Portal vein thrombosis

Portal vein thrombosis has come up in  Question 27 from the first paper of 2012 (where you had to recognise it on CT).  The CT is how it usually recognised, because there may actually be no clinical findings. Generally speaking, these people also have liver disease, and it is not abnormal for them to have asictes and bleeding varices - which is how an acute elevation in portal venous pressure may go unrecognised until abdominal pain prompts a CT of the abdomen. 

Clinical features of portal vein thrombosis

"Nothing" or "vague unwellness" may actually be the only clinical findings, but more generally people will present either with features of portal venous hypertension (such as ascites) or with abdominal pain. Either way, the presentation is usually such that would ultimately lead to a contrast CT of the abdomen.

Historical features

• Abdominal pain, typically radiating to the back
• Abdominal distension
• Diarrhoea
• Worsening of pre-existing portal hypertension (among those chronically drunk and yellow)
  • Variceal bleeding
  • Haemorrhoids
  • Fevers, painful liver (portal thrombophlebitis)

Physical examination

• Ileus
• Ascites
• Features of portal hypertension and portosystemic shunting, eg. caput medusae

Radiological features of portal vein thrombosis

a) can be answered just by describing the pathology on the CT. What should there be? Mathieu et al (1985) and Lee et al (2008) describe several characteristic CT findings:

Direct visualisation of the clot:

• Usually seen as a hypodense filling defect in the portal vein lumen
• The decreased density inside the portal vessels is best seen in the contrast phase
• Non-contract CT will not pick this up: you need portal venous phase contrast enhanced studies.
• You can also sometimes see enhancement of the walls of the portal vein, which is either dilated vasa vasorum or a thin film of contrast flowing around the obstruction
• The thrombus itself should not enhance with contrast. If there is enhancement, particularly in the arterial phase, the thrombus is probably malignant in origin (i.e a HCC has eroded into the lumen).

Sequelae of portal vein obstruction:

• Venous infarcts of the liver and spleen
• Cavernous transformation of the portal vein (looks like a varicocele!)
• Posrtosystemic collateral vessels and arterioportal shunts
• Gut ischaemia
• Ascites

Radiology of the portal vein

Where am I looking? Where's the portal vein? Lets pretend I've never seen a CT. Here's a helpful image from Radiopedia, with nice obvious arrows pointing to the healthy vessels:

Now, here are some shots of the portal vein occluded with clot, stolen shamelessly from Lee et al (2008):

Image A shows the thrombus as a filling defect in the contrast-filled splenoportal junction. Image B is a coronal reconstruction showing some sort of low-density structure in the portal vein.

Causes of portal vein thrombosis

Valla and Condat (2000)  discuss these at great length, and their excellent lists have been incorporated into the table below. Broadly, we can organise the causes into two major categories: "something wrong with the liver" and "something wrong with the clotting cascade"

Causes of Portal Vein Thrombosis

Local causes ("something wrong with the liver") Local inflammatory lesions Neonatal omphalitis Diverticulitis Appendicitis Pancreatitis Duodenal ulcer Cholecystitis Tuberculous lymphadenitis Portal venous system injury Surgical portacaval shunting (or TIPS) Pancreatic islet cell transplantation Splenectomy Colectomy Gastrectomy Really, any abdominal surgery Malignancy Hepatocellular carcinoma Abdominal malignancy of any sort Portal venous stasis Cirrhosis Pancreatic tumour (compression) Retroperitoneal fibrosis

Systemic problems ("something wrong with the clotting cascade") Acquired prothrombotic states Primary myeloproliferative disorders Antiphospholipid syndrome Paroxysmal nocturnal haemoglobinuria Oral contraceptives Pregnancy and post-partum Malignancy Hyperhomocysteinemia Congenital prothrombotic states Antithrombin deficiency Protein C deficiency Protein S deficiency Factor V Leiden mutation Factor II (prothrombin gene) G20210 mutation Systemic inflammatory states Behçet's syndrome Severe sepsis SLE Inflammatory bowel disease, eg. Crohns

Management of portal vein thrombosis

If one does not wish to pay for a pre-digested UpToDate management discussion, one may refer to freely available review articles such as Boyer (2008) or Basit et al (2015). Unless otherwise stated, these three sources have informed the discussion below.  In short, the options are as follows:

Do nothing.

The thrombus sometimes self-resolves. It is though that in the cirrhotic community there may be frequent small thrombi forming and dissolving all the time.  If the patient is asymptomatic, there may be an argument against anticoagulation.  If the patient has significant risks associated with anticoagulation (eg. bleeding varices), one may also rely on portal cavernous transformation to divert venous flow around the thrombus. However, rate of spontaneous resolution is low (~ 17%).

Anticoagulation

Apparently, after six months of treatment 90% of the patients will recanalise their portal vein. Those who benefit most from anticoagulation include patients without varices and patients with prothrombotic states. Ideally, anticoagulation should be started in the first week (according to Basit et al, "Rate of recanalization is 69% if anticoagulation is started in the first week, but 25% if anticoagulation is started in second week").

Warfarin is the standard for non-cirrhotic patients and for those without malignancy. If not warfarin, then low molecular weight heparin. 

Management of portal venous hypertension

Several options are available:

• β-blockers, eg. propanolol
• Diuretics, eg. spironolactone
• TIPS procedure

Experimental or unacceptably risky treatments include:

• Intraportal thrombolysis
• Surgical thrombectomy (only if you are already in the abdomen, eg. for mesenteric ischaemia)