Autoimmune hepatitis (AIH) is a chronic liver disorder marked by considerable clinical heterogeneity, presenting with manifestations that range from asymptomatic disease to acute liver failure and advanced cirrhosis. It is further characterized by diverse laboratory abnormalities and histopathological findings, which collectively contribute to significant diagnostic complexity [1,2]. In recent years, the incidence of AIH has exhibited a consistent upward trend. Substantial evidence has identified cirrhosis as an independent prognostic risk factor that significantly influences clinical outcomes in AIH patients [3,4], with approximately one-third of adult patients already showing signs of cirrhosis at the time of diagnosis [5–7]. However, in contrast to this general figure, our prior study [8] and data from several developing nations [9,10] report a significantly higher prevalence, ranging from 40 % to 60 %. Furthermore, our previous study demonstrated a progressively increasing proportion of decompensated cirrhosis, rising from 42.9 % in 2012 to 72.8 % in 2023 (p = 0.016) [11]. Once AIH progresses to decompensated cirrhosis, the prognosis is markedly poor, with a significantly elevated risk of liver-related mortality. Guidelines highlight substantial differences in survival rates between patients with compensated and decompensated cirrhosis [5,6], advocate for stage-specific management strategies [8].
In recent years, a subset of patients with decompensated cirrhosis has shown improvement to a clinical state resembling compensated cirrhosis following etiological treatment [12–16]. Building on this observation, the Baveno VII consensus introduced the concept of “recompensation” [17], which has since been applied to patient cohorts with decompensated cirrhosis due to alcohol-related liver disease (ALD) [18], virologically suppressed hepatitis B virus (HBV) infection [19], cured hepatitis C virus (HCV) infection [20] and primary biliary cholangitis (PBC) treated with ursodeoxycholic acid [21]. Recent meta-analyses [22] have suggested that approximately one-third of patients with decompensated liver cirrhosis can achieve recompensation, thereby reducing the risk of hepatocellular carcinoma (HCC) and mortality. The Shanghai AIH cohort achieved a recompensation rate of 30.9 % under immunosuppressive therapy, demonstrating the potential of this treatment approach to facilitate such outcomes [23]. Our previous study also demonstrated that immunosuppressive therapy could significantly reduce liver-related adverse events in patients with AIH-PBC overlap syndrome and decompensated cirrhosis [24]. However, liver biopsy was not mandatory in the Shanghai AIH cohort [23]. Given that liver histology is paramount for the diagnosis of AIH, yet is frequently not obtained in patients with decompensated cirrhosis due to legitimate clinical concerns, the previous cohort may lack diagnostic precision. Consequently, it is imperative to investigate a well-defined, biopsy-proven cohort of patients with AIH at the decompensated stage.
Previously, our group and others have found that cytokines, including the key inflammatory mediators tumor necrosis factor-alpha (TNF-α) and interleukin-6 (IL-6) are elevated in AlH patients [25,26] and immune cells such as monocytes [27], dendritic cells [28], and regulatory T cells [29] play critical roles in the pathogenesis of AIH. Emerging evidence suggests that biological agents, including infliximab, exhibit favorable efficacy and safety profiles in AIH patients who respond inadequately to conventional therapy [30]. However, it remains unclear whether inflammatory factor levels are associated with disease progression or recompensation in patients with decompensated AIH.
Consequently, this study aims to enroll patients with biopsy-proven AIH and decompensated cirrhosis to systematically analyze their clinical characteristics and recompensation outcomes. Furthermore, we aim to identify cytokines specifically associated with AIH recompensation, thereby providing evidence-based support for therapeutic decision-making and the optimization of intervention strategies.
2Materials and methods2.1Study populationIn this single-center, retrospective cohort study, we screened patients with AIH and decompensated cirrhosis who received treatment at West China Hospital of Sichuan University between January 2011 and August 2024.
