• Open Access

Hypercoagulability in Dogs with Protein-Losing Nephropathy as Assessed by Thromboelastography


  • This study was presented in abstract form at the 2010 American College of Veterinary Internal Medicine Forum, Denver, CO.
  • This study was supported by the North Carolina State University Firestone Research Endowment.

Corresponding author: Shelly Vaden, North Carolina State University College of Veterinary Medicine, 1060 William Moore Drive, Raleigh, NC 27606; e-mail: slvaden@ncsu.edu.



Dogs with protein-losing nephropathy (PLN) are at risk of thromboembolic disease, but the mechanism leading to hypercoagulability and the population of dogs at risk are unknown.


To characterize thromboelastography (TEG) and its association with serum albumin (SALB), UPC, and antithrombin activity in dogs with PLN.


Twenty-eight client-owned dogs with PLN (urine protein:creatinine ratio [UPC] > 2.0) and 8 control dogs were prospectively enrolled in this observational study.


TEG parameters, antithrombin activity, serum biochemical profiles, and UPC were measured. TEG analyses were run in duplicate with kaolin activation; reaction time (R), clot formation time (K), α-angle (α), maximal amplitude (MA), and global clot strength (G) were analyzed.


Dogs with PLN had lower K (= .004), and higher α (= .001), MA (< .001), and G (< .001) values than controls. No significant correlation between TEG parameters and UPC, SALB, or antithrombin was noted. Twelve PLN dogs (42.8%) were azotemic and 19 (67.8%) were hypoalbuminemic (SALB < 3.0 g/dL); 11 had SALB < 2.5 g/dL.

Conclusions and Clinical Importance

These results indicate that dogs with PLN have TEG values that demonstrate hypercoagulability compared with a control population but that antithrombin, SALB, or UPC cannot be used in isolation to predict this result. A comprehensive evaluation of the coagulation system in individual patients may be necessary to predict the point at which anti-thrombotic therapy is indicated.




urine protein-to-creatinine ratio


protein-losing nephropathy


serum albumin concentration


reaction time


clot formation time


maximum amplitude


global clot strength

Renal disease has been associated with hypercoagulable states in dogs and humans. Specifically, in patients with nephrotic syndrome there is a predisposition toward thrombosis.[1-3] In humans with nephrotic syndrome, reported incidence of thromboembolic complications is up to 42%.[4, 5] In one study, 22.2% of 137 dogs with idiopathic protein-losing nephropathy (PLN) had evidence of thromboembolism.[3] Despite the predisposition to thromboembolic disease, the underlying cause of hypercoagulability in dogs with PLN remains incompletely understood. Antithrombin deficiency and increased platelet aggregation secondary to hypoalbuminemia have been identified in dogs with nephrotic syndrome.[6, 7] However, plasma antithrombin activity, urine protein-creatinine ratio (UPC), and serum albumin concentration (SALB) are not predictive of thromboembolic complications.[3]

Thromboelastography (TEG) is a technique that allows global evaluation of the coagulation system, including clot formation, clot strength, and fibrinolysis.[8] The reaction (R) time, clot formation (K) time, α-angle (α), maximum amplitude (MA), and global clot strength (G) are the most commonly reported values generated by TEG. R is the precoagulation time, a measure of time from clot initiation until the first strands of fibrin are formed. K is a measure of time from initial clot formation to a predetermined clot strength. The α-angle is a measurement of the angle tangent to K, which represents rate of clot formation. MA represents final clot strength. In hypercoagulable states, R and K values decrease, whereas α-angle and MA increase. G is a calculated value, by means of the formula G = 5000 × MA/(100−MA) and also has been described as a way to categorize the overall results as hypo-, normo-, or hypercoagulable.[8-10] TEG is one of the only clinically available tests that can be used to detect a hypercoagulable state. In a study of chronic kidney disease in humans, uremic subjects were hypercoagulable compared with controls when evaluated by TEG.[11] This technique also has been prospectively utilized to identify humans at high risk for thrombosis after various interventions such as surgery or stent placement.[12-15]

Because TEG allows cellular and plasmatic components to interact, it is proposed to be a better method to determine thrombosis risk in patients with multifactorial etiologies of hypercoagulability, as seen in PLN, than solely measuring SALB or plasma antithrombin activity. A previous report of TEG in dogs with PLN documented hypercoagulability, but the study only included dogs with severe hypoalbuminemia. It is unknown if hypercoagulability as assessed by TEG develops before severe hypoalbuminemia.[16] The objective of this study was to characterize TEG and its association with SALB, UPC, and antithrombin activity in dogs with PLN. We hypothesized that dogs with PLN would be hypercoagulable compared with controls as demonstrated by TEG.

