• Open Access

Portal Vein Thrombosis in 33 Dogs: 1998–2011


  • The study was completed at Tufts Cummings School of Veterinary Medicine.

  • Abstract poster presented at ACVIM Forum 2010.

Corresponding author: M. Respess, 200 Westboro Rd, Grafton, MA 01536; e-mail: meghan.respess@tufts.edu



Portal vein thrombosis (PVT) has been reported infrequently in dogs.


To characterize the presentation, associated disease conditions, and outcome in dogs with PVT.


Client-owned dogs with a diagnosis of PVT and a complete medical record.


Records were retrospectively analyzed for presentation, history, physical examination, clinicopathologic data, diagnostic imaging, treatment, and outcome.


Thirty-three dogs were included. Common clinical signs were vomiting, diarrhea, abdominal pain, ascites, and signs of hypovolemic shock. Associated disease conditions included hepatic (14/33), neoplastic (7/33), immune (5/33), and infectious (4/33) diseases, protein-losing nephropathy (3/33), hyperadrenocorticism (2/33), protein-losing enteropathy (1/33), and pancreatitis (1/33). Fourteen dogs were receiving glucocorticoids at the time of diagnosis. Twenty-nine dogs had at least 1 predisposing condition for venous thrombosis, and 11 had 2 or more. Thrombocytopenia (24/33), increased liver enzyme activity (23/33), and hypoalbuminemia (20/33) were common laboratory abnormalities. Clinical syndromes at the time of PVT diagnosis included shock (16/33), systemic inflammatory response syndrome (SIRS), (13/33) and disseminated intravascular coagulation (3/33). Twenty-four dogs had acute and 9 had chronic PVT. Multiple thrombi were found in 17/33 dogs. Nineteen dogs survived to discharge. Dogs treated with anticoagulant therapy were more likely, whereas those with acute PVT, multiple thromboses or SIRS were less likely to survive.

Conclusions and Clinical Importance

Hepatic disease is a common pre-existing condition in dogs with PVT. PVT should be considered in dogs with risk factors for venous thrombosis presenting with abdominal pain, ascites, and thrombocytopenia. Studies evaluating anticoagulant therapy in the management of PVT are warranted.


chronic hepatitis


congenital portosystemic shunt


computed tomography


hepatic arteriovenous fistula


immune-mediated hemolytic anemia


protein-losing enteropathy


protein-losing nephropathy


portal vein thrombosis


systemic inflammatory response syndrome





Portal vein thrombosis (PVT) refers to partial or total obstruction of blood flow caused by thrombosis within the extrahepatic portal venous system. Factors contributing to the development of PVT in humans include inflammatory disorders such as pancreatitis and cholecystitis, abdominal neoplasia, chronic hepatitis (CH) or cirrhosis, vascular injury (trauma, surgery, portal hypertension), inherited or acquired prothrombotic disorders, myeloproliferative disease, and systemic inflammatory response syndrome (SIRS).[1-3]

Thrombosis of the cranial vena cava, splenic vein, and pulmonary circulation has been well described in dogs,[4-9] but less is known about dogs with PVT. Information regarding PVT in dogs is limited to a necropsy study of 11 dogs with PVT, another series of dogs with splenic vein thrombosis (SVT) and concurrent PVT (14 cases), and isolated case reports.[4, 10-21] These studies have identified a number of concurrent prothrombotic conditions including pancreatitis, immune disease, neoplasia, SIRS, DIC, and administration of glucocorticoids,[4, 10-21] but detailed information regarding the clinical syndrome of PVT in dogs remains limited. The aim of this retrospective study was to better describe the clinical presentation, associated diseases, diagnostic imaging findings, and outcome associated with PVT in dogs.

Materials and Methods

Criteria for Case Selection

The medical record system at the Foster Hospital at Tufts Cummings School of Veterinary Medicine was reviewed to identify dogs with PVT from 1998 to 2011. Dogs were included if PVT was confirmed by abdominal ultrasound examination (US), computerized tomography (CT), or direct visualization at surgery or necropsy, and a complete medical record was available.

