Abstract
Background & Aims
In this study, we aimed to evaluate the incidence, predictors, and prognostic significance of recompensation in autoimmune hepatitis (AIH)-related decompensated cirrhosis following immunosuppressive therapy (IST).
Methods
We retrospectively analyzed patients with AIH at first decompensation. Recompensation, defined using modified Baveno VII criteria, required clinical resolution (≥12 months without ascites, variceal bleeding, or hepatic encephalopathy, with liver function restored to Child-Pugh A) along with aetiological suppression (complete biochemical response under IST). Predictors of recompensation were identified using multivariate regression, and survival outcomes were compared among compensated, recompensated, and non-recompensated groups.
Results
A total of 258 patients with AIH-related decompensated cirrhosis were included (median follow-up: 47 months, IQR 28-75). Clinical resolution was achieved by 124 patients (48.1%), while 68 patients (30.9% of 220 treated with IST) met criteria for recompensation. Predictors of recompensation included ascites as the only complication (hazard ratio [HR] 14.40, 95% CI 4.17-49.64, p <0.001), lower IgG levels (HR 0.90, 95% CI 0.89-0.96, p <0.001), higher bilirubin levels (HR 1.04, 95% CI 1.00-1.08, p = 0.030), and higher platelet counts (HR 1.01, 95% CI 1.00-1.01, p = 0.039). Patients achieving recompensation experienced a significantly reduced risk of liver transplantation or death (HR 0.07, 95% CI 0.01-0.50, p = 0.009), with survival outcomes comparable to those of compensated patients.
Conclusions
Recompensation was achieved in approximately one-third of patients with AIH-related decompensated cirrhosis undergoing IST, leading to markedly improved transplant-free survival. Predictors of recompensation included having ascites as the sole complication, lower IgG levels, higher bilirubin levels, and higher platelet counts.
Impact and implications
The predictors and long-term prognostic implications of recompensation in patients with autoimmune hepatitis (AIH)-related decompensated cirrhosis remain unclear. This study demonstrates that recompensation is achievable in patients with AIH-related decompensated cirrhosis and is associated with significant long-term benefits, including improved survival and reduced transplantation needs. We identified ascites (as the sole decompensating event), lower IgG levels, higher bilirubin levels and higher platelet counts as independent predictors of recompensation. These findings can be used by clinicians to identify the patients most likely to benefit from immunosuppressive therapy.
Keywords: Autoimmune Hepatitis, Cirrhosis, Decompensation, Recompensation
Graphical abstract
Highlights
-
•
A substantial proportion of patients with AIH-related decompensated cirrhosis achieve recompensation on immunosuppressive therapy.
-
•
Ascites as the sole decompensating event, lower IgG levels, higher bilirubin levels, and higher platelet counts predicted recompensation.
-
•
Achieving recompensation significantly improves long-term outcomes and reduces transplantation requirements.
Introduction
Autoimmune hepatitis (AIH) is a chronic immune-mediated inflammatory liver disease characterized by elevated serum levels of aminotransferases and immunoglobulin G (IgG), the presence of associated autoantibodies, and specific histological features.1 The incidence of AIH is rising globally, with an annual incidence of 1.31 per 100,000 and a pooled prevalence of 12.99 per 100,000 in Asia.2 Approximately one-third of patients with AIH present with histological cirrhosis.3 When AIH progresses to the decompensated stage, it leads to frequent hospital admissions, impaired quality of life, and significant morbidity and mortality.4
Several studies have demonstrated that patients with decompensated cirrhosis can achieve prolonged stability after aetiological treatment, such as abstinence from alcohol, cure of hepatitis C, and suppression of hepatitis B, resulting in the disappearance of complications and markedly improved long-term prognosis.[5], [6], [7] Based on these findings, a definition of recompensation was proposed in the Baveno VII consensus.8 While recent evidence showed the potential for recompensation in AIH-related decompensated cirrhosis,9 predictors and long-term prognostic implications remain unclear.
Immunosuppressive therapy (IST), comprising prednisolone alone or in combination with other immunosuppressants, is the standard treatment for AIH,3,10,11 and has been associated with regression of liver fibrosis and improved survival.[12], [13], [14] Nonetheless, few studies have explored the effectiveness of IST in AIH-related decompensated cirrhosis. Our previous research demonstrated that 62.5% of patients (40 out of 64) demonstrated reversal of decompensation with IST.15 A related study found that patients with mild or no ascites treated with low-dose steroids had a prognosis similar to those with compensated cirrhosis.16 A recent study further suggested that IST may benefit patients with AIH-related decompensated cirrhosis and active disease, particularly those without overt hepatic encephalopathy (grade 3–4) and with a model for end-stage liver disease (MELD)-Na score <28.9
This study aimed to evaluate the incidence and predictors of recompensation in AIH-related decompensated cirrhosis, following the Baveno VII definition of recompensation, and assess its prognostic significance.
Patients and methods
Study population and data collection
This was a retrospective analysis of a prospectively maintained database registered in the ClinicalTrials.gov database (NCT05785793), which included 1,722 patients diagnosed with AIH recruited from the Department of Gastroenterology and Hepatology at Shanghai Renji Hospital from January 2009 to December 2022. The study was approved by the Renji Hospital ethics committee. Written informed consent was obtained from all patients.
The inclusion criteria were (1) probable or definite diagnosis of AIH according to the International Autoimmune Hepatitis Group: the 1999 revised diagnostic score ≥1017 and/or the simplified diagnostic system ≥6;18 (2) cirrhosis defined by histological, clinical, and/or imaging criteria consistent with cirrhosis.4,19 Decompensated cirrhosis was characterized by the occurrence of complications such as ascites, variceal bleeding, or hepatic encephalopathy.20 Ascites was evaluated based on imaging reports and classified as mild (less than 3 cm), moderate (3-6 cm), or severe (greater than 6 cm).
Patients meeting any of the following criteria were excluded: (1) a concomitant diagnosis of other chronic liver disease (including viral hepatitis, alcohol, and/or nonalcoholic fatty liver disease) and/or variant forms of AIH (overlap with primary biliary cholangitis or primary sclerosing cholangitis); (2) the presence of acute-on-chronic liver failure diagnosed according to the APASL (Asia Pacific Association of Study of Liver Disease) criteria at baseline;21 (3) lack of comprehensive data required for our analysis.
