The study was carried out at The Royal Veterinary College, Hawkshead Lane, North Mymms, Herts AL9 7TA, UK.
Association between Hypercoagulability and Decreased Survival in Horses with Ischemic or Inflammatory Gastrointestinal Disease
Article first published online: 3 NOV 2010
Copyright © 2010 by the American College of Veterinary Internal Medicine
Journal of Veterinary Internal Medicine
Volume 24, Issue 6, pages 1467–1474, November/December 2010
Total views since publication: 19
How to Cite
Dunkel, B., Chan, D.L., Boston, R. and Monreal, L. (2010), Association between Hypercoagulability and Decreased Survival in Horses with Ischemic or Inflammatory Gastrointestinal Disease. Journal of Veterinary Internal Medicine, 24: 1467–1474. doi: 10.1111/j.1939-1676.2010.0620.x
- Issue published online: 3 NOV 2010
- Article first published online: 3 NOV 2010
- Submitted May 4, 2010; Revised August 2, 2010; Accepted September 8, 2010.
Background: Coagulopathies are common in horses with ischemic or inflammatory gastrointestinal (GI) disturbances. There is indirect evidence suggesting that early stages of these diseases are characterized by hypercoagulability (HC).
Hypothesis/Objectives: HC, assessed via thromboelastography (TEG), is common in horses with ischemic or inflammatory GI diseases. The degree of HC is correlated with nonsurvival and thrombotic complications.
Animals: Thirty client-owned horses with ischemic or inflammatory GI disease, 30 client-owned horses with nonischemic or inflammatory GI disease, and 30 healthy horses (control group).
Methods: Prospective, observational clinical study. TEG profiles of 30 horses with ischemic or inflammatory GI disease were obtained on admission and 48 hours after admission, and these were compared with profiles from 30 horses with nonischemic or inflammatory GI disease and 30 healthy controls. Prothrombin time (PT), activated partial thromboplastin time (aPTT), antithrombin activity (AT), and D-Dimer concentrations were also determined in horses with GI disease.
Results: Horses with ischemic or inflammatory GI disease had shorter R times compared with healthy horses (14.8 ± 8.3 versus 22.8 ± 12 minute; P= .011). However, changes were subtle and TEG profiles did not resembled those obtained from animals or humans presumed to be hypercoagulable. Although conventional coagulation testing supported the presence of HC (decreased AT and increased D-Dimer concentrations), TEG and coagulation abnormalities were rarely found in the same horses and the methods were not statistically related.
Conclusions and Clinical Importance: There is evidence of HC in horses with GI disease but techniques for diagnoses require refinement.
activated partial thromboplastin time
disseminated intravascular coagulation
Derangements of the coagulation system, including subclinical and clinical disseminated intravascular coagulation (DIC), are commonly observed in critically ill horses.1–8 The presence of hypercoagulability (HC) has been suspected in numerous studies,7–12 mainly in horses with ischemic or inflammatory gastrointestinal (GI) lesions,8–10 and has been associated with nonsurvival.8–10,13,14 More recently, the direct evidence for HC in the form of multi-organ microvascular fibrin deposits was identified in 40% of horses with ischemic/inflammatory GI disease.7
As HC is difficult to ascertain directly, surrogate markers can be used to suggest the presence of HC. Decreased antithrombin (AT), protein C, and plasminogen activities and increased thrombin-antithrombin complex (TAT) and D-Dimer concentrations have been proposed as indicators of HC in horses with various GI diseases. Presumed HC was detected in most studies early in the disease process, either on admission or during the first 72 hours.8,9,13 These findings support the assumption that HC precedes the profound hypocoaguable state associated with clinical bleeding found in horses with DIC.
