Canine Cushing's syndrome or hyperadrenocorticism (HAC) is a common canine endocrinopathy that is not usually life threatening. However, many clinicians believe that HAC can lead to thromboembolic events, such as pulmonary thromboembolism (PTE).[1-5] It has been presumed that hypercoagulability is the proximate cause of this anecdotal PTE risk, although this has not been documented. A hypercoagulable state caused by increased procoagulant activity, decreased fibrinolysis, or both could predispose dogs to PTE, deep venous thrombosis, or both. Epidemiologic studies in human medicine have documented an increased risk of thromboembolism associated with hypercortisolism for which varying theories of causality have been proposed.[6-9] The belief that canine HAC leads to an increased incidence of thromboembolic disease derives from anecdotes, case series,[1, 10, 11] and studies of dogs diagnosed with PTE, in which evidence supportive of HAC was retrospectively identified.[2, 3, 5] If HAC dogs are indeed at increased risk of PTE, coagulation monitoring, and thromboprophylaxis should be considered as additions to the standard of care.
One recent study using thrombelastography (TEG) and standard coagulation assays did not identify an association between hypercoagulability and HAC in dogs, whereas another study did identify increased maximum amplitude (MA) in HAC dogs compared with controls.a The absence of an association in the first study could have several possible explanations. There truly may be no association between hypercoagulability and HAC in dogs, a hypercoagulable state associated with HAC may be an endothelial or other in vivo phenomenon that was not detected by the study methodology, or hypercoagulability may only exist in a subset of HAC dogs. In humans, dyslipidemias, obesity, and diabetes mellitus (DM) are associated with an increased tendency toward pathologic thrombosis, and those conditions are common in HAC dogs.[13-15]
The goal of this study was to document the existence of a hypercoagulable tendency in dogs with HAC using a coagulation panel that evaluated individual coagulation factors and endogenous anticoagulants (fibrinogen, antithrombin [AT], and Factor VIII coagulant activity [FVIII : C]), fibrinolytic activity (D-dimers [DD], TEG clot lysis parameters [CL30, CL60, LY30, LY60]), and assays of systemic coagulation (prothrombin time [PT], activated partial thromboplastin time [aPTT], thrombin-antithrombin complexes [TAT], and TEG). We hypothesized that dogs with HAC would have a hypercoagulable tendency, caused by increased procoagulant activity, decreased fibrinolysis, or both. A secondary goal was to evaluate common clinical and biochemical markers in dogs with HAC to identify those at increased risk of hypercoagulability.
Materials and Methods
- Top of page
- Materials and Methods
Dogs suspected of having HAC that presented to the University of Georgia Veterinary Teaching Hospital (UGAVTH) between November 2009 and January 2011 were prospectively considered for enrollment. Dogs were included if they were newly diagnosed with HAC, as defined by compatible clinical and clinicopathologic findings, and 1-hour post-ACTH stimulation cortisol concentration >22 μg/dL. Compatible clinical signs included some combination of polydipsia, polyuria, polyphagia, endocrine alopecia, panting, and abdominal distension. Compatible laboratory findings included some combination of isosthenuria, increased serum ALP activity, neutrophilia with lymphopenia, and thrombocytosis. To be enrolled, a combination of ≥3 compatible clinical and laboratory findings and a 1-hour post-ACTH stimulation cortisol concentration >22 μg/dL were required at initial evaluation. Exclusion criteria for this study were patients weighing <2 kg, inability to safely collect 12 mL of blood, history of anesthesia or surgery within 10 days preceding the evaluation, or presence of atypical HAC as defined by post-ACTH stimulation cortisol ≤22 μg/dL with increases in ≥2 post-ACTH stimulation cortisol precursors. Dogs with documented systemic infections, adrenal tumors showing vascular invasion as demonstrated on abdominal ultrasound examination, prior history of treatment for HAC, or current use of any medications known to interfere with coagulation or platelet function also were excluded. Patients with concurrent endocrinopathies, skin infections, musculoskeletal abnormalities, or neurologic signs were not excluded. Informed consent was obtained from all owners and all procedures were approved by the UGAVTH Clinical Research Committee. A complete medical history was obtained and a physical examination was performed on each dog. Comorbid conditions (including concurrent endocrinopathies, infections, hepatic, renal, cardiac or neurologic disease, current medications, hypertension, hypertriglyceridemia, marked hypercortisolemia [poststimulation cortisol >30 μg/dL] and proteinuria), whether previously diagnosed or identified at the time of enrollment, were recorded.
