INTRODUCTION

Acute liver failure is a heterogeneous and devastating liver disease in patients with previously normal liver function. Acute liver failure is characterized by acute onset liver injury, hepatocellular damage, hepatic encephalopathy, and coagulopathy and carries a high rate of mortality in the absence of liver transplantation. The standard treatment is plasma exchange and/or hemodialysis, but there is no proven effective therapy for ALF, including liver transplantation.^{1, 2}

Even though the use of corticosteroids as a treatment for ALF is controversial,^3-10 recent studies suggest that many adult cases of ALF of unknown etiology are actually immune-mediated; indeed, there are features suggestive of autoimmune pathogenesis.¹¹ Therefore, steroids may improve the survival of adult ALF patients with high serum ALT levels (which are indicative of inflammation) and a low MELD score.^{4, 6, 9} Chapin et al. reported that pediatric patients with indeterminate ALF, defined as ALF of unknown etiology, improved after treatment with corticosteroids.⁵ However, few studies show that corticosteroids improve survival in such cases in the absence of liver transplantation^{5, 12}; to date, steroid therapy for PALF remains controversial.

Increasing evidence shows that indeterminate PALF is often associated with immune activation and dysregulation, characterized by increased levels of serum sIL-2R or dense hepatic CD8 T cell infiltration.^13-17 By contrast, one of the most important etiologies of PALF is a congenital metabolic disease accompanied by minimal hepatic cytotoxic T cell infiltration¹⁷; this makes it more difficult to select patients who may benefit from steroid therapy during the early stages of PALF.

At our institution, corticosteroid treatment for PALF is indicated mainly for patients with suspected autoimmune hepatitis, or if the presentation is strongly suggestive of immune activation in the liver (i.e., elevated serum ferritin or sIL-2R, combined with positive pathological findings and/or CT images).

This retrospective case–control study aimed to clarify whether steroid therapy for PALF accompanied by immune activation improves survival of the native liver and to identify predictors of steroid efficacy. To achieve this goal, we compared medical survival rates or clinical characteristics (based on clinical findings suggestive of inflammation of the liver) in patients treated with/without steroids. The results demonstrate the efficacy of steroids for PALF associated with immune-mediated responses.

METHODS

Thirty-eight consecutive patients under 18 years of age, all diagnosed with ALF at Kyoto University Hospital from February 2006 to August 2022, were enrolled.

The eligibility criteria for ALF were as follows^{18, 19}: (1) no medical history of chronic liver disease, (2) biochemical markers compatible with acute liver injury, (3) coagulopathy caused by factors other than vitamin K deficiency, and (4) PT/INR >2.0 or >1.5, with hepatic encephalopathy. These PALF patients were assessed retrospectively to determine whether they received steroid treatment (or not), and for clinical findings suggestive of immune activation in the liver. Twenty-one of the 38 patients received steroid therapy, while 17 did not. As shown in Figure 1, 19 of the steroid-treated patients met the following criteria detecting inflammation in the liver: (1) serum ferritin or sIL-2R more than twice the upper limit of normal, and (2) liver biopsy showing mononuclear cell infiltration, or imaging result suggestive of liver inflammation (including CT findings of periportal edema and gallbladder wall thickening). These patients were classified into the “Steroid group”. The seven patients who did not receive steroid therapy, but who also met the above-mentioned criteria for identifying inflammation in the liver, were classified into the “Nonsteroid group”. Twelve patients were excluded because they did not meet the criteria defining liver inflammation. Among these 12 patients, seven had CT imaging or pathological findings that did not show signs of liver inflammation, and four were either not evaluated for serum ferritin or sIL-2R or did not meet the inclusion criteria levels. Most patients in the Nonsteroid group did not fulfill the criteria for liver inflammation because they were transferred to a transplant surgeon for the purpose of liver transplantation and underwent transplant procedures without evaluation of immune activation in the liver.

The clinical and pathological findings of the Steroid and Nonsteroid groups were compared (serum AST, ALT, gamma-glutamyl transpeptidase, creatinine, albumin, ferritin, sIL-2R, total bilirubin, ammonia, platelet count, PT-INR, D/T ratio, stages of hepatic encephalopathy, MELD score, and PELD score) (Figure 1). The MELD and PELD scores were assessed before administering steroids or blood purification therapy prior to liver transplantation. The three patients that did not receive blood purification therapy were in the steroid group, and their scores were calculated before steroid administration. Differences in treatment regimens between the two groups were also assessed. The stages of hepatic encephalopathy were assessed based on the previous guidelines for the management of PALF.²⁰

RESULTS

DISCUSSION

High-dose steroid therapy improved survival of PALF patients with hepatic inflammation, even in the absence of liver transplantation, and had no adverse effect on survival after liver transplantation. There was no difference in posttransplant survival rates between the Steroid and Nonsteroid groups, nor was there a difference in the incidence of infection. These results suggest that the treatment strategy used at our institution is both timely and appropriate for patients treated with steroids, as well as those not treated with steroids.