The principles of thromboelastography (TEG) have been reviewed recently and its use has been described in healthy horses.15–18 Parameters derived from TEG tracings provide information about the initiation of clot formation, clot development, final clot strength, and fibrinolysis and the factors influencing each step. TEG should therefore provide a measurement of the in vivo phenotype of coagulation disorders and present a close approximation of the global coagulation status of a patient.19 As TEG allows characterization of the entire coagulation process, TEG should be superior in detecting disturbances in coagulation, particularly HC, compared with conventional plasma-based coagulation analytes. Confirming the utility of TEG analysis in equine patients could facilitate the implementation of goal directed therapy in patients with coagulation disorders. Furthermore, determination of optimal timing and the efficacy of established or novel treatments in the clinical and in the research setting would be possible.
This prospective study was designed to test the hypothesis that HC, detectable via TEG, is common in horses with ischemic or inflammatory GI disease and that HC is associated with nonsurvival and thrombotic complications. To test the hypothesis, TEG profiles from horses with ischemic or inflammatory GI disease (Group 1) were compared with profiles from horses with nonischemic or inflammatory GI disease (Group 2) and healthy horses (Group 3). The association between an abnormal TEG profile in horses with GI disease and outcome and the development of thrombotic complications was determined. Conventional coagulation tests were compared with parameters obtained from TEG tracings to investigate whether TEG provides additional insight into the coagulation status of equine patients. This was performed particularly in regards to the presence of HC, which cannot be obtained by measurement of conventional coagulation parameters alone.
Materials and Methods
The study protocol was approved by the institutional Ethics and Welfare Committee and informed client consent was obtained before enrollment of horses into the study. A single blood sample was obtained from healthy horses that either belonged to a research herd or were systemically healthy hospital patients not suffering from any vascular or hematological disorder and that had blood obtained for reasons unrelated to the study. Additional samples from 60 horses with GI disease presenting to the Equine Referral Hospital were obtained on admission (parameters obtained on admission are subscribed with an “ad”) and 48 hours after admission (subscribed with a “48 hour”). Horses with strangulating small and large intestinal lesions, enteritis, colitis, and primary and secondary peritonitis were assigned to Group 1. Diagnoses were established during exploratory laparotomy or postmortem examination. An ischemic or inflammatory lesion was also assumed to be present if peritoneal fluid analysis demonstrated a nucleated cell count >5,000/μL and a protein concentration >2.5 mg/dL, evidence of sepsis (intracellular bacteria on cytology) or intestinal rupture (fecal material and bacteria evident on cytology), and signs of systemic inflammation were present. Horses with noninflammatory or nonischemic GI disease such as large colon displacement, mild to moderate impaction, spasmodic colic, and medical colic that resolved with minimal treatment were included in Group 2 if peritoneal fluid analysis yielded a nucleated cell count <5,000/μL and a protein concentration <2.5 mg/dL and with no evidence of systemic inflammation20 present. Systemic inflammation was assumed to be present if ≥2 of the following criteria were present: fever (rectal temperature >38.6°C/101.5°F) or hypothermia (<36.6°C/98°F), tachycardia (heart rate >48 beats/min), tachypnoea (respiratory rate >16 breaths/min), or hypocapnoea (PaCO2 < 35 mmHg), leucocytosis (>14,300/μL), leucopenia (<5,000/μL), or >10% of immature granulocytes.20
Thrombotic complications were defined as clinically appreciated thrombosis, such catheter-associated thrombosis. Catheter sites were inspected and palpated daily and examined ultrasonographically if clinical signs were suspicious of thrombosis.21
Venipuncture of the jugular vein was performed by an experienced clinician with a 20 or 18 G needle. Blood was collected into a plastic syringe and immediately transferred into 4.5 mL siliconized glass vacutainersa containing a final concentration of 0.32% sodium citrate (0.105 M; dilution 1 : 9). After TEG analysis had been performed, the remaining blood was centrifuged at 1,000 ×g for 15 minute at 4°C within 1 hour of collection and plasma was stored at −80°C in aliquots for determination of prothrombin time (PT), activated partial thromboplastin time (aPTT), AT, and D-Dimer concentrations. Additionally, routine hematology was performed using an automated analyzerb on a blood sample collected in EDTA. If a platelet count <100 × 103/μL was identified, a blood smear was reviewed and a platelet count performed manually.
