Difficult to treat and refractory autoimmune hepatitis: Recent advances in pharmacological management

Abstract

Autoimmune hepatitis (AIH) is a rare cause of chronic liver disease. The exact pathophysiology of AIH is unknown. Breakdown of self-tolerance against hepatic antigens and molecular mimicry are often implicated in the pathogenesis of AIH. Immunosuppressive therapy is the mainstay of treatment; however, 10%–25% of patients with AIH may not respond to primary therapy. Those patients are often salvaged with second- and third-line immunosuppressive therapy. Workup for other concomitant diseases should be done for patients who fail to respond to primary immunosuppressive therapy. Concurrent metabolic dysfunction-associated steatotic liver disease, alcohol-related liver disease, overlap syndrome (AIH with primary biliary cholangitis or sclerosing cholangitis), chronic hepatitis B virus, hepatitis C virus, and human immunodeficiency virus infection should be ruled out in such cases. Targeting the concomitant etiology may lead to resolution of the clinical symptoms and induce biochemical and histological remission. Isolated AIH without other etiologies for liver injury should be managed with a higher dose of steroids, azathioprine, or other immunosuppressive agents. Second- and third-line immunosuppressive agents include mycophenolate mofetil, cyclosporine, tacrolimus, infliximab, and rituximab. Patients with AIH may present with acute severe AIH (AS-AIH) and AIH-related acute on chronic liver failure, and they often require liver transplantation. The terms refractory or difficult-to-treat AIH have been used interchangeably and have no distinct definition. Difficult-to-treat AIH includes patients with intolerable side effects, fulminant disease (AIH with acute on chronic liver failure and AS-AIH), AIH in pregnancy, and HIV infection. Patients who fail to respond to standard first-line immunosuppressive therapy should be classified as refractory AIH. This review addresses the issues in the management of difficult-to-treat AIH with recent advances in pharmacological management.

Key Words: Autoimmune hepatitis; Difficult-to-treat autoimmune hepatitis; Steroids; Azathioprine; Mycophenolate mofetil; Cirrhosis

Core Tip: Immunosuppressive therapy remains the mainstay for patients with autoimmune hepatitis (AIH). However, managing AIH in special situations, such as acute-on-chronic liver failure, pregnancy, old age, and hepatitis B and C co-infection, can be challenging. Between 10% and 20% of patients may not respond to standard first-line immunosuppressive therapy. This review focuses on the individualized management of patients with difficult-to-treat or refractory AIH and recent advances in second- and third-line immunosuppressive therapy.

INTRODUCTION

Autoimmune hepatitis (AIH) is a chronic, immune-mediated inflammatory liver disease. It is characterized by the presence of autoantibodies, elevated IgG levels, and specific histological findings such as interface hepatitis (Figure 1)[1]. It is an uncommon cause of chronic liver disease[1,2]. It is more prevalent among women; however, its clinical presentation can be heterogeneous. Patients may present with an asymptomatic rise of liver enzymes to acute or fulminant liver failure[1,2]. Its epidemiology and presentation differ globally, depending on the patient’s age, gender, and ethnicity[3,4]. Glucocorticoids and azathioprine (AZA) have been used in patients with AIH with a response rate of 60%–70%[5]. These immunosuppressive therapies are not devoid of adverse events. Glucocorticoids are often associated with weight gain, osteoporosis, glucose intolerance, cataracts, and other adverse events[6]. AZA can lead to bone marrow suppression and infection-related complications[7].

Figure 1 Pathogenesis of autoimmune hepatitis and pharmacological targets. Immunopathogenesis of autoimmune hepatitis (AIH) involves adaptive and innate immune systems. Molecular mimicry and breakdown of self-antigenic tolerance lead to activation of B and T cells. B cells in turn stimulate plasma cells which secret autoantibodies. Cytotoxic T cells and autoantibodies lead to hepatotoxicity. Immunosuppressive therapies target various pathways of the pathogenesis in AIH as depicted in the figure. Methotrexate, azathioprine, 6-mercaptopurine, and 6-thioguanine interfere with cell cycles and arrest the proliferation of B and T cells. Infliximab is a chimeric anti-tumor necrosis factor antibody that acts on the downstream inflammatory pathways. Rituximab and belimumab deplete B cells and attenuate B-cell-mediated activations of plasma cells. Calcineurin inhibitors like tacrolimus and cyclosporine preferably act on T cell proliferation. Sirolimus is a mammalian target of rapamycin inhibitor that ameliorates helper T cells-mediated tissue damage in the liver cells. MMF: Mycophenolate mofetil; 6-MP: 6-mercaptopurine; 6-TG: 6-thioguanine; IL: Interleukin; m-TOR: Mammalian-target of rapamycin; IFNγ: Interferon-gamma; CTL: Cytotoxic T lymphocytes; TNF: Tumor necrosis factor.

