EASL Clinical Practice Guidelines

Special patient populations


For patients with stable AIH, the issue of conception and pregnancy frequently arise. An evolving literature on which to base recommendations now exists.

In a large series of patients that attended King’s College Hospital between 1983 and 1998, 18 patients had 35 pregnancies; 31 live births were reported with birth abnormalities seen in only two cases [82]. Flares in disease activity occurred during four pregnancies and in additional four within three-months of delivery. In an expanded series from the same institution, 81 pregnancies were reported in 53 women with 41% of pregnancies occurred in the context of cirrhosis [81]. At conception, 61 pregnancies (75%) were on therapy for AIH and 75% of patients were receiving pharmacotherapy. Of these, 27 patients were on prednisolone monotherapy (mean dose 10 mg/day, range 2.5 mg–40 mg), seven were on azathioprine monotherapy (range 1 mg/kg/day–2 mg/kg/day) and 25 patients were on combination therapy of azathioprine (range 1–2 mg/kg/day) and prednisolone (mean dose 5 mg/day, range 2.5–20 mg). In addition, one patient received tacrolimus (2 mg/day) in conjunction with prednisolone [81]. Among those patients on medication, 46 (74%) were stable on their medication regimen for over one year prior to conception.

The live birth rate (LBR) was 73% (59/81). Prematurity occurred in 12/59 (20%) and six infants (11%) required admission to special care baby unit (SCBU). In mothers who were cirrhotic at the time of conception the LBR was lower and need for admission to SCBU was higher. The overall maternal complication rate was 31/81 (38%) conceptions. A flare in disease activity occurred in 26/81 (33%) pregnancies. A serious maternal adverse event (death or need for liver transplant) during or within 12-months of delivery, or hepatic decompensation during or within three-months of delivery, occurred with nine pregnancies (11%) and was significantly more common in women with cirrhosis. Maternal therapy had no significant impact on the LBR, termination rate, miscarriage rate or gestational period. Flares of the underlying AIH were more likely in patients who were without therapy or who had a disease flare in the year prior to conception. Patients who had a flare in association with pregnancy were more likely to decompensate from a liver standpoint. Importantly, no further birth abnormalities were reported beyond those reported in the original report [81].

In a German series, 42 pregnancies in women with AIH, 11 adverse outcomes were reported with serious maternal complications in four [238]. The unexplained adverse outcomes were associated with the presence of anti-SLA/LP and anti-Ro/SSA antibodies. Flares during pregnancy occurred in 21% of patients, whereas 52% of patients had post-partum flares. In a survey of 63 pregnancies in patients with AIH, higher caesarean section rate but no increase in stillbirth or fetal malformation rate was observed [239].

In all these large series, no apparent relationship existed between azathioprine use during pregnancy and an adverse outcome. In large studies of patients with inflammatory bowel disease (IBD), the relative safety of azathioprine or 6-MP during pregnancy has also been well established [240]. Indeed, results of a meta-analysis evaluating the outcome of women with IBD exposed to thiopurines, the pooled odds ratio for low birth weight, preterm birth, and congenital abnormalities were 1.01 (95% CI (confidence interval) 0.96, 1.06), 1.67 (95% CI 1.26, 2.20), and 1.45 (95% CI 0.99, 2.13), respectively. In men, the pooled OR for congenital abnormality was 1.87 (95% CI 0.67, 5.25) [241].

Therefore continuation of this drug during pregnancy appears to be justified. Moreover, in a small case series of 14 patients in whom immunosuppression was reduced during pregnancy, following delivery (or stillbirth in one patient), 12/14 patients had a rapid flare of AIH [79].

AIH may also present for the first time during pregnancy or (more frequently) in the immediate post-partum period [[80], [242]]. Index presentations should be treated in the same fashion as the non-pregnant patients. Overall, the available large series support a strategy of minimal adjustment to standard immunosuppression (prednisolone/azathioprine) during the course of pregnancy so that the risk of flare can be minimized both during pregnancy and post-partum. Similar, due consideration should be given to calcineurin inhibitor (CNI) therapy, although exact data pertaining to the use of ongoing CNI use in pregnancy can only be derived from transplant series [[79], [80], [243]].

