EASL Clinical Practice Guidelines

Treatment of AIH

The aim of treatment in AIH is to obtain complete remission of the disease and to prevent further progression of liver disease. This requires mostly permanent maintenance therapy, or (only achievable in a minority of patients) induction of a sustained remission following treatment withdrawal.

The current immunosuppressive treatment regimens are based on studies that were mostly published in the 1970s and 80s [[175], [176], [177], [178], [179]]. These studies revealed that untreated, moderate to severe AIH (confluent necrosis on liver biopsy, AST levels >5 times the ULN, γ-globulin levels >2 × ULN) had a very poor prognosis and have demonstrated that immunosuppressive therapy improves liver functions tests, ameliorates symptoms and prolongs survival. As these studies were performed before discovery of the HCV they are likely to have included patients with viral hepatitis C.

The benefits of immunosuppressive treatment in asymptomatic older patients with mild necroinflammatory activity on liver biopsy are not well established in terms of clinical endpoints and their management remains controversial (Fig. 3). Treatment related side effects should be counterbalanced to the risk of subclinical disease progression and evolution into symptomatic disease as well as the prospect of a complete and sustained response to treatment. Ten-year survival in untreated patients with mild disease was reported to be 67–90% [[180], [181]], and in an uncontrolled study untreated asymptomatic patients had similar survival to those receiving immunosuppression [44]. Thus, a decision not to treat might be justified, especially if there are relative contraindications to the use of steroids. In addition, spontaneous resolution of AIH may occur [182]. However, it also has to be acknowledged that untreated AIH has a fluctuating, unpredictable disease behaviour and a substantial proportion of asymptomatic patients become symptomatic during the course of their disease follow-up [[44], [183]], and progression towards end-stage liver disease with liver failure and development of HCC is possible [181]. In addition, as AIH is a lifelong disease, and progressive fibrosis may take many years to become clinically apparent, the observational studies published may have been too short and may have included too few patients in order to demonstrate the benefit of immunosuppressive therapy in milder disease. The fluctuating course of disease and the danger of subclinical disease progression make it of critical importance, that if patients with mild disease are left untreated, they must nonetheless undergo a regular subsequent monitoring including follow-up liver biopsy, if ALT and/or IgG levels increase or fluctuate (Fig. 3). In symptomatic patients and patients with advanced fibrosis or cirrhosis, treatment should always be initiated as this represents a negative prognostic predictor [[44], [104], [160], [184]]. In addition, even in advanced fibrosis and cirrhosis substantial regression of scarring after successful treatment has been reported. In view of the progressive nature of AIH and the effectiveness of immunosuppressive therapy, the consensus group recommends that all patients with active disease should receive treatment.

Fig. 3
Therapeutic algorithm with case-by-case decisions about commencing steroid therapy, informed by baseline assessments. For example, a patient with active disease (elevated transaminases >3 normal values and hepatitis activity index (HAI) >4/18) requires treatment. *Treatment probably no longer indicated in decompensated, burn-out cirrhosis, unless high inflammatory score on liver biopsy.

View Large Image | View Hi-Res Image | Download PowerPoint Slide

Remission induction (Fig. 4)

The survival benefit of corticosteroid therapy with or without azathioprine has been shown in a number of controlled trials performed in the 1960s and 1970s [[175], [176], [177], [178], [179]]. In 1971, Cook et al. demonstrated survival benefits of immunosuppressive therapy comparing prednisolone monotherapy (15 mg/day) with placebo. Mortality rates differed markedly (14% vs. 56%) [175]. The results were confirmed in a study from the Mayo Clinic one year later. This study compared prednisone monotherapy (starting with 60 mg/day and reduced to 20 mg over four weeks), azathioprine monotherapy (100 mg/day), combination therapy (prednisone starting at 30 mg/day reducing to 10 mg/day maintenance combined with azathioprine at 50 mg/day) and placebo, [176] the beneficial effect on survival was similar with prednisone monotherapy and prednisone/azathioprine combination therapy (mortality rate: 6% vs. 7% vs. 41% in the placebo group) [179]. However, the combination regimen was associated with fewer side effects (10% vs. 44%) [185]. Histological remission was achieved in 75% of patients after 18 months of active prednisone based treatment but lagged behind clinical and biochemical remission by several months. This trial further illustrated that azathioprine monotherapy, used as induction therapy, resulted in a high mortality (36%). Murray-Lyon et al. also reported a higher mortality rate of azathioprine monotherapy as induction therapy, compared to prednisone (24% vs. 5%) [177].