Blood was allowed to rest for 30 minutes following collection before TEG analysis was performedc and the TEG procedure was carried out according to the manufacturer's instructions. In brief, the blood-containing vial was inverted twice before 340 μL blood was transferred into a plain cupc containing 20 μL of 0.2 M calcium chloride; the analysis was started immediately and allowed to run until all parameters consisting of reaction time (R), speed of clot formation (K and α), overall clot strength (MA), and evidence of thrombolysis (LY30 and LY60) had been determined (Fig 1). If horses had received heparin or heparinized saline, cups containing heparinasec were used. All TEG analyses were performed by the same operator (B.D.). Evidence of HC and hypocoagulability was defined as ≥25% deviation of TEG parameters above or below the minimum and maximum values established from healthy horses (HC: decreased R or K or increased angle α or MA; hypocoagulability: increased R or K, decreased angle α or MA; Fig 1).
PT, aPTT, AT activity, and plasma D-Dimer concentration were measured in duplicate as described previously.5 Plasma D-Dimer concentration was determined using an immunoturbidititimerd with commercial reagents and controlse and PT and aPTT were determined with a semiautomatic coagulometer with commercial reagents and controls.f AT activity was determined using a chromogenic kitg in a semiautomatic analyzerh as described previously.8 All coagulation tests were performed blindly and in duplicate. Subclinical DIC was defined as the presence of ≥3 out of 5 abnormal conventional coagulation parameters (≥25% deviation from laboratory reference intervals: platelet count 100–600 × 103/μL; PT 10–12.5 seconds; aPTT 35–50 seconds; D-Dimer concentration <1,000 ng/mL; AT 160–250%) without clinical evidence of a coagulopathy on examination. Clinical DIC was defined as the presence of ≥3 out of 5 abnormal conventional coagulation parameters as described above with the presence of clinical signs indicative of a coagulopathy.4
Normality of the data was assessed using the Kolmogorov-Smirnov test. Data are presented as mean±standard deviation of the mean (SD) or median and range, as appropriate. Data were compared using an independent Student's t-test, a paired t-test or repeated measures ANOVA followed by post hoc comparison using Bonferroni's test or their nonparametric equivalents. The association between TEG and conventional coagulation parameters was assessed using logistic regression. A P value ≤.05 was considered to be statistically significant. Statistical analysis softwarei was used for all analyses.
Thirty horses with ischemic/inflammatory GI disease (Group 1) and 30 horses with nonischemic or inflammatory GI disease (Group 2) were enrolled in the study; TEG parameters from 30 healthy horses (Group 3) were used as controls. The mean age of all horses was 12.8 ± 5.8 years. A large variety of breeds was present, including 25 Irish drafts, Irish draft crosses and other draft breeds, 19 Thoroughbreds/Thoroughbreds crosses, 23 Warmblood/Warmblood crosses, 20 pony/pony cross breeds including Cobs and 2 Arabians. Signalment and physical parameters obtained on admission are summarized in Table 1. Group 1 included 16 horses with strangulating small intestinal lesions, 5 horses with large colon torsion, 4 horses with colitis, 2 horses with intestinal rupture and 1 horse each with cecocolic intussusceptions, ileocecal intussusceptions and 1 horse with a colon displacement with ischemic compromise to the bowel wall; 20 horses in Group 1 underwent surgical exploration of the abdomen. Group 2 included 10 medical colic cases which resolved with no or minimal interventions, 10 horses with large colon displacements (4 right dorsal, 3 left dorsal, and 3 nonspecified displacements), 6 horses with impactions (5 pelvic flexure and 1 ileal impaction), and 1 horse each with a retroflexed colon, equine dysautonomia and chronic colic and 1 horse with small intestinal ileus of undetermined origin in conjunction with a retroflexed ascending colon. The reason of the small intestinal ileus in this horse could not be determined; however, as peritoneal fluid analysis was normal, the horse showed no signs of systemic inflammation and ileal biopsies did not reveal intestinal inflammation, the horse was assigned to Group 2. Four horses in Group 2 underwent surgical exploration of the abdomen and 26 were treated medically. Overall, 43 (72%) horses with GI disease survived; 15 horses in Group 1 survived (50%; 10 of these were euthanized on admission or during surgery due to cost constraints or perception of poor prognosis) and 28 horses (93%) survived in Group 2. Of the 2 nonsurviving horses in Group 2, one was euthanized following the owners' request and the other was euthanized following confirmation of equine dysautonomia by ileal biopsy; both were removed from the outcome analysis. There were significantly more nonsurvivors in Group 1 compared with Group 2 (P<.001). Clinical coagulopathy was identified in 2 horses (jugular vein thrombosis).
