• Open Access

Thromboelastographic Evaluation of Dogs with Congenital Portosystemic Shunts


  • This study was supported by a grant from the Companion Animal Health Fund at TCSVM.

Corresponding author: Cynthia R.L. Webster, TCSVM, 200 Westboro Road, North Grafton, MA 01536; e-mail: cynthia.leveille-webster@tufts.edu.



On plasma-based assays, dogs with congenital portosystemic shunts (CPSS) have changes in serum concentrations of both pro- and anticoagulant proteins, but how these abnormalities affect whole blood coagulation assays (eg, thromboelastography) are unknown.


To conduct kaolin-activated thromboelastography (TEG) analysis in dogs with CPSS and to compare TEG coagulation status with clinical presentation, routine serum biochemistry, and plasma-based coagulation tests.


Twenty-one client-owned dogs with CPSS confirmed by ultrasound examination or nuclear scintigraphy.


In a prospective study, signalment, clinical presentation, TEG analysis, CBC, serum biochemistry, and hemostatic tests (platelet count, prothrombin time [PT], activated partial thromboplastin time [aPTT], quantitative fibrinogen, antithrombin [AT] activity, protein C [PC] activity, d-dimers, and factor VIII activity) were analyzed in dogs with CPSS.


Dogs with CPSS had significantly shorter K values and increased angle, maximum amplitude (MA), and G values compared with the reference population. On plasma-based coagulation testing, dogs with CPSS had significantly prolonged PT, lower platelet counts, lower AT and PC activities, and increased d-dimers and factor VIII activity. Evaluation of G value defined 9/21 dogs with CPSS as hypercoagulable. These dogs were more likely to have hepatic encephalopathy (HE) than CPSS dogs that had normal coagulation.

Conclusions and Clinical Importance

TEG analysis detected hemostatic abnormalities consistent with a hypercoagulable state in some dogs with CPSS. The presence of a hypercoagulable state was 40 times more likely in dogs with symptomatic HE.


activated partial thromboplastin time




congenital portosystemic shunts


computed tomography


hepatic encephalopathy


protein C


prothrombin time


Tufts Cummings School of Veterinary Medicine



Dogs with congenital portosystemic shunts (CPSS) have changes in serum concentrations of both pro- and anticoagulant proteins. Studies comparing dogs with CPSS to healthy controls have found prolonged prothrombin time (PT) and activated partial thromboplastin time (aPTT), lower platelet counts and deficiencies of procoagulant factors, including factors II, V, VII, and IX.[1-4] Collectively, these studies suggest that CPSS patients may be hypocoagulable.[1-4] However, other reports documenting decreased antithrombin (AT) and protein C (PC) activities and increased factor VIII activity and von Willebrand factor concentration suggest that these patients may be at risk for hypercoagulability.[2-4] Clinically, dogs with CPSS can demonstrate both bleeding and thrombotic tendencies.[2, 3, 5-7] Routine plasma-based coagulation assays fail to take platelet function and all anticoagulant activities into consideration. As a whole blood assay, thromboelastography (TEG) may provide a more global representation of coagulation status in dogs with CPSS than plasma-based assays.

The objectives of this study were to describe the TEG findings in dogs with CPSS and to compare coagulation status as determined by TEG to clinical presentation, serum biochemistry results, and conventional coagulation tests.

Materials and Methods

Study Population

Twenty-one dogs with CPSS confirmed by ultrasound examination or nuclear scintigraphy were prospectively enrolled. Dogs on medications known to affect coagulation and those with concurrent disease conditions suspected to cause hypercoagulability were excluded. The TCSVM institutional review board approved the study.

