To investigate the changes and prognostic values of serum complement C3 levels in patients with hepatitis B virus-related acute-on-chronic liver failure (HBV-ACLF).

Clinical data of 138 patients with HBV-ACLF admitted to the Third Affiliated Hospital of Sun Yat-sen University from May 2019 to December 2020 were retrospectively analyzed. The medical ethical committee approval was received. Among them, 115 patients were male and 23 female, aged from 19 to 75 years, with a median age of 46 years. White blood cell (WBC), neutrophil-lymphocyte ratio (NLR), total bilirubin (TB), serum creatinine (Scr), Na, complement C3, prothrombin time activity (PTA) and model for end-stage liver disease (MELD) score were detected. All patients were divided into the survival and death groups according to the survival status during hospital stay and within 3 months after discharge. The changes of multiple indexes between two groups were compared by using t test or rank-sum test. The influencing factors of clinical prognosis were identified by multivariate Logistic regression analysis. Prediction model was established by Logistic regression. The diagnostic efficiency was assessed by the receiver operating characteristic (ROC) curve.

The mortality rate of 138 patients was 19.6% (27/138), including 25 male and 2 female. The average age in the survival group was (45±12) years, and (50±12) years in the death group. The levels of WBC, NLR, TB, PTA, complement C3 and MELD score in the survival group were 7(4)×10⁹/L, 3.3(3.1), 308(156) μmol/L, 31%(12%), 0.40(0.18)g/L, 27±4, and 8(3)×10⁹/L, 5.3(7.6), 421(272) μmol/L, 23%(12%), 0.36(0.23) g/L, 31±4 in the death group, and significant differences were observed (Z=-2.311, -2.904, -3.438, -3.075, -1.742; t=-5.167; P<0.05). Multivariate analysis showed that age, TB, PTA and complement C3 were the independent prognostic factors of HBV-ACLF patients (OR=1.054, 1.009, 0.940, 0.017; P<0.05). The prediction model was established according to the equation: Logit(P)=-3.913-0.053×Age-0.009×TB+0.062×PTA+4.053×C3. ROC curve analysis indicated that the area under ROC curve (AUC) of complement C3-based model was 0.833, and the AUC of MELD score was 0.795, and no significant difference was observed (Z=0.682, P>0.05).

The level of complement C3 is an independent prognostic factor for patients with HBV-ACLF. The complement C3-based prediction model plays a critical role in evaluating clinical prognosis and guiding clinical treatment for HBV-ACLF patients.