Dogs were fasted for 12 hours before presentation at UGAVTH. Enrolled dogs underwent clinicopathologic screening and coagulation assessments (detailed below) as well as blood pressure measurement, and abdominal ultrasound examination. At initial evaluation, blood was collectedb for measurement of serum cortisol concentration,c complete blood count (CBC),d serum biochemical profile (including cholesterol and triglycerides),e paired serum insulin and glucose concentrations, and coagulation testing (collected in 3.2% sodium citrate with a final citrate : blood ratio of 1 : 9). Exogenous insulin was not withheld from diabetic dogs before blood sampling. Samples were collected preferentially from the lateral saphenous or cephalic vein; the jugular vein was used only if a peripheral vein was inadequate for sample volume. All samples were collected in a single venipuncture, using a butterfly catheter and syringes, by 1 of the investigators (SLP) or by an experienced registered veterinary technician. Five μg/kg of cosyntropinf was given IV to each patient immediately after the initial blood sample was taken using the butterfly catheter already in place. Urine was collected for urinalysis and urine protein : creatinine ratio. A second blood sample was collected 1 hour later for the measurement of cortisol concentration.
Measured CBC parameters included hematocrit (HCT), leukocyte count and differential, platelet count (PLT), mean platelet component (MPC) and calculated plateletcrit (PCT). The plateletcrit is the platelet equivalent of the hematocrit, representing the relative volume of platelets in a blood sample. The plateletcrit is measured directly on a centrifuged whole blood sample or, like hematocrit, can be approximated by multiplying the platelet count by the average platelet volume (MPV). Citrated plasma was obtained by centrifugation (1,500 × g for 10 minutes) within 30 minutes of collection and was used for measurement of fibrinogen, PT, aPTT, AT, DD, TAT, and FVIII : C. Analyses were batched, and plasma was stored frozen at −80°C until analysis. Serum for the measurement of insulin concentration also was stored at −80°C until batch analysis every 6 months. All other analyses were performed on the same day.
One tube of citrated whole blood was used for TEG analysis.g TEG samples were kept at room temperature (22–24°C) for 30 minutes before analysis. For each sample, 340 μL of citrated blood was added to a prewarmed cup (37°C) containing 20 μL of 0.2 M CaCl2.h,[] Duplicate samples were run for 60 minutes after the measurement of maximal amplitude (MA). Each TEG tracing reported R value (minutes), K value (minutes), α-angle (degrees), and MA value (millimeters). The parameters G, CI, LY30, CL30, LY60, and CL60 also were measured. G, the elastic shear modulus, was calculated using the equation: G = (5,000 × MA/[100 − MA]). CI, the coagulation index, was calculated using the equation −0.245(R) + 0.0184(K) + 0.1655(MA) − 0.0241(α-angle) − 5.0220. All values were compared to the institutional reference intervals.
PT, aPTT, fibrinogen concentrations, and DDi as well as AT activityj were determined at UGAVTH Clinical Pathology Laboratory. TAT complex concentration was measured using a commercial human TAT ELISA, previously reported for use in dogs and cats.k, Factor VIII : C was measured using a modified one-stage aPTT assay as previously described. At 6-month intervals, frozen plasma and serum samples were shipped overnight on dry ice to Cornell University Animal Health Diagnostic Center, Comparative Coagulation Laboratory (TAT and FVIII) and to Michigan State University, Clinical Pathology Laboratory (insulin and glucose) for analysis.
Urine pH and specific gravity were performed immediately after urine collection by cystocentesis, and urine protein : creatinine ratios (UP : C) were measured on an automated biochemistry analyzer.e Creatinine was measured with the Jaffe reaction and urine protein was measured turbidometrically with addition of benzethonium chloride.