Hemostatic Analysis

At the time of admission, whole blood for TEG analysis was drawn by peripheral venipuncture with a vacutainer blood collection needle into tubes containing 3.2% sodium citrate to obtain a dilution of blood to sodium citrate of 9:1. Additional blood was drawn into an EDTA tube for a CBC. After a 30-minute rest period, kaolin-activated TEG was performed by a single operator at room temperature. The remaining citrated plasma was stored at −70°C for analysis of PT1,2 aPTT,1,2 quantitative fibrinogen,1,2 AT activity,1,2 PC activity,1,2 d-dimers,1,2 and factor VIII activity.3 CBC, automated platelet counts, and serum biochemistry panels were performed in the TCSVM clinical pathology laboratory or by a commercial veterinary diagnostic laboratory.

Statistical Analysis

Box and whisker plots and tests for skewness and kurtosis were conducted to evaluate data distribution. Parametric and nonparametric data were expressed as mean and standard deviation or median and range, respectively. TEG values, biochemical data, platelet count, MCV, and coagulation parameters were compared between CPSS dogs and reference ranges with either parametric (Student's t test) or nonparametric (Mann–Whitney) tests. Statistical significance was set at P < 0.05 (2-tailed) and adjusted for multiple comparisons by Bonferroni correction.

Depending on TEG analysis, dogs were labeled as either hypercoagulable (G value > 8446 d/s) or normocoagulable.[8-11] Two-by-two contingency tables were used to analyze categorical data from hypercoagulable and normocoagulable dogs and compared using the Fisher exact test with a 2-tailed P < 0.05 considered significant. Pearson's correlation coefficient was used to examine the relationship between G value and various coagulation and biochemical parameters.


The 21 dogs enrolled in the study represented 5 mixed breed dogs and 12 pure breed dogs including Pug (n = 2), Maltese (n = 2), Toy Poodle (n = 2), Papillon (n = 2) and 1 each of Dachshund, Labrador Retriever, Jack Russell Terrier, Chihuahua, Pomeranian, Cairn Terrier, Bichon Frise, and Nova Scotia Duck Tolling Retriever. There were 11 females and 10 males with a median age and weight of 2 years (range, 0.6–10 years) and 4.2 kg (range, 1.4–33 kg), respectively. At the time of presentation, 17 dogs were symptomatic and exhibited varying clinical signs and findings: 10/21 dogs had neurologic signs or clinical findings compatible with HE (ie, diffuse cerebral signs, increased blood ammonia concentration or both), 6/21 dogs had ammonium biurate crystalluria or calculi, 5/21 exhibited gastrointestinal signs (eg, weight loss, inappetence, vomiting, diarrhea or some combination of these), and 1/21 had ascites. The remaining 4/21 dogs were asymptomatic and referred for diagnostic evaluation of increased liver enzyme activities, increased total serum bile acid concentrations, or both. Using abdominal ultrasound examination (19/21), nuclear scintigraphy (3/21), computed tomography (CT) scan (2/21), or some combination of these, a single extrahepatic portosystemic shunt was identified in 19/21 dogs and an intrahepatic shunt was identified in 2/21 dogs. On presentation, none of the dogs had evidence of bleeding or thrombosis. Likewise, none of the 12 dogs taken to surgery for shunt attenuation developed clinically detectable post-operative bleeding or thrombosis.

An automated platelet count, manual blood smear evaluation, and serum biochemistry panel were performed on all 21 dogs and the remainder of the hemostatic testing on 19 dogs (Table 1). As previously reported,[2] dogs with CPSS had significant decreases in MCV and serum albumin, cholesterol, and globulin concentrations; and increases in ALT activity compared to the reference population (data not shown).

Table 1. Hemostatic testing in dogs with CPSS
VariableaCPSS Median (range)Reference Range Median (range)P-Valueb
  1. a

    Platelet count measured in 21 dogs, all other tests measured in 19 dogs.

  2. b

    P-value for comparison of hemostatic variables in CPSS dogs and reference range using the Mann–Whitney test.