A nuanced and individualized approach is advocated for the management of AIH[1-4]. Up to 10%–20% of patients may not respond to the primary immunosuppressive treatment[5], and 5%–10% of patients may even discontinue therapy owing to its debilitating adverse events[5,6]. Failure to achieve remission is invariably associated with low transplant-free survival[8]. Moreover, concomitant diseases and conditions, such as diabetes, pregnancy, human immunodeficiency virus (HIV) infection, and other comorbidities, may complicate the course of the disease[9-12]. Such patients are classified as difficult-to-treat AIH. Immunosuppressives should be used cautiously in some critical conditions such as acute severe AIH (AS-AIH), acute-on-chronic-liver failure (ACLF), concomitant HIV infection, metabolic dysfunction-associated steatotic liver disease (MASLD), and cirrhosis[4,9,13]. Concurrent viral hepatitis, overlap syndromes, and MASLD should be ruled out in such patients before switching to other immunosuppressive agents and labelling them as refractory AIH[4,9]. Once the diagnosis of refractory AIH is made, an individualized approach is advocated to choose among different second- and third-line immunosuppressive agents. They include mycophenolate mofetil (MMF), tacrolimus (TAC), cyclosporine (Cys), sirolimus, infliximab (IFX), and belimumab. This review discusses evidence-based management strategies in patients with refractory and difficult-to-treat AIH. It will primarily focus on the second, and third-line pharmacological therapy for the management of patients with AIH.

DIFFICULT-TO-TREAT AIH: SPECIAL SITUATIONS AND POPULATIONS

AIH in difficult situations

AS-AIH is defined by the presence of jaundice and coagulopathy [international normalized ratio (INR) ≥ 1.5] in patients with an acute presentation of AIH (≤ 26 weeks between the onset of jaundice and coagulopathy) with no evidence of pre-existing liver disease[13]. A subset of these patients may develop encephalopathy, categorizing them as AS-AIH with acute liver failure (ALF)[4,9]. The diagnosis of AS-AIH is established only in the absence of severe fibrosis or cirrhosis on histology[13,14].

The reported corticosteroid response rate ranges widely from 36% to 100%, with the high variability likely due to differences in the severity of liver disease across studies[13-16]. In AS-AIH, it is crucial to identify patients who will benefit from corticosteroid therapy. The predictors of failure to respond include high INR (> 2.46) and a model for end-stage liver disease (MELD) score > 28.5[17]. The SURFASA scoring system, which combines three parameters, baseline INR, change in INR over 3 days, and change in total bilirubin over 3 days after initiation of steroids, has been validated to predict outcomes[15,18]. A SURFASA score < 0.9 suggests a steroid response of 75%, while a score of > 1.75 indicates a risk of transplant or death of > 85%. AS-AIH has a poor prognosis, and patients should be evaluated for liver transplantation (LT) promptly[18].

AIH with ACLF

AIH-related ACLF is a syndrome characterized by an acute decline in liver function observed in patients with pre-existing chronic liver disease. It is often accompanied by the presence of extrahepatic organ failure[19]. AIH with ACLF (AIH-ACLF) can be distinguished from AIH with ALF by the presence of advanced fibrosis (F3/F4), ductular reactions and large areas of parenchymal collapse with lymphoplasmacytic inflammation[19,20].

The data on steroid use in AIH-ACLF patients are limited. In the largest study, from the Asia Pacific Association for the Study of Liver (APASL) – ACLF Research Consortium (AARC) database, 28 of 82 (34%) patients with AIH-ACLF received steroids. The patients who received steroids had a 90-day survival of 75%, with five patients dying of sepsis. Two had progressive liver failure-related death. Factors associated with a lack of response to corticosteroid therapy included high baseline serum bilirubin, high baseline MELD score (> 27), presence of hepatic encephalopathy (HE) and grade 3/4 fibrosis in liver biopsy[19,20]. Another retrospective study included 29 patients with AIH-ACLF[21]. Following steroid, 90-day and 180-day transplant-free survival were 55.2% [95% confidence interval (95%CI): 39.7%–76.6%] and 30.2% (95%CI: 16.7%–54.6%), respectively[21]. Due to a lack of adequate data on steroids and the poorer outcomes, patients with AIH-ACLF should be evaluated for LT. They can be considered for immunosuppressive therapy only on a case-by-case basis after ruling out and controlling sepsis or overwhelming infection[19-21].

