Vitamin K antagonists versus heparin for the treatment of splanchnic vein thrombosis in the ISTH registry: Results of 12-month follow-up and a propensity score analysis

Introduction: Splanchnic vein thrombosis [SVT] is a challenging disease, because of the concurrent increased risk of bleeding and potentially lifethreatening complications. We aimed to explore the actual management of SVT in a large prospective cohort and to report clinical outcomes during follow-up. Methods: Consecutive SVT patients were enrolled in a multicenter international registry, from 2008 to 2012. Clinical outcomes (major bleeding; vascular events, defined as venous or arterial thrombosis; mortality) were collected. A propensity score [PS], created from baseline characteristics of patients receiving the two main treatments (different dosages of parenteral anticoagulants vs vitamin K antagonists [VKAs]), was used to estimate the effect of therapeutic strategies. We here report the results of the 12-month follow-up. Results: 613 patients were enrolled; mean age 53.1±14.8 years; 62.6% males. Initially, 143 patients were not anticoagulated; 175 received parenteral anticoagulants only (mean duration 6.28±4.21 months) and 295 started VKAs (9.67±3.53 months). Major bleeding occurred in 25 patients, 16 on treatment (4.53/100 patientyears [pt-y]) and 9 off treatment (5.45/100 pt-y); vascular events in 47 patients, 27 and 20 (7.93/100 pt-y vs 11.85/100 pt-y); death in 79 patients, 44 and 35 (12.35/100 pt-y vs 20.30/100 pt-y). Solid cancer, ascites, anemia and thrombocytopenia were inversely associated with the prescription of VKAs in a binary logistic regression. Using PS based on these variables, we matched 102 patients treated with VKAs with 102 patients treated with parenteral anticoagulants. Major bleeding occurred in 5.88% vs 0.98% for VKAs and heparin, respectively; vascular events in 2.94% vs 8.82%; mortality in 5.88% vs 18.63%. Conclusions: In our cohort of SVT patients, the incidence of vascular and bleeding events at 1-year follow-up was relevant. Although the PS matching did not allow for a complete balance of heterogeneity, in designated patients VKA treatment appeared to be sufficiently safe.