Venous Thromboses at Unusual Sites

Overview

Mesenteric vein thrombosis ( Box 17.4 ) is distinct from and much less common than mesenteric artery thrombosis, accounting for only 5% to 15% of cases of mesenteric ischemia. Although it is rare, the mesenteric veins rank as the third most common site of venous thrombosis behind the lungs and the limbs. Mesenteric artery thrombosis is usually seen in older, hypertensive, and often diabetic patients who experience an acute abdominal catastrophic event that rapidly leads to ischemia and death of the abdominal organs supplied by the superior mesenteric artery. A detailed description of mesenteric artery thrombosis is outside the scope of this chapter but is provided in other sources. Compared with mesenteric artery occlusion, thrombosis of the superior and inferior mesenteric veins is less common and less precipitous; it is often subacute. Superior mesenteric vein thrombosis has historically been difficult to diagnose because less sophisticated radiographic images, such as flat plates and upright views of the abdomen, as well as various barium studies, provide nonspecific findings and often do not lead to the diagnosis. The diagnosis of mesenteric vein thrombosis is now more readily made, even if accidentally, by CT with vascular contrast enhancement. The thrombosed superior mesenteric vein is seen as a large, distended vessel that does not fill appropriately with contrast medium ( Fig. 17.1 ). The bowel wall may be thick and edematous. In some cases a misty mesentery has been observed on abdominal CT scan, caused by inflammation and edema from mesenteric vein thrombosis ( Fig. 17.2 ). This finding totally resolves with successful treatment of the thrombosis. Concomitant extrahepatic PVT occurs in 65% to 72% of cases.

Two sequential articles in the Annals of Surgery in 1895 provided seminal insight into mesenteric vein thrombosis. Delatour described a fatal case of mesenteric infarction that occurred after an enlarged spleen had been electively removed to treat a patient with what was most likely a myeloproliferative disorder. He remarked that the hematologic condition so changed her blood that coagulation was easily induced. Elliot was able to remove infarcted bowel in a living patient yet noted that the patient concomitantly had bilateral femoral vein and PVT, which suggested an underlying hypercoagulable condition.

Causes

The causes of mesenteric vein thrombosis are multiple. One must use caution in reading older reviews of mesenteric vein thrombosis that described patients before the modern era of hypercoagulability awareness, because obviously the presumed causes could not include the thrombophilic diseases that had yet to be defined. In one study, mesenteric vein thrombosis accounted for 0.01% of all surgical admissions and surgical autopsies, and, in another study, accounted for 0.06% of all surgical admissions. Clearly, this is a rare disease. On the other hand, in a cross-sectional study, mesenteric vein thrombosis was found to have occurred in 3% of patients who were known to have a deficiency of ATIII, protein C, or protein S ; in another cross-sectional study, mesenteric vein thrombosis was found in 10% of patients with ATIII deficiency, in 6% of those with protein C deficiency, and in 4% of patients with protein S deficiency. One can estimate that the incidence of mesenteric vein thrombosis is increased at least 100-fold in patients who have an identified deficiency of one of these anticoagulant proteins. Pabinger and Schneider reported that 80% of patients who had thrombophilia and mesenteric vein thrombosis had a history of a DVT or PE. Studies of patients with mesenteric vein thrombosis reported in the last decade have discovered prothrombotic disorders in 45% to 82% of cases. In these studies, factor V Leiden was found in 18% to 25% of cases and the prothrombin 20210 mutation in 25% to 45% of cases. Combined defects were seen in more than one-third of cases. Recently, the JAK2 mutation has been detected in 17% of patients with SVT; most had no other evidence of myeloproliferation (e.g., by blood counts) at the time of the thrombosis. A series reported by Acosto and colleagues noted thrombophilia in 67% of cases, a local factor such as surgery or inflammation in 25%, cancer in 24%, and oral contraceptive use in 6%. Thrombosis of the larger distal portions of the mesenteric vein are more commonly secondary to local factors (e.g., malignancy), while those originating from the smaller branching vessels are more commonly related to a prothrombotic state.

