metricas
covid
Buscar en
Annals of Hepatology
Toda la web
Inicio Annals of Hepatology Development and validation of a noninvasive prediction model for significant hep...
Journal Information
Vol. 29. Issue 3.
(May - June 2024)
Share
Share
Download PDF
More article options
Visits
334
Vol. 29. Issue 3.
(May - June 2024)
Original article
Full text access
Development and validation of a noninvasive prediction model for significant hepatic liver fibrosis in Chinese patients with autoimmune hepatitis
Visits
334
Hanzhu Chena,b,#, Wenya Rena,b,#, Xingdi Yangb,#, Piao Hud, Shouhao Wange, Chengan Xub, Fei Lvb, Yue Zhaob, Qiaoqiao Yinb, Wei Zhengb, Jing Xuc,
Corresponding author
xukeren@163.com

Corresponding authors.
, Hongying Panb,
Corresponding author
hypanzjsrmyy@126.com

Corresponding authors.
a The Second Clinical Medical College, Zhejiang Chinese Medical University, Hangzhou, Zhejiang 310053, PR China
b Center for General Practice Medicine, Department of Infectious Diseases, Zhejiang Provincial People's Hospital (Affiliated People's Hospital, Hangzhou Medical College), Hangzhou, Zhejiang 310014, PR China
c Hepatology Department II, Affiliated Hangzhou Xixi Hospital, Zhejiang University School of Medicine, Hangzhou, Zhejiang 310023, PR China
d The First People's Hospital of Xiaoshan District, Xiaoshan First Affiliated Hospital of Wenzhou Medical University, Hangzhou 311200 Zhejiang, PR China
e Hepatology Diagnosis and Treatment Center, The First Affiliated Hospital of Wenzhou Medical University & Zhejiang Provincial Key Laboratory for Accurate Diagnosis and Treatment of Chronic Liver Diseases, Wenzhou, Zhejiang 325035, PR China
This item has received
Article information
Abstract
Full Text
Bibliography
Download PDF
Statistics
Figures (4)
Show moreShow less
Tables (5)
Table 1. Histological characteristics of the patients.
Table 2. Baseline comparison of patients in the training and validation groups.
Table 3. Comparison of baseline characteristics and clinical data of patients in no significant and significant liver fibrosis group in the training group
Table 4. Univariate and multivariate logistic regression analysis of factors affecting significant liver fibrosis in AIH patients in the training set.
Table 5. AUROC values for comparing the different models.
Show moreShow less
Abstract
Introduction and Objectives

Autoimmune hepatitis (AIH) is a prevalent noninfectious liver disease. However, there is currently a lack of noninvasive tests appropriate for evaluating liver fibrosis in AIH patients. The objective of this study was to develop and validate a predictive model for noninvasive assessment of significant liver fibrosis (S ≥ 2) in patients to provide a reliable method for evaluating liver fibrosis in individuals with AIH.

Materials and Methods

The clinical data of 374 AIH patients were analyzed. A prediction model was established through logistic regression in the training set, and bootstrap method was used to validate the models internally. In addition, the clinical data of 109 AIH patients were collected for external verification of the model.The model was expressed as a nomogram, and area under the curve (AUC) of the receiver operating characteristic (ROC), calibration curve, and decision curve analysis were used to evaluate the accuracy of the prediction model.

Results

Logistic regression analysis revealed that age, platelet count (PLT), and the A/G ratio were identified as independent risk factors for liver fibrosis in AIH patients (P < 0.05). The diagnostic model that was composed of age, PLT and A/G was superior to APRI and FIB-4 in both the internal validation (0.872, 95%CI: 0.819–0.924) and external validation (0.829, 95%CI: 0.753–0.904).

Conclusions

Our predictive model can predict significant liver fibrosis in AIH patients more accurately, simply, and noninvasively.

Keywords:
Autoimmune hepatitis
Hepatic fibrosis
Noninvasive detection
Predictive model
External authentication
Abbreviations:
ALT
AST
TBIL
DBIL
IBIL
ALP
GGT
TP
ALB
GLO
A/G
TC
TBA
LDH
Hb
RDW
PLT
PT
INR
FIB
IgA
IgM
IgG
ANA
AMA
SMA
SLA/LP
LKM-1
AIH
HCC
APRI
FIB-4
AUROC
Full Text
1Introduction

Autoimmune hepatitis (AIH) is a liver inflammatory disease caused by immune abnormalities that lacks specific clinical symptoms and signs. It is marked by elevated circulating autoantibodies, increased concentrations of IgG, and unique histological features [1]. Clinically, it is believed to be caused by chronic and progressing inflammation of the liver, and in some cases, it eventually leads to liver inflammation, cirrhosis or failure and even death [2]. The diagnosis of this disease is challenging because the age of onset is wide, serological markers are diverse, and various clinical manifestations can range from silent disease to fulminant liver failure [3].

Liver fibrosis represents an inevitable stage in the progression from AIH to cirrhosis. Approximately 30% of patients with AIH have cirrhosis at diagnosis [1], and another 10% develop cirrhosis during follow-up [4]. Patients with untreated AIH progress to cirrhosis and may have complications such as liver failure, portal hypertension, and liver cancer. Although AIH patients achieve a biochemical response following treatment, due to the fluctuating progression of the disease, up to 40% of patients still have persistent histological activity, and fibrosis progression may still occur [5]. Therefore, patients with AIH require close, lifelong follow-up. Prednisone (or prednisolone) alone or in combination with azathioprine is considered the first-line treatment for AIH and can alleviate the symptoms of most AIH patients, achieve biochemical and histological remission, improve immediate and long-term survival, and prevent or reverse liver fibrosis [6]. In untreated patients, AIH progresses to cirrhosis at varying rates, with a subsequent risk of complications such as portal hypertension, liver failure, and hepatocellular carcinoma (HCC). Early treatment of liver fibrosis not only prevents the progression of fibrosis but also reverses it to a certain extent, resulting in a good long-term prognosis [7]. Therefore, the early diagnosis of hepatic fibrosis could have significant implications for the prognosis of AIH.

