• Autoimmune hepatitis
• 6-mercaptopurine
• azathioprine

We congratulate Gleeson and Heneghan1 on their detailed and comprehensive guidelines on the management of autoimmune hepatitis (AIH). A noteworthy area for comment relates to the use of azathioprine (AZA) in the management of AIH and the utilisation of thiopurine metabolites to detect hepatotoxicity. As general hepatologists we are most familiar and comfortable with the use of AZA compared with other immunosuppressive agents and feel that every effort should be made to maintain our patients on these well-studied and established agents. We should take a leaf out of the IBDologists' book with the following strategies:

1. We believe that the majority of patients, unless contraindicated, should be started on AZA in conjunction with corticosteroids in order to avoid the use of long-term corticosteroids.

2. Approximately 3%–13% of patients with inflammatory bowel disease (IBD) and AIH will develop hepatotoxicity as a direct consequence of AZA.2–4 This hepatotoxicity may be difficult to distinguish from partial or non-response of AIH without a liver biopsy. Measurement of thiopurine metabolites, namely 6-thioguanine (6-TGN) and 6-methyl mercaptopurine (6-MMP), can be a valuable diagnostic guide. Elevated 6-MMP levels (>5700 pmol/8×10⁸) with low or normal 6-TGN (a phenomenon known as ‘shunting’) have been shown to be associated with hepatotoxicity and greater side-effects in IBD patients.3

3. Consideration should be given to the use of the xanthine oxidase inhibitor, allopurinol (100 mg once daily, with a 75% reduction of the AZA dose) to divert the enzymatic pathway preferentially towards increased 6-TGN production. In the last 12 months, three patients from our IBD cohort experienced deranged liver enzymes after starting AZA. Addition of allopurinol and AZA dose adjustment led to normalisation of the liver enzymes and continuation of AZA in all three patients. Ansari et al successfully re-treated 9 out of 11 IBD patients who suffered AZA-induced hepatotoxicity by allopurinol co-therapy.5

4. Recently at our unit, an AIH patient was successfully treated with allopurinol co-therapy. She had a persistently elevated alanine transaminase (ALT)>100 IU/l despite AZA adherence and high-dose corticosteroids for over 6 months. Elevated 6-MMP levels (6362 pmol/8×10⁸) with low 6-TGN levels (100 pmol/8×10⁸) were detected. The addition of allopurinol led to rapid normalisation of liver enzymes and discontinuation of corticosteroids within 12 weeks.

5. Use of thiopurine metabolites to guide AZA dosing for AIH appears to be unhelpful as no direct correlation between 6-TGN levels and response rates has been established.4 ,6

In conclusion, we would encourage perseverance with AZA in cases of non-severe adverse reactions. Though not universally available, judicious measurement of thiopurine metabolite levels in selected cases may help distinguish between drug hepatotoxicity, non-adherence or treatment failure. In patients found to have elevated 6-MMP levels, the addition of allopurinol with appropriate AZA dose reduction may correct hepatotoxicity and induce remission. This will reduce the number of repeat liver biopsies and avoid the need to change to other less well-established immunosuppressive agents.