This young man has presented with a three-week history of epigastric pain, and more concerningly, several episodes of melena three days ago; note that the latter is worrisome of upper gastrointestinal (GI) bleeding. Given his age, two likely etiologies are gastritis and gastric or duodenal ulcers; the possibility of gastric carcinoma should also be kept in mind. Variceal bleeding secondary to portal hypertension and chronic liver disease is another common cause of melena; however, there are no clinical findings suggestive of the latter; note also that his current level of alcohol consumption is unlikely to give rise to cirrhosis. Upper GI endoscopy is an essential next step; this reveals the presence of grade II esophageal varices, a possibility already thought of, but considered unlikely. As mentioned before, esophageal varices are a manifestation of portal hypertension, which itself is most often a sequela of chronic liver disease. However, this young man has no obvious signs and symptoms of the latter, nor any risk factors which would favor it. One possibility is that his level of alcohol consumption is far higher than was self-reported; however, rather than jumping to conclusions, a structured and systematic diagnostic analysis should be employed. In this respect, it is important to appreciate that portal hypertension may arise from one of three major causal categories: - Pre-hepatic: portal vein thrombosis, splenic vein thrombosis, or tumoral compression - Hepatic: cirrhosis, acute fulminant hepatitis, alcoholic hepatitis, primary biliary cirrhosis, Wilson's disease, hemochromatosis. - Post-hepatic: thrombosis of the inferior vena cava, right heart failure, constrictive pericarditis, severe tricuspid regurgitation, Budd-Chiari syndrome, arterial-portal venous fistula. Liver function tests and an abdominal ultrasound will help both determine and categorize the etiology of the portal hypertension, and should be performed next; in this patient, the former is found to be normal. However, sonography reveals the presence of multiple echogenic structures within the portal vein causing partial obliteration of the lumen, and hepatofugal flow (i.e. portal venous blood flow directed away from the liver, which is always abnormal). These findings are suggestive of portal vein thrombosis (PVT). In non-cirrhotic individuals, PVT is most often secondary to hereditary or acquired thrombophilic states, local inflammation, infections, bowel surgery, or splenectomy. In this patient, given the absence of prior surgery or recent infection, a thrombophilia is thus highly likely, and further evalation is essential. The optimal treament of this patient is something of a conundrum. As the PVT is still acute (i.e. less than 60 days since symptom onset) anticoagulation may induce recanalization of the portal vein, resulting in resolution of symptoms. However, the presence of varices confers a significant bleeding risk. Conversely, if anticoagulation is not employed, recanalization is unlikely; he will then potentially progress to develop chronic PVT and/or experience recurrent episodes of thrombosis. One potential management plan is to ligrate the varices, and then commence anticoagulation, with careful monitoring; while not entirely without risk, it could be argued that this is likely to lead to the best long-term outcome. Unfortunately, data on the risk-benefit indices of more aggressive procedures, such as transjugular intrahepatic portosystemic shunt (TIPS) placement or thrombectomy are scant; they are probably best considered if more basic techniques fail.
As the name suggests, portal vein thrombosis (PVT) involves thrombotic occlusion of the portal vein and/or its tributaries, including the splenic, and superior and inferior mesenteric veins. The best-known single etiology of PVT is hepatic cirrhosis, found in 0.6% to 28% of patients. Another well-known cause is primary or metastatic hepatic carcinoma. However, myriad conditions may pose a risk for the disease; these may be broadly classified into local and systemic factors. Local factors include infections, inflammatory disease (e.g. pancreatitis), malignancies, cirrhosis, and direct injury to the portal venous system following surgery or blunt abdominal trauma. Inherited and acquired thrombophilias are the main systemic factors, while pregnancy and oral contraceptive use are also associated with hypercoagulability. The above classifications notwithstanding, PVT is often multifactorial in origin; furthermore, 8% to 15% of cases are idiopathic. To appreciate the sequela of PVT, one must recall that the liver has a dual blood supply: an arterial supply from the hepatic artery, and a unique venous supply via the portal vein. Obstruction of the portal vein may result in the loss of as much as two-thirds of the liver's blood supply; in the acute stages, this is compensated for by dilation of the hepatic artery. A few days later, venous collaterals develop to bypass the obstruction; this process is completed in around three to five weeks. The portal vein may eventually become just a thin, fibrotic cord which is hard to visualize on imaging studies; this late stage heralds the development of a hyperkinetic circulation, characterized by low systemic vascular resistance and increased cardiac output. Clinically, PVT may be classified as either acute or chronic, based on the time which has passed between the onset of symptoms and the initial clinical presentation. If under 60 days, the condition is said to be acute. Acute PVT is often asymptomatic; if present, symptoms include transient abdominal pain, or spikes of high fever with chills if associated with septic thrombi. Rarely, acute intestinal ischemia and bowel infarction may occur. Conversely, the chronic form presents with classical symptoms of portal hypertension, particularly GI bleeding. Abdominal pain is uncommon, unless the thrombus extends into the mesenteric veins, causing bowel ischemia or infarction. On examination, individuals with acute PVT have minimal physical findings. Abdominal distention secondary to ileus may be seen, or there may be guarding and ascites due to intestinal infarction. On the other hand, persons with chronic thrombosis may display distinct signs such as splenomegaly, hepatomegaly, abdominal tenderness, or low-grade fever. Ascites is rare in the absence of underlying cirrhosis. In both forms of the disease, routine laboratory studies are often unremarkable, save possibly for mild elevations in transaminases, alkaline phosphatase, and bilirubin. Definitive diagnosis requires imaging of the portal venous system. Ultrasonography may demonstrate an echogenic mass within the lumen, with distention of the portal vein proximal to the occlusion, and extensive collateral vessels; color flow doppler studies can confirm partial or complete absence of venous flow. Contrast computed tomography (CT) will demonstrate the thrombus as a filling defect within the portal venous system; complete occlusion may produce a "train track" appearance due to contrast around the vessel. With chronic thrombosis, evidence of cavernous transformation may be seen. MRI is an alternative to CT. In addition to the above information, it provides a detailed anatomy of the entire portal venous system, including the intrahepatic portal vessels. Note that angiography is rarely indicated, as CT and MRI are usually sufficient to establish the diagnosis. The management of PVT depends on whether the thrombosis is acute or chronic, and on the presence or absence of symptoms. In individuals with acute thrombosis, the management centers on anticoagulant therapy, so as to prevent extension of the thrombus and allow for recanalization of the obstructed vein; if initiated early, complete recanalization is seen in 38.3% of patients, and partial recanalization in 14%. Note that spontaneous recanalization in the absence of anticoagulation is rare; furthermore, data is scarce on the efficacy of systemic or local thrombolytic therapy and mechanical recanalization. Transjugular intrahepatic portosystemic shunting (TIPS) is an alternative modality of treatment. However, this is technically complex, and recurrent shunt occlusion may occur. Surgical thrombectomy is rarely indicated, save for when a laparotomy is being performed for a suspected bowel infarction. In chronic PVT, the goals of the management to prevent and treat variceal bleeding, and avoid recurrent thrombosis. The use of nonselective beta-blockers and endoscopic band ligation reduces the incidence of bleeding and prolong survival.