What is Portal Vein Obstruction?
In the English literature, portal vein obstruction was first reported in 1868 by Balfour and Stewart, who described a patient presenting with an enlarged spleen, ascites, and variceal dilatation.
The vast majority of cases are due to primary thrombosis of the portal vein; most of the remaining cases are caused by malignant obstruction.
The portal vein forms at the junction of the splenic vein and the superior mesenteric vein behind the pancreatic head, and it can become thrombosed or obstructed at any point along its course. In cirrhosis and hepatic malignancies, the thromboses usually begin intrahepatically and spread to the extrahepatic portal vein. In most other etiologies, the thromboses usually start at the site of origin of the portal vein. Occasionally, thrombosis of the splenic vein propagates to the portal vein, most often resulting from an adjacent inflammatory process such as chronic pancreatitis.
Inherited and acquired disorders of the coagulation pathway are frequent causes of portal vein thrombosis. Inherited disorders include mutations in the prothrombin gene G20210A as well as deficiencies of various intrinsic anticoagulation factors, such as protein C and protein S, and activated protein C resistance. Acquired disorders include antithrombin III deficiency resulting from malnutrition, sepsis, disseminated intravascular coagulation, inflammatory bowel disease, liver disease, or estrogen use.
Coagulation disorders in portal vein thrombosis.
Stasis can be another major category for portal vein thrombosis. The global resistance to hepatic blood flow produced by cirrhosis is a common cause. Sclerotherapy for esophageal varices has been postulated as a possible mechanism though not proven thus far. The portal vein or its tributaries can be obstructed by adjacent tumor compression or invasion. Infectious and inflammatory processes may also lead to venous thrombosis.
Portal vein obstruction does not affect liver function unless the patient has an underlying liver disease such as cirrhosis.1 This is partially due to a rapid arterial buffer response, with compensatory increased flow of the hepatic artery maintaining the total hepatic blood flow. Formation of collaterals occurs rather rapidly as well, and they have been described as early as 12 days after acute thrombosis, though the average time to formation is approximately 5 weeks.
The development of a collateral circulation, with its attendant risk of variceal hemorrhage, is responsible for most of the complications and is the most common manifestation of portal vein obstruction. Other sequelae of the subsequent portal hypertension, such as ascites, are less frequent. Rarely, the thrombosis extends from the portal vein to the mesenteric arcades, leading to bowel ischemia and infarction.
Portal vein obstruction is a relatively rare condition with an overall incidence of 0.05-0.5% in autopsy studies. Incidence varies, depending on the group of patients studied (eg, general population vs patients with cirrhosis) and the method used to diagnose portal vein obstruction (eg, autopsy studies, angiography studies, noninvasive radiological screening).
Incidence of portal vein obstruction in people with cirrhosis has been reported to vary from 5-18%. However, these were patients referred for transplantation and were at an advanced stage of liver disease. No large autopsy studies are available. Extrahepatic portal vein obstruction is estimated to be responsible for 5-10% of all cases of portal hypertension.
In Japan, the frequency of portal vein obstruction in autopsy studies was reported to be 0.05%. In an angiography surveillance study of patients with cirrhosis, the incidence was 0.5%, which is much lower than the reported incidence in the Western literature.
In India, extrahepatic portal vein obstruction is reported more frequently; in one study, the incidence even exceeded reported cases of cirrhosis. Of all cases of portal hypertension in developing countries, 40% are attributed to portal vein obstruction, presumably secondary to an increased incidence of pylephlebitis associated with abdominal infections.
In the absence of cirrhosis, the 2 year bleeding risk from esophageal varices is reported to be 0.25% and of those that bleed the mortality rate is approximately 5%. Those with cirrhosis and varices have a 20-30% 2 year bleeding risk with a mortality rate of 30-70%. This difference is primarily a consequence of the normal hepatic function in the noncirrhotic patient. Variceal size is the major predictive factor for bleeding.
In adults with portal vein thrombosis, the 10-year survival rate has been reported to be 38-60%, with most of the deaths occurring secondary to the underlying disease (eg, cirrhosis, malignancy).
In children with portal vein thrombosis, the prognosis is much better overall, with a 10-year survival rate greater than 70%, which is attributable to the low incidence of underlying malignancy and cirrhosis.
No racial differences have been reported.
No sex differences have been reported overall, except for a slight male predominance in patients whose obstruction is secondary to cirrhosis.
The distribution of the age of presentation of portal vein thrombosis reflects the demographics of the underlying disease process. Primary portal vein thrombosis from coagulopathies occurs with equal frequency in adults and children.2 The frequency of portal vein obstruction from tumor compression or invasion is greater in adults.