Systolic blood pressure was measured by Doppler ultrasonic flow probe.l Patients were allowed to rest calmly for 10–20 minutes before measurement, and then 3–5 measurements were averaged to obtain the recorded blood pressure. All blood pressures were measured by an experienced veterinary technician, using cuff widths approximately 40% of the circumference of the patients' limbs.
Study population results for each coagulation assay were evaluated discretely (normal versus abnormal). Also, we included a subgroup of these tests in a proposed hypercoagulability panel including TAT, FVIII : C, α-angle, MA, CI, and CL60. The hypercoagulability panel was evaluated for each enrolled dog to identify the presence of any individual abnormal parameter implying a hypercoagulable tendency as indicated by increased procoagulant activity or decreased fibrinolysis. Statistical analysis was performed using commercially available software.m Group values are described as mean ± SD if distribution was parametric, and median (range) if distribution was nonparametric. A Student's t-test was used to compare coagulability outcomes (TAT, FVIII : C, α-angle, MA, CI, and CL60) between dogs with and without DM, phenobarbital therapy, blood pressure >160 mmHg and postcortisol concentration >30 μg/dL. Simple linear regression was used to test for relationships between continuous factors and coagulability assays, and among coagulability assays. For outcomes where multiple factors were significant a multiple regression also was performed. All hypothesis tests were 2-sided and the significance level was α = .05.
- Top of page
- Materials and Methods
Seventeen dogs qualified for enrollment. An additional 10 dogs were considered for enrollment based on presenting signs but did not have post-ACTH stimulation serum cortisol concentrations ≥22 μg/dL. Of the 17 enrolled dogs, there were 10 neutered males and 7 spayed females. Median age was 11.2 years (range, 7–13.1 years). Predominantly small and toy breeds were represented including 3 Beagles, 3 Miniature Schnauzers, 2 mixed breed dogs, 2 Cocker Spaniels, and 1 each of Maltese, Dachshund, Shih Tzu, Yorkshire Terrier, Pekingese, Pomeranian and Lhasa Apso. Dogs had a median weight of 10.0 kg (range, 3.3–29.4 kg). Clinical signs had been present for an average of 17 ± 13 weeks. Of comorbid conditions recorded, 3 dogs (17.6%) had a history of DM and 2 dogs (11.8%) were under treatment with phenobarbital for idiopathic epilepsy.
Results of HCT, platelet parameters and selected biochemical markers are summarized in Table 1. Platelet parameters were not available for 1 dog because of analyzerc malfunction. One dog had HCT above, and 1 dog below, the HCT reference interval (RI). For both platelet counts and PCT, the group median (PLT 498 × 103/μL; range, 286–758 × 103/μL and PCT 0.58%; range, 0.34–1.11%) was greater than RI. MPC was not greater than RI in any dog, but 5 dogs (31.2%) had values lower than RI. Sixteen of 17 (94.1%) dogs had increased ALP activity, and 8 (47.1%) dogs had increased ALT activity. Most dogs had increased cholesterol (n = 13; 76.4%), triglycerides (n = 12; 70.6%), and fasting insulin (n = 15; 88.2%) concentrations as well as increased UP : C ratios (n = 13; 76.4%). Only 2 dogs were hyperglycemic, but the majority (n = 13; 76.4%) had increased I : G ratios; one of these 13 dogs also was diabetic and receiving exogenous insulin. Eight (47.1%) dogs had systolic blood pressure >160 mmHg. Inclusion criteria specified a poststimulation cortisol concentration ≥22 μg/dL; however, 6 dogs (35.2%) had a markedly increased poststimulation cortisol concentration ≥30 μg/dL. For those 6 dogs, the markedly increased cortisol concentration median was 40.4 μg/dL (range, 30.7–46.0 μg/dL; data not shown).
