Objectives To evaluate citrated recalcified thromboelastography (TEG) in healthy newborn foals, and to determine intra-assay, inter-individual and intra-individual (at 12 h, 24 h and 7 days after birth) variations. Additionally, to compare TEG variables, haematological values and conventional coagulation profiles from healthy, sick non-septic, and septic foals.
Design Prospective study.
Methods The study group comprised 18 healthy, 15 sick non-septic and 17 septic foals. Two citrated (3.2%; 1 : 9 anticoagulant : blood ratio) blood samples were submitted for haemostatic evaluation using a TEG analyser and conventional coagulation profile. TEG values (R time (R), K time (K), angle (α), maximum amplitude (MA) and G value (G)), complete blood count (CBC) and conventional coagulation profile (prothrombin time (PT), activated partial thromboplastin time (aPTT), fibrinogen concentration (Fib) and antithrombin (AT)) were evaluated. Signalment, presenting complaint, sepsis scores, blood culture results and outcome were taken from the medical records of the sick foals.
Results Mean values ± SD for TEG variables in healthy neonatal foals were: R = 11.82 ± 5.35 min, K = 3.06 ± 1.34 min, α= 51.19 ± 12.66 degrees, MA = 55.06 ± 6.67 mm and G = 6361 ± 1700 dyn/cm2. Mean coefficients of variation for intra-assay/inter-individual/intra-individual in healthy foals were: R = 3.5/45.2/43.1%; K = 5.3/58.7/28.7%; α= 1.5/24.7/11.9%; MA = 0.3/12.1/6.1%; G = 1.6/26.7/14.7%. Septic foals had significantly greater α, MA and G values than sick non-septic foals, and significantly greater MA and G than healthy foals, changes that are consistent with hypercoagulability. Weak correlations were detected between TEG variables and haematological or haemostatic values.
Conclusions TEG could be used to provide additional information about the haemostatic system in equine neonates.
Sepsis is a major cause of morbidity and mortality in human and equine neonates.1–5 During septic shock, the relationships among procoagulant, anticoagulant and fibrinolytic factors that maintain the delicate balance of the haemostatic system are disrupted.6 Haemostatic abnormalities have been reported in neonatal foals with severe sepsis and these findings are frequently associated with poor outcome.7–9 One recent report evaluating six coagulation parameters showed that foals in septic shock were 12.7-fold more likely to have clinical evidence of bleeding than sick non-septic foals.9 Foals with presumed septicaemia typically have prolonged prothrombin time (PT), activated partial thromboplastin time (aPTT) and increased concentration of fibrinogen degradation products (FDPs) with decreased antithrombin (AT) and protein C concentrations.7 Furthermore, neonatal foals with severe septicaemia were recently shown to have fibrin deposits in the vasculature of certain tissues, a finding that is supportive of disseminated intravascular coagulation (DIC) and may lead to organ dysfunction and failure.10 Other coagulation tests, such as plasminogen activator inhibitor (PAI) activity, plasminogen, alpha-2 antiplasmin and plasma D-dimer concentration, have also been evaluated in foals to further characterise haemostasis and to help in the diagnosis of coagulopathies.8,11
Novel coagulation tests are currently being applied in equine medicine to further characterise haemostasis and to help in the diagnosis of coagulopathies.12–14 Dallap Schaer et al. were the first authors to report the evaluation of haemostasis in neonatal foals using a viscoelastic coagulation and platelet function analyser (Sonoclot).15 The thromboelastograph is another viscoelastic whole blood analyser that measures static ex vivo coagulation in whole blood, including primary and secondary haemostasis, and the fibrinolytic system. This contrasts with the traditional haemostatic tests, such as platelet count (PLT), PT and aPTT, which only evaluate one portion of the haemostatic system. Reference intervals have been published for thromboelastography (TEG) in healthy adult horses.12,13,16 However, to our knowledge, citrated recalcified TEG has not been reported in the peer-reviewed literature for healthy or clinically sick equine neonates. In humans, TEG has clinical utility in different pathologies and has been used to detect sepsis in human neonates, with 96% sensitivity and specificity.17–20 In healthy newborn foals, it is well known that haemostatic indices vary during the first month of life,11 so TEG variables in neonates need to be evaluated separately from those in young or adult horses.
