Identifying causal factors underlying splanchnic vein thrombosis is central to the clinical assessment. These factors may be broadly categorized into those which are inherited and therefore intrinsically permanent and those which are acquired. The acquired conditions may be categorized as transient, correctable, or permanent. Examples of transient acquired factors include oral contraception, pregnancy, surgery, and trauma. In general, provoked thromboembolic events attributed to transient or correctable acquired risk factors have a sufficiently low risk of recurrence such that prolonged anticoagulant therapy is neither necessary nor advisable [8]. For unprovoked thrombotic events or for those with acquired and non-correctable risk factors, the risk of recurrence is deemed sufficiently high as to warrant prolonged secondary prevention with anticoagulants assuming the risk of major bleeding is mild to moderate. This distinction can be particularly important for splanchnic vein thrombosis where portal and mesenteric venous hypertension are associated with varix formation, which can greatly augment the bleeding risk [1, 7, 8].

Thrombosis involving the splanchnic veins results predominantly from acquired pathology of the organ drained by the involved venous segment or by inflammation and injury of adjacent organs. As part of the initial evaluation, it is therefore important to scrutinize the history, physical examination, cross-sectional imaging, and laboratory data for clues to organ pathology as a potential acquired cause for the splanchnic vein thrombosis. If identified, the laboratory assessment for an acquired or congenital thrombophilia is not likely warranted [8, 9]. A number of acquired conditions contributing to splanchnic vein thrombosis have been identified including a prevalence estimate from published cohorts (Table 23.1) [10–14]. In a series of 832 patients with splanchnic vein thrombosis, the most common underlying pathology was malignancy, present in 43 % of the cohort [14]. These cancers included solid tumors (27 %), hematologic malignancies (5 %), and myeloproliferative neoplasms (11 %). The three most common solid tumors were pancreatic (38 %), hepatobiliary (32 %), and gastrointestinal (19 %). When multiple splanchnic venous segments are involved, cancer was present in nearly half of patients. Moreover, cancer was associated with a higher prevalence of thrombosis in specific venous segments. For example, nearly half of patients with either splenic or portal vein thrombosis had an underlying malignancy. By comparison, hepatic vein thrombosis was associated with a solid tumor in 13 % of patients and a myeloproliferative neoplasm in 22 %. These high percentages mandate a thorough search for an underlying malignancy in all patients with documented splanchnic vein thrombosis.

A recent meta-analysis reported that the prevalence of myeloproliferative neoplasms may account for one third of patients with splanchnic vein thrombosis [15]. This compares to only 1 % for patients with conventional venous thromboembolism. In fact splanchnic vein thrombosis may be the initial clinical manifestation of myeloproliferative neoplasms in some patients. Myeloproliferative neoplasms, including polycythemia vera, essential thrombocythemia, and primary myelofibrosis, are an important consideration in the search for underlying mechanism of splanchnic vein thrombosis [1, 10, 14, 16–18]. These disorders represent a stem cell-derived clonal myeloproliferation [16–19]. Polycythemia vera results in an increased number of red cells, white cells, and platelets with limited bone marrow fibrosis. Essential thrombocythemia involves primarily platelet expansion. Primary myelofibrosis may be primary or reflect transformation of the polycythemia vera and essential thrombocythemia and is characterized by progressive fibrosis of the bone marrow. Although the most dramatic outcome is transformation to myeloid leukemia, the most common clinical manifestation and cause of death is thrombosis related [18]. These disorders result from an acquired mutation within JAK2, a member of the Janus kinase family of cytoplasmic tyrosine kinases associated with growth factor receptors [17]. This mutation involves a replacement of valine for phenylalanine in position 617 (V617F) of the JAK2 protein resulting in growth factor-independent proliferation of various cell lines [20]. This mutation may explain 90 % of cases of polycythemia vera and up to 50 % of cases of essential thrombocythemia. Screening for this mutation is therefore appropriate in the initial evaluation of patients suspected of having these disorders [20]. In a large series of patients with splanchnic vein thrombosis, myeloproliferative disorders were a prominent cause of both hepatic vein thrombosis (22 %) and multisegmental splanchnic vein thrombosis (17 %) [14]. In other series, investigators performed bone marrow biopsies on 128 patients with splanchnic vein thrombosis [21, 22]. Of these, 4 % carried a diagnosis of MPD prior to development of splanchnic vein thrombosis. An additional 33 % met criteria based on the bone marrow biopsy results.

