Аннотация
Introduction: Compensated cirrhosis carries a significant risk of progression to decompensation, which substantially worsens prognosis. Accurate prediction of decompensation events is critical for guiding surveillance, optimizing intervention timing, and improving patient outcomes. Although many prognostic biomarkers have been studied, findings remain heterogeneous. This umbrella review synthesizes evidence from systematic reviews and meta-analyses to identify and appraise biomarkers predicting decompensation in alcoholic and nonalcoholic compensated cirrhosis. Methods: PubMed, Scopus, and Web of Science were searched to August 15, 2025, for English-language systematic reviews and meta-analyses. From the included meta-analyses, prognostic performance, heterogeneity, and publication bias were recorded. Results: Four systematic reviews were included, two with meta-analyses. Strong predictors across reviews were serum albumin, INR, bilirubin, platelet count, and liver stiffness measurement. HVPG remained a robust invasive predictor, while emerging biomarkers-interleukin-6, keratin-18, and extracellular vesicles-were associated with an increased risk of decompensation, although the certainty of evidence was limited by heterogeneity and methodological constraints. Composite scores enhanced predictive accuracy. AMSTAR 2 ratings ranged from high to low, with common reporting limitations. Overlap analysis indicated moderate redundancy among primary studies. Conclusions: Both established and emerging biomarkers predict decompensation in compensated cirrhosis. Integrated multidomain models combining clinical, biochemical, and imaging-derived measures may provide the greatest predictive value for guiding clinical decision-making.