Liver biopsy is the gold standard for the diagnosis of liver fibrosis. Although the accuracy is extremely high, its repeatability is poor, it is not suitable for early liver fibrosis screening, and it is not suitable for regular monitoring as an invasive examination [8]. Traditional cortisol therapy is aimed at inhibiting inflammatory activity [9], and preventing or reversing liver fibrosis is not the main goal of treatment. At present, anti-fibrosis therapies are emerging [10,11], and in the development of anti-fibrosis interventions for autoimmune hepatitis, the first challenge is to determine the progression or regression of liver fibrosis during treatment in a reliable, noninvasive manner. Therefore, we aim to develop a repeatable and non-introgenic trauma method to assess the level of liver fibrosis in patients, which can be in determining the appropriate timing of treatment, treatment regimen, withdrawal time, etc. The currently proposed noninvasive serological models of liver fibrosis, such as AST to platelet ratio index (APRI), fibrosis-4 (FIB-4), etc., have shown poor accuracy in predicting liver fibrosis in AIH patients in existing studies [12–14]. FibroScan, a device used to detect liver fibrosis in clinical practice, is easily disturbed by liver inflammatory activity. Autoimmune hepatitis is a disease characterized by severe fluctuations in inflammatory activity, and FibroScan tends to overestimate liver fibrosis [15]. In addition, most of the existing serological diagnostic models were developed based on patients with viral hepatitis, and few noninvasive prediction models have been established based on liver fibrosis in patients with autoimmune hepatitis [12,16]. At present, no prediction models for significant liver fibrosis have been reported.

This study retrospectively collected the basic information and clinical data of AIH patients, aiming to establish a significant liver fibrosis diagnostic model suitable for AIH patients to evaluate liver fibrosis in patients with autoimmune hepatitis. At the same time, the accuracy of the model was tested by using DCA curve and calibration curve.

2Materials and Methods2.1Research object

Patients diagnosed with AIH who underwent liver biopsy at Zhejiang Provincial People's Hospital and the First People's Hospital of Xiaoshan District, Zhejiang Province, between January 2010 and October 2022 were enrolled. The diagnostic criteria for AIH were based on the simplified criteria recommended by IAIHG [17]. The exclusion criteria were as follows: (1) combined with viral hepatitis, alcoholic liver disease, nonalcoholic fatty liver disease, drug-induced liver injury, or other autoimmune liver disease; (2) hereditary and metabolic liver diseases, such as hepatolenticular degeneration and hepatic amyloidosis; (3) liver cancer and liver transplantation; (4) incomplete clinical data and laboratory test data; and (5) liver biopsy pathology that could not be used to stage liver fibrosis. In addition, the clinical data collected from Xixi Hospital, Affiliated with Zhejiang University, from January 2010 to May 2023, were used as the dataset for external independent validation.

2.2Data collection

The clinical data of the patients were collected, including age, sex and serological examination results, within a week before liver biopsy. Including serum biochemistry (TBIL, DBIL, IBIL, TP, ALB, GLO, A/G, ALT, AST, ALT/AST, GGT, ALP, LDH, TBA, and CHO), coagulation function test (PT, INR and FIB), serum immunoglobulin (IgA, IgG and IgM), blood routine blood examination (Hb, RDW, and PLT), serum autoantibodies (antinuclear antibody (ANA), smooth muscle antibody (SMA), antimitochondrial antibody (AMA), type 1 liver and kidney microsomal antibody (LKM-1) and soluble liver antigen/liver-pancreas antigen (SLA/LP).

2.3Pathological diagnosis of hepatic fibrosis

Liver biopsy specimens were immersed in 10% neutral formalin and sent to the Department of Pathology. The diagnosis was performed by a pathologist with more than five years of clinical experience and the histopathological results of the patients were collected. These samples were evaluated by experienced pathologists who were blinded to the clinical characteristics of the subjects. According to the Scheuer scoring system [18], liver fibrosis was divided into five stages. The degree of liver fibrosis S0-S1 was defined as no significant liver fibrosis, and S2–4 was defined as significant liver fibrosis [19].

2.4Statistical analysis

All the statistical analyses were performed using SPSS 25.0 software and the R 4.2.3 programming language. Continuous variables are expressed as the mean ± standard deviation or median (quartile range), and categorical variables are described by frequency. Continuous variables were compared by the t-test or rank sum test. The chi-square (X2) test or Fisher's exact test was used to compare categorical variables. Univariate and multivariate logistic regression were used to quantitatively analyze the influence of each factor on the dependent variable in the training data, and the prediction model was constructed. The Bootstrap method was used to verify the validation dataset internally, and the validation dataset was used for external verification. The model was expressed as a nomogram, and AUC, calibration curve, and decision curve analysis were used to evaluate the predictive accuracy of the model. P < 0.05 was considered to indicate statistical significance.

2.5Ethical statement

This study was approved by the local hospital's ethics committee and complied with the ethical guidelines of the Declaration of Helsinki(QT2023316). The researchers only analyzed anonymous data, so the Ethics Committee approved the waiver of informed consent.

3Results3.1Analysis of the clinical laboratory parameters and histological characteristics of the patients

A total of 483 patients diagnosed with AIH met the inclusion criteria in this study (Fig. 1). A total of 62 male patients and 421 female patients were incleded. The age of the 483 patients ranged from 26 to 75 years old, with a median age of 53 years old. We summarized the histological features of AIH patients without significant liver fibrosis and significant liver fibrosis (Table 1). It can be seen that AIH patients with significant liver fibrosis have a higher probability of interface hepatitis, lymphocyte/plasma cell infiltration, rosettes, emperipolesis, and bile duct changes. A total of 374 patients from Zhejiang Provincial People's Hospital and the First People's Hospital of Xiaoshan District were used as the training group. A total of 109 patients from Xixi Hospital, affiliated with Zhejiang University, were included in the external validation group. The comparison of baseline information between the two groups is shown in Table 2.

