How I treat splanchnic vein thrombosis

M.A. is a 73-year-old white man who presented to our Thrombosis Center 2 days after ultrasonographic diagnosis of occlusive portal vein thrombosis (PVT). He had known liver cirrhosis secondary to chronic alcohol consumption (Child class A). The ultrasound test was ordered because the patient complained of abdominal pain during the previous 5 days. The remaining clinical history was unremarkable with the exception of arterial hypertension and recurrent episodes of hemorrhoidal bleeding. The patient was prescribed a calcium-channel blocking drug. Laboratory tests revealed an international normalized ratio of 1.3, a prothrombin time ratio of 1.4, an activated partial thromboplastin time of 35 seconds (laboratory reference range, between 25 and 40 seconds) with a ratio of 1.1, and a platelet count of 90 000/mm³. Estimated glomerular filtration rate was 60 mL/minute. Upper gastrointestinal endoscopy was immediately ordered to check the presence of esophageal varices, and the patient was temporarily started on intermediate-dose low-molecular-weight heparin (LMWH) (1 mg/kg once a day), pending endoscopy. Grade 1 esophageal varices and a mild congestive gastropathy were detected, and treatment with LMWH was continued at the usual full therapeutic dose of 1 mg/kg twice a day. Propranolol for the primary prevention of variceal bleeding was also started. As no bleeding events occurred and as the hemoglobin level remained stable at the follow-up visit on day 10, it was decided to continue treatment with LMWH for a total of 1 month and then to switch the patient to warfarin, with a target therapeutic range international normalized ratio between 2 and 3, for an indefinite period of time. Complete recanalization of the thrombosis was observed at an ultrasound test performed at 3 months.

R.L. is a 46-year-old white woman who is followed at our Thrombosis Center for a previous episode of splanchnic vein thrombosis (SVT) diagnosed in 2013. The event occurred in the superior mesenteric vein and was diagnosed on abdominal computed tomography scan ordered because of pain. One year before SVT diagnosis, the patient was treated for an intraductal and papillary G2 breast cancer, stage IIA. After surgery, radiotherapy and chemotherapy adjuvant hormonal therapy with tamoxifen and gonadotropin-releasing hormone agonist was started. Because of the initial suspicion of cancer-associated thrombosis, mesenteric vein thrombosis (MVT) was treated with full therapeutic dose of LMWH (1 mg/kg twice daily) for 1 month. Thrombophilia screening was negative. Blood count was unremarkable, and Janus kinase 2 (JAK2) V617F mutation was absent. Recurrent breast cancer was ruled out. The patient refused to switch to warfarin. LMWH dose was reduced to 75% of the initial dose for the secondary prevention. Because of the lower thrombotic risk associated with aromatase inhibitors, the treating oncologist modified hormonal therapy, switching from tamoxifen to anastrozole. Antithrombotic therapy was reassessed at regular intervals. After more than 1 year of treatment, based on a comprehensive evaluation of clinical and radiologic parameters, the LMWH was stopped. The following data supported discontinuation: a 1-year treatment duration, a risk factor for SVT (tamoxifen) had been discontinued, and no clinical, blood testing, or radiological signs of cancer recurrence were present.

SVT includes PVT, MVT, splenic vein thrombosis, and the Budd-Chiari syndrome (BCS). BCS is the least frequent manifestation of the SVT spectrum, with an estimated incidence of about 0.5 to 1 case per million people per year.¹  The incidence of PVT and MVT is reported to range between 0.7/100.000 and 2.7/100.000 person-years.^2,3  Recent advances in imaging techniques are facilitating the detection of SVT in both symptomatic and asymptomatic patients. In a recent retrospective review of computed tomography scans of the abdomen performed for reasons other than the search for suspected SVT, the prevalence of unsuspected abdominal vein thrombosis was 1.7% (95% confidence interval [CI], 1.3-2.3), with the majority of events occurring in the splanchnic veins.⁴ 

Management of SVT is often a clinical challenge because of the heterogeneity of patient populations and the severity of short- and long-term outcomes if inadequately treated, as well as because active bleeding or bleeding risk factors at presentation are common in these patients. In this review, we discuss the acute and long-term anticoagulant treatment of SVT.

