Portosystemic shunting is the key driver of portal/splanchnic vein thrombosis development

Introduction The pathophysiological mechanisms underlying portal/splanchnic vein thrombosis (PVT/SVT) development in patients with advanced chronic liver disease (ACLD)/cirrhosis are poorly understood. A recent prospective study found portal hypertension as a key predictor of PVT.

Material and Methods We conducted a combined pro-(VICIS cohort, n=400) and retrospective (n=741) study including patients with ACLD undergoing hepatic venous pressure gradient(HVPG)-measurement at the Vienna Hepatic Haemodynamic Lab. Patients with PVT/SVT and anticoagulation at baseline were excluded. First, baseline characteristics of patients from the prospective cohort with vs. without development of PVT/SVT during follow-up were compared. Second, inverse probability of treatment weighting-(IPTW) for HVPG was applied to validate our findings in the retrospective cohort. Third, we calculated a competing risk regression analysis with PVT/SVT development as outcome of interest as outcome of interest and hepatocellular carcinoma (HCC) development/liver transplantation/death as competing events.

Results In the prospective cohort, 23 patients(5.8%) developed PVT/SVT during a median follow-up of 28.4 (95%CI:26.7-30.5) months. Despite having similar HVPG (18[IQR:15-21]vs.17[IQR:12-21]mmHg,p=0.257) and numerically lower MELD (14[IQR:10-17]vs.11[IQR:9-16]points,p=0.098), patients developing PVT/SVT more commonly had prior variceal bleeding (8[34.8%)]vs.40[10.6%],p=0.002) and varices (21[91.3%]vs.232[61.5%],p<0.001) as well as higher ammonia levels (50.1[IQR:37.9-62.1]vs.38.0[IQR:28.6-50.2]µmol/L,p=0.005). In the retrospective cohort, IPTW-matching for HVPG confirmed that PVT/SVT development (n=49, median follow-up of 78.1[95%CI:74.4-83.1]months) was associated with varices (43[88.6%]vs.502[72.5%],p=0.024) and higher ammonia levels (62.3±29.8vs.47.3±27.5µmol/L,p=0.021), despite less severe hepatic dysfunction based on MELD. Neither advanced laboratory tests of coagulation or inflammation (e.g., IL-6 and procalcitonin), nor biomarkers of (other) pathomechanisms of ACLD (endothelial, circulatory, or hepatic dysfunction) were associated with PVT/SVT development.

Conclusion At the same severity of portal hypertension (HVPG), patients who develop PVT/SVT during follow-up had a higher prevalence of (large) varices and higher ammonia levels despite less pronounced hepatic dysfunction, suggesting a pathophysiological role of portal blood flow diversion by portosystemic collaterals. Our findings support previous observations on the association of low portal blood flow velocity and PVT/SVT development.

Publication History

Article published online:
13 May 2025