Abstract: Objective To explore the difference in routine test indicators and specific cytokine profiles between hepatitis B-related cirrhosis (HBC) and alcoholic cirrhosis (AC) and provide a basis for differential diagnosis and optimized treatment. Methods We collected data from 70 HBC patients and 57 AC patients who were treated at Suzhou Fifth People's Hospital and Shulan (Hangzhou) Hospital between January 2021 and May 2024. Data on complete blood count, liver biochemical parameters, cytokine levels, and immune cell ratios were collected and analyzed using GraphPad and SPSS. Results No significant difference was found between the two groups in serum levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), albumin (ALB), alkaline phosphatase (ALP), platelet (PLT), and C-reactive protein (CRP) (P>0.05). However, total Bilirubin (TB, P=0.0083), direct bilirubin (DB, P=0.0014), gamma-glutamyl transferase (GGT, P<0.0001), and prothrombin time (PT, P=0.0050) were significantly higher in the AC group compared to the HBC group. The red blood cell count (P=0.0391) was significantly higher in the HBC group than in the AC group, while other blood routine indicators showed no significant differences between the two groups. Interleukin-2 (IL-2) and Interferon-γ (IFN-γ) levels were not significantly different between the groups, whereas IL-4 (P=0.0007), IL-17 (P<0.0001), and tumor necrosis factor-α (TNF-α, P=0.0114) were significantly higher in the HBC group, and IL-6 (P=0.0049) and IL-10 (P=0.0267) were significantly higher in the AC group. Receiver operating characteristic (ROC) curve analysis indicated that TB (0.614), DB (0.642), GGT (0.686), PT (0.631), IL-4 (0.676), and IL-6 (0.641) had low predictive diagnostic ability (0.5 < AUC < 0.7) between the two groups, while IL-17 (0.735) had moderate discriminatory ability (0.7 < AUC < 0.85). A binary logistic regression analysis was used to construct a combined predictive model, revealing that combining GGT and IL-17 yielded the optimal predictive model. The prediction formula was logit(P)=0.09+0.01*GGT-0.075*IL-17. This model significantly improved the ability to differentiate between HBC and AC, with a prediction accuracy of 83.6%, sensitivity of 84.2%, specificity of 71.4%, and a cutoff value of -0.085. When 0.01*GGT - 0.075*IL-17 ≥ -0.265, the diagnostic accuracy for AC was 83.6%. Conclusion By analyzing the differences in routine clinical indicators and cytokine expression between the HBC and AC groups, we found significant differences in liver biochemical indicators and cytokines such as IL-17. Further results indicate that the a predictive model combining GGT and IL-17 significantly improves the ability to differentiate between HBC and AC. In summary, the predictive model established using binary logistic regression analysis aids in distinguishing between HBC and AC, which provides a reference for the early diagnosis and optimized treatment of both types of cirrhosis.

Key words: Hepatitis B-related cirrhosis, Alcoholic cirrhosis, GGT, IL-17, Combined predictive model