Comparison of MELD, HVPG, and their changes to predict clinically relevant endpoints in cirrhosis.

AIM: Identification of predictors in the natural history of cirrhosis is based on determinations at a fixed time point. However, changes of these predictors may offer more information. To evaluate the predictive value of Model for End Stage Liver Disease (MELD) and hepatic venous pressure gradient (HVPG) and their changes in cirrhosis.

METHODS

Patients with repeat HVPG measurements between January 2000 and December 2008 were considered for inclusion. Patients were followed until decompensation/death or July 2009. Multivariate Cox regression was used to analyze the predictive value of a single measurement of MELD and HVPG, and changes between measurements. Compensated and decompensated patients were analyzed separately.

RESULTS

One hundred and seventeen patients were included (51 compensated, 66 decompensated). Median time between measurements and follow-up was 13 (2-24) and 11 (6-38) months in compensated and 8 (1-16) and 10 (3-21) months in decompensated patients, respectively. Fifteen compensated patients developed decompensation while twelve decompensated patients died. On multivariate analysis, MELD (HR 1.12 (95% CI 1-1.24)) and HVPG (HR 1.16 (95% CI 1.04-1.29)) were independent predictors of decompensation in compensated, while MELD (HR 1.18 (95% CI 1.09-1.27)) was the only predictor of death in the decompensated.

CONCLUSION

Single and repeat measurements of MELD and HVPG are associated to outcomes in cirrhosis. Use of repeat measurements does not seem to add further information.