Refractory autoimmune hepatitis in children: Considerations for assessment and management

Abstract

Refractory autoimmune hepatitis (AIH) is defined as intolerance of or unresponsiveness to standard immunosuppression and occurs in 10%-20% of children with AIH. Lack of response or slower than expected response to induction of remission with steroids, despite good compliance, might be the first clue to refractory AIH. Refractoriness to treatment is associated with an 11.7 times higher risk for liver transplantation or death due to liver disease. The first and foremost consideration for the management is to assess compliance with treatment. It is then important to re-evaluate the diagnosis, assess alternative aetiologies which can mimic the clinical, serological, and histological features of AIH, and address the presence of extra-hepatic co-morbidities. It is important to consider the specific clinical situations, previous therapy, and prior adverse effects before deciding on the most appropriate treatment regimen in refractory AIH. Consideration also should be given to compliance with previous therapy, need for drug level monitoring, growth potential, available formulations, route of administration of medication, and children’s and families’ preferences before deciding on the therapy. Treatment should be decided and monitored only in specialized hepatology centers.

Key Words: Autoimmune hepatitis; Refractory; Remission; Treatment; Non-compliance

Core Tip: Treatment of autoimmune hepatitis seeks to induce and maintain remission, prevent progression, and potentially reverse fibrosis. In the 10%-20% of children unresponsive to steroids and azathioprine, adherence to medication and diagnosis should be reassessed and mimicking diseases excluded. Mycophenolate mofetil is the preferred second-line agent. Further therapy should be individualized, with adjustments made only after 6-12 months.

INTRODUCTION

Autoimmune hepatitis (AIH) is a chronic immune-mediated inflammatory disease caused by loss of tolerance to hepatocyte autoantigens, leading to an altered immunological response and resulting in progressive liver injury and fibrosis[1,2]. All age groups can be affected. Clinical presentation in children is highly variable, ranging from asymptomatic cases, nonspecific symptoms, complications of cirrhosis and/or portal hypertension, acute severe hepatitis, or acute liver failure. The diagnosis of AIH is based on elevated aminotransferase levels and immunoglobulin G (IgG), in the presence of compatible histological features, and on the exclusion of viral, hereditary, metabolic, cholestatic, and drug-induced diseases that may resemble AIH[2,3]. AIH is classified into two types according to the antibody profile: Type 1 (AIH-1), with positivity for anti-nuclear antibody (ANA) and/or anti-smooth muscle antibody (SMA); and type 2 (AIH-2), characterized by anti-liver–kidney microsomal antibody type 1 and/or anti-liver cytosol type 1 antibody[2,3]. The most widely recommended and preferred drugs for inducing remission and for first-line maintenance treatment are prednisolone (prednisone) and azathioprine (AZA), respectively.

RESPONSE TO TREATMENT IN AIH AND CLINICAL SIGNIFICANCE

The aim of treatment is to induce biochemical and histological remission, improve symptoms, prevent disease progression, and maintain long-term remission. Achieving remission not only halts disease progression but may lead to regression of fibrosis[4]. There is significant heterogeneity in the response to treatment. In most children, a marked reduction in transaminases (more than 50%) is seen within the first week, followed by improvement in bilirubin and coagulation parameters, and lastly, due to its longer half-life, a decrease in IgG levels[4]. IgG levels correlate well with histological activity; significant hepatic inflammation is likely to be present in 50% of patients if either aminotransferases or IgG are elevated[5]. An exception is that patients with cirrhosis in biochemical remission may have persistent elevation of IgG[2]. It may take a few months for aminotransferases to return to baseline, but an 80% decrease is usually achieved within the first 2 months[3]. Histological improvement typically lags behind clinical and biochemical improvement by 3-8 months[6]. The definitions of various treatment outcomes in AIH are detailed in Table 1. Lack of biochemical remission at 6 months is associated with a worse long-term prognosis, including progression to cirrhosis, an increased risk of portal hypertensive complications, the need for liver transplantation (LT), and mortality from liver disease[7]. The risk of developing complications is proportional to the total time spent with elevated aminotransferases[8].

Table 1 Definitions of treatment outcomes in autoimmune hepatitis.

Ref.	Outcomes	Definition
Mack et al[2], 2020; Pape et al[8], 2022	Biochemical remission; complete biochemical response	Normal aspartate aminotransferase, alanine amino transferase and immunoglobulin G levels
Mack et al[2], 2020	Histological remission	Absence of inflammation in liver tissue after treatment
Mieli-Vergani et al[3], 2018	Complete remission	Normal aspartate aminotransferase, alanine amino transferase and immunoglobulin G levels; anti-nuclear antibody/anti-smooth muscle antibody- negative or low titer (< 1:20); anti-liver-kidney microsomal antibody type 1 and anti-liver cytosol type 1 antibody - < 1:10 or negative
Mack et al[2], 2020	Treatment intolerance	Inability to continue maintenance therapy due to drug-related side effects
Mack et al[2], 2020	Incomplete response	Improvement of laboratory and histological findings that are insufficient to satisfy criteria for remission
Pape et al[8], 2022	Non-response	Failure to achieve a more than 50% reduction of alanine amino transferase within 4 weeks of treatment
Mack et al[2], 2020	Treatment failure	Worsening laboratory or histological findings despite compliance with standard therapy
Mack et al[2], 2020	Relapse	Exacerbation of disease activity after induction of remission and drug withdrawal (or nonadherence)
Mieli-Vergani et al[3], 2018	Refractory	Intolerant of or unresponsive to standard immunosuppression