Platelets (×109/L)296 (129–404)353 (180–525)0.005
PT (s)9.5 (7.3–18.4)7 (5–9)0.0005
aPTT (s)15.2 (11.9–30.3)15.2 (9.9–20.4)0.25
Fibrinogen (mg/dL)233 (80–489)241 (73.4–410)0.85
Antithrombin (%)70.3 (32.2–103)93 (75–110)0.0005
Protein C (%)45.5 (23.8–82.3)98 (64.9–130)0.0005
D-dimers (ng/mL)323 (131–917)294 (55–533)0.025
Factor VIII (%)223.0 (104–1000)125 (50–200)0.0005

AT and PC activities were decreased in 11/19 and 17/19 dogs, respectively, whereas d-dimers and factor VIII activity were increased in 7/19 and 14/19 dogs, respectively. The median platelet count in dogs with CPSS was significantly lower than in the reference population but the actual count was decreased below the reference interval in only 3/21 dogs. PT was prolonged in 10/19 dogs. There was no difference in aPTT and fibrinogen concentrations between dogs with CPSS and the reference population. Only 1/19 dogs had an increased fibrinogen concentration or prolonged aPTT.

TEG analysis was performed on all 21 dogs (Fig 1). The K time was significantly shorter in dogs with CPSS compared to reference dogs, although it fell below the reference range in only 2 dogs with CPSS. Similarly, the angle and MA in dogs with CPSS were significantly greater than in reference dogs, although these values exceeded the reference range in only 2/21 and 6/21 dogs, respectively. Nine dogs with CPSS had G values compatible with a hypercoagulable state. Of the remaining 12 dogs, G values were compatible with normal coagulation.

Figure 1.

Thromboelastographic parameters in dogs with congenital portosystemic shunts (CPSS, n = 21) compared with the reference range generated in the Tufts Cummings Coagulation Laboratory (n = 75). Results shown are means with standard deviation. Means between the reference population and dogs with CPSS were compared with the Student's t test and corresponding P-values are shown.

Dogs with CPSS were divided into 2 groups (hypercoagulable and normocoagulable) based on the G value to determine if any clinical, coagulation, or biochemical variables were associated with the presence of hypercoagulability based on TEG. This analysis showed that platelet count was significantly higher, K value shorter, and both angle and MA greater in dogs that were identified as hypercoagulable (Table 2). Neither age, weight, albumin concentration, cholesterol concentration, ALT activity, PCV, MCV, WBC, or neutrophil count were different between dogs with increased or normal G values (data not shown). Dogs with increased G values did have higher platelet counts than dogs with normal G values (median, 323 × 109/L; range, 296–404 × 109/L versus median, 232 × 109/L; range, 129–403 × 109/L, respectively). When linear regression analysis was carried out comparing G values with conventional coagulation parameters, only platelet count (r = 0.66, P = 0.001) and fibrinogen concen-tration (r = 0.42, P = 0.025) showed a significant association.

Table 2. TEG variables in hypercoagulable and normocoagulable dogs with CPSS
TEG VariablesHypercoagulable (n = 9) Mean ± SDNormocoagulable (n = 12) Mean ± sdP-ValueaReference Range
  1. a

    P-value for comparison of TEG parameters in hyper- and normocoagulable dogs with CPSS using the Student's t test.

R (min)3.70 ± 1.124.33 ± 2.10.4171.85–6.85
K (min)1.17 ± 0.241.85 ± 0.740.0170.74–3.48
Angle (°)72.7 ± 3.865.6 ± 7.760.02148.2–76.6
Maximum amplitude (mm)64.8 ± 1.556.87 ± 5.000.000245.1–64.0
LY30 (%)0.89 ± 1.810.96 ± 2.20.940−1.67–3.04
G (d/s)9265 ± 5986724 ± 12120.0000153867–8446

The coagulation status of dogs with signs of HE at the time of presentation was examined. These dogs were more likely to be hypercoagulable based on TEG than dogs without signs of HE (Fisher exact test, P = 0.001; odds ratio, 40; 95% confidence interval, 3–224). Of the 10 dogs with CPSS and HE, 8 had TEG changes consistent with hypercoagulability. Because recent evidence supports a role for inflammation in HE,[11] we determined if increases in WBC, neutrophils or fibrinogen concentration were more common in the dogs with HE. There was no difference in WBC or neutrophil counts (data not shown), but dogs with HE (289 ± 93 ng/dL) had a significantly (P = 0.015) higher fibrinogen concentration than those without HE (191 ± 57 ng/dL).


