Portal Vein Thrombosis Following Splenectomy for Hematologic Diseases: Prevalence and Risk Factors.

Portal vein thrombosis (PVT) is a potentially fatal complication of splenectomy; the prevalence of PVT remains up till now controversial, such as the related-risk factors. The hematologic patients who underwent splenectomy in Niguarda Hospital, Milan, Italy, between January 1995 and November 2004 were retrospectively reviewed to identify the prevalence of post-splenectomy PVT and risk factors associated with its development. Splenectomy was performed because of malignant (n = 41) or non-malignant (n = 44) hematologic diseases. Indications for splenectomy in the groups with and without PVT were compared with Fisher’s exact test. We tested the reliability of platelet count as a marker of PVT, analysing the receiver operator characteristic (ROC) curve. Cut-off level was chosen for best prediction in term of sensitivity and specificity.

Among 85 patients who underwent splenectomy, the associated non-malignant diseases were immune thrombocytopenic purpura (n = 40), hemolytic anemia (n = 2), Evan’s syndrome (n = 1), cryoglobulinemia (n = 1); malignant diseases were lymphoproliferative (n = 39) and myeloproliferative disorders (n = 2). Four cases of PVT (4,71%) were diagnosed. One of 2 patients with a myeloproliferative disorder had PVT. PVT was also present in 3/39 (7.69%) patients with lymphoproliferative malignancies, all with splenic marginal zone B-cell variety non-Hodgkin’s lymphoma (3/16) . The median splenic weight in patients who developed PVT was 1800 g (range 1500–1980 ) and the median post-surgery platelet count was 661 x 10⁹/liter (range 211–1250). No patient with non-malignant diseases had portal vein thrombosis. The incidence of the event was of 0.0000523 new cases for day-patient, with a 0.005% new incidental case probability. Splenic weight, with a cut-off value of 500 grams for the risk analysis (p=0.029), was significantly correlated to thrombosis. PVT was also significantly correlated to the entity of the post-surgery thrombocytosis (p=0.024), setting the threshold value of thrombosis risk at 850.000/uL (Sensibility 0.50 - Specificity 0.95).

Patients with myeloproliferative and lymphoproliferative disorders seem at higher risk for the development of PVT than those with non-malignant diseases. A splenic weight over 500 grams and a post-surgery thrombocytosis over 850.000/uL are significantly associated with the occurrence of PVT. The presence of these two risk factors should require the prophylactic use of adequate doses of low molecolar weight heparin.