Fig. 1.

Describes the flow chart of the participant selection process. (A) Set from Zhejiang Provincial People's Hospital and the First People's Hospital of Xiaoshan District; (B) Set from Xixi Hospital Affiliated to Zhejiang University. AIH, Autoimmune hepatitis.

(0.18MB).
Table 1.

Histological characteristics of the patients.

Histological feature  Nonsignificant liver fibrosis  Significant liver fibrosis  p value 
Interface hepatitis  89 (80.91%)  357 (95.71%)  <0.001 
Lymphocyte/plasma cell predominance  64 (58.18%)  314 (84.18%)  <0.001 
Rosettes  17 (15.45%)  222 (59.52%)  <0.001 
Emperipolesis  2 (1.82%)  83 (22.25%)  <0.001 
Bile duct changes  27 (24.55%)  162 (43.43%)  <0.001 
Steatosis  23 (20.91%)  85 (22.79%)  0.678 
Table 2.

Baseline comparison of patients in the training and validation groups.

Variables  Training set (n = 374)  Validation set (n = 109)  P value 
Male sex (%)  44 (11.76%)  18 (16.51%)  0.192 
Age (year)  54 (45–57)  50 (45–56)  0.457 
Significant fibrosis       
No  57 (15.24%)  53 (48.62%)  <0.001 
Yes  317 (84.76%)  56 (51.38%)   
Laboratory data       
ALT (U/L)  56.00 (28.00–126.25)  28.00 (18.00–58.00)  <0.001 
AST (U/L)  58.00 (27.00–115.00)  31.00 (22.00–57.50)  <0.001 
ALT/AST  1.01 (0.78–1.37)  0.86 (0.60–1.17)  <0.001 
TBIL (μmol/L)  19.00 (12.58–36.60)  16.40 (11.50–30.55)  0.088 
DBIL (μmol/L)  6.87 (4.07–21.97)  4.40 (2.35–11.75)  <0.001 
IBIL (μmol/L)  10.65 (7.40–16.75)  11.80 (8.75–16.90)  0.067 
GGT (U/L)  108.00 (40.00–291.25)  52.00 (20.00–116.00)  <0.001 
ALP (U/L)  137.00 (97.00–242.00)  109.00 (79.00–155.50)  <0.001 
TP (g/L)  72.52±8.25  67.60±7.62  <0.001 
ALB (g/L)  37.35 (33.75–41.03)  36.30 (32.95–40.60)  0.379 
GLO (g/L)  35.00 (30.58–39.53)  29.40 (25.75–35.05)  <0.001 
A/G  1.07 (0.90–1.26)  1.29 (1.04–1.53)  <0.001 
TC (mmol/L)  5.67 (4.60–6.78)  4.76 (3.93–5.69)  <0.001 
TBA (μmol/L)  16.40 (8.65–52.60)  19.10 (10.35–48.50)  0.682 
LDH (U/L)  184.00 (161.75–226.50)  180.00 (147.50–235.00)  0.442 
ALP/AST  3.19 (1.75–4.82)  5.60 (2.74–7.89)  0.513 
Hb (g/L)  127.00 (113.00–136.00)  131.00 (114.00–140.00)  0.073 
RDW (%)  13.60 (13.00–15.13)  14.00 (13.05–15.85)  0.196 
PLT (×109/L)  196.00 (147.75–244.00)  187.00 (128.50–221.00)  0.041 
PT (s)  11.45 (10.90–12.00)  11.30 (10.45–12.30)  0.224 
INR  1.02 (0.95–1.08)  1.04 (0.97–1.15)  <0.001 
FIB (g/L)  2.26 (1.79–2.77)  2.44 (1.95–3.05)  0.050 
IgA (g/L)  3.03 (2.33–3.75)  2.49 (1.99–2.96)  <0.001 
IgM (g/L)  1.73 (1.21–3.72)  1.29 (0.94–2.01)  <0.001 
IgG (g/L)  14.80 (12.60–18.13)  13.00 (11.50–15.50)  <0.001 
ANA (+)  354 (94.65%)  92 (84.40%)  <0.001 
AMA (+)  80 (21.39%)  2 (1.83%)  <0.001 
SMA (+)  17 (4.55%)  6 (5.50%)  0.317 
SLA/LP (+)  4 (1.07%)  0 (0.00%)  0.579 
LKM-1 (+)  0 (0.00%)  0 (0.00%)  – 

ALT, alanine transaminase; AST, aspartate aminotransferase; TBIL, total bilirubin; DBIL, direct bilirubin; IBIL, indirect bilirubin; GGT, -glutamyl transferase; ALP, alkaline phosphatase; TP, total protein; ALB, albumin; GLO, globulin; A/G, albumin-to-globulin ratio; TC, total cholesterol; TBA, total bile acid; LDH, lactate dehydrogenase; Hb, hemoglobin; RDW, red blood cell distribution width; PLT, platelet count; PT, prothrombin time; INR, national standardized ratio; FIB, fibrinogen; IgA, immunoglobulin A; IgM, immunoglobulin M; IgG, immunoglobulin G; ANA, antinuclear antibody; AMA, antimitochondrial antibody; SMA, smooth muscle antibody; SLA/LP, soluble liver antigen/liver-pancreas antigen; LKM-1, liver kidney microsomes type-1.

