Professor El-Omar has selected Professor Dermot Gleeson from the Liver Unit, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK, to do the next #GUTBlog.

The #GUTBlog focusses on the latest British Society of Gastroenterology (BSG) guideline entitled “British Society of Gastroenterology guidelines for diagnosis and management of autoimmune hepatitis” which was published in paper copy in GUT in September 2025.

Professor Gleeson writes:

“The 2011 BSG Autoimmune Hepatitis (AIH) guidelines were written by two Hepatologists and based on results of 12 randomised trials (RCTs) and of cohort studies from a few large centres. The 2025 guidelines are multi-author, multi-multidisciplinary, informed by a commissioned systematic literature review, moulded by many Microsoft Teams discussions and grounded by the active and persistent input of our two patient authors.

The continuing dearth of RCTs in AIH is stark (only 5 more published since 2011). However more robust data on AIH epidemiology, treatment and outcome have emerged from multicentre studies and from national registries. The guidelines are supported by a new meta-analysis of first-line treatments published in BMJ Open Gastroenterology in 2025 (1). We incorporate other meta-analyses, on mycophenolate and tacrolimus as second-line agents, and on AIH in pregnancy. We cite evidence-based consensus statements regarding histological diagnostic criteria, and early biochemical markers of treatment response which predict longer-term outcomes.

Diagnosis of AIH remains partly based on raised serum transaminases, absence of active hepatotropic viral markers, exclusion of biliary obstruction and usually, raised serum IgG (90%) and serum autoantibodies (80%). However, despite revised histological diagnostic criteria, with downgrading of rosettes and emperipolesis, liver biopsy is still recommended for confirmation, unless unsafe. Differentiation from drug-induced injury, Wilson’s disease and MASLD (metabolic dysfunction-associated steatotic liver disease) can be challenging.

Early treatment is summarised in guideline figure 2. Corticosteroids confer a survival advantage in most patients and are recommended in all but very mild cases. Prednisolone is preferred unless there is major concern about steroid-related cosmetic adverse effects, when budesonide is used. Initial prednisolone dose should not exceed 40 mg/day. Early addition of a steroid sparing agent (SSA), usually azathioprine, confers a survival advantage over steroids alone, with less adverse events. Despite a promising recent RCT, more information on mycophenolate is needed before its recommendation as a first-line SSA. However, in azathioprine-intolerant patients mycophenolate is a good option (conception must be avoided, given its teratogenicity).

In patients starting treatment, immunisations should be up to date (section F). In those with high clinical fracture risk (FRAX) scores, early institution of bisphosphonates can be considered pending DEXA scan results (Guideline figure 3).

Inadequate early treatment response is defined by consensus, as failure to normalise serum transaminases and IgG within 6 months (section G). Repeat biopsy to confirm histological remission in those with adequate early response is no longer routinely needed.

If treatment response is inadequate, options include optimisation of azathioprine metabolite levels, tacrolimus, infliximab and rituximab (section H) .

AIH is usually a life-long condition, needing long-term follow-up, including serial non-invasive fibrosis assessment. Historically, drug treatment has been continued for many years (sometimes for life). However, we have become more aware of its later adverse effects, especially the metabolic effects of steroids and a probable increased cancer incidence with azathioprine (section I).

Thus, recent consensus favours an attempt to phase out treatment: initially steroids following an adequate early response – and considering the possibility of adrenal insufficiency. And then after 2-3 years, the SSA – if enough criteria detailed in guideline table 12 are met.

Outcomes of pregnancy in AIH are usually good (section M), but complication rates are increased, requiring obstetric surveillance. Continuing steroids and azathioprine throughout pregnancy is associated with better foetal outcomes.

AIH can follow taking several drugs (section O, table 13); now termed Drug induced Autoimmune-Like Hepatitis (DI-AILH). Cessation of a potential drug precipitant of AIH should be routine. Otherwise, management is standard, apart from preferment of prednisolone alone, with addition of a SSA only if biopsy shows advanced fibrosis or there is relapse after stopping prednisolone.

A detailed account of AIH variant or “overlap” syndromes, with primary biliary (PBC) or sclerosing (PSC) cholangitis, is included. The guidelines end with proposals for clinical standards, service organisation and research priorities (boxes 3-5).

AIH was the first liver disease in which randomised trials (in the 1970s) showed medical treatment to be effective. Sadly, management has changed little since. Possible reasons include the competing goals of developing treatments for other commoner liver diseases. But also, a misperception that AIH had been effectively “cured”. More recently, we have realised that despite treatment, patients with AIH still develop de-novo cirrhosis have overall mortality rates about twice the population rates and have impaired quality of life. There is still an “unmet need” for better treatments, especially ones which are kinder to patients.
Fortunately, several novel agents are currently being evaluated. There is a collaborative research initiative called the UK AIH. With the BSG support, a one-day meeting to further stimulate collaborative activity is being organised for early 2026.”

References:

(1) Gleeson D, et al. What is the optimal first-line treatment of autoimmune hepatitis? A systematic review with meta-analysis of randomised trials and comparative cohort studies. BMJ Open Gastroenterol. 2025 Mar 28;12(1):e001549. doi: 10.1136/bmjgast-2024-001549.