Systematic review: managing suboptimal treatment responses in autoimmune hepatitis with conventional and nonstandard drugs


  • This uncommissioned systematic review was subject to full peer-review.

Correspondence to:

Dr A. J. Montano-Loza, Zeidler Ledcor Centre, 130 University Campus, University of Alberta, Edmonton, AB T6G 2X8, Canada.




Corticosteroid treatment for autoimmune hepatitis has been shown by randomised controlled clinical trials to ameliorate symptoms, normalise liver tests, improve histological findings and extend survival. Nevertheless, suboptimal responses to corticosteroid treatment still occur.


To describe the current definitions, frequencies, clinical relevance and treatment options for suboptimal responses, and to discuss alternative medications that have been used off-label for these occurrences.


Literature search was made for full-text papers published in English using the keyword ‘autoimmune hepatitis’. Authors' personal experience and investigational studies also helped to identify important contributions to the literature.


Suboptimal responses to standard therapy include treatment failure (7%), incomplete response (14%), drug toxicity (13%) and relapse after drug withdrawal (50–86%). The probability of a suboptimal response prior to treatment is higher in young patients and in patients with a severe presentation, jaundice, high MELD score at diagnosis, multilobular necrosis or cirrhosis, antibodies to soluble liver antigen, or inability to improve by clinical indices within two weeks or by MELD score within 7 days of conventional corticosteroid treatment. Management strategies have been developed for the adverse responses and nonstandard drugs, including mycophenolate mofetil, budesonide, ciclosporin, tacrolimus, sirolimus and rituximab, are emerging as rescue therapies or alternative frontline agents.


Once diagnosed, the suboptimal response should be treated by a highly individualised and well-monitored regimen, preferentially using first-line therapy. Nonstandard drugs warrant consideration as salvage or second-line therapies.


Autoimmune hepatitis can result in cirrhosis, death from liver failure or need for liver transplantation. Untreated severe autoimmune hepatitis has a mortality of 40% at 6 months, and those surviving commonly develop cirrhosis and subsequent complications of portal hypertension.[1] Randomised controlled trials have shown that treatment with corticosteroids alone or in combination with azathioprine is generally effective in improving clinical, laboratory and histological features and in preventing death from hepatic failure.[2-4] However, not all patients respond to conventional treatment with prednisone and azathioprine, and those who do respond may develop side effects related to the treatment or relapse after drug withdrawal. Suboptimal responses in patients with autoimmune hepatitis include treatment failure, incomplete response, drug toxicity and relapse after treatment withdrawal.[5] While there is consensus on the ideal first-line therapies for autoimmune hepatitis, there is little agreement regarding the treatment of patients with suboptimal responses. To complicate matters, suboptimal responses have varying presentations; nonstandard drugs are emerging that have been used for off-label indications; and robust clinical evidence for new treatments is lacking.

The roles of alternative salvage therapies with drugs such as mycophenolate mofetil, ciclosporin and tacrolimus among others are not clear. Small case series from diverse medical centres have reported some success with nonstandard agents, but they have involved small numbers, varying dosing schedules and different durations of follow-up. Currently, these medications are used for off-label indications, and the current guidelines of the American Association for the Study of Liver Diseases[5] do not include them in formal management algorithms. In this review, we will discuss the current definitions, frequencies, clinical relevance and treatment options for patients with autoimmune hepatitis and suboptimal responses after conventional treatment. We will present the existing evidence for the use of nonstandard medications in autoimmune hepatitis, discuss each agent in detail and provide recommendations for their use.


PubMed was searched in from January to May 2012 for full-text papers published in English-language journals, using the keywords ‘autoimmune hepatitis’.

The author's personal experience and investigational studies also identified important contributions to the literature.

Citations were chosen on the basis of their relevance to the text.

Treatment Failure in Autoimmune Hepatitis

Definition, frequency and clinical relevance of treatment failure

Treatment failure is defined as worsening of clinical, laboratory or histological findings in any combination despite compliance with standard therapy[5] (Table 1). At least 7% of patients with autoimmune hepatitis fail to respond to conventional therapy,[6-8] and these patients can progress to cirrhosis, develop hepatocellular carcinoma, require liver transplantation, or die more from liver failure with mortality rates of approximately 30%.[9, 10]

Table 1. Treatment schedules for adults patients with autoimmune hepatitis recommended by the AASLD
Weeks administeredCombination therapyPrednisone therapy
Prednisone (mg daily)AzathioprinePrednisone (mg daily)
North America (mg daily)Europe (mg/kg/day)
  1. Adapted from the American Association for the Study of the Liver Diseases (AASLD) guidelines.[5]

Maintenance until end point10501–220
Relative contraindications



Active malignancy

Short course (less than 6 months)

Thiopurine methyltransferase deficiency

Post-menopausal state





Emotional lability

Risk factors for treatment failure

Disease onset at an early age (≤40 years),[11] acute severe (fulminant) presentation,[3, 12] jaundice or high serum bilirubin level at diagnosis,[8, 13] model for end-stage liver disease (MELD) score ≥12 points at presentation,[7] HLA DRB1*03[10] and lack of improvement in laboratory indices within 1–2 weeks of corticosteroid treatment identify individuals with a high risk for treatment failure.[6, 8, 13] Inability to improve the MELD, MELD plus sodium, or the United Kingdom end-stage liver disease (UKELD) scores within 7 days of corticosteroid treatment predict a poor outcome for treatment-naïve, jaundiced individuals with a sensitivity of 85% and specificity of 68%.[8]

The immediate response to conventional corticosteroid therapy has been the most reliable determinant of outcome. Patients who demonstrate improvement in any laboratory index of liver inflammation within 2 weeks of corticosteroid therapy survive for at least 6 months in 98% of instances, whereas patients with jaundice, histological features of multilobular necrosis and unimproved or worsening laboratory indices, especially an unimproved or worsening abnormal serum bilirubin level, after 2 weeks of corticosteroid treatment, have a dismal prognosis (100% mortality).[13]

