Abstract
BACKGROUND: Hepatic decompensation carries profound implications for patient quality of life and risk of mortality. We lack comparative data on how noninvasive tools perform in risk stratification for those with compensated cirrhosis. We performed a systematic review to assess the performance of laboratory and transient elastography-based models for predicting hepatic decompensation in patients with compensated cirrhosis. METHODS: The following databases were searched by an informationist to identify relevant studies, including adult patients with compensated cirrhosis from inception to August 2023: Medline, Embase, Scopus, Web of Science, and ClinicalTrials.gov. Title and abstract screening followed by full-text review were performed by 2 independent reviewers, and data abstraction was completed using standardized forms. Studies of patients with decompensation at baseline (defined by ascites, variceal bleeding, and HE) or any primary hepatic malignancy were excluded. The primary outcome was hepatic decompensation, as defined above. Pooled HRs were calculated using the common-effect inverse-variance model. RESULTS: Forty-four full-text studies met the inclusion criteria. Across 52,589 patients, the cumulative incidence of any decompensation was 17.9% over a follow-up time of 111,401 patient years. Pooled risk estimates for all-cause decompensation demonstrated that MELD (HR: 1.08; 95% CI: 1.06-1.10), albumin-bilirubin (HR: 2.13, 95% CI: 1.92-2.36), fibrosis-4 (HR: 1.04, 95% CI: 1.03-1.06), albumin-bilirubin-fibrosis-4 (HR: 1.25, 95% CI: 1.18-1.33), and liver stiffness by transient elastography (HR: 1.04; 95% CI: 1.04-1.05) predict decompensation. CONCLUSIONS: Available blood and imaging-based biomarkers can risk-stratify patients for hepatic decompensation. Changes in albumin-bilirubin appear to have the highest discrimination in predicting decompensation events.