Abstract
Although the Systemic Immune-Inflammation Index (SII) has established prognostic value in several diseases, its role in hepatitis B virus (HBV) infection remains undefined. HBV-related inflammation drives progression to cirrhosis, hepatic failure, and hepatocellular carcinoma. However, the association between SII and all-cause mortality in this population has not been well established. Therefore, our study aims to investigate the association of SII with all-cause mortality in patients with HBV infection. We retrospectively analyzed a cohort of US adults (≥18 years) with HBV infection from the 1999 to 2018 National Health and Nutrition Examination Survey, focusing specifically on HBeAg-positive individuals. All-cause mortality (ascertained through the National Death Index as of December 2018) was evaluated in relation to SII using multivariable Cox regression with restricted cubic splines. Prognostic accuracy was evaluated using time-dependent receiver operating characteristic analysis, complemented by Kaplan-Meier survival estimates. Subgroup and sensitivity analyses verified the robustness of the results against potential confounders. The cohort comprised 3332 HBV-infected individuals followed for a mean 105.83 ± 65.21 months, during which 682 all-cause mortality events were recorded. In the fully adjusted Cox model, a higher Log-Transformed Systemic Immune-Inflammation Index (lnSII) level remained independently associated with a significant 53% increase in all-cause mortality (hazard ratio = 1.53, 95% confidence interval: 1.13-2.08). Restricted cubic spline regression confirmed a dose-dependent positive association between lnSII and mortality (P < .001). Time-dependent receiver operating characteristic analysis demonstrated sustained high discriminatory accuracy across all intervals (area under the curve: 0.895 at 1-year, 0.874 at 3-year, 0.870 at 5-year, 0.868 at 10-year). Subgroup analyses revealed pronounced risk amplification among Hispanic individuals (hazard ratio = 2.90, 95% confidence interval: 1.14-7.38), while a U-shaped mortality relationship manifested in the low hepatic fibrosis stratum (Fibrosis-4 index < 1.30; P = .006), with optimal survival at lnSII = 6. Robustness was confirmed through extensive sensitivity analyses. Elevated SII is an independent predictor of all-cause mortality in HBV-infected patients, with the greatest risk in the top quartile. Its consistent accuracy for both short- and long-term outcomes calls for integrating this marker into clinical practice to improve risk-stratified management.