Serological markers are also useful in identifying patients who are likely to be problematic.[14] Antibodies to soluble liver antigen (anti-SLA) are associated with more severe histological changes, longer duration of treatment, higher risk of relapse after drug withdrawal and greater risk of liver transplantation or death from liver failure than patients without these antibodies.[15, 16] Moreover, anti-SLA are associated with HLA DRB1*03 and antibodies to ribonucleoprotein/Sjogren's syndrome A antigen (anti-Ro/SSA).[17-19] The major limitations of the serological markers as predictive indices are that they are not pathogenic and they are variably expressed.[14] Antibodies to SLA are detected in only 15% of white North American patients and their absence lacks a prognostic implication.[20]

Management of treatment failure

Corticosteroid responsiveness is a defining feature of autoimmune hepatitis and failure of the disease to improve suggests an alternative diagnosis.[21] Many diseases can resemble autoimmune hepatitis, including Wilson disease, chronic hepatitis C, non-alcoholic fatty liver disease, primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC). These conditions may be unrecognised at presentation, and they should be re-considered at the time of treatment failure. Moreover, autoimmune hepatitis may undergo transitions during its course, with a cholestatic syndrome such as PBC or PSC emerging that might be refractory to the original treatment.[22, 23] Consequently, all patients with treatment failure warrant histological evaluation and endoscopic or magnetic resonance cholangiography.[5]

Furthermore, another condition may be superimposed on the original process, such as PSC,[24, 25] viral infection,[26] drug toxicity[27] or fatty liver disease,[28] which could alter treatment outcome. Corticosteroid treatment is by itself a cause for obesity, diabetes and fatty liver, and worsening liver tests during treatment may not be a call for more corticosteroids.[7] Lastly, in younger patients, non-adherence to therapy should be considered as cause of treatment failure, and this must be excluded before treatment escalation.

Once the original diagnosis has been corroborated, the recommended initial step is to start therapy with high-dose prednisone (60 mg/day) or a lower dose of prednisone (30 mg/day) in combination with azathioprine (150 mg/day) for at least 1 month.[5] The doses of medication are then reduced by 10 mg for prednisone and 50 mg for azathioprine for each month of clinical and laboratory improvement until conventional maintenance doses of medication are achieved (prednisone, 20 mg daily, or prednisone, 10 mg daily, with azathioprine, 50 mg daily).[5] Approximately 70% of patients improve their clinical and laboratory parameters within 2 years, but histological remission is achieved in only 20%.[6] Most patients remain on therapy indefinitely, and they are at risk for progression of their liver disease and the development of treatment-related side effects.[6, 7] Refractory progressive disease and manifestations of liver failure during this treatment compel a liver transplant evaluation.[5]

Recently, a dosage of prednisolone (or prednisone) of 1 mg/kg/day plus azathioprine has been proposed as first-line treatment of patients with the new diagnosis of autoimmune hepatitis. The prednisolone is then reduced to 10 mg/day over 2–3 months as aminotransferases are normalised.[29-31] A previous study showed that patients who received this dosage had faster normalisation of aminotransferases (77% at 6 months) compared to 39% with standard dose prednisolone in a different randomised trial.[29] The frequency and rapidity of histological resolution, treatment tolerance and long-term outcome, including progression to cirrhosis, relapse after drug withdrawal and treatment failure, require further definition.

Incomplete Response in Autoimmune Hepatitis

Definition and frequency of incomplete response

An incomplete response is defined as improvement of the clinical, laboratory and histological findings that are insufficient to satisfy criteria for treatment withdrawal.[32] Patients who are unable to improve to normal or near-normal liver tests and liver tissue within 36 months of treatment have a higher frequency of cirrhosis (54% vs. 18%) and need for liver transplantation (15% vs. 2%) than those who improve within 6 months,[33] and the probability of these patients improving further by continuing the same regimen beyond 36 months is less than 3%.[5, 33] The frequency of an incomplete response is 14%.[33]

Clinical relevance and risk factors for incomplete response

The clinical relevance of an incomplete response relates to the duration of treatment expended in the effort to induce normal or near-normal liver tests.[33] Prognosis worsens if full suppression of disease activity cannot be accomplished within 36 months. Elderly patients (≥60 years) respond more quickly to treatment than young adults (<40 years), and 94% satisfy criteria for drug withdrawal within 24 months. In contrast, only 64% of young adults satisfy these same criteria for drug withdrawal within 24 months and only 81% respond to this degree within 36 months.[33] The young adults have HLA DRB1*03 more commonly (69% vs. 31%) and HLA DRB1*04 less frequently (20% vs. 76%) than the elderly patients, and age-related differences in the vigour of the immune response or the nature and degree of antigen exposure may account for these findings.[33] Protracted immunosuppressive therapy without resolution of disease also subjects patients to treatment-related side effects, progression to cirrhosis and the risk of hepatocellular cancer.[9]

Current management guidelines are that treatment be continued until normal serum AST and ALT, γ-globulin and immunoglobulin G (IgG) levels and histological improvement to normal or near-normal liver tissue.[5] Normal laboratory indices before termination of treatment reduce the relative risk of relapse after drug withdrawal by 3- to 11-fold compared with patients who do not achieve these results, and 87% of patients who remain inactive long-term have normal laboratory indices prior to drug withdrawal.[34]

The pursuit of normal liver tests and liver tissue is the ideal goal of therapy, but it can lead to protracted treatment, drug-related side effects and an incomplete response. The attainment of normal liver tissue is a daunting challenge. Only 18 of 83 patients (22%) without cirrhosis at presentation revert to normal liver tissue after 56 ± 5 months of conventional corticosteroid treatment, and multiple liver biopsy examinations are necessary to document this result (mean, 8 ± 1 liver tissue samples per patient).[32] Furthermore, the attainment of normal liver tissue does not preclude relapse after treatment withdrawal (20–28% occurrence)[32, 34]; patients with portal hepatitis at the termination of therapy can still improve to normal liver tissue spontaneously after corticosteroid withdrawal (18% occurrence)[32, 35]; and persistent portal hepatitis after treatment can remain asymptomatic without disease progression.[36] In patients with cirrhosis at presentation (≤28% occurrence) or during therapy (≥20% occurrence), the ideal goal of normal liver tests and liver tissue is unachievable, and by this standard, they will always have an incomplete response.[32, 37, 38]