Acute pancreatitis was also an important cause of splanchnic vein thrombosis accounting for 7–13 % of cases [10, 11, 14]. In one series, acute pancreatitis explained nearly half of all patients with splenic vein thrombosis [14]. This diagnosis should be relatively straight forward to confirm or exclude based on the acute presentation, clinical risk factors, examination, imaging, and commonly available laboratory testing. Cirrhosis is an important association, present in 16–24 % of cases. In particular, cirrhosis may account for up to one third of cases of portal vein thrombosis. This is an important diagnosis to confirm given the complexity of anticoagulant management in such patients. Recent abdominal surgery may explain 10–28 % of cases, particularly procedures involving splenectomy or liver transplantation [14]. Other important associations include sepsis of intra-abdominal origin (8–11 %), inflammatory bowel disease (1–8 %), and connective tissue diseases (up to 6 %).

Paroxysmal nocturnal hemoglobinuria is a rare but serious disorder involving clonal expansion of hematopoietic stem cells which lack the glycosylphosphatidylinositol membrane protein which anchors important proteins to the cell surface [22, 23]. Of these, complement-inhibiting proteins, CD55 and CD59, are particularly important for inhibition of complement fixation and mediated cellular lysis. Clinical manifestations of this disease include hemolytic anemia and both venous and arterial thromboembolism. Patients with this disease have been reported to suffer splanchnic vein thrombosis. Indeed, up to 10 % of patients with hepatic vein thrombosis may have this disease as an underlying mechanism. In large series, the prevalence of PNH in splanchnic vein thrombosis may account for 1–2 % of underlying causes [9, 14]. This disorder can be readily detected using flow cytometry and antibodies against CD55 and CD59 surface antigens.

Acquired risk factors for splanchnic vein thrombosis can be further stratified by thrombus location, which may be helpful to address the risk for specific cases (Table 23.2) [14]. Although splanchnic vein thrombosis can be viewed as a composite of venous segments, thrombosis within each individual venous segment contains unique features in terms of underlying etiology, risk of recurrence, and long-term prognosis [14]. Cancer is a prominent cause of thrombosis in each of the splanchnic venous segments and particularly when multiple venous segments are involved. Beyond cancer, hepatic vein thrombosis, for example, affects primarily younger women and has myeloproliferative disorder, cirrhosis, and oral contraception as prominent causes. Portal vein thrombosis, the most common within the splanchnic territory, has poor survival because of the relatively high incidence of cirrhosis. Isolated splenic vein thrombosis is uncommon, mainly occurs in middle-aged males with pancreatitis, and has a strong association with pancreatitis. Infection and recent abdominal surgery are common among patients with mesenteric vein thrombosis. For mesenteric vein thrombosis, a relatively high percentage of patients without an identifiable underlying etiology was evident.

In summary, the effective evaluation of patients with splanchnic vein thrombosis must begin with a thorough history and physical examination supported by appropriate laboratory testing and cross-sectional imaging. This evaluation should focus on the exclusion of underlying cancers including solid tumors, myeloproliferative neoplasms, and hematologic malignancies as well as other abdominal pathologies. This search will help to determine which patients are most likely to benefit from thrombophilia testing.

When considering thrombophilia testing, it is important to discern which patients to test, when to perform the testing, and what tests to include in the panel. Identifying those patients most likely to benefit from thrombophilia testing is an important first step for several reasons. First, indiscriminate thrombophilia testing in unselected patients with incident venous thrombosis is neither indicated nor cost-effective [9]. Second, thrombophilia test results have been shown to neither influence the initiation nor the intensity of anticoagulant therapy. Both the initiation and intensity of anticoagulation are the same irrespective of thrombophilia status. For example, heparin therapy in patients with antithrombin deficiency results in a therapeutic failure in only a small minority of patients [24]. Warfarin skin necrosis is very rare even in those patients with protein C or protein S deficiencies [25, 26]. High-intensity warfarin (INR 3.1–4.0) has not been shown to be superior to standard warfarin (INR goal 2.0–3.0) in patients with antiphospholipid antibodies and previous thrombosis [27]. Third, thrombophilia testing in unselected patients has not been shown to reduce thrombosis recurrence [28–30]. In a case-control study of patients from the MEGA cohort, the odds ratio for venous thrombosis recurrence was similar for those patients undergoing thrombophilia testing compared to those who were not tested [28]. In a separate cohort of 570 otherwise unselected patients with first objectively proven venous thrombosis, testing for thrombophilia did not aid in the prediction of venous thrombosis recurrence at 2 years [29]. Similar results were noted in a prospective study of 474 consecutive patients from the Leiden Thrombophilia Study [30]. Whereas the majority of patients with splanchnic vein thrombosis have clearly identified provoking factors, the role of thrombophilia testing should therefore be pursued with constraint [23]. Within large cohorts of patients with splanchnic vein thrombosis, only between 15 and 20 % lacked an identifiable provoking cause [9, 12, 13]. This percentage of unprovoked venous thrombosis is lower than has been observed for leg DVT/PE which has been reported at 26 % [13]. It is therefore our practice to reserve thrombophilia testing for patients with idiopathic or unprovoked venous thromboembolism.