In this study, TEG analysis identified the presence of hemostatic abnormalities consistent with systemic hypercoagulability in some dogs with CPSS. Abnormalities consistent with hypercoagulability included shorter K values and increased angle and MA compared to a reference population. The G value (a mathematical derivation of MA) was significantly increased in 42.8% (9/21) dogs with CPSS. Collectively, TEG analysis suggests that dogs with CPSS have a tendency to be hypercoagulable.

Dogs with high G values suggestive of a hypercoagulable state were more likely to have HE. This observation may fit with the emerging concept of the role of inflammation in the pathogenesis of HE.[11] Our finding of higher fibrinogen concentrations in dogs with HE, as well as a recent study showing increased C reactive protein concentrations in dogs with HE,[12] suggest a link between HE and inflammation. Because inflammation can predispose to a procoagulant state,[13] this may be an important factor that shifts dogs with CPSS toward a hypercoagulable state. Inflammation in dogs with CPSS may arise from bacterial translocation from the intestinal tract or from systemic absorption of proinflammatory bacterial products such as endotoxin. Additional studies are necessary to confirm the association of HE with alterations in inflammation and coagulation.

Several factors could contribute to the presence of a hypercoagulable state in dogs with CPSS. Hypercoagulability is associated with altered vascular flow, increased procoagulant activity, decreased anticoagulant activity, disordered fibrinolysis, or some combination of these abnormalities. Altered vascular flow patterns exist in the portal circulation of dogs with CPSS, and turbulence in the portal vasculature is a hallmark ultrasonographic finding in dogs with CPSS.[14] In addition, increases in procoagulant factor VIII activity were identified in 14/19 (73%) of the dogs in this study and may reflect endothelial cell dysfunction induced by altered shear stress in the portal vasculature, or alternatively may be associated with the hepatic arteriolar proliferation that occurs as a consequence of hypo-perfusion. Dogs with CPSS also had decreases in the anticoagulants, AT activity (11/19, 58%), and PC activity (17/19, 89%). Thus, dogs with CPSS have both local vascular factors as well as systemic coagulation changes that might predispose them to hypercoagulability.

Concurrently, dogs with CPSS have changes in coagulation parameters that could predispose them to hypocoagulability. Our studies and those of others have shown prolonged PT and aPTT and mild to moderate decreases in platelet count in dogs with CPSS.[1, 3] Furthermore, some studies show mild decreases in platelet aggregability[15] and decreases in the concentrations of factors II, V, VII, and X.[3] Despite the presence of these abnormalities traditionally associated with hypocoagulablity, TEG analysis suggests that concurrent procoagulant alterations (decreases in AT and PC activities and increases in factor VIII activity) may have a greater impact on global coagulation status in dogs with CPSS. In this study, 10/19 (53%) dogs with CPSS had prolonged PT, but 7/10 (70%) were normocoagulable, and 3/10 (30%) were hypercoagulable. One dog with CPSS had prolonged aPTT and was normocoagulable. Thus, our results indicate that prolongations in PT, aPTT, or both are not necessarily associated with a hypocoagulable state in dogs with CPSS.

