The use of 6-mercaptopurine (6-MP) in AIH is basically reserved for patients with AZA intolerance which is defined as any adverse event possibly related to the treatment as assessed by the treating physician leading to potential discontinuation of the drug [14]. As mentioned above, more than 25% of AIH patients experience side effects from AZA, leading to drug discontinuation in most of them. According to a recent position statement on second- and third-line treatments [24], the use of 6-MP may be attempted, as 50–75% of patients intolerant to AZA are able to adequately tolerate 6-MP. However, this option has not been endorsed by the recent the American Association for the Study of Liver Diseases (AASLD) clinical practice guidelines [2] and other centres, as this recommendation was mainly based only on a small retrospective study with 40% of patients achieving CBR (8 out of 20 patients) [25]. The recommended initial dose of 6-MP ranges from 0.5–1.0 mg/kg per day and in case of good tolerance the titration should be according to the 6-thioguanine nucleotides (6-TGN) blood levels which however, is not widely available in everyday practice [4, 25].

Mycophenolate mofetil (MMF) is the first selective, potent, reversible, and non-competitive inhibitor of type II isoform of inosine-5’-monophosphate dehydrogenase discovered more than 100 years ago, exerting antiviral, anti-fungal, antibacterial, anti-tumor, and immunosuppressant properties [26]. As a result, MMF leads to selective immunosuppressive effect by depleting guanosine-triphosphate specifically in activated B- and T-cells without affecting the type I isoform and therefore, resulting in low rates of adverse events [5, 26]. The administration of MMF is principally justified as second-line treatment in patients intolerant to AZA according to the European Association for the Study of the Liver (EASL) and AASLD clinical practice guidelines as well as a recent position statement [2, 4, 24], even though recent systematic reviews and meta-analysis from the Australian Liver Association Clinical Research Network [27] and US [28] showed that MMF led to response rates ranging from 32.0–57.1% in insufficient responders to AZA and 62–82% in intolerant patients.

However, some national guidelines propose the use of MMF as an alternative first-line agent based on its high rates of CBR among treatment-naive patients and its low rates of adverse events and/or intolerance (< 10%) [5]. Indeed, the Hellenic Association for the Study of the Liver (HASL) has included since 2019 [5], apart from AZA, MMF as an acceptable first-line treatment option for induction and maintenance of response in AIH patients. The latter recommendation was based on several prospective cohort studies and a recent meta-analysis which showed that MMF could be an alternative and safe first-line treatment option for AIH patients [29–32].

Additionally, a recent strict prospective propensity matching trial of MMF versus AZA as first-line treatments, found that MMF was associated with higher rates of CBR bearing a higher overall efficacy than AZA, even when patients who were intolerant or unresponsive to AZA were included in the analysis [33]. Regarding the long-term response, a recent study from the same group with 21 years of follow-up, has shown lower primary non-response rates at week 4, lower rates of intolerance and serious adverse events, and more frequent eligibility to stop immunosuppression compared to those treated with standard of care in MMF group compared to the AZA group, while 78% of treatment-naive MMF-treated patients-maintained remission off prednisolone [22]. Furthermore, results from the unique controlled randomised trial of AZA versus MMF (CAMARO trial) [34] for the induction of response in treatment-naive AIH patients showed very recently, superiority of MMF to achieve CBR at 6 months (72.2% vs. 32.3%; P = 0.004) with significantly lower adverse events [35]. Taking together, these results indicate that the new guidelines will probably change in the near future both for first- and second-line treatments.

Up to the present, several studies have evaluated the efficacy of MMF in patients with AZA intolerance/insufficient response and/or as first-line option in treatment-naive patients (Table 1). Most of these studies have shown high rates of CBR, which in most cases are comparable or even better than those reported with AZA, ranging from 60–97% [22, 29–31, 33, 35–44]. On the contrary, two older studies have reported lower rates of complete response or even treatment failure following MMF initiation [45, 46]. The latter findings could be attributed to the heterogeneity of the study populations and the variability of study protocols (combination or not with corticosteroids, corticosteroids dose, duration of treatment, etc.).

To sum up, MMF seems an effective and safe alternative first-line treatment in AIH but can also be considered as second-line drug in patients with poor response or intolerance to AZA [22, 24, 35].

TAC is a calcineurin inhibitor with greater immunosuppressive activity than cyclosporine. A number of studies have been conducted in AIH, evaluating the use of TAC in the setting of treatment failure, incomplete response, and AZA intolerance. In all of them TAC has been administered in combination with corticosteroids and AZA or MMF. Biochemical response rates vary amongst different studies probably because of the difference in dose and trough levels. One multicentre study reported biochemical response (normalization of transaminases) in 94.1% of patients who were intolerant to standard treatment and 56.5% in those who had insufficient response to previous therapy [43]. The respective results in the same study for MMF administration were 91.9% and 34% [43]. Based on these results the AASLD recommended both MMF- and TAC-based therapies as appropriate second-line treatments [2].