Materials and Methods

Dogs with PLN were eligible for enrollment, if they had a UPC > 2.0 with no evidence of pre- or postrenal proteinuria (eg, Bence Jones proteinuria, urinary tract infection). All dogs had results of a CBC, serum biochemistry profile, and urinalysis evaluated. Because this was a prospective observational study, coagulation profile testing, ultrasound examination, thoracic radiographs, and other diagnostic tests were performed at the discretion of the clinician responsible for patient management. All dogs with PLN were required either to have had a urine sample submitted for microbial culture that yielded no growth at time of documentation of proteinuria or to have been given broad spectrum antibiotics for at least 1 week before presentation.

Dogs were excluded if evidence of a concurrent disease associated with hypercoagulability (eg, pancreatitis, hyperadrenocorticism, neoplasia, protein-losing enteropathy, parvoviral enteritis, immune-mediated hemolytic anemia) was present, if the UPC was found to be < 2.0, if the dog was receiving heparin, or if urine culture results were positive for bacterial growth.

Dogs with PLN were further subdivided into groups based on the presence or absence of documented evidence of thrombosis. Thrombosis was diagnosed by imaging modalities (eg, ultrasound examination, magnetic resonance imaging) or by identification of a thrombus during post-mortem examination. Dogs that had suspected but unconfirmed thromboembolic disease (as determined by the clinician managing the case) were removed from the data set for statistical evaluation of thromboembolic complications.

Control dogs consisted of a population of healthy research dogs used in another study.[17] All control dogs had a CBC, serum biochemistry profile, coagulation profile, and urinalysis performed to exclude subclinical disease before inclusion in this study. The study protocol was approved by the North Carolina State University Institutional Animal Care and Use Committee.

Blood samples were collected from all dogs by jugular venipuncture with a 21-gauge Vacutainer® winged blood collection set by 1 of 2 investigators (EML, JMW).1 Blood was collected directly into 2 3.2% sodium citrate tubes (2.7-mL glass tube, 9 : 1 ratio of blood to anticoagulant) with the vacuum assistance of the Vacutainer® collection set after an initial volume of 1 mL was collected into a serum tube and discarded.2 One sodium citrate tube was allowed to rest in an upright position at 23°C for 30 minutes before performing TEG, and in the PLN dogs, a 2nd sodium citrate tube was immediately centrifuged and examined to confirm absence of hemolysis. Plasma was frozen at −80°C for later analysis of antithrombin activity.

Kaolin-activated TEG was performed by 1 of 2 investigators (EML, JMW) 30 minutes after collection of blood samples with a TEG analyzer according to manufacturer's instructions on a system consisting of two integrated thromboelastography analyzers.3 Briefly, 20 μL of 0.2 M calcium chloride was placed in a prewarmed (37°C) cup of the TEG analyzer. Citrated whole blood (1 mL) was transferred to a vial containing 1% kaolin and inverted 5 times.4 Then, 340 μL kaolin-activated citrated whole blood was pipetted into the TEG cup for a total of 360 μL, and the TEG tracing was initiated and allowed to run for 60–120 minutes at 37°C. Paired samples, 1 aliquot on each TEG analyzer, were run concurrently from each blood sample. The results of TEG variables R, K, α-angle, MA, and G were recorded, evaluated for consistency between tracings, and duplicates were averaged for final analysis.

Coagulation profile testing was performed at the discretion of the clinician managing the case, and was performed through the North Carolina State University Clinical Pathology Laboratory and included determination of PT, aPTT, D-dimer concentration, fibrinogen, and platelet count. Samples were delivered to the laboratory within 15 minutes of collection, and plasma was immediately centrifuged and separated. After separation, plasma samples were allowed to remain at 23°C for no longer than 30 minutes, after which coagulation tests were immediately performed.