Table 1. Selected platelet, biochemical, and physiologic parameters in 17 dogs with hyperadrenocorticism
|Parameter||Reference Interval||Mean||SD||Median||Minimum||Maximum||n > RI|
|Hematocrit (%)||36.6–59.6|| || ||48.6||36.4||59.7||0|
|Platelet count (×103/μL)a||156–394b|| || ||498||286||758||13|
|Plateletcrit (%)a||0.18–0.43b|| || ||0.58||0.34||1.11||13|
|MPC (g/dL)a||18.9–23.7|| || ||21.25||16.6||23.3||0|
|ALP (U/L)||13–122|| || ||847||90||2,951||16|
|ALT (U/L)||12–108|| || ||106||21||391||8|
|Cholesterol (mg/dL)||129–264|| || ||326||189||1,117||13|
|Triglycerides (mg/dL)||26–138|| || ||174||79||1,240||12|
|UP : C||0–0.3|| || ||2.0||0.0||24.8||13|
|Insulin (pmol/L)||58–229|| || ||374||159||1,441||15|
|Glucose (mmol/L)||4.4–6.7|| || ||4.9||4.4||15.3||2|
|I : G||14–43|| || ||68||21||314||13|
|Blood pressure (mmHg)||110–160||160||28|| ||96||205||8|
Coagulation parameters are summarized in Table 2. Four (23.5%) dogs had prolonged aPTT, 1 dog (5.9%) had increased DD, and 3 (17.6%) dogs had fibrinogen concentrations above RI, whereas 1 (5.9%) had a fibrinogen concentration below RI. Ten dogs (58.8%) had AT activity above RI. Nine (52.9%) dogs had TAT concentrations above the RI, and 1 dog had a TAT concentration below RI. There was one dog with a TAT concentration of 498.5 μg/dL; the next highest concentration was 32.6 μg/L. Five (29.4%) dogs had FVIII : C activities above RI.
Table 2. Coagulation parameters in 17 dogs with hyperadrenocorticism
|Parameter||Reference Interval||Mean||SD||Median||Minimum||Maximum||n < RI||n > RI|
|aPTT (seconds)||9.4–15.1||13.3||2.1|| ||9.7||17.4||0||4|
|PT (seconds)||5.8–9.8|| || ||7.0||6.3||9.4||0||0|
|Fibrinogen (mg/dL)||150–490||374||155|| ||92||706||1||3|
|D-dimers (ng/mL)||0–250|| || ||37||0||461||0||1|
|AT (%)||80–120|| || ||124||109||188||0||10|
|TAT complex (μg/L)||1.0–8.0|| || ||8.3||0.0||498.5||1||9|
|Factor VIII : C (%)||50–200|| || ||107.0||35.5||1,304.0||1||5|
|R (minutes)||2.1–11.0||4.5||2.0|| ||2.0||8.6|| || |
|K (minutes)||1.2–4.6||1.8||0.8|| ||0.9||4.2|| || |
|α- angle (degrees)||39–74|| || ||66.9||42.0||75.9||0||3|
|MA (degrees)||44.5–61.7||65.6||5.0|| ||55.7||73.5||0||13|
|G (dyne/cm5)||4,010–8,100||9,998||2,422|| ||6,286||14,095||0||13|
|LY30 (%)||0–1.7|| || ||0||0.0||1.0|| || |
|CL30 (%)||94.2–100|| || ||100||97.2||100.0|| || |
|LY60 (%)||0.2–5|| || ||0.4||0||3.4||3||0|
|CL60 (%)||87–97|| || ||97.6||90.9||100||0||8|
On the TEG, 3 (17.6%) individual dogs had α-angles above RI and 13 (76.4%) dogs had MA values above RI. The elastic shear modulus (G) parameter results mirrored those of the MA, and were not analyzed further. Nine dogs (52.9%) showed a CI above RI. Lysis parameters CL and LY were assessed at 30 and 60 minutes after generation of MA. All dogs were within RIs at 30 minutes. Three dogs were outside of the LY RI, and 8 were outside of the CL RI at 60 minutes. All dogs outside of the lysis RIs had tracings indicative of decreased fibrinolysis.