3.2Establishment of predictors and models for significant hepatic fibrosis

The Scheuer scoring system was used to classify liver fibrosis, and the 374 patients in the training group were then divided into non-significant liver fibrosis group (n = 57) and the significant liver fibrosis group (n = 317). There were statistically significant differences observed in sex, age, TP, GLO, A/G, Hb, PLT and IgG between the two groups (P<0.05) (Table 3). Univariate and multivariate logistic regression analyses showed that age, A/G and PLT were independent risk factors for significant liver fibrosis in AIH patients (P < 0.05) (Table 4). The corresponding prediction model was constructed with the above three indicators and named the AP model. The formula is as follows: AP = 7.624 + 0.081 × Age - 4.848 × A/G - 0.019 × PLT.

Table 3.

Comparison of baseline characteristics and clinical data of patients in no significant and significant liver fibrosis group in the training group

Variables  Nonsignificant liver fibrosisn=57(%)  Significant liver fibrosisn=317(%)  P value 
Male sex (%)  1(1.75%)  43(13.56%)  0.011 
Age (year)  47(40-57)  54(48-57)  0.013 
ALT (U/L)  78.00(31.50-203.50)  54.00(28.00-121.00)  0.087 
AST (U/L)  68.00(33.50-168.50)  56.00(27.00-112.00)  0.064 
ALT/AST  0.95(0.75-1.27)  1.04(0.81-1.38)  0.249 
TBIL (μmol/L)  22.40(12.10-54.00)  18.60(12.71-32.25)  0.168 
DBIL (μmol/L)  7.50(3.82-27.95)  6.79(4.15-18.60)  0.529 
IBIL (μmol/L)  10.20(8.00-19.25)  10.50(7.15-16.35)  0.255 
GGT (U/L)  104.00(30.00-383.00)  112.00(41.00-290.50)  0.825 
ALP (U/L)  141.00(83.50-377.50)  137.00(99.00-232.50)  0.934 
TP (g/L)  70.09±9.90  72.96±7.85  0.042 
ALB (g/L)  39.00(32.10-44.05)  37.20(33.95-40.50)  0.112 
GLO (g/L)  30.30(26.65-36.75)  35.30(31.65-40.00)  <0.001 
A/G  1.25(1.03-1.46)  1.04(0.90-1.20)  <0.001 
TC(mmol/L)  4.96(4.25-6.89)  5.81(4.73-6.78)  0.071 
TBA(μmol/L)  13.20(4.75-96.60)  16.60(9.50-48.55)  0.459 
LDH(U/L)  184.00(163.00-238.50)  184.00(160.00-223.50)  0.680 
ALP/AST  2.58(1.33-4.48)  3.27(1.76-5.04)  0.121 
Hb (g/L)  117.00(103.00-135.00)  128.00(115.00-137.00)  0.018 
RDW(%)  14.00(12.90-17.05)  13.60(13.00-15.00)  0.086 
PLT(×109/L)  246.00(208.50-342.50)  181.00(141.50-235.00)  <0.001 
PT(s)  11.40(11.10-11.85)  11.50(10.90-12.00)  0.902 
INR  1.02(0.96-1.07)  1.02(0.94-1.08)  0.340 
FIB(g/L)  2.25(1.97-2.90)  2.27(1.77-2.78)  0.352 
IgA(g/L)  2.94(1.06-4.33)  3.04(2.47-3.71)  0.119 
IgM(g/L)  2.45(1.23-3.81)  1.67(1.21-3.70)  0.467 
IgG(g/L)  13.60(10.59-15.55)  15.30(12.80-18.40)  0.001 
ANA(+)  57(100.00%)  297(93.69%)  0.103 
AMA(+)  17(29.82%)  63(19.87%)  0.092 
SMA(+)  0(0.00%)  17(5.36%)  0.149 
SLA/LP(+)  0(0.00%)  4(1.26%)  1.000 
LKM-1(+)  0(0.00%)  0(0.00%)  – 

ALT, alanine transaminase; AST, aspartate aminotransferase.

Table 4.

Univariate and multivariate logistic regression analysis of factors affecting significant liver fibrosis in AIH patients in the training set.

VariablesUnivariableMultivariable
Coefficient  OR (95% CI)  P value  Coefficient  OR (95% CI)  P value 
Male sex (%)  2.173  8.788 (1.185–65.155)  0.033       
Age (year)  0.052  1.053 (1.020–1.087)  0.002  0.081  1.084 (1.035–1.135)  0.001 
TP (g/L)  0.042  1.043 (1.008–1.080)  0.016       
GLO (g/L)  0.101  1.106 (1.057–1.158)  <0.001       
A/G  −2.862  0.057 (0.017–0.192)  <0.001  −4.848  0.008 (0.001–0.041)  <0.001 
Hb (g/L)  0.019  1.019 (1.004–1.035)  0.016       
PLT (×109/L)  −0.013  0.987 (0.983–0.991)  <0.001  −0.019  0.981 (0.976–0.986)  <0.001 
IgG (g/L)  0.112  1.118 (1.033–1.211)  0.006       
Constant        7.624  2046.292  <0.001 

TP, total protein; GLO, globulin; A/G, albumin-to-globulin ratio; Hb, hemoglobin; PLT, platelet count; IgG, immunoglobulin G.

The cut-off value and Youden index of the model were calculated. In the training group, when the patients were higher than the cut-off value (>1.917), there was significant liver fibrosis, and when the patients were lower than the cut-off value (<1.917), there was no significant liver fibrosis.

3.3Model comparison and verification

Based on the prediction model, we generated a clinical nomogram for model visualization and easy operation (Fig. 2). The training set was internally validated by bootstrap sampling 1000 times, and the AUCs were 0.872 (95% CI: 0.819–0.924) and 0.829 (95% CI: 0.753–0.904), which were better than FIB-4 and APRI. This suggests that the diagnostic performance of the established prediction model is higher than that of the above serological model (Table 5). In the calibration plot, the calibration line and, the ideal line apparent curve (actual) and the deviation correction curve (1000 bootstrap adjustments) are very close to the ideal curve and show good consistency in the training set and the verification set (Fig. 3). At the same time, we established a decision curve analysis (Fig. 4) to evaluate clinical utility. We can conclude that the model has obvious net benefits and high clinical application value.