PT and aPTT were determined by semiautomated electromechanical clot detection method with the STart® 4 hemostasis analyzer5 by thromboplastin6 and actin-activated cephaloplastin,7 respectively. D-dimers were determined by latex auto-agglutination.8 Platelet concentration was determined by an automated hematology analyzer9 and results verified by visual inspection. Fibrinogen concentration was estimated by heat precipitation.

Frozen plasma samples were sent to an outside laboratory10 for determination of antithrombin activity by the anti-IIa chromogenic assay. A standard curve for antithrombin activity was generated from dilutions of pooled samples of normal canine plasma which were assigned a value of 100% antithrombin activity.

Data were tested for normality by a Shapiro-Wilk test. Parametric data were compared by a Student's t-test, and non-parametric data were compared with a Wilcoxon Sum Rank test.11 Plasma antithrombin activity and SALB were correlated with TEG parameters by Spearman's rank correlation test.11 Descriptive statistics were used to describe other parameters. Data were expressed as median (range). Statistical significance was set at < .05.


Study Subjects

Thirty-two dogs with PLN (30 client-owned dogs and 2 dogs from a research colony of Soft Coated Wheaten Terriers) and 8 healthy control dogs from another research colony were enrolled in the study. Four PLN dogs were excluded; 2 had a UPC < 2.0, 1 had neoplasia discovered after collection of samples, and the remaining dog had a positive urine culture. Twenty-eight PLN dogs were included in the final analysis, including 14 spayed females, 12 castrated males, 1 intact male and 1 intact female dog. Breeds represented in this group included 3 Cairn Terriers, 3 Soft Coated Wheaten Terriers, 2 each of Cocker Spaniels, Golden Retrievers, Labrador Retrievers, Welsh Terriers, Yorkshire Terriers, and mixed breed dogs, and 1 each of Airedale, Beagle, Bernese Mountain Dog, Chihuahua, Dalmatian, English Springer Spaniel, Miniature Schnauzer, Papillon, Pembroke Corgi, and Pomeranian. Control dogs consisted of 8 adult intact males including 5 Beagles, 2 Foxhounds, and 1 mixed breed dog. PLN dogs ranged in age from 1 to 13 years (median, 8). Control dogs ranged in age from 3 to 4 years (median, 3.5). Five PLN dogs were being treated with enalapril, 1 was being treated with losartan, and 1 was being treated with clopidogrel at the time of sample collection. No dogs were receiving aspirin or heparin. Three dogs were undergoing antibiotic therapy (amoxicillin-clavulanate, 2; enrofloxacin, 1). One dog was given a single SC dose of dexamethasone 48 hours before sample collection.

Clinicopathologic and TEG Data

The results of clinicopathologic data are summarized in Table 1. Fourteen PLN dogs (50%) were azotemic. Only 4 PLN dogs had AT activity below the reference range. Dogs with PLN had significant alterations in TEG parameters compared with controls (Fig 1). Twenty-seven of the 28 dogs with PLN had at least 1 TEG variable that was hypercoagulable relative to controls. TEG data in PLN dogs and controls, and data classified according to clinicopathological results, are presented in Table 2. When hypoalbuminemic dogs were removed from analysis, PLN dogs with normal SALB (≥3.0 g/dL) still had significantly higher K, α-angle, MA, and G values than did control dogs (Fig 2, Table 2; < .05). In addition, PLN dogs with normal antithrombin activity (≥65%) had significantly higher K, α-angle, MA, and G values than did control dogs (< .05). The correlation (R) between antithrombin activity and SALB in dogs with PLN was .30 (= .167). In dogs with PLN, no significant relationship among UPC, SALB, antithrombin, platelet count, or any other clinicopathologic variable, and any TEG parameter was identified by a Spearman's rank correlation test.

Figure 1.

TEG parameters in dogs with PLN (n = 28) vs. control (n = 8). Statistical significance between groups is denoted by the symbol a. (A) Reaction time (R) is not significantly different between PLN and control dogs (= .32). (B) Clot formation time (K) is significantly lower in dogs with PLN than control dogs. (= .004). (C) α-angle is significantly higher in dogs with PLN than control dogs (= .001). (D) Maximum amplitude (MA) is significantly higher in dogs with PLN than controls, (a:< .001).

Figure 2.