The inclusion criteria were as follows: (1) diagnosis of AIH grounded in histological examination of liver biopsy samples, was primarily made according to the simplified AIH scoring system (IAIHG, 2008) [19], as well as the comprehensive diagnostic scoring system revised in 1999 [31]; (2) presence of cirrhosis with at least one decompensation event (e.g., ascites, variceal bleeding, and/or hepatic encephalopathy); (3) receipt of immunosuppressive regimens for at least one year following AIH diagnosis; and (4) no prior use of immunosuppressive therapy at external institutions before admission.
The exclusion criteria were as follows: (1) concurrent chronic liver diseases, such as PBC, ALD, or viral hepatitis; (2) liver biopsy performed more than three months after initiation of immunosuppressive therapy; (3) compensated cirrhosis or classified as Child-Pugh grade A; (4) presence of liver failure at baseline, as defined by the European Association for the Study of the Liver (EASL) guidelines [32]; and (5) those whose baseline data or post-treatment follow-up data were unavailable.
2.2DefinitionsIn this study, decompensation events were defined as the occurrence of ascites, variceal bleeding, and/or hepatic encephalopathy in patients with cirrhosis. For the assessment of ascites based on imaging reports, the classification was as follows: <3 cm was considered mild ascites, 3–6 cm was considered moderate ascites, and 6 cm or greater was considered large ascites [33]. Given the established pathophysiological relationship in which ascites directly contributes to body weight, significant multicollinearity was anticipated between these two variables. To address this issue and achieve statistical stability, we introduced an adjusted “dry body weight” metric. The conversion from ultrasonographic descriptions to estimated volume was performed as shown in Table S1. This approach enables the “dry body weight” variable to more accurately represent the patient's nutritional and somatic mass, thereby effectively decoupling the two for regression analysis. Patients with esophageal variceal bleeding typically present with symptoms of upper gastrointestinal bleeding, such as hematemesis and melena, which can be confirmed via endoscopy. The stratification of hepatic encephalopathy in our study was primarily based on the SONIC classification system, which categorizes patients according to the presence and severity of cognitive and neurological abnormalities. In our cohort, all identified cases of hepatic encephalopathy were classified as overt hepatic encephalopathy (OHE) [34].
Complete biochemical response (CBR) [35] was defined as normalization of alanine aminotransferase (ALT), aspartate aminotransferase (AST), and immunoglobulin G (IgG) after six months of immunosuppressive therapy. Non-response [35] was defined as a reduction in transaminase levels of <50 % after four weeks of treatment. Recompensation was defined according to the Baveno VII consensus [17] as (1) removal, suppression, or elimination of the primary etiology of cirrhosis, defined as achieving CBR or histological remission [35]; (2) resolution of decompensation for at least one year, with fulfillment of criteria including ascites resolution (discontinuation of diuretics), hepatic encephalopathy resolution (discontinuation of lactulose or rifaximin), and absence of variceal rebleeding; and (3) stable improvement in liver function, assessed using serum albumin, bilirubin, and international normalized ratio (INR) as evaluation parameters. The “recompensation date” is the first date on which a patient met all the predefined criteria for recompensation.
2.3Liver biopsy and ELISAPlasma samples were collected and stored at baseline. Enzyme-linked immunosorbent assay (ELISA) kits (Multi Sciences, Hangzhou, China; EK180, EK182, EK110, and EK117) were used to measure the levels of cytokines—IFN-γ, TNF-α, IL-10, and IL-17A—in available plasma samples from a subset of patients, according to the manufacturer’s protocols (see Fig. 5 for sample size).
