If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Norwegian PSC Research Centre, Department of Transplantation Medicine, Division of Surgery, Inflammatory Diseases and Transplantation, Oslo University Hospital Rikshospitalet, Oslo, NorwayInstitute of Clinical Medicine, University of Oslo, Oslo, NorwayEuropean Reference Network on Hepatological Diseases (ERN RARE-LIVER)
Department of Clinical Science, University of Bergen, Bergen, NorwayNorwegian PSC Research Centre, Department of Transplantation Medicine, Division of Surgery, Inflammatory Diseases and Transplantation, Oslo University Hospital Rikshospitalet, Oslo, NorwayEuropean Reference Network on Hepatological Diseases (ERN RARE-LIVER)
Most patients with autoimmune hepatitis respond well to standard immunosuppressive therapy with steroids and azathioprine, and while untreated disease is usually fatal, patients who respond well to therapy have an excellent prognosis. However, insufficient response to standard therapy or intolerable side effects requiring dose adaptions or treatment changes occur in 10–20% of patients. While there is fairly good agreement on second-line treatment options, there is very wide variation in the indication and use of possible third-line therapies. Herein, the European Reference Network on Hepatological Diseases (ERN RARE-LIVER) and the International Autoimmune Hepatitis Group (IAIHG) outline a treatment algorithm for both children and adults that should help to standardise treatment approaches, in order to improve patient care and to enable the comparison of treatment results between scientific publications.
In their letter, Staffan Wahlin and Cumali Efe argue that tacrolimus should be regarded as an equal alternative second-line therapy to mycophenolate mofetil. The considerations put forward by the authors are some of many that were discussed in detail amongst the members of the European Reference Network for Hepatological Diseases (ERN RARE LIVER) and the International Autoimmune Hepatitis Group, and our position paper is the result of this discussion process, weighing different arguments and arriving at consensus.
We read with great interest and appreciation “Second-line and third-line therapy for autoimmune hepatitis”, which was published recently in the Journal of Hepatology.1 The authors provide clinically useful recommendations for the treatment of patients with autoimmune hepatitis (AIH) who do not respond to or cannot tolerate standard treatment.
Studying the effect of prednisone or prednisolone in AIH was the subject of a number of randomised trials in the 1960s and 1970s, some of the earliest randomised trials in medical history, with a clear result across all trials: steroid therapy improves survival impressively.
There is still some debate regarding the optimal dosage of these drugs, and to what extent the azathioprine dose should be increased in order to be able to withdraw steroids (vs. lower dose combination therapy).
There is consensus that azathioprine should be added within 2 weeks of starting steroids, if total bilirubin levels are <6 mg/dl, and that azathioprine should be the backbone of any maintenance therapy.
The exact titration of doses during maintenance therapy, and the decision on whether to use higher doses of azathioprine as a possible monotherapy or low-dose prednisone (e.g. 5 mg/day) combined with low dose azathioprine (around 1 mg/kg body weight) depend on individual risk factors and preferences, and should be decided together with the patient. There is also consensus that relapse rates after withdrawal of therapy are extremely high, and withdrawal should only be attempted in patients who have been in stable remission on low-dose therapy for at least 2 years.
Liver biopsy before treatment withdrawal has been recommended by many guidelines on AIH, but is considered optional in the EASL Clinical Practice Guideline. Due to the limitations of liver biopsy with respect to sampling errors, and due to the increasing evidence for biochemical remission as a reliable predictive marker, liver biopsy has become less important in the assessment of remission. Patients with alanine aminotransferase (ALT) levels below half the upper limit of normal in combination with IgG levels below 12 g/L have a very good chance of successful treatment withdrawal
and therefore probably do not need liver biopsy before weaning of immunosuppression. The main role of liver biopsy on follow-up is to distinguish remaining AIH activity from other causes of elevated liver enzymes (drug toxicity, associated non-alcoholic steatohepatitis [NASH] etc.), and to adapt management accordingly. Furthermore, whenever there is doubt that biochemical parameters are reliable in the assessment of remission, e.g. if IgG levels were not elevated at the beginning of treatment, or if levels of transaminases are close to the upper limit of normal, liver biopsy may also provide valuable information for patient management. Treatment withdrawal should be avoided during puberty, irrespective of the duration of remission, and it should only be considered with great caution in patients with type 2 AIH and all patients in whom the disease first manifested in childhood.