AIH with decompensation

There are a few retrospective studies that have looked at the efficacy of immunosuppressive treatment in patients with decompensated cirrhosis. The largest multicenter study from the International Autoimmune Hepatitis Group included 232 patients with AIH-related decompensated cirrhosis. At diagnosis, 89% had ascites, and 41% had overt HE. A total of 214 (92%) treated patients had higher aminotransferases, bilirubin and modified hepatic activity index. Those with no HE or HE grades 1/2 showed a 60% LT-free survival on treatment. Patients with MELD-Na > 28 had worse LT-free survival rates. Also, a decline in MELD-Na after 4 weeks of treatment had a 100% negative predictive value for death or LT. Around half of the patients achieved recompensation on follow-up[22]. In another small retrospective study of 64 patients treated with corticosteroids, 63% recompensated, while 14% had liver-related death or LT. Changes in total bilirubin and MELD scores at day 7 predicted treatment outcome[23]. Another study compared patients with decompensated cirrhosis and mild or no ascites vs gross ascites vs compensated cirrhosis; 31% of patients with mild ascites vs 4% with gross ascites achieved a biochemical response. Although the initial steroid dose was lower in patients with decompensated cirrhosis (20–25 mg/day), infections were more frequent[24]. Even though there is the possibility of recompensation on immunosuppressive therapy, it is recommended that patients with decompensated cirrhosis should be evaluated for LT.

Asymptomatic patients

Twenty-five percent of patients with AIH may be asymptomatic and may just have elevated liver enzymes[4,25-27]. An Italian retrospective study of 305 patients compared the treatment response and outcomes in patients with asymptomatic vs symptomatic AIH. Thirty percent of the patients were asymptomatic at presentation and had lower enzymes; however, treatment response, liver disease progression, and progression to cirrhosis were similar between the two groups[27]. Patients should therefore be started on immunosuppressive therapy when indicated, irrespective of the presence of symptoms[4,25].

AIH in special populations

Pregnancy: Flares during pregnancy in AIH are not uncommon, with a reported incidence of 21%–33%[28]. Postpartum flares are frequent, occurring in 41% of patients at a median of 11 weeks after delivery[28,29]. AIH may present initially during pregnancy, and standard diagnostic algorithms and treatment regimens should be used. A complete biochemical response during conception and pregnancy is associated with reduced AIH flares and better pregnancy-related outcomes[29,30]. To conceive, patients should be in a biochemical remission for at least 1 year[25]. The commonly used first-line immunosuppressive drugs, steroids and thiopurines can be safely used during pregnancy and lactation[4]. Treatment with MMF or TAC is contraindicated in pregnancy due to teratogenic effects (Table 1), and contraceptives should be used in women of childbearing age on MMF. MMF should be discontinued at least 3 months before conception[4,25].

Table 1 Second-line immunosuppressives used in refractory autoimmune hepatitis.

Drugs	Mycophenolate mofetil	Tacrolimus	Cyclosporine	Methotrexate	Sirolimus
Position in managing refractory AIH	Second line	Second line	Second line	Second line	Second line
Mechanism of action	Prodrug converted to mycophenolic acid, which inhibits purine biosynthesis and attenuate proliferation of B and T cells	Calcineurin inhibitors which suppress IL-2 production	Calcineurin inhibitors which suppress IL-2 production	Inhibition of folic acid synthesis	m-TOR inhibitor
Dose	20–40 mg/kg/d	2–6 mg/kg/d	2–4 mg/kg/d	7–20 mg/wk	1–2 mg/kg/d
Pregnancy category1	D	C	C	X	C
Adverse events	Diarrhea, cytopenia and colitis	Nephrotoxicity, gingival hypertrophy, hypertension and dyselectrolytemia	Nephrotoxicity, hyperkalemia	Bone marrow suppression, pulmonary fibrosis, and liver injury	Arthralgia, peripheral edema, skin rashes ang gastrointestinal problems
Comments	More effective for azathioprine intolerant than azathioprine refractory AIH. Current RCT shows it can be safe first-line alternative to azathioprine[61-63]	Tacrolimus was effective in 18 of 23 patients with first-line-intolerant or refractory AIH[65]	Five of six patients with steroid refractory AIH responded to cyclosporine[66,120]	Six of 11 patients responded to methotrexate after initial non-response to steroid with or without azathioprine[68,69]	Four of five patients with steroid refractory AIH responded to sirolimus[92]

¹These capital letters denote pregnancy category of the drugs.

m-TOR: Mammalian-target of rapamycin; AIH: Autoimmune hepatitis; RCT: Randomized controlled trial; IL: Interleukin.