2.4Statistical analysisContinuous variables are presented as medians with interquartile ranges, and categorical variables are reported as numbers and percentages. Comparisons of continuous variables were performed using independent samples t-tests or Mann–Whitney U tests, depending on data distribution, while categorical variables were compared between groups using Pearson’s chi-square test or Fisher’s exact test. Continuous variables that met the assumption of normality are expressed as mean ± standard deviation (X̄±s), whereas those with skewed distributions are described as median (first quartile, third quartile) (M (Q1, Q3)). The time origin (time zero) was defined as the date of treatment initiation for treated patients and the date of diagnosis for untreated patients. The primary endpoint was death or liver transplantation. Recompensation status was incorporated into the Cox regression model as a time-dependent covariate. A competing risks Cox regression model, with liver transplantation or death treated as competing events, was used to estimate the hazard ratio (HR) for mortality between treated and untreated patients. Additionally, Kaplan-Meier curves were constructed to estimate cumulative survival probabilities in patients who achieved recompensation versus those who remained decompensated, and log-rank tests were applied for subgroup comparisons. A two-sided p-value < 0.05 was considered statistically significant. All statistical analyses were conducted using SPSS software (version 26.0; IBM Inc., Chicago, IL, USA) and R software (version 4.3.3).
2.5Ethical statementThis study was conducted in accordance with the 1964 Declaration of Helsinki and its later amendments and was approved by the Ethics Committee of West China Hospital of Sichuan University (No.2024–2005).
3Results3.1Baseline characteristics of patients with AIH decompensated cirrhosisDuring the study period, we screened and evaluated 437 patients with AIH treated at West China Hospital of Sichuan University between January 2011 and August 2024. According to the inclusion and exclusion criteria, 273 patients were excluded: 179 who did not progress to cirrhosis or progressed after treatment initiation; 64 with missing or incomplete clinical and follow-up data, including those who died or underwent liver transplantation within the first year following diagnosis of decompensated cirrhosis; and 30 with other concomitant chronic liver diseases. Additionally, 73 patients with compensated cirrhosis were excluded, resulting in a final cohort of 91 patients diagnosed with decompensated cirrhosis and classified as Child-Pugh grade A. Among these, 10 patients did not receive immunosuppressive therapy due to severe illness or advanced age (Fig. 1A). The analysis showed that the majority of patients in the cohort were from western and southwestern regions of China (Fig. 1B).
(A) Patient selection process. (B) Geographic distribution of patients who have received immunosuppressive therapy with decompensated AIH-related cirrhosis in the cohort. Shades of blue (from light to dark) correspond to increasing numbers of patients. Due to the substantially larger number of patients in Sichuan Province (n = 69) relative to other provinces, the bar for Sichuan was truncated for visual clarity.
We compared baseline characteristics between the treated and untreated patient groups (Table S2). The Child–Pugh grade was significantly higher in the untreated group than in the treated group (p = 0.002), indicating worse baseline liver function. Statistically significant differences were also observed in white blood cell (WBC) levels (×10⁹/L, 3.92 vs. 5.59, p = 0.028). Liver biopsy findings revealed significant differences in the higher proportion of G-4 (30.0 % vs. 5.7 %, p = 0.018) and the positivity rates of IgG4-positive plasma cells (50.0 % vs. 22.2 %, p = 0.027) between the two groups. Significant differences were also found in key biochemical markers: the untreated group had lower albumin (ALB) levels (g/L, 29.60 vs. 34.20, p = 0.045) and higher total bilirubin (TBIL) (μmol/L, 72.05 vs. 23.15, p = 0.049). There was no statistically significant difference in all-cause mortality or liver transplantation between the groups (16.0 % in the immunosuppressive therapy group vs. 40.0 % in the untreated group, p = 0.261), despite a numerically higher rate in the untreated group — a finding potentially attributable to the limited sample size.
3.2Incidence and characteristics of patients with hepatic recompensationBecause etiological treatment is one of the criteria for recompensation, subsequent analyses focused on 81 patients with AIH and decompensated cirrhosis who received immunosuppressive therapy. Among these patients, 28 (34.6 %) achieved recompensation, whereas 53 (65.4 %) remained decompensated (Figure S1). The median time from treatment initiation to recompensation was 32.1 months (IQR, (20.4 - 44.2) months).