Although mesenteric vein thrombosis may appear to occur spontaneously in patients with thrombophilia, an additional provocation can be detected in about one-half of cases. More likely than not, this accounts for the fact that older reviews cite such causes of mesenteric vein thrombosis as cirrhosis, heart failure, intraabdominal malignancy, peritonitis, intraabdominal abscess, abdominal trauma, and abdominal surgery. Malignancy is identified in in 4% to 16% of cases of acute mesenteric vein thrombosis.

Although these provocations may well induce thrombosis in hypercoagulable patients, they prove to be the sole cause in only a minority of cases. Synergy between an underlying hypercoagulable disorder and a provocation is frequent. For example, several cases have been observed in which patients with APLS whose long-term oral anticoagulant therapy had unwisely been held in preparation for colonoscopy with polypectomy developed abdominal pain several days after the procedure and were subsequently given a diagnosis of mesenteric vein thrombosis. The unopposed hypercoagulable disorder coupled with inflammation and bacterial showering from the biopsy site may have provoked thrombosis in these high-risk patients.

Signs and Symptoms

Symptoms of mesenteric vein thrombosis are vague and nonspecific. Abdominal pain is usually of insidious onset. The patient cannot find a position or maneuver that makes the pain disappear. In contrast to most abdominal catastrophes, bowel movements continue, and frequently the patient continues to eat. Nausea is seen in less than one-half of cases. Typically, the pain continues for days to weeks, and several medical evaluations may yield no diagnosis. When transmural ischemia occurs, gastrointestinal bleeding, perforation, and peritonitis may ensue. Hematochezia, hematemesis, or melena occurs in only 15% of patients, although occult blood may be detected in one-half of cases. Fever is rare and when present, typically low grade. Most commonly what at first appears as gastrointestinal bleeding is actually a sloughing of the involved mucosa. This has been called currant jelly stool, and it is not typical melena but shiny, purple, and very mucoid. Because it is not a true hemorrhage and its pathophysiology is thrombosis rather than hemorrhage, it is strongly advisable to continue anticoagulant therapy. When intra-abdominal disease is the main underlying cause, thrombosis usually occurs in proximal large vessels and then extends distally; however, the reverse is often true when underlying thrombophilia is the main cause.

Abdominal pain appears to be far worse than one can account for based on physical examination findings. Rebound tenderness is not present unless the bowel is infarcted. Laboratory data show hemoconcentration, and white cell count is usually in the range of 15,000 to 30,000/µL. The disrupted mucosal barrier may lead to translocation of bacteria in the abdominal cavity and subsequent sepsis or multiorgan failure. An elevated lactic acid level, although nonspecific, may imply ischemic bowel. Serial measurements may prove useful when the findings of the abdominal examination are nonspecific. However, normal serum lactate levels do not exclude MVT and an elevated level implies the process is late in the course and is associated with high mortality. Leukocytosis and hemoconcentration are also common findings. Routine abdominal radiographs often show normal findings in the early stages, while thumb printing, which are semiopaque indentations in the bowel lumen, secondary to edema and air, may be seen in the later stages. The natural history of the process is such that if a diagnosis is not made after 10 to 20 days, the intestines may undergo ischemic infarction. At that time, the symptoms evolve into those of a classic surgical abdomen, with rebound tenderness, rigidity, and increased morbidity and mortality due to venous infarction of the intestine. The presence of colonic involvement or short bowel syndrome after surgery strongly predicts worse 30-day mortality.

Diagnosis

Diagnosis is best made by helical CT or CT angiography, which confirms the diagnosis in 90% of cases. Doppler ultrasonography is also accurate but is operator dependent and may be limited by bowel gas. MRI with gadolinium enhancement and MRA are also highly accurate but may be less readily available. Helical CT remains the diagnostic test of choice.

If the patient undergoes surgery before the correct diagnosis has been made, the surgeon usually finds a dusky but not frankly gangrenous intestinal wall (unless through-and-through infarction has taken place) and bounding mesenteric arterial pulses. When ischemia or infarction is present, viability is not as sharply demarcated as in cases of acute arterial infarction. Second-look laparotomy has been advocated as an approach that is useful for minimizing the amount of bowel resected during the initial operation. When the bowel wall is transected, tiny worm-like clots extrude from engorged veins at the edge of the resected bowel in a peculiar but pathognomonic way. Histologic examination shows extensive hyperemia and hemorrhage, and the degree of infarction is determined by the duration of ischemia.