SVT has long been defined as “primary” or “secondary,” depending on the presence or absence of associated abdominal (local) or systemic risk factors. Although such a distinction is not always clear in clinical practice, emerging risk factors have decreased the number of truly unprovoked (primary) events, such that in recent reports, they account for 15% to 27% of patients.^5,6 

The relative incidence of risk factors varies with age, economic status, geographical area, and site of thrombosis. For example, a local precipitating factor is rare in BCS,⁷  but not in PVT and MVT.⁵  Thus, hematologic disorders, autoimmune diseases, and the use of hormonal therapy are the most common risk factors in BCS, whereas liver cirrhosis, abdominal cancer, intraabdominal inflammatory conditions, and surgery are the most common risk factors in PVT/MVT.^5-7  A list of risk factors associated with SVT is reported in Table 1.

During the last several years, myeloproliferative neoplasms (MPNs) have emerged as a leading systemic cause of SVT. Results from the largest cohort studies report a prevalence of MPNs of about 10% when all SVT patients are included, and up to 50% in cohorts including patients with BCS only.^5-7  SVT may actually represent the first clinical manifestation of MPN, particularly when the JAK2V617F mutation is present. The prevalence of JAK2V617F mutation in patients with SVT has been reported to be 32.7% (95% CI, 25.5%-35.9%),⁸  ranging between 27.7% (95% CI, 20.8%-35.8%) in nonmalignant, noncirrhotic patients with PVT to 41.1% (95% CI, 32.3%-50.6%) in patients with BCS.⁹  JAK2V617F screening in patients with SVT without overt MPN features identified MPN in 15.4% and 17.1% of screened patients, respectively.⁹  Thus, routine screening for JAK2V617 mutation facilitates the diagnosis of MPN in this setting, but whether MPN-specific treatment should be initiated in positive patients with no overt MPN is a question that remains to be answered.⁹ 

Sex-specific risk factors such as oral contraceptives, hormone replacement therapy, pregnancy, and puerperium are particularly relevant to the pathogenesis of BCS,^5,7  with a reported prevalence of nearly 40% in both retrospective and prospective cohort studies.^1,7  In PVT, the reported prevalence of these risk factors is lower, ranging from 4% to 15%.^2,5 

Less common risk factors include autoimmune diseases such as Behçet’s disease and hematologic disorders such as paroxysmal nocturnal hemoglobinuria (PNH). Thrombosis is one of the most important complications of PNH and can be the presenting feature.¹⁰  Thus, in patients with SVT and other compatible clinical findings suggesting the presence of intravascular hemolysis, the presence of PNH should be considered.¹¹ 

Among inherited thrombophilias, deficiencies of antithrombin, protein C, and protein S are infrequently diagnosed. Diagnosis of inherited disorders of these proteins is complicated by liver impairment, resulting in reduced synthesis of coagulation factors. The factor V Leiden mutation is more strongly associated with BCS than with PVT, whereas the converse is true for the prothrombin G20210A mutation.^12,13 

BCS, PVT, and MVT may present as distinct entities, although concomitant involvement of two or more splanchnic veins is not uncommon.¹⁴  Acute and chronic SVTs are different stages of the same disease.