We systematically extracted demographic, clinical, biochemical, immunological, histological, radiological, and endoscopic parameters, along with detailed treatment records, from a centralized electronic database. The baseline timepoint was established as the initial presentation of decompensating events, with all variables subsequently reassessed at the end of the follow-up period.
Outcome
The primary outcome was the cumulative incidence of recompensation, defined by modified Baveno VII criteria: (1) clinical resolution, evidenced by sustained absence of decompensation events (ascites, hepatic encephalopathy, and variceal bleeding) for ≥12 months following cessation of diuretics and prophylactic therapies, accompanied by restoration of Child-Pugh A liver function; and (2) aetiological suppression, demonstrated by complete biochemical response (CBR) with normalization of aminotransferase and IgG levels under IST.8 Given the propensity of ascites and HE to fluctuate, the 12-month event-free requirement was critical to differentiate definitive recovery from transient improvement, ensuring rigorous and uniform assessment across all decompensated events. Recompensation date was established at the initial documentation of concurrent fulfillment of all criteria, with analyses restricted to immunosuppression-treated patients.
Secondary outcomes included transplant-free survival and the cumulative incidence of further decompensation, the latter defined as the recurrence of decompensated complications after clinical resolution.
Follow-up duration was stratified by clinical outcomes. In recompensated patients, it was calculated from the initial decompensation to either (1) the date of recompensation (for analyses of recompensation incidence, liver function improvement, and predictive factors) or (2) death, liver transplantation (LT), or last follow-up (for survival analysis). In non-recompensated patients, follow-up spanned from the onset of decompensation to death, transplantation, or last follow-up, whichever occurred first.
Statistical analysis
Categorical variables were expressed as number (%). Continuous variables were described as mean ± SD or median (IQR) based on normality of distribution. Categorical variables were assessed using the chi-square test or Fisher’s exact test, as appropriate. Continuous data were compared between two groups with the Mann-Whitney test (non-normally distributed data) or two-sided Student’s t test (normally distributed data). The Kruskal-Wallis test was used for more than two groups. The cumulative incidences of recompensation, further decompensation and LT-free survival were estimated by Kaplan-Meier curves using a log-rank test. A Fine-Gray regression model considering LT, TIPS (transjugular intrahepatic portosystemic shunt) and death as competing events was used to investigate predictors of recompensation. Univariate and multivariate time-dependent Cox regression analyses were performed to evaluate the impact of recompensation on mortality, with recompensation modeled as a time-varying covariate. Variables without collinearity were included in the multivariate regression model using stepwise forward selection.
Statistical analysis was performed using R Version 4.2.2, SPSS 25.0 (SPSS Inc, Chicago, IL) and GraphPad prism 9.5.1 (Graph Pad Software, La Jolla, CA). p values <0.05 were considered statistically significant.
Results
Patient characteristics
As shown in Fig. 1, a total of 1,722 patients diagnosed with AIH from January 2009 to December 2022 were assessed; 1,140 patients were excluded due to the absence of cirrhosis, concurrent chronic liver diseases, presentation with acute-on-chronic liver failure at diagnosis, or insufficient clinical data. At the time of AIH diagnosis, 379 (65.1%) patients were at the compensated stage and 203 (34.9%) were at the decompensated stage. During follow-up, 55 initially compensated patients progressed to decompensation, resulting in a total of 258 decompensated cases included in the study. Of these, 220 patients received IST, whereas 38 did not.
Fig. 1.
Study flowchart.
AIH, autoimmune hepatitis; IST, immunosuppressive therapy; DC, decompensated cirrhosis.
This cohort, as presented in Table 1, consisted predominantly of females (80.6%) with a median age of 60 years (IQR 51-67 years). At the earliest decompensation event, the median MELD score was 11.7 (IQR 9.4-15.5), and the median Child-Pugh score was 8 (IQR 7-10). Specifically, 53 patients (20.5%) were classified with Child-Pugh A, 138 (53.5%) with Child-Pugh B, and 67 (26.0%) with Child-Pugh C cirrhosis. Ascites was the most frequent initial decompensating event, occurring in 80.6% of cases, followed by variceal bleeding (19.4%) and hepatic encephalopathy (4.3%). Details of decompensating events stratified by Child-Pugh grade are shown in Table S1. At baseline, the serum IgG level was 22 g/L (IQR 18-25 g/L). Median alanine aminotransferase (ALT), aspartate aminotransferase and total bilirubin levels were 45 IU/L (IQR 26-87 IU/L), 57 IU/L (IQR 37-126 IU/L) and 1.82 mg/dl (IQR 1.17-2.89 mg/dl), respectively.
Table 1.
Baseline characteristics of decompensated patients, grouped by use of immunosuppressive therapy.