Treatment

The initial and main goal of treatment is to preserve the bowel and prevent intestinal infarction. Surgery is to be avoided as the primary diagnostic method unless frank peritonitis and obvious bowel infarction are detected. A thickened or edematous bowel wall on imaging does not itself imply infarction, but rather indicates engorgement and is a nearly universal finding. Surgical exploration has been replaced by radiologic exploration. Supportive care measures include placement of a nasogastric tube to decompress the bowel and decrease mechanical pressure, empiric broad-spectrum antibiotic therapy, fluid and electrolyte replacement, and pain control. Once the thrombotic cause of the disorder has been discovered, it should be approached vigorously with aggressive anticoagulant therapy, even in the face of hemodynamically stable active gastrointestinal bleeding. In the acute setting, unfractionated heparin may be preferred over longer acting parental agents, as the need for surgery or invasive procedures is often unpredictable. However, others prefer low-molecular-weight heparin (LMWH) even in the initial stages, unless the patient is at high bleeding risk. Thrombolytic therapy has also been used with success. Systemic thrombolysis has yielded excellent results without excessive hemorrhage. Some clinicians have advocated catheter-directed thrombolytic therapy, but this is not a universally accepted approach. Early surgical series that were reported before anticoagulant therapy or modern radiologic imaging became available described a mortality rate of about 65%, which was an improvement over the initial natural history of the disease when treated with only supportive care, for which mortality was estimated to be 95%. If the diagnosis is made at surgery and is followed promptly by anticoagulant therapy, observed mortality drops to approximately 35%. If the diagnosis is made radiographically and the patients are treated promptly with anticoagulants with no surgical intervention, mortality averages about 10%.

Because the pathogenesis of this abdominal catastrophe is thrombosis, it is appropriate that therapy should be directed toward minimizing thrombotic potential. Such an approach is appropriate and also targets, even if inadvertently, thromboses at other sites that the patient may harbor, and of which the practitioner may be unaware. In more modern series, recanalization rates of up to 80% have been noted with anticoagulant therapy, which should now be considered the standard of care in the treatment of mesenteric vein thrombosis. A study involving a cohort of patients with mesenteric vein thrombosis showed that patients had a low risk of recurrence while receiving oral anticoagulation, but the incidence increased after treatment was stopped (mean treatment time, 12 months; range, 5 days to 45 months). Mesenteric vein thrombosis has a high rate of recurrence, which usually takes place within 30 days of presentation and typically occurs at the anastomotic site in patients who have undergone surgery; this further supports the need for early and adequate anticoagulation.

Because of the seriousness of this event and the fact that most patients have already sustained a previous DVT or PE, unless a definitive provocation (e.g., the use of oral contraceptives) is found and reversed, many clinicians favor indefinite anticoagulant therapy with warfarin, even in the presence of esophageal varices, to maintain an international normalized ratio (INR) in the range of 2.0 to 3.0.

Overview

Splenic vein thrombosis ( Box 17.5 ) is often subtle and is rarely recognized at the time of initial thrombosis. However, vague acute abdominal pain and new-onset splenomegaly may lead to imaging studies that show acute splenic vein thrombosis. More often, splenic vein thrombosis is discovered during the evaluation of a patient with vague chronic abdominal pain or chronic splenomegaly or thrombocytopenia, or it may be detected serendipitously during the evaluation of pancreatitis or pancreatic carcinoma. Splenic vein thrombosis may complicate sclerotherapy for esophageal varices.

Causes

Splenic vein thrombosis has been described in all types of hypercoagulable disorders, whether congenital or acquired. However, local intraabdominal events, whether or not they are accompanied by hypercoagulable disorders, directly lead to a significant percentage of splenic vein thromboses. The chief offenders are pancreatitis and pancreatic carcinoma because of the intimate contact of the splenic vein with the pancreas. Pancreatitis is found to be the initiating event that leads to the discovery of isolated splenic vein thrombosis in up to 65% of cases. Splenic vein thrombosis has also been described in association with abdominal trauma and after abdominal surgery. It complicates 11% of all splenectomies but occurs at a much higher rate in patients who have had their spleens removed for hematologic indications. Thus, the condition that warrants splenectomy (e.g., a myeloproliferative disorder) probably is more responsible for the hypercoagulable state than the operation per se. Splenic vein thrombosis has been related historically to splenectomy for immune thrombocytopenic purpura (ITP). That connection now is believed most likely to be due to the association of ITP with APLS, which explains both the ITP and the splenic vein thrombosis.

