Impact of Pharmacogenetics on Pharmacokinetics of First-Line Antituberculosis Drugs in the HIRIF Trial

Abstract

Background: Variability in the pharmacokinetics (PK) of first-line antituberculosis drugs (rifampicin [RIF], isoniazid [INH], and pyrazinamide [PZA]) is high and may be influenced by pharmacogenetic polymorphism. We performed a pharmacogenetic substudy in 90 participants with PK data from the HIRIF trial in Peru.

Methods: Relevant single-nucleotide polymorphisms (SNPs) in the NAT2, SLCO1B1, AADAC, and AOX1 loci were genotyped using real-time polymerase chain reaction (PCR).

Results: The proportions of slow, intermediate, and fast acetylators predicted by a conventional 6-SNP NAT2 panel were 32.5%, 48.2%, and 19.2%, respectively. A single NAT2 tag SNP (rs1495741) agreed with the panel-predicted phenotype in 91% and was a better predictor of INH area under the curve (AUC). Accounting for discrepancies possibly caused by rare alleles not represented in the panel or that could be unequivocally resolved using observed AUC, sensitivity of the tag SNP was 97.7%. A previously described SNP in SLCO1B1 (rs4149032) was present at an allele frequency of 0.31 and appeared to influence RIF AUC and maximum concentration (Cmax) at a dose of 20 mg/kg, despite an extreme distribution of alleles across the randomized arms. The AADAC SNP (rs1803155) predominated in the study population and was not linked to RIF PK, although an effect could have been missed due to sample size and allele frequency. There was no association between PZA PK and a common SNP in AOX1 (rs55754655).

Conclusions: A tag SNP approach may offer simpler and cheaper prediction of INH PK. Further exploration of the impact of SLCO1B1 SNPs on RIF PK is required in this and other populations. Clinical Trials Registration. NCT01408914.

Keywords: NAT2; SLCO1B1; isoniazid; pharmacogenetics; pharmacokinetics; rifampicin; tuberculosis.

Figure 1.

Linkage disequilibrium (LD) heatmap for NAT2 alleles.

Figure 2.

Distribution of isoniazid (INH) area under the curve (AUC) with superimposed nonparametric scaled density estimates for rs1495741 genotypes (NAT2 tag single-nucleotide polymorphism).

Figure 3.

Rifampicin (RIF) plasma area under the curve (AUC) by dose level, stratified by SLCO1B1 genotype (rs4149032), wild type (left) and mutant (right), using a dominant genetic model. Boxplots show median, interquartile range, range (whiskers) and any outliers.