Other older and smaller studies observed improvement in liver tests in most of the patients, but without achieving CBR [37, 47–52]. Regarding safety, dose reduction or discontinuation occurred in approximately 25% of cases and the most frequently reported side effects were neurological symptoms and mild increase in creatinine levels, none of these leading to treatment discontinuation [48, 49, 52, 53].

However, EASL, HASL, the European Reference Network on Hepatological Diseases (ERN RARE-LIVER), and the International Autoimmune Hepatitis Group (IAIHG) do not recommend TAC as a potential second-line option in AIH patients [4, 5, 24] because the abovementioned large study by Efe et al. [43] was a retrospective analysis of very heterogeneous group of patients treated with multiple treatment schedules and with irregular follow-up data. Recently, there are several similar studies in the literature, with more systematic data collection [54–56]. The feeling of these studies is that calcineurin inhibitors are very potent drugs but require close monitoring of drug levels because of a quite small therapeutic window while toxicity remains considerable. Actually, a recent two-centre study confirmed the limited effectiveness and the risks of this kind of therapy in AIH [57]. In the previous study [43], only neurotoxicity was mentioned, but renal injury, diabetes mellitus, hypertension, hyperlipidaemia, and malignancies also should be mentioned. Therefore, as in other immune mediated diseases, MMF can be tried in all AIH patients who are intolerant/insufficient responders to thiopurines and third-line treatment with TAC must be administered in cases with ongoing insufficient response. According to experts’ opinion, standardization of the dose so as to achieve higher initial trough levels (6–8 ng/mL) and tapering of trough levels after response has been achieved is strongly recommended [24].

Experience in the use of cyclosporine comes mainly from the paediatric population, where cyclosporine is principally used in severe disease or to prevent steroid side effects. Evidence of cyclosporine in adults with drug intolerance or insufficient response to first- or second-line treatments comes from old and small case series with biochemical response (normalization of transaminases) rates around 80% [58–60]. However, these numbers change when patients who are treatment intolerant and non-responders are simultaneously analysed. In the most recent study, including larger number of patients, response rate was 60% [61]. Therefore, it is clear that the limited number of patients and the heterogeneity of the cohorts preclude the extraction of safe conclusions.

Taking into consideration the abovementioned data, cyclosporine remains a potential third-line option in AIH. As far as the treatment schedule, recent suggestion by the ERN RARE-LIVR and IAIHG is the use of cyclosporine at a dose of 2 mg/kg per day twice daily and serum trough levels < 120 ng/mL [24] but renal function should be assessed prior to and during treatment.

TNFa inhibitors may also have a therapeutic role in AIH, although still their use remains a matter of debate. In a small single-centre retrospective study from Germany, infliximab was administered in 11 difficult to treat AIH patients including seven patients with established cirrhosis. The dose used was 5 mg/kg at day 0, week 2, and week 6 and every 4–6 weeks thereafter according to laboratory and clinical evolution of each patient. Among patients included, 6/11 (54%) achieved CBR [62]. However, a considerable number of patients (7/11, 64%) developed infectious complications needing hospitalization, while treatment had to be withdrawn in 3 cases because of side effects [62].

Although this study supports the role of anti-TNFa in the pathophysiological basis of AIH management, it is well known that TNFa blockade can sometimes result in exactly the opposite outcome inducing an immune-mediated liver disease resembling AIH or even true AIH [63–67]. Indeed, a large international pharmacovigilance database including 389 patients with anti-TNFa blockade-associated AIH, reported that infliximab was the most frequently associated drug [68]. In other words, the use of TNFa-blockade drugs seems like a “two-edged sword” management because of the unforeseen safety of these agents. This therapeutic paradox is mainly attributed to the disruption of the regulatory role of TNFa signalling throughout the immune system. Therefore, the AASLD guidelines have considered infliximab as a drug with definite association with AIH development [2].

Considering the above-mentioned data, infliximab may do have a role in the management of refractory AIH, however due to the probability of unforeseen complications more evidence regarding its efficacy and safety are warranted.

Everolimus and sirolimus inhibit the mammalian target of rapamycin (mTOR) modulating the proliferation and survival of activated lymphocytes. The use of both drugs is better studied for preventing rejection in solid organ transplantation. The experience of their use for poor responders to standard treatment in AIH is very limited.

Only one report is available about the use of everolimus in AIH, which consists of a retrospective analysis of 7 patients with intolerance or insufficient response to second- and third-line (calcineurin inhibitors) agents. All patients had improvement of liver biochemistry with trough levels ranging between 3–6 ng/mL, while 3/7 of patients achieved CBR after 5 months of treatment with no severe side effects [69].