|Group 1||Group 2||Group 3|
|Age (years)||13.8 ± 6.7||11.9 ± 4.7||12.4 ± 7.8|
|Heart rate (beats/min)||62 ± 22b,c||44 ± 10a,b||38 ± 6a,c|
|Respiratory rate (breaths/min)||20 (10–68)b,c||16 (12–52)a,c||12 (8–20)a,c|
|Temperature (°C)||37.5 ± 1.2||37.0 ± 0.5||37.6 ± 0.4|
|PCV (%)||42.3 ± 11.9c||38.7 ± 6.3c||31.7 ± 6.8a,b|
|Plasma protein (mg/dL)||6.6 ± 1.2||6.7 ± 0.9||6.9 ± 0.3|
|White blood cell count (/UL)||6900 ± 3100||6500 ± 2400||5800 ± 1700|
|Neutrophils (%)||68.2 ± 16.7||71.1 ± 11.6c||62.5 ± 11.2b|
|Lymphocytes (%)||28.4 ± 17.3||25.4 ± 10.7c||34.1 ± 10.5b|
|Blood lactate concentration (mmol/L)||2.9 (0.8–17.8)b||0.8 (0.6–4.8)a||NA|
TEG parameters on admission from horses with GI disease were available for all but 1 horse in Group 2 (technical error) while TEG parameters at 48 hours could be obtained from 14 horses in Group 1 (12 horses were euthanized within 24 hours of admission, 2 horses died shortly after admission, and from 2 horses a 2nd sample was not available) and 22 horses from Group 2 (8 horses were discharged the day after admission). Hematological and TEG parameters for all groups on admission are summarized in Table 2. Groups 1 and 2 had a significantly higher PCV compared with Group 3 (P= .001 and .046, respectively). Horses from Group 2 had a higher neutrophil (P= .049) and lower lymphocyte (P= .039) count than horses from Group 3. Compared with horses in Group 3, horses from Groups 1 and 2 had a significantly higher heart rates (P<.001 and P= .005, respectively), and higher respiratory rates (P<.001 and P= .016, respectively). Horses from Group 1 had higher heart rate (P<.001), higher respiratory rate (P= .008) and a higher blood lactate concentration (P<.001) on admission compared with horses from Group 2 (Table 1).
