Diagnostic evaluation of non-invasive liver fibrosis models and establishment of a new model in chronic hepatitis B patients complicated with nonalcoholic fatty liver disease
10.3760/cma.j.cn311367-20231227-00230
- VernacularTitle:慢性乙型肝炎合并非酒精性脂肪性肝病患者无创肝纤维化模型的诊断评价及新模型的建立
- Author:
Yinghui GAO
1
;
Mingyue DENG
1
;
Ruirui ZHU
1
;
Zhixian GUO
1
;
Jingya YAN
1
;
Xuemeng ZHAO
1
;
Zhiqin LI
1
Author Information
1. 郑州大学第一附属医院感染科，郑州 450052
- Publication Type:Journal Article
- Keywords:
Chronic viral hepatitis B;
Non-alcoholic fatty liver disease;
Liver fibrosis;
Noninvasive diagnostic model
- From:
Chinese Journal of Digestion
2024;44(10):686-692
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To evaluate the diagnostic efficacy of FibroScan combined with various noninvasive diagnostic models for liver fibrosis in patients with chronic hepatitis B (CHB) complicated with nonalcoholic fatty liver disease (NAFLD), and to establish a new predictive model with common clinical indicators.Methods:From January 2016 to May 2024, the clinical data of 118 CHB patients complicated with NAFLD from the First Affiliated Hospital of Zhengzhou University, who underwent liver biopsy and FibroScan examination were respectively analyzed. According to the Scheuer scoring system, different diagnostic endpoints were set based on the degree of liver fibrosis (S0 to S1, ≥S2, ≥S3, and S4), fibrosis stage ≥S2 was designated as the criterion for significant liver fibrosis. Fibrosis-4 (FIB-4), γ-glutamyl transpeptidase (GGT) to platelet ratio (GPR), GGT-age-platelet-international normalized ratio (GAPI) model, S index, King index and Forns index were calculated according to the common clinical indicators. The independent t test or Mann-Whitney U test was used to compare the two groups. Spearman rank correlation was used to analyze the correlation between each noninvasive diagnostic method and the degree of liver fibrosis. Receiver operating characteristic curve (ROC) was plotted, and the DeLong test was performed to compare the area under the curve (AUC), and to evaluate the predictive value of FibroScan combined with various noninvasive diagnostic models for the diagnosis of liver fibrosis. The laboratory indicators were compared between patients with non-significant liver fibrosis and patients with significant liver fibrosis. And the indicators with statistically significant differences ( P<0.05) in the univariate analysis were further analyzed by multivariate logistic regression to establish a new predictive model for liver fibrosis. Hosmer-Lemeshow test was used to assess the model′s goodness of fit. Results:The results of Spearman rank correlation showed that FIB-4, GPR, FibroScan, GAPI model, S index, King index, and Forns index were positively correlated with the stage of liver fibrosis ( r=0.413, 0.458, 0.512, 0.473, 0.533, 0.380 and 0.478, all P<0.001). The results of ROC analysis indicated that among combination of FibroScan and other diagnostic models, the AUC values of FibroScan+ FIB-4, FibroScan+ Forns index were relatively high in ≥S2 and ≥S3, which were 0.804 and 0.907, respectively. The platelet count ((200.65±50.89)×10 9/L vs. (169.96±63.68)×10 9/L), total cholesterol ((4.69±0.77) mmol/L vs. (4.32±1.00) mmol/L), high-density lipoprotein (HDL) (1.28 (1.05, 1.46) mmol/L vs. 1.08 (0.92, 1.21) mmol/L), total protein (74.00 (70.63, 77.08) g/L vs. 68.80 (64.60, 73.55) g/L), albumin (47.06 (44.65, 48.81) g/L vs. 44.70 (41.55, 46.20) g/L), and globulin (26.80 (24.48, 29.70) g/L vs. 25.80 (23.05, 27.60) g/L) of the non-significant liver fibrosis group were higher than those of the significant liver fibrosis group, and the differences were statistically significant ( t=2.74, 2.09; Z=-3.30, -3.88, -3.95, -2.01; P=0.007, =0.040, =0.001, <0.001, <0.001, =0.044). GGT (27.50 (17.00, 41.75) U/L vs. 37.00 (22.50, 87.00) U/L), the liver stiffness measurement (LSM) in the non-significant hepatic fibrosis group was lower than the significant liver fibrosis group (6.85 (5.60, 9.26) kPa vs. 11.60 (7.08, 17.26) kPa), and the differences were statistically significant ( Z=-2.73, -4.39; P=0.006, <0.001). The result of multivariate logistic regression analysis revealed that globulin, albumin, HDL, and LSM were independent factors of liver fibrosis ( OR (95% confidence interval)=0.865 (0.759 to 0.985), 0.804 (0.691 to 0.935), 0.128 (0.023 to 0.711), and 1.251 (1.091 to 1.433), respectively; P=0.029, 0.025, 0.019, 0.001). A novel model, GLAH, was established with globulin, LSM, albumin, and HDL. The AUC for diagnosing liver fibrosis degree ≥S2, ≥S3, and S4 was 0.847, 0.938, and 0.909, respectively, which were higher than those of the above models. The positive predictive value for diagnosing liver fibrosis degree ≥S2 with GLAH>1.12 as the cutoff value was 95.8%. The negative predictive value for excluding fibrosis stage ≥S2 with GLAH<-1.41 was 92.3%. This approach could reduce the number of liver biopsies by 48.3% (57/118), with an accuracy of 94.7% (54/57). Conclusions:The clinical value of FibroScan combined with FIB-4 or Forns index is better in the diagnisis of fibrosis stage ≥S2 and ≥S3. The GLAH model has higher diagnostic value and can accurately predict the degree of liver fibrosis in CHB patients complicated with NAFLD, thus reducing the need for liver biopsy.