Signs and Symptoms

A common presentation of chronic splenic vein thrombosis is variceal bleeding, especially from the stomach and lower esophagus. Varices develop in 17% to 55% of patients with isolated splenic vein thrombosis. When varices occur in the setting of normal serum liver function test results and no hepatomegaly but isolated splenomegaly, the diagnosis of splenic vein thrombosis should be strongly considered. This syndrome has been referred to as left-sided portal hypertension. Up to 71% of patients with splenic vein thrombosis develop splenomegaly, although few show significant cytopenias. Because removal of the spleen relieves venous collateral outflow, splenectomy cures the portal hypertensive gastropathy, and esophageal hemorrhage usually ceases at this time as well. Prophylactic banding of esophageal and gastric varices should be considered, especially if anticoagulant therapy is to be initiated. Splenectomy remains a treatment choice when variceal bleeding is frequent, but it is controversial as a means of prophylaxis with non-bleeding varices.

Diagnosis

The signs and symptoms of splenic vein thrombosis are nonspecific. The diagnosis is most commonly made through ultrasonography, CT, or MRI.

Treatment

Should the diagnosis of acute splenic vein thrombosis be made, anticoagulation should be initiated, not only to stop the process but also to limit potential propagation of thrombosis into the mesenteric and portal veins. Long-term anticoagulant therapy is indicated, particularly if the patient has an underlying hypercoagulable condition. Most practitioners would maintain the INR at 2.0 to 3.0 if using warfarin. Continued follow-up with the hematologist is also appropriate because a significant number of patients with idiopathic splenic vein thrombosis who do not have manifestations of a myeloproliferative disorder at the time of the thrombosis do develop them over time; for this reason, evaluating for the JAK2 mutation should be considered in all cases.

Overview

Portal vein thrombosis ( Box 17.6 ) refers to thrombosis in the trunk of the portal vein or its intrahepatic branches. PVT typically occurs in association with cirrhosis or malignancy of the liver, but it may occur without associated liver disease, in which case the term extrahepatic portal venous obstruction has been applied. PVT is often classified as acute or chronic, extra or intrahepatic, and occlusive or nonocclussive. PVT has been noted in up to 35% of cases of decompensated cirrhosis and up to 16% of cases of compensated cirrhosis. Worldwide, PVT is an important cause of prehepatic portal hypertension and variceal bleeds. A recent meta-analysis suggested PVT significantly adversely affected mortality and hepatic decompensation in patients with cirrhosis, but this continues to be debated.

As a consequence of portal vein obstruction, the liver loses about two-thirds to three-quarters of its blood supply, but given compensatory increase in supply from the hepatic artery, liver dysfunction is not typically significant. As a compensatory response, the hepatic artery vasodilates to increase blood flow. In addition, there is rapid development of venous collaterals bypassing the obstruction, usually within days to weeks. The majority of patients with chronic PVT develop esophageal varices, but some develop splenomegaly and pancytopenia. Gastroesophageal varices may be detectable as early as 1 month after acute PVT; thus early endoscopy with prophylactic banding should be considered. Biliary strictures, termed portal biliopathy, are common and occur when pericholedochal and periportal collaterals cause external compression on or ischemia in the bile duct system.

Similar to splenic vein thrombosis, PVT usually is not recognized in its acute phase, although when acute, it may manifest as intestinal congestion leading to abdominal pain, diarrhea, rectal bleeding, and, in its most severe forms, life-threatening intestinal ischemia and infarction. Chronic PVT is more commonly associated with portal hypertension, esophageal varices, and bleeding. It is characterized by relatively painless increasing splenomegaly and ascites without concomitant worsening of hepatic function. PVT appears to have been described first by Balfour and Stewart in 1869.

