TEN was induced by carbamazepine and allopurinol, respectively. To identify genetic markers that covered the MHC region, we carried out a case-control association enrolling 20 Caucasian patients with SJS/TEN. Our patient series included 10 cases related to paracetamol, 7 to allopurinol and 3 to different drugs (plaquenil, itraconazol, nabumetone). Healthy controls were represented by 115 Caucasian bone marrow or stem cell donors. The HLA-A*, B*, C*, DRB1*, DQB1*, DQA1* and DPB1* genotyping were determined. The frequencies of HLA-A*33:03 as well as C*03:02 and C*08:01 were significantly higher in SJS/TEN patient subgroup showing allopurinol drug-induced severe cutaneous adverse reactions (SCAR) as compared to controls (28.6% vs 0%, P = 0.00002, Pc = 0.0011; 28.6% vs 0%, P = 0.00002, Pc = 0.001; 28.6% vs 0%, P = 0.00002, Pc = 0.001, respectively). In the same subgroup the frequencies of B*58:01, DRB1*15:02 and DRB1*13:02 alleles, although considerably higher than in control group (42.8% vs 5.2%, P = 0.003; 28.6% vs 1.7%, P = 0.005; 28.6% vs 3.5%, P = 0.037, respectively), appeared no more statistically different after P correction (Pc = 0.248; Pc = 0.29; Pc = 1.00, respectively). In addition, in 10 of the 20 SJS/TEN patient subgroup with paracetamol-induced SCAR no statistically significant association with HLA alleles could be found. However, in the same SJS/TEN patient subgroup showing allopurinol drug-induced SCAR, haplotype analysis indicated that B*58:01, DRB1*13:02 and DRB1*15:02 alleles, that in a single allele analysis lost statistical significance after P correction, may still confer susceptibility, because the B*58:01-DRB1*13:02 and DRB1*15:02-DQB1*05:02 are positively associated with the disease (14.2% vs 0.43%, P = 0.00001, Pc = 0.00028; 14.2% vs 0.43%, P = 0.00001, Pc = 0.00028, respectively). Our results show that in contrast to SCAR-related to paracetamol, where HLA alleles do not appear to be involved, HLA molecules behave as a strong risk factor for SCAR-related to allopurinol even when a limited number of patients are considered.