|Group 1 (n = 30)||Group 2 (n = 29)||Group 3 (n = 30)|
|R (min)||14.6 ± 8.3c||19.7 ± 20.5||22.8 ± 12a|
|K (min)||6.3 ± 6.3||7.6 ± 7.4||13.0 ± 19.5|
|Angle α (°)||34.0 ± 13.0||33.7 ± 14.7||30.6 ± 17.0|
|MA (mm)||51.6 ± 11.8||54.6 ± 8.5||57.3 ± 14.0|
|LY30 (%)||0.7 ± 0.7b||1.2 ± 1.0a||1.0 ± 0.8|
|LY60 (%)||3.3 ± 2.2||4.4 ± 2.5||3.4 ± 2.3|
|Platelet count (× 103/L)||129 ± 33.8||140 ± 45||147 ± 31|
|PT (sec)||12.6 ± 1.7||11.9 ± 2.0||11.6 ± 0.9 (n = 6)|
|aPTT (sec)||45.9 ± 8.1||44.8 ± 10.2||42.9 ± 4.2 (n = 6)|
|D-Dimer (ng/mL)||3,069 ± 3,375b,c||1,013 ± 2,482a||164 ± 68a (n = 6)|
|Antithrombin (%)||227 ± 53b||251 ± 38a||227 ± 40 (n = 6)|
|Group 1 (n = 14)||Group 2 (n = 22)|
|Admission||48 hours||Admission||48 hours|
|R (min)||13.4 ± 4.3||18.2 ± 7.1*||15.2 ± 5.7||19.9 ± 7.9*|
|K (min)||4.5 ± 1.7||6.4 ± 4.5||5.2 ± 2.9||7.4 ± 4.6*|
|Angle α (°)||35.6 ± 8.4||34.7 ± 16.4||38.5 ± 11.6||27.7 ± 14.5*|
|MA (mm)||55.9 ± 8.2||57.3 ± 7.4||56.2 ± 8.4||56.1 ± 10.8|
|LY30 (%)||0.6 ± 0.5||1.0 ± 1.2||1.1 ± 0.8||1.0 ± 0.8|
|LY60 (%)||3.4 ± 1.4||3.5 ± 3.01||4.6 ± 2.2||4.1 ± 2.3|
|Platelet count (× 103/L)||121 ± 39||122 ± 36||147 ± 48||132 ± 38|
|PT (sec)||13.1 ± 1.8||13.0 ± 2.4||11.7 ± 0.9||12.3 ± 1.6|
|aPTT (sec)||48.5 ± 7.3||46.7 ± 6.7||44.9 ± 13.0||46.4 ± 7.3|
|D-Dimer (ng/mL)||3,797 ± 3,091b||2,368 ± 3,036||1,243 ± 3,113a||3,030 ± 3,354|
|Antithrombin (%)||218 ± 45b||210 ± 61b||247 ± 44a||248 ± 39a|
Comparison of TEG Parameters
Group 1 had a significantly shorter Rad than Group 3 (P= .011) but not a shorter Rad than Group 2 (P= .08). Kad time in Group 1 compared with Group 3 was not different (P= .056). Further differences included a smaller LY30ad (P= .046) but not LY60ad (P= .071) in Group 1 compared with Group 2 (data summarized in Table 2A).
When only horses that had samples obtained at both time points were considered, R and K significantly increased in all horses from admission to 48 hours (Rad 14.5 ± 5.2 to R48 19.2 ± 7.5; P < .001 and Kad 4.9 ± 2.5 to K48 7.0 ± 4.5; P= .005) and the angle α significantly decreased (angle αad 37.3 ± 10.3 to angle α48 30.5 ± 15.5; P= .009). When only horses from Group 1 are considered, only R changed significantly (P= .013) while in Group 2 R, K, and angle α changed significantly (P< .001, P= .005 and P= .009, respectively; see Table 2B). TEG parameters obtained on admission and 48 hours after admission were significantly correlated (Rad and R48: r= 0.463 P= .005; Kad and K48: r= 0.425, P= .011; anglead and angle48: r= 0.427, P= .01; MAad and MA48: r= 0.631, P < .001; LY30ad and LY3048: r= 0.533, P= .001; LY60ad and LY6048: r= 0.516, P= .001). Correlation coefficients for most comparisons were reasonable, albeit not high.
Based on TEG profile, 24 horses were identified as HC: 14 in Group 1 (7 survivors and 7 nonsurvivors) and 10 in Group 2. Evidence of HC was documented in 21 of these horses on admission: 13 in Group 1 (7 nonsurvivors and 6 survivors) and 8 in Group 2. HC was noted after 48 hours in 10 horses: 3 in Group 1 (all survivors) and 7 in Group 2. HC was not associated with survival and not more common in either group.