The final decision to modify immunosuppression either pre-conception or during pregnancy should be based on the perceived risk to the patient and the pregnancy. Unquestionably, patients with established cirrhosis are at greatest risk of adverse outcomes both during pregnancy and in the first year following delivery and therefore this group should have closest attention during this time. Notwithstanding the fact that azathioprine is considered a category D drug by the US Food and Drug Administration [244], azathioprine and 6-MP have both demonstrated relative safety in pregnancy. In contrast, MMF which is in the same category, has significantly greater risk of teratogenicity and should be withdrawn prior to considering pregnancy and avoided entirely [245].

With regard to breastfeeding in the context of autoimmune hepatitis and immunosuppression, little data exist. The data available is derived from use for other conditions such as IBD or rheumatological disorders. In general, azathioprine or 6-MP is considered safe for breastfeeding although small amounts of metabolite can be detected in breastmilk. However, this does not appear to result in complications for the feeding infant [[246], [247]].


As discussed in section ‘Epidemiology of AIH’, AIH is seen in all ages and races, and a considerable cohort of patients is increasingly derived from transition or adolescent clinics. The general principles of management of AIH presenting in childhood are similar to those presented in adult patients with some caveats. Indications for therapy are similar [248]. Multiple large series have demonstrated that the disease entity appears to be more aggressive at presentation than that seen in adults with AIH. Whether this relates to the presence of other autoimmune diseases, delays in diagnosis or genetic susceptibility is unclear [249]. Similarly, the potential overlap with an alternative disease entity, specifically AISC complicates the diagnostic pathway [[52], [248], [250], [251]].

Since at diagnosis, more than 50% of children will have evidence of cirrhosis, and the milder forms of disease are not usually seen, this justifies initiation of early treatment following diagnosis [[52], [248], [250], [251]]. The aggressive course of disease and reports that delays in diagnosis may affect prognosis, legitimizes early drug therapy. As with adult AIH, treatment may be withheld in only rare circumstances, most specifically in the context of burnt out cirrhosis without inflammatory activity, and only in consultation with a specialised hepatologist.

Treatment regimens in childhood AIH

In comparison to adult AIH, treatment regimens in children have been derived primarily from large single-centre practices and reflect the experience and experience bias of these institutions [[52], [248], [250], [251], [252], [253], [254], [255], [256], [257]]. Until recently, no randomised trial in the management of AIH in childhood had been undertaken [258]. However, despite the lack of controlled trials in children with AIH, multiple reports have documented the efficacy of induction treatment regimens at a level similar to adult patients with normal liver enzymes in up to 90% of patients after six-nine months of therapy [[52], [248], [250], [251], [252]].

Until now, prednis(ol)one has been the mainstay of therapy in virtually all reported regimens for children, and it is typically administered initially in a dose of 1–2 mg/kg daily (up to 60 mg daily) [[52], [248], [250], [251]]. Variability exists regarding the tapering of steroid dose. Opinion varies regarding this aspect of care with some institutions advocating rapid switch to alternate day steroid therapy in order to minimize the effects on growth retardation. In contrast, other centres favour a low dose daily schedule of prednis(ol)one.

Since a concern exists for physicians, parents and children of high dose therapy on linear growth, bone growth, and cosmesis, early introduction of azathioprine (1–2 mg/kg daily) or 6-MP (1.5 mg/kg daily) for all children is usually recommended unless contraindications exist [[52], [248], [250], [251]].

A recent important study evaluated an alternative induction regimen in children and adolescents with AIH [258]. It currently represents the only double blind, controlled trial of therapeutics in AIH in childhood. Using budesonide as an alternative to prednisone, 46 patients were enrolled in a six-month, prospective, double blind, randomised, active-controlled, multicentre phase IIb study that compared budesonide in 19 patients (dosed at 3 mg twice or three times daily) with 27 patients that received prednisone (40 mg/day tapered to 10 mg/day). Both groups received azathioprine (1–2 mg/kg/day), followed by a further six months of open-label budesonide therapy.

