Historically, the presence of autoantibodies with elevated serum immunoglobulin in adult patients with active chronic hepatitis (ACH) was able to predict the favorable response to immunosuppression and ultimately lead to the description of AIH
[30,31,32][10][11][12]. Soon after it was noted that some patients with autoantibodies-negative ACH or with cryptogenic chronic hepatitis also had a comparable response to immunosuppression
[33][13]. A systematic review of patients successfully treated with immunosuppression led to the full description of the so-called seronegative AIH (sAIH). Patients with sAIH were found to be undistinguishable from those with “seropositive” AIH in terms of demographic, laboratory and histological features as well as in terms of treatment outcomes
[34][14]. Moreover, this group of patients shared several features of immune dysregulation with seropositive-AIH including hypergammaglobulinemia, association with characteristic HLA haplotypes, propensity to relapse and seroreactivity against investigational autoantigens such as auto-antibodies against the liver membrane lipoprotein preparation known as liver-specific membrane lipoprotein (LSP) or against the hepatic asialoglycoprotein receptor (ASGP-R)
[35][15]. Frequency of sAIH varied from 5 to 35% of a reported cohort of AIH being higher in those presenting with acute or fulminant hepatitis
[36][16]. Criteria for diagnosis were limited to the finding of suggestive histological features in patients with acute or chronic hepatitis once viral, toxic or metabolic causes had been excluded
[3]. In the 2019 AASLD guidelines, sAIH was finally endorsed and its prevalence was estimated as 20% of AIH
[3]. More recently sporadic cases of antibody-negative disease were reported in children and adolescents with concomitant celiac disease
[37,38][17][18].
5. Long-Term Follow-Up
JAIH is a severe disease with high-risk of relapse due to poor compliance
[64,65][26][27] and wide variability of reported sustained remission rates after treatment withdrawal ranging from 3 to 87%.
[66,67][28][29]. In children, the official recommendations prior to attempting treatment withdrawal include normalization of transaminases for 2 to 3 years, normal serum IgG, negative or low titres of serum autoantibodies and no inflammation on liver histology
[2]. In adults, on the other hand, liver biopsy is no longer considered a mandatory prerequisite for treatment withdrawal
[3]. In addition, there are only limited pediatric reports on the long-term follow-up beyond 10 years of management
[67,68][29][30].
WResearche
rs recently published data on a long-term observational study of 117 children with AIH, excluding fulminant, seronegative, drug-induced hepatitis and sclerosing cholangitis, who were diagnosed and treated in a single pediatric hepatology center in France
[51][31]. Cirrhosis was present in 80 children at diagnosis. Management consisted of immunosuppression with prednisone with/without azathioprine in most. Attempts at treatment withdrawal under medical supervision were carried out in one of two ways: before 1981, withdrawal was performed after liver histology with a combination of normal aminotransferase activity and serum gamma globulins. Later, having found that liver histology did not reliably predict the lack of relapse after withdrawal
[48][32], another approach was undertaken: when alanine aminotransferase activity remained normal on therapy for at least 1 year, no liver biopsy was performed, and prednisone was progressively decreased over one year, while checking aminotransferases for normalization before each new decrease. Once prednisone was stopped, azathioprine was continued for another 6 months and then stopped if aminotransferases remained normal. A complete biochemical response was defined as normalization of serum alanine aminotransferase and gammaglobulins no later than 6 months after initiation of treatment following the recent recommendations of the IAIHG
[69][33]. Data were available until death and for 8 to 38 years (median, 20 years) after starting treatment in surviving patients.
[51][31].
6. JAIH and Future Pregnancies
JAIH affects a number of young women of childbearing age, some of whom have a long-lasting history of disease and immunosuppression
[77][34]. Moreover, with at least one third of patients having cirrhosis at diagnosis, the population of pregnant women with AIH is very heterogeneous in terms of severity of disease and treatment regimens. There are many aspects of JAIH that may affect pregnancy: presence or not of cirrhosis and related complications; presence of potentially dangerous comorbidities due to concomitant autoimmune diseases such as type 1 diabetes or antiphospholipid syndrome; biochemical remission or active disease at the beginning of pregnancy; ongoing immunosuppression regimen and its potential toxicity or teratogenicity for the fetus. The literature addressing this issue can be roughly divided into two categories: population-based cohort studies
[78,79,80][35][36][37] and single-center retrospective reports
[81,82,83,84,85][38][39][40][41][42]. In population-based studies data, are retrieved from health registries via the ICD code, in one case matched with histopathology report data
[78][35]. The major pitfalls of these studies are the limited amount of clinical information available and possible coding errors. On the other hand, accurate epidemiological analysis could be conducted thanks to the huge size of the study cohort and possibility to build a control group. Single-centers experiences are valuable because diagnosis is well-documented and clinical details are usually available. Limited size and retrospective nature are the limitations. Fertility in AIH depends on disease activity and it is decreased in poorly controlled AIH
[83,85][40][42] and in cirrhotic patients, 10% of whom are reported to use assisted reproductive technology
[83][40]. During pregnancy, the increase in estrogen and progesterone induces maternal immunity to enter a tolerogenic status to safely host the semi-allograft fetus. Th0 cells are stimulated to differentiate into Treg and Th2 phenotypes that produce high levels of IL-10, IL-4, IL-5 and TGFβ, anti-inflammatory and tolerogenic cytokines
[86][43]. However, after delivery, pregnancy hormones levels abruptly fall and the reactivity of immune system is rapidly restored. Average transaminase levels in AIH patients are significantly lower during gestation, as reported in a recent metanalysis
[87][44].