(A) MA is significantly higher, as denoted by the symbol a, in both normoalbuminemic (n = 9: P < .001) and hypoalbuminemic (n = 19: < .001) PLN dogs compared with control (n = 8) dogs. (B) MA is significantly higher, as denoted by the symbol a, in PLN dogs with normal (n = 24: < .05) or low (n = 4: < .05) antithrombin compared with control dogs (n = 8).

Table 1. Selected clinicopathologic variables
 PLN Median (Range)Control Median (Range)Reference Range P
  1. ND = not determined.

  2. Reference ranges provided by the North Carolina State University Clinical Pathology laboratory except where noted. Median not determined for D-dimer concentration because the laboratory reports a range for each patient rather than an absolute number.

  3. a

    Test performed and reference range provided by Animal Health Diagnostic Center, Cornell University.

PCV (%)42 (23–51)50.5 (48–58)39–58.0001
Platelet count (103/μL)370 (98–722)224 (180–341)190–468.007
BUN (mg/dL)40.0 (8–180)15 (11–24)6–26.04
Creatinine (mg/dL)1.7 (0.4–8)0.8 (0.7–1.1)0.7–1.5.24
Albumin (g/dL)2.6 (1.5–4.3)3.6 (3.5–3.9)3.0–3.9.001
Cholesterol (mg/dL)296 (161–481)195 (142–349)124–344.0024
Antithrombin (%)a82.5 (23–145)ND65–145ND
Urine Specific Gravity1.017 (1.008–1.049)1.043 (1.013–1.057)1.015–1.045<.001
Urine protein-creatinine6.3 (2.1–25.0)ND<0.5ND
D-dimer (n = 4)ND (<250–2000)NDNDND
Fibrinogen (n = 4)450 (300–700)ND100–300ND
PT (n = 4)8.5 (7.9–8.6)ND6.8–10.7ND
aPTT (n = 4)12.7 (10.9–25.8)ND7.5–13.8ND
Table 2. TEG results classified by clinicopathologic data
TEG ValueControl Median (Range)All PLN Median (Range)PLN Normal Albumin Median (Range)PLN Low AlbuminMedian (Range)PLN Normal AT Median (Range)PLN Low AT Median (Range)PLN High Platelets Median (Range) PLN Normal/Low Platelets Median (Range)PLN Normal PCV Median (Range)PLN Low PCV Median (Range)PLN Non-Azotemic Median (Range)PLN Azotemic Median (Range)PLN Male Median (Range)PLN Female Median (Range)PLN ≤ 6 years of AgePLN > 6 years of Age
  1. Platelet count could not be determined in 1 dog because of clumping.

  2. a

    Values that were statistically different (P < .05) from control population.

  3. b

    Statistically significant differences between the groups based on specified parameters (P < .05).

R4.4 (3.8–8.3)4.0 (1.2–8.3)3.9 (3.1–8.3)4.0 (1.2–7.6)4.0 (3.0–8.3)3.8 (1.2–6.0)3.8 (1.2–7.6)4.1 (2.5–8.3)4.0 (3.0–8.3)4.0 (1.2–7.6)4.3 (3.0–6.0)3.7 (1.2–8.3)4.2 (3.1–8.3)3.8 (1.2–7.6)3.8 (3.2–5.2)4.3 (1.2–8.3)
K1.9 (1.4–3.1)1.3a (0.8–3.3)1.3a(0.9–3.3)1.2a(0.8–1.9)1.2a(0.8–3.3)1.2 (0.8–1.6)1.2a(0.8–1.5)1.3a(0.8–3.3)1.3a(0.9–3.3)1.1a(0.8–1.9)1.3a(0.8–1.7)1.2a(0.8–3.3)1.3a(0.9–3.3)1.2a(0.8–1.9)1.1a(0.8–1.9)1.3a(0.8–3.3)
α-angle64.5 (49.7–69.1)71.6a(50.8–82.7)69.2a(50.8–75.8)72.2a(64.0–82.7)71.6a (50.8–80.1)72.9a (68.1–86.9)73.1a (65.6–82.7)69.8a (50.8–79.3)70.2a (50.8–75.8)74.4a (64.0–82.7)70.2a (66.1–86.9)73.6a (50.8–82.7)72.2a (50.8–77.4)71.1a (64.0–82.7)74.9a (63.0–80.1)69.8a (50.8–82.7)
MA57.1 (48.0–59.2)71.5a (60.2–86.9)67.4a (62.1–76.9)73.1a (60.2–86.9)71.5a (62.1–82.4)71.9a (60.2–86.9)76.9a (67.0–86.9)70.9a,b(60.2–80.4)70.9a (62.1–80.4)76.3a (60.2–86.9)71.5a (66.1–86.9)73.3a (60.2–82.4)71.3a (60.2–86.9)73a (66.1–82.4)74.2a (66.1–80.6)70.9a (60.2–86.9)
G6.7 (4.7–7.3)12.5a (7.6–33.0)11a (8.2–16.7)13.6a,b(7.6–33.0)12.5a (8.1–23.5)13.2a (7.6–33.0)16.7a (10.2–33.0)12.2a (7.6–20.6)12.2a (8.2–20.6)16.1a (7.6–33.0)12.5a (9.8–33)13.9a (7.5–23.5)12.4a (7.6–33)13.5a (9.8–23.5)14.5a (9.8–20.9)12.2a (7.6–33.0)

