Abstract / Summary
The optimal anticoagulant strategy for portal vein thrombosis (PVT) in patients with cirrhosis remains uncertain. This systematic review and meta-analysis compared the clinical, radiological, and safety outcomes associated with direct oral anticoagulants (DOACs) versus vitamin K antagonists (VKAs) in cirrhotic PVT. Multiple databases were searched from inception to 31 August 2026 for studies comparing DOACs with VKAs in cirrhotic PVT. Pooled risk ratios (RRs) with 95% confidence intervals (CIs) were calculated using random-effects models with restricted maximum likelihood estimation and the Hartung–Knapp adjustment where applicable. The certainty of evidence was assessed using the GRADE approach. Six retrospective cohort studies including 3,345 patients (1,651 DOAC users and 1,694 VKA users) were included. DOAC use was associated with higher rates of any recanalization (2 studies; RR = 2.85, 95% CI: 1.81–4.49) and lower observed risks of liver transplantation (2 studies; RR = 0.44, 95% CI: 0.33–0.59) and intracranial hemorrhage (2 studies; RR = 0.52, 95% CI: 0.32–0.85). Associations with all-cause mortality (3 studies; RR = 0.81, 95% CI: 0.64–1.02) and PVT progression (3 studies; RR = 0.19, 95% CI: 0.01–2.45) were not significant, and excluding the largest cohort shifted the mortality estimate to RR = 1.04 (95% CI: 0.63–1.72). No clear difference was observed for major bleeding (RR = 0.96, 95% CI: 0.28–3.32) or gastrointestinal bleeding (RR = 1.09, 95% CI: 0.79–1.50), while estimates for complete recanalization and esophageal variceal bleeding were imprecise. The certainty of evidence was very low for all outcomes. In cirrhotic PVT, DOAC use was associated with higher rates of any recanalization than VKAs, without a clear increase in bleeding. Associations with lower mortality and less PVT progression were not significant, and those with transplantation and intracranial hemorrhage relied largely on a single cohort. Given the very low certainty of evidence and potential residual confounding, these associations cannot establish causality, and prospective randomized studies are needed. CRD420261469885.