Based on TEG profile, 2 horses were identified as hypocoagulable: 1 in Group 1 and 1 in Group 2 (both on admission) and 5 horses had tendency toward increased fibrinolysis (3 horses in Group 1: 1 nonsurvivor and 2 survivors and 2 horses in Group 2, all survivors).
HC, hypocoagulability, or increased fibrinolysis, as determined by TEG, on admission, 48 hours after admission, or at any time point were not associated with nonsurvival and were not more common in any group. An abnormal TEG profile (either hyper- or hypocogulation or increased fibrinolysis on admission, 48 hours after admission or at any time point (either admission or 48 hours or both), was also not associated with nonsurvival and was not more common in any group.
Abnormalities of any of the individual TEG parameter were not associated with nonsurvival and not more common in either group; an abnormal R value on admission had a tendency to occur more commonly in Group 1 than Group 2 (P= .082).
Comparison of Conventional Coagulation Parameters between Groups
Group 1 had a higher D-Dimer concentration (P= .014) on admission compared with Group 2 and lower AT activity on admission (P=.049) and 48 hours after admission (P=.038) (Tables 2A and B). None of the conventional coagulation parameters changed significantly over time in both groups combined or either group analyzed separately and none of the conventional parameters obtained at both times were significantly correlated.
In contrast to TEG profiles, an abnormal coagulation profile (abnormalities in any of the following parameters: platelet count, D-Dimer concentration, PT, aPTT, or AT activity) on admission was significantly associated with nonsurvival (odds ratio = 4.8; P= .021) and significantly more common in Group 1 compared with Group 2 (odds ratio = 4.2; P=.009). Abnormal coagulation profiles at 48 hour or at any time (either admission or 48 hours or both) were neither associated with nonsurvival nor more common in any group (of all nonsurvivors, only 1 horse was still alive at 48 hours to be sampled).
D-Dimer concentrations on admission were associated with nonsurvival (odds ratio for survival: 0.47; P= .007). Abnormalities of aPTT (odds ratio 4.7; P= .049) and D-Dimer concentrations (odds ratio for being in Group 2: 0.39; P= .001) on admission were more common in Group 1 than Group 2. Abnormal TEG profiles and abnormal coagulation profiles were not correlated at any time.
Subclinical DIC on admission was identified in 6 horses in Group 1 (3 survivors and 3 nonsurvivors) and in 1 horse in Group 2 with a diagnosis of small intestinal ileus. After 48 hours, 11 horses in Group 1 (1 nonsurvivor and 10 survivors) and 5 horses in Group 2 had laboratory evidence of subclinical DIC; 3 of these horses showed subclincal DIC on both occasions, the 1 nonsurvivor in Group 1 and 3 horses in Group 2. Subclinical DIC was more common in Group 1 at 48 hours (P= .024) and at any time point (P= .028).
Comparison of Survivors and Nonsurvivors
Comparison of all horses with GI disease demonstrated a higher heart rate (67±27 versus 49±14/min; P=.022), respiratory rate (27±16 versus 19±11/min; P=.033), and blood lactate concentration (6.7±5.7 versus 1.9 ± 2.0 mmol/L; P=.008) in nonsurvivors on admission. Furthermore, MAad (47±13 versus 55± 8 mm; P=.037), LY60ad (2.8±2.7 versus 4.3±2.2%; P=.035), and platelet countad (114±33 versus 143±40 × 103/μL; P= .004) were lower in nonsurvivors.
