Stopping Immunosuppressive Treatment in Autoimmune Hepatitis (AIH): Is It Justified (and in Whom and When)?

Laura Harrison; Dermot Gleeson


Liver International. 2019;39(4):610-620. 

In This Article

Abstract and Introduction


Background: Initial treatment of autoimmune hepatitis (AIH) with prednisolone ± azathioprine is based on randomised controlled trials. Many patients receive long-term immunosuppressive treatment to prevent disease relapse; this strategy has a weaker evidence base.

Aim: To consider whether immunosuppressive treatment (IST) withdrawal in AIH is justified and to develop a rationale for patient selection.

Methods: We reviewed published papers between 1972 and 2018, which addressed the outcomes of IST withdrawal and/or complications of IST in AIH.

Results: (1) AIH relapse rates after withdrawal of IST vary between 25% and 100%. There is heterogeneity in these studies regarding relapse definition, IST duration prior to withdrawal and criteria for biochemical and histological remission prior to withdrawal. (2) Factors associated with relapse following IST withdrawal include: (a) absence of an identifiable initial disease trigger, (b) presence of other autoimmune diseases, (c) longer time to biochemical remission and (d) elevated serum transaminases on treatment withdrawal. Reports of associations between relapse and age, IST duration and failure of histological remission have been inconsistent. (3) Continued IST reduces risk of AIH relapse over at least 5 years. However, there is no evidence that routine (as opposed to selective) long-term IST improves disease outcome. (4) Patients with AIH have an increased risk of extrahepatic cancer, notably non-melanoma skin cancer, to which long-term IST may contribute. Long-term corticosteroid therapy is associated with weight gain, low-trauma fractures, diabetes and possibly vascular disease.

Conclusions: While further studies are needed, evidence supports a strategy of IST withdrawal in some patients with AIH who have achieved remission.


Autoimmune hepatitis (AIH) is a chronic noninfectious liver disease affecting about 24 per 100 000 people in Western Europe.[1] AIH is a heterogeneous condition which can affect all ages and races. Eighty per cent of patients are female. Its mode of presentation ranges from acute liver failure to asymptomatic liver test abnormalities. There is no single reliable diagnostic test; therefore the diagnosis is based on a synthesis of clinical, demographic, biochemical, immunological and histological information. The process has been formalised by development of diagnostic scores developed by the International Autoimmune Hepatitis Group (IAIHG).[2,3]

There are many unanswered questions regarding the management of AIH. Only 12 randomised control trials have been published.[4,5] Most of these trials have focussed on the initial treatment of AIH and have clearly shown that immunosuppressive treatment reduces 5-year mortality. Based on these studies, initial therapy of AIH is steroid-based (usually prednisolone). Budesonide may be as effective in the short term,[5] but long-term studies are needed to support its role in AIH treatment).

Usually a second drug is added, often termed a steroid-sparing agent, the benefit of which is to enable similar short-term efficacy to higher doses of steroid monotherapy. The most often used standard steroid-sparing agent is azathioprine, with mycophenolate (MMF) or 6-mercaptopurine substituted in azathioprine intolerant patients.[6] There is currently insufficient evidence for the use of alternative drugs to azathioprine, such as calcineurin inhibitors or biological agents, as first-line agents.

After 2 years of treatment, biochemical remission (defined as normalisation of both transaminase levels and IgG levels)[7] is achieved in 80%-90% patients. Failure to achieve biochemical remission is associated with a poor outcome.[8–10] Histological remission is achieved in a lower percentage of patients: 55%-70%.[11–14]

Notwithstanding this high initial remission rate, long-term management of AIH remains suboptimal. Despite treatment, de novo cirrhosis develops in 18 (6–54) (median (range))% patients[15–23] and also premature death with standardised mortality ratios of 2.0-4.0.[1,15,24,25] In part, this is because AIH frequently relapses after stopping treatment and sometimes despite continuing treatment. Multiple relapses are associated with fibrosis progression and a poor outcome.[15,19,26] However, even independently of clinical and biochemical relapse, fibrosis may still progress and cirrhosis develop. This is probably because of incomplete suppression of liver inflammation. Even in patients achieving biochemical remission, failure to obtain histological remission was associated in one study with failure of fibrosis to regress and with poorer long-term survival.[12]

One important question relates to how long should immunosuppressive treatment (IST) of AIH be continued. Recent management guidelines have not resulted in a consensus. The recent EASL guidelines recommend consideration of IST withdrawal[27] after at least 3 years of treatment and at least 24 months after biochemical remission is attained. The BSG Guidelines recommend an "individualised" approach to treatment. Demonstration of biochemical and histological remission and treatment duration of at least 24 months are the minimal criteria set out by the AASLD guidelines for consideration of IST withdrawal.[28]

In this review, we explore the available published evidence on withdrawal of IST in AIH in an attempt to address:

  1. The relative benefits and risks of withdrawing vs its long-term continuation

  2. Which patients are suitable for IS withdrawal

  3. When is the optimal time to stop treatment?