First decompensation outweighs portal vein thrombosis in cACLD: Molecular profiling and derivation of the VAIL score.

BACKGROUND &

aimsThe temporal relationship between hepatic decompensation and portal vein thrombosis (PVT) in compensated advanced chronic liver disease (cACLD) remains unclear. We investigated their temporal sequence, prognostic significance, and molecular signatures and developed a biologically grounded predictive model for decompensation.

methodsIn this prospective study, 367 rigorously defined patients with cACLD without previous decompensation or PVT were followed for up to 11 years. Baseline assessment included hepatic venous pressure gradient, liver stiffness, upper endoscopy, Doppler ultrasound, and serum measurements of 35 biomarkers spanning angiogenesis, inflammation, endothelial injury, hypoxia, and gut permeability. Temporal associations were analysed using Kaplan-Meier and time-dependent Cox models. Biomarkers independently associated with decompensation were integrated to derive the VAIL (VEGF-D, Angiopoietin-1, I-FABP, IL-6) score, which was compared with established prognostic scores.

resultsDuring follow-up, 102/367 (27.7%) patients developed hepatic decompensation, and 22/367 (6.0%) developed non-malignant PVT. Among patients experiencing both events, decompensation preceded PVT in 53.3%, whereas PVT preceded decompensation in 33.3%. Time-dependent Cox analysis showed that hepatic decompensation was the strongest determinant of mortality (hazard ratio [HR] 11.86, 95% CI 5.68-24.78; p <0.001), whereas PVT was not independently associated with mortality (HR 0.56, 95% CI 0.12-2.46; p = 0.443). Biomarker profiling showed a restricted signature for isolated PVT (Tie2 and mesothelin) but broad multisystem dysregulation preceding decompensation. The VAIL score showed the most robust predictive performance (optimism-corrected coefficient 1.76, 95% CI 1.48-4.36; p <0.001) after bootstrap validation, outperforming all established prognostic scores.

conclusionsIn cACLD, hepatic decompensation is an early, biologically active, and prognostically dominant event, whereas PVT is a late, biologically restricted complication. The VAIL score provides superior mechanism-based risk stratification for predicting clinical deterioration. IMPACT AND IMPLICATIONS: This study provides prospective evidence that hepatic decompensation represents an early, biologically active, and prognostically dominant transition in compensated advanced chronic liver disease, whereas portal vein thrombosis arises later and reflects a distinct process with limited independent prognostic impact. These findings are important for clinicians and researchers as they clarify the temporal and biological hierarchy of key disease events, helping refocus risk assessment on pathways driving decompensation rather than on downstream vascular complications. The identification of reproducible biomarker signatures and the derivation of the VAIL score suggest that biomarker-informed risk stratification could support earlier identification of patients at high risk of decompensation and guide preventive strategies in clinical practice. Although confirmation in multicentre cohorts is required, this work provides a biologically grounded framework for future studies to test whether targeted, biomarker-guided interventions can modify the disease trajectory in advanced chronic liver disease. CLINICAL TRIALS NUMBER: ClinicalTrials.gov (NCT03083002).