Packed cell volume (PCV) previously has been reported to affect TEG results when increased or when severe anemia is present.[18],12,13 PCV was significantly different in control versus PLN dogs (< .001), and was below the reference range in 12/28 dogs with PLN, with the lowest PCV being 23%. When anemic dogs were removed from the analysis, results of initially significant TEG parameters remained statistically significant (< .05).

Platelet count also was significantly different between control and PLN dogs (Table 1; = .007) and also has been reported to affect TEG results.14 Platelet count was outside of the reference range in 10/28 dogs (thrombocytopenia, 3 dogs; thrombocytosis, 7 dogs) and could not be determined because of clumping in 1 additional dog. When dogs with abnormal platelet counts were excluded from analysis, results of initially significant TEG parameters remained statistically significant (< .05).

Proteinuria was assessed in control dogs by dipstick analysis. Three dogs were negative for protein (USG 1.013–1.041), 4 had trace proteinuria (USG 1.032–1.057), and 1 dog had 1+ proteinuria (USG 1.052).

Coagulation Panel Testing and Presence of Thromboembolic Disease

Coagulation panel testing (fibrinogen, D-dimer, PT, aPTT) was performed concurrently with TEG analyses in only 4 dogs. Results are included in Table 1, but no conclusions can be drawn because of the small number of patients that had coagulation profiles examined at the time TEG was performed.

Four (14%) PLN dogs had documented thromboembolic (TE) disease. Two dogs were diagnosed antemortem, and 2 dogs had thrombi demonstrated during postmortem examination. Dogs with documented TE disease had significantly lower SALB than did dogs in which no thrombi were documented (< .05). UPC was significantly higher in dogs with documented TE disease (< .05) and was correlated with presence of thromboembolism. Antithrombin activity was not significantly different between the 2 groups (= .34).


The results of this study indicate that dogs with PLN have TEG values that are hypercoagulable compared with a control group, and there was no correlation between TEG values that were significantly different between control and PLN dogs and SALB, antithrombin activity, and UPC. Considering hypercoagulability to be defined as ≥1 TEG values above (α-angle, MA) or below (R or K) the reference range, established as the mean plus or minus 2 standard deviations of our control group, 27 of 28 PLN dogs had a hypercoagulable TEG result. PLN dogs with normal SALB or antithrombin activity also were hypercoagulable, as assessed by TEG, relative to controls, indicating that hypercoagulability may develop before alterations in SALB or antithrombin activity. The number of dogs with TE was too low to adequately assess correlation between TE and any TEG parameters.

The TEG parameter MA was significantly higher in PLN than in control dogs. MA indicates final clot strength, is a representation of platelet and fibrin binding, and is affected by platelet count and activation.[8, 9, 18] This finding is supported by results of studies in humans, demonstrating that platelet activation is of critical importance in the pathophysiology of PLN-induced hypercoagulability.[4, 5] Also, R was not significantly different between PLN and control dogs, suggesting that the intrinsic pathway and initial fibrin generation may not be involved to the same extent in the pathophysiology of hypercoagulability as compared with other coagulation parameters.[8]

K value was decreased and α-angle increased in dogs with PLN when compared with controls. K is a measure of time from initial clot formation to a predetermined clot strength, whereas α-angle is a measure of clot kinetics. Both are influenced by factors II and VIII, platelet count and function, thrombin formation, fibrin precipitation, fibrinogen concentration, and hematocrit.[8, 9, 17] In this study, alterations in platelet count and PCV could not fully account for the altered K and α-angle values in dogs with PLN, because the difference from control dogs remained statistically significant when dogs with abnormal platelet count or PCV were removed from analysis. An insufficient number of dogs had fibrinogen concentration measured to allow determination of fibrinogen's effect on TEG parameters. Because PLN dogs tended to have normal R and increased MA, an interaction between platelets and fibrin in the amplification and propagation phases of clot formation is suspected as the etiology of hypercoagulability in this study.