Laboratory Testing

Antithrombin and D-dimer concentrations were measured at the Comparative Coagulation Section at the New York State College of Veterinary Medicine. Prothrombin time (PT), activated partial thromboplastin time (aPTT), and kaolin-activated thromboelastography (TEG) analysis were performed at the Clinical Pathology Laboratory at Tufts University as previously reported.[22]

Data Collection

Signalment, history, presenting complaint, physical examination findings, clinicopathologic data, diagnostic imaging findings, treatment, outcome, and final diagnosis were recorded. Necropsies were performed by a board-certified pathologist. A board-certified radiologist reviewed US images and recordings (O.T.). Thrombus location, echogenicity and homogeneity, and degree of luminal filling were recorded, as was the presence of peritoneal effusion, additional thrombi, portosystemic shunting or extension of PVT into the mesenteric vasculature. Pulsed and color Doppler findings were available inconsistently, and therefore portal flow dynamics was not included.

By means of previously published criteria (see )[4] concurrent conditions were recorded including immune disease, protein-losing nephropathy (PLN), protein-losing enteropathy (PLE), disseminated intravascular coagulation (DIC), hyperadrenocorticism (HAC), SIRS, hepatopathy, and neoplasia. Diagnosis of infection was based on positive culture or serology results, or the presence of intracellular bacteria on cytology. History of prior abdominal surgery involving the portal vasculature was recorded. One or more of these conditions were identified in each dog based on these criteria.

Dogs with acute abdominal pain, hypovolemic shock, or both were classified as having acute PVT, and others as chronic.[1-3] Diagnosis of hypovolemic shock was based on presence of tachycardia, tachypnea, mucous membrane color and capillary refill time, abnormal pulse quality, decreased mentation, decreased base excess or increased blood lactate concentration.[23] Survival of dogs with PVT was defined as discharge from the hospital.

Statistical Analysis

The median and range were calculated. Distribution of data was examined by box and whisker plots. Differences in continuous biochemical variables between dogs with different disease outcomes (survival versus nonsurvival) were analyzed by the non-parametric Wilcoxon Rank-Sum test; a P-value ≤.05 was applied. Associations between categorical variables and disease outcome (survival versus nonsurvival) were determined by Pearson's Chi-square test or Fisher's exact test if an expected value was <5. A 2-sided P-value ≤.05 was applied. Statistical analyses were completed by commercial software.1



Thirty-three dogs were identified, including 17 males, (4 intact) and 16 females (3 intact), with a median age of 7.3 years (range, 1 month to 13 years). Several breeds were represented with most being medium to large-sized breeds with a median weight of 29.6 kg (range, 1.4–66.2 kg). Diagnosis of PVT was determined by abdominal US (27/33), necropsy (4/33), direct visualization at surgery (1/33), or CT angiogram (1/33).

History and Physical Examination Findings

Pre-existing conditions previously reported in association with venous thrombosis in dogs were identified in 29/33 dogs (Table 1). More than 1 condition was identified in 21/29 dogs. Fourteen of 29 dogs were diagnosed with hepatic disease, including CH (8/14), hepatic neoplasia (4/14) (cholangiocarcinoma, undifferentiated sarcoma, neuroendocrine tumor, unclassified cystic liver mass), congenital portosystemic shunt (CPSS) (2/14), and congential hepatic arteriovenous fistula (HAV) with multiple acquired portosystemic shunts (1/14). One dog with intrahepatic CPSS had a longstanding coil within the shunting vessel and had recently undergone splenectomy whereas another dog with an extrahepatic CPSS had concurrent infectious disease. Nonhepatic neoplasms were identified in 3/29 dogs and included a pancreatic carcinoma (1/3), gastrointestinal lymphosarcoma (1/3), and splenic sarcoma (1/3). Immune disease was diagnosed in 5/29 dogs including immune-mediated hemolytic anemia (IMHA) (3/5) and immune-mediated thrombocytopenia (2/5). Four of 29 dogs had documented infectious disease including septic arthritis (1/4), chronic pyoderma (1/4), urinary tract infection (1/4), and concurrent infection with Anaplasma phagocytophilum and Hemotrophic mycoplasma (1/4). Previous abdominal surgery or invasive procedures directly affecting portal circulation were performed in 6/29 dogs and included splenectomy (4/6), coil occlusion for intrahepatic CPSS 5 years earlier (1/6), and liver lobectomy 2 days earlier (1/6). Splenectomy procedures ranged from 1 day to 5 years before, with most occurring within 3 months. Hind limb amputation also was performed 3 days before PVT diagnosis in 1/6 dogs having undergone prior abdominal surgery or invasive procedure. Fourteen of 29 dogs were receiving glucocorticoids, either prednisone (13/14; 0.875 mg/kg/d; range, 0.1–6 mg/kg/d) or dexamethasone (1/14; 0.25 mg/kg/d). Seven dogs were receiving anticoagulant therapy for a pre-existing condition including low molecular weight heparin2 (2/7), aspirin3 (2/7), unfractionated heparin4 and warfarin5 (1/7), clopidogrel6 (1/7), and urokinase7 infusion (1/7).