Patients in the Steroid group had significantly higher levels of serum ALT, which correlate with the grade of liver injury and hepatic inflammation; thus, ALT levels are used as surrogates of inflammatory responses during ALF.²⁴ The high levels of transaminases and sIL-2R (the latter may be an indicator of lymphocyte activation), meant that the Steroid group was characterized by a higher degree of inflammation than the Nonsteroid group. Histopathological examination of the liver revealed lobular inflammation in all patients in the Steroid group, but not in the Nonsteroid group, although portal inflammation was observed in both groups. Hepatocytes were almost entirely absent from liver biopsy samples taken from patients in the Nonsteroid group; therefore, it was impossible to determine whether there was any lobular inflammation. These differences could also be attributed to the fact that most of the Nonsteroid group samples comprised recipient livers resected before transplantation.

At the referring hospitals, most patients in the Nonsteroid group who met the criteria for detecting liver inflammation did not receive steroid treatment or died until transfer for liver transplantation. The absence of established guidelines for steroid therapy in PALF, along with concerns regarding infantile cases and posttransplant complications, might have contributed to the hesitance to administer steroids. This is the reason why patients in the Nonsteroid group had more severe diseases than those in the group that prioritized medical intervention with steroids. Indeed, markers of hepatic synthetic function (e.g., bilirubin levels and the D/T ratio) were lower in the Nonsteroid group, but the MELD and PELD scores (markers of disease severity and mortality, respectively) were higher. Our study also reveals that among the steroid-treated group, patients who did not respond to steroids had impaired liver function and extensive liver necrosis. This suggests that earlier administration of steroids might have prevented disease progression and saved patients with a native liver.

Administration of corticosteroid to patients with ALF remains controversial.^3-10 Early studies suggested no survival benefit after steroid administration.^{3, 8} However, more recent research indicates that patients with ALF related to an inflammatory process may respond to corticosteroid administration.⁶ Karkhanis et al. undertook a retrospective study of 361 adult patients with ALF to evaluate the efficacy of steroids for improving overall survival and spontaneous recovery.⁹ Their findings indicated that steroids did not improve overall survival or spontaneous recovery in many patients. Nonetheless, patients who received steroid therapy and showed ALT in the higher 50% of values, did show improved spontaneous recovery compared with those not receiving steroid therapy. Meanwhile, treatment with corticosteroids was significantly associated with lower overall survival in patients with a high MELD score.^{1, 3} These results suggest that ALF patients with higher ALT levels, a surrogate marker for liver inflammation, and lower MELD scores may be more likely to respond favorably to steroid treatment. A Japanese prospective study also reported the efficacy of high-dose corticosteroid for early-stage adult ALF; the drugs suppressed liver inflammation and trended to increase survival, although the effect did not reach statistical significance.¹⁵

In terms of PALF, the efficacy of steroids remains unclear. Few studies have reported the effectiveness of steroid therapy for PALF.^{5, 12} A retrospective study of patients with indeterminate PALF (median of 4 years of age [range, 1–16 years]) reported that 12 of 28 patients who received intravenous mPSL recovered with a native liver.⁵ Cases were defined as having indeterminate ALF if, despite a thorough diagnostic work-up, no identifiable cause for their disease could be determined. However, this study did not compare indeterminate PALF patients who received intravenous mPSL with Nonsteroid patients who did not. As a result, it is unclear whether steroid therapy contributes to the improvement of recovery of patients with a native liver. Here, we demonstrate that steroid treatment improves survival of PALF patients with high serum ferritin and sIL2-R levels, coupled with histopathological or imaging findings suggestive of immune activation.