The first objective of this study was to evaluate TEG variables in healthy newborn foals and test intra-assay, inter-individual and intra-individual (at 12 h, 24 h and 7 days after birth) variations. The second objective was to describe TEG findings in sick non-septic and in septic neonatal foals, and compare them with those in healthy neonates. The third objective was to determine correlations between TEG variables (R-time (R), K-time (K), angle (α), maximum amplitude (MA), and G value (G)) and conventional haemostatic values (PT, aPTT, fibrinogen concentration (Fib) and PLT).
Materials and methods
This study was approved by the University Teaching Hospital Executive Committee and observed the humane treatment of animals in veterinary clinical investigations. Owners were informed and owner consent was obtained before inclusion in the study.
The control group included 18 healthy foals ≤8 days old, of either sex and any breed. Health status was based on physical examination, results of a complete blood count (CBC) within normal limits, serum immunoglobulin G (IgG) concentration >800 mg/dL, sepsis score ≤821 and a conventional coagulation profile (aPTT, PT, Fib) within reference ranges. IgG concentrations were determined by SNAP Foal IgG Test Kit (IDEXX Laboratories, Westbrook, ME, USA).
Sick non-septic and septic foals
Clinically ill foals admitted to hospital between January 2008 and May 2009 were also included in this study. Sick foals were classified into two groups: sick non-septic and septic. The sick non-septic group included foals that presented for conditions other than septicaemia (including orthopaedic problems, hypoxic ischaemic encephalopathy or meconium impaction). These foals had negative bacterial blood cultures and/or sepsis scores ≤8.21 The septic group included foals that had a positive blood culture and/or a sepsis score ≥1221 and a documented septic focus. Foals that had received fluids, antimicrobials, anticoagulants or plasma products prior to admission were excluded from the study.
Complete history, signalment, presenting complaint and physical examination findings, including the presence of a septic focus, such as pneumonia, enteritis, uveitis, omphalitis or arthritis, were obtained at presentation for classification purposes. Clinicopathological information recorded included: sepsis score, packed cell volume (PCV), plasma total proteins (TP), PLT, white blood cell (WBC) count, IgG (Hitachi 911 System, Boehringer Mannheim, Indianapolis, IN, USA) and results of bacterial blood cultures (aerobic and anaerobic). Sepsis scores were calculated as described by Brewer and Koterba.21 We routinely perform blood cultures for all sick neonatal foals that present to the hospital.
In the healthy group, blood was collected by direct venipuncture technique using a Vacutainer needle (Vacutainer, Becton Dickinson, Franklin Lakes, NJ, USA). An EDTA tube (2 mL, EDTA Monoject™ Lavender Stopper, Tyco Healthcare Group LP, Mansfield, MA, USA) was attached to collect blood for the CBC, and two Vacutainer glass tubes containing 3.2% buffered sodium citrate (4.5 mL, 0.105 M; Vacutainer, BD, Franklin Lakes, NJ) were attached, achieving a final proportion of anticoagulant to blood of 1 : 9, and then mixed gently. Whole blood from hospitalised foals was collected on admission prior to administration of any fluids or medications. Samples (in the non-septic and septic groups) were drawn from an aseptically placed jugular catheter, using a plastic syringe and discarding the first 10 mL. Next, blood was immediately placed into one EDTA (2 mL, EDTA Monoject Lavender Stopper) and two citrated tubes (4.5 mL, 3.2% 0.105 mol/L buffered sodium citrate Vacutainer). A further 20 mL of blood was drawn from the catheter for blood culture (BBL Septi-Chek blood culture system, Becton Dickinson) in all sick foals. CBC was performed within 60 min of sampling, using an automated haematology analyser (Cell-Dyn 3500R Automated Haematology Analyzer, Abbott Laboratories, Abbott Park, IL, USA) and a manual differential WBC count was also performed. One citrated tube was left to stand on a rack at room temperature for 30 min before TEG analysis. The other citrated tube was centrifuged immediately after sampling at 1380g for 12 min at room temperature, and plasma was harvested and frozen at −80°C. A haemostatic panel (including PT, aPTT, Fib and AT) was performed for all samples using a coagulation analyser (ACL® 7000 Coagulation Analyzer, Beckman Coulter, IL, USA).