The prognosis of autoimmune hepatitis improves when the ALT level is maintained at ≤40 IU/L,[39] and the frequency of relapse after drug withdrawal is reduced by normalising AST and ALT levels, decreasing abnormal fluctuations of these enzymes, and eliminating plasma cells from the liver tissue.[40, 41] In a Swedish study of 473 patients with autoimmune hepatitis, individuals with ALT levels that persisted above 5-fold the upper limit of the normal (ULN) range had a greater mortality and need for liver transplantation than patients with normal or improved (<5-fold ULN) ALT levels.[42]

An incomplete response that persists indefinitely can be deleterious. Treatment must be individualised to the rapidity of response recognising that treatment beyond 24 months in the elderly and beyond 36 months in the young adult exceeds expectation and denotes a suboptimal response.[33] Serological markers at presentation, such as antibodies to α-actinin which are directed against a predominant epitope on filimentous actin,[43, 44] are being evaluated as predictors of response, and they may facilitate the early recognition of patients with a slow response that is destined to be incomplete.[45]

Management of incomplete response

A practical approach to reducing the frequency of an incomplete response is to broaden the objectives of treatment. The histological pattern that has been classically associated with the identity and the aggressiveness of autoimmune hepatitis is interface hepatitis, and treatment until disappearance of interface hepatitis from the liver tissue is an important and achievable goal.[46] Improvements to portal hepatitis, inactive cirrhosis or minimally active cirrhosis are histological end-points that reflect clinically inactive nonprogressive disease, and they constitute practical clinical objectives. The elimination of plasma cells from the liver tissue is another treatment objective that is achievable and is associated with a decreased frequency of relapse after drug withdrawal.[37, 38] Pursuit of these objectives will reduce disease progression, facilitate a reprieve from continuous corticosteroid therapy and decrease the frequency of an incomplete response.[46]

Patients with incomplete response by rigid or modified criteria are candidates for indefinite maintenance therapy with azathioprine alone[47, 48] or low-dose prednisone.[49, 50] Azathioprine (2 mg/kg daily) is the preferred long-term maintenance therapy.[5] The dose of azathioprine is increased or added to the conventional regimen, and the dose of corticosteroids is reduced or withdrawn in a tapered fashion if the AST level can be maintained below three-fold the ULN. Previous studies have demonstrated that histological features of interface hepatitis disappear from the liver biopsy tissue when AST levels are <3-fold ULN[51] and the disease does not progress.[33, 47]

Low-dose prednisone (≤10 mg daily, median dose, 7.5 mg daily) can also be used long-term to maintain the AST level <3-fold ULN, and this is an especially appropriate regimen for patients with cytopaenia or known azathioprine intolerance.[51] Low-dose therapy maintained for 7–43 years (mean, 16 years) has resulted in normal aminotransferase levels in 81%, progression to cirrhosis in 54%, minimal side effects (weight gain, osteoporosis) in 26%, severe side effects (cryptococcal meningitis, aseptic hip necrosis) in 5%, liver transplantation in 2% and no deaths from hepatic failure.[49, 50]

Drug Toxicity in Autoimmune Hepatitis

Drug toxicity means the development of a side effect due to prednisone or azathioprine intolerance.

Prednisone related side effects

The side effects associated with prednisone in adults with autoimmune hepatitis are typically mild, and they consist of cosmetic changes, including facial rounding, weight gain, dorsal hump formation, striae, hirsutism, or alopecia, and somatic changes, including emotional instability, glucose intolerance or cataract formation. These mild cosmetic and somatic side effects are present in 80% of patients after 2 years.[52]

Side effects related to prednisone therapy that are severe enough to warrant premature discontinuation of the medication occur in 13%,[53] and they include osteopaenia or osteoporosis, vertebral compression, diabetes, psychosis and hypertension. Other side effects that are uncommon include inflammatory and neoplastic diseases, such as pancreatitis, opportunistic infection and malignancy.[52]

Azathioprine-related side effects

The frequency of azathioprine-related side effects in patients with autoimmune hepatitis is 10%,[54] and the most frequent side effects are haematological. Mild cytopaenias develop in up to 46%, mainly in patients with cirrhosis, and severe leucopaenia or thrombocytopaenia that justifies premature discontinuation of medication or dose reduction occurs in ≤6%.[55, 56] Severe myelosuppression is rare, but it may be fatal.[56, 57]

Other complications are usually mild somatic reactions, such as nausea, emesis, skin rash, fever and arthralgias,[52] but serious complications, such as pancreatitis, cholestatic hepatitis, sinusoidal obstruction syndrome, extrahepatic malignancies, intestinal villous atrophy with malabsorption and nodular regenerative hyperplasia, are possible.[27, 58-61] Azathioprine has been associated with congenital defects in mice,[62] and its metabolites cross the placenta.[63] Accordingly, it has been designated by the Food and Drug Administration of the United States as a Category D drug in pregnancy. However, studies of pregnant patients with autoimmune hepatitis have not shown any association between azathioprine use during pregnancy and adverse outcomes,[64-68] and larger studies of patients with inflammatory bowel disease also support the safety of this medication during pregnancy.[69]

Management of drug toxicity

The presence of drug toxicity justifies discontinuation of the offending drug or a reduction in dose. Therapy with the tolerated agent, either prednisone or azathioprine, can be maintained in adjusted dose to prevent worsening in the clinical and laboratory features.[5]

Relapse After Corticosteroid Withdrawal

Definition, frequency and clinical relevance of relapse

Relapse is defined as an exacerbation of disease activity after induction of remission and drug withdrawal.[5, 70] Relapse is characterised by an increase in the AST level to more than 3 times ULN and/or an increase in the γ-globulin level to more than 2 g/dL (or 20 g/L). An elevation of the serum IgG concentration could also be used as a surrogate marker of relapse.[51] Laboratory changes of this degree are associated with the histological re-appearance of interface hepatitis, and a liver tissue examination is not required if there are no other bases for the flare.[48]

