Background and clinical presentation

AIH is a chronic liver disease presenting in all ages, both sexes and all races. Clinical presentation (summarised in table 2) ranges from asymptomatic to acute severe disease to fulminant liver failure. Approximately 25% of patients present acutely with AIH, and this can be indistinguishable from viral hepatitis. Thirty per cent of patients are cirrhotic at presentation including a significant number who are asymptomatic.

In patients with increased genetic susceptibility, environmental toxins such as drugs and virus can trigger AIH by molecular mimicry. Viral triggers include hepatitis A, cytomegalovirus and herpes simplex virus. Drug triggers include most commonly nitrofurantoin and minocycline.1 Biological therapy, for example, antitumour necrosis factor (TNF) monoclonal antibodies are also implicated to initiate AIH.

There is no single pathognomonic test for AIH. It remains a diagnosis of exclusion based on several indicative clinical, serological, biochemical and histological findings. Other aetiologies of liver diseases need exclusion. Clinically, AIH is characterised by elevated serum aminotransferase activity, autoantibodies and hypergammaglubulinaemia. Unless contraindicated, a liver biopsy is undertaken in all patients, ideally pretreatment to aid confirmation of diagnosis and staging. Histologically, AIH is characterised by interface hepatitis, dense plasma-rich infiltrates, hepatocyte rosetting and emperipolesis (figure 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Histological features of AIH. (A) H&E stained section showing portal inflammation and interface hepatitis. (B) Haematoxylin Van Gieson stained section (collagen fibres stained red). Interface hepatitis is associated with delicate fibrous entrapment of periportal hepatocytes, which show ballooning and rosetting. Images provided by Professor Stefan Hübscher, Consultant Histopathologist, Department of Cellular Pathology, Queen Elizabeth Hospital, Birmingham.

Autoantibodies, usually detected by indirect immunofluorescence, aid both in diagnosis and subclassification. Type 1 AIH is characterised by antinuclear antibody (ANA) and/or antismooth muscle antibody (anti-SMA). Neither antibody is specific for AIH as they can be present in patients with other liver and autoimmune diseases. Classically, ANA in AIH has a homogenous pattern; however, a course or finely speckled pattern can occasionally be seen.2 Up to 50% of patients with ‘definite’ primary biliary cholangitis (PBC) also test positive for disease-specific ANA, the characteristic immunofluorescence patterns being perinuclear/rim-like membranous (anti-gp210) and multinuclear dot (anti-Sp100). SMA antibodies can have three patterns: V (vessels), G (glomeruli) and T (tubules). The VG and VGT patterns, which correlate with F-actin antigenicity, are frequently associated with but not pathognomonic of AIH as compared with the V pattern.2 Type 2 AIH occurs mostly in children and is characterised by presence of antiliver/kidney microsomal type 1 (anti-LKM 1) and/or antiliver cytosol type 1 antibodies. Anti-LKM 1 antibodies are described mainly in the paediatric population in Europe and are seemingly rare in adults in the USA (only 3% of adults in North America3).

A simplified scoring system devised and validated by the international AIH group (2008) (table 3) based on two international multicentre patient cohorts4 had 88% sensitivity and 97% specificity (cut-off ≥6) and 81% sensitivity and 99% specificity (cut-off ≥7) when applied to the validation cohort (n=109). Recently it has been validated in Chile (n=294 patients) and found to be highly accurate for AIH diagnosis in the Chilean-Hispanic cohort.5 Although the scoring system is useful in AIH diagnosis, clinical judgement is required as neither autoantibodies nor histological features are absolutely specific to AIH and cannot be used in isolation to make a diagnosis.

Some patients with AIH can have features of primary sclerosing cholangitis (PSC-AIH) or PBC-AIH leading to several terms ‘overlap’, ‘variant’ or ‘cross- over’ syndrome. Such patients can have variable presentations. Some are identified at the time of diagnosis (simultaneous presentation), while others are diagnosed with one disease, which is stable for years before developing features of another disease (sequential presentation). Clinically variant syndrome is considered when a patient’s disease course deviates from that expected for the dominant underlying disease. A holistic approach is required to confidently diagnose overlap taking in to consideration a combination of biochemical, serological, histological and radiological (cholangiography) features. There is no single diagnostic feature. AIH, PSC and PBC are heterogeneous diseases with variations in their clinical features; for example, 8%–12% of patients with all the features of AIH can be persistently AMA positive,1 while interface hepatitis is common across the three diseases. These variations make the differential diagnosis between the diseases challenging and can lead to overdiagnosis of variant syndrome. Caution should be exercised when considering ‘overlap’ particularly in patients with a simultaneous presentation; often this is a reflection of disease variation rather than true overlap, and many would delay ‘diagnosing’ overlap until the dominant disease process is treated and the clinical picture rereviewed. Chazouillères et al 6 has developed the most commonly used criteria to aid in the classification of PBC-AIH. The diagnosis of PSC-AIH requires combination of cholangiographic or histological features typical of PSC, alongside robust histological features of AIH. Further discussion of variant syndrome is beyond the scope of this review.