The primary efficacy endpoint was defined as complete biochemical remission (normal serum ALT and AST) without predefined steroid specific side effects. The results identified no statistically significant difference in the percentage of patients who met the primary endpoint between the budesonide (3/19; 16%) and prednisone groups (4/27; 15%) after six months of treatment, nor in the percentage of patients who experienced biochemical remission (budesonide, 6/19 [32%]; prednisone, 9/27 [33%]), lack of steroid specific side effects (budesonide, 10/19 [53%]; prednisone, 10 of 27 [37%]). Overall, there was significantly less weight gain in the budesonide group (1.2 ± 3.5 kg vs. 5.1 ± 4.9 kg in the prednisone group (p = 0.006). In the subsequent six-month open-label limb of the study, a total of 42 patients received open-label budesonide after which, 46% of these patients achieved complete remission. Based on these data, oral budesonide with azathioprine can both induce and maintain remission in paediatric patients with AIH and may be considered an alternative therapy to prednisone in non-cirrhotic patients. Long-term effects of this regimen on bone growth and linear growth remain to be assessed.

The use of azathioprine in isolation as a maintenance regimen in children is limited. In five eligible patients, who were withdrawn from steroids 4/5 remained in remission [252]. Cyclosporine A has also been used as alternative induction regimen, although, no advantage to conventional therapy can be identified over standard initiation regimens [255]. MMF has been utilised in children as a rescue regimen in unresponsive disease [253]. Of 26 patients with autoimmune liver disease, 18 responded with 14 having complete normalisation of transaminases in follow-up [253]. Interestingly, of the non-responders, six had AISC.

Like all adult patients, all children should be assessed for evidence of immunity against hepatitis A and hepatitis B virus (HBV) infection and vaccinated accordingly [259]. Similarly, assessment of bone density should be undertaken at diagnosis and during follow-up as described in adults.

Co-morbidity and old age

Individualisation of therapy may be appropriate for certain groups of patients. In making the decision to devise strategies in management, one needs to consider the presence of co-morbidity in conjunction with the severity of disease and the goals of treatment. In that regard, it is possible to identify certain categories of patients, i.e. patients with evidence of osteoporosis at disease onset, patients with established diabetes mellitus and metabolic syndrome, patients with co-existing viral hepatitis such as HBV or HCV infection. In a similar way, consideration needs to be given to strategies of treating patients who present in older age, who may have a more responsive disease process and who will also have less symptoms at onset of disease [[35], [44], [260]].

Old age

Older patients are often less symptomatic at presentation [[35], [44]] and will also be more biochemically responsive to therapy [[35], [260], [261]]. Moreover, genetic influences in particular a higher prevalence of the HLA DR4 alleles in this group seems to at least contribute in part in Western populations to this clinical phenotype of more responsive disease [260]. In a large US single-centre experience, the prevalence of cirrhosis at accession was identified to be 33% in patients >60 years compared to a prevalence of 10% in a comparison group of patients who were aged <30 years at presentation [260]. Similar findings regarding the prevalence of cirrhosis or more advanced fibrosis (F3) at accession have been reported from Japan [261]. These patients also had a higher prevalence of concurrent autoimmune disease [[260], [262]]. In these studies, patients >60 years were all treated with standard regimes of corticosteroids and azathioprine.

In contrast to the elderly patient with established cirrhosis or marked inflammation, an unresolved question particularly among elderly patients, is the presence of mild interface activity with low necroinflammatory scores on liver biopsy [263]. In that context, an unanswered question is whether these patients should be treated, especially if other co-morbidity exists. In a series of patients from the 1970s, ten-year survival was 90% [180], although a more recent study from the Mayo Clinic described a range of outcomes including progressive liver failure, ascites, and HCC during a follow-up interval of ten years (range, 2.7–19.9 years) [181]. These untreated patients with mild disease also were less likely to improve (12% vs. 63%) and had a ten-year survival of 67% compared to a treated control group whose survival was 98% [181]. It is noteworthy, however, that these outcomes reflect only a small number of patients. In a Canadian study, the absence of symptoms at presentation did not translate to poorer outcomes when compared to treated patients, although in follow-up, 25% of patients developed symptoms [44]. Pragmatically, clinical judgement is required if embarking on a watchful waiting strategy. Close follow-up of these patients is essential, as activation and relapse can occur any time, occasionally decades after the initial presentation. If liver function tests remain abnormal or worsen, then repeat liver biopsy is appropriate and treatment initiation should be undertaken.