There is wide consensus that remission can be defined both biochemically and histologically.
However, this may be very closely linked to IgG levels, and has thus far not been confirmed as an independent indicator of remission. Histological remission is defined as a hepatitis activity index (HAI-score) of up to 3 out of 18, as this degree of disease activity is widely regarded as non-progressive.
A recent case series has demonstrated that patients reaching full biochemical remission not only do not progress, but even show regression of liver fibrosis, as measured by repeated evaluation of transient elastography, e.g. by Fibroscan.
On the other hand, this study (and others) have confirmed that remaining disease activity is associated with progressive fibrosis and development of cirrhosis, thus underlining the need to achieve full biochemical remission if possible.
Patients with AIH require alternatives to standard immunosuppressive treatment in two clinical scenarios: insufficient response, i.e. not achieving remission, or intolerance due to side effects.
While there is growing consensus on the standard management of autoimmune hepatitis, there is a considerable degree of uncertainty regarding how to deal with patients insufficiently responding to standard therapy, or not tolerating standard therapy. The EASL clinical practice guideline and the ESPGHAN position statement recommend mycophenolate mofetil (MMF) for those patients not tolerating azathioprine, and this also appears to be a widely held consensus beyond Europe.
While the data for MMF in patients intolerant to azathioprine are encouraging, MMF seems of little benefit in those patients not responding sufficiently to first-line therapy, but there are only limited data on this issue.
Besides establishing budesonide as an alternative drug that can induce remission with less steroid-specific side effects, this trial demonstrated the very high rate of patients not achieving biochemical remission within 6 months (using normal ALT as an endpoint). While some authors have criticised this trial for applying a rather low dose of prednisone and for tapering prednisone too strictly and too quickly in some patients, overall the treatment reflected widely used standards, and yet in both treatment arms more than 30% of patients were not in biochemical remission at month 6. Together with epidemiological studies showing increased liver-related mortality in AIH despite therapy, this trial highlights that there is still an unmet need for more effective and better tolerated therapeutic options in AIH.
Therefore, in addition to the need for new drugs and new therapeutic approaches, we need recommendations, and in due course better data, on how to deal with patients who are not responding sufficiently or are intolerant to standard therapy in AIH.
The scarcity of data, with only small studies and case series published so far (Table S1), and the wide variation of inclusion criteria and drug usage, do not currently allow for recommendations regarding which particular third-line therapy should be used in an individual patient. Multicentre studies are needed to accumulate data with sufficient power to make robust conclusions. In order to achieve this most efficiently, we need to structure, standardise and report our approach to patients with AIH in need of third-line therapy. The following position statement is the result of repeated discussions within the scientific community coordinated by the European Reference Network on Hepatological Diseases (ERN RARE-LIVER) and the International Autoimmune Hepatitis Group (IAIHG); it was developed in order to help in the management of these patients, and in order to agree on a standard process, in particular for the use of third-line therapies, that enables the collection of comparable data from treatment centres. It is hoped that, based on this treatment algorithm, such data can be collected in the future, helping us to improve management of this important patient population, with the final goal of establishing third-line treatment options based on high-quality data in adequately powered studies. This position statement will cover both adult and paediatric data and address the two key problems, sometimes intertwined: insufficient response and intolerance (Fig. 1 and 2).
The treatment aim in every patient with autoimmune hepatitis is prevention of progressive liver disease. Lack of progression, and even regression of fibrosis, is observed in more or less all patients with full biochemical remission, defined as normal transaminases and normal IgG levels.