| Characteristic | All N = 258 |
Treated n = 220 | Untreated n = 38 | p value |
|---|---|---|---|---|
| Gender (female) | 208 (80.6%) | 177 (80.5%) | 31 (81.6%) | 0.871 |
| Age (years) | 60 (51, 67) | 59 (50, 66) | 66 (59, 73) | <0.001 |
| Antibodies | ||||
| ANA (positive) | 237 (93.3%) | 207 (94.1%) | 30 (88.2%) | 0.258 |
| ASMA (positive) | 22 (8.5%) | 16 (7.3%) | 6 (15.8%) | 0.109 |
| Anti-SLA/LP (positive) | 12 (4.7%) | 11 (5.0%) | 1 (2.6%) | >0.999 |
| IgG (g/L) | 22 (18, 25) | 22 (18, 26) | 22 (19, 24) | 0.720 |
| ALT (IU/L) | 45 (26, 87) | 48 (28, 105) | 32 (22, 48) | 0.007 |
| AST (IU/L) | 57 (37, 126) | 61 (38, 131) | 46 (34, 77) | 0.065 |
| Total bilirubin (mg/dl) | 1.82 (1.17, 2.89) | 1.80 (1.16, 3.01) | 1.95 (1.32, 2.83) | 0.867 |
| Albumin (g/L) | 32.3 (28.5, 36.1) | 32.7 (28.5, 36.1) | 32.1 (29.0, 35.6) | 0.897 |
| Hemoglobin (g/L) | 112 (93, 125) | 114 (97, 125) | 97 (79, 113) | <0.001 |
| Platelet (×109/L) | 74 (54, 114) | 72 (55, 111) | 85 (48, 122) | 0.990 |
| Child-Pugh score | 8.00 (7.00, 10.00) | 8.00 (7.00, 10.00) | 8.00 (7.00, 9.75) | 0.532 |
| Child-Pugh grade | 0.721 | |||
| Class A | 53 (20.5%) | 47 (21.4%) | 6 (15.8%) | |
| Class B | 138 (53.5%) | 116 (52.7%) | 22 (57.9%) | |
| Class C | 67 (26.0%) | 57 (25.9%) | 10 (26.3%) | |
| MELD score | 11.7 (9.4, 15.5) | 11.6 (9.3, 15.3) | 12.2 (10.0, 15.7) | 0.661 |
| Ascites | 208 (80.6%) | 181 (82.3%) | 27 (71.1%) | 0.106 |
| Ascites grade | 0.033 | |||
| Mild | 89 (34.5%) | 81 (36.8%) | 8 (21.1%) | |
| Moderate | 74 (28.7%) | 66 (30.0%) | 8 (21.1%) | |
| Severe | 45 (17.4%) | 34 (15.5%) | 11 (28.9%) | |
| Variceal bleeding | 50 (19.4%) | 38 (17.3%) | 12 (31.6%) | 0.039 |
| Hepatic encephalopathy | 11 (4.3%) | 8 (3.6%) | 3 (7.9%) | 0.210 |
| Follow-up duration (months) | 47 (28, 75) | 53 (31, 79) | 27 (18, 43) | <0.001 |
ALT, alanine aminotransferase; ANA, antinuclear antibody; ASMA, anti-smooth muscle antibody; AST, aspartate aminotransferase; HE, hepatic encephalopathy; IgG, Immunoglobulin G; INR, international normalized ratio; MELD, model for end-stage liver disease; SLA/LP, soluble liver antigen/liver-pancreas antibodies.
Continuous variables are presented as median (IQR) and categorical values as proportions (%).
Qualitative and quantitative differences were analyzed by the chi-squared or Fisher’s exact test for categorical variables, and the Student’s t test or Mann-Whitney U test for continuous variables, as appropriate.
Statistically significant values are shown in bold.
Immunosuppressive therapy
In our analysis, 220 (85.3%) patients with AIH-related decompensated cirrhosis received IST. Compared to untreated patients, those treated with IST were younger (median age 59 vs. 66 years, p <0.001), exhibited higher levels of ALT (48 vs. 32 IU/L, p = 0.007) and hemoglobin (114 vs. 97 g/L, p <0.001). Additionally, the IST-treated group had a lower incidence of variceal bleeding as the first decompensating event (17.3% vs. 31.6%, p = 0.039). Notably, while no statistically significant difference was observed in hepatic encephalopathy rates between groups, it is noteworthy that only a small proportion of patients with prior hepatic encephalopathy received IST (8 cases, 3.6%) (Table 1).
All 220 patients initiated their treatment with prednisolone as induction therapy, commencing at a median initial dose of 20 mg (IQR 15-20 mg), equivalent to 0.33 mg/kg/day (IQR 0.28-0.43 mg/kg/day). The dose was subsequently tapered to a maintenance dose of 5 mg. One hundred and eight patients (49.1%) received immunosuppressants as maintenance therapy. Specifically, 55 patients (25.0%) initiated IST a median of 15 months (IQR 4-34 months) prior to decompensation, whereas the remaining 165 patients (75.0%) started IST a median of 2 months (IQR 0-5 months) after their first decompensation (Table S2). This interval was necessitated by the imperative to manage and stabilize complications, such as infections and hemorrhage, and to confirm the diagnosis of AIH before starting IST. Patients who began IST prior to decompensation exhibited lower IgG levels (19 vs. 22 g/L, p <0.001) and lower aminotransferase levels (ALT: 42 vs. 54 IU/L, p = 0.042; aspartate aminotransferase: 45 vs. 72 IU/L, p <0.001) at the time of their first decompensation. They also had lower disease severity scores (Child-Pugh score: 7 vs. 8, p <0.001; MELD score: 10.3 vs. 11.2, p = 0.032) (Table S3).
Among patients treated with IST, the most frequently reported severe side effect was infection, affecting 21 individuals (9.5%), most often presenting as lung infections. Most patients recovered after temporarily stopping IST and receiving antimicrobial or antifungal treatment, but nine patients died as a direct or indirect result of these infections. Additionally, prednisolone was discontinued in six cases due to severe osteoporosis, in six cases due to necrosis of the femoral head, and in one case due to a fracture.
Incidence of recompensation
During a median follow-up of 47 months (IQR 28-75 months), 68 of 220 patients (30.9%) undergoing IST achieved recompensation according to Baveno VII criteria (Fig. 2). The cumulative incidences of recompensation at 2 years, 3 years, and 5 years were 24.0%, 31.0%, and 36.6%, respectively.
Fig. 2.
The cumulative incidence of recompensation.
The cumulative incidence of recompensation was plotted using the Kaplan-Meier method.
In our cohort, clinical resolution occurred in 124 patients (48.1%) during follow-up, with cumulative incidence rates of 40.0%, 49.0%, and 59.2% at 2, 3, and 5 years, respectively (Fig. S1A). Among the 220 IST-treated patients, 121 (55.0%) achieved clinical resolution, and 87 patients (39.5%) attained CBR. Notably, patients who attained CBR demonstrated a significantly higher rate of clinical resolution (78.2%, 68/87) compared to those without CBR (39.8%, 53/133), corresponding to a 2.84-fold increased likelihood of clinical resolution (95% CI 1.97-4.09; p <0.001) (Fig. S1B).
Patients who achieved recompensation demonstrated significant improvement in liver function as evidenced by changes in Child-Pugh classification between initial decompensation and recompensation. Specifically, 35 patients (51.5%) improved from class B to A, and 18 patients (26.5%) improved from class C to A (Fig. 3).
Fig. 3.
The changes in Child-Pugh class from baseline to the endpoint.
The colors of the columns represent patients with different Child-Pugh grade, with yellow representing Child-Pugh A, blue representing Child-Pugh B, and red 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 = 220). (B) Patients with recompensation (n = 68). (C) Patients without recompensation (n = 152). CTP, Child-Pugh.