REFRACTORY AIH

The European Society for Paediatric Gastroenterology, Hepatology, and Nutrition position statement[3] defines refractory AIH as patients’ intolerance of, or unresponsive to, standard immunosuppression. The term difficult-to-treat AIH includes refractory AIH as well as other clinical situations, such as acute severe hepatitis with or without liver failure, AIH in pregnancy, and AIH in patients with viral hepatitis[9,10]. The 2019 practice guidelines by the American Association for the Study of Liver Diseases (AASLD)[2] do not include a separate category for refractory AIH but instead classify these patients into three groups: Treatment failure, incomplete response, and treatment intolerance (Table 1). The general approach to managing these three categories is similar; therefore, it is practical to group them together. Refractoriness occurs in 10%-20% of children with AIH[11]. The most important predictors of treatment outcome are the speed of response to steroids and the response to first-line therapy. Aminotransferase levels should improve within two weeks of starting treatment[2]. Lack of response, or a slower-than-expected response, to steroids despite good compliance may be the first clue to refractory AIH[3,4]. Refractoriness to treatment is associated with an 11.7 times higher risk of LT or death due to liver disease[8,12]. Patients with high aminotransferase levels, elevated IgG prior to LT, and/or moderate-to-severe inflammatory activity in the explant liver have a higher risk of recurrent AIH after LT[13].

APPROACH TO REFRACTORY AIH

Prompt biochemical response, with a steady and consistent decline in aminotransferases on standard immunosuppression, is the norm and has been suggested as a diagnostic criterion for AIH. Therefore, the first and most important step in suspected refractory AIH is to assess compliance with treatment (Table 2). Next, it is essential to re-evaluate the diagnosis and to consider alternative or overlapping diagnoses before escalating immunosuppressive therapy. The presence of hepatic or extrahepatic comorbidities that may cause derangement of aminotransferases should also be assessed. Only after these possibilities have been excluded should a diagnosis of refractory AIH be made. Figure 1 details an algorithm for the evaluation of children not responding to or intolerant to first-line treatment.

Figure 1 Algorithm for the evaluation of children with autoimmune hepatitis not responding to or intolerant to first-line treatment. AIH: Autoimmune hepatitis; ESPGHAN: European Society of Pediatric Gastroenterology, Hepatology and Nutrition; ANA: Antinuclear antibody; SMA: Smooth muscle antibody; LKM-1: Liver kidney microsomal antibody type 1; LC-1: Liver cytosol antibody type 1; SLA: Soluble liver antigen; pANCA: Perinuclear antineutrophil cytoplasmic antibody; IgM: Immunoglobulin M; DNA: Deoxyribonucleic acid; PCR: Polymerase chain reaction; RNA: Ribonucleic acid; TSH: Thyrotropin; MRCP: Magnetic resonance cholangiopancreatography.

Table 2 Considerations for assessing patients with refractory autoimmune hepatitis.

	Considerations for assessing patients with refractory autoimmune hepatitis
1	Confirm compliance/rule out non-adherence
2	Re-evaluate the diagnosis of AIH and evaluate for liver diseases which can mimic biochemical and/or histological features of AIH
3	Assess for extra-hepatic co-morbidities

AIH: Autoimmune hepatitis.

Step 1: Confirm compliance with medication regimen/rule out non-adherence

The most important consideration in refractory AIH is confirming adherence to therapy. Non-adherence to treatment must be ruled out in all who do not respond to first-line treatment before diagnosing and evaluating for resistant AIH[2]. The risk of non-adherence while on chronic immunosuppressive medications is higher in children and adolescents than in adults[14,15]. Assessment of adherence can be challenging, as there is no gold-standard tool. A combination of direct and indirect methods is recommended[16]. Inaccurate estimation of medication adherence can result in falsely diagnosing resistant AIH, leading to unnecessary diagnostic testing, physical and emotional distress, increased risk from invasive procedures, including liver biopsy (LB), and inappropriate intensification of immunosuppressive therapy.

Adherence assessment should include a combination of self-report questionnaires, structured interviews, pharmacy pick-up/refill rates, and therapeutic drug monitoring[16,17]. All patients on AZA should have consistent therapeutic 6-thioguanine nucleotide levels. Occasionally, directly observed therapy may be required to confirm adherence and assess treatment response. It is also important to evaluate factors contributing to non-adherence, including high medication costs, false beliefs about treatment (e.g., “medications are harmful to the liver”) held by the child or parent, cultural influences, psychiatric disorders, social adjustment difficulties, psychological distress, behavioural problems, poor family functioning, and intolerance due to side effects of immunosuppressive therapy (e.g., weight gain, abdominal pain)[14,15].

Intolerance to AZA: AZA is an antagonist of purine metabolism, inhibiting RNA and DNA synthesis and thereby affecting more rapidly dividing cells, including lymphocytes[1]. Intolerance is defined as the inability to continue therapy due to side effects. Mild nausea is the most common side effect. Significant side effects - most commonly, gastrointestinal - requiring discontinuation of therapy occur in 5%-15% of children[1,2]. AZA hypersensitivity syndrome, characterised by systemic symptoms such as fever, myalgia, rash, arthralgia, and nausea, may develop within the first few days to weeks after starting treatment and occurs in a minority of patients[1]. Other rare but important side effects requiring discontinuation of AZA include bone marrow suppression, hepatotoxicity, acute cholestatic hepatitis, pancreatitis, and opportunistic infections[1,2]. The most important dose-limiting adverse effect is bone marrow suppression, which is idiosyncratic and may be exacerbated by concomitant cytopenia due to hypersplenism.