Table 1. Diseases and conditions in dogs with PVT.
ParameterNumber (%)
Glucocorticoids14 (42)
Hepatic disease14 (42)
SIRS13 (39)
Previous invasive procedure6 (18)
Immune5 (15)
Infectious4 (26)
Nonhepatic neoplasia3 (9)
PLN3 (9)
Hyperadrenocorticism2 (6)
PLE1 (3)
Pancreatitis1 (3)

The most common clinical signs were hypovolemic shock (16/33), vomiting (15/33), abdominal pain (13/33), diarrhea (12/33), and abdominal distension (11/33) (Table 2). Thirteen of the 16 dogs with shock also met the criteria for SIRS. Twenty-four of 33 dogs had acute PVT. Nine of 33 dogs had chronic PVT. Development of PVT occurred during hospitalization in 4 dogs undergoing treatment for PTE and IMHA, aortic thromboembolism, septic arthritis, or splenic vein thrombosis (SVT). Evidence of gastrointestinal hemorrhage (eg, melena, hematochezia) was present in 10/33 dogs; 2/10 dogs were receiving anticoagulant treatment with clopidogrel7 or urokinase.8 One of these dogs had multiple mesenteric thrombi identified on ultrasound examination and at necropsy, and the other had diffusely distended small intestines and bowel edema on ultrasound examination.

Table 2. Clinical characteristics of dogs with PVT.
CharacteristicAll PVT N = 33Acute PVT N = 24Chronic PVT N = 9
  1. a

    Identified on physical examination, abdominal ultrasound examination, or both.

Abdominal pain13130
Palpable fluid wave1165
Pitting edema312
Central nervous system  signs220
Peripheral lymphadenopathy220
Cardiopulmonary arrest110

Clinical Pathology

Clinicopathologic findings are summarized in Table 3. Complete blood counts were available for 32/33 dogs, with thrombocytopenia (24/32), anemia (15/32), and leukocytosis (14/32) being the most common abnormalities. Leukopenia was present in 4/32.

Table 3. Statistical evaluation of selected laboratory results in dogs with PVT.
Median (Range)No.LowHighMedian (Range)No.LowHighReference RangeP-Value
PCV (%)37 (13–56)198/191/1933 (18–59)148/141/1437–55.729
WBC (×103/μL)12.6 (5.68–678)192/198/1913.8 (3.3–26.5)132/136/136–17.699
Platelets (×103/μL)105 (11–434)1915/190/1957 (2–1000)1212/131/13200–500.051
TP (g/dL)5.9 (4–8.1)196/192/134.7 (2–6.1)148/130/135.2–7.2.059
BUN (mg/dL)14 (5–75)194/192/1922 (6–102)141/144/148–33.129
Creatinine (mg/dL)0.8 (0.4–5.5)192/192/191.1 (0.5–5.5)120/121/120.5–1.5.501
Albumin (g/dL)2.8 (1.8–4.6)1910/191/192.2 (0.9–3.0)129/120/123–4.2.281
Globulin (g/dL)2.9 (2.0–5.1)193/191/192.3 (1.3–3.7)124/120/122.3–4.2.071
Alk Phos (U/L)314 (75–3549)190/199/19171 (62–7161)120/124/1220–320.429
ALT (U/L)425 (24–7623)190/1911/19339 (5–1860)121/129/1210–95.792
AST (U/L)78 (20–15,162)190/1912/19113 (24–7917)120/127/1215–52.301
GGT (U/L)14 (3–228)170/1711/176 (1–212)120/123/121–10.548
Total bilirubin (mg/dL)0.3 (0.1–6.3)190/198/190.5 (0.1–16)120/125/120.1–0.5.413

Biochemical profiles were evaluated in 31/33 dogs. Twenty-three of 31 dogs had increased serum liver enzyme activity or hyperbilirubinemia. Median increases in serum transaminases alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, and gamma glutamyltranspeptidase activities were 3.7-, 1.8-, 1.1-, and 1.1-fold, respectively. Total serum bilirubin concentrations were increased in 13/31 dogs, with a median of 0.45 mg/dL (range, 0.1–16 mg/dL). Hypoalbuminemia (19/31), hypoproteinemia (14/31), and hypoglobulinemia (7/31) also were present.