Predictors of recompensation
Compared to non-recompensated patients, those achieving recompensation exhibited distinct baseline characteristics, including lower IgG levels (19 vs. 22 g/L, p = 0.002), higher ALT levels (66 vs. 44 IU/L, p = 0.005), and higher platelet counts (95 vs. 68×109/L, p = 0.029), despite comparable demographics and disease severity (Child-Pugh and MELD scores) (Table 2).
Table 2.
Characteristics of patients with recompensated and non-recompensated cirrhosis.
| Characteristics | Recompensated n = 68 (30.9%) | Non-recompensated n = 152 (69.1%) | p value |
|---|---|---|---|
| Gender (female) | 55 (80.9%) | 122 (80.3%) | 0.915 |
| Age (years) | 60 (49, 66) | 59 (51, 66) | 0.863 |
| IgG (g/L) | 19 (17, 23) | 22 (19, 27) | 0.002 |
| ALT (IU/L) | 66 (36, 213) | 44 (26, 79) | 0.005 |
| AST (IU/L) | 73 (36, 202) | 58 (38, 105) | 0.124 |
| Total bilirubin (mg/dl) | 1.94 (1.30, 3.75) | 1.72 (1.16, 2.79) | 0.273 |
| Albumin (g/L) | 33.4 (29.3, 36.4) | 32.2 (28.3, 35.9) | 0.270 |
| Hemoglobin (g/L) | 116 (102, 130) | 113 (95, 125) | 0.234 |
| Platelet (×109/L) | 95 (63, 115) | 68 (53, 108) | 0.029 |
| Child-Pugh score | 8.00 (7.00, 10.00) | 8.00 (7.00, 10.00) | 0.923 |
| Child-Pugh grade | 0.969 | ||
| Class A | 15 (22.1%) | 32 (21.1%) | |
| Class B | 35 (51.5%) | 81 (53.3%) | |
| Class C | 18 (26.5%) | 39 (25.7%) | |
| MELD score | 12.3 (9.2, 15.8) | 11.4 (9.4, 15.0) | 0.777 |
| First decompensation event | |||
| Ascites | 64 (94.1%) | 117 (77.0%) | 0.002 |
| Mild | 34 (50.0%) | 47 (30.9%) | |
| Moderate | 21 (30.9%) | 45 (29.6%) | |
| Severe | 9 (13.2%) | 25 (16.4%) | |
| Variceal bleeding | 4 (5.9%) | 34 (22.4%) | 0.003 |
| Hepatic encephalopathy | 0 (0.0%) | 8 (5.3%) | 0.061 |
| Decompensation event during follow-up | |||
| Ascites only | 63 (92.6%) | 75 (49.3%) | <0.001 |
| Variceal bleeding | 5 (7.4%) | 55 (36.2%) | <0.001 |
| Hepatic encephalopathy | 0 (0.0%) | 29 (19.1%) | <0.001 |
| Immunosuppressive therapy | |||
| Duration of IST (months) | 63 (40, 88) | 43 (24, 73) | <0.001 |
| Treat after decompensation | 52 (76.5%) | 113 (74.3%) | 0.736 |
| Duration between first decompensation and IST (months) | |||
| Duration of IST after first decompensation | 1 (0, 2) | 2 (0, 7) | 0.007 |
| Duration of IST before first decompensation | 11 (4, 16) | 15 (3, 41) | 0.307 |
| IST regimen | 0.230 | ||
| Prednisolone alone | 35 (51.5%) | 77 (50.7%) | |
| Prednisolone + AZA | 7 (10.3%) | 11 (7.2%) | |
| Prednisolone + MMF | 26 (38.2%) | 56 (36.8%) | |
| Prednisolone + TAC/CsA | 0 (0.0%) | 8 (5.3%) | |
| Initial dose of prednisolone (mg) | 20 (20, 25) | 20 (15, 20) | 0.072 |
| Initial dose of prednisolone (mg/kg/d) | 0.34 (0.29, 0.44) | 0.33 (0.26, 0.41) | 0.072 |
ALT, alanine aminotransferase; ANA, antinuclear antibody; ASMA, anti-smooth muscle antibody; AST, aspartate aminotransferase; AZA, azathioprine; CsA, cyclosporin; IgG, Immunoglobulin G; INR, international normalized ratio; IST, immunosuppressive therapy; MELD, model for end-stage liver disease; MMF, mycophenolate mofetil; SLA/LP, soluble liver antigen/liver-pancreas antibodies; TAC, tacrolimus.
Continuous variables are presented as median (IQR) and categorical values as proportions (%).
Qualitative and quantitative differences were analyzed by the chi-squared or Fisher’s exact test for categorical variables, and the Student’s t test or Mann-Whitney U test for continuous variables, as appropriate.
Statistically significant values are shown in bold.
Clinically, recompensated patients more frequently presented with ascites as their initial decompensating event (94.1% vs. 77.0%, p = 0.002), which was generally milder, and had significantly lower rates of variceal bleeding (5.9% vs. 22.4%, p = 0.003) and hepatic encephalopathy (0.0% vs. 5.3%, p = 0.061). During follow-up, 92.5% of recompensated patients experienced ascites as their sole complication, with a lower incidence of variceal bleeding (7.4% vs. 36.2%, p <0.001). Only five patients with prior variceal bleeding managed to achieve recompensation, and none with hepatic encephalopathy attained recompensation.
Regarding IST, recompensated patients initiated IST earlier after decompensation, compared to non-recompensated patients (median 1 vs. 2 months, p = 0.007). However, initial and maintenance prednisolone doses, as well as the proportion of additional immunosuppressant use, were comparable between the two groups (Table 2). Furthermore, the timing of IST initiation, before vs. after decompensation, did not significantly influence recompensation rates (p = 0.810) (Fig. S2).
In our competing-risk multivariate analysis (considering LT, death, and TIPS as competing events), ascites as the sole decompensating event emerged as the strongest independent predictor of recompensation, and was associated with a markedly higher likelihood of recompensation than variceal bleeding or hepatic encephalopathy (hazard ratio [HR] 14.40, 95% CI 4.17-49.64, p <0.001). Baseline IgG levels (HR 0.92, 95% CI 0.89-0.96, p <0.001), total bilirubin (HR 1.04, 95% CI 1.00-1.08, p = 0.030) and platelet counts (HR 1.01, 95% CI 1.00-1.01, p = 0.039) independently predicted recompensation (Table 3).