It is uncertain whether histological proof of remission by demonstration of no or only minimal inflammatory activity (HAI <3/18) on liver biopsy is more reliable than biochemical remission. On the one hand, histology is considered the gold standard; on the other hand, due to the risk of sampling error with biopsy, biochemical remission parameters might provide a better overall picture. The two approaches can be considered complementary. For example, in patients with borderline biochemical parameters, a liver biopsy may help to assess the remaining disease activity more reliably. Furthermore, liver biopsy may be able to detect possible azathioprine hepatotoxicity or comorbidities, such as NASH, responsible for continued abnormal transaminase levels.
When full biochemical remission is achieved, histological confirmation of remission is not required. Similarly, if both IgG and transaminases show a clearly insufficient response, histological confirmation of insufficient response is also not required. For all other cases, histological
Biochemical remission of AIH is defined as normal transaminases as well as normal IgG levels and histological remission is defined as a hepatitis activity index (HAI-score) of up to 3 out of 18.
diagnosis is recommended, as the resulting treatment decisions may have important consequences, and are likely to determine the long-term treatment of this chronic disease.
While insufficient response in the initial treatment period is not well defined, and needs to be assessed individually depending on the severity of disease and comorbidities, we have agreed that insufficient response is failure to achieve full biochemical remission within the first 6 months of treatment.
In some patients with initially severe disease and a clear tendency toward improvement, achieving full biochemical remission may even take a few months longer than 6 months, but for the vast majority of patients the 6-month time point is appropriate for assessing response and should therefore be standard. Histological remission does take longer than biochemical remission, and thus histological evaluation for degree of response may need to be delayed for a further 6 months (up to 1 year) if there is uncertainty in the interpretation of biochemical response
There are two main reasons for this recommendation:
Azathioprine drug metabolism varies widely between individuals. It is the group of 6-thioguanine nucleotides (6-TGN) that contain the main active drug 6-tioguanine-triphosphate (6-TGTP), while the 6-methyl-mercaptopurine metabolites (6-MMP) are not immunosuppressive, but can reflect drug toxicity.
Non-adherence to recommended treatment is another important reason for insufficient treatment response and can be detected by absent or very low levels of all drug metabolites.
Unfortunately, 6-TGN measurements are not trivial and can only be undertaken in a few specialised laboratories. Future studies will need to show the reliability and reproducibility of measurements across different laboratories, but data from transplant immunology as well as from inflammatory bowel disease have shown the value of these measurements in patient management.
A larger, albeit retrospective study of patients with AIH from Sheffield, UK, showed that patients with 6-TGN levels below 220 pmol/8 × 108 red blood cells were more likely to experience relapse or have insufficiently suppressed disease activity.
Therefore, it appears reasonable to try to achieve such a level in all patients with insufficient response and lower levels on drug measurement. This can normally be achieved by addressing adherence to therapy and adapting the dose up to 2 mg or even 2.5 mg/kg body weight, as is customary in treatment of patients with inflammatory bowel disease.
For those patients with suboptimal levels of 6-TGN and high levels of 6-MMP, the alternative drug metabolite, the underlying problem is altered drug metabolism in the presence of good patient adherence. In these patients, again a lesson learned in inflammatory bowel disease, the combination of azathioprine with allopurinol can be very effective, as allopurinol blocks the 6-MMP pathway.
The dose required for the desired drug interaction is closer to 2–3 mg/kg body weight, thus allowing a very good safety margin. In order to achieve therapeutic levels, azathioprine then needs to be carefully increased keeping the allopurinol dosage constant, and drug levels need to be monitored along the way.
In countries in which 6-TGN measurements are unavailable, some colleagues have used mean corpuscular erythrocyte volume (MCV) as a surrogate marker for azathioprine dosage, as MCV should increase during sufficiently dosed azathioprine.
That said, this approach has not been validated systematically and can only be regarded as a simplistic alternative when drug level testing is not available. However, normal MCV levels in azathioprine-treated patients should, once other causes of microcytosis such as iron deficiency or thalassaemia have been excluded, always raise suspicion of below target drug levels, be they due to non-adherence or to altered drug metabolism.