In the acute phase, the presentation of portal vein obstruction is relatively uncommon and easily missed because the patient may be asymptomatic. Symptoms most often begin in the chronic or subacute stage. Schistosomiasis can cause presinusoidal portal obstruction by blocking intrahepatic portal venules with parasite eggs. It does not cause extrahepatic portal vein obstruction, though the clinical manifestations are often similar.
Patients can present emergently with sudden onset of right upper quadrant pain, nausea, and/or fever. Alternatively, the symptoms of the primary infectious and inflammatory condition that led to portal vein obstruction predominate (eg, right lower quadrant pain in appendicitis).
Progressive ascites, intestinal ischemia resulting from propagation of thrombus, or intestinal suffusion secondary to acute portal hypertension can also be the presenting manifestations. Occasionally, variceal bleeding can occur acutely with development of portal vein thrombosis, particularly in the setting of preexisting varices with cirrhosis.
Spontaneous resolution of acute/recent thrombosis undoubtedly occurs and symptoms abate. In other patients, the acute symptoms often subside as collaterals develop, and the diagnosis may be missed. These patients then present at a later stage with manifestations of portal hypertension.
These groups of patients most often present with complications related to portal hypertension. In 90% of cases, variceal bleeding is the presenting complaint. On average, this occurs 4 years after the thrombotic event and has been described as long as 12 years later. Ascites is less frequent, and hepatic encephalopathy is rare in the absence of preexisting cirrhosis.3
The specific etiology of the portal vein obstruction not only influences the initial clinical presentation but also the time course and prognosis.
In the presence of cirrhosis with underlying hepatic insufficiency, sudden worsening of hepatic function, development of hepatic encephalopathy, and development of ascites are all more frequent, leading to worse outcomes.
With intra-abdominal malignancies, bleeding is less commonly the first manifestation because many of these patients do not survive long enough to develop sequelae of portal hypertension. These patients most often present with sudden ascites, anorexia, right upper quadrant or epigastric pain, and weight loss. Portal vein obstruction may also be discovered incidentally on imaging studies obtained for pain or ascites.
Rarely, patients with portal vein obstruction present with a fever of unknown origin.
Splenomegaly is found in 75-100% of patients, most presenting in the chronic stage. Mild hepatomegaly is often present, as is right upper quadrant epigastric tenderness, especially in the acute setting.
Ascites is found infrequently. Stigmata of chronic liver disease, such as spider angiomata or palmar erythema, are usually found in the presence of underlying liver disease.
The presence of caput medusae indicates posthepatic or intrahepatic portal hypertension because it forms by recanalization of the umbilical vein, which connects with the left hepatic branch of the portal vein.4 It should not be observed in isolated extrahepatic portal vein obstruction because the obstruction is below the origin of the umbilical vein.
In children, growth retardation may be present.
Abnormalities of the extrahepatic biliary tree may occur in 80% of cases due to compression by choledochal or periportal varices or from ischemic stricturing. These findings manifest by jaundice, cholangitis, hemobilia, cholecystitis, or a hilar mass that can be mistaken for a cholangiocarcinoma.
In children and neonates, the most common etiology is intra-abdominal infection, accounting for 50% of all cases in this age group.
In this age group, neonatal sepsis with umbilical catheter placement has been reported to be the cause of portal vein thrombosis in 10-26% of cases.
Appendicitis is a commonly reported risk factor in children with portal vein thrombosis.
Congenital anomalies of the portal venous system, often associated with cardiovascular anomalies (eg, ventricular and atrial septal defects, deformed inferior vena cava) and biliary tract abnormalities, have been reported in 20% of children with portal vein obstruction and thrombosis.
In adults, cirrhosis is the major etiology, accounting for 24-32% of cases of portal vein thrombosis.
Neoplasms are another major cause, accounting for 21-24% of cases of portal vein obstruction, with hepatocellular carcinoma and pancreatic carcinoma causing most of these cases. These tumors can cause compression or direct invasion of the portal vein and lead to thrombosis by inducing a hypercoagulable state.5 Local ablative therapies for hepatocellular or metastatic disease have been linked to its development.
Although less common than in children, infections (predominantly intra-abdominal) still play an important role, with a particular association to Bacteroides fragilis bacteremia.
Myeloproliferative disorders and inherited or acquired coagulation disorders account for 10-12% of cases in adults.
Approximately 8-15% of cases have been reported to be idiopathic in the recent literature. For other less common etiologies, such as abdominal trauma, surgery, and inflammatory bowel disease, see Media file 1 or below.
The etiology of portal vein obstruction.
Other Problems to Be Considered
Congenital hepatic fibrosis
Primary biliary cirrhosis
Liver function test results are only mildly elevated in the absence of underlying cirrhosis or massive hepatic malignancy.
Inherited coagulation disorders, such as activated protein C resistance, are listed in Media file 2. Tests for these disorders should be ordered in any case of portal vein thrombosis in which the diagnosis is unclear.