Each dog's hypercoagulability panel was evaluated for the presence of any individual marker of a hypercoagulable tendency. Using this approach, 15 of 17 HAC dogs (88.2%) exhibited a hypercoagulable tendency. Five dogs had 2 abnormal parameters (1 increased MA and CI, 2 increased MA and CL60, 1 increased TAT and FVIII : C, 1 increased FVIII : C and CL60). Five dogs had 3 abnormal parameters (2 increased TAT, MA and CL60, 1 increased MA, CI, and CL60, 1 increased TAT, MA, and CI, 1 increased α-angle, MA, and CI). The remaining 5 dogs had elevations in 4 parameters (3 with increases in TAT, FVIII : C and MA and CI, and 2 with increases in TAT, MA, CI, and CL60).
No significant effects of DM, phenobarbital therapy, hypertension, increased UP : C, post-ACTH cortisol concentration ≥30 μg/dL, or duration of clinical signs on assays included in the hypercoagulability panel (TAT, FVIII : C, α angle, MA, CI, or CL60) were found. Linear regression indicated inverse relationships between fibrinogen concentration and TAT complexes (R2 = .26, P = .043), fibrinogen concentration and FVIII : C (R2 = .38, P = .008), and FVIII : C and MA (R2 = .29, P = .024). Significant positive relationships with fibrinogen concentration were found for MA (R2 = .31, P = .018) and CL30 (R2 = .29, P = .030). MA was associated with PCT (R2 = .30, P = .028) and with serum ALT (R2 = −.27, P = .031), but not with HCT (P = .533). In multiple regression, a positive relationship was found between PCT (P = .0430) and fibrinogen concentration (P = .0475) with MA levels.
- Top of page
- Materials and Methods
In this study of 17 dogs with HAC, the majority (15; 88.2%) exhibited a hypercoagulable tendency characterized by an abnormality in at least 1 component of the proposed hypercoagulability panel. In fact, each of these 15 dogs exhibited >1 abnormality in the hypercoagulability panel, but no consistent pattern of concurrent coagulation abnormalities was detected. TEG MA (13 dogs) was the most commonly increased parameter, with increased TAT complexes (9 dogs), suggestive of ongoing thrombin generation, as the next most common indicator. Only one of the dogs with increased TAT complexes had a concurrent increase in DD, indicating ongoing fibrinolysis. Fewer dogs (5 dogs) had increased FVIII : C or increased TEG α-angle (3 dogs). Eight dogs had evidence of decreased fibrinolysis, and all of these dogs also had other indications of increased procoagulant activity by TEG or other tests.
These findings suggest that the HAC dogs in this study exhibited a hypercoagulable tendency. However, not all parameters were abnormal in all dogs, and abnormalities in 1 coagulation assay did not predict abnormalities in others. Increased insulin : glucose ratios, which have been associated with hypercoagulable tendencies, were documented in 13 dogs (including 1 of 3 diabetic dogs), but did not correlate with any of the coagulation assays. Furthermore, neither comorbid conditions, nor increased laboratory parameters, nor duration of clinical signs were found to predict hypercoagulability in any of the subgroup of assays included in the hypercoagulability panel. However, because dogs were recruited for their suspected HAC and not the presence of any comorbidity or laboratory abnormality, only small numbers of dogs with each potential predictive variable were enrolled. This resulted in a low-powered analysis of predictive variables, which is discussed further below.
The interest in hypercoagulability in HAC derives from a series of reports of PTE in dogs with HAC. The actual risk of PTE is unknown in this population, and a prospective study of PTE in HAC dogs has not been reported. It has been presumed that hypercoagulability is the proximate cause of potential PTE risk, but this has not been documented, nor has the etiology of the presumed hypercoagulability been elucidated. Excess circulating cortisol is commonly theorized as the cause of hypercoagulability, because it can result in increased fibrinogen concentrations, decreased AT concentrations, decreased fibrinolysis,[8, 28] or thrombocytosis. A relative increase in the concentrations of circulating procoagulant molecules, such as FVIII and fibrinogen, also has been suggested as a cause.[7, 28] Two recent studies documented the development of hypercoagulability in dogs that were receiving prednisone therapy,[23, 29] whereas alternate reports documented increased MA in HAC dogs compared with controlsa and delayed fibrinolysis in dogs with spontaneous HAC.n
Because thromboembolic events in HAC dogs occur suddenly and sporadically, little information is available to document their predisposing causes. Although hypercoagulability seems a plausible risk factor, it also has been challenging to document. Whole blood TEG is useful for evaluation of hemostatic status in dogs with systemic diseases other than HAC.[30-32] Unlike assessment of PT, aPTT, fibrinogen, and AT concentrations, from which a hypercoagulable state can only be inferred, TEG can demonstrate a tendency toward hypercoagulability. Furthermore, the cellular components of the blood that contribute to coagulation (eg, platelets and microparticles) are integrated into whole blood TEG analysis. In this report, a significant positive relationship was found between fibrinogen concentration and MA, as has been reported in other species.[33-35] PCT also had a weak correlation with MA, which may indicate the contribution of platelets to final clot strength.