BCS can be classified as fulminant, acute, subacute, or chronic.¹⁵  Fulminant BCS is extremely uncommon and is usually associated with fatal hepatocellular necrosis and hepatic encephalopathy. Acute BCS is characterized by symptoms such as ascites and hepatic necrosis without the formation of venous collaterals; chronic BCS may present with symptoms of portal hypertension and hepatic cirrhosis.¹⁶  Up to 15% of BCS is asymptomatic.¹⁷  The clinical presentation depends on the extent and rapidity of development of hepatic outflow obstruction, as well as the degree of liver decompression via collateral blood flow. Hepatomegaly, splenomegaly, right upper abdominal quadrant pain, and ascites are frequently seen in these patients, whereas mild jaundice and slight elevation of the aminotransferase liver function tests occur rarely.^7,18 

PVT may present as an acute or chronic disease, and patients may have several different symptoms or may be occasionally asymptomatic. The extent of the obstruction in the portal venous system, the speed of its development, and the presence of concomitant provoking diseases explain the heterogeneity of the clinical presentation. Acute PVT may be characterized by a sudden onset of abdominal pain, with fever and other nonspecific abdominal symptoms including nausea, vomiting, and diarrhea.⁵  Signs of portal hypertension are typically lacking in noncirrhotic patients. In general, liver function is preserved because of the rapid development of collateral veins and the compensatory increase in the hepatic arterial blood flow. Portal cavernoma (cavernous transformation of the portal vein) is the typical sign of chronic PVT, in which the portal vein is obstructed (partially or in total) and is replaced by fibrous tissue with the development of a hepatopetal network of periportal collateral veins. Portal hypertension is common in these patients, and its complications such as bleeding from esophageal or gastric varices and hypersplenism are the main presenting clinical manifestation.¹⁹  Less common manifestations include nonspecific dyspeptic symptoms, ascites, and hepatic encephalopathy.^20,21  These latter two manifestations are rarely observed, but subclinical cognitive changes are frequently present.²²  Finally, portal cholangiopathy with mild cholestasis or more severe biliary complications is increasingly recognized in patients with chronic PVT.²³  Clinical manifestations include jaundice, pruritus, fever, and abdominal pain.

In cirrhotic patients, PVT is frequently asymptomatic and is detected during hepatic ultrasound examination. However, PVT should be suspected in case of gastrointestinal bleeding or with the development or abrupt worsening of ascites or hepatic encephalopathy.^5,6 

MVT may present as an acute, subacute, or chronic disease.²⁴  Significant abdominal pain is typical of acute MVT, and it can be associated with diarrhea, nausea, vomiting, and lower gastrointestinal bleeding.²⁴  Severe abdominal pain, usually radiating to the back and ileus as a result of intestinal ischemia, may be present when the proximal mesenteric venous arches are involved.²⁵  Progression to intestinal infarction should be suspected in case of hematochezia, ascites, metabolic acidosis, or acute renal or respiratory failure. Mortality rates of these patients may be high, and surviving patients may have severe morbidity even after a prompt surgical resection. Chronic presentation with no acute abdominal pain and extensive venous collateral circulation is not uncommon.²⁴ 

Doppler ultrasound has a sensitivity of about 90% in the diagnosis of BCS and PVT.^26,27  However, its accuracy may be affected by patient characteristics, expertise of the operator, and awareness of the clinical suspicion of SVT. Beyond the occlusion of the hepatic veins, patients with BCS may also present with ultrasonographic liver findings including caudate lobe enlargement, focal liver enhancement, hypervascular nodules, and intrahepatic collateral vessels.²⁶  In cases of PVT, presence of hyperechoic material within the portal vein, distension of the portal vein and its tributaries, and total or partial absence of flow in these vessels is seen.²⁷  Computed tomography and magnetic resonance imaging have better accuracy compared with Doppler ultrasound in the diagnosis of BCS and PVT, and concomitant diseases and/or alternative diagnoses may be seen with these techniques.

Ultrasound is not helpful in the diagnosis of MVT because its accuracy is limited by overlying bowel gas. Thus, computed tomography or magnetic resonance imaging scans should be considered where MVT is suspected.²⁸  With these techniques, signs suggestive of intestinal infarction, such as thinning or thickening of the intestinal wall, lack of mucosal enhancement after contrast injection, or the presence of intramural gas, may be detected.²⁹ 

The prognosis of SVT mainly depends on the site and extension of disease and on the presence of underlying disorders.