Alternative first-line medications for AIH: Steroids, most commonly prednisolone, are used as first-line therapy either alone or in combination with AZA[1]. Steroids are the backbone of AIH treatment, as they are highly effective in inducing biochemical remission. Strict monitoring of serum transaminases, with rapid weekly tapering of the steroid to a minimum dose (most commonly 5 mg once daily) along with staged introduction of AZA, is the most common practice in paediatrics. It is important to taper the steroid to the lowest effective dose as quickly as possible, as systemic side effects can lead to reduced health-related quality of life (HRQoL) and poor adherence. These side effects include weight gain, cushingoid features, insomnia, osteoporosis, hyperglycaemia, hypertension, psychosis, hypertrichosis, and acne. Mycophenolate mofetil (MMF) has also been used as first-line maintenance therapy in children, but this practice is not common and is not recommended by most international guidelines[2,3].

Step 2: Re-evaluate the diagnosis of AIH and evaluate for liver diseases that can mimic biochemical and/or histological features of AIH

All children with refractory AIH should undergo a thorough re-evaluation to confirm the initial diagnosis and reconfirm the diagnosis of AIH. This should include a detailed history, complete clinical examination, review and reassessment of the autoantibody profile, histopathology, and evaluation for concomitant hepatic or extra-hepatic co-morbidities. The European Society of Paediatric Gastroenterology, Hepatology and Nutrition scoring system to classify AIH as probable or definite (Table 3) can be useful. A score of ≥ 7 is consistent with probable AIH, while ≥ 8 would be classed as definite AIH. The histopathology of the initial LB should be re-evaluated by an experienced liver pathologist. It is important to repeat the LB before changing the immunosuppressive regimen or escalating immunosuppression (Figure 2). This is to assess for an alternate aetiology, confirm histological features consistent with AIH, quantify the degree of inflammation, assess fibrosis stage, and plan the treatment regimen. The clinical, laboratory, radiological, and histological assessments should also evaluate for liver diseases which can mimic biochemical and/or histological features of AIH as detailed below (Table 4).

Figure 2 Histopathology of autoimmune hepatitis in a 13-year-old patient with good compliance on azathioprine but with mildly elevated aminotransferases (55-70 IU/L; upper limit of normal 40 IU/L) and immunoglobulin G (17 g/L; upper limit of normal 16 g/L). A and B: Classical features of autoimmune hepatitis in the index biopsy showing lymphoplasmacytic infiltrate affecting both portal and perivenular regions; C: Interface hepatitis with emperipolesis and hepatocyte resetting. Hematoxylin-eosin original magnifications, × 100, × 200, × 200; D and E: Follow-up biopsy demonstrating ongoing portal infiltrate with an interface component; F: Resolution of acute necro-inflammatory and perivenular inflammation, leaving bridging fibrosis. Hematoxylin-eosin original magnifications × 100, × 100, Haematoxylin van Gieson × 40.

Table 3 Scoring system by the European Society of Paediatric Gastroenterology, Hepatology and Nutrition for the diagnosis of juvenile autoimmune hepatitis.

Criteria	Titre or level	Points
Positive autoantibodies1,2
Antinuclear antibodies (ANA) or anti-smooth muscle	≥ 1:20	+ 1
antibodies (SMA)	≥ 1:80	+ 2
Anti-liver kidney microsomal type 1 (anti-LKM1)	≥ 1:10	+ 1
	≥ 1:80	+ 2
Anti-liver cytosol type 1 (anti-LC1) antibody		+ 2
Anti-soluble liver antigen (anti-SLA)		+ 2
Anti-perinuclear neutrophil cytoplasmic antibody (pANCA)		+ 1
Immunoglobulin G (IgG)
	> ULN	+ 1
	> 1.2 × the ULN	+ 2
Histological features3
	Compatible with AIH	+ 1
	Typical of AIH	+ 2
Absence of viral hepatitis, Wilson disease, metabolic associated steatotic liver disease, drug exposure		+ 2
Presence of extrahepatic autoimmunity		+ 1
Family history of autoimmune disease		+ 1
Cholangiography
	Normal	+ 2
	Abnormal	- 2

¹Antibodies measured by indirect immunofluorescence on a composite rodent substrate (i.e., kidney, liver or stomach).

²Addition of points achieved for antinuclear antibody, smooth muscle antibody, anti-liver kidney microsomal antibody type 1 antibodies, liver cytosol antibody type 1, and anti-soluble liver antigen cannot exceed a maximum of 2 points.

³Typical histological features are those contained in the revised diagnostic criteria of the International Autoimmune Hepatitis Group, principally interface hepatitis.

AIH: Autoimmune hepatitis; ANA: Antinuclear antibody; SMA: Smooth muscle antibody; LKM-1: Liver kidney microsomal antibody type 1; LC-1: Liver cytosol antibody type 1; pANCA: Perinuclear antineutrophil cytoplasmic antibody; IgG: Immunoglobulin G; ULN: Upper limit of normal.

Table 4 Differential diagnosis in refractory autoimmune hepatitis.

Differential diagnosis in refractory autoimmune hepatitis
Primary sclerosing cholangitis
Viral infections
Wilsons disease
Hereditary hemochromatosis
Coeliac disease
Alpha one antitrypsin deficiency
Inflammatory bowel disease
Metabolic dysfunction-associated steatotic liver disease
Drug-induced autoimmune-like hepatitis
Genetic defects of immune regulation
Extrahepatic immune disorders (thyroid disorders, systemic lupus erythematosus)

PSC/overlap syndrome: Lack of response to immunosuppressive therapy in AIH might be due to the development or progression of biliary features. This is most frequently observed in AIH-1 but can also rarely occur in AIH-2[18,19]. Long-term follow-up data, as reported by Warner et al[20], showed that approximately 20% of children with childhood onset AIH-1 developed biliary features by adulthood; 50% of these showed a phenotypical transition to primary sclerosing cholangitis (PSC), while the remainder developed an overlap syndrome.