Old age: Elderly patients with AIH achieve a similar complete biochemical response as younger patients; however, they have lower rates of relapse[31,32]. In a recent systematic review, treatment-related adverse events were similar in all age groups[31]. Due to the presence of multiple comorbidities and the risk of osteoporosis with corticosteroids in these patients, recent guidelines recommend treating elderly patients with moderate activity on histology and/or with advanced fibrosis[4,26]. Continuing a lower dose of steroids after achieving remission is advisable to reduce the incidence of osteoporosis[26].

AIH in HIV: AIH in people living with HIV is usually diagnosed after initiation of antiretroviral therapy when the immune status of the patient has been restored[12]. Immune reconstitution inflammatory syndrome after initiation of antiretroviral therapy has been hypothesized to be the trigger for AIH[33,34]. Patients started on corticosteroids and AZA for maintenance showed excellent response to therapy in a case series. It was without HIV viral rebound or infectious complications[33,34].

AIH in hepatitis B and C co-infection: Concurrent infection with hepatitis B virus (HBV) and hepatitis C virus (HCV) is rare; however, it can be found in high-prevalence regions. With the advent of highly efficacious direct-acting antivirals (DAAs), HCV co-infection can be tackled with ease. DAAs have also been shown to resolve autoimmune features by themselves[35,36]. HBV co-infection needs extensive evaluation and therapy in most cases before immunosuppressive therapy. Hepatitis B surface antigen (HBsAg) positive, HBsAg negative with antibody against total core antigen (anti-HBc) positive patients at high risk of reactivation should be initiated on antiviral therapy before immunosuppression[35]. HBsAg-negative and anti-HBc-positive patients should be monitored for HBsAg and HBV deoxyribonucleic acid (DNA) while on immunosuppressive therapy. Patients with chronic hepatitis D virus infection demonstrate a higher frequency of antinuclear antibody and anti-smooth muscle antibody positivity (67% and 43%, respectively) and higher levels of serum IgG, which suggest an autoimmune phenomenon at play[36,37]. However, the relevance of this is unknown.

First-line immunosuppressive therapy in AIH: Current concept

The aim of first-line immunosuppressive therapy is to induce remission and halt fibrosis progression in patients with AIH[4,25,26]. First-line therapy comprises glucocorticoids with or without AZA (Figure 2). The standard dose of prednisone is 20–40 mg/day or 1–2 mg/kg/day. Some societies advocate the use of AZA along with a lower dose of steroids from the beginning of the treatment[4,25]. However, given the side effects profile, it is logical to assess for steroid responsiveness after 2 weeks of the therapy. If the patient responds to the steroid, AZA (50-150 mg/day) is added once bilirubin comes below 5 mg/dL[25]. Thiopurine methyl transferase (TPMT) mutation testing is mandatory before starting AZA[38]. Owing to its higher prevalence, nudix hydrolase (NUDT-15) mutation analysis is a prerequisite for starting AZA in Asian countries[39,40]. Following the initiation of the treatment, liver function tests are done after 1–2 weeks, and a biochemical response is assessed after 4–8 weeks (Figure 2). If patients respond to treatment, the steroid is tapered to 5 mg/day to minimize the risk of steroid-related adverse events[4,26,41]. A dose of ≥ 10 mg/day is associated with higher adverse events[41,42].

Figure 2 Indication of second- and third-line immunosuppressive therapy in patients with autoimmune hepatitis. Following primary immunosuppressive therapy, response to treatment should be assessed after 2–4 weeks. After checking drug compliance, concomitant etiology should be ruled out if patients do not respond to primary therapy. Subsequently, second- and third-line immunosuppressive therapy should be chosen. 6-MP: 6-mercaptopurine; 6-TG: 6-thioguanine; MASLD: Metabolic dysfunction-associated steatotic liver disease; AIH: Autoimmune hepatitis; PBC: Primary biliary cholangitis; DILI: Drug-induced liver injury; ArLD: Alcohol-related liver disease.

Steroid monotherapy is advisable when the anticipated duration of treatment is < 6 months or the patient has contraindications for AZA therapy[4].