Fig. 4
Therapeutic strategy in autoimmune hepatitis. Treatment requires induction of remission and prolonged maintenance therapy. Induction is delivered by steroids and thiopurines are added as steroid sparing strategy. Laboratory endpoints are normalisation of IgG and ALAT. MMF, mycophenolate mofetil.

View Large Image | View Hi-Res Image | Download PowerPoint Slide

Comparing another strategy with titrated doses of prednisone (starting also at 60 mg/day) to maintain serum transaminase activity at less than twice the upper limit of normal, Summerskill et al. showed that this led to less severe side effects compared with the fixed dose regimen (starting with 60 mg/day and reducing to 20 mg over four weeks) with comparable effect on clinical symptoms and biochemical parameters. However, histological remission was achieved in only 25% and 30% after 24 and 36 weeks respectively [186]. In 1982 Tage-Jensen et al. demonstrated once more superiority of prednisone monotherapy over azathioprine monotherapy in inducing remission [187].

Despite the limitations of these early studies (no testing for HCV available at that time), they provide compelling evidence that the use of a prednis(ol)one/azathioprine combination regimen has the best profile in combining high efficacy with minimal side effects. Although prednisolone monotherapy and prednisolone/azathioprine combination therapy is considered equally effective [34], frontline combination therapy with use of azathioprine may be preferable, particularly in patients with a high anticipation of side effects such as in post-menopausal women, individuals with emotional instability, pre-existent osteoporosis, brittle diabetes, labile hypertension or obesity. Similarly, young female patients are often concerned about weight gain and cosmetic side effects due to steroid treatment which might adversely affect adherence and outcome. A pragmatic approach to this issue is needed to ensure the best long-term outcomes. A proposed dosing regimen is shown in Table 7. Caution for the use of azathioprine is advisable in patients with malignancy, cytopenia and established thiopurine methyltransferase (TPMT) deficiency (see below) as well as pregnancy, and in these situations, a risk-benefit analysis should be undertaken at an individual patient level (see below).

Table 7
Treatment proposal for adult patients with AIH (e.g. 60 kg).

Reduction of prednisolone to 7.5 mg/day if aminotransferases reach normal levels and after three-months to 5 mg/day, tapering out at three-four months intervals depending on patient’s risk factors and response. *Azathioprine dose of 1–2 mg/kg according to body weight.

Additional to the classical prednisolone/azathioprine regimen that are also recommended in the AASLD [34] and BSG guidelines, [40] several slight modifications have been proposed and are being used in clinical practice in many expert centres. A higher dose of predniso(lo)ne (up to 1 mg/kg/day) in combination with azathioprine was shown to result in a more rapid normalisation of serum transaminases in patients without cirrhosis [188]. Although the absence of an early fall of transaminases [[185], [189]] and failure to normalise [[190], [191], [192]] are negative predictors of treatment success, this strategy seems promising. It needs to be confirmed that this translates into better long-term outcome and is really suitable for all patients.

Another reasonable strategy is to delay institution of azathioprine and start with prednisone monotherapy. Delaying introduction of azathioprine (usually by about two weeks) can be pragmatically helpful in managing patients with AIH, as it may on the one hand help to resolve diagnostic uncertainties while on the other hand avoids the diagnostic dilemma of discrimination between azathioprine-induced hepatotoxicity from primary non-response. Although hepatotoxicity of azathioprine is uncommon, its frequency is increased in patients with advanced liver disease [3]. In general, treatment of AIH should be response guided and treatment regimens should be individualised according to the response of the patient and tolerance of treatment. A suggested induction strategy is shown in Fig. 4.

