This work was performed at the Department of Small Animal Clinical Sciences, The Faculty of Life Sciences, University of Copenhagen, Frederiksberg, Denmark. Preliminary data were presented in abstract form at the annual meeting of the Veterinary Cancer Society, October 2005.
Corresponding author: Annemarie T. Kristensen, DVM, PhD, Department of Small Animal Clinical Sciences, The Faculty of Life Sciences, University of Copenhagen, Dyrlaegevej 16, DK-1870 Frederiksberg C, Denmark; e-mail: firstname.lastname@example.org.
Background: Abnormal routine coagulation assay results have been reported to be common in veterinary patients with neoplasia, but the overall hemostatic functional state, including hypercoagulability, has not been described.
Hypothesis: The overall hemostatic functional state, including hypercoagulability, can be assessed in dogs with neoplasia by tissue factor (TF)-activated thromboelastography (TEG).
Animals: Thirty-six dogs with malignant neoplasia and 13 dogs with benign neoplasia presented to the Small Animal Veterinary Teaching Hospital, The University of Copenhagen, Frederiksberg, Denmark.
Methods: Prospective study evaluating the overall hemostatic functional state in dogs with neoplasia by a newly validated TF-activated TEG assay and routine coagulation parameters activated partial thromboplastin time (aPTT), prothrombin time (PT), platelet count, and D-dimer concentration.
Results: Hemostatic dysfunction was observed in 28/49 (57%) dogs with neoplasia. Twenty-four were dogs with malignant neoplasia, the majority of which 18/36 (50%) were hypercoagulable, whereas 6/36 (17%) were hypocoagulable. All hypocoagulable dogs had metastatic disease. The proportion of dogs with altered hemostasis was significantly different between dogs with malignant and benign neoplasia.
Conclusions and Clinical Importance: TF-activated TEG detected hypercoagulable and hypocoagulable states in this population of dogs with neoplasia. The most common hemostatic abnormality in dogs with malignant neoplasia was hypercoagulability. These findings suggest that this novel hemostatic function test may be of value as a cage side method for the assessment of overall hemostatic function in dogs with cancer, including the detection of both hyper- and hypocoagulable states as well as mixed disorders.
Disorders of hemostasis are common in human as well as veterinary cancer patients. In humans, overt deep venous thrombosis (DVT) often is observed as a primary clinical presenting problem before diagnosis.1–3 Laboratory evidence of hemostatic abnormalities have been reported in 58–98% of human cancer patients.1
In dogs with cancer, a variety of hemostatic disorders have been reported, most commonly thrombocytopenia and prolongation of the activated partial thromboplastin time (aPTT), prolongation of the prothrombin time (PT), or both, as well as decreased fibrinogen concentration, increased D-dimer concentration, or both.4–16 In 1 study evaluating coagulation tests in 100 dogs with cancer, 83% had abnormal coagulation tests and 36% were thrombocytopenic.4 In another study, 10% of 2,059 dogs with cancer were thrombocytopenic.6 In studies of specific types of cancer, 1 study showed, in a group of 24 dogs with hemangiosarcoma, that 50% had disseminated intravascular coagulation (DIC) at presentation, 75% were thrombocytopenic, and 25% died because of hemostatic problems.9 In another study, the incidence of DIC in 206 dogs with malignant tumors was 9.6% overall and 12.2% in those with malignant solid tumors.11 An association between stage of disease and presence and severity of hemostatic changes was reported in a study from 1999, in which it was shown in 60 dogs with untreated mammary carcinoma that hemostatic abnormalities increased with stage of disease, and that hemostatic abnormalities were observed more frequently in dogs with tumors with distant metastasis, invasion, and extensive tumor necrosis.16 Common to all the above-mentioned studies is the use of assays that are based on isolated parts or phases of the hemostatic process (ie the use of citrated plasma for the traditional clotting assays in which only the initiation of the clotting process is recorded). These assays are most likely unable to provide an evaluation of the overall hemostatic system in any disease state, because not all components of the hemostatic system are part of the assays.