Comparing survivors and nonsurvivors of Group 1 only, lactate remained significantly higher (P= .048) and MAad (P=.027) and platelet countad (P=.001) significantly lower in nonsurvivors on admission (Table 3).
|Group 1 (n = 15)|
|R (min)||13.3 ± 4.3||15.9 ± 11|
|K (min)||4.3 ± 1.7||8.8 ± 8.4|
|Angle α (°)||36.1 ± 8.7||31.9 ± 16.3|
|MA (mm)||56.4 ± 7.9||46.9 ± 13.3*|
|LY30 (%)||0.8 ± 0.5||0.6 ± 0.9|
|LY60 (%)||3.9 ± 1.6||2.8 ± 2.7|
|Platelet count (× 103/L)||147 ± 31||111 ± 26*|
|PT (sec)||12.2 ± 1.4||13.0 ± 1.9|
|aPTT (sec)||46 ± 7.1||45.7 ± 9.3|
|D-Dimer (ng/mL)||2,769 ± 3,485||3,223 ± 3,065|
|Antithrombin (%)||226 ± 37||228 ± 67|
This study investigated the clinical utility of TEG in horses with GI disease. The results of the study are in agreement in many aspects with previous findings9,13,14,22–24 in that a tendency toward HC (decreased R compared with healthy horses) was identified in horses with ischemic or inflammatory lesions8 which was supported by the finding of significantly higher D-Dimer concentrations and lower AT activity in this group. However, most TEG values remained within the normal range. When HC was defined as >25% deviation above the upper limit of the normal range, HC was not more common in either group and was not associated with survival. In other species, HC profiles are usually characterized by changes in >1 parameter and an increase in MA is considered by many to be the most significant indicator of HC.25 Increases in MA were rarely observed in the horses investigated and TEG profiles did not resemble hypercoaguable profiles obtained in other species.26 It is therefore possible that true HC was not demonstrated in any of the horses studied. HC can be inferred indirectly by determination of AT, TAT, and protein C concentrations or by the detection of products of fibrin/fibrinogen degradation.27 However, none of these parameters is specific for HC and, at least in other species, abnormalities of these values can be identified during normo-, hypo-, or hypercoaguable states, as diagnosed by TEG.27–29 The theoretical advantage of TEG is the ability to monitor and document the interaction of plasma components and blood cells during coagulation with the hopes that the TEG tracing will closely mimic the in vivo phenotype of a coagulation disorder. The inability to demonstrate HC convincingly by use of TEG in the presented study may be related to the technique used rather than being indicative of an absence of HC since fibrin formation in the microvasculature has been clearly demonstrated in horses with GI disease7,11 and changes suggestive of HC have been identified in other equine studies.8,18 TEG can be performed on native or citrated blood with or without the addition of coagulation activators to enhance the speed of the analysis. The use of recombinant human tissue factor (rhTF) activated TEG has been advocated for equine blood as it has been shown to decrease standard deviation of most TEG parameters and inter-operator variability.17 However, the addition of rhTF also significantly reduced R and K and increased angle α, which could influence the ability to identify changes in these parameters toward HC, and the effect of rhTF on TEG profiles obtained from horses with coagulopathy has to the authors' knowledge not been investigated.17 As this study aimed to identify HC in horses with GI disease, TEG was performed without the addition of an activator due to concerns that this could decrease the sensitivity of TEG for detection of HC. It is possible that addition of rhTF would have decreased the overlap between healthy and abnormal horses and thereby could have improved rather than decreased the sensitivity and specificity of detection of HC. Further clinical studies comparing TEG with and without addition of rhTF are necessary to establish the optimum method for clinical use.
Another possibility is that events occurring during TEG analysis, despite the use of whole blood, do not reflect in vivo coagulation. This could be due to the use of citrated plasma and/or the absence of an endothelial surface, which could play a significant role in activation of coagulation in diseases with systemic inflammation. The presence of calcium citrate has been shown to impair natural coagulation processes in human blood and it has been proposed that the use of citrated blood does not replicate the natural dynamics of clot formation.30 Use of a lower citrate concentration (1.5%) in association with blood collection into glass vacutainers significantly changed TEG profiles toward a more HC appearance.31
Changes in blood hematocrit influence TEG parameters in humans. Decreases in hematocrit appear to change the TEG profile toward a more HC picture (shorter R and K and increase in angle α and MA) and vice versa.32 However, another study suggests that this could be an in vitro phenomenon related to the methodology used rather than an actual in vivo change in coagulation.33 To the authors' knowledge, the influence of hematocrit on coagulation parameters has not been reported in horses. If similar changes in TEG profiles are observed, the increasing PCV observed in many colic cases could mask HC and render TEG less suitable to assess HC in patients with increased hematocrit.
