Exploring Associations of 6-Thioguanine Nucleotide Levels and Other Predictive Factors with Therapeutic Response to Azathioprine in Pediatric Patients with IBD Using Multilevel Analysis

Background:Metabolite monitoring and response predictors to azathioprine (AZA) in pediatric inflammatory bowel disease (IBD) are debatable. In an attempt to optimize thiopurine therapy and understand the mechanism of action of thiopurines, we correlated metabolites and other factors with AZA efficacy in children with IBD.Methods: Data from 86 children with IBD with 440 metabolite measurements were retrospectively analyzed using multilevel logistic regression analyses. A therapeutic response was defined as a pediatric Crohn's disease activity index 10 for Crohn's disease or a pediatric ulcerative colitis activity index 10 for ulcerative colitis without any treatment with steroids, antitumor necrosis factor, other immunomodulators, or exclusive enteral nutrition.Results: The 6-thioguanine nucleotide levels >250 pmol per 8 x 10(8) red blood cells correlated with a higher response (odds ratio, 4.14; 95% confidence interval, 1.49-11.46, P = 0.007), whereas 6-methyl-mercaptopurine and 6-methyl-mercaptopurine:6-thioguanine nucleotide ratio showed no correlation. Other novel response predictors in children with IBD were relative leukopenia (odds ratio, 14.01; 95% confidence interval, 3.77-52.10; P <0.001) and the absence of lymphopenia (odds ratio, 3.71; 95% confidence interval, 1.26-10.89; P = 0.017). Lower thiopurine methyltransferase activity (P = 0.015), lower platelet count (P = 0.020), and higher aspartate aminotransferase level (P = 0.009) also predicted therapeutic response. Age, gender, patient adherence, the duration of AZA therapy, IBD type, erythrocyte count, and erythrocyte sedimentation rate did not predict efficacy. The high interindividual variability accounting for 57.7% of variance in therapeutic response was observed.Conclusions: The significant 6-thioguanine nucleotide level-response relationship may support metabolite monitoring to improve thiopurine efficacy in pediatric IBD. The reported response predictors may be helpful for treatment optimization in AZA-treated children with IBD, but should be proved in prospective studies.