The recompensated group had higher dry body weight (kg, 61.00 vs. 54.00, p < 0.001), and higher body mass index (BMI) (kg/m², 24.03 vs. 21.89, p < 0.001). This group also exhibited higher levels of ALT (×ULN, 4.95 vs. 1.65, p < 0.001), AST (×ULN, 3.69 vs. 1.94, p = 0.024), and ALB (g/L, 33.80 vs. 27.30, p < 0.001), while the persistently decompensated group had a higher prevalence of diabetes (25.5 % vs. 6.7 %, p = 0.042) (Table 1). Other biochemical markers (such as hemoglobin (HB) and TBIL), immunological profiles (including antinuclear antibody (ANA) or smooth muscle antibody (SMA) positivity), and clinical indicators (such as cirrhosis stage, Child-Pugh score, and MELD score) were not comparable between the two groups. Recompensated patients tended to have higher IgG levels (g/L, 24.25 vs. 23.10, p = 0.134), although this difference did not reach statistical significance (Table 1).
Baseline characteristics of patients with recompensated and remained decompensated AIH cirrhosis.
Abbreviations: AMA, anti-mitochondrial antibody; ANA, anti-nuclear antibody; ALB, albumin; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CI, confidence interval; EAD, extrahepatic autoimmune disease; GGT, glutamyl transpeptidase; GLB, globulin; HB, hemoglobin; IgG, immunoglobulin G; INR, international normalized ratio; LC-1, anti-liver cytosol antibody type-1; LKM, anti-liver-kidney microsomal antibody; MELD, model of end-stage liver disease; SCR, Serum Creatinine; SMA, Anti–smooth muscle Antibodies; SLA, Anti-soluble liver antigen antibody; OHE, overt hepatic encephalopathy; PLT, platelet; TC, total cholesterol; WBC, white blood cells.
* p value <0.05 was considered to indicate statistical significance, **p value < 0.01, ***p value<0.001.
The complications and their distribution in the two patient groups are summarized in Table 1 and Figure S2. The proportion of patients with ascites as the sole complication was significantly higher in the recompensated group (96.4 % vs. 66.0 %, p = 0.04). Notably, there was a significant difference in ascites classification between the groups (p < 0.001). Specifically, the proportion of patients with mild ascites was significantly greater in the recompensated group (37.9 % vs. 3.8 %, p = 0.001), whereas the proportion with large ascites was significantly higher in the persistently decompensated group (52.8 % vs. 20.7 %, p = 0.001). No significant difference was observed in the proportion of moderate ascites between the groups (p > 0.05). Regarding other complications, the proportion of patients with either bleeding alone or ascites combined with bleeding was significantly higher in the persistently decompensated group (26.4 % vs. 3.6 %, p = 0.015). With respect to overt hepatic encephalopathy (OHE), no significant difference was observed in the proportion of affected patients between the two groups (p > 0.05).
SMA positivity was assessed in a subset of patients (n = 31), recompensated patients (n = 5), remained decompensated patients (n = 26) with 1 case (3.84 %) testing positive. The denominator reflects only tested cases rather than the total cohort.
Cirrhosis stage [36]: Stage 3: esophageal variceal bleeding for the first time; Stage 4: the first occurrence of non-hemorrhagic decompensation events (such as ascites, jaundice, hepatic encephalopathy); Stage 5: two or more decompensated complications (such as rebleeding + ascites, infection, renal failure, etc.)
All 81 patients underwent liver biopsy procedures (Table S3). The persistently decompensated group exhibited a higher proportion of mild interface inflammation (26.4 % vs. 7.1 %, p = 0.044) and a lower proportion of moderate interface inflammation (37.7 % vs. 64.3 %, p = 0.035). However, no statistically significant differences were observed in other pathological features between the two groups (p > 0.05).
3.3Survival analysis of patientsThroughout the entire follow-up period, the median follow-up duration was 47.5 months (IQR: 12.0, 171.0) in the recompensated group and 76.0 months (IQR: 23.0, 118.0) in the persistently decompensated group (p = 0.143). In the persistently decompensated group, a total of 4 patients underwent liver transplantation and 10 patients died; among these deaths, 4 were attributed to liver-related causes, while the remaining 6 resulted from non-liver-related events: 1 due to sepsis, 2 due to extrahepatic malignancies, and 3 due to out-of-hospital deaths from unknown causes, such as accidents. In the recompensated group, only one patient died, also from an out-of-hospital event.