Twenty-five dogs had PT and aPTT determined at diagnosis, with mild to moderate prolongation (<50% above the upper limit of the reference range) in 11/25. Three of the 11 dogs currently were receiving anticoagulant treatments including low-molecular heparin,2 aspirin,3 unfractionated heparin,4 warfarin5 or some combination of these. One dog that had been receiving aspirin for 3 weeks (0.6 mg/kg PO q24 h) had mild prolongation of PT and aPTT. Antithrombin activity (AT) was obtained in 6 dogs. Two dogs with PLN and 1 dog with both CH and PLE had AT activities of 57, 66 and 70% (reference range, >80%), respectively. D-dimers were elevated (>250 ng/mL) in 2 of 3 dogs tested. Based on the data obtained, 3 dogs met the criteria for DIC.[24] TEG was evaluated in 7 dogs. Based on increased G values (median, 12.9 Kd/s; range, 11.3–13.7 Kd/s; reference range, 5.7–6.9 Kd/s), 3 dogs were hypercoagulable, including 2 acute and 1 chronic PVT. The remaining 4 dogs were hypocoagulable based on decreased G values (median, 3.45 Kd/s; range, 2.4–4.9 Kd/s) and included 3 dogs with thrombocytopenia and 1 receiving aspirin3 (0.6 mg/kg PO q24 h) for 3 weeks before diagnosis. No dog tested had a normal TEG tracing.

Urinalysis results were available for 14/33 dogs. Common abnormalities included bilirubinuria (10/14) and proteinuria (6/14). Two dogs with proteinuria had urine protein-to-creatinine ratios of 3.2 (normal <0.5). Abdominal fluid analysis was available for 16/26 dogs, and 15/16 had a low protein (<2.5 mg/dL) modified transudate and 1 dog had hemoabdomen.


Of 31 dogs undergoing US, PVT was successfully identified on the initial examination in 20/31 (64%) as compared to 27/31 (87%) when a more experienced operator performed the examination. Eleven thrombi were located at the level of the porta hepatis. Thrombus echogenicity was classified as homogeneous (10/31) or heterogeneous (11/31) and hyperechoic (9/31) or poorly echoic with an echogenicity close to that of the vessel lumen (11/31). A hyperechoic lining surrounding the thrombus was noted in 5/9 hyperechoic thrombi. Luminal filling of the portal vein by the thrombus was estimated at <50% (2/24), 51–75% (5/24), and >75% (17/24) in 24 dogs in which these data were available. Focal distention of the portal vein at the level of the thrombus was noted in 12/31 dogs. Peritoneal effusion was present in 25 dogs, and was described as marked (8/25), moderate (13/25), or mild (4/25). Some dogs with effusion had pancreatic edema (6/25) and increased mesenteric fat echogenicity (6/25). Hepatic parenchymal abnormalities were noted in 20 dogs including heterogenous and hyperechoic (10/20), nodular (7/20), solitary mass (5/20), and hypoechoic (3/20). Gastrointestinal abnormalities included both focal and diffuse thickening consistent with edema (9/31) and intestinal wall mass (1/31). Vascular abnormalities included extrahepatic CPSS (1/31), HAV (1/31), and multiple acquired portosystemic shunts (1/31). Additional thromboses were identified in the splenic vein (13/31), mesenteric vein (4/31), pancreaticoduodenal vein (2/31), caudal vena cava (1/31), and previously identified ATE (1). The 4 cases in which PVT was not seen on US had a diagnosis established at necropsy (2/33), surgery (1/33), or on CT (1/33). Two dogs that did not have an US examination had a diagnosis established at necropsy.

Seven dogs underwent necropsy examination. Concurrent thrombi were identified in 5 dogs in the splenic (5/7), pulmonary (3/7), or mesenteric veins (3/7). Mircrothrombi were identified in the myocardium (1/7) and pancreas (1/7). Other necropsy findings included hepatic cirrhosis and nodular regeneration (2/7), hepatocellular necrosis (1/7), and acute pancreatic necrosis (1/7).


Management ranged from supportive care including anticoagulant or thrombolytic therapy to surgical intervention. Three dogs with acute PVT underwent surgical embolectomy, with only 1 dog that had venotomy and arterial-venous fistula ligation surviving to discharge.