Fig. 2.

A nomogram was established to predict significant liver fibrosis in patients with AIH. A/G, albumin-to-globulin ratio; PLT, platelet count.

(0.07MB).
Table 5.

AUROC values for comparing the different models.

AUROCSEP95% CI
LOWER  UPPER 
Training set           
FIB-4  0.643  0.038  0.001  0.567  0.718 
APRI  0.523  0.039  0.583  0.446  0.600 
AP  0.872  0.027  <0.001  0.819  0.924 
Validation set           
FIB-4  0.784  0.044  <0.001  0.698  0.869 
APRI  0.756  0.046  <0.001  0.666  0.847 
AP  0.829  0.039  <0.001  0.753  0.904 

FIB-4, fibrosis-4; APRI, the AST/platelet ratio index; AP, AP model.

Fig. 3.

Calibration diagram of the prediction model. Calibration curve for (a) the model (B = 1000rpetitions, boot; n = 374; Mean absolute error = 0.009) and (b) external validation for the model (B = 1000rpetitions, boot; n = 109; Mean absolute error=0.041).

(0.12MB).
Fig. 4.

Constructs the decision curve of the AP model, FIB-4 and APRI net benefit prediction. Net benefit of screening patients according to the AP model, FIB-4 and APRI in the (a) training set, (b) validation set. FIB-4, fibrosis-4; APRI, the AST/platelet ratio index; AP, AP model.

(0.24MB).
4Discussion

AIH is an incurable chronic inflammatory liver disease that can lead to cirrhosis, HCC, decompensation and death after immunosuppressive therapy. According to statistics, approximately one-third of patients with AIH already have signs of liver cirrhosis when they are first diagnosed, and AIH patients who receive immunosuppressive therapy still progress to liver cirrhosis at a rate ranging from 0.1 to 8.1% per year [20]. In clinical practice, it is impossible to use a sequential biopsy to regularly monitor the stages of liver fibrosis. Therefore, noninvasive assessment of fibrosis is of great significance. First, a small number of AIH patients showed mild elevations in transaminases and serum IgG levels at the time of first detection, while inflammatory activity and fibrosis progression in the liver may not be as easy as serological manifestations [21]. In these patients, the indications for starting immunotherapy are not obvious, but no treatment may also lead to the rapid progression of liver fibrosis. Secondly, after receiving immunosuppressive drugs for biochemical remission, normal transaminase levels do not rule out the activity of liver inflammation and progressive fibrosis [22]. Therefore, developing accurate, noninvasive methods to assess disease progression and guide treatment is crucial. We established a model of significant liver fibrosis in patients with AIH In order to predict early liver fibrosis in AIH.

First, we screened out AIH patients through simplified criteria. The inclusion criteria were determined by referring to the AIH diagnostic scoring system and previously reported factors related to liver fibrosis. Then, the model was established by univariate and multivariate logistic regression analysis. We found that three indicators, including age, A/G, and PLT were independent predictors of significant liver fibrosis in patients with autoimmune hepatitis. Age was a risk factor, while A/G and PLT were protective factors. These indicators are often routinely tested on admission and are very easy to obtain.

Age is the main factor of increased susceptibility to fibrosis [23]. The models currently reported, such as FIB-4 [24], HB-F [25], and King's score [26], all include age as a variable and a number of studies have shown that liver fibrosis is positively correlated with age [27–29]. Our study also confirmed that age is an independent risk factor for the development of significant liver fibrosis in patients with AIH. Some studies have found that [30] the age distribution of the onset of AIH is a double-peak distribution, and the two peaks are located around puberty and 40–60 years old. In this study, 74.6% of the patients with AIH were between 40 and 60 years old, which was consistent with the above conclusion. The reason for the bimodal distribution of age may be that when women enter puberty and menopause, a significant change in estrogen levels will lead to the occurrence of autoimmune diseases [31,32].

In this study, A/G is a protective factor for significant liver fibrosis in AIH patients. It has been reported in the literature that A/G levels may decrease with the progression of liver fibrosis [33,34]. The elevation of IgG and/or γ-globulins is one of the characteristic serological changes of AIH [35]. With the development of liver fibrosis, the liver synthesis function of patients with liver cancer is damaged; therefore, the synthesis of ALB is reduced, the production of GLO is increased, and the A/G ratio is gradually reduced. The A/G ratio is a good indicator for observing the progression of chronic liver disease [36–39]. Currently, there is no cohort study on AIH to verify the relationship between A/G and liver fibrosis. However, studies on viral liver and nonalcoholic fatty liver have shown that A/G levels are negatively related to the degree of liver fibrosis in patients with chronic hepatitis B and nonalcoholic fatty liver [25,33,34]. For example, Sripongpun et al. [34] reported a new fibrosis-8 score (FIB-8) model. In their model, A/G levels were a protective factor for significant liver fibrosis in patients with nonalcoholic fatty liver disease.

PLT is also an independent risk factor for significant liver fibrosis in AIH patients. Thrombocytopenia is a common complication associated with liver cirrhosis, which is exacerbated by decreased circulating platelet count due to an enlarged spleen, increased platelet consumption, and decreased synthesis and increased degradation of thrombopoietin (TPO) [40,41]. In addition to thrombocytopenia, qualitative platelet defects may occur, which may be due to decreased TXA 2 synthesis, decreased ATP and serotonin concentrations in dense granules, increased platelet inhibitor production, and impaired response to profibrotic agents [42], ultimately leading to PLT decline. Numerous studies have shown that PLT counts in patients with AIH are inversely correlated with the degree of fibrosis [43,44], which is consistent with our current study.

