Non-malignant portal vein thrombosis in liver cirrhosis – predictors of risk and the use of anticoagulation

Introduction:

The occurrence of a portal vein thrombosis (PVT) in patients with liver cirrhosis is a controversial discussed topic. The effect on the natural course of the underlying liver disease and the possible role as a negative predictor has not yet been fully clarified.

Aim of study:

This study aimed to determine possible predictors for higher mortality in patients with liver cirrhosis and portal vein thrombosis and to investigate the impact of established therapies.

Methods:

Approximately 30.000 cross-sectional imaging CT- and MRI-scans in a single transplant center (University hospital Mainz) were retrospectively screened for portal vein thrombosis with underlying liver cirrhosis. Cases with any malignancy were excluded from further analysis. Mortality was correlated with clinical and laboratory findings at the time of PVT diagnosis and throughout the end of observation period.

Results:

55 patients with liver cirrhosis and PVT were included into analysis. The median follow-up time was 129 weeks and median age at diagnosis of PVT 57.6 years. Median MELD at PVT diagnosis was 13. During follow-up 35 patients died (63.6%) after 68 weeks in median. Using univariable cox proportional-hazard regression models, MELD (HR 1.07, p = 0.009), lower serum albumin levels (HR 0.885, p = 0.001), ascites (HR 2.22, p = 0.031), spontaneous bacterial peritonitis (SBP) during follow-up (HR 2.25, p = 0.020), overt hepatic encephalopathy at diagnosis of PVT or during follow-up (HR 3.11, p = 0.005), and hepatorenal syndrome at diagnosis of PVT or during follow-up (HR 3.43, p = 0.001) were identified as predictors for higher mortality. Therapeutic anticoagulation tended to improve prognosis (HR 0.5, p = 0.055). In a multivariable cox-regression model, SBP during follow-up (HR 7.75, p = 0.0001) and lower serum albumin levels (HR 0.78, p = 0.0001) were shown to be independently associated with a lethal course. Additionally, after adjusting for the aforementioned variables, therapeutic anticoagulation was an independent predictor for better prognosis (HR 0.128, p = 0.0001).

Conclusion:

Spontaneous bacterial peritonitis and lower albumin levels are predictors for high mortality in patients with liver cirrhosis and PVT. Additionally, anticoagulation was an independent predictor for better prognosis in these patients. Consequently, future studies should evaluate the use of anticoagulation in every patient with liver cirrhosis and PVT.