PCV and platelet count have been reported to affect TEG parameters.[18],10,11,12 In a study of the isolated effect of decreased hematocrit on TEG, clot kinetics were increased when the hematocrit decreased from 40 to 30%, but a significant change in MA or G was not seen until hematocrit decreased to 20%.[19] None of the patients in this study had a hematocrit < 20%. Platelet count has been demonstrated to directly influence TEG variables. Thrombocytopenia does not typically introduce a substantial change until counts decrease below 100,000/μL[19]; only 1 dog in this study had a platelet count < 100,000/μL (98,000/μL). The effects of thrombocytosis are not as clearly defined but do appear to contribute to a hypercoagulable state.[21] Although PCV and platelet counts in PLN dogs were significantly different than controls, TEG parameters that were initially different between PLN and control dogs remained significantly different when dogs with abnormalities in PCV and platelet counts were removed from analysis, suggesting that these alterations were not solely responsible for the differences in TEG parameters seen in this study.

The results of this study indicate that in dogs, as in humans, the cause of hypercoagulability in PLN is multifactorial. Evaluation of only 1 component (or its marker) of the coagulation system, such as antithrombin activity, may not be sufficient to fully assess a patient's global coagulation status. Although not well-studied in dogs, humans with PLN have a multifactorial etiology of hypercoagulability, including platelet activation, hyperfibrinogenemia, increased cofactors necessary for clotting (factors V and VIII), decreased antithrombin, thrombocytosis, decreased red blood cell deformability, and increased von Willebrand factor. Fibrinogen and α2-macroglobulin accumulate, and fibrinolysis is inhibited.[4, 5] Vascular stasis associated with intravascular volume depletion also may predispose to thromboembolism in some patients. Therefore, measurement of a single parameter to determine hypercoagulability and predisposition for thromboembolism is unlikely to be sufficient for determination of thromboembolic risk; a more comprehensive evaluation of the coagulation system is preferred.

In contrast to a commonly cited report,[3] SALB and antithrombin activity were only loosely correlated statistically in this study, indicating that SALB cannot reliably predict antithrombin activity in an individual patient. This finding also is supported by 2 recent studies. In a study of dogs with glomerular disease, only moderate to weak correlations between SALB and antithrombin activity were found (R2 = 0.39 in dogs with nephrotic syndrome and 0.22 in dogs with non-nephrotic glomerular disease).[22] In a study of the prognostic utility of antithrombin activity in dogs, only a weak correlation between SALB and antithrombin activity was identified (R = 0.3).[23] Reports of an association of antithrombin deficiency and risk of thromboembolism also are conflicting in humans.[4] Regardless, antithrombin has not been demonstrated to be predictive of thromboembolic risk.

The current standard of care in dogs with PLN is to treat with an inhibitor of platelet aggregation (eg, aspirin) after hypoalbuminemia has developed.[24] This recommendation, drawn from the predisposition for thrombosis in dogs with PLN reported in retrospective studies (rather than prospective risk-based analyses) is based upon a previously reported correlation between hypoalbuminemia and low-antithrombin activity[3] and the assumption that low antithrombin is the major if not sole contributor to the hypercoagulable state. Additional studies with larger numbers of patients are required to determine if a correlation exists between hypercoagulability demonstrated by TEG and thromboembolic disease. However, based on the lack of correlation between TEG variables and SALB, UPC, or antithrombin activity in this study, the recommendation to initiate platelet inhibition after development of hypoalbuminemia may require modification. This conclusion is supported not only by the findings of this study but also by other recent studies that failed to demonstrate a strong correlation between SALB and antithrombin,[22, 23] in addition to the weak correlation between the severity of hypoalbuminemia and hypercoagulability as assessed by TEG. Weak correlation between the severity of hypoalbuminemia and TEG findings also was supported by a recent study in dogs with protein-losing enteropathy.[25] Antiplatelet therapy continues to be recommended in hypoalbuminemic patients. However, many dogs with normal SALB concentrations also were hypercoagulable as assessed by TEG, and these patients may benefit from antiplatelet therapy for prevention of thromboembolism. Additional studies are necessary to evaluate risk of thromboembolic disease and preventative treatment.