In this study, we constructed a new noninvasive model for predicting significant liver fibrosis in AIH, which was named AP model. The AUROC of AP model was superior to APRI and FIB-4 in internal validation (0.872,95% CI: 0.819–0.924) and external validation (0.829,95% CI: 0.753–0.904). In the training cohort, when the cut-off value was 1.917, the sensitivity of AIH was 75.7% and the specificity was 86.0%. This method successfully predicted significant fibrosis in 240 patients (75.71%) in the training set and 40 patients (71.43%) in the validation set. It shows that the model has good practicability in the noninvasive diagnosis of significant liver fibrosis in AIH.

This study also found that APRI and FIB-4 had poor accuracy in predicting significant liver fibrosis in AIH. The AUROC of APRI was 0.523(95%CI: 0.446–0.600) in the training group and 0.756 (95%CI: 0.666–0.847) in the validation group. The AUROC of FIB-4 was 0.643 (95%CI: 0.567–0.718) in the training group and 0.784 (95%CI: 0.698–0.869) in the validation group. The AUROC of the two models is smaller than that of the AP model. There was no strong correlation between APRI and FIB-4 scores and fibrosis stage in the training cohort. However, in the external validation cohort, these two scores have certain diagnostic values. The reason may be that the results are more heterogeneous because of the small sample size. It may also be because APRI and FIB-4 were originally developed based on the data of patients with chronic hepatitis C, and the diagnostic efficacy for AIH was poor [45–47]. In previous studies, the prediction accuracy of APRI and FIB-4 for AIH liver fibrosis also showed heterogeneity. From the meta-analysis of APRI and FIB-4 in the detection of AIH liver fibrosis by Bingtian Dong et al. [43], they believed that APRI (AUC: 0.66, 95%CI: 0.61–0.70) and FIB-4 (AUC: 0.75,95%CI: 0.71–0.79) were not ideal for the diagnosis of AIH liver fibrosis. The current number of published original articles on APRI and FIB-4 for assessing AIH liver fibrosis is limited, and the accuracy of APRI and FIB-4 requires large-scale and multi-center studies to further evaluate.

In addition, we also used a variety of methods to analyze the efficacy of the model. The calibration curve showed a strong consistency between the predicted value and the actual observed value. DCA showed that the model had high clinical applicability.

Our study has several advantages. First, according to current reports, this study established a noninvasive model to predict significant liver fibrosis in AIH patients for the first time [16,48-50]. Second, it was a multi-center and retrospective study, which reduces the differences in patient composition between different hospitals. Third, an independent cohort was recruited for external verification to provide a more convincing conclusion for the accuracy of predicting liver fibrosis. Fourth, the model is based on the relevant clinical data of AIH patients, which is more applicable to AIH patients. Fifth, the model was used to predict significant liver fibrosis in AIH patients, which is conducive to early intervention for liver fibrosis to obtain a better prognosis. Sixth, the variables in the model were determined by serological indicators, which are common laboratory tests and can be easily obtained at all levels of medical institutions. It is scientific to use the results of liver biopsy as an index to evaluate the degree of liver fibrosis. Therefore, the model could be used as a noninvasive indicator in primary hospitals to assess whether patients have significant liver fibrosis.

However, our research still has certain limitations. First, a total of 473 patients were included in this study, so the sample size was not particularly large. In future research, we will continue to train and verify our model through subsequent large-scale multi-center research and external verification studies. Second, the population included in this study is Chinese patients, so it is not clear whether the model is applicable to other races. Future studies need to evaluate the applicability of the model in a more diverse patient population.

5Conclusions

We constructed a noninvasive, highly accurate model to distinguish whether AIH patients have significant liver fibrosis. This model is clinically beneficial to reduce the need for liver biopsy, helps the clinician to identify and treat these high-risk patients early and follow-up of AIH patients, and assists in formulating strategies and timing of immunosuppressive therapy in order to obtain a better prognosis and even reverse liver fibrosis.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Author contributions

Hongying Pan is the guarantor of the article. Specific author contributions: Hongying Pan and Jing Xu were responsible for project conceptualization; Hanzhu Chen: performed the research; Hanzhu Chen, Shouhao Wang and Chengan Xu: collected and analyzed the data; Hanzhu Chen wrote the manuscript; Wei Zheng, Qiaoqiao Yin, Fei Lv and Yue Zhao: critically reviewed the manuscript. All authors approved the final version of the article including the authorship list.