Budesonide has been used instead of prednisolone in patients with AIH. A randomized controlled trial (RCT) has demonstrated the superiority of budesonide over prednisone. After 6 months, 60% of patients in the budesonide arm achieved biochemical remission compared to 39% in the prednisone arm, with fewer steroid-related side effects (28% vs 53%). However, the study did not include patients with ALF or cirrhosis. As budesonide undergoes extensive fast-pass metabolism in the liver, it is not recommended in patients with cirrhosis[43]. Budesonide is also associated with the risk of portal vein thrombosis in patients with cirrhosis[44,45].

Patients who fail to respond to steroids: What is next?

Nonresponse to steroids is defined as the lack of reduction of liver enzymes by more than 50% after 4 weeks of therapy[46]. However, according to the APASL guidelines, failure of reduction of liver enzymes by 25% within 2 weeks of treatment initiation is steroid nonresponsiveness (Figure 2)[4]. Patients who respond to steroids but do not meet the criteria for a complete biochemical response after 6 months are categorized as partial responders[4,26]. Early response to steroids is associated with longer transplant-free survival[27]. Non-responders and partial responders are at the highest risk for progression of liver disease. A long-term follow-up study from Bologna, Italy, has shown that 55% of patients who fail to respond completely to steroids (partial or non-responders) experienced the advancement of liver diseases[27].

Delaying second-line therapy in such patients can be associated with worsening of pre-existing liver dysfunction. They should be promptly evaluated for the other concomitant etiology and counseled for second-line therapy (Figure 2)[4,25-27].

Work-up for concomitant etiology

Patients who fail to respond to the initial immunosuppressive therapy should be investigated for concurrent etiology for liver injury, such as MASLD, primary biliary cholangitis (PBC), primary sclerosing cholangitis (PSC), drug-induced liver injury (DILI), complementary and alternative medicine (CAM) or herb-related liver injury, alcohol-related liver injury, and chronic viral hepatitis B and C (Figure 2)[4,8-10].

Twenty to 30% of patients with AIH may have underlying histological evidence of MASLD. A subset of such patients may have steatohepatitis. Concomitant MASLD in AIH is associated with higher liver enzymes, advanced fibrosis, and decompensation[10,47]. Steroids may worsen underlying metabolic dysfunction in such patients[47,48]. Overlap syndromes with PBC and PSC should be ruled out using standard criteria[49-51]. Adding ursodeoxycholic acid (UDCA) to immunosuppression often leads to clinical and biochemical remission in such patients. Patients with alcohol-related liver disease (ArLD) may have circulating autoantibodies[49,50]. Liver histology should be interpreted cautiously in patients with significant alcohol intake[52-54]. However, AIH may coexist in patients with ArLD. Rigopoulou et al[54] reported 12 patients with AIH/ArLD overlap among 317 patients with AIH. Patients with coexistent ALD had a poor outcome compared to AIH alone. Concurrent DILI should be considered if there is a history of related drugs[55-57]. Roussel Uclaf Causality Assessment Method score can be calculated to establish a definite association[55,56]. Liver injury following CAM and herbs is much more common in the Asian subcontinent[55]. A detailed history of the exposure is important. Drug-induced AIH-like injury is a separate entity, which may sometimes require immunosuppression[58].

The seroprevalence of hepatitis B and C ranges from 2% to 8%, depending on the population[59]. Routinely, HBsAg and anti-HCV are used to screen for the presence of these infections. However, routine screening with HbsAg may miss a subset of patients infected with the surface-mutant variant of hepatitis B[4,25,26]. Immunosuppression may lead to reactivation of the underlying HBV infection. Thus, screening with HBsAg, total anti-HBc, and anti-HBsAg titer should be the standard of care. Anti-HCV serology is recommended to screen for HCV infection. Patients with chronic kidney disease and HIV infection may not mount a sufficient antibody response. Testing HCV RNA is recommended in such populations[4].

Second-line therapy

MMF: MMF is the most commonly used second-line therapy in patients with AIH. However, current guidelines recommend the use of MMF as an alternative first-line therapy in patients with AZA-intolerant or nonresponsive AIH (Figure 2)[4,25,60]. MMF is a prodrug, and it is converted to mycophenolic acid, which inhibits purine synthesis. Attenuation of purine synthesis leads to reduced B- and T-cell proliferation[60]. A recent RCT has clearly shown that when combined with steroids, MMF performed better than AZA in treatment-naïve patients with AIH. At 24 weeks, normalization of liver enzymes and IgG was significantly higher in patients receiving MMF (56.4% vs 29%; P = 0.004). No adverse events were reported in the MMF group; however, serious adverse events were noted in 13% of patients in the AZA group[61]. Retrospective data have shown long-term safety and efficacy of MMF as a first-line therapy in patients with AIH[62].