The etiology of ALF in children is different from that in adults. In Japan, a greater proportion of cases are attributed to metabolic disease (10%–11.1% vs. 0%, retrospectively) or remain indeterminate (47%–49% vs. 29%–34%) and fewer cases are caused by viral infection (6%–7% vs. 46%–50%).^{1, 25-30} Congenital metabolic disorders, which are also important causes of PALF, are not typically associated with hepatic inflammatory cell infiltration¹⁷; hence, the use of corticosteroids in PALF should be approached with careful consideration. By contrast, indeterminate ALF characterized by CD103⁺ CD8⁺ T cell-predominant hepatic infiltrate³¹ stands as the most common cause of PALF; therefore, steroids or other immunosuppressive therapies might effectively mitigate CD103⁺ CD8⁺ T cell-mediated inflammation.³¹ Another study focusing on biomarkers related to PALF shows that four circulating lymphocyte activation markers (i.e., perforin- or granzyme-positive CD8 cells, total CD8 cells, and sIL-2R), are associated with immune activation and can predict the outcomes of early targeted immunosuppression.³² However, the activation of peripheral blood lymphocytes does not always reflect inflammation in the liver. Further investigations are needed to identify more accurate biomarkers that can be used for early detection of immune activation in PALF.

In some situations, it is more prudent to use biomarkers linked to liver inflammation to identify ALF candidates suitable for early immunosuppressive therapy (including steroid administration); this is because it is not safe to wait until all full work-up results are available. In our institution, we used steroids for PALF cases showing elevation of serum sIL-2R or serum ferritin levels, both of which are indicators of lymphocyte or macrophage activation; we also base our assessment on imaging findings suggestive of hepatitis, or inflammatory cell infiltration (mainly lymphocytes), on liver biopsy. As a matter of fact, the “Steroid group” in the present study, which was classified according to this protocol, later turned out to have immune activation-based disease caused by viral infection, HLH, AIH, or drug-induced hypersensitivity; all these conditions respond to steroids. By contrast, the “Nonsteroid group” had metabolic diseases (e.g., Wilson disease and mitochondrial hepatopathy), which do not respond to steroid treatment. Therefore, the reasonable approach is to detect early inflammation-mediated disease using our protocol; in such cases, immunosuppressive therapies, including steroid administration, are effective.

We also focused on identifying factors that predict response to steroid treatment.

Patients in the nonresponder group were younger than those in the responder group. In particular, the nonresponder group contained a higher percentage of infants. Also, patients in the nonresponder group were more severely ill, with higher MELD or PELD scores and more severe coagulopathy, than those in the responder group. Moreover, a significantly higher proportion of patients in the nonresponder group had massive necrosis on liver biopsy. These results are consistent with those of a previous study showing that corticosteroid administration to adult ALF patients led to significantly lower overall survival in those with high MELD score.⁹

Furthermore, higher levels of serum ferritin were found in the nonresponder group, even though there was no difference in transaminase or sIL-2R levels between responders and nonresponders. Serum ferritin level is a widely available surrogate marker of macrophage activation, as well as of iron overload.³³ Hyperferritinemia in PALF is common in those with neonatal hemochromatosis, viral infections, and HLH,^{24, 34} which is consistent with our finding that PALF in some nonresponders was caused by HLH and HSV-1 infection. High levels of serum ferritin prior to liver transplantation in patients with chronic liver failure are associated with increased mortality following liver transplantation.³⁵ Hence, hyperferritinemia in PALF could be a poor prognostic factor.

Taken together, the data suggested that liver transplantation should be a priority for infant patients with ALF coupled with severe coagulopathy or massive necrosis of liver tissue, as they are unlikely to respond to steroid therapy.

The present study has several limitations. First, the study is retrospective, and steroid use was not randomized or standardized. Second, the sample size is small. Finally, there were few available pretreatment liver biopsies, especially from the Nonsteroid group. Histopathologic analysis of the liver provides more accurate information about hepatitis, but biopsy increases the risk of bleeding, especially in patients with coagulation abnormalities. Therefore, in PALF patients with severe coagulopathy, liver resected at the time of transplantation was used for assessment, if available; failing that, we had to rely on the combination of CT images and serum levels of inflammatory markers.

CONCLUSION

In cases of PALF, high-dose steroid therapy in patients with findings suggestive of inflammation-mediated disease can be an effective treatment and increase the recovery rate with the native liver. Liver inflammation detected by serum ferritin and sIL-2R levels, liver biopsy, or imaging are recommended to select candidates for early corticosteroid treatment.

In severe infant PALF with severe coagulopathy, steroids are not recommended, and early liver transplantation should be prioritized.

CONFLICT OF INTEREST STATEMENT

Authors declare no conflict of interests for this article.

ETHICS STATEMENTS

Approval of the research protocol by an institutional reviewer board: This retrospective study was approved by the Institutional Review Board of the Kyoto University Hospital and by the Ethics Committee of Kyoto University Hospital (registered under protocol numbers R2481 and G1233).

Informed consent: Informed consent was obtained in the form of opt-out on the website.

Registry and the registration no. of the study/trial: N/A.

Animal studies: N/A.