Background/aims: In acute variceal bleeding (AVB) hepatic venous pressure gradient (HVPG) is associated with prognosis. However, this has not been studied in patients receiving the currently recommended therapy. We evaluate here the performance of early HVPG measurement as a predictor of treatment failure in patients with acute variceal bleeding managed with the current standard treatment and whether clinical variables might be of similar predictive accuracy.
Methods: We included 117 patients with AVB in whom HVPG was measured within 48 h of admission. The main endpoint was 5-day failure, a composite of uncontrolled bleeding, early rebleeding or death within 5 days.
Results: Eighteen patients (15%) had 5-day failure. Multivariate analysis identified three variables independently associated with 5-day failure: HVPG 20, systolic blood pressure at admission <100 mmHg and non-alcoholic cause of cirrhosis. The discriminative capacity of this model was good (c statistic: 0.79). When only clinical variables were included in the analysis, Child-Pugh class, systolic blood pressure at admission and etiology were the independent predictors. This model had also a good discriminative ability (c statistic: 0.80).
Conclusions: HVPG independently predicts short-term prognosis in patients with acute variceal bleeding treated with pharmacologic and endoscopic therapy, but similar predictive accuracy can be achieved using only simple clinical variables that have universal applicability.