Nineteen dogs were treated with some combination of anticoagulant, antiplatelet or thrombolytic management. Single agent modalities were used in 8 dogs including low-dose aspirin5 (0.5 mg/kg PO q24 h), low molecular weight heparin2 (100–150 IU/kg SC q8–24 h), and unfractionated heparin3 (200 IU/kg SC q12–24 h). The remaining dogs were treated with a combination of these drugs or warfarin4 (0.5 mg/kg PO q24 h) and clopidogrel3 (2.5–3.75 mg/kg PO q24 h). Thrombolytic medications included urokinase4 (1) (4,400 IU/kg/h for 12 hours) and streptokinase8 (1) (9,000 IU/kg IV loading dose over 1 hour then 4,500 IU/kg/h over 12 hours) (1).


Nineteen of 33 dogs survived to discharge, including 11/24 with acute PVT and 8/9 with chronic PVT. Of the 14 nonsurvivors, 12/14 were euthanized for failure to respond to management or grave prognosis and 2 died of circulatory collapse. On initial presentation, both hypovolemic shock and SIRS were identified in 10/14 of the nonsurvivors. Follow-up US examinations were available for 7 dogs and were performed at the clinicians’ discretion. One dog with chronic PVT was unchanged at 7 months, and in another PVT had resolved. Complete resolution was present in 2 dogs with acute PVT, 1 by day 9 (with PVT and mesenteric thrombi), and the other by 2 months (concurrent PVT and SVT). One dog had evidence of SVT 10 days after PVT, and multiple acquired portosystemic shunts and ascites were found in 2 others at 8 and 12 months after diagnosis of PVT. On a 2 week follow-up US for SVT, an additional dog was found to have PVT.

Risk Factors for Mortality

The following factors were analyzed for risk of mortality: acute versus chronic PVT, the presence of pre-existing liver disease, neoplasia, moderate to marked abdominal effusion, SIRS, ≥1 risk factor for thrombosis, treatment with anticoagulation agents or glucocorticoids, previous abdominal surgery or invasive procedure, and laboratory data including platelet and white blood cell count, serum total protein, albumin, globulin, BUN, creatinine, ALT, ALP, AST, GGT, and total bilirubin (Tables 3 and 4). Dogs with acute PVT (= .047), SIRS (= .012), or thrombi at >1 location (= .05) were more likely to be non-survivors whereas those that received anticoagulant medications were more likely to survive to discharge (= .029). The presence of decreased platelets (= .051) approached significance in predicting non-survival. Dogs with acute PVT were more likely to have >75% occlusion of the portal vein (= .003) than those with chronic PVT at the time of diagnosis.

Table 4. Statistical evaluation of survival to discharge in dogs with PVT.
Parameter EvaluatedSurvivors N = 19Nonsurvivors N = 14P-Value
Hepatobilary disease86.97
Moderate to marked  ascites146.073
Multiple thrombi (>1)710.050
1 or > thrombotic risk  factors1010.28


The results of this study suggest (i) that underlying diseases associated with PVT are similar to those reported previously for venous thrombosis except a higher percentage of dogs with PVT have hepatic disease; (ii) based on clinical presentation and laboratory evaluation, a high clinical suspicion of PVT should be considered in dogs presenting with abdominal pain, ascites, thrombocytopenia, increased serum liver enzyme activity, and hypovolemic shock; (iii) dogs with chronic PVT are more likely to survive than dogs with acute PVT; (iv) US is a specific and relatively sensitive, but operator-dependent, method to diagnose PVT; (v) the presence of thrombi at locations other than the portal vein is associated with nonsurvival; and (vi) medical management with anticoagulant agents, thrombolytic agents or both may improve outcome.

At least 1 potentially predisposing prothrombotic condition was identified in 87% of dogs in this study, and 63% had ≥2 predisposing conditions. Concurrent conditions included glucocorticoid therapy, hepatic disease, SIRS, previous abdominal surgery or invasive procedure, and immune disease, and were similar to those previously reported in dogs with venous thrombosis in the cranial vena cava, lung, and splenic vein.[4-21] One difference in dogs with PVT was a high percentage of concurrent hepatic disease (14/33, 42%). Although hepatobiliary disease has been reported in dogs with PVT,[4, 10, 15, 17-21] only 1/11 and 2/14 dogs in the 2 largest case series were reported to have concurrent hepatopathy.[4, 10] This discrepancy may reflect inclusion criteria, one being a necropsy study and the other reporting PVT in association with SVT.