PVT has been described in the recent literature in association with most of the acquired or congenital hypercoagulable disorders. However, it also continues to be seen in patients with common portal hypertension, especially when caused by hepatic cirrhosis. A review by Amitrano and colleagues revealed an 11% incidence of PVT in 701 patients with cirrhosis examined by Doppler ultrasonography.

Causes, Signs, and Symptoms

Almost one-half of patients in whom PVT was detected remained asymptomatic. It is likely that the two risk factors of hypercoagulability and portal hypertension act synergistically in the genesis of PVT. Infectious, inflammatory, and malignant conditions are the most common local risk factors. Pancreatitis, pancreatic cancer, and intraabdominal surgery have all been described as precipitating factors. Cases previously labeled as idiopathic have been shown to be associated with thrombophilic mutations in 60% of patients, with an additional local provocation in 30% of cases. Because protein C, protein S, and ATIII levels are low in patients with liver disease, it is likely that deficiencies in these clotting factors represent a secondary phenomenon, especially if levels of all three are depressed. In patients with liver disease, it has been suggested that a value below 0.7 for the ratio of protein C antigen, protein S antigen, or ATIII value to [(factor II + factor X)/2] is highly suspicious for a primary deficiency rather than a secondary phenomenon. A recent meta-analysis suggested that antithrombin III, protein C, and protein S levels were not related to the development of PVT in cirrhosis. A congenital or acquired thrombophilic condition was detected in 37% of patients with PVT in a case-controlled multicenter trial. In other series, approximately 70% to 80% of patients with PVT were noted to have thrombophilic mutations. In a study of patients with cirrhosis, factor V Leiden and the prothrombin 20210 gene mutation were found in 13% and 35%, respectively, of patients with PVT, but in only 8% and 2.5%, respectively, of those without PVT. A systematic review and meta-analysis showed an association of both factor V Leiden and the prothrombin 20210 mutation in PVT both with and without cirrhosis. On the other hand, in a review by Pinto and coworkers, hereditary thrombophilia did not seem to play a significant role in PVT in children and adolescents. However, in an Egyptian study, El-Karaksy and colleagues identified hereditary thrombophilia in 62% of 40 children with PVT investigated at a single center. Factor V Leiden (present in 30% of patients), protein C deficiency (present in 28%), and the prothrombin 20210 gene mutation (present in 15%) were the most common conditions diagnosed. Other clinicians argue that even in patients with cirrhosis, thrombophilia likely plays a contributing role and should be investigated.

Other causes of PVT include intraabdominal neoplasia, especially carcinoma of the pancreas; infection, particularly spontaneous bacterial peritonitis; and abdominal trauma, including surgery. In addition, sclerotherapy for esophageal varices has been implicated in several cases. A local factor can be identified in at least 30% of patients. PVT may complicate neonatal umbilical vein catheterization, but this will not be discussed further. Hypercoagulability should be considered strongly in patients who have no history of hepatic disease, intraabdominal infection, or an inflammatory or neoplastic process. As with splenic vein thromboses, the MPD are notoriously common in patients with PVT, particularly when they are followed long enough to manifest a previously occult myeloproliferative disorder. In fact, in the Western hemisphere, latent myeloproliferative disease has been reported in 58% of patients with idiopathic PVT, which makes it the most common prothrombotic disorder detected. For this reason, investigation for the JAK2 mutation should be considered in all patients with idiopathic PVT. When splenectomy is performed in patients with MPD, the incidence of postoperative portal vein and mesenteric vein thromboses is approximately 30% for patients not given prophylactic anticoagulation. Accordingly, when splenectomy is undertaken in patients with a known or suspected hypercoagulable or myeloproliferative disease, antithrombotic prophylaxis is strongly indicated.

Diagnosis

The diagnosis of PVT is usually made using noninvasive radiologic methods. With color Doppler techniques, the direction of portal vein blood flow can be determined reliably. Ultrasonography with Doppler imaging has a 98% negative predictive value for PVT, which makes it the imaging modality of choice. Contrast CT and MRI/MRA are equally powerful and may be better at determining extent of thrombosis but are more expensive ( Fig. 17.3 ).