These patients with AIH initially respond to immunosuppressive therapy with excellent biochemical response, but the progression of biliary disease is not halted and may become unresponsive to immunosuppressive medications with the change of phenotype (Figure 3)[18,19]. It has been postulated that even biliary ductular reaction considered insignificant in the context of the overall histological features in the initial LB might be an indicator of future biliary disease progression[20]. AIH with associated biliary features at the time of initial diagnosis, for example presence of copper-associated protein, is more likely to show progression to a biliary phenotype[18]. The following are the criteria for diagnosis of AIH–PSC overlap syndrome in children - the presence of typical features of AIH, absence of anti-mitochondrial antibodies, imaging features consistent with PSC on endoscopy or magnetic resonance cholangiopancreatography, or LB showing consistent histological features[2].

Figure 3 Sequential liver biopsies in a patient with refractory autoimmune hepatitis with a change of phenotype to primary sclerosing cholangitis over time. A: Index biopsy showing features compatible with autoimmune hepatitis; portal plasma cell-rich infiltrate with interface hepatitis; B: Perivenular infiltrate with hepatocyte dropout [A and B: Hematoxylin-eosin (HE) × 200]; C: Early bridging fibrosis was present at time of index biopsy (Masson trichrome × 40); D and E: Follow-up biopsy after 10 years; resolution of inflammation (D) (HE × 40); subtle biliary features with an “onion skin” sclerosing bile duct lesion (E) (HE × 200); F: Sclerosing bile duct on collagen stain (Haematoxylin van Gieson × 200).

The 2019 AASLD practice guidelines[2] recommend that AIH–PSC overlap syndrome should be considered in all patients with AIH and inflammatory bowel disease (IBD), and those non-responsive to conventional steroid therapy. Persistently raised gamma-glutamyl transferase values can be an initial clue to a change of phenotype. Gamma-glutamyl transferase levels are likely to be significantly higher in children with a phenotype change to PSC[18]. Assessment of the biliary system with ultrasonography or preferably magnetic resonance cholangiopancreatography might provide further diagnostic clues. LB is essential in this group to decide whether patients are likely to benefit from continuation or escalation of immunosuppression, or even discontinuation of immunosuppressive therapy. Ursodeoxycholic acid or oral vancomycin therapy might be indicated.

DI-ALH: Is defined as liver injury with laboratory and/or histological features that may be indistinguishable from those of AIH[21]. Drug-induced autoimmune-like hepatitis (DI-ALH) has been reported in 9% of the total cases of drug-induced liver injury[22]. The clinical manifestations most commonly occur within 3 months of drug exposure, but can appear after prolonged intervals, even > 500 days[21,23]. DI-ALH can have high IgG levels, positive ANA, anti-SMA, anti-liver–kidney microsomal antibody type 1 antibody, and histological features mimicking the morphological pattern of AIH[21,22]. The medications most associated with DI-ALH are nitrofurantoin, hydralazine, interferon, imatinib, adalimumab, infliximab, and statins[21,22]. Herbal, dietary supplements, and vaccines, especially the severe acute respiratory syndrome coronavirus 2 vaccine, are other reported associations[21,22]. DI-ALH after severe acute respiratory syndrome coronavirus 2 vaccination was reported after the second dose and was associated with elevated IgG and autoantibodies in 57%[21]. The liver injury in DI-ALH usually resolves spontaneously within six months or after starting corticosteroids once the causal drug is withdrawn[21,22]. In all children with refractory AIH, a thorough pharmacological history for the preceding 6-12 months with exposure to drugs, vaccines, herbal, and dietary supplements is essential. The strength of the temporal relationship of any drug, treatment duration, and the longest latency of clinical presentation should be explored in detail. Other features that should raise suspicion of DI-ALH include dermatological changes, renal injury, and a history of previous drug-induced liver injury[22].

Wilson’s disease: Children with Wilson’s disease (WD) can have clinical, immunological, and histological features indistinguishable from AIH[24]. Very rarely, WD and AIH can co-exist, necessitating treatment for both conditions[25]. IgG, ANA, and anti-SMA can show significantly elevated titres in WD[26]. The histological features can also mimic AIH with parenchymal mononuclear inflammatory infiltrate and interface hepatitis[24]. All children with refractory AIH should be evaluated for WD with ceruloplasmin, serum copper, 24-hour urine copper estimation, and testing for pathogenic mutations in the ATPase copper-transporting beta (ATP7B) gene. A liver tissue specimen for parenchymal copper concentration should be taken (if not done earlier) when performing LB in refractory AIH.

Metabolic dysfunction-associated steatotic liver disease: The incidence of metabolic dysfunction-associated steatotic liver disease (MASLD) is increasing worldwide, and the incidence in AIH is likely to be similar to the general population[27]. MASLD can be present at the diagnosis of AIH or develop later. MASLD or metabolic dysfunction-associated steatohepatitis (MASH) can be associated with elevated liver enzymes. Autoantibodies and immunoglobulin [mostly immunoglobulin A (IgA)] can be elevated in some patients with MASLD. Histological findings of MASLD/MASH can be present in 17%-30% of patients with AIH[27,28]. Most studies have reported that there are no significant differences in treatment outcomes of AIH with or without MASLD[27,28]. However, the presence of MASH in patients with AIH has been reported to be associated with non-response, though this was a small cohort, and studies in larger groups are needed[27]. Children with refractory AIH should be assessed for hepatic steatosis (imaging and LB) and the presence of metabolic risk factors (diabetes, hypertension, obesity).