The mechanism of hypercoagulability in dogs with liver disease is unknown.[1-3, 25] The occurrence of PVT in association with hepatic disease has been reported in cats[26] and with cirrhosis in humans.[1-3] In addition, hyperbilirubinemia and increased serum ALP activity have been reported as risk factors for thromboembolism in dogs with IMHA.[6] In the present study, at least 1 additional predisposing factor for thrombosis was found in 12/14 dogs with hepatic disease, including glucocorticoid therapy (7/14) and previous abdominal surgery (6/14). In humans with hepatic disease, hypercoagulability is promoted by decreases in proteins C and S, and AT.[1-3] In addition, in humans, the presence of portal hypertension or previous surgical procedures that injure the portal vasculature predispose to thrombosis[1-3] caused by vascular shear stress, endothelial cell activation, and increased release of Factor VIII and vWF.[27, 28]

PVT is a rare postoperative complication after correction of congenital vascular disease in dogs,[15, 17] usually occurring when postoperative prothrombotic complications such as pancreatitis or sepsis are present. Likewise, the 3 dogs in this study with congenital vascular disease also had other associated prothrombotic conditions including recent splenectomy, the presence of an intravascular coil, portal hypertension or infectious disease. Possible mechanisms for hypercoaguability with congenital vascular disease include altered blood flow in the portal circulation causing sheer stress, local trauma associated with invasive procedures, systemic changes in blood coagulants or some combination of these. Dogs with CPSS have been shown to have increased Factor VIII and decreased Protein C activity.[25, 27]

Many dogs with PVT were receiving glucocorticoids (42%). Glucocorticoids may contribute to hypercoagulability,9,[29, 30] and glucocorticoid therapy previously has been associated with PVT and SVT in dogs.[4, 10] Increases in plasminogen activator inhibitor, Factor VIII, and vWF have been demonstrated in human patients receiving glucocorticoids.[29] The mechanism for hypercoagulability in dogs, however, has been poorly characterized. Increases in procoagulant Factors II, V, VII, IX, X, XII, and fibrinogen, and decreased antithrombin concentration, were observed in a dog model of hyperadrenocorticism in humans.[31] Clinicians should carefully consider the risk for thrombosis in dogs, particularly those with other underlying prothrombotic conditions when contemplating the use of glucocorticoids.

Clinical signs of PVT in this population of dogs ranged from none to cardiovascular failure, collapse, and death. In humans, PVT is classified as acute or chronic.[1-3] Chronic PVT is often found during investigation of complications associated with portal hypertension, or even serendipitously.[1-3] Alternatively, acute PVT is associated with the onset of abdominal pain, and, with thrombus propagation to the mesenteric vasculature, signs of intestinal infarction including melena, hematochezia, or both may be present and can progress to shock, SIRS, and hemodynamic instability. Two dogs with gastrointestinal bleeding were receiving anticoagulant or thrombolytic therapy. Both dogs had evidence of intestinal ischemia on ultrasound examination or necropsy; therefore, bleeding tendencies were likely the result of underlying thrombosis, but it also may have been associated with anticoagulant or thrombolytic therapy. Acute PVT was distinguished from chronic PVT based on the basis of abdominal pain, hypovolemic shock or both. Shock, SIRS, or both were identified in over half of the dogs in this study. SIRS was associated with a greater risk for non-survival. Nineteen of 24 dogs with acute PVT also had ascites. Our findings suggest that an intensive investigation for abdominal thrombosis and PVT should be conducted in dogs presenting with abdominal pain, ascites, hypovolemic shock, and SIRS, especially if the dog has a known pre-existing prothrombic condition. Moreover, the occurrence of these signs in dogs with concurrent hepatic disease, especially those on glucocorticoid therapy, should not be confused with decompensating chronic liver failure.

Ascites was the most common presenting sign in dogs with chronic PVT (6/9). These dogs had a variety of underlying diseases including CH, chronic renal failure, HAV, neoplasia, septic arthritis, and HAC. Dogs with chronic PVT were more likely to survive to discharge than were dogs with acute disease, possibly reflecting the compensated state of their portal vasculature.

Most dogs in this study had increases in serum liver enzyme activity, possibly reflecting a primary liver disorder, glucocorticoid induction, mild ischemic injury attributable to obstruction of portal blood flow, or some combination of these. Mild to moderate increases in serum liver enzyme activities with mild serum hyperbilirubinemia were present in many dogs, indicating mild functional liver impairment, sepsis, or hemolysis. Despite these indications of hepatic impairment, signs consistent with acute hepatic failure (defined by prolonged PT, aPTT, or both and the presence of encephalopathy) were not evident in this study suggesting that primary hepatic disease was not the cause for mortality in this population of dogs. The fact that the degree of increases in liver enzyme activities and bilirubin were not associated with survival supported this conclusion.