Coagulation disorders in portal vein thrombosis.
In the presence of underlying hepatic insufficiency, levels may be low, presumably secondary to decreased production in the liver. Caution is therefore needed in making the diagnosis of an inherited thrombophilic disorder in this scenario.
Some authors suggest checking for an inherited coagulation disorder even when a local factor for portal vein thrombosis is obvious and, conversely, also checking for local factors even in the presence of inherited coagulation disorders because more than one risk factor may be present in a single patient.
Ultrasound: This is the first-line diagnostic modality because of its accuracy, affordability, and noninvasiveness.
The thrombus is observed as an echogenic lesion within the portal vein, though a recently formed thrombus may be anechoic (ie, not observable on standard grey-scale ultrasound).
The addition of color Doppler imaging is especially helpful in the detection of portal vein flow and the diagnosis of portal vein obstruction.
The sensitivity is around 70-90%, with a specificity of 99%. With Doppler, the false-positive rate is 9% in patients with cirrhosis because of sluggish or turbulent portal vein flow.
Major limitations are obesity and nonvisualization secondary to bowel gas.
The presence of pulsatile, arterial flow in the thrombus correlates with a malignant, not bland, thrombus.
MRI and magnetic resonance angiography (MRA): This is the next step if further portal venous information is needed. MRI is helpful if hepatic parenchymal detail is required (in hepatic malignancies), and, unlike CT scan, MRI can also quantitate portal and hepatic vessel flow, which is required in the planning of interventions, such as shunt surgery, transjugular intrahepatic portosystemic shunt (TIPS), or liver transplantation.
Acute clot (<5 wk) appears hyperintense on both T1- and T2-weighted images, whereas older clots appear hyperintense only on T2-weighted images. Tumor thrombi can be differentiated from bland thrombi because they appear more hyperintense on T2-weighted images and enhance with gadolinium.The overall sensitivity, specificity, and accuracy of the MRA are 100%, 98%, and 99%, respectively. There is a high sensitivity for detection of submucosal, serosal, paraesophageal collaterals.
CT scan: Contrast-enhanced CT scan shows a thrombus as a nonenhanced intraluminal-filling defect.
Portal vein thrombosis with cavernous transformation. The long arrow indicates the splenic vein at the junction with the superior mesenteric vein just below the site of thrombosis. The short arrow points to a serpiginous mass consistent with periportal collaterals, the so-called cavernous transformation of the portal vein.
Hepatocellular carcinoma with portal vein thrombosis.
Contrast-enhanced CT scan has the advantage over ultrasound in displaying varices (sensitivity, 65-85%) and parenchymal hepatic abnormalities.
The combination of CT scan and Doppler ultrasound is common in the evaluation of portal vein obstruction.
This examination is not usually required to confirm the diagnosis of portal vein thrombosis in the presence of CT scan or MRI.
Angiography’s major value lies in preoperative planning before shunt surgery or liver transplantation; however, it is not a prerequisite, and many transplant centers use MRI/MRA for this purpose.
Even angiography can provide false-positive results in portal hypertension in the presence of extensive portosystemic collaterals in which mesenteric flow is directed away from a patent portal vein.
Endoscopic ultrasound (EUS): Although not a common diagnostic modality, EUS has recently been found to be 81% sensitive and 93% specific in patients with portal vein thrombosis as compared to patients with thrombus confirmed by contrast-enhanced CT scan or surgery.
Usually, no specific alterations occur in the histology. In rats, apoptosis has been described in the underperfused portion, with increased mitotic activity in the remaining well-perfused liver.
The primary goals are to alleviate acute bleeding and to prevent further bleeding.
In the acute setting, these goals are best accomplished with variceal banding or sclerotherapy, often requiring several sessions to obliterate the bleeding. This has a success rate of 95% for the acute bleed.
Octreotide infusion has also been used in acute bleeding, with control of the acute bleed in 85% of patients. The rate of recurrent bleeding with this approach is 16-28%.
In the setting of portal vein obstruction, the role of propranolol to prevent rebleeding has not been studied, though it is used routinely.
Treatment of underlying etiology:
Anticoagulation in patients with acute/recent portal vein thrombosis, studied only retrospectively, has been shown to recanalize in more than 80% of cases. This is essential to prevent advancement of thrombosis or rethrombosis in patients with inherited coagulation disorders in which lifelong anticoagulation therapy is recommended once variceal control has been achieved.
Anticoagulation therapy has also been recommended after shunt surgery to prevent rethrombosis. There was a study in which 84 of 136 nonmalignant, noncirrhotic patients with portal vein thrombosis were anticoagulated with similar bleeding risks but less risk for thrombotic propagation. Debate remains regarding the risk-to-benefit ratio of anticoagulation in chronic portal vein thrombosis and should be decided on a case-by-case approach at this time.