Table 3.
Univariate and multivariable competing-risk multivariate analysis (transplantation, TIPS, and death considered as competing events) of recompensation.
| Characteristic | Univariable |
Multivariable |
||||
|---|---|---|---|---|---|---|
| HR | 95% CI | p value | HR | 95% CI | p value | |
| Age (years) | 1.00 | 0.99, 1.02 | 0.690 | |||
| Female | 1.07 | 0.60, 1.93 | 0.810 | |||
| IgG (g/L) | 0.95 | 0.91, 0.98 | 0.007 | 0.92 | 0.89, 0.96 | <0.001 |
| ALT (IU/L) | 1.00 | 1.00, 1.00 | <0.001 | |||
| Total bilirubin (mg/dl) | 1.06 | 1.03, 1.10 | <0.001 | 1.04 | 1.00, 1.08 | 0.030 |
| Albumin (g/L) | 1.03 | 0.98, 1.07 | 0.260 | |||
| Hemoglobin (g/L) | 1.01 | 1.00, 1.02 | 0.230 | |||
| Platelet (×109/L) | 1.00 | 1.00, 1.01 | 0.032 | 1.01 | 1.00, 1.01 | 0.039 |
| Ascites only∗ | 9.82 | 3.93, 24.52 | <0.001 | 14.40 | 4.17, 49.64 | <0.001 |
| Variceal bleeding | 0.18 | 0.07, 0.45 | <0.001 | |||
| Hepatic encephalopathy | 0.00 | 0.00, 0.00 | <0.001 | |||
ALT, alanine aminotransferase; IgG, immunoglobulin G; HR, hazard ratio; TIPS, transjugular intrahepatic portosystemic shunt.
Variable selection for the multivariate regression model was performed using stepwise forward selection.
Ascites only was defined as the absence of variceal bleeding and hepatic encephalopathy.
Impact of recompensation on prognosis
During a median follow-up of 53 months (IQR 31-79 months) in IST-treated patients, patients without recompensation exhibited significantly higher rates of all-cause mortality or LT (20.4%, 31/152) compared to recompensated cases (1.5%, 1/68, p <0.001). The single mortality in the recompensated group resulted from acute severe pancreatitis. In the non-recompensated group, liver-related deaths (n = 24) were attributed to liver failure (n = 14), infections (n = 9), and hepatocellular carcinoma (HCC) (n = 1). Non-liver-related deaths (n = 4) included respiratory failure (n = 2), extrahepatic malignancy (n = 1), and fracture (n = 1). Three patients without recompensation underwent LT due to liver failure. Importantly, recompensation conferred a 20.5-fold survival benefit (95% CI 2.79-150.63, p = 0.003) (Fig. 4). However, CBR status among clinically resolved patients showed no prognostic impact (HR 0.46, 95% CI 0.03-7.64, p = 0.588) (Fig. S3), underscoring clinical resolution as the principal determinant of outcomes.
Fig. 4.
Long-term prognosis of the compensated, recompensated, and non-recompensated groups.
The transplant-free survival was plotted using the Kaplan-Meier method. Log-rank test was used to compare overall differences in transplant-free survival among patients with different cirrhosis stages. Pairwise comparisons were conducted using Cox proportional hazards models with the recompensated group as the reference group.
We also analyzed 379 patients with compensated AIH-related cirrhosis before IST initiation. The baseline date was defined as the earliest date of cirrhosis diagnosis. During a median follow-up period of 69 months (IQR 43-92 months), 14 patients (3.7%) died or underwent LT. Baseline characteristics are detailed in Table S4. There was no significant difference in LT-free survival between the recompensated and compensated groups (HR 2.10, 95% CI 0.27-16.15, p = 0.476) (Fig. 4).
Given the potential mortality risks of IST (e.g. infections, cardiovascular events), we employed stringent endpoints (LT or all-cause mortality) to more accurately isolate the prognostic benefit of recompensation. In a time-dependent Cox regression analysis with recompensation modeled as a time-varying covariate (Table 4), achievement of recompensation emerged as a powerful independent protective factor against the composite endpoint (HR 0.07, 95% CI 0.01-0.50, p = 0.009), after adjusting for age, gender, platelet counts, and Child-Pugh score at baseline. This association persisted in sensitivity analysis excluding patients with <2 years of follow-up to address potential immortal time bias (HR 0.07, 95% CI 0.01-0.57, p = 0.012) (Table S5), further supporting the survival benefit of recompensation. Moreover, HCC occurred exclusively in non-recompensated patients (6/152, 3.95%), suggesting recompensation may reduce HCC risk in decompensated AIH-related cirrhosis.
Table 4.
Multivariable regression models for the assessment of recompensation associated with transplantation or death in AIH-related decompensated cirrhosis.
| Variables | Model 1 |
Model 2 |
Model 3 |
|||
|---|---|---|---|---|---|---|
| HR (95%CI) | p value | HR (95%CI) | p value | HR (95%CI) | p value | |
| Time-dependent covariate | ||||||
| Recompensation | 0.06 (0.01–0.44) | 0.006 | 0.05 (0.01–0.440) | 0.004 | 0.07 (0.01–0.50) | 0.009 |
| Baseline covariates | ||||||
| Age (years) | 1.03 (1.00–1.07) | 0.047 | 1.03 (1.00–1.07) | 0.093 | ||
| Gender (Male) | 0.30 (0.09–1.00) | 0.051 | 0.33 (0.10–1.15) | 0.081 | ||
| Platelet (×109/L)∗ | 1.00 (1.00–1.01) | 0.701 | ||||
| Child-Pugh score | 1.06 (0.86–1.31) | 0.597 | ||||
HR, hazard ratio.
Recompensation was modeled as a time-varying covariate.
Model 1: Crude. Model 2: Adjust: baseline age, gender. Model 3: Adjust: baseline age, gender, platelet, Child-Pugh score.
Platelet count was log-transformed when inserted into the model.
Further decompensation
During follow-up, 5.9% (4 out of 68) of patients who had previously achieved recompensation experienced further decompensation, with all cases occurring after recompensation and all developing mild ascites. Among 121 patients who reached clinical resolution after IST, 15 (12.4%) experienced further decompensation, with ascites being the main issue. Additionally, three patients with a history of variceal bleeding had recurrent episodes during follow-up, and two patients experienced new episodes of hepatic encephalopathy. More details on further decompensation are available in Table S6.
