Treatment of autoimmune hepatitis: Budesonide does not solve our problems.

Abstract

When randomized trials were developed as an investigative tool, hepatologists with an interest in autoimmune hepatitis (AIH) were leading the field performing three convincing randomized trials, which demonstrated immunosuppressive treatments with steroids to be lifesaving.1 All three trials showed a marked reduction in the high mortality rate of untreated AIH by applying prednisone or prednisolone. Ever since, steroids have been and continue to be the drug of choice for remission induction, while azathioprine has become the drug of choice for maintenance of remission.2,3 Response to steroid treatment in AIH is so universal that documented treatment response is used as a diagnostic criterion. In everyday clinical practice, however, even though response rates are excellent, achieving the treatment target of histological and biochemical remission (normal transaminases and normal IgG) remains often a challenge. A significant proportion of patients show an insufficient response to therapy and continue to have significant inflammation.1 Furthermore, steroids are highly unpopular with patients, as steroid-driven treatment regimens frequently evoke side effects, at least during the first weeks of treatment. Early side effects of steroids are usually, but not always, transient and mild. Some patients require steroids for several years in order to lessen the disease, and in these long-term side effects such as osteoporosis, cataract formation, and skin affection represent a major treatment burden. Budesonide is a corticosteroid with a high first-pass effect in the liver. It is rapidly transformed by cytochrome P450 3A4 (CYP3A4) into the largely inactive metabolites 6β-hydroxybudesonide and 16α-hydroxyprednisolone, thus reducing systemic steroid effects by approximately 90%.4 Hence, budesonide became a very attractive steroid for the treatment of inflammatory bowel diseases and other intestinal inflammatory disorders such as eosinophilic esophagitis, combining effectiveness with a low degree of systemic side effects. After some smaller pilot studies of using budesonide to treat AIH provided encouraging results, a large randomized trial was organized by Michael Manns together with the drug company Falk—indeed the largest randomized trial in AIH.5 This trial showed that budesonide indeed can be effective in AIH. In fact, the combined endpoint of achieving remission without steroid side effects was reached more often in the budesonide group than in the fixed-dose prednisone group. Like any good trial, there were many aspects that could be discussed controversially, such as the drug dosages, the dosing schedule, and the combined endpoint, which all may have favored budesonide over prednisone. Furthermore, the low overall response rates in both treatment arms were disappointing and possibly related to the trial design. Nonetheless, there can be no argument that this trial showed budesonide to be an effective treatment option for remission induction of AIH. This may be surprising as there is no evidence that AIH is a solely presinusoidal disease. Therefore, a steroid with such a high first-pass effect in hepatocytes may not be ideal to suppress an inflammatory process that involves all anatomical parts of the liver lobule, especially in the acute phase, and that furthermore has marked systemic components, such as hypergammaglobulinemia. And indeed, we do not really know if it is primarily the presinusoidal effect of budesonide that leads to treatment response in AIH, altered metabolism of the drug with reduced first-pass effect in the context of ongoing liver inflammation, or the systemic effects of the drug (Figure 1). Based on the randomized 2010 trial, budesonide was included in the guidelines on the treatment of AIH, with most guidelines, however, being hesitant in prioritizing one steroid over the other.FIGURE 1: Potential effector mechanisms of budesonide and its alterations in hepatic inflammation. Visualization of budesonide metabolism and effector function in the liver. (A) In liver homeostasis budesonide is rapidly metabolized by CYP3A4 into 6β-hydroxybudesonide and 16α-hydroxyprednisolone, which both have only a minimal corticosteroid effect, resulting in a high first-pass effect of budesonide in the noninflamed liver. (B) Upon hepatocyte-directed immune response, local release of Th1-cytokines (especially interferon gamma and TNFα) results in tissue damage and in further release of damage-induced cytokines (e.g. IL-1 and IL-6 as well as TNFα). Inflammation-induced cytokines in turn markedly downregulate CYP3A4, thus decreasing the first-pass effect of budesonide. This decreased first-pass effect of budesonide may help to exert a systemic steroid effect on arterial and lymphogenic inflammatory cell infiltration, but also leads to systemic steroid side effects. Abbreviations: A, arteria interlobularis; APC, antigen-presenting cell; CYP3A4, Cytochrome P450 3A4; CV, central vein; LV, lymphatic vessel; PV, portal bloodflow via vena interlobularis.Very little experience beyond that initial trial has been reported in the medical literature. In this issue of Hepatology the Spanish collaborative AIH consortium reports real-life experience in a cohort of 105 patients treated with budesonide as initial therapy compared with 276 patients primarily treated