Article Text
PDF
Abstract
Background Liver stiffness measurement (LSM) holds promise in monitoring disease progression or regression. We aim to assess the prognostic significance of the dynamic changes in LSM over time concerning liver-related events (LREs) and death in patients with chronic hepatitis B (CHB) and compensated advanced chronic liver disease (cACLD).
Methods This retrospective study included 1272 patients with CHB and cACLD who underwent at least twice measurement, including LSM and fibrosis score based on four factors (FIB-4) (IDDF2024-ABS-0174 Figure 1). ΔLSM was calculated as [(follow-up LSM-baseline LSM)/baseline LSM×100], expressed as a percentage. All the participants received antiviral treatment for at least 6 months. We recorded LREs and all-cause mortality during a median follow-up time of 46 months. Hazard ratios (HR) and confidence interval (CI) for outcomes were calculated using Cox regression. Individuals were stratified into different risk groups based on baseline LSM and Δ LSM (low risk, baseline LSM ≤ 15kpa + Δ LSM < -20%; intermediate risk, baseline LSM > 15kpa + Δ LSM < -20%; high risk, baseline LSM ≤ 15kpa + Δ LSM ≥ -20%; extremely high risk, baseline LSM > 15kpa + Δ LSM ≥ -20%).
Results Baseline FIB-4, baseline LSM, ΔFIB-4, and ΔLSM were independently and simultaneously associated with LREs (adjusted HR, 1.04, 95% CI, 1.00-1.07; 1.02, 95% CI, 1.01-1.03; 1.06, 95% CI, 1.03-1.09; 2.05, 95% CI, 1.71-2.46, respectively). Baseline LSM combined with ΔLSM achieved the highest Harrell’s C (0.751), integrated AUC (0.776), and time-dependent AUC (0.737) for LREs. By using baseline LSM and ΔLSM, we proposed a risk stratification method to improve clinical applications. The risk stratification proposed based on LSM performed well in terms of prognosis: low risk (n=390; reference), intermediate risk (n=446; HR=3.38), high risk (n=272; HR=5.64), extremely high risk (n=164; HR=11.11) (IDDF2024-ABS-0174 Figure 1).
Conclusions Baseline and repeated non-invasive test measurements allow individual risk stratification of patients with CHB and cACLD. Combining baseline and dynamic changes in LSM improves prognostic prediction.