It is also conceivable that the attempt to stratify colic cases into those likely to suffer from an HC state (ischemic/inflammatory; Group 1) and those unlikely to be affected by coagulation disturbances failed to achieve this purpose. The fact that horses in Group 1 had higher heart and respiratory rates, higher D-Dimer concentrations, lower AT activities, and higher mortality when compared with horses from Group 2 suggests that these horses suffered from more severe disease. Nevertheless, it is uncertain whether the use of a different classification, for example, separating “ischemic and inflammatory cases” into separate groups, would have demonstrated more pronounced differences in TEG profiles. Although both ischemic and inflammatory lesions have been shown to be associated with coagulopathies,8,34 differences in time course, extend of disease and other, as of yet described factors, might exist.9
Although sample size calculations performed before the study indicated that 30 horses per group would be sufficient to reject the null hypothesis, variability was larger than anticipated and inclusion of a larger number of horses may have resulted in a clearer distinction between TEG parameters of the 2 groups.
The finding that MA was significantly lower in nonsurvivors may reflect the lower platelet counts observed in this group but could also indicate that these horses had already progressed toward a hypocoaguable state. The fact that many nonsurvivors were euthanized or died within the first 24 hours made meaningful comparison of survivors and nonsurvivors at 48 hours impossible and it was not feasible to establish whether changes over time could be used as prognostic indicators. Comparison of survivors and nonsurvivors was further confounded by the influence of finances and/or a perceived poor prognosis on the decision to euthanize horses before or during surgery. Although it is reasonable to assume that lesions in this groups of horses were more severe compared with horses in Group 2, it is unknown whether these horses would have survived with appropriate treatment.
Another limitation of the study was the significant overlap in the various parameters measured between healthy and affected horses. TEG profiles of healthy horses obtained with the technique described varied widely, and the technique is currently of limited clinical use as treatment decisions should probably not be based on TEG findings alone. Changes to rhTF or kaolin activated TEG may result in less variability and make the technique more applicable for use in equine patients.
In summary, TEG identified a tendency toward HC, as evidenced by short R times, in horses with ischemic or inflammatory GI disease compared with healthy horses. However, changes were subtle, and in most horses only 1 or 2 parameters, often R and K, were above or below the normal range and none of the horses demonstrated a TEG profile that resembled TEG profiles obtained from patients presumed to be HC in other species. Although conventional coagulation testing appeared to support the diagnosis of HC, the abnormalities were not necessarily found in the same horses.
aBD Biosciences, Oxford, UK
bScil animal care company GmbH, Viernheim, Germany
cHaemoscope TEG Haemostasis analyzer TEG 5000 and TEG analytical software; Haemoscope; London, UK
dMiniquant-1, Biopool, Trinity Biotech, Wicklow, Ireland
eMiniquant, Biopool, Trinity Biotech
fStago ST4, Stago Diagnostics, Asnières-Sur-Seine, France
gSTA Antithrombin III, Stago Diagnostics
hCobas-Bio, Roche, Basel, Switzerland
iSPSS 17.0; SPSS Inc, Chicago, IL
The study was supported by an American College of Veterinary Emergency and Critical Care research grant.
- 8Hypercoagulation and hypofibrinolysis in horses with colic and DIC. Equine Vet J Suppl 2000:19–25., , , et al.
- 16Evaluation of thromboelastography initiated with tissue factor on citrated whole blood from healthy horses. J Vet Emerg Crit Care 2006;16:S12., ,
- 20Equine neonatal sepsis: The pathophysiology of severe inflammation and infection. Compend Contin Educ Pract Bet 2001;23:661–670.,
- 31Influence of citrate concentration and material of blood tubes on thromboelastographic parameters in horses. J Vet Emerg Crit Care 2009;19:A14., ,