References
[1]
EASL clinical practice guidelines: autoimmune hepatitis.
J Hepatol, 63 (2015), pp. 971-1004
[2]
C. Harrington, S. Krishnan, C.L. Mack, P. Cravedi, D.N. Assis, J. Levitsky.
Noninvasive biomarkers for the diagnosis and management of autoimmune hepatitis.
Hepatology, 76 (2022), pp. 1862-1879
[3]
M.N. Rahim, R. Miquel, M.A. Heneghan.
Approach to the patient with acute severe autoimmune hepatitis.
[4]
M. Werner, H. Prytz, B. Ohlsson, S. Almer, E. Björnsson, A. Bergquist, et al.
Epidemiology and the initial presentation of autoimmune hepatitis in Sweden: a nationwide study.
Scandinav J Gastroenterol, 43 (2008), pp. 1232-1240
[5]
H.K. Dhaliwal, B.S. Hoeroldt, A.K. Dube, E. McFarlane, J.C. Underwood, M.A. Karajeh, et al.
Long-term prognostic significance of persisting histological activity despite biochemical remission in autoimmune hepatitis.
Amer J Gastroenterol, 110 (2015), pp. 993-999
[6]
A.J. Czaja.
Review article: permanent drug withdrawal is desirable and achievable for autoimmune hepatitis.
Alimen Pharmacol Therap, 39 (2014), pp. 1043-1058
[7]
L. Muratori, A.W. Lohse, M. Lenzi.
Diagnosis and management of autoimmune hepatitis.
BMJ (Clinical research ed), 380 (2023),
[8]
G. Shiha, A. Ibrahim, A. Helmy, S.K. Sarin, M. Omata, A. Kumar, et al.
Asian-pacific association for the study of the liver (APASL) consensus guidelines on invasive and noninvasive assessment of hepatic fibrosis: a 2016 update.
Hepatol Int, 11 (2016), pp. 1-30
[9]
G.C. Cook, R. Mulligan, S. Sherlock.
Controlled prospective trial of corticosteroid therapy in active chronic hepatitis.
[10]
Z. Ghiassi-Nejad, S.L. Friedman.
Advances in antifibrotic therapy.
Expert Rev Gastroenterol Hepatol, 2 (2008), pp. 803-816
[11]
F. Tacke, R. Weiskirchen.
An update on the recent advances in antifibrotic therapy.
Expert Rev Gastroenterol Hepatol, 12 (2018), pp. 1143-1152
[12]
L.P. Llovet, J. Gratacós-Ginès, L. Téllez, A. Gómez-Outomuro, C.A. Navascués, M. Riveiro-Barciela, et al.
Noninvasive prediction of outcomes in autoimmune hepatitis-related cirrhosis.
Hepatol Commun, 6 (2022), pp. 1392-1402
[13]
X. Yuan, S.Z. Duan, J. Cao, N. Gao, J. Xu, L. Zhang.
Noninvasive inflammatory markers for assessing liver fibrosis stage in autoimmune hepatitis patients.
Europ J Gastroenterol hepatol, 31 (2019), pp. 1467-1474
[14]
L. Liu, J. Cao, Z. Zhong, Z. Guo, Y. Jiang, Y. Bai, et al.
Noninvasive indicators predict advanced liver fibrosis in autoimmune hepatitis patients.
J Clin Laborat Anal, 33 (2019), pp. e22922
[15]
P. Romanque, F. Stickel, J.F. Dufour.
Disproportionally high results of transient elastography in patients with autoimmune hepatitis.
Liver Int, 28 (2008), pp. 1177-1178
[16]
Z. Zhang, J. Wang, H. Wang, Y. Qiu, L. Zhu, J. Liu, et al.
An easy-to-use AIHF-nomogram to predict advanced liver fibrosis in patients with autoimmune hepatitis.
[17]
E.M. Hennes, M. Zeniya, A.J. Czaja, A. Parés, G.N. Dalekos, E.L. Krawitt, et al.
Simplified criteria for the diagnosis of autoimmune hepatitis.
Hepatology, 48 (2008), pp. 169-176
[18]
P.J. Scheuer.
Classification of chronic viral hepatitis: a need for reassessment.
J Hepatol, 13 (1991), pp. 372-374
[19]
EASL 2017 Clinical Practice Guidelines on the management of hepatitis B virus infection.
J Hepatol, 67 (2017), pp. 370-398
[20]
A. Tansel, L.H. Katz, H.B. El-Serag, A.P. Thrift, M. Parepally, M.H. Shakhatreh, et al.
Incidence and determinants of hepatocellular carcinoma in autoimmune hepatitis: a systematic review and meta-analysis.
Clin Gastroenterol Hepatol, 15 (2017), pp. 1207-1217
[21]
J. Hartl, R. Miquel, K. Zachou, G.W. Wong, A. Asghar, S. Pape, et al.
Features and outcome of AIH patients without elevation of IgG.
[22]
A. Laschtowitz, K. Zachou, V. Lygoura, S. Pape, F. Derben, E. Jaeckel, et al.
Histological activity despite normal ALT and IgG serum levels in patients with autoimmune hepatitis and cirrhosis.
[23]
M. Mahrouf-Yorgov, A. Collin de l'Hortet, C. Cosson, A. Slama, E. Abdoun, J.E. Guidotti, et al.
Increased susceptibility to liver fibrosis with age is correlated with an altered inflammatory response.
Rejuven Res, 14 (2011), pp. 353-363
[24]
C.W. Hsu, K.H. Liang, S.F. Huang, K.C. Tsao, C.T. Yeh.
Development of a noninvasive fibrosis test for chronic hepatitis B patients and comparison with other unpatented scores.
BMC Res Notes, 6 (2013), pp. 212
[25]
T. Prasoppokakorn, W.K. Chan, V.W. Wong, P. Pitisuttithum, S. Mahadeva, N.R. Nik Mustapha, et al.
Validation model of fibrosis-8 index score to predict significant fibrosis among patients with nonalcoholic fatty liver disease.
World J Gastroenterol, 28 (2022), pp. 1563-1573
[26]
T.J. Cross, P. Rizzi, P.A. Berry, M. Bruce, B. Portmann, P.M. Harrison.
King's Score: an accurate marker of cirrhosis in chronic hepatitis C.
Europ J Gastroenterol Hepatol, 21 (2009), pp. 730-738
[27]
H.B. Liu, J.P. Zhou, Y. Zhang, X.H. Lv, W. Wang.
Prediction on liver fibrosis using different APRI thresholds when patient age is a categorical marker in patients with chronic hepatitis B.
Clinica Chimica Acta; Int J Clin Chem, 412 (2011), pp. 33-37