Kaplan-Meier survival curves were constructed to assess the composite endpoint of all-cause mortality or liver transplantation (Fig. 2). The log-rank test yielded a p- value of 0.044, indicating a statistically significant difference in survival probability between the recompensated and persistently decompensated groups.
Furthermore, the median follow-up duration since achieving recompensation was 56.0 months (IQR, (36.1–87.3) months). However, during the follow-up period, two patients were observed to redevelop decompensation after achieving recompensation. One developed mild ascites, and the other experienced gastrointestinal bleeding. Detailed information on these cases is provided in Table S9.
Patients who achieved recompensation showed significant improvement in liver function, as reflected by changes in the Child–Pugh classification from initial decompensation to the time of recompensation (Fig. 3). Specifically, among these patients, 23 (82.1 %) improved from class B to class A, and 5 (17.9 %) improved from class C to class A. This shift indicates a clear trend toward enhanced hepatic functional reserve following recompensation.
The changes in Child-Pugh class from baseline to the endpoint. The colors of the columns represent patients with different Child-Pugh grades, with red representing Child-Pugh A, blue representing Child-Pugh B, and yellow representing Child-Pugh C. The length of the column represents the proportion of patients. The thicker the line, the greater the number of patients involved. (A) All patients (n = 81). (B) Patients with recompensation (n = 28). (C) Remained recompensated patients (n = 53). CTP, Child-Turcotte-Pugh.
A total of 81 patients received immunosuppressive therapy. As shown in Table 2, there was no significant difference in the initial glucocorticoid dose between the recompensated (n = 28) and remained decompensated (n = 53) groups (all p > 0.05). The proportion of patients receiving an initial dose of 40 mg was 7.1 % vs. 18.9 %, 32 mg was 17.9 % vs. 15.1 %, 24 mg was 64.3 % vs. 54.7 %, and other doses were 10.7 % vs. 11.3 % in the recompensated and remained decompensated groups, respectively. Regarding combination immunosuppressant use, the proportion of patients receiving azathioprine (Imuran) was 57.9 % vs. 61.5 % (p = 0.312), mycophenolate mofetil (CellCept) was 36.8 % vs. 33.3 % (p = 0.957), and other immunosuppressants was 5.3 % vs. 5.1 % (p = 0.899) between the two groups. No statistically significant differences in treatment regimens were observed between the groups (all p > 0.05). We also examined the impact of treatment response on subsequent outcomes. A significantly higher proportion of patients who achieved recompensation also attained CBR within 6 months of treatment initiation (53.6 % vs. 20.8 %, p = 0.003). In contrast, no significant difference was observed in the distribution of non-responders between the groups (28.6 % vs. 20.8 %, p = 0.430).
Comparison of immunosuppressive treatment regimens and response between recompensated and remained decompensated patients.
Abbreviations: AZA, Azathioprine; CBR, Complete biochemical response; MMF, Mycophenolate mofetil. **p value < 0.01.
Meanwhile, several adverse reactions were reported in patients receiving glucocorticoid therapy. Among these patients, 8 with pre-existing diabetes experienced elevated blood glucose levels, necessitating adjustments to their hypoglycemic regimens. Additionally, 11 patients developed severe osteoporosis or fractures despite receiving calcium supplementation with calcium carbonate. Five patients developed a severe respiratory tract infection during glucocorticoid treatment.
Percentages are calculated solely among patients who received immunosuppressive therapy (Recompensated patients: n = 19; Remained decompensated patients: n = 39), with multiple agents per patient counted individually.
