Genomics study of anti-tuberculosis drug-induced hypersensitivity reactions

Abstract

Introduction: All first-line anti-tuberculosis drugs can be associated with all phenotypes of cutaneous adverse drug reactions (CADR). Second-line drugs are associated with much poorer outcomes. Thus, identifying the offending drug in poly-pharmacy is difficult. Re -challenge with the drug is the gold standard in identifying the offender, however poses unacceptably high risk of CADR recurrence. Population and drug-specific genomics help identify those susceptible to adverse reactions to a drug facilitating avoidance of the drug. Objective: To investigate the genomic susceptibility in patients with confirmed rifampicin and or isoniazid-associated hypersensitivity reactions using both genome- wide association studies and candidate gene approaches. Methods: A case control study using 14 patients with previous tuberculosis-associated CADR who were re-challenged with first-line anti-tuberculosis drugs and subsequently developed re-challenge reactions to either isoniazid or rifampicin as cases. These were compared with 30 controls who had tolerated rifampicin and isoniazid during the re- challenge process (12 patients, Group 1a) and consecutive patients who had been on TB treatment for at least 12 weeks without developing any adverse drug reaction (1 8 patients, Group 1b) and 200 black South Africans from the general population. HLA genotypes of the samples were determined by SeCore® HLA Sequence based typing (Invitrogen, Life technologies, USA), and potential ambiguities were resolved by sequencing-based typing. Results: We found HLA-B*58:02 (OR=3.6; 95% CI: 1.4-8.99) and HLA-DRB1*09:01 (OR=15.3; 95% CI: 2.1-113.1) to be significantly more prevalent in patients who developed rifampicin and isoniazid-associated CADR as compared to black South African general population. However, we found no significant associations between HLA genotype and rifampicin/isoniazid-associated CADR when we compared the cases to our study controls that had tolerated rifampicin and isoniazid. HLA-B *58.02 was not found to be statistically associated with HIV positive status (p=0.42) and DRESS phenotype ( p= 0.6279). The majority of our cases were black Africans. Approximately 80% of our cases and controls were HIV-infected. DRESS/DIHS was the prevalent phenotype of CADR, accounting for approximately 80% of cases and controls. Discussion: To our knowledge, this is the first study to show an association between HLA-B*58:02 and HLA-DRB1*09:01 alleles and severe cutaneous adverse drugs reactions secondary to rifampicin and isoniazid in an African population. We identify 2 candidate HLA alleles that need confirmation of their association in African patients who develop rifampicin or isoniazid-associated CADR in larger studies. The value of identifying candidate alleles could lead to CADR preventative screening prior to initiating anti-tuberculosis therapy in black South Africans. The HLA-B*58:02 noted in our cases and controls tolerant of the drugs might not be associated with CADR but could be a reflection of the HIV status and control in HIV-TB co-infected persons. Conclusion: HLA-B*58:02 and HLA-DRB1*09:01 may be associated with rifampicin and isoniazid-associated CADR. Alternately HLA-B*58:02 may be associated with HIV status rather than CADR. A sufficiently powered study is needed to confirm this association.