Relapse occurs in approximately 50–86% of patients, and the frequency depends mainly on the laboratory and histological findings prior to drug withdrawal.[32, 34, 36, 71] The frequency of relapse can be reduced to as low as 20% if normal liver tissue can be achieved.[32, 34] In most instances, relapse cannot be predicted, and its occurrence should be anticipated by maintaining laboratory surveillance at 3-week intervals for at least 3 months and then at 3- to 6-month intervals thereafter.[5] The high frequency of relapse should not dissuade drug withdrawal as 21% of patients remain inactive long-term without medication.[36]

Repeated relapses and re-treatments are associated with progressive increase in the frequency of cirrhosis (38% vs. 4%) and requirement for liver transplantation or death from hepatic failure (20% vs. 0%),[69] and the prevention of these serious consequences requires a change in management strategy after the first relapse.[5]

Risk factors for relapse

The major risk factor for relapse is residual liver inflammation after treatment withdrawal manifested by a lymphoplasmacytic infiltrate with mild interface hepatitis in the biopsy specimen and mild elevation of the serum AST or γ-globulin level.[34, 40, 41] These findings are not highly predictive of outcome as the liver inflammation may not exacerbate, may remain mild and asymptomatic indefinitely or resolve spontaneously.[72] Furthermore, normal liver tests and liver tissue do not preclude relapse, and an exacerbation is possible in at least 20% of patients with an ideal end point.[32, 34]

Serological markers may be reflective of residual liver inflammation and the propensity to relapse. The presence of antibodies to asialoglycoprotein receptor (anti-ASGPR) has characterised patients who relapse, and treatment until disappearance of these antibodies has been associated with histological resolution and sustained long-term remission after drug withdrawal.[72-74] Similarly, anti-SLA has been associated with HLA DRB1*03 and a high frequency of relapse,[15] and antibodies to α-actinin may also be instruments by which to judge residual disease activity.[45] Early studies have also demonstrated that presence of type 1 liver-kidney microsomal antibody (anti-LKM1) at presentation is associated with a higher risk of relapse after treatment withdrawal.[75]

Management of relapse

Relapse justifies the re-institution of the original therapy with prednisone alone or in lower dose combined with azathioprine.[5] Conventional treatment is continued until clinical and laboratory resolution is achieved, and then azathioprine is added to the regimen or the dosage is increased to 2 mg/kg daily as prednisone is gradually withdrawn.[5] Azathioprine is then continued indefinitely as the sole drug. More than 80% of adult patients managed by maintenance therapy with azathioprine remain in remission during long-term follow-up.[47, 48] The risk of serious consequences from the treatment and the liver disease increases with repeated relapse and re-treatments,[69] and these observations have justified the institution of maintenance therapy after the first relapse.

Another option after the first relapse is to administer prednisone in the lowest dose possible to maintain the aminotransferases levels within normal limits or at least below 3-fold ULN.[49] Dosages of prednisone less than 10 mg daily are generally well tolerated, and most patients can be managed long-term in this fashion (median dosage, 7.5 mg daily). Patients with severe cytopaenia and previous azathioprine intolerance are candidates for this treatment.[46] Budesonide may be an alternative to prednisone, especially in noncirrhotic patients with uncomplicated disease, as it has fewer glucocorticoid side effects.[76] The substitution of budesonide for prednisone can be complicated by severe corticosteroid withdrawal symptoms and exacerbations of concurrent immune diseases that no longer have the same systemic exposure to glucocorticoid.[77] In some patients, maintenance therapy may eventually become unnecessary, especially in patients who have remained inactive for more than 12 months.[71]

Alternative Drug Therapies for Suboptimal Responses

The success of standard corticosteroid therapy in improving the clinical manifestations and enhancing the survival of patients with autoimmune hepatitis has dampened interest in pursuing alternative treatments that have greater expense, risks and uncertainties than the current therapies.[78] Furthermore, autoimmune hepatitis is relatively rare; suboptimal responses can be managed by conventional medication (Table 2); funding priorities for research have been directed elsewhere; therapeutic animal models that mimic the human disease are unavailable; and a collaborative network of investigators that can evaluate new treatments in a rigorous and timely fashion does not exist.[79] Nevertheless, new therapies in autoimmune hepatitis are needed to prevent suboptimal responses and to treat those that occur. The availability of agents that promise immunosuppressive actions that are superior to prednisone and azathioprine and the successes of these agents in animal models and humans with other immune-mediated inflammatory diseases encourage their evaluation in autoimmune hepatitis.[80-82] The most promising medications that have emerged to treat or prevent the suboptimal response are ciclosporin, tacrolimus, mycophenolate mofetil (MMF) and budesonide. Rituximab and sirolimus have had more theoretical than actual bases for their consideration,[80-82] but they exemplify a period of investigation in autoimmune hepatitis that is responding to clinical need by applying new drugs in small but highly selected clinical situations, and invigorating a call for large formal clinical trials. The nonstandard medications are unlicensed for use in autoimmune hepatitis, and their consideration in this disease has been based on reported experiences in small, single-centre trials.

