Do HLA-A markers predict skin-reactions from aromatic antiepileptic drugs in a Norwegian population? A case control study.

Abstract Purpose Cutaneous adverse reactions (cADRs) from carbamazepine (CBZ) have been associated with human leukocyte antigens (HLA). Our aims were to assess the clinical usefulness of HLA-A*31:01 as a predictor of CBZ-induced cADRs in the Norwegian population and to explore whether cADRs from aromatic antiepileptic drugs (AEDs) in general might be linked with a common HLA-A-marker. Materials and methods 86 ethnic Norwegians with a history of non-bullous cADRs from aromatic AEDs were included. 114 subjects tolerant to at least one aromatic AED were used as drug-specific controls. Complete HLA-A genotyping was performed. 1026 blood donors were used as population controls. Results Comparing all cADR subjects with controls and blood donors, there were no statistical differences for any HLA-A allele, except for HLA-A*24 ( p =0.022 vs. controls and p =0.014 vs. blood donors). When comparing tolerant controls with patients having had a cADR to one of the two most used drugs, CBZ ( n =48) and lamotrigine ( n =28), we found no significant associations for CBZ to HLA-A*31:01 or HLA-A*24:02 , but for lamotrigine there was an association with HLA-A*24:02 ( p =0.027). In patients developing cross-reactivity ( n =14) to aromatic AEDs, the presence of HLA-A*31:01 or HLA-A*24:02 was not different compared to patients with a single cARD tolerant to at least one other drug. Conclusion We question the clinical usefulness of HLA-A*31:01 as a marker for CBZ rash in the Norwegian population. A previously suggested protective effect of aromatic AED cross-reactivity from HLA-A*24:02 was not confirmed. The association between HLA-A*24:02 and lamotrigine-induced rash should be further investigated.