Because TEG is an ex vivo system, DD, TAT complexes, and FVIII : C concentrations were evaluated to document in vivo coagulation. DD are breakdown products of cross-linked fibrin that have been used in the diagnosis of disseminated intravascular coagulopathy and PTE in humans and animals,[36-39] and which may be persistently increased in some human HAC patients. In this study, DD was only increased in a single dog. However, if HAC dogs exhibit decreased fibrinolysis, DD may not be increased in circulation even if clots have formed. Increased TAT complexes have been documented in dogs without clinical evidence of thrombosis.[25, 41] Over half of the dogs in this study had increased TAT complexes, and a significant inverse relationship was found with fibrinogen concentration. One possible explanation of this relationship is that ongoing in vivo coagulation may lower fibrinogen concentration by consumption. If so, decreased fibrinogen concentrations may suggest a hypercoagulable tendency. Further investigation would be required to explain this relationship.
One previous study has evaluated FVIII : C in HAC dogs and failed to find a significant difference from controls. In the propagation phase of coagulation, FVIII is needed for generation of large amounts of thrombin. In humans, HAC and increased insulin, fibrinogen, and triglyceride concentrations are associated with increased circulating FVIII, and increased FVIII activity has been associated with both arterial and venous thrombosis. In our study, only 5 dogs had increased FVIII : C activity, and these dogs did not all have TEG evidence of hypercoagulability (3 of 5 had MA above the RI).
Most dogs in this study had increased PLT and PCT, but normal MPC, and PCT was positively associated with MA. Because MA is an indicator of clot strength, increased platelet surface area likely allowed for greater fibrin formation and binding, resulting in increased clot strength. MPC measures the cytometric complexity of the platelets, with lower MPC associated with platelet activation and degranulation. Increased MA is the only TEG parameter that has been linked to thrombosis in humans. Due to lack of standardization in the veterinary literature, MPC was regarded as an observational variable only, and was not included in the hypercoagulability panel. Although 4 of the 5 dogs with low MPC also had other indicators of a hypercoagulable tendency on the hypercoagulability panel, no pattern of association with specific coagulation assays was identified.
Hypercortisolism is associated with insulin resistance, proteinuria, increases in cholesterol and triglyceride concentrations, and hypertension. Some people with hyperinsulinemia have increased TAT complexes and FVIII activity, indicating ongoing coagulation. Because HAC is a known cause of peripheral insulin resistance leading to hyperinsulinemia, insulin concentrations, and insulin : glucose ratios were assessed as possible hypercoagulability markers. No significant relationship was found between insulin concentration or insulin : glucose ratio and coagulation parameters. There were 3 known diabetic patients in the study, all of whom had increased insulin concentrations. However, 2 of these 3 patients had normal insulin : glucose ratios, and it does not seem likely that exogenous insulin biased the results. We did not withhold insulin from diabetic dogs before enrollment in this study because the insulin : glucose ratios actually experienced by diabetic HAC dogs under normal management conditions would be relevant to their risk of hypercoagulability.
Long-term hypercortisolism can lead to glomerular changes that can produce increased UP : C ratio and urinary loss of AT.[28, 48] Despite 13 dogs with increased UP : C, all dogs had normal AT concentrations. This may indicate that AT loss was not occurring, or that any loss was compensated by increased AT production. A recent study of dogs with increased UP : C and decreased AT because of glomerulonephritis did show hypercoagulability measured by TEG, but a relationship between TEG parameters and UP : C or AT was not identified in the current study.