In addition to optimising azathioprine dosing, patients with insufficient response should be re-evaluated diagnostically (Fig. 1). This entails questioning the initial diagnosis (Is this really AIH?) as well as excluding alternative and additional diagnoses. Primary sclerosing cholangitis (PSC) as well as primary biliary cholangitis (PBC) should be excluded in any patient with persisting cholestatic features. For possible PBC, tests should be done not only for antimitochondrial autoantibodies (AMA), but also for the PBC-associated antinuclear autoantibodies (ANA), anti-SP100 and anti-gp210, all of which can be tested by highly reliable immunoserological tests.
Drug-induced liver injury is another important differential diagnosis to consider, including the possibility of azathioprine-induced liver toxicity. However, the diagnosis of azathioprine hepatotoxicity is often given to patients only based on raised liver enzymes during therapy, and may be a misinterpretation of insufficient response, leading to an unnecessary change to second-line therapy.
For AIH patients not achieving remission under standard treatment, optimisation of azathioprine dosage on the basis of 6-TGN levels should be aimed at first, before starting alternative immunosuppressants.
In view of the excellent treatment efficacy of azathioprine in most patients, drug toxicity should not be simply assumed, but needs to be proven, either by liver biopsy, or by drug withdrawal associated with improvement, and re-exposure associated with worsening of liver tests. Histological features of azathioprine drug toxicity are usually quite characteristic and, at least by an experienced liver pathologist, can be well differentiated from AIH activity.
If in doubt, liver biopsy should be sent to a reference centre for re-evaluation. Furthermore, viral infections, in particular Epstein-Barr virus and cytomegalovirus infection in previously seronegative patients, must also be excluded as reasons for raised liver enzymes in patients being treated for AIH.
Non-alcoholic fatty liver disease (NAFLD) and NASH are frequent diseases in the general population, but they may be even more frequent in AIH as a side-effect of initial steroid therapy. Thus, NAFLD may develop in a patient in whom initial liver biopsy did not show any suspicious features, and should be considered in insufficient responders, especially if the liver test pattern is compatible, usually normal alkaline phosphatase (ALP), high gamma-glutamyltransferase (GGT) and ALT higher than aspartate aminotransferase (AST).
Ultrasound screening showing increased echogenicity of the liver parenchyma, as well as high values of ‘controlled attenuation parameter’ (CAP) on Fibroscan can help in the diagnosis – and again liver biopsy may be required to make the final diagnosis and to try to differentiate the relative role of NAFLD/NASH and AIH activity in explaining the laboratory values.
Relapse during maintenance therapy may also be considered a version of insufficient response but should be viewed slightly differently depending on the reasons for relapse. The most common form of relapse is due to lowering of maintenance therapy down to a level lower than required in the individual patient. Dose adaption and perhaps a transiently higher steroid dose, depending on the degree of relapse, will solve the problem in the majority of patients, who can then be managed on acceptable levels of standard maintenance therapy. Relapse may also be due to non-adherence; this problem is particularly common in adolescents and young adults but can occur in any age group. Patient preferences and the risks and benefits of therapy should be discussed in these patients; psychological help is sometimes necessary. Repeated relapses despite adequate maintenance therapy and adherence also represent insufficient response.
In patients with insufficient response, in whom alternative and additional liver disease has been excluded, standard drug therapy should be intensified taking into consideration disease activity, comorbidities, and drug-related side effects (Fig. 1). Third-line therapy should be attempted in patients who are adherent to therapy, and in whom the disease is active and presumably progressive despite intensified therapy; in such cases, third-line therapy should be attempted according to one of the protocols given in Table 1. In non-adherent patients, psychological support and interventions to improve adherence are to be preferred, which may or may not also include third-line therapy. As an alternative to starting third-line therapy directly after unsuccessful intensification of first-line treatment, mycophenolate mofetil (MMF) can be considered as a second-line drug, although its effectiveness in poor responders to first-line therapy seems to be limited.
Fibroscan can be very helpful in assessing disease progression: Fibroscan measures the combination effect of inflammatory infiltration in the liver and fibrosis, and should thus improve upon remission induction.
