American Society of Hematology

Skating on Thin Ice: Blood Thinners in Myeloproliferative Neoplasms

Margaret V. Ragni, MD, MPH

Published on: March 01, 2012

Dr. Ragni indicated no relevant conflicts of interest.

Hoekstra J, Bresser EL, Smalberg JH, et al. Long-term follow-up of patients with portal vein thrombosis and myeloproliferative neoplasms. J Thromb Haemost. 2011;9:2208-2214.

Portal vein thrombosis (PVT) complicates up to 30 percent of myeloproliferative neoplasms (MPNs), a group of disorders that includes polycythemia vera, essential thrombocythemia, and primary idiopathic myelofibrosis.1 PVT may be the presenting sign of an MPN, yet our understanding of the natural history of PVT/MPN is limited by the paucity of published longitudinal studies. Now, in a retrospective analysis covering 28 years, Hoekstra and colleagues from the Netherlands have reported on the long-term follow-up of MPNs complicated by PVT.

Of 44 patients with MPN and PVT (median age at diagnosis 48 years old) identified by computerized hospital records covering the period between 1980 and 2008, 70 percent presented with PVT prior to diagnosis of MPN (for this subgroup, the median time between presentation with PVT and diagnosis of MPN was seven months). Splenomegaly was present in 75 percent of the cohort with abdominal pain reported in 70 percent and ascites documented in 34 percent. PVT was confirmed by radiographic or laparoscopic studies with thrombosis confined to the portal vein in approximately half. Of the 29 patients in whom testing was feasible, JAK2V617F was detected in 90 percent, a value that is 2.6-fold higher than the 34 percent prevalence of this mutation in an all-inclusive study of patients with PVT.1

Twenty patients were found to have additional risk factors for thrombosis, including oral contraceptives or pregnancy in 37 percent, an underlying prothrombotic condition in 24 percent, and local prothrombotic factors in 16 percent, including abdominal surgery and intra-abdominal infection. At a median of 2.3 years from PVT diagnosis, 17 patients (39%) experienced nonfatal gastrointestinal bleeding, of which two were anticoagulated. Of the 15 with bleeding in the absence of anticoagulation, eight (47%) had variceal bleeding at PVT presentation. In contrast, at a median 7.5 years from PVT diagnosis, 12 (27%) developed thrombosis recurrence, fatal in one-fourth (3/44, 18% overall), one of whom was receiving anticoagulation. Anticoagulation with oral vitamin K antagonists (VKA), low-molecular-weight heparin, or unfractionated heparin was initiated after a diagnosis of PVT in 52 percent and was continued long-term in 34 percent (VKA in all cases). Antiplatelet drugs were given in 36 percent of the cohort (27% after diagnosis and 9% before diagnosis of PVT). The four patients who were treated before diagnosis of PVT were prescribed the antiplatelet agent at the time of diagnosis of the MPN.

Whether long-term anticoagulation improves survival in patients with MPN-associated PVT is unknown, and because MPN/PVT is uncommon, organizing a sufficiently powered randomized clinical trial to address this issue would be challenging. In patients with MPN/PVT, long-term anticoagulation appears to be a double-edged sword, as 39 percent of the group studied by Hoekstra and colleagues developed gastrointestinal bleeding (similar findings were reported from a larger PVT trial2) while 27 percent developed recurrent thrombosis, with a mortality rate of 25 percent in this group with recurrent thrombosis. These data underscore the difficulty of balancing risks of thrombosis recurrence with anticoagulation-associated bleeding risks in patients with MPN/ PVT. Moreover, the results of a multicenter, observational study of 102 non-cirrhotic PVT patients (20% of whom also had a diagnosis of MPN) showed that anticoagulation failed to restore PVT patency in two-thirds. In this group with persistent thrombosis, ascites or splenic vein obstruction was more commonly observed (hazard ratio 3.5).2

We would be treading on thin ice to base recommendations for prophylactic anticoagulation on the data provided by Hoekstra et al. that showed that MPN/PVT can be complicated by either gastrointestinal hemorrhage (non-fatal in this report) or recurrent thrombosis (fatal is some cases) or by the findings of Plessier et al.2 (that identified patients with ascites or splenic vein obstruction as being at high risk for persistent thrombosis). Further, whether limiting anticoagulant choices to anti-Xa inhibitors (that have lower bleeding risks than VKA) will improve outcomes remains to be seen. We also don’t know how new approaches to treatment of MPNs will affect outcome. Although the study of Hoekstra provides valuable information on the natural history of MPN/PVT, how to best manage these complex patients remains an unanswered question.

  1. Kiladjian JJ, Cervantes F, Leebeek FWG, et al. The impact of JAK2 and MPL mutations on diagnosis and prognosis of splanchnic vein thrombosis: a report on 241 cases. Blood. 2008;111:4922-4929.
  2. Plessier A, Darwish-Murad S, Hernandez-Guerra M, et al. Acute portal vein thrombosis unrelated to cirrhosis: a prospective multicenter follow-up study. Hepatology. 2010;51:210-218.
back to top