Therapy is aimed to obtain complete biochemical and histological remission. In patients showing a prompt response with complete normalisation of transaminases and IgG within the normal limits of a follow-up, liver biopsy to demonstrate also histological remission is normally not required, as chances of significant inflammatory activity requiring increased immunosuppression are very low. Follow-up biopsy is, like any invasive procedure, recommended, if a change of management is somewhat likely to result from the procedure; this is particularly the case in patients with sub-optimal response to immunosuppression, and in patients with treatment side effects. In these cases, individual risk assessment of disease progression needs to be weighed against (possible) treatment side effects, and assessment of disease grading can help in this. Ongoing studies suggest that Fibroscan can also be used in follow-up. Increase in liver stiffness can be due to either disease reactivation with increased inflammatory infiltrates and oedema, or due to progressive fibrosis (or both).

In a recent prospective, double blind, randomised, phase IIb trial of patients without cirrhosis [193], budesonide (9 mg/day) and azathioprine (1–2 mg/kg/day) given for six months was compared to conventional prednisone and azathioprine-based immunosuppression. Budesonide/azathioprine was shown to normalise transaminases more commonly and had fewer side effects compared to prednisone (40 mg/day tapered to 10 mg/day) azathioprine (1–2 mg/kg/day) combination, and was superior in a combined endpoint. A complete biochemical remission without steroid specific side effects was achieved in 47% of patients given budesonide vs. 18.4% given prednisolone. However, follow-up data on histology and long-term data are not available. Remission rates in the control arm in this trial were very low, and clearly lower than in earlier published case series, presumably due to fixed dose reduction schedule in the prednisone group and a fairly low used prednisone dose. While budesonide in this trial was given at a high dose until response was observed and treatment was response guided, the control arm was not, introducing a bias in the trial. Nonetheless, this trial has demonstrated efficacy of budesonide in AIH. The successful use of budesonide in AIH has also been reported in other small case series [[194], [195], [196]], but failure was also described in other series [197].

The decision to use budesonide in a particular patient should balance the anticipated beneficial side effect profile against uncertainties in strength and long-term efficacy. Budesonide has a 90% first pass hepatic clearance and should not be used in cirrhotic patients, or those with peri-hepatic shunting, because of the high risk of side effects in patients not protected by effective first pass metabolism [[198], [199]]. In addition, the decision should be influenced by the presence of concurrent extra-hepatic immune-mediated diseases as exacerbations of these might be possible [[197], [200]]. Budesonide (9 mg/day) plus azathioprine may therefore be considered in treatment-naive non-cirrhotic patients with early stage disease and in whom untoward major steroid specific side effects are highly anticipated. However, we have little information about the best approach to dose reduction in budesonide treatment. Due to the short half-life of budesonide, it is not clear if reduction to twice daily (6 mg) and to once daily (3 mg) dosing is reasonable, or if the thrice daily dosing should be continued, reducing individual dose from 3 mg to 2 mg and 1 mg.

A complete clinical, biochemical and histological remission with a sustained off-treatment response after treatment withdrawal is the most desirable treatment endpoint. However, this cannot be reached in the majority of patients. In 80 to 90% of patients, transaminases promptly improve after institution of immunosuppressive treatment. In approximately 20% of patients, a sustained remission following withdrawal of immunosuppressive treatment can be achieved by finite treatment therapy (with a median follow-up of more than six years) [201], which might be improved by the application of stringent endpoint criteria before treatment withdrawal (see below). Primary non-response to immunosuppressive treatment is experienced in only a small proportion of patients with AIH. Non-response (or very slow response) should therefore always lead to a careful reconsideration of the diagnosis and/or re-evaluation of adherence to treatment. In particular, young, non-Caucasian patients with an acute or fulminant presentation and the finding of confluent necrosis in liver tissue have a higher risk of treatment failure [[20], [202], [203], [204], [205]]. Patients with liver failure and lack of improvement of serum bilirubin and MELD score during treatment should be referred early to a transplant centre and LT should be considered since this disease phenotype has a high mortality without LT [[42], [202], [203], [204], [205], [206]]. In patients without liver failure and not responding to initial treatment an increased dose regimen or alternative treatment strategies can be applied (as discussed in detail in section ‘Special patient populations’) [207].