Limitations of this study included small sample size and lack of standardized diagnostic testing in each patient, especially lack of coagulation panel testing. Control dogs were not age- or sex-matched to PLN dogs. Sex does not appear to affect TEG results in dogs, but the effect of age has not been evaluated in dogs to the authors' knowledge.[26] In humans, reports of age-related alterations in TEG results are conflicting, but age generally is considered to be weakly associated with hypercoagulability in the elderly.[26-29] Neonates and children also may differ in their hemostatic profiles. However, these changes do not appear to substantially affect kaolin activated TEG.[31] The median age of PLN dogs in this study was 8 years compared with 3.5 years in controls, but results of all initially significant TEG findings remained significant when PLN dogs > 6 years of age were removed from analysis to exclude older dogs (Table 2). Control dogs were assessed for proteinuria by dipstick analysis alone. According to a previously published study assessing the use of urine dipstick analysis for exclusion of proteinuria, it is unlikely that any of these dogs had clinically relevant proteinuria as it was suggested that these dipstick results should be interpreted as negative.[32]

In addition, only a small number of dogs had complete evaluation for thromboembolic disease, and TEG findings could not be correlated with presence of thromboembolism in enough dogs to draw strong conclusions. Finally, several different types of TEG assays have been reported, and the optimal assay for clinical use has not been determined.[25] Kaolin-activated TEG has low-preanalytic variation, and we attempted to control for variability by standardized blood collection and assay techniques. Blood collections were not considered to be traumatic and any effect of venipuncture quality was mitigated by discarding the first sample obtained by vacutainer.[33]

In conclusion, dogs with PLN demonstrate a tendency for hypercoagulability as assessed by TEG that is not directly associated with antithrombin activity or severity of hypoalbuminemia. TEG is recommended for evaluation for hypercoagulability in patients with PLN. The point at which to initiate antiplatelet therapy in PLN may require modification because the presence of hypercoagulability as assessed by TEG preceded development of hypoalbuminemia or decreased antithrombin activity. Prospective studies correlating TEG findings with the incidence of TE disease are required.


The authors thank Tonya Harris for her assistance in completing the project.

Conflict of Interest Declaration: Authors disclose no conflict of interest.


  1. 1

    Vacutainer® blood collection set, Becton Dickinson, Franklin Lakes, NJ

  2. 2

    Vacutainer® evacuated blood collection tubes, Becton Dickinson

  3. 3

    TEG 5000 Hemostasis Analyzer, Haemoscope, part of Haemonetics Corporation, Braintree, MA

  4. 4

    Disposable TEG cups and pins, 0.2M calcium chloride, and kaolin vials obtained from Haemoscope, part of Haemonetics Corporation

  5. 5

    STart® 4 Hemostasis Analyzer, Diagnostica Stago, Inc, Parsippany, NJ

  6. 6

    Thromboplastin C Plus reagent, Dade Behring, Deerfield, IL

  7. 7

    Actin Activated Cephaloplastin reagent, Dade Behring, Deerfield, IL

  8. 8

    Minutex D-dimer Latex, Biopool US Inc, Ventura, CA

  9. 9

    Bayer Advia 120 Automated Hematology Analyzer, Siemens Diagnostics, Deerfield, IL

  10. 10

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

  11. 11

    SigmaStat, Jandel Scientific, San Rafael, CA

  12. 12

    Jaquith SD, Brown AJ, Scott MA. Effects of decreased hematocrit on canine thromboelastography. J Vet Emerg Crit Care 2009;19(Suppl 1):A4 (abstract)

  13. 13

    Vilar P, Hansell J, Westendorf N, Iazbik MC, Marín L, Couto CG. Effects of hematocrit on thromboelastography tracings in dogs. J Vet Intern Med 2008;22:774 (abstract)

  14. 14

    Jaquith SD, Brown AJ, Scott MA. Effects of decreased platelet count on canine thromboelastography. Vet Clin Pathol 2010;39:538 (abstract)