Table 2. Conventional and empiric treatments of suboptimal responses in autoimmune hepatitis
Suboptimal responsesConventional treatmentsAlternative drug therapies
Treatment failurePrednisone (30 mg daily) and azathioprine (150 mg daily)

Ciclosporin (5–6 mg/kg daily)

Tacrolimus (4 mg twice daily)

Drug toxicity

Azathioprine (2 mg/kg daily) if prednisone intolerance

Prednisone (20 mg daily) if azathioprine intolerance

Budesonide if prednisone intolerance (3 mg twice daily)

Mycophenolate mofetil (1 g twice daily) if azathioprine intolerance

Incomplete response

Prednisone maintenance ≤10 mg daily if serum AST <three-fold normal

Azathioprine maintenance (2 mg/kg daily) if serum AST <three-fold normal

Budesonide maintenance (3 mg twice daily)

Mycophenolate Mofetil maintenance (1 g twice daily)


After first relapse:

Re-start original prednisone ± Azathioprine regimen until normal test

Consider Treatment Withdrawal only if Complete Remission >12 months

After second relapse:

Azathioprine maintenance (2 mg/kg daily) if serum AST <three-fold normal

Prednisone maintenance reduced to ≤10 mg daily if serum AST <three-fold normal

Mycophenolate mofetil maintenance (1 g twice daily)

Ciclosporin maintenance (5–6 mg/kg daily)


Ciclosporin binds to cyclophilin and inhibits the phosphatase activity of calcineurin. Consequently, it impairs transcription of interleukin-2 and the downstream activation events depending on this cytokine. In dosages of 5–6 mg/kg daily, it has been used successfully as ‘salvage’ therapy in patients who have failed or have been intolerant of conventional corticosteroid therapy, and it has also been used as first-line therapy.[83-89]

Ciclosporin was first used in 1985, in a 51-year-old man with 5-year history of autoimmune hepatitis refractory to corticosteroid therapy. Presenting symptoms improved for the first time since the onset of his illness, liver enzyme levels fell to near-normal values and virtually no side effects were experienced.[83] Since 1985, other single case reports of patients who failed treatment with corticosteroids but responded well to treatment with ciclosporin have been described.[84-86] Sherman et al. reported six patients with autoimmune hepatitis who had previously undergone treatment with corticosteroids with or without azathioprine and had failed to achieve laboratory resolution. Five of six patients normalised or nearly normalised aminotransferases levels within 10 weeks, and biochemical remission was sustained for periods of up to 1 year.[87] A similar study described five patients who did not respond to corticosteroid and azathioprine therapy and who were treated with ciclosporin at 2–3 mg/kg/day. Biochemical remission was induced in four of the five patients within 3 months. Ciclosporin was generally well tolerated, and none of the patients developed renal insufficiency.[88]

Malekzadeh et al. described 19 patients with autoimmune hepatitis who were managed with ciclosporin. Fifteen patients completed a 26-week course of treatment, and four discontinued ciclosporin due to side effects. Of the patients who completed therapy, all achieved remission and the histological activity scores in the liver tissue specimens of 14 patients after treatment showed improvement.[89]

Importantly, ciclosporin has not been compared with conventional treatments in randomised trials, and its use, especially as a first-line drug, remains controversial. Ciclosporin has been used in autoimmune hepatitis in a dosage of 5–6 mg/kg/daily and target trough levels have been between 100 and 250 ng/mL.[86, 88, 89] Laboratory tests of kidney function (creatinine, blood urea nitrogen), electrolyte status, glucose, liver inflammation (AST, ALT, bilirubin, IgG) and ciclosporin trough levels should be performed once a week during the first month, every 2–4 weeks from the 2nd to the 6th month, and once a month thereafter.

Side effects of ciclosporin include renal insufficiency, hypertension, malignancy and neurotoxicity. Manifestations of neurotoxicity develop in 10–60% of patients, and they are usually mild (tremor, neuralgia, headache, focal neurological deficits, or peripheral neuropathy).[90] In 5% of patients, psychoses, hallucinations, blindness, seizures, ataxia, muscle weakness or akinetic mutism can be severe and potentially permanent or fatal.[90, 91]


Tacrolimus was first described in the therapy for autoimmune hepatitis in 1995.[92] Treatment-naïve patients were given twice-daily dosages of tacrolimus starting with 0.075 mg/kg. The dosage was then titrated to achieve a trough level of 0.6–1.0 ng/mL.[92, 93] Twenty one patients were studied and most of them had clinical evidence of jaundice (mean bilirubin level 2.0 mg/dL or 34 μmol/L), a surrogate marker of disease severity. The target dose was achieved at 3 weeks, and the AST and ALT levels improved by 75% in conjunction with a decrease in bilirubin concentrations. Renal function was unimpaired and the medication was well tolerated.[92]

To date, there have been four case series of tacrolimus in patients with autoimmune hepatitis and steroid-refractory disease. Larsen and colleagues identified nine patients who had steroid-refractory disease [those who did not respond to an additional course of prednisone, mycophenolate mofetil (MMF), and/or ciclosporin]. These patients were started on tacrolimus adjusted to keep the concentration below 6 ng/mL and for a median duration of 18 months. Prednisone therapy was continued in all patients and the median dosage of tacrolimus at the end of follow-up was 3 mg/day. All patients had a reduction in serum aminotransferase levels; the degree of fibrosis decreased; and the histological features of liver inflammation also improved in most patients. Furthermore, tacrolimus was well tolerated in all patients without major side effects, such as arterial hypertension and renal insufficiency.[93] In another study, 13 patients with autoimmune hepatitis were treated with tacrolimus because of a failure to normalise serum aminotransferase levels with prednisone or because of intolerance to other immunosuppressive agents. Approximately 60% had bridging fibrosis and 36% had cirrhosis. They were given a dosage of 2–6 mg/day with a mean trough serum concentration of 6.0 ng/mL. Normalisation of liver enzymes was seen in 12 of the 13 patients, and tacrolimus was discontinued in two patients after development of haemolytic uremic syndrome or and squamous cell carcinoma.[94]

The side effects of tacrolimus are similar to those of ciclosporin. Diabetes mellitus, neurotoxicity, nephrotoxicity, diarrhoea, pruritus and alopecia may occur more frequently with tacrolimus, whereas hypertension, hyperlipidaemia, hirsutism, gingivitis and gum hyperplasia occur more commonly with ciclosporin.[95, 96]

Mycophenolate mofetil

MMF is the ester pro-drug of mycophenolic acid that inhibits inosine monophosphate dehydrogenase, leading to de novo inhibition of DNA synthesis. Lymphocytes do not possess alternative pathways of purine synthesis, and they are preferentially targeted by this drug. MMF is used in the treatment of autoimmune disorders such as lupus, psoriasis and rheumatoid arthritis and in the transplant population for immunosuppression.[97] The safety profile of MMF is tolerable, with gastrointestinal effects of nausea and diarrhoea being the most commonly encountered symptoms. MMF is potentially teratogenic and is associated with first trimester pregnancy loss. In Canada, use outside of indications can be prohibitively expensive necessitating special release.