Viral infections: Viral hepatitis B and C should be excluded in all children prior to starting treatment for AIH. But children on treatment for AIH can contract hepatitis B or C. Reactivation of the hepatitis B virus is a known complication of immunosuppressive medications and can lead to a spectrum of manifestations, including acute hepatitis. Though chronic hepatitis E infection has mostly been described in solid organ transplant recipients, patients receiving immunosuppressive or immunomodulatory therapies for autoimmune diseases are also at risk. Chronic viral hepatitis can present with biochemical changes identical to AIH and can be associated with autoantibodies. Other viral infections, especially Epstein-Barr virus and cytomegalovirus infection in previously seronegative patients, must also be excluded[29]. All children with refractory AIH should be assessed for chronic hepatitis B, C, and E.

Genetic defects of immune regulation: Genetic defects of immune regulation have been rarely reported as a cause of refractory AIH. The classic example of a genetic defect of immune regulation is autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy, caused by biallelic mutations in the autoimmune regulator (AIRE) gene. The classic triad of manifestations is chronic mucocutaneous candidiasis, hypoparathyroidism, and adrenal insufficiency[30]. Autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy-associated hepatitis can present with mild asymptomatic laboratory derangements through to life-threatening fulminant hepatic failure requiring LT. Hepatitis can be the initial manifestation[30]. The laboratory and histological features are identical to classical AIH. Most patients with autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy-associated hepatitis respond well to the classical immunosuppressive regimen of AIH, but a small fraction (10%) can be refractory to treatment[30]. AIH-2 has been reported in immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome, a rare disorder caused by mutations of the FOXP3 gene and characterized by autoimmune enteropathy, early-onset type 1 diabetes mellitus, thyroiditis, and eczema[31]. An autosomal dominant gain-of-function mutation in the signal transducer and activator of transcription 1 (STAT1) gene has also been reported as a cause of refractory AIH[32]. The Janus kinase/signal transducer and activator of transcription apparatus constitutes essential intracellular signalling pathways that induce phosphorylation and stimulation of various proinflammatory cytokines, interleukins, and growth factors; mutations lead to an imbalance of pro- and anti-inflammatory cytokines[32]. It is important to consider the possibility of monogenic disorders of immune regulation in children with refractory AIH, especially if < 5 years of age, anti-LKM positive, and in the presence of co-existent extrahepatic manifestations.

Step 3: Assess for extrahepatic co-morbidities

IBD: IBD can be present at the time of diagnosis of AIH or can develop over time[18]. IBD occurs in 20% of children with AIH, and the incidence is much higher (45%) in children with AIH-PSC overlap syndrome[2]. Children with histological progression of biliary features are four times more likely to have IBD compared to those without[18]. Ulcerative colitis is the most common type of IBD. Untreated IBD can be a cause of persistently deranged liver enzymes. Children with co-existent IBD can be asymptomatic or clinically silent and remain undetected for months or years unless specifically screened for[33]. Treatment with immunosuppressive therapy can also mask symptoms and signs of IBD, leading to a delay in diagnosis[34]. All children with refractory AIH should be screened for concomitant IBD with faecal calprotectin. If faecal calprotectin is elevated, colonoscopy and mucosal biopsies, with special focus on the cecum/ascending colon, should be performed[34].

Coeliac disease: Coeliac disease (CD) is more frequently seen in children with AIH than in the general population[2]. Gluten exposure, either due to non-compliance with a gluten-free diet or undiagnosed CD, can be associated with elevated transaminases in 50% of cases[35]. Transaminases will normalise only after 6-12 months of a strict gluten-free diet. Since laboratory and serological features can mimic AIH, all children with refractory AIH should be screened for CD, and compliance with a gluten-free diet should be assessed in those already diagnosed[2]. The recommended screening tests are tissue transglutaminase-IgA antibody and total IgA level. If both transglutaminase-IgA and total IgA are normal, CD is unlikely.

Other extrahepatic autoimmune diseases: Concurrent extrahepatic autoimmune disease occurs in 15%-50% of children with AIH[1,2]. Autoimmune thyroid disorders (Hashimoto’s disease, autoimmune thyroiditis, Graves’ disease) are the most common[2,36]. Autoimmune skin disorders, type 1 diabetes mellitus, and systemic lupus erythematosus are also associated. In a large adult cohort, the presence of extrahepatic autoimmune disease had no impact on the clinical course or prognosis in AIH[36]. All children with refractory AIH should be screened for thyroid disorders and systemic lupus erythematosus. Assessment for other systemic autoimmune disorders should be guided by clinical judgement and patient symptomatology.

THERAPEUTIC OPTIONS FOR REFRACTORY AIH

The options for treatment in children are MMF, calcineurin inhibitors, rituximab, infliximab, and sirolimus[37]. Table 5 details the immunosuppressive therapies used in children with refractory AIH.

Table 5 Immunosuppressive therapies used in children for refractory autoimmune hepatitis.