Hypertension may contribute to hypercoagulability by direct endothelial damage and by alterations in blood flow. In rat models, both hypertension and hyperlipidemia cause hypercoagulability, with hyperlipidemia exerting the stronger effect. The combination of insulin resistance, dyslipidemia, hypertension, and hypercoagulability in people with HAC and Type II DM may increase the risk of thromboembolic and cardiovascular events. We did not find a correlation between blood pressure, cholesterol concentration, or triglyceride concentration and any coagulation outcome.
Finally, we considered the possibility that chronic, cumulative effects of a constellation of minor changes could produce a hypercoagulable tendency. There were no significant relationships, however, between duration of clinical signs and coagulation outcomes.
There are limitations to our study because hypercoagulability is challenging to document. Investigators use various protocols to perform the TEG assay and a universally accepted standard does not exist. In particular, the role of coagulation activators such as kaolin or tissue factor frequently is debated. When coagulation in the TEG is initiated by recalcification alone (ie, the contact activation pathway), the effects of preanalytic variables such as blood collection technique and rest temperature may become more pronounced, which was the reason for the strict protocol followed in this study.[21, 53] Additionally, variations of HCT, PLT, and fibrinogen concentration have been shown to influence TEG measurements.[33, 54] Notably, higher HCTs have been associated with relatively hypocoagulable tracings and lower HCTs with relatively hypercoagulable tracings, even when HCT falls within RI, both in vivo and in vitro.[12, 35, 54-56],o Whereas these variations in TEG tracing can be created in vitro, they also identify an incompletely understood in vivo phenomenon. That is, the mechanisms by which changes in blood viscosity and coagulation factor concentration that accompany anemia or polycythemia affect clot formation in vivo remain unclear. Thus, although the correlation of TEG parameters with HCT may be an artifact of the assay, it also represents actual variable coagulation status of patients across a range of HCT. In the present study, HCT ranged from 36 to 59% and there was no significant relationship between MA and HCT (P = .533), perhaps because of the influence of other factors, or perhaps due to the fact that they were all fairly closely grouped within the reference range. The use of TAT complexes and FVIII : C concentrations to document ongoing coagulation has been limited in veterinary medicine to this point and future work is necessary to refine our understanding of the settings in which these assays are best performed. Although we documented a hypercoagulable tendency by ≥2 parameters of the hypercoagulability panel in most HAC dogs of this report, the variability in the patterns of normal and abnormal results within each individual dog is difficult to explain. This variability precludes the ability to identify a single best coagulation assessment to employ in HAC dogs, and an individual test to do so may not exist. A panel of tests, similar to our proposed hypercoagulability panel, may be required to identify hypercoagulability in HAC dogs.
An important limitation of this study is the relatively small number of dogs that was evaluated, especially given that a minority of HAC dogs would be expected to exhibit a hypercoagulable tendency. Furthermore, our initial power calculation was based solely on increased MA values and did not separately account for the potential prevalences of the various comorbid conditions. Although several different comorbid conditions were documented, no single condition was documented frequently in our study population. Thus, these comparisons, in particular, were underpowered, and the absence of a detectable risk attributed to any of these comorbid conditions may reflect Type II error. Consequently, continued evaluation of dogs with naturally occurring HAC should be undertaken to generate a larger data bank for this disease. In addition, the analysis of outcome measures (ie, confirmed thrombosis) in relation to coagulation analysis may help to identify more important factors that are associated with thrombosis for future investigations.
Although the majority of the HAC dogs of this report exhibited a hypercoagulable tendency, reports of thrombotic events among HAC dogs remain rare and thrombosis was not seen in these dogs during the period of study. The lack of true outcome measures (ie, thrombosis associated with a hypercoagulable TEG tracing) is an important hurdle to our understanding of the best use for TEG technology in veterinary medicine. In addition, our analysis of comorbid conditions and biochemical variables did not identify a clear indicator of which dogs would have the greatest number or severity of coagulation abnormalities. Future studies evaluating thrombotic outcome measures, in combination with comprehensive coagulation screening, may help to elucidate the unresolved links between HAC, hypercoagulability, and thrombosis.