The reasons for starting third-line therapy should be recorded, its pros and cons discussed with the patient, and the results of this discussion recorded. A liver biopsy before the start of third-line treatment is recommended in order to prove the necessity of third-line therapy, to exclude alternative diagnoses, and to have detailed information on disease activity (grading) and fibrosis (staging) prior to starting these experimental therapies. Patient (and in children also parents') preferences need to be taken into consideration when deciding on third-line therapy, both on initiation and on the drugs chosen. The recommendations for patients with insufficient response are summarised in Fig. 1.
Table 1Third-line therapy in autoimmune hepatitis: Expert opinion and suggestions for standards.
Dose - Adults
Dose - Children
0.1 mg/kg twice daily, or prolonged-release formulation of tacrolimus in lower dose. Serum trough levels <8 ng/ml
0.05 mg/kg/day. Initial serum trough levels 6–8 ng/ml, tapering to 3–5 ng/ml after full biochemical remission has been achieved.
The best studied alternative. Variable and generally lower doses have been applied in retrospective and prospective studies (e.g. 0.5–6 mg/day or 2–3 mg twice daily). Several of these studies have reported a significant, but insufficient effect (full remission was not achieved). Hence, by expert opinion, we recommend standardisation of the dose at a higher serum trough level as given in this table and tapering the trough levels after remission has been achieved. Renal function (eGFR) should be assessed prior to and during treatment; consider dose reduction in cases of eGFR reduction >25%.
4 mg/kg twice daily to initial trough levels at 200–250 ng/ml, tapering to trough levels <120 ng/ml after full biochemical remission has been achieved.
Doses of 2–5 mg/kg/day were assessed in a small open-label clinical trial. A dose of 2.5 mg/kg twice daily is frequently advised for non-transplant indications (e.g. nephrotic syndrome) and tapering the trough levels after remission has been achieved. Renal function (eGFR) should be assessed prior to and during treatment; consider dose reduction in cases of eGFR reduction >25%.
5 mg/kg/day; at week 0, 2, 6, and every 4–8 weeks thereafter
The proposed dose corresponds to standard recommendations for other indications (e.g. inflammatory bowel disease, ankylosing arthritis). Maintenance therapy in AIH appears to usually require a 4-week interval contrary to most IBD patients
1,000 mg at week 0 and 2, to be repeated whenever transaminases rise (e.g. after 6–12 months)
Surveillance of CD20+ B-cells is recommended. Supplement with immunoglobulins may be necessary. Paediatric data limited to single case reports.
7.5–15 mg per week
10 mg/m2 per week
Data are scarce and limited to case series. Some case reports associate development of autoimmune hepatitis with ongoing methotrexate therapy for other indications.
1–1.5 mg/kg/day or pulse therapy 1 g i.v. every 4 weeks
Drug intolerance precludes the use of azathioprine in a fair proportion of patients, estimated to be between 3% and 5%. Reliable prospective studies of the true incidence of azathioprine intolerance are missing. It is an idiosyncratic reaction usually manifesting within the first 2 weeks of therapy, and clinically characterised by general malaise and nausea, often associated with any of the following symptoms: fever, diarrhoea, muscle and body pain, vomiting. Symptoms may mimic acute gastroenteritis, which is a differential diagnosis. Symptoms typically subside within 2 to 3 days of stopping treatment and resume rather more quickly upon re-exposure. Many colleagues will attempt a trial of re-exposure to be certain that it really is azathioprine intolerance, and the first recommendation is to undertake such an attempt by switching from azathioprine to 6-mercaptopurine (6-MP; Fig. 2), thus limiting possible intolerance to 6-MP and its metabolites and avoiding the pre-drug azathioprine. 6-MP is the first metabolite of azathioprine on the way to its active agent 6-TGN.
6-MP is thus just as effective a drug as azathioprine, and in some countries the first drug of choice in this drug class of purine analogues. As only azathioprine, and not 6-MP, is licensed for use as an immunosuppressive agent in most countries, and as azathioprine tablets are available in a wider variety of doses in most countries, most physicians prefer azathioprine, but from a pharmacological point of view, 6-MP is just as effective. The problem of inadequate dosing alternatives may also sometimes hamper optimal therapy in very young children, as azathioprine fluid, which allows exact dosing, is not available everywhere. At the same time steroid maintenance therapy can be problematic in young children due to its negative growth effects, hence some specialists choose MMF as an alternative first-line therapy due to its wider availability in liquid form, allowing individual dosing. Systematic data are lacking, and drug preference should be individualised depending on the steroid dose required and possible side effects of MMF.
