PURPOSE: The present study aimed to investigate the role of hepatic venous pressure gradient (HVPG) for prediction of long-term mortality in patients with decompensated cirrhosis. MATERIALS AND METHODS: Clinical data from 97 non-critically-ill cirrhotic patients with HVPG measurements were retrospectively and consecutively collected between 2009 and 2012. Patients were classified according to clinical stages and presence of ascites. The prognostic accuracy of HVPG for death, survival curves, and hazard ratios were analyzed. RESULTS: During a median follow-up of 24 (interquartile range, 13-36) months, 22 patients (22.7%) died. The area under the receiver operating characteristics curves of HVPG for predicting 1-year, 2-year, and overall mortality were 0.801, 0.737, and 0.687, respectively (all p<0.01). The best cut-off value of HVPG for predicting long-term overall mortality in all patients was 17 mm Hg. The mortality rates at 1 and 2 years were 8.9% and 19.2%, respectively: 1.9% and 11.9% with HVPG < or =17 mm Hg and 16.2% and 29.4% with HVPG >17 mm Hg, respectively (p=0.015). In the ascites group, the mortality rates at 1 and 2 years were 3.9% and 17.6% with HVPG < or =17 mm Hg and 17.5% and 35.2% with HVPG >17 mm Hg, respectively (p=0.044). Regarding the risk factors for mortality, both HVPG and model for end-stage liver disease were positively related with long-term mortality in all patients. Particularly, for the patients with ascites, both prothrombin time and HVPG were independent risk factors for predicting poor outcomes. CONCLUSION: HVPG is useful for predicting the long-term mortality in patients with decompensated cirrhosis, especially in the presence of ascites.
Adult ; Aged ; Ascites/mortality ; Female ; Hepatic Veins/*physiopathology ; Humans ; Kaplan-Meier Estimate ; Liver Cirrhosis/blood/complications/diagnosis/*mortality/*physiopathology ; Liver Failure/diagnosis/*mortality/physiopathology ; Male ; Middle Aged ; Predictive Value of Tests ; Prognosis ; Proportional Hazards Models ; ROC Curve ; Retrospective Studies ; Risk Factors ; Severity of Illness Index ; Venous Pressure

BACKGROUND/AIMS: This study compared the prognostic values of the Model for End-stage Liver Disease (MELD) and the hepatic venous pressure gradient (HVPG) in the prediction of death within 3 and 12 months in patients with decompensated liver cirrhosis. METHODS: We used data from 136 consecutive patients with decompensated cirrhosis who underwent HVPG between January 2006 and June 2008. Cox regression analysis was used to investigate the independent relationships with death of MELD and HVPG. The prognostic accuracies of MELD and HVPG were analyzed by calculating the area under the receiver operating characteristic curve (AUROC) for the occurrence of death within 3 and 12 months. RESULTS: Both MELD and HVPG were independent predictors of death [hazard ratio (HR)=1.11 and 1.12, respectively; 95% confidence interval (CI)=1.04~1.20 and 1.08-1.16]. Analysis of the AUROC demonstrated that the prognostic power did not differ between MELD and HVPG for predicting the 3-month survival (HR=0.76 and 0.68, respectively; 95% CI=0.62~0.89 and 0.52~0.84; P=0.22) or the 12-month survival (HR=0.72 and 0.73, 95% CI=0.61~0.83 and CI=0.61~0.84). CONCLUSIONS: Both MELD and HVPG are independent prognostic factors of death within 3 and 12 months in patients with decompensated liver cirrhosis, and their accuracies are similar. However, HVPG has a limited role in the prediction of death in decompensated cirrhosis due to its invasiveness and limited use.
Adult ; Aged ; Area Under Curve ; Cohort Studies ; Female ; Hepatic Veins/*physiopathology ; Humans ; Liver Cirrhosis/diagnosis/*mortality/physiopathology ; Liver Failure/diagnosis/*mortality/physiopathology ; Male ; Middle Aged ; Models, Biological ; Predictive Value of Tests ; Prognosis ; Proportional Hazards Models ; Regression Analysis ; Retrospective Studies ; Severity of Illness Index ; Survival Analysis ; Venous Pressure