A recent meta-analysis has revealed that MMF works best in patients who initially respond to AZA but have to discontinue it because of side effects[63]. In such patients with AZA intolerance, the pooled response rate with MMF was 82%. However, the response rate was lower (32%) in patients who failed to respond to AZA initially. Pooled discontinuation and adverse events with MMF were 0.08% and 0.14%, respectively[63]. Leukopenia, diarrhea, and teratogenicity are the main concerns with the use of MMF (Table 1)[60].

Calcineurin inhibitors: Cys and TAC have been used in patients with AIH with varied response rates and adverse events. They act by inhibiting interleukin (IL)-2 production and T-cell proliferation[64]. Calcineurin inhibitors (CNIs) have been used as a second-line therapy in pediatric patients with AIH (Table 1). A systematic review and meta-analysis included 76 pediatric patients who failed to respond to steroid or AZA[65]. Overall response rate at 6 months was 36% with MMF and 50% with TAC, and 83% for Cys. The Cys and TAC groups had a significantly higher rate of adverse events (78% and 42%, respectively).

Other real-world data on CNIs as a second- or third-line agent included 20 patients with difficult-to-treat AIH. Ten patients received TAC, and others were managed with Cys. The duration of therapy was longer when the drugs were used as second- or third-line agents. On follow-up, 55% of patients achieved normalization of liver enzymes (Tables 1 and 2)[66].

Table 2 Third-line immunosuppressives used in refractory autoimmune hepatitis.

Drugs	Infliximab	Rituximab	Belimumab
Position in managing refractory AIH	Third line	Third line	Third line
Mechanism of action	Monoclonal antibody to TNF-α	Chimeric monoclonal antibody against CD-20	B-cell-activating factor inhibitor
Dose	5 mg/kg	2 doses of 1000 mg IV administered 2 wk apart, or 375 mg/m2 rituximab	10 mg/kg
Pregnancy category1	B	C	C
Adverse events	Infectious complications can cause drug induced AIH like injury allergic reaction	Infectious complications	Infection, allergic reactions, progressive multifocal leukoencephalopathy, and mental health issues
Comments	Response rate 78% when used as a third line agent[94]	As a third line agent biochemical remission: 67%, 71% flare-free survival after median 6 yr[88-90,121]	Steroid refractory AIH. Two patients had complete biochemical response[99]

¹These capital letters denote pregnancy category of the drugs.

AIH: Autoimmune hepatitis; TNF: Tumor necrosis factor.

Methotrexate: Methotrexate (MTX) inhibits the synthesis of tetrahydrofolate and interferes with the production of DNA[67]. Compared to other second- or third-line therapies, MTX is associated with a lower response rate in patients with AIH[68]. The conventional side effects of MTX include bone marrow suppression, teratogenicity, and liver fibrosis (Table 1). Despite these limitations, it has been used with a success rate of ~50%[68,69]. Long-term use of MTX is linked to pulmonary and hepatic fibrosis, although recent data suggest an overestimation of liver injury related to MTX[70-72]. Routine (2–3-monthly) liver function test monitoring is recommended in patients receiving MTX[71].

6-mercaptopurine: AZA and 6-mercaptopurine (6-MP) are purine anti-metabolites[73,74]. AZA is converted to 6-MP. 6-MP is methylated by TPMT to form methyl-6-MP, or it may undergo oxidation to produce 6-thiouric acid. The immunosuppressive effects of 6-MP are mediated by its active form, 6-thioguanine (6-TG) nucleosides, and 6-methylthioinosine monophosphate. By interfering with DNA synthesis, they attenuate B- and T-cell proliferation[75,76].

In patients with AZA intolerance, 6-MP is a choice. However, 6-MP is ineffective in patients who fail to respond to AZA. A retrospective study by Hübener et al[77] included 20 patients with AIH who had AZA-intolerance: 15 (75%) responded to 6-MP; however, a complete response was seen in eight patients (40%). In this cohort, two patients were AZA-non-responders. None of them responded to 6-MP either (Figure 2).

Long-term use of 6-MP is associated with various side effects, including liver injury[78,79]. Periodic monitoring and strong suspicion can avoid liver toxicity with 6-MP[80].