The efficacy of sirolimus in AIH has been reported in only 5 AIH patients refractory to AZA and MMF. CBR was achieved in 2/5 of patients with serum trough levels between 10–20 ng/mL, while side effects occurred in 2/5 of patients [70]. However, a report of two additional cases of refractory AIH was more disappointing as in one case sirolimus was discontinued due to side effects, while in the second case biochemical response was not achieved [71].

So, based on these data the use of mTOR inhibitors in AIH is not recommended, until more data regarding their use are collected.

MTX is a folic-acid antagonist used mainly in the treatment of rheumatic diseases such as rheumatoid arthritis and lupus, but also as anti-neoplastic drug. Data on efficacy and safety of MTX in AIH patients are limited to case reports and one case series [72, 73]. The unique case series published from Haridy et al. [74], reports the use of MTX in 11 patients with AIH previously unresponsive to first-line treatment. The authors reported CBR in 54.5% of patients treated with MTX at a median dose of 10 mg/week during 36 months of follow-up as well as, a significant reduction of corticosteroids dosage [74]. However, nearly half of the patients (5/11, 45.5%) had to stop treatment during the first 12 months because of an elevation in liver enzymes, which in two cases occurred during the first 3 months after treatment initiation and was considered as DILI mediated by MTX [74].

So, MTX could serve as third-line treatment in AIH, but we need more data for clear-cut recommendations. In addition, strict surveillance is recommended since current data on safety and efficacy are extremely limited and there is a high risk of adverse events including the probability of DILI development.

Rituximab is a monoclonal antibody directed against B-cell surface receptor CD20. The use of rituximab for refractory AIH was first reported in a case of a 68-year-old woman with Waldenström macroglobulinaemia and diagnosis of AIH with poor response to induction treatment with corticosteroids. Rituximab was used for the induction of response which was afterwards maintained with low dose corticosteroids in combination with AZA for a long follow-up period of 4 years [75]. In a single centre open label pilot study from Canada, including 6 patients with either refractory disease (3/6) or intolerance to first-line treatment (3/6), rituximab was used with good tolerance and rapid improvement of liver biochemistry and IgG in all subjects by 12 weeks of drug infusion. Biochemical response and B cell depleting effect were maintained throughout 72 weeks of follow-up [76]. Moreover, recent results from the IAIHG in 22 patients with AIH refractory to second- and/or third-line therapy showed significant improvement of transaminases that was maintained during the 24 months of follow-up. In addition, both prednisone doses and flares were reduced (71% of patients without disease flare at 2 years) [77]. However, CBR (normalization of both transaminases and IgG levels) was not achieved perhaps because only two doses, 2 weeks apart, were administered to the patients.

According to these data, rituximab seems to be a safe and effective treatment option for refractory AIH in combination with standard treatment. The recommended dose based on published data for adults is 1,000 mg at week 0 and 2, to be repeated afterwards according to liver biochemistry (response guided). Serious adverse events seem to be rather infrequent, however, surveillance of B-cells depletion (particularly CD20 cells) is recommended as well as supplement with immunoglobulins when it is considered necessary [24].

B cell activating factor (BAFF) is a cytokine belonging to the TNF family expressed by T and dendritic cells and is proven to be crucial for the development and differentiation of B cells. BAFF levels have already been shown to be associated with liver inflammation and to be reduced in AIH patients after response to corticosteroids. Thus, inhibition of BAFF could be a pathogenetically justified third-line treatment option in AIH [78].

Belimumab was first administered as third-line agent in combination with corticosteroids and MMF in 2 patients with advanced fibrosis/cirrhosis and refractory disease. Results were more than encouraging as both patients achieved CBR with histological reversion of cirrhosis in one available liver biopsy while both remained in CBR under low-dose corticosteroids, without serious adverse events [79]. Recently, a small case series of 6 patients with AIH and/or primary biliary cholangitis (PBC) and other concomitant autoimmune diseases, treated with belimumab was published [80]. In this retrospective analysis an improvement of liver biochemistry was reported after belimumab initiation along with reduce of corticosteroids doses in patients with AIH, whereas no improvement of liver function tests was observed regarding patients with PBC [80]. However, due to the heterogeneity of the cohort, the co-existence of other autoimmune diseases and the combination of different immunosuppressive agents in each case, these results should be interpreted with caution.

Beyond their limitations, these preliminary findings indicate belimumab as a promising treatment option for patients with AIH and refractory disease. Currently, the effectiveness of lanalumab, another BAFF receptor inhibitor, in treating patients with AIH who are not responding to conventional medication was investigated in a multicentre phase II/III trial (NCT03217422) and results are going to be presented soon.