3.5Factors influencing recompensationWe established a Cox regression model to assess the influence of various variables on the probability of patients with decompensated AIH liver cirrhosis achieving recompensation. In the univariate analysis, variables including BMI (kg/m², HR = 1.987, 95 % CI: 1.905–2.076, p < 0.001), diabetes (HR = 0.086, 95 % CI: 0.053–0.275, p = 0.026), ALT (×ULN, HR = 1.584, 95 % CI: 1.216–1.882, p < 0.001), ALB (g/L, HR = 1.124, 95 % CI: 1.003–1.259, p < 0.001), and CBR (HR = 1.384, 95 % CI: 1.295–1.563, p = 0.003) were significantly associated with the outcome. Variables with p < 0.05 in the univariate analysis were included in the multivariate analysis to adjust for potential confounding effects and more rigorously evaluate independent predictors.
In the multivariate analysis, BMI (kg/m², HR = 1.161, 95 % CI: 1.022–1.326, p = 0.025), diabetes (HR = 0.582, 95 % CI: 0.294–0.896, p = 0.004), ALT (×ULN, HR = 1.168, 95 % CI: 1.216–1.365, p = 0.028), ALB (g/L, HR = 1.388, 95 % CI: 1.195–1.635, p < 0.001), and CBR (HR = 1.895, 95 % CI: 1.154–2.312, p = 0.014) were identified as independent predictors of recompensation. The detailed results of the regression analyses are presented in Fig. 4 and Table 3.
Integrated forest and ridge plot from the multivariable analysis of recompensation. Color intensity reflects the hazard ratio (HR) value, with blue indicating protective effects (HR<1), and red indicating harmful effects (HR>1). The height of the density ridges represents the probability distribution of the HR estimates.
Analysis of factors influencing recompensation in AIH liver cirrhosis patients in the decompensated stage.
Abbreviations: AMA, anti-mitochondrial antibody; ANA, anti-nuclear antibody; ALB, albumin; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CBR, complete biochemical response; CI, confidence interval; EAD, extrahepatic autoimmune disease; GGT, glutamyl transpeptidase; GLB, globulin; HB, hemoglobin; IgG, immunoglobulin G; INR, international normalized ratio; LC, anti-liver cytosol antibody; LKM, anti-liver–kidney microsomal antibody; Scr, serum creatinine; SMA, anti–smooth muscle antibodies; SLA, anti-soluble liver antigen antibody; PLT, platelet; PT, prothrombin time; TC, total cholesterol; WBC, white blood cells.
* p-value < 0.05 was considered to indicate statistical significance, **p-value < 0.01, ***p-value < 0.001.
The core pathological feature of AIH is immune-mediated hepatocyte inflammation. We measured baseline plasma levels of IFN-γ, IL-10, IL-17A, and TNF-α. The baseline characteristics are presented in Table S4. As shown in Fig. 5A, the expression level of IFN-γ was significantly higher in the persistently decompensated group than in the recompensated group. Similarly, significantly higher levels were observed for IL-17A and TNF-α in the persistently decompensated group compared to the recompensated group (Figs. 5B and 5D, respectively). For IL-10 (Fig. 5C), although a trend toward increased expression was observed in the persistently decompensated group, the difference was not statistically significant. Specific values are provided in the supplementary tables (Tables S5–S8).
4Discussion
This study demonstrates that 34.6 % of patients with biopsy-proven AIH-related decompensated cirrhosis achieved hepatic recompensation under immunosuppressive therapy, as defined by the Baveno VII criteria. Several baseline factors—including higher BMI, elevated albumin and ALT levels, absence of diabetes, and achieving CBR at 6 months—were significantly associated with recompensation. Additionally, patients who recompensated exhibited lower baseline levels of plasma inflammatory cytokines.
The Baveno VII consensus [17] first introduced the concept of “recompensation”, while subsequent studies have explored this concept in cirrhosis of various etiologies [18,19]. Its applicability to autoimmune-mediated liver diseases—which account for over 20 % of liver transplant waiting lists [37]—remains unclear. Based on the Baveno VII criteria, among 81 patients with AIH-related decompensated cirrhosis receiving immunosuppressive therapy, 28 (34.6 %) achieved recompensation—a rate comparable to the 30.9 % reported in a recent AIH cirrhosis cohort but differing from rates observed in other etiologies (e.g., 60.4 % in HBV-related cirrhosis, 18.2 % in ALD-related cirrhosis) [16,18–21]. These differences may be attributed to the potent immunosuppressive effects of AIH treatment, in which early control of inflammatory activity may alter the natural history of fibrotic progression toward cirrhosis.