Medication		Route of administration	Dose in children	Trough level	Adjuvant therapy indicated
Mycophenolate mofetil	Second line	Oral	10 mg/kg/dose twice daily; maximum dose 20 mg/kg/dose twice daily	Not applicable	Yes- steroid
Tacrolimus	Second or third line	Oral	0.05 mg/kg/day- adjusted as per trough level	Initial serum trough levels 6-8 ng/mL, tapering to 3-5 ng/mL	Yes- steroid
Cyclosporine A	Third or fourth line	Oral	4 mg/kg twice daily- adjusted as per trough level	Initial trough levels at 200-250 ng/mL, tapering to trough levels < 120 ng/mL after full biochemical remission	Yes- steroid
Sirolimus	Third/fourth line	Oral	Initial dose 1-2 mg/m2 body surface area	Trough level of 4 ng/mL to 8 ng/mL	Yes- steroid
Rituximab	Third/fourth line	Intravenous infusion	375 mg/m2 2-4 doses every alternate week or fortnightly; further doses of rituximab might be needed 4-6 monthly	Not measurable. Surveillance of CD20+ B-cells is recommended; immunoglobulin supplementation may be necessary	Yes- steroid
Infliximab	Third/fourth line	Intravenous infusion	5 mg/kg infliximab is infusions at 0, after two weeks, and after six weeks of initial infusion, and thereafter every four to eight weeks depending on laboratory and clinical course	Data not available	Yes- steroid

MMF

MMF inhibits purine synthesis in B and T lymphocytes and is the most common second-line agent used in the management of refractory AIH[38]. MMF should be the first therapeutic option for patients’ intolerant or not responding to AZA[2]. Santiago et al[39], in a meta-analysis of adult patients treated with MMF, reported response rates of 82% and 32% for AZA intolerance and for treatment failure, respectively. There was a low incidence of adverse events (14%) and discontinuation of therapy (8%)[39]. In children, MMF induced a complete biochemical response in 89% of those with AZA intolerance, while, similar to adult studies, the response rate was lower (22%) in non-responders to AZA[40]. Adverse events were reported in 44%, but cessation of therapy was only needed in 11%[40].

The major disadvantages of MMF are uncertainties in how to monitor drug levels and its teratogenic potential. Across Europe, only 22% of paediatric centres used drug level monitoring, with a preference for predefined or weight-adjusted doses[38]. The trough level of mycophenolic acid depends on multiple factors, including polymorphisms, renal function, albumin levels, and concomitant medications, and there is no recommended range in paediatric AIH. MMF is teratogenic, with risk of congenital malformations, spontaneous abortions, and is hence contraindicated in pregnancy. Due to the risk of genotoxicity, contraception is also needed in men[29]. The risk of teratogenicity should be discussed with all adolescents and families, contraceptive advice should be offered, and the discussion documented. The recommended dose in children is initially 10 mg/kg body weight twice daily, increased to 20 mg/kg body weight twice daily (1-2 g/day) based on treatment response, with close monitoring for side effects, especially cytopenia and gastrointestinal symptoms (nausea, vomiting, diarrhoea).

Calcineurin inhibitors

Tacrolimus and, rarely, cyclosporine A (CsA) are commonly used as immunosuppressive agents post-transplantation; both are structurally distinct but have a similar mechanism of action. Tacrolimus causes impairment of gene expression in target cells by binding to FK506-binding protein (CsA binds to cyclophilin), inhibiting calcineurin phosphatase, and suppressing interleukin-2 synthesis and T-cell proliferation. Tacrolimus also potentiates the actions of glucocorticoids by binding to FK-binding proteins contained within the hormone–receptor complex, preventing degradation. Across Europe, tacrolimus is most commonly used as a third-line and less commonly as a second-line agent in refractory AIH[38]. As a second-line agent, tacrolimus can be used in specific situations - for example, when there are concerns about potential non-compliance, when frequent monitoring of drug levels is beneficial, or in patients concerned about the teratogenic potential of MMF. With tacrolimus, it is possible to titrate the trough level within a narrow range. Tacrolimus induced biochemical remission in approximately 90% and 50% of children when used after intolerance to previous medication and non-responsiveness to previous treatment, respectively[11,40]. The recommended target serum trough level before attaining remission is 6-8 ng/mL; the target level can be tapered to 3-5 ng/mL after full biochemical remission has been achieved[29]. Frequent monitoring of trough levels is essential until the levels are consistently within range.

CsA is very rarely used as a third-line agent in AIH. In a systematic review and meta-analysis of second-line agents in AIH, CsA induced remission in 83% at 6 months, but studies evaluating the use of CsA are all historical[11]. The long-term use of CsA is complicated by the high frequency of side effects, including cosmetic (gingival hyperplasia, hirsutism), nephrotoxicity, and neurotoxicity. Due to the undesirable side-effect profile - especially cosmetic - there is a high risk of non-adherence. Frequent drug level monitoring is needed to keep the drug level in the therapeutic range to decrease the risk of long-term complications. The target therapeutic level in AIH is not well defined, but initial trough levels of 100-300 ng/mL, tapering to trough levels < 100 ng/mL or 100-200 ng/mL after full biochemical remission, have been used[38]. The only indication for CsA use in children with AIH as a second or third-line immunosuppressant, as per the 2019 AASLD practice guidelines[2], is in the presence of concurrent diabetes mellitus.

Anti-B-cell therapy

Though AIH is characterised by elevated immunoglobulins, autoantibodies, and a plasma cell-rich histological infiltrate, it has traditionally been considered a T-cell-mediated disease, with onset and chronicity driven by T-helper cells and impaired T-regulatory cells, respectively[1]. The use of anti-B-cell therapy in AIH is supported by evidence from murine models showing B-cell facilitation of the autoimmune process through antigen presentation, stimulation of T cells, and inflammatory cytokine production, causing hepatocyte injury and cell death[1]. Rituximab, an anti-CD20 monoclonal antibody, is the only anti-B-cell therapy used in children[41]. Rituximab depletes B cells and prevents CD4 T-cell activation, thereby preventing hepatocyte damage. B-cell activating factor, a cytokine required for B-cell development and differentiation, has also been proposed as a therapeutic target in AIH[6]. Belimumab, an anti-B-cell activating factor monoclonal antibody, has been reported to induce remission in adults with AIH, but its use has not been reported in children[42].