Thioguanine: Patients with AIH who are intolerant or refractory to AZA or 6-MP can be managed by 6-TG[81]. A study by Legué et al[81] included 17 patients with AZA-intolerant AIH. Following 6-TG therapy, 16 patients had normalization of liver enzymes within 3 months. In another study, 6-TG was safe and effective in patients with AZA or 6-MP-intolerant or refractory cases (Figure 3). Overall response rate was 83%[82].

Figure 3 Managing patients with autoimmune hepatitis who are intolerant or refractory to azathioprine or 6-mercaptopurine. Azathioprine and 6-mercaptopurine are associated with bone marrow suppression and hepatotoxicity. Baseline TPMT and NUDT-15 genetic polymorphism testing is advocated to avoid myelosuppression. It should be avoided in patients who are homozygous for both enzymes. About 10%–20% of patients who are intolerant to azathioprine can be salvaged with 6-MP or MMF; however, AZA-refractory cases do not respond well to 6-MP. They will often require a third line of therapy or MMF. Metabolite assays can help guide us regarding the need for allopurinol. 6-MMP: 6-methylmercaptopurine; 6-MP: 6-mercaptopurine; 6-TG: 6-thioguanine; TPMT: Thiopurine methyltransferase; AZA: Azathioprine; MMF: Mycophenolate mofetil.

Loss of remission and clinical flare should be evaluated using the 6-MP/6-TG ratio. Addition of allopurinol should be considered when the ratio is < 4[83,84]. No correlation exists between serum 6-TG level and liver enzyme levels and remission in patients with AIH[85].

Third line of therapy

Rituximab: B-cell-driven autoimmune injury plays a role in the pathogenesis of AIH (Figure 1). Thus, targeting B cells is a potential opportunity to treat AIH[86,87]. Rituximab targets CD20 on the cell surface. Rituximab is a monoclonal antibody and is effective in patients with AIH[87]. Various case series with long-term data have established its role as second- and third-line therapy in AIH. A case series by Burak et al[88] described six cases of steroid-refractory AIH successfully managed with rituximab (Table 1). A large series of 22 patients with AIH showed promising results in patients with difficult-to-treat type 1 AIH. Prednisolone dose could be reduced in 62% of patients after 12 months, and at 22 months, 71% of patients had flare-free survival. No adverse event was reported with the use of rituximab[89]. A large retrospective data from the ColHai registry included 14 patients with refractory AIH and 19 patients with concomitant autoimmune diseases. Overall, 91% of patients achieved a complete biochemical remission with rituximab[90]. Patients with AIH and other concomitant autoimmune disorders can be successfully salvaged by rituximab[88-90].

Mammalian target of rapamycin inhibitors: Sirolimus is a macrolide. It is a mammalian-target of rapamycin (m-TOR) inhibitor, widely used in solid organ transplantation (Figure 1). By inhibiting m-TOR, it attenuates the sensitivity of B cells and T cells to IL-2[91]. It has been used in refractory AIH. A series of five patients with refractory AIH treated with sirolimus showed a success rate of 80% (four patients had a biochemical response, i.e., reduction of alanine aminotransferase > 50%). However, only two patients maintained sustained remission on follow-up (Table 1)[92].

Anti-tumor necrosis factor-

IFX: IFX is a chimeric monoclonal antibody against tumor-necrosis factor (TNF)[93]. It has been used as third-line of therapy in patients with AIH. In an elegant study by Efe et al[94], 20 patients who failed standard therapy, MMF, 6-MP, and TAC received IFX (Figure 2). Another group of 22 patients received IFX because they had concomitant extra-hepatic autoimmune diseases. After a median follow-up period of 17 (3–104) months, a complete biochemical response was maintained in 78% of patients[94]. The major limitations of IFX include infectious complications. Paradoxically, anti-TNF has been associated with the development of drug-induced AIH-like liver injury[95,96]. Despite these limitations and risks of infections like reactivation of latent tuberculosis in endemic countries, IFX is an attractive option for patients with difficult-to-treat AIH (Figure 2)[97].

Belimumab: Belimumab is a B-cell activating factor inhibitor used in managing systemic lupus erythematosus[98]. Arvaniti et al[99] reported two cases of refractory AIH who were managed with belimumab (Table 1). They were used as a third line of therapy[99]. Patients with SLE and concomitant AIH can be salvaged using belimumab.