It therefore appears prudent to attempt a challenge with 6-MP in a patient suspected of being intolerant to azathioprine (Fig. 2); patients with only milder, mainly gastrointestinal symptoms due to azathioprine are particularly likely to tolerate 6-MP therapy. If symptoms of intolerance recur, this drug class should probably be avoided completely, even though there are some reports of better tolerance with thioguanine, an alternative agent from this drug class.
If 6-MP is tolerated, it should be used, and its dose adapted based on optimal 6-TGN blood levels (starting usually at a dose of mercaptopurine of 0.5–1 mg/kg body weight).
In patients intolerant to both azathioprine and 6-MP, MMF should be used as a second-line drug, starting at a dose of 500 mg twice daily and increasing to 1 g twice daily; in children, MMF should be started at a dose of 5 mg/kg body weight twice daily up to a maximum of 20 mg/kg body weight twice daily (Fig. 2). MMF seems to be generally better tolerated than azathioprine, but does cause gastrointestinal symptoms in a fair proportion of patients, as well as impaired wound healing, increasing the risk of organ perforation, for example, in peptic ulcer disease. Furthermore, MMF is strictly contraindicated in pregnancy, and for men reliable contraception is recommended because the risk of genotoxicity cannot be completely ruled out.
Long-term family planning needs to be discussed with every patient at reproductive age before switching to MMF. This way treatment changes can be avoided if family planning becomes relevant after starting MMF. This is particularly relevant in adolescents and young adults, who may have been put on MMF at some stage, and in whom planned or unplanned pregnancies may lead to a complicated situation. The effectiveness for AIH appears to be good, and it has been reported that up to two-thirds of patients intolerant to azathioprine achieve full biochemical remission using an MMF regimen.
In paediatrics and adolescents, growth development, problems with adherence and availability of liquid formulations of immunosuppressive drugs for small children can be additional reasons to adapt the treatment strategy.
Side effects of corticosteroids may also be severe and be considered to cause intolerance in some patients. During adolescence, non-adherence to steroid-based therapy may be a major problem, both due to real as well as to perceived or feared potential side effects. As high-dose steroids are usually only required for short periods, dose adaptation should first be tried in this situation, preferably by optimising azathioprine treatment as rapidly as possible in order to spare steroids. Budesonide can be considered as an alternative in patients with severe steroid side effects, but is contraindicated in all patients with cirrhosis, in whom the first-pass effect is impaired, increasing the risk of steroid-related side effects and portal vein thrombosis.
While being an option for remission induction, budesonide is more difficult to taper due to its short half-life and limited dose availability on the market. Both prednisolone and budesonide can cause considerable long-term side effects, and in view of the better long-term effectiveness of systemic immunosuppressants, steroids should in any case not be the mainstay of therapy in AIH.
Nonetheless, in patients intolerant to azathioprine, and also intolerant to MMF, steroid monotherapy may be a valid treatment alternative, if bone mineral density is good and remains good, and if the prednisolone dose required can be kept at a maximum of 10 mg/day (in children probably at a maximum of 2.5–5 mg/day). This approach is summarised in Fig. 3 and should be discussed as a possible option with the patient before starting an alternative third-line therapy. In this discussion, it needs to be mentioned that depression is quite frequently observed in patients with AIH, and this seems to be closely associated with steroid use, thus hinting towards a pathogenetic role of steroids in AIH-associated depression.
Therefore, existing depression and depression risk should be assessed carefully, and psychological support may be required in these patients. An algorithm summarising all recommendations is given in Fig. 3.