The efficacy of MMF as first-line therapy for autoimmune hepatitis has been highlighted by a recent prospective trial from Greece evaluating 59 treatment-naïve patients who were given prednisolone in conjunction with MMF, 1.5–2 g/day. Eighty-eight per cent responded within 3 months, and 12% had a partial response. Complete remission was achieved in 59%, and it required normalisation of the aminotransferase and immunoglobulin G levels. Prednisolone was withdrawn in 58% in 8 months.[98] Severe side effects compelled discontinuation of MMF in only 3% of patients. As conventional therapy with prednisone in combination with azathioprine leads to a partial or complete response in over 90% within the first year, the advantages of an alternative, more expensive, first-line treatment are uncertain.

MMF as treatment for autoimmune hepatitis was first described in a case report of a patient refractory to conventional treatment.[99] A subsequent report described seven patients who were initially treated with prednisolone monotherapy and who responded with a decrease in aminotransferases levels.[100] Maintenance treatment with azathioprine, however, was not successful as aminotransferase levels increased during treatment or intolerance to azathioprine (nausea, vomiting and neutropaenia) developed. MMF was started at a dosage of 250 mg twice daily and increased to 1 g twice daily in addition to prednisone (15–20 mg/day). MMF was tolerated by all patients; 71% normalised aminotransferase levels after 3 months; and the average steroid dosage was reduced from a median of 20 to 2 mg/day. Moreover, all but one patient were successfully weaned from prednisone.[100]

Another study evaluated 15 patients with autoimmune hepatitis who were intolerant of or unresponsive to corticosteroids with or without azathioprine, and approximately 75% of patients normalised their aminotransferase levels during therapy with MMF. Histological scores for liver inflammation and fibrosis improved, but most patients still required dual therapy with prednisone.[101]

Hlivko et al. reported an experience with 29 patients who were treated with MMF either empirically as a first choice or as a salvage therapy for intolerance or unresponsiveness to the standard regimen, and they demonstrated an improvement in 84%.[102] Similarly, a retrospective review from Germany identified 37 patients treated with MMF who included eight unresponsive to treatment and 28 intolerant to azathioprine. This study found that 75% of those unresponsive to azathioprine did not respond to MMF, and only 25% reached biochemical remission. In contrast, 43% of the 28 patients who were treated because of intolerance to azathioprine achieved remission with MMF. Overall, 61% of patients did not respond favourably to the drug.[103] A smaller Canadian study identified 11 patients who had failed conventional therapy and in whom treatment with MMF induced a complete laboratory response in 64%.[104] A similar study identified 12 patients who had been converted to therapy with MMF because of refractory disease, and of these none achieved biochemical resolution.[105]

The discrepancies between studies in the frequency of treatment response probably reflect different clinical situations that have been treated with MMF. The drug seems to be more effective in rescuing patients from azathioprine intolerance in comparison with patients with steroid-refractory disease. Of those patients who were on MMF because of intolerance to azathioprine, 88% improved. In contrast, all patients treated for steroid-refractory disease showed biochemical improvement, but none had a complete response.[106] These findings have been supported by another retrospective analysis of 30 patients in which 73% with azathioprine intolerance improved compared to only 20% with azathioprine-refractory disease.[101] Therefore, MMF is emerging as a viable alternative treatment in those patients who are intolerant of first-line therapy with prednisone and azathioprine. For the 5–10% who experience side effects of conventional therapy, MMF may become the therapy of choice.[106]

Side effects have developed in 3–34% of patients with autoimmune hepatitis who have been treated with MMF, and the most common findings are nausea and leucopaenia.[105-107] MMF has been associated with miscarriages and congenital defects in humans, and it should not be used in pregnancy or in patients contemplating pregnancy.[108-110]


Budesonide is a next-generation glucocorticoid, and it has a high first-pass clearance by the liver and metabolites that are devoid of glucocorticoid activity. Preliminary studies in patients with mild disease have suggested that it is useful in selected patients as a frontline treatment,[111-113] and these small reports have justified the performance of a randomised clinical trial. The European trial compared the combination regimen of budesonide (3 mg thrice daily) and azathioprine (1–2 mg/kg daily) with that of prednisone (40 mg daily, tapered to 10 mg daily) and azathioprine (1–2 mg/kg daily) in 203 treatment-naïve patients with autoimmune hepatitis. The primary end-point was to achieve complete remission without the typical steroid-induced side effects of acne, hirsutism, diabetes, striae, glaucoma and facial swelling. Biochemical remission was achieved more frequently after 6 months in the patients treated with budesonide compared with patients treated with prednisone (47% vs. 18%), and side effects were fewer (28% vs. 53%). Normalisation of bilirubin and IgG levels occurred with similar frequencies in both treatment groups.[76] The frequency of histological resolution and the durability of the response with budesonide therapy are unknown, and the low frequency of response (18%) and high occurrence of side effects (53%) in the patients treated for only 6 months with conventional therapy are unexplained. Furthermore, budesonide has not been effective in patients who are refractory to corticosteroid treatment or corticosteroid-dependent.[77]

The side effects of budesonide are the same as those of prednisone, but they occur less frequently in noncirrhotic patients.[76] The high first-pass clearance of the drug by the liver, its low systemic availability and metabolites devoid of glucocorticoid activity contribute to its greater tolerance. These favourable attributes can be troublesome in certain situations. Concurrent immune-mediated diseases, which are common in autoimmune hepatitis, can exacerbate because of low systemic exposure to the drug[77]; patients on prednisone cannot be easily switched to budesonide without severe corticosteroid withdrawal symptoms[77]; and patients with cirrhosis are at risk for glucocorticoid-induced complications because of decreased hepatic metabolism and high plasma concentrations of the active drug.[114-116]

