Portal vein thrombosis (PVT) typically presents with abdominal pain, ascites, and splenomegaly, but it is frequently asymptomatic. Current guidelines recommend that anticoagulation be used for symptomatic patients with PVT. However, it remains controversial regarding the benefits of treatments for asymptomatic patients with PVT.
We retrospectively enrolled 933 patients with suspicious PVT in the Taipei Veterans General Hospital from January 2002 to December 2011. At total of 93 patients were confirmed using either doppler sonography, computed tomography, or magnetic resonance imaging. Response to treatment was defined as the recanalization or cavernous formation of portal veins. Logistic regression was used to investigate the clinic-laboratory parameters that were predictive for the resolution of PVT.
Among the 93 patients, the median age was 63 years old (range 1–91), and 61 (66%) of the patients were male. Abdominal pain was the most common symptom, occurring in the 53 patients (57%). Twenty-nine (31%) patients were asymptomatic at the time of diagnosis. For 35 of the treated patients, anticoagulation (71%), anti-platelet agents (20%), and catheter-directed urokinase infusion (26%) were the common modalities. Bleeding was noted in 6 (17%) of the patients, and all of the bleeding incidents were in the gastrointestinal tract. In the multivariate analysis, treatment was the only independent factor for thrombus improvement [odds ratio 8.54, 95% confidence interval 2.61–28.0, P < 0.001]. The results were the same when we analyzed the symptomatic and asymptomatic subgroups. The cumulative probability of improvement was higher among the treated patients compared to untreated patients at 2 years (62.8% vs. 26.2%, P< 0.001), and the benefits of the treatment were evident among the symptomatic and asymptomatic patients.
Asymptomatic PVT patients may benefit from treatment. Further large-scale or prospective studies are necessary.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.