Polymorphisms in cytochrome P450 2C19 enzyme and cessation of leflunomide in patients with rheumatoid arthritis

Abstract

Introduction: Rational selection of disease modifying anti-rheumatic drugs in the treatment of rheumatoid arthritis (RA) has many potential advantages, including rapid disease control, reduced long-term disability and reduced overall cost to the healthcare system. Inter-individual genetic differences are particularly attractive as markers to predict efficacy and toxicity, as they can be determined rapidly prior to drug selection. The aims of this study, therefore, were to investigate the association between differences in genes associated with the metabolism, clearance and efficacy of leflunomide with its cessation in a group of rheumatoid arthritis patients who were treated with an intensive contemporary, treat-to-target approach.

Methods: This retrospective cohort study identified all individuals who received leflunomide and were enrolled in the Early Arthritis inception cohort at the Royal Adelaide Hospital between 2001 and July 2011. Inclusion criteria were age (>18) and a diagnosis of rheumatoid arthritis. Patients were excluded if a DNA sample was not available, if they withdrew from the cohort or if clinical data were insufficient. Subjects were followed for 12 months or until either another disease modifying antirheumatic drug was added or leflunomide was ceased. The following single nucleotide polymorphisms (SNPs) were determined: CYP2C19*2 (rs4244285), CYP2C19*17 (rs12248560), ABCG2 421C>A (rs2231142), CYP1A2*1F (rs762551) and DHODH 19C>A (rs3213422). The effects of variables on cessation were assessed with Cox Proportional Hazard models.

Results: Thirty-three of 78 (42.3%) patients ceased leflunomide due to side effects. A linear trend between cytochrome P450 2C19 (CYP2C19) phenotype and leflunomide cessation was observed, with poor and intermediate metabolizers ceasing more frequently (adjusted Hazard Ratio = 0.432 for each incremental change in phenotype, 95% CI 0.237 to 0.790, P = 0.006). Previously observed associations between cytochrome P450 1A2 (CYP1A2) and dihydro-orotate dehydrogenase (DHODH) genotype and toxicity were not apparent, but there was a trend for ATP-binding cassette sub-family G member 2 (ABCG2) genotype to be associated with cessation due to diarrhea.

Conclusions: CYP2C19 phenotype was associated with cessation due to toxicity, and since CYP2C19 intermediate and poor metabolizers have lower teriflunomide concentrations, it is likely that they have a particularly poor risk:benefit ratio when using this drug.

Figure 1

Rates of discontinuation of leflunomide. The solid line shows the actual rates of continuation of leflunomide and the dashed lines the 95% confidence interval.

Figure 2

Cox proportional hazard model to estimate the rates of discontinuation according to CYP2C19 phenotype (Hazard Ratio = 0.432 for each incremental change in phenotype, 95% Confidence Interval 0.237 to 0.790, P = 0.006 after adjustment for current leflunomide dose, continuation of triple therapy and positive rheumatoid factor at diagnosis).