The presence of hypercoagulability has been difficult to assess antemortem in dogs, and the importance of hypercoagulability and thromboembolic disease has not been well-defined in veterinary medicine. In a study of the causes of thromboembolic disease in 29 dogs with postmortem confirmed pulmonary thromboembolism (PTE), cancer was the most common finding and was observed in 41% of the patients.17 In addition, the study showed that risk factors for PTE in these 29 dogs included factors that may be common in dogs undergoing treatment for cancer, such as IV catheters (77%), cortico-steroids (48%), cytotoxic drugs (38%), recent surgery (21%), and blood transfusion (10%).17 In another study of PTE, neoplasia also was reported to be among the most common findings in these dogs.18 In a study of portal vein thrombosis in 11 dogs, the 2 most common causes were cancer and pancreatic necrosis with 36% each.19 The above-mentioned studies indicate that neoplasia is a risk factor for hypercoagulability and associated complications such as thromboembolic disease.
For both dogs and humans with cancer, bleeding and thrombosis can lead to acute, life-threatening complications. Despite this, assessment of hemostasis and in particular hypercoagulability is not routine in the diagnostic evaluation and management of veterinary cancer patients. Interestingly, experimental studies in mice indicate that coagulation plays an important role in disease progression and metastasis.20–28 Many of the cells and proteins involved in maintaining hemostasis also are involved in the processes of cancer growth, invasion, and metastasis as well as angiogenesis.20,21,23 Modulating the hemostatic response in cancer may therefore also slow disease progression and prevent development of complications. Clinical studies in human patients with carcinomas, invasive solid tumors, and several other cancer types using low molecular weight heparins and warfarin concurrent with other treatment and comparing cancer patients receiving anticoagulant treatment to patients that did not have treatment indicated longer survival times in these patients, suggesting that anticoagulant treatment may slow disease progression as well as control the complications associated with the prothrombotic state.1–3 DVT is an important clinical complication in human cancer patients, but whether the clinical relevance of DVT is similar in dogs with cancer is unknown.
Assessment of hypercoagulability and thrombosis in dogs is not easily accomplished with routinely used coagulation assays. The D-dimer concentration mainly has negative predictive value. It is not widely available in a practice setting, and usually is sent to a commercial laboratory along with other coagulation testing.29,30 Smaller instruments such as the SCA 2000 are available for larger veterinary practices, but they provide only aPTT, PT, and activated coagulation time. There is a need for improved assay methods that enable easier and cage side assessment of hemostatic changes, including hypercoagulability, and the overall hemostatic state in dogs with cancer, with the ultimate goal of improving the management of these patients. Thromboelastography (TEG) enables global assessment of hemostatic function with whole blood (WB) by evaluation of the interaction of both cellular and plasma components during initiation, amplification, and propagation of clot formation as well as during fibrinolysis.31,32 The advantages of TEG include the assessment of overall hemostatic function by evaluation of the visco-elastic properties of the developing blood clot, including clot formation, kinetics, strength, stability, and resolution as well as immediate and cage side detection of hyper- as well as hypocoagulable states.31,32 TEG methods include TEG (Haemascope) or ROTEG (Pentapharm).31,32 We recently have validated a recombinant human tissue factor (TF)-activated TEG assay.31,32 The TF-induced TEG was chosen because cell-based TF is believed to be one of the key factors in the initiation of hemostasis.
The objective of the present study was to characterize the overall hemostatic state in dogs presenting with a variety of tumors, as well as to evaluate whether or not hypercoagulability would be observed by the newly validated TEG assay.
Materials and Methods
Animals and Blood Samples
This study was approved by the Small Animal Ethics and Administrative Committee at the Department of Small Animal Clinical Sciences, The Faculty of Life Sciences, University of Copenhagen, Denmark.