In elderly patients or in those with co-morbidity, the choice of steroid therapy should be considered carefully. For patients without cirrhosis, when severe steroid–related side effects are likely to exacerbate poorly controlled diabetes, osteoporosis, and psychosis, budesonide 9 mg/day plus azathioprine 1–2 mg/kg/day may be an appropriate choice [193]. Although, these endpoints have not been evaluated at a long-term, current management would favour this approach. Clearly attempts at early steroid withdrawal should be undertaken.

Presence of osteopenia/osteoporosis

In the classic early standard immunosuppression trials in AIH, side effects were particularly problematic in the steroid-only regimens. The presence of Cushingoid facies and buffalo hump were reported in up to 50% of patients, diabetes mellitus in between 15–20% of patients [[176], [177], [178], [186]]. Reports of hypertension, psychosis, cataract development, osteoporosis and vertebral collapse related to osteoporosis were of the order of between 5 and 10% and although less commonly seen on combined regimens, the prevalence of these side effects is nonetheless approximately 5% [[176], [177], [178], [186]].

Worldwide, it is estimated that over 200 million individuals suffer with osteoporosis with the major complication relating to increased bone fragility and subsequent reduced quality of life, morbidity and mortality [264]. Although steroid induced osteoporosis represents only a fraction of the osteoporotic population, its impact is important in terms of complications. For example, patients with AIH who may receive several courses of high dose steroids (daily dose ⩾15 mg and cumulative exposure >1 g) have a substantially increased risk of fracture [265]. There is a considerable scope to intervene in susceptible patients using preventive measures. Moreover despite numerous data being available in relation to intervention to bone health, it is estimated that only between 5 and 62% of patients on glucocorticoid therapy in the United States and Europe receive appropriate preventive therapies [266].

In 2003, the combined guidelines of the American College of Gastroenterology (ACG) and the American Gastroenterological Association (AGA) were published, recommending lifestyle modifications and medical therapy, including supplemental calcium, vitamin D and bisphosphonates, to prevent or treat bone loss in patients who receive long-term steroid therapy [267]. Indeed implementation of these guidelines concurrently in a population of patients with IBD led to the detection of osteopenia and osteoporosis with initiation of specific therapies in a majority of gastroenterology clinic patients who met the guidelines’ criteria for dual energy X-ray absorptiometry (DEXA) scanning [268]. Interestingly, evaluation of a similar type of strategy at a US based Gastroenterology and Hepatology tertiary centre suggested that tertiary care physicians were most likely to recommend bone health medications, lifestyle changes or DEXA scan to patients if they had AIH, IBD, or were current smokers, had liver disease or had a history of osteoporosis or osteopenia. Remarkably, previous fracture, advanced age, steroid use and reduced body mass index were not associated with tertiary physician bone health recommendations in the study [269].

Pragmatically, and consistent with previous recommendations relating to cholestatic liver disease and in particular PBC, for patients with AIH bone mineral density assessment (DEXA) is a useful guide for treatment and should be undertaken when possible in probably all patients at presentation, along with follow-up assessment between one and five years depending on outcome and general osteoporosis risk, although, no specific data can be used to support this assertion. The use of calcium and vitamin D supplements is supported by epidemiological data with proven reduction or reversal in the natural rate of bone loss, but no trial data supports or refutes this approach in AIH, while as a general principle hormone replacement therapy is effective in post-menopausal female patients. There are trial data to support the use of bisphosphonates (particularly alendronate) when osteoporosis is present [[270], [271]].