CTPV with EHPVO in the cirrhotic patient was a relatively rare disease but could result in serious complications, such as esophagogastric variceal bleeding, recurrent ascites and hepatic failure. There was no consensus on the treatment of PVT with EHPVO. Anticoagulation treatment was recognized as the basic treatment for PVT in non-tumor patients [1, 9]. However, CTPV with EHPVO was always excluded from the studies which explored the effect of anticoagulation treatment for PVT. As far as we know, our study is firstly to explore the value of anticoagulation for CTPV cirrhotic patients with EHPVO. Our study showed that anticoagulation treatment could significantly prevent the progression of PVT and occurrence of hepatic decompensation, and improve prognosis, although had little effect on recanalization of main portal vein thrombosis.

CTPV was always secondary to long-term portal vein thrombosis and portal vein occlusion, and was considered to be an irreversible condition. One randomized controlled trial [13] showed that nadroparin-warfarin sequential anticoagulation could achieve a 62.5% recanalization rate (complete or partial) in cirrhotic patients with portal vein thrombosis. One systematic review and meta-analysis [12] performed by Loffredo demonstrated that patients treated with anticoagulation treatment had a higher recanalization rate compared with patients who did not accept anticoagulants (71% vs 42%, P < 0.0001). Regrettably, little recanalization rate was shown in CTPV cirrhotic patients with EHPVO who received anticoagulation therapy. Only several cases [7, 11] were reported, which showed that CTPV was recanalized after long-term anticoagulation treatment. Our results indicated that anticoagulants had little effect on portal thrombosis recanalization, but could prevent progression and maintain stability. We speculate that the following reasons could explain it. CTPV with EHPVO always had a long time course of the disease and was characterized by fibrous thrombosis and disruption of blood flow, which was difficult to be reversed by anticoagulant, especially with spleen and mesenteric venous thrombosis. Warfarin prevented thrombosis progression and maintained blood flow in the collateral circulation around the portal vein by adjusting the balance of hemostasis and anticoagulation.

Theoretically, recanalization of the portal vein achieved the restoration of blood flow to the liver and relieved portal hypertension, which avoided hepatic decompensation. A randomized controlled trial conducted by Zhou et [13] demonstrated that the Child-Pugh score and albumin level of cirrhotic patients with portal vein thrombosis were increased after six months of warfarin treatment. Our results indicated warfarin could prevent hepatic decompensation for CTPV cirrhotic patients with EHPVO, although had little effect on portal vein thrombosis recanalization. The unobstructed blood flow of cavernous transformation around the main portal vein was maintained by warfarin and prevented portal hypertension and ensured adequate hepatic blood perfusion, which decreased the incidence of hepatic decompensation. The Kaplan-Meier curve indicated patients in the anticoagulation group had a better prognosis. The primary cause of this death was correlated with hepatic decompensation, indicating a decrease in portal pressure and adequate blood flow to the liver could prevent hepatic decompensation and further increase the prognosis for CTPV cirrhotic patients with EHPVO. Our cox regression model showed that anticoagulant was the only predictor for hepatic decompensation and survival, with no other hepatic indicators.

Warfarin was chosen as an anticoagulant due to its low price and ease of administration. However, clinicians may be hesitant to use warfarin because CTPV cirrhotic patients with EHPVO always have esophagogastric varices. Several studies [12, 14, 15] have demonstrated that anticoagualtion treatment did not increase the bleeding risk of portal hypertension in patients with PVT. Our results demonstrated that the variceal bleeding rate was lower in the anticoagulation group than in the control group. Another disadvantage of warfarin treatment was the continuous adjustment of dosage by monitoring INR value, which reduced patient adherence. In clinical practice, the INR value of most patients will be stable at targeted value after a short time adjustment, and then patients need to monitor the INR value every month. There were no serious complications due to warfarin treatment. After a reduced dosage of warfarin, the most adverse event will diminish, such as gingival hemorrhage, urticaria, transit nausea and transit diarrhea.

As far as we know, this is the first study to evaluate the efficacy of warfarin for CTPV cirrhotic patients with EHPVO. However, there were several limitations in our research. Firstly, our study was one single-center retrospective study, with inevitable selection bias. Randomized controlled trials about this need to be further explored. Secondly, the assessment of changes in liver function after warfarin treatment was deficient, because our patients were always followed up at the local hospital instead of our center unless developed hepatic decompensation. Besides, the evaluation of esophagogastric varices by endoscopy was lacking.