Rituximab is recommended only as a third or fourth line-line drug in refractory AIH[38]. Evidence for the use of rituximab in refractory AIH is limited to case reports and case series[41,43]. In a large series of adult patients, Than et al[43] reported that rituximab caused a significant decline in aspartate aminotransferase, alanine aminotransferase, and IgG within a month of starting treatment, sustained for 2 years, with the absence of clinical disease flares and a reduction in prednisolone dose, without significant side effects. Saul et al[41] reported the successful use of rituximab in two children, achieving complete biochemical remission. Rituximab should always be used in combination with other immunosuppressive agents. The initial course used in children is a dose of 375 mg/m², 2-4 doses given every other week. Further doses may be required every 4-6 months. Patients and families should be counselled regarding the increased risk of infections, and immunoglobulin infusion (1 g/kg) can be given after rituximab infusion to reduce infection risk.

Mammalian target of rapamycin inhibitors

Mammalian target of rapamycin inhibitors acts late in the cell cycle, preventing interleukin-2–mediated proliferation of activated CD4 and CD8 T cells. Sirolimus is the only drug in this group that has been used in refractory paediatric AIH, with its use extrapolated from experience in liver transplant recipients, particularly in recurrent and de novo AIH. Complete biochemical remission has been reported in 40%-50% of patients treated with sirolimus, without significant side effects[44,45]. Sirolimus should only be used as a third- or fourth-line medication in refractory AIH. A starting dose of 1-2 mg/m² and a target trough level of 4-8 ng/mL have been suggested.

TNF-α inhibitors

The only tumor necrosis factor alpha (TNF-α) inhibitor used in refractory AIH in children is infliximab, a recombinant humanised chimeric antibody that directly neutralises soluble TNF-α. Evidence for its use in AIH is limited to case reports and small series[46,47]. Infliximab induced complete remission in approximately 55% of patients unresponsive to second- or third-line treatment[47]. Another study reported complete remission in 75% of non-responders to second-line therapy and 46% of non-responders to third-line therapy[46]. Infliximab should only be used as a third- or fourth-line medication in refractory AIH. Patients and families should be counselled regarding the increased risk of infectious complications, the possibility of developing autoimmune-like hepatitis while on treatment, and the need for multiple infusions over prolonged periods. A suggested regimen is infliximab 5 mg/kg administered at week 0, 2, and 6, followed by maintenance infusions every 4-8 weeks depending on laboratory and clinical course.

CHOICE OF THERAPEUTIC AGENTS IN REFRACTORY AIH

There is a lack of universal consensus on the management of paediatric patients with refractory AIH. Most experts and protocols recommend MMF as the first choice for second-line treatment[2,3,37]. The use of other medications is based on case reports, case series, and treatment centre preferences rather than clinical trials[48]. The choice of treatment options in paediatric patients has been guided by the experience of using these medications in adults. Hence, it is difficult to provide a generalised recommendation for refractory AIH in children if refractory to MMF or if treatment with MMF is not an option.

Figure 4 illustrates a treatment algorithm for children with refractory AIH. Choice of therapy has traditionally been based on local practices and experience of using the drugs in other settings, for example, post-solid organ transplantation. It is important to consider the specific clinical situations, previous therapy for AIH, compliance, options for therapeutic drug monitoring, and adverse effects profile before deciding on the most appropriate treatment regimen. Patients and families should receive extensive counselling on the benefits, risks, and long-term complications of medication(s) prior to initiation of second-line or higher therapy, and the discussion should be documented. Consideration also should be given to adherence, the need for drug level monitoring, growth potential, and availability of liquid formulations of medication while deciding on possible treatment options. Children’s and families’ preferences need to be taken into consideration before deciding on the therapy.

Figure 4 Suggested treatment algorithm for children with refractory autoimmune hepatitis. ¹If mycophenolate mofetil used as second line agent. ²In children with concomitant type 1 diabetes mellites. Modified from Mack et al[2] and Mieli-Vergani et al[37].

It is also vital that after initiating any of the regimens, a sufficient period of time is allowed before considering changing or intensifying therapy. Patients with advanced fibrosis or severe interface hepatitis may require a longer duration to achieve a complete biochemical response, and the use of more potent immunosuppressive drugs could increase the risk of unnecessary short- and long-term complications[49]. Hence, the risks vs benefits of escalating or augmenting therapy should be carefully considered prior to any changes. Overall, children who are intolerant to AZA are more likely to respond to a change of treatment than non-responders. Patients with insufficient response to first- and/or second-line treatments are expected to exhibit lower rates of response to third-line therapies compared to intolerant patients and often require combinations of two or three drugs to achieve complete biochemical response[49]. The recommended time between changes is 6 months, but it might be prudent to wait at least 12 months before changing treatment[11]. It is recommended that second-line or higher treatment options be decided and monitored only in specialised hepatology centres[29].

NON-PHARMACOLOGICAL INTERVENTIONS

Non-pharmacological interventions are a crucial part of managing any chronic liver disease and become even more important in children with refractory AIH, whose HRQoL is likely lower than those responding to treatment[37]. Improved HRQoL has been shown to increase treatment adherence and improve disease outcomes in chronic illnesses. The role of multidisciplinary teams in supporting patients and families is vital. There should be open and honest discussions with patients about how various team members can provide support. Mental health services and HRQoL assessments should be offered to all patients[49]. Positive effects on stress reduction and self-control have been reported to improve outcomes[49].