Budesonide in combination with azathioprine seems to be most appropriate in treatment-naïve, noncirrhotic patients with uncomplicated disease and in patients with mild disease at risk for corticosteroid-induced complications (diabetes, obesity, osteoporosis and hypertension).[80-82]


Rituximab is a chimeric monoclonal anti-CD20 antibody that can induce depletion of B lymphocytes by targeting their cell-surface receptor. This interaction may affect complement activation, antibody-dependent cytotoxicity and the induction of apoptosis. Its rationale in autoimmune hepatitis is based on a presumption that B-cell expansion and hyperactivity can be suppressed by blocking the CD20 cell-surface receptor of these cells. The unregulated proliferation of activated plasma cells can result in the production of immunoglobulins that adhere to normal membrane constituents of the hepatocytes. These aggregates can then become targets of natural killer cells with Fc receptors, and the antigen–antibody complex on the hepatocytes surface can induce an antibody-dependent cytotoxicity.[117, 118]

These presumptions have been supported by the successful treatment of autoimmune hepatitis and idiopathic thrombocytopaenic purpura[119]; autoimmune hepatitis and cryoglobulinemic glomerulonephritis[120]; autoimmune hepatitis and previous B cell lymphoma[121]; and autoimmune hepatitis and Evans syndrome (autoimmune haemolytic anaemia and idiopathic thrombocytopaenic purpura).[122]

A recent study has extended treatment with rituximab to patients without co-existent B-cell disease. In an abstract describing 6 patients with autoimmune hepatitis intolerant of or refractory to therapy with prednisone and azathioprine, rituximab administered intravenously (1000 mg at days 1 and 15) was associated with biochemical resolution and improvement of IgG and γ-globulin levels with no serious adverse events.[123] These findings indicate that rituximab may have a role in the treatment of refractory autoimmune hepatitis and that formal study in a research protocol should be considered.

Rituximab has been licensed for use in adults with CD20-positive B-cell lymphoma or rheumatoid arthritis, but it has also been used for off-label indications that include refractory non-Hodgkin lymphoma, leukaemia, haemophilia, chronic immune thrombocytopaenic purpura, mixed cryoglobulinaemia and nephrotic syndrome. The complications of treatment with monoclonal antibodies to CD20 have included late onset neutropaenia, especially after stem-cell transplantation and in rheumatic diseases,[124] hepatitis B virus reactivation,[125] infection,[126] interstitial pneumonitis,[127] intestinal perforation[128] and possible multifocal leucoencephalopathy associated with activation of a latent polyoma virus (JC virus).[129, 130] Complications have typically been infrequent and mild in clinical trials, and they have mainly been infusion reactions, bacterial infections, neutropaenia, anaemia, rash, fever, diarrhoea and reactivation of viral infections.[131]


Sirolimus or rapamycin is a drug frequently used in liver transplantation, mainly for patients with renal insufficiency after transplantation, and it has been especially useful in the treatment of transplanted patients with nephrotoxicity associated with the use of calcineurin inhibitors.[132] Sirolimus acts by inhibiting the mammalian target of rapamycin (mTOR), a protein that modulates the proliferation and survival of activated lymphocytes.[133] It also increases the apoptosis of antigen-sensitised CD4+ and CD8+ lymphocytes and thus decreases the production of perforin and granzyme B. The apoptosis of hepatocytes is diminished, and the immune-mediated pathway of liver injury is interrupted.[134-136] Sirolimus has been reported to be effective as salvage treatment in five patients with posttransplant autoimmune hepatitis (de novo autoimmune hepatitis) who were nonresponders to conventional therapies with prednisone, azathioprine, MMF and the calcineurin inhibitors.[137] Consequently, this nonstandard off-label medication is a theoretical salvage therapy for refractory autoimmune hepatitis in untransplanted patients. There have been no reported experiences with sirolimus in autoimmune hepatitis outside liver transplantation, and there are currently no guidelines or clinical justifications for its use.

Side effects include dyspnoea, cough, fatigue and fever, which commonly reflect a pulmonary toxicity that has occurred in 2–7% of patients receiving the drug after organ transplantation.[138] Hyperlipidaemia, myelosuppression, proteinuria, delayed wound healing, oedema, stomatitis and skin rash have also been described, and sirolimus may enhance the toxicity of the calcineurin inhibitors.[138] The most common adverse effects of sirolimus are hyperlipidaemia and myelosuppression.[139]

Other medications that have been used in patients resistant to conventional treatment with good results, such as sustained biochemical remission and histological improvement include infliximab,[140] methotrexate[141] and ciclophosphamide[143]; however, these are only case series and more data are needed on these agents before any being considered as salvage therapy for autoimmune hepatitis.

Monitoring Patients with a Suboptimal Response

Patients with autoimmune hepatitis who are receiving treatment for refractory disease should have aminotransferase, bilirubin and IgG (or γ-globulin) levels determined at 3-week intervals until laboratory improvement and tolerance of the new treatment schedule have been demonstrated.[5] Patients on long-term corticosteroid treatment should be monitored for hyperglycaemia and bone disease with bone densitometries at baseline and then at annual intervals.[5] Patients receiving azathioprine or MMF should be monitored for myelosuppression by performing leucocyte and platelet counts on a weekly basis at the initiation of the treatment and at 1- to 3-month intervals thereafter, depending on haematological stability on the new schedule.

Azathioprine-treated patients with cytopaenia before or during therapy are candidates for assessment of the blood thiopurine methyltransferase activity.[5] Determinations of this enzyme activity or the genetic phenotypes associated with enzyme deficiencies have not correlated well with azathioprine intolerance,[142, 144, 145] but bone marrow failure has occurred in patients receiving azathioprine [53-57] and the absence of thiopurine methyltransferase activity can contribute to this risk. Normal enzyme activity does not preclude the need for regular haematological surveillance during azathioprine therapy, but an absence of enzyme activity would limit the use of the drug.