WB was collected prospectively by careful jugular venipuncture, using minimum stasis of blood and a 21-G butterfly needle from 49 dogs with a variety of malignant (n = 36) and benign (n = 13) tumors (carcinomas, n = 13; hemangiosarcoma, n = 8; lymphoma, n = 8; perianal adenoma, n = 2; benign mammary tumors, n = 3; variety of others, 1 each) admitted to The Small Animal Veterinary Teaching Hospital, Department of Small Animal Clinical Sciences, The Faculty of Life Sciences, University of Copenhagen, from November 2004 through July 2006. As many dogs as possible were included prospectively in the study period if diagnostic evaluation, including blood samples, was pursued and a diagnosis was obtained by histopathology, cytology, or both, except in 1 dog with a cardiac tumor that was diagnosed by echocardiography. Staging was not performed in all dogs. Blood samples were collected before treatment and none of the dogs had received medications that are known to influence hemostasis. Blood samples were collected into serum, EDTA, and citrated vacutainer plastic tubes, in that order. The 3-mL citrate tubes were inverted carefully 5 times after sampling to ensure mixing of 3.2% trisodium citrate and blood in a 1 : 9 ratio, and stored at room temperature for subsequent TEG analysis. Blood samples collected in serum tubes and 3.2% citrate were centrifuged at 4,000 ×g for 120 seconds and the serum and plasma were separated and collected within 30 minutes of sampling and stored at −80 °C until further analysis. TEG analyses with TF as activator at a concentration of 1 : 50,000 were performed 30 minutes after collection of blood.
The TEG analysis was performed with a computerized thromboelastographa according to the previously published method.31,32 Canine citrated WB samples were activated with a solution of recombinant human TFb prediluted 1 : 2,780 in a HEPES buffer with 2% bovine serum albumin. The final TF dilution was 1 : 50,000. After 30 minutes of storage at room temperature, 20 μL 280 mM CaCl2 was pipetted into a prewarmed (37 °C) TEG cup.a Subsequently, 20-μL prediluted TF was mixed with 320-μL canine citrated WB and the premix was added to the cup, giving a total volume of 360 μL in each cup, after which the pin was gently lowered into the cup and measurements were initiated. The TEG analyses were run for 120 minutes. Data were obtained continuously by electronic transfer to personal computer from the analyzer.
Interpretation of TEG Results
Five TEG parameters (R, K, α, MA, and G) were investigated in the present study (see Fig 1).
R is the reaction time and represents the time of latency from the time the blood was placed in the TEG analyzer until a preset fibrin formation is reached, measured as an increase in amplitude of 2 mm. R is primarily related to plasma clotting factors and inhibitor activity. The clotting time K is the time to clot formation measured from the end of R until an amplitude of 20 mm is reached. It is a measure of the time it takes from initial clot formation until a predetermined clot strength is reached. K is primarily related to clotting factors, fibrinogen, and platelets. The α or angle represents the rapidity of fibrin build-up and cross-linking and is mainly dependent on platelets, fibrinogen concentration, and clotting factors. MA is the maximum amplitude reached. The MA is dependent on platelet count and function as well as fibrinogen concentration, and it is a direct function of fibrin and platelet bonding, which represents the ultimate strength of the fibrin clot. The TEG maximum amplitude also is a measure of clot stiffness and can be used to derive clot shear elastic modulus G, where G= 5,000 ×MA/(100 −MA) and is a measure of the overall coagulant state as normo-, hyper-, or hypocoagulant.
The results of TEG variables R, K, α, MA, and G for dogs with a variety of tumor types were compared with a previously established reference range for healthy dogs.31 The dogs were divided into 3 groups based on the TEG G value: normo-, hypo-, and hypercoagulable.
Routine tests, including aPTT, PT, platelet concentration, fibrinogen concentration, and plasma D-dimer, were assessed. aPTT, PT, and fibrinogen concentration were evaluated with an automated chemical analyzer,c platelet concentration with the Advia 120 automated hematology instrument,d and concentrations of D-dimer by turbidometric immunoassay.e Measurement of the aPTT involved use of synthetic phospholipids and silica to trigger coagulation.f The PT and fibrinogen concentration were evaluated simultaneously by use of a rabbit brain calcium thromboplastin.g The automated chemical analyzer records light scattering before and after clot formation, and fibrinogen concentration was calculated by use of these values and a calibration curve. A pooled sample of plasma from 8 clinically healthy dogs was analyzed together with the samples. Plasma samples were thawed at 37 °C in a water bath immediately before analysis and centrifuged at 3,000 ×g for 5 minutes (to avoid remnants of cryoprecipitate in plasma after thawing); the supernatants were used for analysis.
For the coagulation parameters PT, aPTT, and D-dimer, a Mann-Whitney test was used to determine whether differences between medians of measurements in dogs with benign and malignant neoplasia were significant.
