Azathioprine (AZA) is an effective steroid-sparing therapy for chronically active Crohn's Disease (CD) and prevents relapse of steroid-induced remissions. However, despite efficacy in controlled trials, there is no dose-ranging data or prospective evidence on how to optimize therapy. It is now recognized that AZA, an inactive pro-drug, undergoes a series of enzymatic reactions leading to 6-thioguanine nucleotides (6- TGN), considered the active but, myelotoxic, metabolites. In a competing enzymatic pathway, thiopurine methyltransferase (TPMT) catalyzes formation of 6-methyl-mercaptopurine ribonucleotides (6-MMPR), metabolites that are therapeutically inactive and potentially hepatotoxic. Co-dominantly inherited polymorphic alleles confer high (TPMTH) and low (TPMTL) functional TPMT activity that impact AZA's therapeutic response and toxicity. Retrospective observations suggest that low levels of 6-TGN metabolites (due to under-dosing or high TPMT activity) are associated with poor therapeutic response. The study hypotheses are that pharmacogenetic variability in the metabolism of AZA impacts short and long-term therapeutic efficacy and tolerance in the treatment of CD, and that optimal dosing and response to treatment can be predicted based upon baseline TPMT activity and early metabolite levels after initial dose-escalation. Our objective is to determine optimal dosing and prediction of response by prospectively assessing TPMT enzyme activity levels and serial 6-TGN measurements for the management of steroid refractory and steroid-dependent CD in adults and children. To test the hypothesis, we propose a double blind, multi-center trial randomizing adult and pediatric patients with steroid-refractory and steroid dependent CD to either current standard AZA therapy dosed at 2.5mg/kg, or AZA, at a dose determined by their TPMT activity, and subsequently adjusted to maintain the 6-TGN levels within the proposed therapeutic range. Our primary endpoint will be to determine if there is a difference between fixed and individualized AZA therapy in the frequency of steroid-free disease remissions at 16 weeks. Secondary endpoints will be to assess the frequencies of remissions at 28 weeks and 52 weeks and to assess adverse events, corticosteroid requirements, and health related quality of life endpoints. Predictive models will be performed to assess responsiveness based upon initial and inducible TPMT activity and achievable 6-TGN and 6-MMPR metabolite based upon incremental AZA dosing.