Kaplan-Meier analysis suggested a survival advantage in the recompensated group, informing clinical risk stratification. As patients without recompensation predictors warrant prioritized transplant assessment, the observed instances of re-decompensation also caution that this state is dynamic and not irreversible. Therefore, a structured post-recompensation management protocol is mandated, featuring lifelong etiology-directed treatment, regular monitoring, and proactive infection management.
Further analysis revealed that elevated baseline levels of BMI, albumin, and ALT serve as predictors of hepatic recompensation. The association with higher BMI may reflect the “obesity paradox”, wherein nutritional reserves can mitigate the poor prognosis associated with sarcopenia [38] and malnutrition [39], which are highly prevalent in advanced cirrhosis. For example, when malnutrition coexists with visceral fat accumulation, long-term mortality increases significantly (p = 0.036) [40]. Elevated ALT levels indicate the presence of ongoing, treatment-responsive inflammation, suggesting a critical therapeutic window for immunosuppressive intervention [41].
Meanwhile, higher albumin levels signify preserved synthetic liver function, which supports systemic stability and facilitates hepatic regeneration [42,43]. These findings indicate that recompensation is most achievable in AIH patients when sufficient metabolic-nutritional competence coexists with a treatable inflammatory drive. Consequently, this underscores the necessity for an integrated management strategy: implementing nutritional support to correct deficits while avoiding excessive metabolic risk, combined with targeted anti-inflammatory therapy to fully exploit the regenerative potential implied by active hepatic inflammation.
Our study identifies diabetes as a significant negative predictor of hepatic recompensation. Extensive research has established associations between diabetes and cirrhosis of various etiologies [44]. It can impair hepatic regeneration by promoting intrahepatic lipid accumulation and has been identified as an independent risk factor for accelerated fibrosis in AIH (OR 2.445, p < 0.05) [45]. Given that corticosteroid-based induction therapy is the standard for AIH, these findings underscore the critical need for active screening and stringent glycemic control in this patient population. Integrating diabetes management into cirrhosis care is thus essential, as it may enhance recompensation rates and improve long-term prognosis.
Our findings solidify the essential role of immunosuppressive therapy in AIH-related decompensated cirrhosis. While immunosuppressive therapy remains the cornerstone of AIH management, specific guidance for its application in decompensated cirrhosis remains limited in international consensus. The 2021 Chinese guidelines provide a cautious, grade B1 recommendation for low-dose corticosteroids in this population (Grade B1) [46]. Multiple retrospective studies have confirmed the efficacy of immunosuppressive therapy in this patient population. For example, recent research further demonstrates the clinical benefits of immunosuppressive treatment in patients with AIH decompensated cirrhosis who show evidence of active disease [23]. In a clinical study by Wang et al. [47]. involving 82 patients with AIH-related decompensated cirrhosis, 40 of 64 patients (62.5 %) receiving corticosteroid treatment achieved reversion to compensated status, demonstrating significant clinical improvement. Previous studies have identified inadequate treatment response as the most significant prognostic factor associated with reduced survival in AIH [48]. Therefore, we innovatively investigated whether CBR at 6 months post-treatment influences recompensation. The results from Cox regression analysis were encouraging, demonstrating that, in addition to baseline characteristics, treatment efficacy can serve as a dynamic predictor of recompensation. This provides a clear clinical directive: initiate immunosuppressive therapy where indicated and diligently assess early treatment response to guide management and improve adherence.
