A prospective, multicentre study in acute noncirrhotic, nonmalignant portal vein thrombosis: comparison of medical and interventional treatment

Background To evaluate medical versus interventional treatment (transjugular thrombus fragmentation, local thrombolysis with or without stent implantation) in patients with acute noncirrhotic, nonmalignant portal vein thrombosis (PVT). Methods This prospective, observational study enrolled 65 patients with acute (<28 days since begin of symptoms, no cavernoma) PVT in nine centres. Thirty patients received medical treatment and 35 patients received interventional treatment. PVT was graded into grade 1: short thrombosis and incomplete occlusion of the vessel lumen and grade 2: extended thrombosis or complete occlusion. Treatment response was classified as partial or complete, if thrombosis was reduced by one grade or to Partial and complete response rates were 7% and 30% in the medical compared to 17% and 54% (P < 0.001) in the interventional treatment group. In the multivariate analysis, interventional treatment showed a strong positive (OR 4.32, P < 0.016) and a myeloproliferative aetiology a negative (OR 0.09, P = 0.006) prediction of complete response. Complications were rare in the medical group and consisted of septicaemia and upper gastrointestinal bleeding of unknown origin in one patient each. Interventional treatment was accompanied by mild and self-limiting bleeding complications in nine patients, moderate intra-abdominal bleeding requiring transfusions (2 units) in one patient and peritoneal bleeding requiring surgical rescue in one patient. Four patients in each group developed intestinal gangrene requiring surgery. One patient died 52 days after unsuccessful interventional treatment. Conclusions Compared to medical treatment alone, interventional treatment doubled response rates at the cost of increased bleeding complications.