Objectives: To compare the cost effectiveness of azathioprine (AZA), methotrexate (MTX) and no immunosuppression for maintaining remission of moderate to severe inflammatory bowel disease (IBD) in New Zealand Caucasians, and to determine whether prospective testing for poor metabolisers of AZA by genotype or phenotype is cost effective.
Methods: Pharmacoeconomic models were developed to compare treatment costs and effects (QALYs) in theoretical populations of 1,000 IBD patients over a 1-year period. Efficacy and tolerability profiles for AZA and MTX were taken from the literature. The costs (year 2004 values) of the drugs and treatment of adverse effects were estimated from New Zealand drug and service costs. Representations of the patients' health-related quality of life (HR-QOL) were obtained from clinicians via the EQ-5D health state classification system and valued using the New Zealand EQ-5D social tariff. The effects of genotyping or phenotyping a population for thiopurine methyltransferase (TPMT) status were compared using the prevalence of TPMT deficiency in Caucasians, the relative risks of neutropenia and the associated costs.
Results: Net cost savings (vs no immunosuppressant treatment) of approximately 2.5 million and 1 million New Zealand dollars were realised for AZA and MTX, respectively, for the theoretical 1,000 patients, and AZA generated 877 QALYs compared with 633 for MTX. Phenotype and genotype testing generated net cost savings (vs no testing) of 120,000 and 11,000 New Zealand dollars, respectively. Savings related to phenotype tests were greater because of the lower assay costs of phenotype testing and a greater likelihood of pre-empting neutropenia.
Conclusion: Our model suggests that both MTX and AZA may generate significant net cost savings and benefits for patients with IBD in New Zealand, with AZA likely to be more cost effective than MTX. Prospective testing for poor metabolisers of AZA may also be cost effective, with phenotype testing likely to be more cost effective than genotype testing.