For all other coagulation parameters, a χ2 test was used to detect statistically significant differences in distributions of data stratified into benign versus malignant neoplasia. For the individual TEG parameters, a χ2 test was used to detect statistically significant differences in distributions of data stratified into normal, hyper-, and hypocoagulable based on the TEG G value and furthermore grouped as benign versus malignant. Statistical significance was set at P= .05.
According to the established TEG reference ranges, in this group of dogs with a variety of tumor types of which 13 could be characterized as benign and the remaining as malignant, a total of 28/49 dogs (57%) had abnormal TEG results (Tables 1 and 2). Overall, a total of 22/49 dogs could be characterized as hypercoagulable (45%), 6 were characterized as hypocoagulable (12%), and 21 were characterized as normocoagulable (43%) (Tables 1 and 2). In dogs with malignant neoplasia, 18/36 (50%) dogs were hypercoagulable, 6/36 (17%) were hypocoagulable, and 12/36 (33%) were normocoagulable (Tables 1 and 2). In dogs with benign neoplasia, 4/13 (31%) were hypercoagulable, no dogs were hypocoagulable, and 9/13 dogs (69%) were characterized as normocoagulable with the TEG assay (Tables 1 and 2). A graphical display of representative hypercoagulable, normocoagulable, and hypocoagulable profiles from dogs in this study is presented in Figure 2.
Table 1. Results of thromboelastography (TEG) in 36 dogs with malignant neoplasia.
Hypercoagulable 18/36 (50%)
Hypocoagulable 6/36 (17%)
The dogs were divided in hyper- and hypocoagulable groups based on the G value.
G (× 103 d/s)
Table 2. Results of thromboelastography (TEG) in 13 dogs with benign neoplasia.
Hypercoagulable 4/13 (31%)
Hypocoagulable 0/13 (0%)
The dogs were divided in hyper- and hypocoagulable groups based on the G value.
NA, not applicable.
G (× 103 d/s)
In the malignant group, the hypercoagulable dogs consisted of 7 dogs with carcinomas, 5 dogs with hemangiosarcoma, 3 dogs with lymphoma, 2 with mastocytoma, and 1 with osteosarcoma. Some of these dogs did have evidence of metastatic disease, but consistent diagnostic evaluation was not available for all of the dogs. In the malignant group, the hypocoagulable dogs all had metastatic cancer, 3 had metastatic carcinoma, 2 had metastatic hemangiosarcoma, and 1 had generalized lymphoma.
In the benign group, 4 dogs were found hypercoagulable, all of which had epithelial neoplasia. In these dogs, only the TEG parameters MA and G were increased above the reference range.
The proportions of dogs with increased or decreased α, K, MA, and G values were significantly different in dogs with malignant neoplasia compared with dogs with benign disease, indicating that fewer dogs were hypercoagulable and no dogs were hypocoagulable in the benign group (Fig 3). The R values were not significantly different between the groups. A pattern between type of neoplasia and TEG profile was not discernable because of the limited number of cases of each type of cancer.
Of the traditional coagulation assays, there was a statistically significant increase in PT and D-dimer concentration and a decrease in the platelet count when the dogs with malignant neoplasia were compared with the dogs with benign disease (Fig 4). The percentages of dogs classified as hyper-, hypo-, or normocoaguable according to the TEG G value that had abnormal traditional coagulation assay results are listed in Table 3.
Table 3. Results of traditional hemostasis assays in dogs with neoplasia characterized as hypercoagulable (n = 22), hypocoagulable (n = 6), or normocoagulable (n = 21) according to the thromboelastography G-value.
APTT, activated partial thromboplastin time (seconds); PT, prothrombin time (seconds); Fibg., fibrinogen (g/dL); Plt, platelet concentration (× 103/μL); D-dimer (ng/mL).