A number of drugs have been reported to help in the treatment of patients with AIH who do not respond sufficiently to first- and second-line therapy (Table S1). Various schedules of third-line drugs have been used, often depending on local expertise as well as reimbursement rules. To study and evaluate the effectiveness of each third-line protocol, a collaborative effort between all expert centres managing such patients is required. We suggest that patients should only receive third-line therapy on the basis of the algorithm presented in Fig. 3. Patients should be stratified according to why they are receiving third-line therapy: because of insufficient response to standard therapy (expected to be difficult-to-treat), or because of intolerance to standard therapy (expected to be easier-to-treat). Furthermore, third-line protocols should be as standardised as possible, and therefore we have listed the possible regimens including dosing and dosing intervals in Table 1. Results of case series of patients treated according to these protocols should be shared and the results published. Collection of data can be undertaken via the ERN RARE-LIVER.
This paper does not give firm guidance on the choice of third-line therapy, nor on the question of whether third-line drugs should be combined with standard drugs, or given as an alternative, as no reliable data for such a recommendation are available. However, experience suggests that for patients who are intolerant to first- and second-line treatment, a well-tolerated single third-line drug will probably be sufficient to control liver inflammation. However, for cases with insufficient response under first- and second-line therapy, double or even triple immunosuppression may frequently be needed to induce remission, at least during the initiation of the third-line drug. But the aim of starting third-line treatment is always that the third-line drug will maintain remission on its own. However, for future evaluation of the effectiveness of third-line drugs, it is important to stratify results according to the reason (intolerance or insufficient response) for giving third-line drugs, and to record if (and in what form) combination therapy has been applied. When initiating third-line therapy, in addition to checking and proving the need according to the above recommendations, re-evaluation of the overall health status of the patient is required, not only in order to evaluate the indication for third-line therapy, but to also determine and limit the risks. First of all, in older patients and patients with comorbidities, overall life expectancy and quality of life need to be assessed: Is AIH really the most important disease in this patient, and would progressive disease limit life-expectancy and quality of life more than the side-effects of higher dose standard therapy? Secondly, particularly in children, growth and developmental assessments need to be performed regularly to determine the risks of disease, side-effects of previous therapy and potential risks of alternative third-line therapies. Body weight should be tested in all patient populations, and additional measures for weight control may be required in overweight patients, particularly in those still requiring steroids. Furthermore, cardiovascular and renal risk factors should be checked, as many of the third-line drugs have relevant cardiovascular and renal side-effects. Chronic infections need to be excluded, particularly tuberculosis in the case of anti-TNF therapy. Finally, vaccination status should be assessed and (re-)vaccinations performed as required, preferably prior to commencing third-line therapy.
No clear preference can be given for single immunosuppressants as third-line treatment since comparative studies are missing and none of the therapies used are approved, yet. However, standardised approaches are needed to make studies comparable and to enable comparative analysis of efficacy in the future.
Checking the need for third-line therapy by carefully re-assessing disease history, patient adherence, patient preferences and comorbidities will benefit difficult-to-treat patients. Applying the above recommendations should not only lead to better patient care but should also form the basis for scientific progress in evaluating the utility and risks of third-line therapies in AIH.
Special recommendations for paediatric and adolescent patients with AIH
While generally speaking AIH is probably the same disease across all age groups, a number of special considerations for paediatric patients need to be stressed. Children and adolescents with AIH may differ from adult patients in several ways: they often present with a cholestatic variant of AIH, certain autoantibodies are more frequently found in children, such as anti-liver-kidney microsome antibodies or anti-liver cytosol type 1,
and the questions of growth and development, as well as psychosocial aspects and questions of adherence, may present special challenges in paediatric care.
Up to about 50% of children presenting with features of AIH may later be found to have underlying cholestatic liver disease, termed by some autoimmune sclerosing cholangitis or AIH/PSC overlap syndrome. Therefore, sclerosing cholangitis should be excluded by cholangiography (magnetic resonance cholangiopancreatography), and possibly by repeating liver biopsy.
As typical findings of sclerosing cholangitis may be subtle and can be missed by both methods, follow-up cholangiography and sometimes also follow-up liver biopsy may be required, and are strongly recommended for all children and adolescents failing to reach remission. These children should also be screened for underlying inflammatory bowel disease by measuring faecal calprotectin followed by a colonoscopy if elevated.