Management

Management is based on immunosuppression and, in the presence of evident inflammation biochemically or histologically, is usually offered to all patients regardless of symptoms. Current treatment regimens are based on historic studies that showed untreated, moderate to severe AIH has a poor prognosis with a 5-year and 10-year survival of 50% and 10%, respectively.1 Such patients are offered treatment given clear survival benefit and excellent patient outcome with immunosuppression; 10-year survival rates exceed 80%.7 Older patients without histological inflammation or those with minimal changes and no fibrosis may present and be monitored but not immediately treated.8 The decision not to treat however remains controversial. Historically, this is because the benefits of treating asymptomatic patients with biopsy proven mild inflammatory activity are less well established, 10-year survival rates being 67%–90%. In contrast, a study showed the 10-year survival rates in untreated patients with mild disease was significantly lower than treated patients with severe disease (67% vs 98%, respectively).9 The disease course of mild AIH is unknown, while some do well without immunosuppression; it should be emphasised that untreated AIH has a fluctuating and unpredictable disease trajectory. Mild disease activity can be combined with episodes of severe activity and a substantial proportion of asymptomatic patients may become symptomatic (26%–70%) during follow-up; therefore, risk of disease progression leading most to treat from the outset.10 11 Furthermore, papers tend to quote 10-year survival because that is the data to hand, but a patient perspective requires longer survival data. Clinicians need to carefully balance treatment-related side effects versus risk of disease and symptom progression. Patients not on treatment need close monitoring.

The overarching treatment goal is to achieve complete biochemical (normalisation of serum aminotransaminase activity and immunoglobulin G (IgG) concentration) and histological remission. Histological remission lags behind biochemical remission by 3–6 months. Remission is achievable in 75%–80% of patients after 24 months of treatment. In some, it may not be possible to achieve complete remission, the goal thereby being to achieve lowest biochemical activity with minimum treatment-related side effects.1

Induction and maintenance remission of AIH

Prednisolone±azathioprine is used to induce and maintain remission. Two strategies are proposed to induce remission: (1) prednisolone monotherapy and (2) prednisolone/azathioprine combination therapy. Combination therapy is reportedly as efficacious as prednisolone monotherapy with fewer side effects (10% vs 44%, respectively), making it the preferred treatment strategy.

Although guidelines for the management of AIH exist,1 7 the optimum prednisolone dosing strategy remains poorly defined, as no trials have directly compared prednisolone dosing. Treatment decisions should not be static but instead be pragmatic and personalised to a patient’s age, disease presentation, comorbidities, as well as preference, so as to achieve the best long-term outcomes. Starting prednisolone doses range from 20 mg/day up to 1 mg/kg/day. Higher prednisolone doses (e.g. 1 mg/kg/day) are however rarely required and are associated with side effects. The addition of azathioprine allows for use of lower prednisolone doses (e.g. 20–30 mg/day) to induce remission. Azathioprine can be commenced at the outset (50 mg/day) or 2–4 weeks after corticosteroid introduction at 1–2 mg/kg/day.1 The European Association for the Study of the Liver (EASL) guidelines recommend prednisolone as initial induction therapy followed by addition of azathioprine after 2 weeks of prednisolone initiation.1 Delay in azathioprine introduction can be pragmatically helpful in diagnostic uncertainty (corticosteroid responsiveness aids in diagnosis confirmation). It also avoids the diagnostic dilemma of differentiating azathioprine-induced hepatotoxicity from primary non-response1 and other general side effects, for example, nausea. In icteric patients, azathioprine initiation is usually delayed until the bilirubin is <100 μmol/L.1 12

