The cost-effectiveness of a pharmacogenetic test: A trial-based evaluation of TPMT genotyping for azathioprine

摘要

Background Thiopurine-methyl transferase (TPMT) testing prior to the prescription of azathioprine in autoimmune diseases is one of the few examples of a pharmacogenetic test that has made the transition from research into clinical practice. TPMT testing could lead to improved prescribing of azathioprine resulting in a reduction in adverse drug reactions as well as an improvement in effectiveness. When allocating scarce resources robust evidence on cost-effectiveness is required. Objective This study aimed to evaluate the cost-effectiveness of a TPMT genotyping test to inform azathioprine prescribing in autoimmune diseases. The secondary aim of this study was to demonstrate the complexity of undertaking a trial-based evaluation of a pharmacogenetic test. Methods A prospective economic evaluation was conducted alongside the TARGET (TPMT: Azathioprine Response to Genotype and Enzyme Testing) study, a pragmatic controlled trial that randomized (1:1) patients to undergo TPMT genotyping before azathioprine (n = 167) or current practice (n = 166). Assuming the UK health service perspective and a time horizon of 4 months, resource-use and health status data were collected prospectively for all recruited patients. Results The mean incremental cost for TPMT genotyping and subsequent care pathways compared with current practice for the 4-month follow-up was -￡421.06 (95 confidence interval -￡925.15 to ￡89.75). Mean incremental quality-adjusted life-years were close to zero but negative: -0.008 (95 confidence interval -0.017 to 0.0002). Assuming a threshold of ￡20,000 per quality-adjusted life-year, the expected incremental net benefit of introducing the test is ￡256.89 (95 CI -￡425.94 to ￡932.86). Conclusions TPMT genotyping potentially offers a less expensive alternative than current practice, but it may also have a small but negative effect on health status. These findings are associated with significant uncertainty, and the causal effect of TPMT genotyping on changes in health status and health care resource use remains uncertain. The results from this study therefore pose a difficult challenge to decision makers.