Thrombocytopenia was present in 75% of dogs. Thrombocytopenia is a common laboratory abnormality in dogs with thrombosis of the cranial vena cava (13/17)[5] and PTE (8/22)[7] and is associated with a higher risk of thromboembolism in IMHA.[6] In this study, dogs with lower platelet counts showed a tendency (= .051) for nonsurvival. This observation may reflect severe platelet consumption caused by excessive clotting in nonsurvivors. Dogs with thrombi elsewhere were less likely to survive and had lower platelet counts (median, 58,000/μL) than dogs with thrombi confined to the portal vein (median, 135,000/μL).

The current study shows that abdominal US examination with color flow Doppler correctly identified PVT in (27/31) 87% of dogs when an experienced operator performed the examination. Seven PVT, however, were missed on initial US evaluation. The reasons included poor visualization of the portal vein attributable to deep-chested conformation, presence of large amounts of ascites, or interference by gas or food in the stomach. Decreased echogenicity of the thrombus also hindered visualization. Prior or concurrent anticoagulant treatment also may have altered thrombus morphology and have interfered with detection. Successful visualization may be enhanced when the operator is prompted to thoroughly evaluate the portal venous system due a high clinical index of suspicion for thrombosis. Abdominocentesis and evaluation of portal vein dynamics also may improve recognition.[14] Identification of PVT on US examination should prompt examination for additional thrombi or propagation of the PVT to the mesenteric vasculature. Concurrent thrombosis in splenic or mesenteric vessels was present in (17/33) 51.5% of dogs in this study. In humans, CT angiography is the gold standard for imaging thrombus propagation and has the added advantage of allowing identification of intestinal ischemia.[1-3]

In humans, anticoagulation is recommended for acute PVT to prevent thrombus propagation and to promote recanalization of the portal vein.[2-6] Initial treatment with low molecular weight heparin for 1–2 weeks is followed by 3–6 months of vitamin K antagonist therapy.[1-3] In the absence of intestinal infarction, alternative therapies such as surgical embolectomy or pharmacologic thrombolysis provide no benefit over anticoagulation, and actually may carry more risk than benefit to the patient.[1-3] Anticoagulation in humans with chronic PVT is more controversial. Long-term anticoagulation is recommended in patients with a permanent noncorrectable risk for venous thrombosis. The decision to anticoagulate other patients is not as straightforward. Humans with chronic PVT can develop portal hypertension that may cause recurrent bleeding from esophageal varices and hypersplenism with pancytopenia.[1-3] Because dogs with portal hypertension do not develop these complications,[27] the risk of anticoagulation may outweigh the benefits. Although we demonstrated an association of anticoagulation with survival in this study, this observation is difficult to interpret. First, the population is small. Second, as a retrospective analysis, treatment was not standardized. Finally, serial laboratory testing to assess the true efficacy of anticoagulant therapy was not done. The survival effect of anticoagulation may in part be attributable to the fact that these dogs received more aggressive intensive care. Recommendations for treatment of dogs with either acute or chronic PVT will require prospective evaluation of the benefits and long-term consequences of both anticoagulation and thrombolysis.

There are several limitations of our study including small population size and the retrospective design with reliance on medical records for extraction of information. These factors along with the lack of serial imaging and follow-up data limited our ability to prognosticate and draw conclusions regarding treatment. One difficulty in particular was the determination as to whether SIRS is a pre-existing condition or a clinical syndrome associated with sudden obstruction to portal blood flow. Also, because identification of chronic PVT may be incidental, many cases may go unrecognized, and the dogs with chronic PVT in this study may represent a selected population of animals in which clinical signs warranted abdominal US examination for a separate reason.

Although PVT is an uncommon occurrence in the dog, clinicians should be aware of its increased risk in selected populations. This study suggests that PVT should be considered in dogs with known prothrombotic conditions such as hepatic disease, abdominal neoplasia, or immune disease presenting with gastrointestinal signs, ascites, and abdominal pain. The presence of thrombocytopenia, increased liver enzyme activities, and hypoalbuminemia should further increase clinical suspicion. Heightened awareness is indicated for dogs receiving concurrent glucocorticoid therapy. Evaluation of coagulation status in dogs with conditions known to predispose to thromboembolism should be considered. Conventional coagulation testing is limited in its ability to assess hypercoagulability. TEG, however, provides a global assessment of hemostatic function and provides evaluation of all stages of coagulation, including primary and secondary hemostasis, and fibrinolysis. TEG has proven to be a reliable predictor of hypercoagulability in dogs with IMHA,[22, 30] PLE,[32] PLN,10,[33] CH9 and neoplasia.[34] Future studies should be aimed at characterization of the natural history of both acute and chronic PVT in the dog, including serial imaging to determine if recanalization or progression occurs. Prospective, randomized clinical trials of anticoagulation in dogs with defined hypercoagulability and risk factors for PVT are indicated.