Hypercoagulable (G > 7.2 × 103 d/s)
Hypocoagulable (G < 3.2 × 103 d/s)
Normocoagulable (3.2 < G < 7.2 × 103 d/s)
In this group of dogs with cancer, it was confirmed that hemostatic dysfunction is common and was present in 28/49 tested (57%) by TF-activated TEG. The proportions of dogs with increased or decreased α, K, MA, and G values was significantly different in dogs with malignant neoplasia compared with dogs with benign disease, indicating that fewer dogs were hypercoagulable and no dogs were hypocoagulable in the benign group. A pattern between type and stage of neoplasia and TEG profile was not discernable because of the limited number of cases of each type of neoplasia and the fact that not all dogs underwent complete staging. Of the traditional coagulation assays, there was a significant increase in PT and D-dimer concentration and a decrease in the platelet count when the dogs with malignant neoplasia were compared to those with benign disease. When the results of the traditional assays were evaluated according to whether or not the TEG G value grouped the dogs as hyper- or hypocoagulable, the results were quite variable for PT, fibrinogen, and D-dimer concentration in the hypercoagulable group, but surprisingly 83% of these dogs had prolonged aPTT, traditionally indicating a hypocoagulable state. None of the dogs was thrombocytopenic, whereas in the hypocoagulable group 83% were thrombocytopenic.
That a high proportion of dogs (57%) with neoplasia had an abnormal hemostatic state according to the TEG was consistent with other studies,4–6,9–11,14 as were the results of the routine coagulation tests with significantly increased PT, increased D-dimer concentration, and decreased platelet count in dogs with malignant neoplasia.4–6,9–11,14 In contrast to previously reported studies of hemostatic changes in dogs with cancer, our study indicated that when TF-activated TEG, a global hemostatic assay, was used to assess the overall hemostatic state, the most common abnormality was hypercoagulability, and interestingly the majority of these dogs simultaneously had prolonged aPTT—a result traditionally indicating hypocoagulability. The most likely reason that previous studies failed to show hypercoagulability is the fact that a classification of hypercoagulability cannot be made for the results based on the methods used. Except for the platelet count, they are primarily plasma-based assays that measure a short clotting time (seconds) or a single component (D-dimers and fibrinogen), whereas the TEG measurements are based on the presence of all hemostatic components. In the present study, the aPTT was prolonged in 83% of the dogs that TEG had classified as hypercoagulable. A prolonged aPTT would normally indicate a hypocoagulable state, and the TEG results of hypercoagulability in a system containing all of the components of the hemostatic system indicate that when evaluated in isolation or as a measure of overall hemostatic state, aPTT or other traditional coagulation assays may incorrectly indicate hypocoaguability.
Hypocoagulability was observed only in dogs with malignant neoplasia and only in 17% of them. This result is much lower compared with the finding in previous studies that up to 83% of dogs had prolongation of one or more traditional coagulation tests, thrombocytopenia, or both, potentially indicating a hypocoagulable state, but it is important to emphasize that the TEG and traditional assays look at very different measures of coagulation.4–6 Because TEG is based on the simultaneous interaction of all of the cells and proteins participating in hemostasis, it could be argued that TEG evaluation may be a better test for evaluating a true hypocoagulable state, but further prospective studies including larger numbers of hypocoagulable dogs and correlating the results of their TEG tracing to clinical signs of bleeding are necessary. Confounding factors in this study potentially could be age, IV catheters placed, or medications given before admission. TEG results for different ages of dogs have not been described, nor has the effect on TEG of placement of IV cathethers or medications routinely given to cancer patients been reported. A small case series in dogs with parvovirus enteritis supports that TEG, although using a different activator from that used in the TEG reported here, enabled detection of hypercoagulability in a different subset of sick dogs,33 but whether there will be differences in TEG between healthy dogs and any with various types of illnesses is at present unknown.