To assess the prognostic utility of biochemical markers, we stratified patients by aminotransferase and IgG levels at clinical resolution (Fig. S4). The analysis demonstrated that failure to achieve CBR was associated with a 4.5-fold increased risk of further decompensation (HR 4.52, 95% CI 1.44-14.26, p = 0.005). Among CBR components, elevated IgG independently predicted decompensation risk (HR 3.32, 95% CI 1.18-9.33, p = 0.016), while aminotransferase elevation showed no significant association (HR 2.01, 95% CI 0.64-6.33, p = 0.224). These findings suggest that persistent IgG elevation, unlike aminotransferase levels, may reflect more advanced cirrhotic changes and serve as a clinically meaningful biomarker for ongoing disease progression risk even after clinical resolution.
Discussion
Our large-scale real-world study demonstrated that 30.9% of immunosuppression-treated patients with AIH-related decompensated cirrhosis achieved recompensation by Baveno VII criteria, and 48.1% attained clinical resolution. These findings validate the therapeutic potential of IST in this population, particularly among patients with favorable predictors including isolated ascites, lower IgG levels, and elevated bilirubin and platelet counts.
While the Baveno VII recompensation criteria were initially validated in viral and alcohol-related cirrhosis, our findings extend their applicability to AIH, albeit with distinct features. Our study indicated that about one-third of patients with AIH-related decompensated cirrhosis achieved recompensation after IST, a rate exceeding those reported for alcohol-related (18.1%)22 and PBC-related cirrhosis (16.7%),23 similar to that for HCV (24.7%),24 but below that for HBV (56.2%),25 likely reflecting differential treatment efficacy between immunosuppression and potent antiviral therapies. However, these cross-etiology comparisons require cautious interpretation due to the limited data available, highlighting the need for further validation studies.
We identified four independent predictors of recompensation: (1) ascites as the sole decompensating event, (2) lower baseline IgG levels, (3) higher bilirubin levels, and (4) higher platelet counts. Isolated ascites was the strongest predictor, whereas portal hypertension severity – evidenced by thrombocytopenia or variceal bleeding – reduced the likelihood of recompensation. Notably, none of the 29 patients with hepatic encephalopathy achieved recompensation, aligning with prior evidence that overt HE (grade 3-4) portends high mortality and LT risk despite IST,9 highlighting the constrained efficacy of IST in this clinical context. Notably, lower IgG correlated with recompensation, aligning with studies linking elevated immunoglobulins to fibrosis progression via FcγR-mediated stellate cell activation.26,27 These findings suggest that elevated IgG may not merely reflect disease severity but actively contribute to progression. Paradoxically, higher baseline bilirubin predicted recompensation, possibly reflecting preserved regenerative capacity in actively inflamed livers when treated. These findings underscore the prognostic utility of biochemical markers in assessing therapeutic response. Furthermore, early IST increased the likelihood of recompensation, likely by halting ongoing injury and facilitating recovery.
Recompensation conferred a 14-fold reduction in LT/mortality risk, with recompensated patients achieving outcomes comparable to compensated cirrhosis. These findings align with prior studies across etiologies,28,29 suggesting that recompensation may reset the prognostic trajectory to that of compensated patients. Of particular note, only one recompensated patient died (due to extrahepatic causes), underscoring the pronounced impact of recompensation on liver-related outcomes. At our center, LT rates (1.97%) were lower than reported,30,31 with all transplants occurring in non-recompensated patients. In China’s resource-limited setting, recompensation may help mitigate transplant demand; however, further studies are warranted to assess its broader impact on transplantation rates. HCC occurred exclusively in non-recompensated patients (3.95%), aligning with prior reports (2.6%),32 but the small event numbers necessitate further validation of any protective effect against hepatocarcinogenesis.
This study offers preliminary insights into the criteria for aetiological suppression in AIH-related recompensation. Although liver biopsy remains the gold standard for assessing disease activity – particularly given the inadequate correlation between conventional biochemical markers (ALT/IgG) and histological activity in this population33 – the practical limitations of repeated histological sampling in decompensated patients motivated our evaluation of IAIHG-recommended biochemical markers for monitoring treatment response. Our findings indicate that while clinical resolution ensures comparable long-term transplant-free survival regardless of CBR, patients achieving CBR exhibited significantly greater clinical stability, with reduced rates of further decompensation. Previous studies have also shown that lower IgG levels correlate with improved clinical outcomes31 and reduced histological activity34 in AIH. Thus, although clinical resolution remains the primary focus for recompensation, CBR may serve as a valuable adjunctive marker of aetiological suppression, thereby reinforcing the Baveno VII criteria. However, current definitions of aetiological suppression require refinement, as standard CBR thresholds may be overly stringent. Ongoing investigations are exploring alternative biomarkers, including modified ALT/IgG cut-offs and non-invasive measures, to better assess histological activity in this patient population.
Our study has several limitations. First, while our cohort represented diverse disease severities, the retrospective design may introduce selection bias, necessitating validation through prospective multicenter studies. Second, although we systematically excluded patients with contraindications to immunosuppression (pre-existing infections, severe osteoporosis, uncontrolled diabetes), comprehensive documentation of treatment-related adverse effects (including infections, fracture, osteoporosis, elevated blood sugar, glaucoma, and sleep disorders, severely impacting quality of life) was limited by the retrospective nature of our study. We only captured severe adverse events leading to IST discontinuation, potentially underestimating the true burden of treatment complications. Third, while liver histology remains essential for AIH diagnosis, routine biopsy in decompensated cirrhosis was precluded by safety concerns, reflecting real-world clinical challenges. Our diagnostic approach strictly adhered to IAIHG criteria, ensuring all cases fulfilled criteria for probable or definite AIH. These limitations highlight the critical need for reliable non-invasive biomarkers to assess histological activity.
In conclusion, our study validated the Baveno VII recompensation criteria in AIH-related decompensated cirrhosis, demonstrating that nearly one-third of patients achieved recompensation with IST, resulting in excellent transplant-free survival rates comparable to those in compensated patients. We identified four key predictors of recompensation: ascites as a sole complication, lower IgG, higher bilirubin, and higher platelet counts, which collectively provide a clinically useful profile for selecting patients most likely to benefit from IST.