We acknowledge Melanie Tong, veterinary technician specialist in anesthesia, for assistance in data collection. The study was not funded by a grant.


  1. 1

    Statistix 9.0, Analytical Software, Tallahassee, FL

  2. 2

    Dalteparin, Fragmin, Eisai Inc, Woodcliff Lake, NJ

  3. 3

    Aspirin, Rugby Laboratories Inc, Duluth, GA

  4. 4

    Heparin 1,000USP units/mL, APP Pharmaceuticals LLC, Schaumburg, IL

  5. 5

    Warfarin, Coumadin, Taro Pharmaceuticals Industrial LLC, Haifu Bay, Israel

  6. 6

    Clopidogrel bisulfate, Plavix, Bristol-Meyers Squibb/Sanofi Pharmaceuticals Partnership, Bridgewater, NJ

  7. 7

    Urokinase, Abbokinase, Abbott Laboratories Inc, North Chicago, IL

  8. 8

    Streptokinase, Sanofi-Aventis US LLC, Kansas City, MO

  9. 9

    Rose L, Bedard C, Dunn M. Effect of prednisone administration on TEG parameters in healthy Beagles. J Vet Intern Med 2008;22:738–739 (abstract)

  10. 10

    Hilling KM, Labato MA, DeLaforcade AM, Shaw SP. Documentation of hypercoagulability in protein-losing nephropathy via thromboelasography in 10 dogs. J Vet Intern Med 2009;23:690–691 (abstract)

Appendix 1

  1.  Hepatopathy, including chronic hepatitis or cirrhosis (CH), acute hepatic necrosis, congenital vascular anomalies including portosystemic shunt (CPSS) or hepatoarteriovenous fistula (HAV) identified based on clinical pathology, ultrasound findings, histopathology or a combination of these. Pertinent clinical pathology to identify hepatopathies included elevation in serum transaminases, serum alkaline phosphatase and gamma glutamyl transpeptidase, hyperbilirubinemia in combination with a test of hepatic functional capacity (blood urea nitrogen, albumin, cholesterol or glucose) or hepatic histopathology.
  2.  Immune, including immune-mediated hemolytic anemia (IMHA) based on anemia with autoagglutination or spherocytosis or a positive Coombs test, and immune-mediated thrombocytopenia (IMTP) based on marked thrombocytopenia (<20,000/μL platelets) without identification of an underlying cause.
  3.  Protein-losing nephropathy (PLN), based on proteinuria and a urine protein-to-creatinine ratio ≥3 or urine protein concentration >500 mg/dL with concurrent isothenuria and serum hypoalbuminemia.
  4.  Protein-losing enteropathy (PLE), based on presence of diarrhea, weight loss, anorexia, serum panhypoproteinemia, histopathology or a combination of these.
  5.  Hyperadrenocorticism (HAC), based on the presence of polyuria and polydipsia, weight gain, excessive panting, alopecia, and ultrasound findings, in addition to either low-dose dexamethasone suppression test, ACTH stimulation test, or urine cortisol:creatinine ratio.
  6.  Neoplasia based on histopathology.
  7.  Disseminated intravascular coagulation (DIC), based on the presence of ≥3 of the following plus a concurrent inflammatory disease identified during the hospitalization period: (i) thrombocytopenia (platelets <150,000/μL); (ii) prolonged prothrombin time or activated partial thromboplastin time; (iii) decreased fibrinogen (<90 mg/dL); (iv) decreased antithrombin (AT <80%); (v) positive fibrin degradation projects or positive D-dimers.[24]
  8.  Systemic inflammatory response syndrome (SIRS), based on the presence of ≥2 of the following: (i) fever (temperature > 103°F) or hypothermia (temperature < 100°F); (ii) tachycardia (HR > 20 beats/min); (iii) tachypnea in the absence of panting (respiratory rate > 40 breaths/min); (iv) leukocytosis (WBC > 16,000/μL), leukopenia (WBC < 6,000/μL), or >3% bands.[35]