with prednisone with at first sight surprising results: the remission rates achieved by budesonide were markedly lower than the remission rates achieved with prednisone therapy. Only in a subpopulation of patients with low baseline transaminases (<2× ULN), response rates of budesonide were not inferior to prednisone-treated patients.6 This discrepancy in effectiveness of budesonide between the real-life experience and the initial budesonide study is in line with other reports investigating the use of budesonide in real life,7,8 and thus seriously challenges the assumed superiority or even just equivalence of budesonide to prednis(ol)one treatment in AIH. If budesonide is maybe less effective than prednis(ol)one in AIH, is it at least better tolerated? This is also called into question by the present study: the overall rate of side effects was at first sight marginally higher in the prednisone group, but this difference disappeared after patients with cirrhosis were excluded from the analysis.6 As budesonide is contraindicated in cirrhosis, and as cirrhosis is associated with the most important steroid-associated risks osteoporosis and diabetes, the benefit of budesonide concerning steroid side effects becomes very questionable. In fact, when looking at the original budesonide randomized trial, the marked difference in steroid side effects was entirely due to “moon face” changes and acne, two rather transient side effects of steroid therapy of little long-term significance. Indeed, diabetes as a possible side effect in that trial was only recorded in four patients in the budesonide group, but none in the prednisone group.5 The risk of more serious side effects in budesonide-treated AIH was also recently pointed out by a Dutch study reporting cataract formation and fractures upon use of budesonide.9 These observations raise the question: Does budesonide act more like systemic steroids, but with less flexibility in the drug dosages, in particular in active AIH? The explanation probably lies in the characteristic pharmacokinetics of budesonide and its alterations in hepatic inflammation: budesonide is metabolized by CYP3A4, an enzyme, whose activity is markedly reduced in the context of inflammation (Figure 1). CYP3A4 activity can be reduced by more than 95%, at least in vitro, by pro-inflammatory cytokines.10 Thus, altered budesonide pharmacokinetics are likely to be even more marked in intrahepatic inflammation. In the inflamed liver, there may indeed be very little first-pass effect of budesonide, leading to high systemic steroid levels and thus side effects in patients with active disease. Indeed, there has been a careful case report of an AIH patient treated with budesonide showing very high systemic drug levels in the absence of portosystemic shunting.11 This fits with the personal experience of sometimes marked steroid side effects on patients on 9 mg budesonide/day.12 As inflammation can alter systemic bioavailability of budesonide, this variability might explain the wide variation in response rates reported.5–8 High systemic drug levels of budesonide in active liver inflammation may lead to higher systemic effects and may in turn increase the effectiveness of budesonide in induction therapy. However, this calls into question the rationale for applying a drug with a very high first-pass effect in the absence of hepatic inflammation. In any case, the debate on which steroid to use in AIH is somewhat focusing on the wrong question: whether or not budesonide may be a reasonable alternative for predniso(ol)one treatment, budesonide does not and will not solve our problems in optimizing treatment of AIH. In other immune-mediated diseases, steroids are no longer a mainstay of therapy and only used for the initial therapy in acute presentations until more effective and more specific immunosuppressive and immunomodulatory drugs take effect. In hepatology, we have fallen behind on testing and adapting such new approaches. This might partly be due to the relative effectiveness of the traditional treatment approach and partly due to the relatively low incidence of AIH. Limitations in our understanding of disease pathogenesis and its multiple immunoregulatory mechanisms make targeted interventions all the more challenging. Undoubtedly, like in many rare diseases, too little industrial interest in investigating novel therapies for AIH hinders progress. In addition, uncertainties on how to best conduct trials in AIH, which inclusion and which exclusion criteria, and which endpoints to use present a further hurdle for progress. The question we need to ask and answer in the future should no longer be which kind of steroids to use in AIH, but how to reduce steroid use altogether and which kind of nonsteroidal therapies to use. Early initiation and optimal dosing of azathioprine or 6-mercaptopurine, including drug monitoring with 6-thioguanine measurements, is already a good option in 2023. In some patients, additional very low-dose steroids, well below the Cushing threshold, may be justified. Patients with insufficient response despite optimal standard therapy should enter systematic clinical studies. As a community, we have to join forces with patients, industry, and regulators to better understand the immunopathogenesis of this enigmatic disease, and to design more specific and probably more individualized therapies for patients with AIH.