In a single-center inception cohort study of 832 patients with SVT at any anatomic site, the 10-year survival rate was 60% (95% CI not reported), and older age, active cancer, and MPN were independent predictors of mortality.⁵  Patients with isolated PVT had the lowest survival rate, whereas patients with isolated hepatic vein thrombosis had the highest rate. The cumulative incidence of recurrence at 10 years was 24%, and the annual incidence of major bleeding events was 6.9/100 patient-years. Gastroesophageal varices and warfarin were independent predictors of bleeding.

Two large studies have reported mortality rates in patients with BCS. In a multicenter chart review of 237 patients with BCS, overall survival at 1, 5, and 10 years was 82% (95% CI, 77%-87%), 69% (95% CI, 62%-76%), and 62% (95% CI, 54%-70%), respectively. Encephalopathy, ascites, prolongation of the prothrombin time, and elevated bilirubin were independently associated with poor prognosis.¹⁷  In a multicenter study of 163 consecutive cases of patients with BCS, the survival rate was 87% (95% CI, 82%-93%) at 1 year and 82% (95% CI, 75%-88%) at 2 years.⁷  The frequency of recurrent thrombosis or major bleeding was not reported in these studies.

In a retrospective cohort study of 173 patients with PVT followed-up for a median of 2.5 years, the overall survival was 69% (95% CI, 61%-76%) at 1 year and 54% (95% CI, 46%-62%) at 5 years. These rates were 92% (95% CI, 86%-98%) and 76% (95% CI, 66%-86%), respectively, after the exclusion of patients with cancer or cirrhosis.²  Age, bilirubin, cirrhosis, and malignancy were significant predictors of mortality in multivariate analysis.²  In a retrospective cohort study of 136 patients with nonmalignant, noncirrhotic PVT, 84 of whom received some form of anticoagulant therapy, the incidence of thrombotic events was 5.5/100 patient-years (95% CI, 3.8-7.2), and the incidence of any gastrointestinal bleeding was 12.5/100 patient-years (95% CI, 10-15). Large varices predicted bleeding events, whereas the presence of an underlying prothrombotic state as well as the absence of anticoagulant therapy were associated with an elevated risk for recurrence.³⁰  In a prospective cohort study of 102 patients with PVT without cirrhosis or solid cancer, 95 of whom were receiving anticoagulant therapy, the 1-year recanalization rate of the portal vein was achieved in 38% of patients, and all recanalizations occurred in the first 6 months of treatment.³¹  In a retrospective cohort of 120 patients with PVT who had neither cirrhosis nor solid malignancy, of whom about half received anticoagulation, the overall risk for recurrent thrombosis was 3% (95% CI, 0%-7%) at 1 year, 8% (95% CI, 3%-14%) at 5 years, and 24% (95% CI, 13%-36%) at 10 years.³²  Of the 22 recurrent events, 16 occurred outside the abdomen, including the pulmonary arteries, cerebral veins or arteries, or the limbs. The overall risk for gastrointestinal bleeding was 33% (95% CI, 24%-41%) at 1 year, 43% (95% CI, 33%-53%) at 5 years, and 46% (95% CI, 36%-56%) at 10 years, with the majority of bleeds being variceal.³² 

A few small cohort studies have reported survival in patients with MVT. In a retrospective study of 51 patients, 30-day mortality was 20%.³  Intestinal infarction at the time of diagnosis was independently associated with an unfavorable short-term prognosis, and cancer was independently associated with mortality.³  Another retrospective cohort study of 68 patients reported the same 30-day mortality rate and found malignancy, advanced age, and symptom duration to be independent predictors of mortality.³³ 

The quality of evidence guiding the treatment of SVT is low because it is based on the results of observational studies only.