Experimental therapy: Despite substantial advances in the management of AIH and recent development of newer immunosuppressive therapies, 10%–20% of patients with AIH fail to respond to currently available therapy[100]. This group of patients with decompensated liver diseases and liver failure should be evaluated for LT. However, recent experimental therapies have shown promising results in refractory AIH, which include tofacitinib, fecal microbiota transplantation (FMT), regulatory T (Treg) cell therapy, low-dose IL-2, bacteriophage therapy, ginsenosides, and preimplantation factor[100].

Tofacitinib: Tofacitinib is a Janus kinase (JAK) inhibitor mostly used in autoimmune disorders and inflammatory bowel disease[101]. Inhibiting JAK, it attenuates downstream signaling pathways and IL production[101]. The ameliorative effects of tofacitinib in immune-mediated liver disease have already been established in animal models[102]. There are reports of successful management of immune-checkpoint-mediated liver injury with tofacitinib[103]. Gökçe et al[104] reported a case of refractory AIH successfully managed with tofacitinib. However, large prospective data are needed to advocate its routine use as a third-line therapy in AIH[105].

FMT: Gut dysbiosis and increased intestinal permeability are some of the major drivers of liver dysfunction in various liver disorders. Increased intestinal permeability, small intestinal bacterial overgrowth and dysbiosis lead to bacterial translocation and lipopolysaccharidemia in patients with cirrhosis[106]. FMT has been tried in various gastrointestinal and liver disorders with promising results. It is currently indicated in patients with recurrent Clostridium difficile infection[106]. By correcting gut dysbiosis, FMT restores the intestinal gut–liver axis and ameliorates the effects of gut-derived toxins in liver dysfunction[106]. In animal models, FMT attenuate the progression of experimental AIH by modulating immune balance[107,108]. Scarcity of long-term data, regulatory and ethical issues are the main concerns regarding the use of FMT in patients with AIH.

Treg therapy and low-dose IL-2: Treg cells are involved in hepatic immune tolerance, and their reserve is important in maintaining immune homeostasis in hepatic parenchyma[109]. Depleted and dysfunctional Treg cells are involved in the pathogenesis of AIH[109]. In a proof-of-concept study by Oo et al[110], successful homing of Treg cell was seen in the hepatic parenchyma following intravenous infusion. Treg cell therapy is an emerging cell-based immunotherapy in patients with refractory AIH[111,112] (Figure 1).

IL-2 stimulates the proliferation of Treg cells. Low-dose IL-2 has been used in patients with refractory AIH[113,114]. Lim et al[115] reported two cases of refractory AIH, which were salvaged using low-dose IL-2.

Bacteriophages: Bacteriophages outnumber the intestinal bacteria and viruses[116]. They have an immunomodulatory role in the intestine and the liver[116]. Phages have been shown to reduce the expression of Toll-like receptor 4 and downregulate nuclear factor B to attenuate the inflammatory pathways and cellular damage[117]. Recent data suggest that phages may ameliorate the immune-mediated tissue damage in various autoimmune disorders and nonalcoholic steatohepatitis[118,119]. Prospective data on the role of phage therapy in patients with refractory AIH are needed to establish its clinical efficacy.

CONCLUSION

An individualized approach should be followed in managing patients with difficult-to-treat and refractory AIH. Second-line therapy is reserved for patients who fail to respond to initial immunosuppression therapy. Patients who are intolerant to AZA can be managed with MMF or 6-MP. However, AZA-non-responders can be salvaged with MMF or a third-line immunosuppressive drug. Genetic polymorphism involving TPMT and NUDT15 may affect the metabolism of AZA and 6-MP, and screening for these mutations is recommended. Metabolite assays and adding allopurinol can be considered when patients experience secondary loss of response with AZA and 6-MP. Patients with concomitant autoimmune diseases can be managed with IFX, belimumab, or rituximab. Frequent monitoring for adverse events is recommended while using immunosuppressive therapy. Worsening liver enzymes should prompt physicians to look for DILI. Few drugs used to treat AIH have been associated with variable degrees of liver injury; i.e., AIH-like DILI (IFX), nodular regenerative hyperplasia (AZA), and hepatic fibrosis (MTX). Data on the safety and efficacy of sirolimus, 6-MP, belimumab, and MTX come from a small number of patients. Large prospective data are needed for newer second- and third-line drugs for managing refractory AIH. Tofacitinib is a promising experimental drug. Low-dose IL-2, phage therapy, and FMT have been tried in patients with difficult-to-treat AIH; however, their role is yet to be established in prospective clinical data.

ACKNOWLEDGEMENTS

The authors express their heartfelt gratitude and thank all the individuals for their tireless efforts, insightful comments, suggestions and intellectual assistance to this work.