In most, a fall in serum aminotransferase activity is seen within 2 weeks of treatment initiation. The prednisolone dose is gradually tapered to the minimum required to maintain normal liver biochemistry and IgG concentration (refer to figure 2 for a proposed treatment algorithm). Dose titrations need to be individualised (considering patient age, comorbidities and severity of presentation) and response guided (falls in alanine aminotransferase (ALT) and IgG). In young patients with aggressive disease, some initiate prednisolone at 30–40 mg/day, while others use higher doses, for example, 1 mg/kg/day. A high and fast response rate was shown to be associated with better outcomes.13 As always, treatment-related side effects must be balanced with benefit.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Proposed treatment algorithm for AIH based on the authors experience. Although guidelines exist for treating AIH, an individualised and response-guided approach to patient management is required. Adherence to therapy is as important as the clinical management of a patient with AIH. Benefits of therapy should be balanced against the risks of potent immunosuppressants. Specialist opinion is advisable: if a patient does not respond well to first-line therapy, there is uncertainty in the management of an AIH patient and in acute severe AIH. A liver biopsy can guide treatment decisions; however, it is rarely needed if a patient relapses or flares as elevated biochemistry (ALT and IgG) indices are indications to treat/escalate therapy. AIH, autoimmune hepatitis; ALT, alanine aminotransferase; AST, aspartate aminotransferase; IgG, immunoglobulin G concentration; LC 1, liver cytosol type 1; LKM 1, liver kidney microsomal type 1 antibody; MMF, mycophenolate mofetil; SLA/LP, soluble liver antigen/liver-pancreas antibody; ULN, upper limit of normal.

Remission is usually maintained using prednisolone (5–12.5 mg/day) alongside azathioprine (1–2 mg/kg/day) or azathioprine monotherapy (up to 2 mg/kg/day). Both strategies are similar in efficacy and superior to prednisolone monotherapy in maintaining remission.14 Studies have shown azathioprine monotherapy (2 mg/kg/day) prevented corticosteroid specific side effects and relapse on corticosteroid withdrawal as well as achieved remission rates of up to 83%.15 16 Careful monitoring is required with azathioprine given risk of cytopenias and doses adjusted accordingly. Serum thiopurine metabolite levels can be helpful optimising dosing with thioguanine nucleotide concentrations >220 pmol/8×10⁸ erythrocytes predictive of remission.17 Therapy needs to be tailored to an individual patient, the most appropriate maintenance regimen will depend on both the severity of therapy related side effects and disease activity. In our practice, we rarely start with a goal of 2 mg/kg/day of azathioprine because of the long-term side effect profile, but in individual patients titrate azathioprine dose as needed for that individual, thereby balancing risk and benefit.

The optimum treatment duration is controversial, but for non-cirrhotic type 1 AIH patients, corticosteroids are usually continued for 12–18 months and azathioprine for 3–5 years. Therapy is continued for minimum 24 months after complete biochemical remission prior to withdrawal. A single attempt at treatment withdrawal is reasonable in non-cirrhotic patients assuming all inflammatory parameters remain repeatedly normal. Approximately 50% of patients in complete biochemical remission have persisting histological activity18; therefore, a confirmatory liver biopsy prior to withdrawal is considered helpful by some. Treatment withdrawal is undertaken by stepwise reduction with patients monitored throughout.

Relapse occurs in 50%–90% of patients following treatment withdrawal and is more likely in the first year. Late relapses also occur; 81% of patients in a study relapsed after 3 years.19 If relapses or flares occur, corticosteroids are reinstituted at induction doses (eg, 20–30 mg/day), and azathioprine can be increased to 2 mg/kg/day.

Budesonide as an alternative induction agent to prednisolone

EASL recommends that budesonide 9 mg/day (a corticosteroid with a short half-life and low systemic exposure) alongside azathioprine can be used to induce remission in treatment-naïve, non-cirrhotic patients with uncomplicated AIH when there is concern regarding side effects and comorbidities potentially exacerbated by prednisolone, for example, diabetes and osteoporosis.

Budesonide alongside azathioprine effectively induced and maintained remission in non-cirrhotic AIH patients with low corticosteroid-specific side effect rates compared with prednisone20; however, another study found budesonide was ineffective in corticosteroid refractory/dependent AIH.21 Switching from prednisolone to budesonide can be considered to reduce side effects in patients on long-term therapy; care is needed as occasionally this may result in corticosteroid withdrawal symptoms. Given budesonide has a 90% first-pass hepatic clearance, it is contraindicated in patients with cirrhosis or those with portal hypertension/perihepatic shunting. Concurrent extrahepatic immune-mediated diseases need consideration as exacerbations are reported with budesonide, in keeping with the first-pass hepatic metabolism.