Based on the TEG G value, which is partly based on the MA and is the parameter usually used to evaluate hypercoagulability in people according to the manufacturer's guidelines and recent publications,34,35 the overall hemostatic state most commonly observed in these dogs with neoplasia was hypercoagulability. This finding is consistent with the well-known finding in human cancer patients of hypercoagulability and increased risk of DVT (Trousseau's syndrome).1–3 The hypercoagulable state and the resulting thromboembolic complications observed in cancer are believed to result from circulating TF (either soluble or cellborne), cancer procoagulant factor (a factor that has not been defined completely yet), and platelet hyperactivity.1,2 Interestingly, in 2 previous veterinary studies evaluating platelet function in dogs with malignancies by platelet aggregation, significant platelet hyperaggregability was observed in both studies.12,13 Routine coagulation tests are plasma-based, whereas the TEG assay is a WB-based assay that includes both cellular and plasma components important for initiation, amplification, propagation, and lysis of the forming blood clot. Our findings and the finding in 2 previous studies of platelet hyperaggregability in dogs with malignant disease suggest that an assay that includes cellular as well as plasma components of the hemostatic process potentially may give a more reliable evaluation of the overall hemostatic state.12,13 The advantage of the TF-activated TEG assay is that it can be used cage side and is easy to perform. Limitations of the assay include that it may not be able to discriminate specific defects (eg, von Willebrand factor deficiency). Also, in the study reported here, parameters evaluating the fibrinolytic phase of the hemostatic process were not included, but the fibrinolytic process may be studied by means of specific fibrinolysis parameters such as LY30 or LY60 by TF-TEG.
The finding that 67% of dogs with malignant disease have hemostatic dysfunction of which 50% of the dogs were hypercoagulable suggests that evaluation of hemostasis should be included in the routine evaluation of dogs with cancer. It also suggests that additional studies are needed to address whether hypercoagulable canine cancer patients also have thromboembolic disease and whether therapeutic intervention by anticoagulant treatment is warranted. In humans, some recent clinical studies have shown increased survival and quality of life in patients with malignancy at risk for DVT when they are treated long term with either a low molecular weight heparin or warfarin.1–3 In experimental studies in mouse models, coagulation has been shown to play an important role in disease progression and metastasis.20–28 Hemostatic cells and proteins promote cancer growth, invasion, and metastasis as well as angiogenesis.20,21,23 Modulating the hemostatic response in cancer therefore may also slow disease progression and complications. Many drugs that modulate hemostasis are available, but limited evidenced-based information is available regarding their use in veterinary medicine. In human medicine, anticoagulant treatment is used routinely for thromboprophylaxis in cancer patients. With the apparent ability to detect hypercoagulability by TEG, the potential for prospective evaluation of anticoagulant treatment in dogs with cancer should be possible.
The number and types of each tumor were limited and precluded an assessment of whether the type and degree of hemostatic dysfunction were associated with a certain type or stage of disease. Additional studies are needed to assess this possibility, but a consistent finding was that only dogs with metastatic disease were observed to have an overall hypocoagulable state, which was consistent with the clinical picture in these dogs, all of which showed symptoms of bleeding. Another interesting finding was that in the dogs with benign neoplasia the only abnormality was a mild hypercoagulable state in 4 dogs, all of which had epithelial neoplasia. This finding could indicate the presence of undetected underlying malignant disease. Another possibility is that epithelial tumors express and shed TF, causing hypercoagulability, because it has been shown in human tissue that epithelial cells express TF.36
In conclusion, the TF-activated TEG assay confirmed that a majority of dogs with neoplasia have hemostatic dysfunction, but most important documented for the 1st time that the most common abnormality is hypercoagulability, similar to what is observed in humans. Our findings suggest that all dogs with neoplasia should receive assessment of their hemostatic functional state before treatment and that TF-activated TEG is useful in the evaluation of hypercoagulability. The findings also indicate that dogs with malignant neoplasia and hypercoagulability may serve as a model for human disease. Additional studies are needed to show whether anticoagulant treatment is of benefit in these dogs and whether or not the TEG assay will be of value in monitoring patients receiving such treatment.
aTEG 5000 Hemostasis Analyzer, Haemoscope Corporation, Niles, IL
bInnovin, Dade Behring, Marburg, Germany
cACL9000, Instrumentation Laboratory, Warrington, UK
dAdvia 120, Bayer Health Care Diagnostics, Berlin, Germany
eNyco Card Reader, NYCOMED, Rosalde, Denmark
fIL Test APTT-SP, Instrumentation Laboratory
gIL Test PT-Fibrinogen, Instrumentation Laboratory
The authors wish to thank Natascha Errebo for expert technical assistance, the participating staff and patients at the Small Animal Veterinary Teaching Hospital, The Faculty of Life Sciences, University of Copenhagen, Denmark, as well as Fisher Medical A/S and Novo Nordisk A/S for supplying the TEG-cups used in this study.