In the 2 studies by Darwish Murad et al on patients with BCS, the frequency of use of anticoagulant drugs ranged between 72% and 86%.^7,17  A small minority of patients received thrombolytic therapy, and portosystemic shunting was performed in 34% to 49% of the patients.^7,17  There was a trend toward a more conservative approach, with noninvasive therapies increasing in frequency over time. Previous studies retrospectively assessing factors associated with survival in adults with BCS suggest that anticoagulation is associated with increased long-term survival in this population.³⁴ 

Studies conducted in noncirrhotic patients with PVT have reported varying rates of use of anticoagulant therapy, ranging from 55% to more than 90%, whereas the use of thrombolysis (mainly administered systematically, when reported) or invasive procedures was infrequently reported.^2,30-32  Anticoagulant drugs were associated with a significant increase in gastrointestinal bleeding in some, but not all, studies.

Optimal management of PVT in cirrhotic patients is even less well studied. The largest report is from a prospective cohort of 56 patients with nonmalignant PVT, 33 of whom received LMWH and 6 of whom underwent a portosystemic shunt procedure. The recanalization rate was higher in patients treated with LMWH than in untreated patients, and thrombus progression was significantly less frequent in treated patients.³⁶  Variceal bleeding occurred in 5 untreated patients and in 1 patient treated with LMWH.³⁶ 

The risk for recurrence and bleeding appears to be higher in patients with PVT, either isolated or associated with MVT or splenic vein thrombosis, than in patients with isolated MVT. Anticoagulant therapy is effective in reducing the risk for recurrence independent of the site and extent of SVT.³⁵  A retrospective cohort of 77 patients with MVT treated with vitamin K antagonists reported an annual risk for recurrence of 4.6% person-years in the about 40% of patients who stopped anticoagulant treatment.³⁷  No data are available describing the short- or long-term risk for recurrence in patients with incidentally detected SVT who are not treated.

Given the lack of adequate evidence, the following recommendations are substantially based on expert opinion, taking into account published peer guidelines.

The 9th edition of the American College of Chest Physicians guidelines on Antithrombotic Therapy and Prevention of Thrombosis recommend anticoagulation over no anticoagulation in symptomatic patients with SVT, and no anticoagulation for incidentally detected thrombosis.³⁸  However, the authors encourage clinicians to consider anticoagulant therapy in case of extensive thrombosis that appears to be acute, with progression of thrombosis, and in patients with active cancer.³⁸  Recommendations on treatment duration follow those made for patients with deep vein thrombosis of the lower limbs or pulmonary embolism.³⁸  Treatment of 3 months’ duration is recommended in the case of provoked thrombosis. Treatment of longer than 3 months’ duration is recommended in the case of unprovoked venous thromboembolism if the bleeding risk is perceived to be low or moderate.³⁸  Finally, extended duration therapy is recommended for cancer associated venous thromboembolism.³⁸ 

The American Association for the Study of Liver Diseases also recommends anticoagulation for patients with acute PVT and BCS.³⁹  For both patients with acute and patients with chronic PVT, treatment is recommended for at least 3 months, with an indefinite duration for patients with permanent risk factors or with extension to the mesenteric veins. Indefinite duration treatment is recommended for all patients with BCS.³⁹  Portosystemic shunting is recommended for patients with BCS whose clinical conditions do not improve with anticoagulant therapy.³⁹ 

We have recently completed an international registry describing treatment strategies and factors associated with therapeutic decisions in patients with SVT.⁶  Of the 613 patients enrolled in the study, 22.2% did not receive anticoagulant therapy within the first month after diagnosis of SVT. Incidental diagnosis, single vein thrombosis, gastrointestinal bleeding at presentation, thrombocytopenia, cancer, and cirrhosis were significantly associated with no treatment.⁶  Of the 470 anticoagulated patients, 406 received LMWH (66.2%) during the acute phase, and 270 were switched to vitamin K antagonists (VKA), whereas 175 continued receiving long-term parenteral anticoagulants. Factors associated with extended-duration parenteral anticoagulation were incidental diagnosis, solid cancer, cirrhosis, and low platelet count.⁶  Therapeutic strategies according to the site of thrombosis are reported in Table 2. Of 176 patients with incidentally diagnosed SVT, 110 (62.5%) received anticoagulant treatment.