Abbreviations
AIH, autoimmune hepatitis; ALT, alanine aminotransferase; CBR, complete biochemical response; HCC, hepatocellular carcinoma; HR, hazard ratio; IgG, immunoglobulin G; IST, immunosuppressive therapy; MELD, model for end-stage liver disease.
Financial support
This work was supported by the National Key R&D Program of China (2023YFC2507502 to XM), National Natural Science Foundation of China grants (#82470612 to XX, #82130017, 81830016 and 81620108002 to XM), Shanghai Top-Priority Medical Center Construction Project (Grant Number: 2023ZZ02022), and Innovative research team of high-level local universities in Shanghai (SHSMU-ZDCX20210301 to XM).
Authors’ contributions
Study concept and design: Xiong Ma, Xiao Xiao, Yu Chen, Haoyu Wen, You Li. Data acquisition: Yu Chen, Haoyu Wen, You Li, Tianyu Mao, Chenyi Jiang, Huayang Zhang, Bo Li, Jun Zhang, Canjie Guo, Li Sheng, Jing Hua, Hai Li, Qixia Wang, Qi Miao, Li Yan, Min Lian, Xiao Xiao, Xiong Ma. Analysis and interpretation of data: Xiao Xiao, Weituo Zhang, Yu Chen, Haoyu Wen, Yujie Zhou, Xiting Pu, Ruqi Tang, Zhengrui You. Manuscript writing, critical revision of the manuscript, and statistical analysis: Xiong Ma, Xiao Xiao, M. Eric Gershwin, Christopher L. Bowlus, Weituo Zhang, Yu Chen, Haoyu Wen, You Li. All the authors vouch for the veracity and completeness of the data and analyses presented. The final version of the manuscript has been reviewed and approved by all authors.
Data availability
More additional data other than what were presented in the manuscript are available to request from the corresponding author. The data are not publicly available due to privacy/ethical restrictions.
Conflict of interest
The authors of this study declare that they do not have any conflict of interest.
Please refer to the accompanying ICMJE disclosure forms for further details.
Acknowledgements
We thank all the investigators and patients who participated in this study.
Footnotes
Author names in bold designate shared co-first authorship
Supplementary data to this article can be found online at https://doi.org/10.1016/j.jhepr.2025.101496.
Contributor Information
Xiao Xiao, Email: xxiao_sh@163.com.
Xiong Ma, Email: maxiongmd@hotmail.com.
Supplementary data
The following are the Supplementary data to this article:
References
- 1.Muratori L., Lohse A.W., Lenzi M. Diagnosis and management of autoimmune hepatitis. Bmj. 2023;380 doi: 10.1136/bmj-2022-070201. [DOI] [PubMed] [Google Scholar]
- 2.Lv T., Li M., Zeng N., et al. Systematic review and meta-analysis on the incidence and prevalence of autoimmune hepatitis in Asian, European, and American population. J Gastroenterol Hepatol. 2019;34:1676–1684. doi: 10.1111/jgh.14746. [DOI] [PubMed] [Google Scholar]
- 3.Wang G., Tanaka A., Zhao H., et al. The Asian Pacific Association for the Study of the Liver clinical practice guidance: the diagnosis and management of patients with autoimmune hepatitis. Hepatol Int. 2021;15:223–257. doi: 10.1007/s12072-021-10170-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Gines P., Krag A., Abraldes J.G., et al. Liver cirrhosis. Lancet. 2021;398:1359–1376. doi: 10.1016/S0140-6736(21)01374-X. [DOI] [PubMed] [Google Scholar]
- 5.Aravinthan A.D., Barbas A.S., Doyle A.C., et al. Characteristics of liver transplant candidates delisted following recompensation and predictors of such delisting in alcohol-related liver disease: a case-control study. Transpl Int. 2017;30:1140–1149. doi: 10.1111/tri.13008. [DOI] [PubMed] [Google Scholar]
- 6.El-Sherif O., Jiang Z.G., Tapper E.B., et al. Baseline factors associated with improvements in decompensated cirrhosis after direct-acting antiviral therapy for hepatitis C virus infection. Gastroenterology. 2018;154:2111–2121.e2118. doi: 10.1053/j.gastro.2018.03.022. [DOI] [PubMed] [Google Scholar]
- 7.Kim T.H., Um S.H., Lee Y.S., et al. Determinants of re-compensation in patients with hepatitis B virus-related decompensated cirrhosis starting antiviral therapy. Aliment Pharmacol Ther. 2022;55:83–96. doi: 10.1111/apt.16658. [DOI] [PubMed] [Google Scholar]
- 8.de Franchis R., Bosch J., Garcia-Tsao G., et al. Baveno VII - renewing consensus in portal hypertension. J Hepatol. 2022;76:959–974. doi: 10.1016/j.jhep.2021.12.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Arvaniti P., Rodríguez-Tajes S., Padilla M., et al. Hepatic encephalopathy and MELD-Na predict treatment benefit in autoimmune hepatitis-related decompensated cirrhosis. Clin Gastroenterol Hepatol. 2025 doi: 10.1016/j.cgh.2025.02.010. Epub ahead of print. [DOI] [PubMed] [Google Scholar]
- 10.Mack C.L., Adams D., Assis D.N., et al. Diagnosis and management of autoimmune hepatitis in adults and children: 2019 practice guidance and guidelines from the American association for the study of liver diseases. Hepatology. 2020;72:671–722. doi: 10.1002/hep.31065. [DOI] [PubMed] [Google Scholar]
- 11.European Association for the Study of the Liver EASL clinical practice guidelines: autoimmune hepatitis. J Hepatol. 2015;63:971–1004. doi: 10.1016/j.jhep.2015.06.030. [DOI] [PubMed] [Google Scholar]
- 12.Schvarcz R., Glaumann H., Weiland O. Survival and histological resolution of fibrosis in patients with autoimmune chronic active hepatitis. J Hepatol. 1993;18:15–23. doi: 10.1016/s0168-8278(05)80005-4. [DOI] [PubMed] [Google Scholar]
- 13.Dufour J.F., DeLellis R., Kaplan M.M. Reversibility of hepatic fibrosis in autoimmune hepatitis. Ann Intern Med. 1997;127:981–985. doi: 10.7326/0003-4819-127-11-199712010-00006. [DOI] [PubMed] [Google Scholar]