Other differential diagnoses must also be considered in an age-related fashion: In toddlers, careful evaluation of metabolic disorders must be undertaken. In both children and young adults, Wilson's disease must be excluded by measuring ferroxidase and 24-hour urine copper excretion, by slit-lamp examination for Kayser-Fleischer rings, by genetic analysis and, if in doubt, by measuring liver tissue copper content quantitatively in a liver biopsy sample.
Previous studies on the treatment of paediatric AIH include only one randomized trial.
That is why treating this patient group is mainly based on case reports, observational and retrospective studies, adult studies, and experience. Despite these difficulties, a paediatric scoring system for autoimmune liver disease and a treatment algorithm have been proposed by The European Society for Paediatric Gastroenterology Hepatology and Nutrition (ESPGHAN), which may help in the management of paediatric AIH.
If remission is not achieved, monitoring of azathioprine metabolites should rely on the 6-MMP/6-TG ratio being below 4 rather than on sole 6-TGN levels. On their own, 6-TGN levels appear to only poorly correlate with biochemical remission in paediatrics.
Although studies on the combination of allopurinol and low-dose azathioprine in children with AIH have not been published, it seems reasonable to use the same approach in children as in adults, with special attention to 6-MMP/6-TGN ratio.
Relapse during maintenance therapy in childhood and adolescence requires a special diagnostic and therapeutic approach: one reason may be accidental undertreatment, because as children grow, continued re-evaluation and adaptation of dosage per body weight is needed. Non-adherence is a challenge in any age group, but is much more commonly seen and at the same time more complex in adolescence. It is probably the most common cause of relapse during this phase in life. Managing adherence well must be considered a multidisciplinary task.
In smaller children, non-adherence may also occur, for example if the child dislikes the medication or cannot swallow the tablet. Furthermore, psychosocial problems of parenting may also be responsible for non-adherence. In type 2 AIH, representing about 10% of the paediatric patients with AIH,
repeated relapse is not uncommon, and second- or third-line treatment is frequently needed.
Because of the longer expected lifespan of a child with AIH, the long-term effects of immunosuppressive treatment should be taken into consideration in a paediatric patient with AIH, and therapy may need to be adapted in different life phases accordingly. Intolerance and side effects of therapy may have a more severe impact on adherence and quality of life in paediatric patients. Studies on 6-MP in paediatric AIH are lacking, however, data from acute lymphoid leukaemia and inflammatory bowel disease in childhood support the safety of 6-MP in this age group. Therefore, 6-MP is also an option in children if azathioprine is not tolerated. MMF is available as a liquid formulation, is often used for transplantation in children and its dosing can be easily regulated in small children. In general, if monotherapy with prednisolone on a higher dose appears to be required to reach and maintain remission, other alternatives should be discussed with the parents and child, in particular in adolescents. Growth and bone mineralisation should be monitored closely, and third-line treatment may be considered if side effects on standard therapy become worrisome, even if the child is in remission. Non-adherence due to the cosmetic side effects of prednisolone can be a major issue during adolescence and can also be considered a form of drug intolerance leading to second- and third-line therapy. Patient involvement in the decision process seems important. Furthermore, as patients transition into adult life, a return to standard therapy should probably be attempted.
The number of paediatric studies on third-line therapy is very low and the number of patients included small (Table S1). This highlights the need for multicentre studies and close collaborations together with adult hepatologists. Two observational paediatric studies on calcineurin inhibitors indicate that remission can be achieved in about 75% of cases within 6 months and that well-monitored usage of tacrolimus is safe in children.
However, we lack long-term data, which are urgently needed due to the possible toxic effects of these drugs. Other third-line drugs (Table S1) in paediatrics can only be supported by case reports or adult data. The lack of data in this area underlines the need for prospective databases and multicentre paediatric studies.
The authors thank EASL and ESPGHAN for supporting ERN RARE-LIVER in their efforts for the care of patients with rare liver diseases, and to the European Commission and the Free and Hanseatic State of Hamburg for financial support for the ERN RARE-LIVER.