Although the AASLD guideline [7] recommends avoiding AZA in AIH patients with decompensated cirrhosis because of the risk of myelosuppression, in our cohort, AZA was used safely in 61.6 % of patients receiving combination immunosuppressive therapy. This was possible because our center routinely performs NUDT15 polymorphism testing in patients with decompensated cirrhosis, which effectively mitigated the risk of myelosuppression, a finding previously established in our pharmacogenetic research [49].
Our study identified significantly lower baseline levels of pro-inflammatory cytokines (IL-17A, TNF-α, IFN-γ) in patients who achieved recompensation, suggesting that a subdued systemic inflammatory state is essential for establishing a favorable immune microenvironment for hepatic recovery. Mechanistically, this synchronized reduction is pivotal, as TNF-α directly promotes hepatocyte injury and fibrogenesis [50,51]. IFN-γ orchestrates potent Th1 immunostimulatory responses [52] and IL-17 [51,53] acts prominently by amplifying inflammatory cascades and directly activating profibrotic pathways. Collectively, the attenuation of this core inflammatory triad represents a low-grade inflammatory milieu that may constitute a necessary precondition for initiating the liver's intrinsic regenerative program and achieving functional and structural recompensation. This could explain why only a subset of patients with milder immune activation ultimately achieve recompensation following etiological control.
These cytokines represent promising biomarkers for predicting recompensation and could guide early intervention. Furthermore, future therapies targeting these pathways (e.g., infliximab for AIH) may [30], when combined with etiological control, actively promote a pro-recompensation immune microenvironment and improve clinical outcomes.
This study specifically investigated the potential for hepatic recompensation in patients with biopsy-proven AIH-related decompensated cirrhosis; however, several limitations should be acknowledged. First, the data were derived from a single center and exhibited population bias: 76.5 % of the decompensation events were due to ascites, while only 18.5 % and 4.9 % presented with bleeding and hepatic encephalopathy, respectively. This distribution may not fully represent the overall spectrum of AIH decompensated cirrhosis, potentially limiting the generalizability of our findings to broader clinical practice. Additionally, untreated patients exhibited poorer liver function, as reflected by Child-Pugh class, and had a mortality rate as high as 40 %, suggesting that our conclusions may be primarily applicable to early-stage disease. Third, in the absence of repeated liver biopsies, assessment of liver function improvement relied solely on serological markers following immunosuppressive therapy, without histological confirmation of true etiological control, which might have led to an overestimation of the recompensation rate. Finally, 19.8 % and 14.8 % of the cohort had comorbid diabetes and hypertension, respectively, making it impossible to entirely exclude potential metabolic influences on cirrhosis progression. Future studies should focus on multi-center, large-sample prospective investigations to validate these findings, ultimately refining diagnostic and therapeutic strategies for AIH decompensated cirrhosis and improving patient outcomes.
5ConclusionsThis study demonstrates that 34.6 % of patients with AIH-related decompensated cirrhosis achieved recompensation following immunosuppressive therapy. Predictors of recompensation included higher baseline levels of BMI, albumin, and ALT achievement of complete biochemical response at 6 months and absence of diabetes. Furthermore, recompensated patients had lower baseline concentrations of plasma inflammatory cytokines.
Author contributionsYujie Zhang and Ruoting Men contributed equally to this work and served as co-first authors. Yujie Zhang was responsible for the conceptualization of the study, data collection, and analysis. Ruoting Men assisted with the statistical analysis and interpretation of the results. Yi Shen and Yunke Peng provided critical insights and revisions during the writing process. Xiaoyan Ao and Xianglin Wang contributed to the methodological framework and supervised the overall research direction. Li Yang participated in the study design and contributed to the review of the manuscript. Xiaoli Fan was responsible for the overall design of the study, provided critical revisions to the manuscript, and served as the guarantor of the article. All authors have read and approved the final version of the manuscript.
FundingThis study was funded by grants from the 135 projects for disciplines of excellence, West China Hospital, Sichuan University (No. ZYGD23031 to Li Yang) and the Sichuan Science and Technology Program (2026NSFSC0576 to Xiaoli Fan and No 2023NSFSC1618 to Yi Shen).
None.