- 14.Czaja A.J., Carpenter H.A. Decreased fibrosis during corticosteroid therapy of autoimmune hepatitis. J Hepatol. 2004;40:646–652. doi: 10.1016/j.jhep.2004.01.009. [DOI] [PubMed] [Google Scholar]
- 15.Wang Z., Sheng L., Yang Y., et al. The management of autoimmune hepatitis patients with decompensated cirrhosis: real-world experience and a comprehensive review. Clin Rev Allergy Immunol. 2017;52:424–435. doi: 10.1007/s12016-016-8583-2. [DOI] [PubMed] [Google Scholar]
- 16.Sharma S., Agarwal S., Kaushal K., et al. Presence and type of decompensation affects outcomes in autoimmune hepatitis upon treatment with corticosteroids. JGH Open. 2020;5:81–90. doi: 10.1002/jgh3.12451. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Alvarez F., Berg P.A., Bianchi F.B., et al. International Autoimmune Hepatitis Group Report: review of criteria for diagnosis of autoimmune hepatitis. J Hepatol. 1999;31:929–938. doi: 10.1016/s0168-8278(99)80297-9. [DOI] [PubMed] [Google Scholar]
- 18.Hennes E.M., Zeniya M., Czaja A.J., et al. Simplified criteria for the diagnosis of autoimmune hepatitis. Hepatology. 2008;48:169–176. doi: 10.1002/hep.22322. [DOI] [PubMed] [Google Scholar]
- 19.Devarbhavi H., Asrani S.K., Arab J.P., et al. Global burden of liver disease: 2023 update. J Hepatol. 2023;79:516–537. doi: 10.1016/j.jhep.2023.03.017. [DOI] [PubMed] [Google Scholar]
- 20.Tapper E.B., Parikh N.D. Diagnosis and management of cirrhosis and its complications: a review. JAMA. 2023;329:1589–1602. doi: 10.1001/jama.2023.5997. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Sarin S.K., Choudhury A., Sharma M.K., et al. Acute-on-chronic liver failure: consensus recommendations of the Asian Pacific association for the study of the liver (APASL): an update. Hepatol Int. 2019;13:353–390. doi: 10.1007/s12072-019-09946-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Hofer B.S., Simbrunner B., Hartl L., et al. Hepatic recompensation according to Baveno VII criteria is linked to a significant survival benefit in decompensated alcohol-related cirrhosis. Liver Int. 2023;43:2220–2231. doi: 10.1111/liv.15676. [DOI] [PubMed] [Google Scholar]
- 23.Hofer B.S., Burghart L., Halilbasic E., et al. Evaluation of potential hepatic recompensation criteria in patients with PBC and decompensated cirrhosis. Aliment Pharmacol Ther. 2024;59:962–972. doi: 10.1111/apt.17908. [DOI] [PubMed] [Google Scholar]
- 24.Premkumar M., Dhiman R.K., Duseja A., et al. Recompensation of Chronic Hepatitis C-related decompensated cirrhosis following direct-acting antiviral therapy: prospective Cohort study from an HCV elimination program. Gastroenterology. 2024;167:1429–1445. doi: 10.1053/j.gastro.2024.08.018. [DOI] [PubMed] [Google Scholar]
- 25.Wang Q., Zhao H., Deng Y., et al. Validation of Baveno VII criteria for recompensation in entecavir-treated patients with hepatitis B-related decompensated cirrhosis. J Hepatol. 2022;77:1564–1572. doi: 10.1016/j.jhep.2022.07.037. [DOI] [PubMed] [Google Scholar]
- 26.Watt K., Uhanova J., Gong Y., et al. Serum immunoglobulins predict the extent of hepatic fibrosis in patients with chronic hepatitis C virus infection. J Viral Hepat. 2004;11:251–256. doi: 10.1111/j.1365-2893.2004.00507.x. [DOI] [PubMed] [Google Scholar]
- 27.Shen H., Zhang M., Kaita K., et al. Expression of fc fragment receptors of immunoglobulin G (FcγRs) in rat hepatic stellate cells. Dig Dis Sci. 2005;50:181–187. doi: 10.1007/s10620-005-1298-5. [DOI] [PubMed] [Google Scholar]
- 28.Hui V.W., Wong G.L., Wong V.W., et al. Baveno VII criteria for recompensation predict transplant-free survival in patients with hepatitis B-related decompensated cirrhosis. JHEP Rep. 2023;5 doi: 10.1016/j.jhepr.2023.100814. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Deng Y., Kang H., Xiang H., et al. Durability and on-treatment predictors of recompensation in entecavir-treated patients with hepatitis B and decompensated cirrhosis. JHEP Rep. 2024;6 doi: 10.1016/j.jhepr.2024.101091. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Plagiannakos C.G., Hirschfield G.M., Lytvyak E., et al. Treatment response and clinical event-free survival in autoimmune hepatitis: a Canadian multicentre cohort study. J Hepatol. 2024;81:227–237. doi: 10.1016/j.jhep.2024.03.021. [DOI] [PubMed] [Google Scholar]
- 31.Slooter C.D., van den Brand F.F., Lleo A., et al. Lack of complete biochemical response in autoimmune hepatitis leads to adverse outcome: first report of the IAIHG retrospective registry. Hepatology. 2023;79:538–550. doi: 10.1097/HEP.0000000000000689. [DOI] [PubMed] [Google Scholar]
- 32.Colapietro F., Maisonneuve P., Lytvyak E., et al. Incidence and predictors of hepatocellular carcinoma in patients with autoimmune hepatitis. J Hepatol. 2023;80:53–61. doi: 10.1016/j.jhep.2023.09.010. [DOI] [PubMed] [Google Scholar]
- 33.Laschtowitz A., Zachou K., Lygoura V., et al. Histological activity despite normal ALT and IgG serum levels in patients with autoimmune hepatitis and cirrhosis. JHEP Rep. 2021;3 doi: 10.1016/j.jhepr.2021.100321. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Hartl J., Ehlken H., Sebode M., et al. Usefulness of biochemical remission and transient elastography in monitoring disease course in autoimmune hepatitis. J Hepatol. 2018;68:754–763. doi: 10.1016/j.jhep.2017.11.020. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
More additional data other than what were presented in the manuscript are available to request from the corresponding author. The data are not publicly available due to privacy/ethical restrictions.