For patients receiving calcineurin inhibitors, ciclosporin should be used in a dosage of 5–6 mg/kg daily with the goal of achieving target trough levels between 100 and 250 ng/mL.[87, 89, 90] Tacrolimus should be used in a dose of 2 mg bid with the goal of achieving target trough levels between 5 and 8 ng/mL.[94, 95] Patients must be monitored for toxic blood levels and other side effects as discussed above. Lastly, all patients with cirrhosis should undergo hepatic ultrasonography at 6-month intervals to detect hepatocellular carcinoma as in other causes of liver cirrhosis.[5]


Suboptimal responses to conventional corticosteroid treatment include treatment failure, incomplete response, drug toxicity and relapse during drug withdrawal. The possibility of a suboptimal response can be anticipated prior to treatment in young patients and in patients with acute severe (fulminant) presentation, jaundice, high MELD score at diagnosis, multilobular necrosis or cirrhosis or the presence of antibodies to SLA or ASGPR. Antibodies to LKM1 may characterise patients with a propensity for relapse after drug withdrawal. A suboptimal response can be anticipated after the institution of conventional treatment by lack of improvement in laboratory parameters within 2 weeks, failure of the MELD scores to improve within 7 days and by the inability to achieve clinical, laboratory and histological improvement to normal or near-normal levels within 6 months (Table 3).

Table 3. Risk factors for suboptimal responses in autoimmune hepatitis
Suboptimal responsesAt diagnosis After conventional treatment
  1. Anti-α-actinin, antibodies to anti-α-actinin; antibodies to soluble liver antigen; Anti-SLA, antibodies to soluble liver antigen; Anti-LKM1, antibodies to liver-kidney microsome type 1; Anti-ASGPR, antibodies to asialoglycoprotein receptor.

Treatment failure

Earlier age

Acute or Severe Presentation

Jaundice or High Bilirubin

MELD >12

Multilobular necrosis

Lack of improvement in laboratory parameters 2 weeks after initiation of corticosteroids

Day 7 change in MELD, MELD-Na, UKELD

Drug toxicityCirrhosisCirrhosis
Incomplete response



Absence of Rapid treatment response (defined as response within 6 months)

Anti-SLA positivity

Anti-ASGPR positivity

Anti-LKM1 positivity

Lack of normalisation of aminotransferases level, γ-globulin and IgG during remission

Once diagnosed, the suboptimal response should be treated by a highly individualised and well-monitored regimen, preferentially using first-line therapy. Treatment failure should be managed with high-dose prednisone alone or in combination with high-dose azathioprine. Ciclosporin or tacrolimus should be considered if the disease remains refractory or drug intolerances develop. A liver transplantation evaluation is also appropriate in these patients (Figure 1). Drug toxicity can be managed by using higher doses of the tolerated medication. Second-line options for drug toxicity are budesonide for prednisone intolerance or MMF for azathioprine intolerance (Figure 1). An incomplete response requires downward adjustments in the dose of prednisone as tolerated to maintain aminotransferase levels below three-fold ULN. Azathioprine can be used in the absence of prednisone, depending on individual patient response and tolerance. Persistent aminotransferase levels more than three-fold ULN or side effects with the maintenance regimen then are opportunities to consider budesonide as a substitute for prednisone or MMF as a substitute for azathioprine (Figure 1). Relapse after corticosteroid withdrawal necessitates re-institution of the original corticosteroid regimen until clinical symptoms and liver tests resolve. To avoid the consequences of liver failure or need for liver transplantation, a chronic maintenance regimen with azathioprine should be instituted after the first relapse. Treatment withdrawal from the maintenance regimen can be considered if completely inactive disease has been documented for more than 12 months (Table 4). A second relapse compels long-term treatment with azathioprine or prednisone alone to maintain serum aminotransferase levels <three-fold ULN (Figure 1).

Figure 1.

Algorithm of treatment decisions in patients with autoimmune hepatitis and suboptimal responses to conventional treatment, including treatment failure, incomplete response, drug toxicity and relapse after corticosteroid withdrawal. In this algorithm, we incorporate nonstandard off-label medications in the treatment of autoimmune hepatitis based on experiences in multiple small single-centre studies of ciclosporin, tacrolimus and mycophenolate mofetil, and one large randomised clinical trial of budesonide combined with azathioprine.

Table 4. Alternative drug therapies for autoimmune hepatitis
DrugDoseActionsPrincipal indication
Ciclosporin 5–6 mg per kg dailyCalcineurin inhibitor; impairs transcription of IL 2; prevents T lymphocyte proliferation; increases hepatic TGF-βSalvage therapy for treatment failure
Tacrolimus4 mg twice dailyCalcineurin inhibitor; impairs transcription of IL 2; limits expression of IL 2 receptors; increases hepatic TGF-βSalvage therapy for treatment failure
Mycophenolate mofetil1 g twice dailyPurine inhibitor; independent of thiopurine methyltransferase; impairs DNA synthesis; reduces T lymphocyte proliferationAzathioprine intolerance
Budesonide3 mg thrice dailyCorticosteroid actions; high first pass clearance by liver; metabolites devoid of glucocorticoid activityInitial therapy combined with azathioprine for non-cirrhotic patients

Lamentably, most of the information regarding the management of suboptimal responses in autoimmune hepatitis is derived from retrospective noncontrolled studies. There is a need for larger, controlled and prospective studies in autoimmune hepatitis comparing different treatment strategies, and a multicentre, collaborative network of clinical investigators are the remaining requisites.

Finally, liver transplantation is the best rescue treatment for liver failure in autoimmune hepatitis, and its success should not be compromised by the use of ineffective or hazardous treatments.[10, 146-148] The implementation of a new nonstandard therapy should not delay or supersede the decision to proceed with liver transplantation in the appropriate candidate.


Declaration of personal interests: Aldo J. Montano-Loza received a clinical research award in 2011, from the American College of Gastroenterology. Declaration of funding interests: None.