The management of SVT requires a multidisciplinary approach that may include a gastroenterologist/hepatologist, a hematologist and/or a thrombosis expert, and an interventional radiologist and/or a surgeon.

Anticoagulant treatment should be considered for all patients with objectively diagnosed SVT. Available data suggest that anticoagulation may improve survival, reduce the risk for recurrence, and improve recanalization, an important outcome in patients with PVT because it prevents portal hypertension and the formation of cavernoma. Improved rates of recanalization may also result in long-term reductions in bleeding risk by reducing portal venous pressure.

Given the heterogeneity of patients with SVT, risks and benefits of anticoagulant treatment need to be evaluated on an individual basis, taking into account severity of presentation, site and extension of thrombosis, and clinical and laboratory baseline characteristics. A thorough assessment of any potential risk factor for SVT, including those that may not be immediately detectable at the time of diagnosis, will help in determining the intensity and duration of antithrombotic treatment.

The decision to not administer antithrombotic drugs at the time of diagnosis may be justified by a very high risk of bleeding or evidence of active bleeding, or by a short-term poor prognosis in patients with end-stage cancer or cirrhosis, in particular when the thrombus has limited extension or is incidentally diagnosed (Table 3). This decision needs periodic reassessment.

The use of thrombolytic agents should be limited to very select cases with severe presentation and acceptably low risk of bleeding, as may be the case in some patients with MVT with signs of intestinal ischemia. In the absence of evidence from the literature, the choice between local or systemic lysis should be based on local experience and preferences.

All other patients should be started on anticoagulant drugs as soon as possible, ideally following the same strategy recommended for patients with deep vein thrombosis of the lower limbs.³⁸  This includes initial treatment with parenteral anticoagulation (preferably LMWH, given its advantages over unfractionated heparin) and early initiation of VKA (the same day parenteral therapy is started) in all patients. Patients with cancer-associated SVT should receive LMWH as a sole treatment for at least 3 to 6 months.³⁸ 

Patients with SVT pose a number of specific challenges that often require some changes to this strategy. In general, we prefer not to start VKA therapy on the first day of treatment, but we wait at least 48 to 72 hours to rule out early bleeding complications. In very high-bleeding-risk patients (eg, patients with recent gastrointestinal bleeding), we prefer the initial use of unfractionated heparin over LMWH because of its shorter half-life and because of its reversibility. In patients without cancer with persisting risk factors for bleeding (eg, patients with low platelet count and/or patients with cirrhosis and varices), we prefer to continue with parenteral treatment alone (LMWH) without starting VKA during the first weeks (or longer) because of the shorter half-life of LMWH compared with VKA. In cirrhotic patients, we request endoscopic screening of esophageal varices, and we may delay the initiation of anticoagulant treatment on the basis of the results of endoscopy. In the management study by Senzolo et al,³⁶  patients with previous variceal bleeding and those with grade 2 esophageal varices with red signs and grade 3 varices were banded before starting anticoagulation, and treatment was started no earlier than 15 days after the last banding session. Last, but not least, we use reduced doses of LMWH (at least 50% of the full therapeutic dose) in patients with platelet count below 50 000 mm³, as well as in patients with severe renal impairment. However, in patients with estimated glomerular filtration rate between 15 and 30 mL/minute and SVT, we prefer unfractionated heparin, given the high risk of bleeding of these patients (Table 3).