Treatment withdrawal (Fig. 5)

The majority of AIH patients respond well to steroid based immunosuppressive treatment and serum transaminases improve to levels within the normal range [[190], [208], [209], [210]]. Complete normalisation of transaminases as well as normalisation of IgG levels should be the aim of treatment in patients with autoimmune hepatitis as persisting elevations of transaminases are predictive of; (i) a relapse after treatment withdrawal; (ii) activity on liver biopsy; (iii) progression to cirrhosis; and (iv) poor outcome [[185], [190], [191], [192], [208], [211], [212]].

Fig. 5
Follow-up of autoimmune hepatitis patients who have achieved remission. Note that drug-free remission of autoimmune hepatitis is infrequent and cannot be achieved in the majority of patients.

View Large Image | View Hi-Res Image | Download PowerPoint Slide

Histological resolution of disease typically lags behind after reaching the biochemical endpoint [176]. There is no clear evidence of optimal treatment duration. However, treatment should be continued long enough to reach histological remission as residual interface hepatitis is still found in patients with normalised ALT levels and is predictive of disease relapse [212]. Together with normalised serum transaminases normalisation of serum IgG appears to also be predictive of histological resolution [108].

Treatment should be continued for at least three years and for at least 24 months after complete normalisation of serum transaminases and IgG levels (biochemical remission). Longer treatments may decrease the frequency of relapse and may therefore be considered. For patients with severe initial presentation and low tolerance of induction treatment, performance of a liver biopsy prior to treatment withdrawal is advisable as histological findings are predictive of fibrosis progression and relapse [108]. In patients with continued histological disease activity (HAI >3), immunosuppressive treatment should not be discontinued, as relapse is almost certain to occur. A recent paper showed that ALT levels below half the upper limit of normal together with IgG levels below 12 g/L were highly predictive for successful treatment withdrawal [213]. A trial of treatment withdrawal should be undertaken by stepwise reduction of immunosuppressive agents, and patients monitored closely. Flares of AIH activity during maintenance therapy or following treatment reduction require increased doses of immunosuppression and preclude complete drug withdrawal (Fig. 5).

A relapse of the disease is frequent (50–90%) after drug withdrawal and typically occurs in the first 12 months after stopping treatment [[201], [214], [215], [216]]. However, later relapse can also occur and underscores the need for regular lifelong monitoring of patients even without immunosuppressive therapy [217]. A relapse is defined as reappearance of ALT elevation >3 times the ULN according to the IAIHG criteria, but may also present with milder ALT elevations and/or increase in IgG levels. A liver biopsy is usually not necessary to confirm the relapse as ALT elevations are highly predictive. A higher frequency of relapse has been reported to be associated with; (i) slow response to immunosuppressive treatment; (ii) persistent elevation of serum transaminases and/or serum globulins and IgG; (iii) residual inflammation on liver biopsy; and (iv) shorter treatment duration [[191], [209], [211], [218], [219]]. In patients with an identifiable (and avoidable) trigger of initial presentation, a relapse seems to be more uncommon [74].

Treatment of the relapse corresponds to initial treatment with prednis(ol)one and azathioprine and is equally effective in inducing a remission as in primary induction therapy. However, close monitoring after treatment withdrawal with early detection of relapse allows lower doses of immunosuppressants to re-induce full remission. Importantly, patients with multiple relapses were shown to experience more side effects and have adverse outcomes [[201], [208], [216], [220]]. Therefore, long-term, probably permanent, maintenance treatment is advisable in patients after a relapse.

Maintenance treatment

After reinstitution of therapy and achievement of clinical and laboratory resolution, the dose of azathioprine should be increased to 2 mg/kg/day as the dose of predniso(lo)ne is gradually withdrawn. Azathioprine should subsequently be continued indefinitely as a continuous maintenance therapy. The data regarding maintenance therapy with azathioprine monotherapy at a dosage of 2 mg/kg/day are encouraging in terms of prevention of steroid specific side effects and relapse, with no relapse during azathioprine maintenance therapy after one year and 83% remission rate over a median follow-up of 67 months [[179], [221], [222]]. During azathioprine maintenance therapy, patients should be monitored for cytopenias. An increased risk of malignancies is possible, although this is, as discussed, controversial [223]. In a small case series of azathioprine withdrawal after a median of five years, 50% of patients relapsed after a median of seven years [35]. The major advantage of the azathioprine regimen is the avoidance of corticosteroids and their possible detrimental long-term side effects.