Fatigue is a common but difficult-to-quantify symptom in children with autoimmune liver diseases. Counselling and physiotherapy support should be made available for such patients. Appropriate nutrition is a critical aspect of managing refractory AIH, and regular nutritional assessment and support should be part of routine follow-up. Patient and family involvement in daily care and research should be encouraged. The benefits of participation in patient support groups and connecting with peers of similar age should be communicated to all patients and families[49]. Support from these groups can significantly enhance HRQoL and should be offered as part of comprehensive care[49].

LONG-TERM OUTCOMES IN CHILDREN WITH REFRACTORY AIH

There is a lack of multi-centre prospective registries or studies addressing the efficacy of alternative agents in AIH. Zizzo et al[11], in a systematic review and meta-analysis, reported data from multiple studies on the use of alternative agents: MMF (5 studies), calcineurin inhibitors (6 studies), one study each of MMF and calcineurin inhibitors, rituximab, sirolimus, and budesonide. The response rate for MMF (36%) was lower than cyclosporine and tacrolimus (84% and 50% respectively), while adverse events were the highest with cyclosporine (82%), followed by tacrolimus (42%) and MMF (38%)[11]. These results need to be interpreted with caution due to the small size of the studies.

Patients who do not respond to medical treatment and have progression of liver disease will require assessment for LT. Approximately 10% of children and adolescents with AIH progress to end-stage liver disease requiring transplantation within 15 years of diagnosis[37]. The indications for LT are decompensated chronic liver disease, uncontrolled portal hypertension, especially intractable/recurrent variceal bleeding which is not controlled by medical or endoscopic management and not amenable to transjugular intrahepatic portosystemic shunt, development of hepatopulmonary syndrome, porto-pulmonary hypertension, or hepatocellular carcinoma. Refractoriness to treatment is associated with a higher risk for LT or the probability of death due to liver disease[8,12]. In a historical cohort (n = 98) from our unit, patients with AIH and overlap syndromes, 18% needed second-line or higher-grade therapy, with 33% attaining remission, while 33% needed LT at a median interval of 32 months from the time of diagnosis[50]. Patients in this cohort received only MMF, tacrolimus, and CsA as second-line or higher therapy[50].

FUTURE THERAPIES AND DIRECTION IN THE TREATMENT OF AIH

A wide range of potential therapies is currently being evaluated based on our present understanding of the immunopathogenic mechanisms in AIH. These include pharmacological and biological agents that can restore homeostatic mechanisms modulating immune responses, reduce oxidative or nitrosative stress, and inhibit hepatic fibrosis[2]. Future management directions probably include molecular interventions (recombinant proteins and monoclonal antibodies), cellular therapies (adoptive transfer and antigenic manipulation such as T-regulatory cells, mesenchymal stromal cells, or myeloid-derived suppressor cells), and pharmacological agents (antioxidants, antifibrotics, and antiapoptotic drugs)[51]. Among these potential therapies, only a few are currently in ongoing trials (Table 6).

Table 6 Examples of ongoing trials in autoimmune hepatitis.

Ref.		Treatment	Mechanism of action	Stage of development
Mack et al[2], 2020; Whitehead and Kriegermeier[42], 2020; Reau et al[48], 2024	1	Zetomipzomib	Immunoproteasome inhibitor	Phase 2
	2	Ianalumab, Belimumab	Antibody-dependent cellular cytotoxicity mediated B-cell depletion through B-cell activating factor inhibition	Phase 2/3
	3	JKB-122	Toll-like receptor-4 inhibition	Phase 2
	4	Mesenchymal stromal cells	Induction of regulatory T-cell differentiation and suppression of lymphocyte activation	Phase 1/2
	5	Synthetic preimplantation factor (s-PIF)	Creation of immunosuppressive and immunomodulatory environment of pregnancy	Phase 1

At present, therapy for AIH is not specific or targeted and relies on the broad effects of immunosuppressive medications[48]. The only personalised management available is to analyse the thiopurine methyltransferase enzyme assay before starting AZA to identify individuals who may need a reduced dose or alternative treatment. This helps to minimize the risk of adverse drug reactions and maximize the effectiveness of treatment. Ideally, specific targeted immunomodulatory treatment options for patients with refractory AIH would decrease the need for broad immunosuppression, reduce the risk of short- and long-term complications, and enable individualized therapeutic approaches. Unfortunately, unlike other autoimmune diseases such as IBD or rheumatoid arthritis, where targeted therapies are available, this remains an aspiration in AIH[48].

Precision medicine holds substantial promise for transforming the management of AIH by tailoring diagnosis, prognostication, and therapy to the individual patient’s unique clinical, genetic, and biomarker profile[52]. In AIH, the integration of demographic information, clinical parameters, biomarker data, and genetic insights into predictive, self-learning systems represents a likely path forward. Such a precision medicine framework would incorporate genomics, proteomics, metabolomics, and microbiome profiling - to stratify risk of disease progression, guide the timing and selection of therapeutic interventions, predict treatment response, and identify high-risk subgroups[52]. This paradigm shift from treatment toward prevention aims to enable earlier detection and targeted intervention, supported by the application of artificial intelligence and machine learning to uncover hidden patterns and refine prognostic models[52].

CONCLUSION

In summary, for the 10%-20% of children who do not respond to first-line treatment with AZA and steroids, non-compliance should be ruled out, the diagnosis of AIH revisited, and evaluation performed for liver diseases that can mimic the biochemical and/or histological features of AIH, as well as for the presence of extrahepatic complications. MMF remains the most commonly used agent in refractory AIH. Further treatment strategies should be individualized, ideally based on the experience of specialized centers, and decided in close consultation with the patient and family.