Second-line therapy

Second-line therapy may be considered in incomplete or non-responders (persistence of biochemical and/or histological inflammatory activity) or those intolerant to standard therapy (10%–20% of patients). For incomplete or non- responders, adherence should be confirmed, and alternative diagnosis (particularly hepatitis E infection) as well as variant syndrome should be excluded. Active disease should be excluded in those with concurrent inflammatory bowel disease as such patients have been shown to enter biochemical remission less often as well as fail treatment more commonly and progress to cirrhosis.22 Following this, standard therapy should be optimised prior to second-line therapy. A liver biopsy may guide treatment decisions. In patients requiring high dose long-term corticosteroid therapy, azathioprine can be increased to 2 mg/kg/day alongside prednisolone, while those on budesonide switched to prednisolone. Use of second-line agents is not standardised; there is variability in clinical practice and efficacy. Expert advice is advisable when second-line therapy is being considered or when there is uncertainty in the management of a patient with active AIH.

Mycophenolate mofetil (MMF 1–2 g daily) is the most commonly used second line therapeutic agent in AIH, largely reserved for azathioprine/6-mercaptopurine intolerant patients. Remission rates when MMF is used in azathioprine-intolerant patients range 43%–88%. Previous studies where MMF was used in azathioprine non-responders are disappointing (remission rates between 0% and 25%). However, recent studies have reported higher remission rates. A retrospective multicentre study (n=201 patients with AIH) found that both tacrolimus and MMF as second-line therapies were equally effective in maintaining biochemical remission (91.9% vs 94.1%, respectively) in patients who were previous complete responders but intolerant to standard therapy.23 However, in non-responders, tacrolimus led to a complete response in a significantly greater proportion of patients as compared with MMF (56.5% vs 34%, respectively). In a second retrospective study of 105 patients with AIH, Roberts et al found no differences in the proportion of patients who achieved biochemical remission within the first 2 years of treatment between patients receiving MMF for non-response and those who were intolerant to standard therapy (57% vs 62%, respectively).24 Reduced biochemical remission rates and higher infection rates were however reported in cirrhotics receiving MMF as compared with non-cirrhotics. Prospective studies are required to better assess the role of MMF as second-line therapy in patients with inadequate response to standard therapy. High rates of intolerance to MMF are reported. It is contraindicated for those contemplating pregnancy making other second-line therapies preferable. A Greek group explored using MMF as first-line therapy to induce AIH remission.25 26 In their observational study (n=131), initial biochemical response was achieved in 93.6% (prednisolone +MMF) versus 72.7% (conventional therapy) of patients within 2 months.25 Further studies are required before MMF can be advocated as first-line therapy.

Calcineurin inhibitors: evidence of ciclosporin’s use in adults with non-responsive AIH is sparse; remission rates of 80% are reported.1

Tacrolimus as salvage therapy in difficult-to-treat AIH improved liver biochemistry, histology and enabled corticosteroid dose reduction.27 Target serum tacrolimus concentrations in AIH are unknown but most maintain a concentration of ≤6 µg/L.27–29 Attention to the side effects is important, including hypertension and diabetes.

Other second-line agents include: 6-mercaptopurine (6-MP), cyclophosphamide, anti-TNF agents and rituximab. Single-centre experience, for example, in treating refractory AIH patients with rituximab has produced encouraging results with significant biochemical improvements seen following rituximab therapy.30

Surveillance considerations and side effect/monitoring of therapy are discussed in tables 4 and 5 respectively

Liver transplantation

Liver transplantation should be considered in (1) acute or fulminant liver failure, (2) patients with a Model for End-Stage Liver Disease (MELD) score of ≥15, (3) decompensated liver disease while on treatment and (4) rarely hepatocellular carcinoma development. Untreated patients with AIH have a survival of <30%.7 Patients, who present with acute or fulminant liver failure, should be referred urgently to a transplant centre31 as emergency transplantation may be required. Overall mortality in such patients ranges between 19% and 45% and rate of liver transplant required range from 9% to 81%.1 Corticosteroid therapy in an acute setting can be complicated by an increased infection risk; however, a trial for all but the sickest patients may be considered following advice from a transplant centre; the appropriate dose in such settings is not however agreed. Patients need close monitoring for deterioration and sepsis. Failure to improve prognostic scores within 7 days should prompt early consideration of liver transplantation.1 Outcomes post-transplantation are good with 5-year and 10-year survival rates of approximately 75%.7 Post-transplant immunosuppression needs to be broadly more intense than other indications, with use of either long-term prednisolone or azathioprine/MMF alongside a calcineurin inhibitor. Recurrence rates at 5 years range 20%–50% of transplants.