[28]
W. Udomsinprasert, A. Sobhonslidsuk, J. Jittikoon, S. Honsawek, U. Chaikledkaew.
Cellular senescence in liver fibrosis: implications for age-related chronic liver diseases.
Expert Opin Therap Targets, 25 (2021), pp. 799-813
[29]
V. Gordon, R. Adhikary, V. Appleby, D. Das, J. Day, T. Delahooke, et al.
Diagnosis, presentation and initial severity of autoimmune hepatitis (AIH) in patients attending 28 hospitals in the UK.
Liver Int, 38 (2018), pp. 1686-1695
[30]
A. Floreani, P. Restrepo-Jiménez, M.F. Secchi, S. De Martin, P.S.C. Leung, E. Krawitt, et al.
Etiopathogenesis of autoimmune hepatitis.
J Autoimmun, 95 (2018), pp. 133-143
[31]
L. Henze, D. Schwinge, C. Schramm.
The effects of androgens on T cells: clues to female predominance in autoimmune liver diseases?.
Front Immunol, 11 (2020), pp. 1567
[32]
V.R. Moulton.
Sex hormones in acquired immunity and autoimmune disease.
Front Immunol, 9 (2018), pp. 2279
[33]
J. Li, R.C. Mao, X.L. Li, J.W. Zheng, X. Qi, Q. Yuan, et al.
A novel noninvasive index for the prediction of moderate to severe fibrosis in chronic hepatitis B patients.
Diges Liver Disease, 50 (2018), pp. 482-489
[34]
Sripongpun P. A.MD., Alexis Touros and W. Ray Kim. FIB-8 score: a model incorporating additional common variables into the FIB-4 score affords better prediction of significant fibrosis in non-alcoholic fatty liver disease. AASLD Abstract publication; 2019.
[35]
A. Komori.
Recent updates on the management of autoimmune hepatitis.
Clin Mol Hepatol, 27 (2021), pp. 58-69
[36]
M. Domenicali, M. Baldassarre, F.A. Giannone, M. Naldi, M. Mastroroberto, M. Biselli, et al.
Posttranscriptional changes of serum albumin: clinical and prognostic significance in hospitalized patients with cirrhosis.
Hepatology, 60 (2014), pp. 1851-1860
[37]
R. Spinella, R. Sawhney, R. Jalan.
Albumin in chronic liver disease: structure, functions and therapeutic implications.
Hepatol Int, 10 (2016), pp. 124-132
[38]
N.K. Gatselis, K. Zachou, G.K. Koukoulis, G.N. Dalekos.
Autoimmune hepatitis, one disease with many faces: etiopathogenetic, clinico-laboratory and histological characteristics.
World J Gastroenterol, 21 (2015), pp. 60-83
[39]
P.Y. Kwo, S.M. Cohen, J.K. Lim.
ACG clinical guideline: evaluation of abnormal liver chemistries.
Amer J Gastroenterol, 112 (2017), pp. 18-35
[40]
J. Muciño-Bermejo, R. Carrillo-Esper, M. Uribe, N. Méndez-Sánchez.
Coagulation abnormalities in the cirrhotic patient.
Annal Hepatol, 12 (2013), pp. 713-724
[41]
T. Milovanovic Alempijevic, M. Stojkovic Lalosevic, I. Dumic, N. Jocic, A. Pavlovic Markovic, S. Dragasevic, et al.
Diagnostic accuracy of platelet count and platelet indices in noninvasive assessment of fibrosis in nonalcoholic fatty liver disease patients.
Canad J Gastroenterol Hepatol, 2017 (2017),
[42]
S. Offermanns.
Activation of platelet function through G protein-coupled receptors.
Circul Res, 99 (2006), pp. 1293-1304
[43]
B. Dong, Y. Chen, G. Lyu, X. Yang.
Aspartate aminotransferase to platelet ratio index and Fibrosis-4 index for detecting liver fibrosis in patients with autoimmune hepatitis: a meta-analysis.
[44]
H. Wang, J. Wang, J. Xia, X. Yan, Y. Feng, L. Li, et al.
Red cell distribution width to platelet ratio predicts liver fibrosis in patients with autoimmune hepatitis.
Medicine, 99 (2020), pp. e21408
[45]
C.T. Wai, J.K. Greenson, R.J. Fontana, J.D. Kalbfleisch, J.A. Marrero, H.S. Conjeevaram, et al.
A simple noninvasive index can predict both significant fibrosis and cirrhosis in patients with chronic hepatitis C.
Hepatology, 38 (2003), pp. 518-526
[46]
R.K. Sterling, E. Lissen, N. Clumeck, R. Sola, M.C. Correa, J. Montaner, et al.
Development of a simple noninvasive index to predict significant fibrosis in patients with HIV/HCV coinfection.
Hepatology, 43 (2006), pp. 1317-1325
[47]
L.A. Adams, M. Bulsara, E. Rossi, B. DeBoer, D. Speers, J. George, et al.
Hepascore: an accurate validated predictor of liver fibrosis in chronic hepatitis C infection.
Clin Chem, 51 (2005), pp. 1867-1873
[48]
Z. Zhang, J. Wang, H. Wang, Y. Li, L. Zhu, Y. Chen, et al.
Develop and validate a novel online AIHI-nomogram to predict severe liver inflammation in patients with autoimmune hepatitis.
[49]
K. Gutkowski, M. Hartleb, T. Kacperek-Hartleb, M. Kajor, W. Mazur, W. Zych, et al.
Laboratory-based scoring system for prediction of hepatic inflammatory activity in patients with autoimmune hepatitis.
Liver Int, 33 (2013), pp. 1370-1377
[50]
H. Yang, L. Huang, Y. Xie, M. Bai, H. Lu, S. Zhao, et al.
A diagnostic model of autoimmune hepatitis in unknown liver injury based on noninvasive clinical data.
Scient Reports, 13 (2023), pp. 3996

Hanzhu Chen, Wenya Ren and Xingdi Yang contributed equally to this paper.

Copyright © 2024. Fundación Clínica Médica Sur, A.C.
Article options
es en pt

¿Es usted profesional sanitario apto para prescribir o dispensar medicamentos?

Are you a health professional able to prescribe or dispense drugs?

Você é um profissional de saúde habilitado a prescrever ou dispensar medicamentos