Loss of remission is rare during pregnancy, although it has been reported in as many as 10–15% of cases in the same studies
[81,83][38][40]. Relapse after delivery, on the other hand, has been reported to occur in up to 50% of cases
[83][40]. More recent studies, however, report lower post-partum relapse rates, which are likely related to the improved knowledge on drugs’ safety and general AIH management during pregnancy. It is known that patients who are in remission before conception have a lower risk of post-partum relapses
[87][44] and that poorly controlled AIH during pregnancy is associated with a greater incidence of adverse outcomes
[81][38]. In terms of maternal outcomes, more recent population studies have not shown an increased rate of caesarean sections in AIH women
[78,79][35][36]. No difference was found in the incidence of maternal death excluding patients with decompensated cirrhosis
[87[44][45],
88], while almost all studies report an increased risk of gestational diabetes
[78,79,80][35][36][37]. Even patients who are off treatment during pregnancy have a higher risk of diabetes
[79][36], suggesting that it is not related to use of steroids but rather to the implicit association of diabetes with other autoimmune diseases. An increased risk of gestational hypertension, preeclampsia, eclampsia and HELLP syndrome was reported in many single-center studies and population-based studies in the US and Sweden
[78,79][35][36]. It is still debated whether to consider AIH patients as belonging to the high-risk group for other inflammatory diseases and whether low-dose aspirin prophylaxis should be recommended. In terms of fetal outcomes, preterm birth was more common in AIH than in the general population, with a rate of 9–20%
[78,79,80][35][36][37]; interestingly, immunosuppression or cirrhosis did not increase this risk. Conversely, the occurrence of a relapse during pregnancy was associated with an increased risk of preterm birth and higher rate of admission to a neonatal intensive care unit
[79][36]. Birth weight < 2500 g but not small-for-gestational-age newborns were associated with AIH in the US and Sweden
[78,79][35][36] suggesting that low weight is related only to the greater proportion of preterm. No differences in terms of risk of congenital malformations, neonatal mortality and stillbirth were found in any recent population-based studies
[78[35][36][37][44],
79,80,87], but only in older ones, which report a rate of fetal loss of up to 27%
[83][40]. The presence of anti-phospholipids and anti-Ro/SSA antigens must be monitored for they carry the risk of adverse outcomes. Only the presence of cirrhosis was also associated with an increased risk of miscarriage in more recent studies
[81,84][38][41]. There is increasing evidence that remission of disease should be achieved before conception and maintained during pregnancy. A safe pregnancy should be programmed; immunosuppression should not be weaned during gestation with low-dose steroids and azathioprine being safe during pregnancy
[85,89,90][42][46][47]. Breastfeeding is also considered safe. A small amount of both drugs can be detected in breast milk but should not have any effect on the infant; however, it is recommended to wait four hours before nursing
[91][48].
Although ciclosporin exposure has been associated with hypertension and intrauterine growth retardation and tacrolimus with nephrotoxicity and gestational diabetes when used during pregnancy, calcineurin-inhibitors are now considered safe both during pregnancy and the breastfeeding period. Data are mainly derived from studies on transplanted patients and rheumatologic ones
[92][49]. It must be considered that pregnancy increases the unbound and active form of tacrolimus, resulting in the underestimation of the whole blood level. Consequently, the dosage should not be increased to avoid toxicity.
Mycophenolate is strongly discouraged due to its known teratogenicity and replacement with azathioprine should be considered
[93][50].
In conclusion, in the majority of women with AIH, pregnancy does not carry substantial risks but remission of disease is strongly recommended. Preconception counseling and multidisciplinary care during gestation and after delivery must be provided.