There is scant information also to drive the optimal duration of treatment. The true incidence of recurrence after discontinuation of anticoagulant treatment is not defined, but the effect of recurrences may be severe, as in about one fourth of cases, these occur as hepatic, mesenteric, or splenic infarctions.³⁰  Bleeding rates appear to be higher than in other clinical settings, but the incidence of life-threatening bleeding is unknown and case-fatality rate of thrombosis after treatment discontinuation was significantly higher than that of gastrointestinal bleeding in a study on patients with MVT.³⁷  Because the majority of SVT events have underlying persisting or permanent risk factors, we favor indefinite treatment duration for most patients, provided the risk of bleeding is acceptably low. These include patients with cirrhosis, active solid cancer, active hematologic cancer, and MPNs, both overt and in the presence of JAK2V717F mutation, given the strong association between this mutation and SVT. These also include patients with chronic inflammatory disorders or autoimmune diseases (Table 4). We prescribe similar treatment durations for patients with incidentally detected SVT who have received an initial course of anticoagulant therapy, because we have found no evidence of a lower risk for recurrence in these patients compared with in patients with symptomatic SVT.

Clinical vignette 1 reflects the high frequency of risk factors for bleeding complications in patients with liver cirrhosis and SVT. This patient had a number of risk factors for bleeding, but no major contraindications to anticoagulants. Thus, anticoagulant treatment was started at a reduced dose. LMWH was maintained during the first month (the highest-risk period for bleeding) after which the patient was switched to warfarin as he no longer desired parenteral therapy; however, long-term LMWH would be acceptable in such patients.

Clinical vignette 2 reflects the complexity often encountered when making decisions around the duration of anticoagulation for secondary prevention. Despite the absence of recurrent breast cancer and the discontinuation of a provoking factor (tamoxifen), treatment was continued for 1 year because of the unusual site of thrombosis. The decision to extend therapy was driven mainly by patient preference, but it also reflected the clinician concerns for SVT recurrence, which commonly leads to prolonged anticoagulant treatment in this setting.

The pharmacokinetic and pharmacodynamic characteristics of the direct oral anticoagulant drugs (DOACs) suggest their potential to improve the therapeutic management of SVT. Thanks to their oral route of administration, their short half-life, and their predictable dose-response, the DOACs are a potential alternative to both heparins and VKA for the acute and long-term treatment of the disease. However, patients with SVT have not been enrolled in phase 3 clinical trials of the DOACs, and the reported increased risk for gastrointestinal bleeding, at least with some DOACs, is a matter of concern. Furthermore, DOACs remain contraindicated in patients with acute or chronic severe liver impairment as a result of their partial metabolism through the cytochrome P450 3A4 system. A few anecdotal case reports of patients with PVT or MVT treated with rivaroxaban were recently published.^40,41 

Defining evidence-based management of SVT remains a largely unmet clinical need. Good-quality research is warranted, but given the epidemiology of SVT, and given that study of each specific patient subgroup (eg, cirrhotic PVT or cancer-associated PVT) should be conducted independently, it seems unlikely that adequately sized RCTs will ever be carried out. Large, prospective observational studies are more feasible and may add useful information on short- and long-term outcomes in patients on and off treatment. Management studies, prospectively assessing the safety and efficacy of specific treatment protocols, may provide valuable evidence to support recommendations. To allow interstudy comparisons, better standardization of outcome definitions, in particular for bleeding events, is also warranted.

Contribution: W.A., F.D., and A.S. contributed to the narrative review of the literature and wrote the paper.

Conflict-of-interest disclosure: W.A. has participated in advisory boards for Bayer HealthCare, Boehringer Ingelheim, Bristol-Myers Squibb, Pfizer, and Daiichi Sankyo and has received travel or research support from Bayer HealthCare, GlaxoSmithKline, Alexion Pharmaceuticals, Pfizer, Bristol-Myers Squibb, Daiichi Sankyo, and Boehringer Ingelheim. The remaining authors declare no competing financial interests.

Correspondence: Walter Ageno, Department of Clinical and Experimental Medicine, University of Insubria, Via Guicciardini 9, 21100 Varese, Italy; e-mail: agewal@yahoo.com or walter.ageno@uninsubria.it.