De novo AIH: this is when an AIH or AIH-like syndrome develops after transplantation for other liver diseases. The prevailing view is that de novo AIH represents a variant of rejection. It is managed by addition of corticosteroids to a patient’s immunosuppression regime. In the case of inadequate response following corticosteroid introduction and optimisation of calcineurin inhibitor concentration, azathioprine (1–2 mg/kg/day) is instituted.7 Usually good response is seen following addition of corticosteroids and azathioprine; however, in some progression to cirrhosis and subsequent graft failure can occur.

Case examples

Case 3

Ms C (aged 27 years) has rheumatoid arthritis (RA) and type 1 ANA positive (1:40) AIH (in biochemical remission). In March 2014, she started abatacept injections (CTLA4-Ig protein that blocks T cell activation) for active RA (serum liver tests/IgG were normal at that point). Serum liver tests in September 2015: ALT 859 U/L, ALP 197 U/L, bilirubin 49 µmol/L, albumin 38 g/L, INR 1.2 and IgG 5.76 g/L. Medications then included prednisolone 5 mg/day, azathioprine 1 mg/kg/day and abatacept injections. Imaging: unremarkable. The possibility of drug-induced liver injury was considered; therefore, the abatacept injections were stopped. Liver biopsy revealed an acute hepatitic process. She was treated for an AIH flare; prednisolone increased to 40 mg/day (figure 3). Despite this, serum liver tests deteriorated further: ALT 619 U/L, AST 700 U/L, ALP 243 U/L, bilirubin 163 µmol/L, albumin 36 g/L, INR 1.6, and IgG 6.84 g/L requiring a hospital admission by December 2015. Repeat liver biopsy revealed a severe acute AIH. High dose corticosteroids (intravenous methyl prednisolone 80 mg daily for 3 days then oral prednisolone 80 mg/day for further 3 days then 40 mg/day of prednisolone as maintenance), alongside salvage therapy with rituximab infusion 1000 mg, prophylactic antibiotics and antifungals were administered. Serum liver tests improved by mid-January 2016: ALT 86 U/L, AST 55 U/L, ALP 148 U/L, bilirubin 54 µmol/L, albumin 40 g/L, INR 1.2 and IgG 5.61 g/L. Prednisolone was weaned down to 10 mg/day, and azathioprine increased to 2 mg/kg/day. Six monthly rituximab infusions were continued. Serum liver tests have subsequently normalised for more than a year.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Course of serum alanine transaminase activity from March 2014 to November 2017. Ms C was maintained on prednisolone (5–7.5 mg/daily) and azathioprine (1 mg/kg/day) throughout her disease course unless otherwise indicated. Maintenance rituximab infusions have been continued until present. ALT, alanine aminotransferase; PO, per mouth.

This case highlights (1) biological therapy may precipitate an AIH flare. (2) In acute severe AIH, following guidance from a transplant centre, a single trial of high dose (e.g. ≥1 mg/kg/day sometimes intravenously) corticosteroids to all patients apart from the sickest is reasonable. Lack of improvement in prognostic scores within 7 days should prompt early consideration of transplantation. (3) Similar to other case series, rituximab was efficacious as salvage maintenance therapy in our patient with treatment refractory AIH. Significant biochemical and histological improvements in treatment refractory AIH patients are reported following rituximab therapy. Given limited studies/experience, rituximab in AIH should preferably be initiated in a specialist centre.

Speciality certificate exam (SCE) questions

1. A 60 year-old woman presents with fatigue and joint pains for 6 weeks. She has a medical history of type 2 diabetes and hypertension. Her body mass index (BMI) is 36 mg/kg². Abdominal examination was unremarkable. Her investigations were as follows:-

Hb 12 g/L, WCC 4.0×10*9/L, platelets 250×10*9/L INR 1.1, sodium 140 mmol/L, urea 3.1 mmol/L, creatinine 43 μmol/L, Alb 32 g/L, ALT 294 U/L, AST 230 U/L, ALP 178 U/L, bilirubin 11 μmol/L, ANA 1:1600 and IgG 20 g/L. Rest of the liver screen including viral screen was negative.

Liver biopsy: lymphoplasmacytic inflammation with moderate interface activity and no fibrosis.

Which one of the following treatment regimens would you initiate for her?

1. Azathioprine 50 mg daily.

2. Prednisolone 20 mg daily.

3. Prednisolone 10 mg daily and azathioprine 50 mg daily.

4. Budesonide 9 mg daily and azathioprine 50 mg daily.

5. Tacrolimus 1 mg twice daily.