Abstract: Objective To explore the predictive value of albumin-to-globulin ratio for assessing liver fibrosis in patients with chronic HBV infection and to determine clinical diagnostic thresholds. Methods A total of 297 patients with chronic HBV infection who underwent liver puncture at the First Affiliated Hospital of Guangxi Medical University from February 2004 to December 2024 were included. Baseline demographic characteristics (gender, age), pathological examination results, and biochemical indicators [total bilirubin(TBil), albumin(Alb), globulin(Glo), alanine aminotransferase (ALT), aspartate aminotransferase (AST), and platelet count (PLT)] were collected, and the albumin-to-globulin ratio, fibrosis 4 index (FIB-4), and aspartate aminotransferase-to-platelet ratio index (APRI) were calculated. A combined model of albumin-to-globulin ratio with FIB-4 was constructed. Based on the METAVIR score, patients were divided into a non-significant liver fibrosis group (stages S0~2) and a significant liver fibrosis group (stage S≥3). Univariate and multivariate logistic regression was used to identify the independent predictive factors that significantly associated with liver fibrosis in HBV-infected patients. The predictive value of these indicators for liver fibrosis was assessed using receiver operating characteristic (ROC) curves, and the area under the curve (AUC), optimal thresholds, sensitivity, and specificity were calculated. Mann-Whitney U tests were used to compare non-normally distributed quantitative data between the two groups, while the χ² test was used for comparisons of categorical data between groups. Results In the 297 chronic HBV infection patients, there were 79 patients with significant liver fibrosis (S3~S4) and 218 patients with non-significant liver fibrosis (S0~S2). The levels of Glo, ALT, and AST, as well as FIB-4 and APRI in patients with significant liver fibrosis were higher than those in the non-significant fibrosis group [32.2 (27.7, 35.1) g/L vs. 27.5 (25.4, 30.7) g/L, 49 (27, 85.5) U/L vs. 32 (22, 49) U/L, 44 (30, 79) U/L vs. 27 (21, 37) U/L, 1.546 (1.0205, 2.5972) vs. 0.9492 (0.6998, 1.4103), 0.6439 (0.4206, 1.2075) vs. 0.3555 (0.2451, 0.5708), Z=-5.418, -3.658, -5.332, -5.369, -5.696, respectively, P<0.05], while Alb, albumin/globulin ratio, and PLT were significantly lower in the non-significant fibrosis group [41.5 (36.9-43.8) g/L vs. 42.6 (39.8, 45.5) g/L, 1.2707 (1.1138, 1.4765) vs. 1.5295 (1.3742, 1.6873), 172 (138, 231)×10⁹/L vs. 199 (165, 239)×10⁹/L, Z=-2.758, -6.349, -2.645, P<0.05]. Multivariate logistic regression analysis indicated that the albumin-to-globulin ratio is an independent predictor of significant liver fibrosis in chronic HBV infection patients (OR=0.091, P<0.05). The optimal cutoff values, AUC, sensitivity, specificity, and accuracy of albumin-to-globulin ratio in predicting the degree of liver fibrosis in patients with chronic HBV infection are 1.3601, 0.741, 64.6%, 76.6%, and 73.4%, respectively; for APRI, they were 0.4744, 0.716, 70.9%, 67.4%, and 68.35%; for FIB-4, they were 1.5911, 0.704, 49.4%, 83.9%, and 74.75%; for the combined model, they were 0.2424, 0.767, 74.7%, 72.9%, and 73.4%. Pairwise comparisons of the AUCs of the four diagnostic models showed no statistically significant differences. The differences in sensitivities among the four diagnostic methods were statistically significant (χ²=21.321, P<0.001), with the sensitivities of the combined model and the APRI method being significantly higher than that of FIB-4 (P<0.0083), while there were no statistically significant differences in pairwise comparisons among the combined model, APRI, and the albumin-to-globulin ratio. The comparison of the specificities of the four diagnostic methods revealed statistically significant differences (χ²=29.739, P<0.001), with the specificity of FIB-4 being significantly higher than that of the other three methods (P<0.0083), and the specificity of the albumin-to-globulin ratio being significantly higher than that of APRI (P<0.0083), whereas the difference in specificity between the combined diagnosis and the albumin-to-globulin ratio was not statistically significant. Conclusion The albumin-to-globulin ratio can serve as an indicator for predicting the degree of liver fibrosis in patients with HBV infection, and the combined diagnostic model of the albumin-to-globulin ratio and FIB-4 performs best in diagnosing patients with significant fibrosis.

Key words: Hepatitis B virus, Liver fibrosis, Albumin to globulin ratio