Although we found that overall dogs with CPSS had significant decreases in albumin, cholesterol, and globulin concentrations as well as decreases in PCV and MCV and increases in serum transaminase activities when compared to reference dogs, none of these clinicopathologic variables differed between dogs with CPSS and hypercoagulable or normocoagulable G values. When we examined the relationship between G value and coagulation parameters, platelet count and fibrinogen concentration were significantly and positively correlated. This correlation was expected, because G is mathematically derived from MA, and MA is largely determined by platelet count and fibrinogen concentration.[16] None of the dogs in the study, however, had thrombocytosis. In fact, overall, dogs with CPSS had lower platelet counts than the reference population, although only 3/21 dogs had a platelet count below the reference range. A similar trend for lower platelet count in CPSS dogs has been reported previously.[1, 3, 4] Dogs with G values in the hypercoagulable range, however, had higher platelet counts than dogs with normal G values. Although it is not clear how platelet counts within the reference range might contribute to hypercoagulability, the association could be explained by the presence of platelet hyperreactivity. We did not investigate platelet function in this study, but a single previous study has shown a tendency toward hypoaggregability in dogs with CPSS.[15] We also were able to show a weak but statistically significant association of G value with serum fibrinogen concentration. Only 1/19 dogs tested was hyperfibrinogenemic and there was no difference in overall fibrinogen concentration between CPSS dogs and reference dogs. Hypercoagulable dogs did show a tendency toward a higher fibrinogen concentration than dogs that had a normal G (data not shown). The higher serum fibrinogen concentration could reflect the presence of a systemic inflammatory state, which might tip the coagulation balance in favor of hypercoagulability.[13]

Although none of the dogs in this study had evidence of overt thrombotic disease at the time of TEG analysis, 7/19 (37%) dogs had increases in d-dimers, 2 of which also had an abnormal percentage of clot lysis at 30 min (LY30). These changes could indicate the presence of enhanced fibrinolysis consistent with the presence of occult thromboembolism. Previous studies in dogs with CPSS also have shown that a small population has increased d-dimers (9–20%), but these increases in d-dimers were not correlated with clinically evident thromboembolism.[3, 4]

Our study has several limitations. We used a kaolin-activated TEG assay, and results with tissue factor activation may be different.[16] Future studies are necessary to determine if TEG in dogs with CPSS using native whole blood and tissue factor activation is similar. Additionally, although we established reference ranges for kaolin-activated TEG, we did not age-, weight-, or breed-match our CPSS population. Another limitation of this study is the small sample size. The investigation was conducted as a pilot study to determine if TEG analysis might give a better indication of the global coagulation status in dogs with CPSS. Because our data suggest that some dogs with CPSS have hemostatic abnormalities consistent with a hypercoagulable state, future studies can be designed with sufficient power to discern the etiologic factors behind this finding. Results of this study suggest that these changes may be related to the presence of HE, either because of the presence of a systemic inflammatory state (association with increased fibrinogen concentration), endothelial dysfunction (increase in factor VIII activity), or platelet hyperreactivity. Future studies should evaluate the role of endothelial cell products (eg, von Willebrand factor concentration, tissue factor pathway inhibitor, plasminogen activator inhibitor 1), continue to investigate platelet function, and examine other factors involved in fibrinolysis (eg, plasminogen, tissue factor pathway inhibitor). These studies take on increased importance as literature emerges that medical management of CPSS can result in prolonged survival.[17],4 This may lead to more dogs being medically managed for longer periods of time, and attention to coagulation status could become an important part of this management.

In conclusion, TEG has identified a relative hypercoagulable state in some dogs with CPSS. Additional studies are warranted to determine if TEG analysis should become another tool to determine if dogs with CPSS are at high risk for perioperative hemostatic complications and for assessing response to medical management of HE.

Conflicts of Interest

Authors disclose no conflict of interest.


  1. 1

    ACL Elite Analyzer, Beckman Coulter, Brea, CA

  2. 2

    TEG 5000 Thromboelastograph, Haemonetics Corp, Braintree, MA

  3. 3

    Cornell University, Animal Health Diagnostic Center, Comparative Coagulation Laboratory, Ithaca, NY

  4. 4

    Center SA, Randolph J, Warner K, et al. Long-term survival of dogs (n = 597) with congenital or acquired portosystemic shunting:1980–2010. J Vet Intern Med 2012;26:781A ACVIM Forum 2012, New Orleans, LA (abstract)