An alternative strategy is to administer prednisolone at the lowest dose possible to maintain the serum transaminase levels within normal limits. The major advantages of the low dose prednisolone schedule are avoidance of long-term azathioprine therapy in fertile young adults and elimination of the theoretical risks of onco- and teratogenicity. The benefit and potential risks of a maybe lifelong azathioprine or prednisolone maintenance therapy should be discussed with the patients and the choice of which strategy to apply in patients who have achieved remission should be individualised. Factors taken into consideration include the stage of liver disease and severity of initial presentation, tolerance of induction treatment, bone density measurements, associated diseases and risk factors as well as the age of the patient.

Monitoring during treatment

Patients initiated on prednisolone/azathioprine combination therapy should have baseline clinical and laboratory parameters monitored during the first four weeks. As the steroid dose is tapered, monitoring intervals can be extended to one- to three-months. Patients with AIH require lifelong monitoring, as disease flares and relapses are frequent even after complete remission. A relapse after treatment withdrawal occurs commonly within 12 months. Patients should therefore be closely monitored after treatment withdrawal. During maintenance treatment, patients should be seen in three- to six-month intervals.

TPMT deficiency

TPMT is an enzyme involved in azathioprine metabolism. Azathioprine is converted to 6-mercaptopurine (6-MP), and this intermediate metabolite is subsequently converted in the liver to either 6-thioguanine, 6-thiouric acid or 6-methy mercaptopurine. [224] Genotyping of or measuring activity of the TPMT enzyme, which catalyses conversion of 6-mercaptopurine into inactive products, may to some extent predict azathioprine toxicity [[225], [226], [227]].

The 6-thioguanine nucleotides (6-TGN) which are responsible for immunosuppressive and anti-inflammatory properties of azathioprine can also cause myelosuppressive toxicity [228]. Impaired conversion of 6-MP to 6-thiouric acid can increase the conversion of 6-MP to the 6-TGN and can therefore increase toxicity of fixed dose azathioprine [229]. Low TPMT activity is associated with various alleles with the most common being the ∗3A allele. Homozygosity with full TPMT deficiency is rare (0.3%) and is associated with a very low enzyme activity, serious toxicity may be encountered due to an accumulation of active 6-MP metabolites, although TPMT genotyping has shown rather variable results in predicting toxicity [[230], [231]]. This is likely due to alternative pathways of metabolism, variable penetrance and possible substrate induction of TPMT activity [232]. Therefore, patients who develop azathioprine related side effects cannot be reliably identified by measuring TPMT activity or genotyping and patients with azathioprine intolerance were shown to have normal or near normal TPMT activity [[230], [233]].

In addition, it has been shown in patients with inflammatory bowel disease that this toxicity may be avoided by use of low doses with careful monitoring of metabolites in the blood [[234], [235], [236]]. Heterozygosity for the low-activity allele with intermediate enzyme activity is found in about 10% of the patients. In patients with autoimmune hepatitis neither heterozygosity nor 6-MP metabolite levels were shown to be reliable predictors of azathioprine efficacy or toxicity [[231], [233], [237]] and cytopenia is frequently encountered due to cirrhosis.

However, given the potentially serious consequence of azathioprine treatment in patients with TPMT deficiency, the benefit in terms of safety and reassurance may outweigh the arguments against universal testing so that, if available, TPMT testing may be performed prior to initiation of azathioprine therapy in patients with AIH. In patients with TPMT deficiency, prednisolone monotherapy regime or a lower dose of prednisolone combined with mycophenolate mofetil (MMF) may be used. However, as azathioprine toxicity is more frequently encountered in the absence of TPMT deficiency, close monitoring of all patients started on azathioprine is mandatory, even following TPMT activity testing.