TEG analysis. A viscoelastic coagulation analyser (TEG® 5000 Thrombelastograph Haemostasis Analyzer; software version 4.2.2 Haemoscope Corporation, Niles, IL, USA) with two channels was used for all TEG analysis. TEG maintenance tests were performed daily before sample analysis (TEG® 5000 User's Manual Haemoscope 1995–2005). A single test (citrated recalcified TEG) was run for all foals, except for the 10 healthy foals, which were analysed in duplicate using both TEG channels simultaneously in order to evaluate the thromboelastograph's precision. In addition, five of the healthy foals had repeat TEG analyses at 12 h, 48 h and 7 days of age to evaluate intra-individual variation over time.
We added 20 µL of 0.2 mol/L CaCl2 (Haemoscope Corporation) to a new preheated (37°C) reaction cup (TEG reaction cups and pins; Haemoscope Corporation), which had previously been placed in one of the TEG channels. The rested citrated blood tube was mixed gently and then 340 µL of blood were added to each reaction cup. Special care was taken when the pipette was introduced in the cup, releasing the blood against the cup wall and avoiding air bubbles. All samples were handled and run by the same experienced operator (J.L.M). Tracings were obtained after 120 min of running time at 37°C. The following TEG parameters were measured and recorded:
• R time (R): time (min) from the addition of agonist (CaCl2) until the clot started forming, which evaluates the activity of plasma coagulation factors
• K time (K): time (min) necessary for an amplitude of 20 mm to be reached, which represents the clot kinetics
• angle (α) (degrees): the slope of the tangent on the elasticity curve, which correlates to the fibrinogen concentration, representing the rapidity of fibrin build-up and cross-linking
• maximum amplitude (MA) (mm): corresponds to the ultimate strength of the clot and depends on the contribution of platelet aggregation and fibrinogen activation
• G value (G) (dyn/cm2): represents the viscoelastic shear of the clot and it is calculated using the MA.
All data were analysed using a statistical software package (Prism version 4.0, GraphPad Inc., San Diego, CA, USA). Descriptive statistics were performed for all data analysed and normality was assessed using the D'Agostino & Pearson omnibus test. Intra-assay (between same sample), inter-individual (between animals) and intra-individual (at 12 h, 24 h, and 7 days after birth) variations were analysed in the healthy group using coefficients of variation (CV). Inter-individual variation was calculated from the first sample taken for each healthy foal. Intra-individual variation was evaluated for the five healthy foals to monitor the changes in coagulation over the first week of life. When data passed the normality test, the three groups were compared through one-way analysis of variance, followed by the Bonferroni post-test when differences were detected. If the data did not follow a Gaussian distribution, the groups were compared using the Kruskal-Wallis test, followed by Dunnet's post-test when required. Correlations between TEG variables and the haematological values and the conventional coagulation profiles from all samples were analysed using Pearson's test (when data followed Gaussian distribution) or non-parametric Spearman test (when not normally distributed), and linear regression was performed for the variables when a statistically significant correlation was found. Statistical significance was set at P < 0.05.
A total of 50 newborn foals were evaluated in this study: 18 were categorised as healthy (36%), 15 as sick non-septic (30%) and 17 as septic (34%). There were 28 (66%) colts and 22 (44%) fillies. The mean ± SD age was 2.2 ± 1.6 days and the mean weight was 48.9 ± 8.4 kg. When age and weight were compared between groups, statistical differences in age were found between healthy (mean 1.72 ± 1.22 days) and septic foals (mean 3.16 ± 1.88 days).
The breed distribution included 23 Quarter Horses, 9 Standardbreds, 6 Thoroughbreds, 4 mixed breeds, 3 Warmbloods, 1 Saddlebred, 1 Percheron, 1 Arabian, 1 Tennessee Walker and 1 Haflinger. Primary problems in the sick non-septic group included: colic (2), ruptured bladder (2), dysphagia (2), flexural limb deformity (2), hypoxic ischaemic encephalopathy (2), patent urachus (2), meconium impaction (1), penile malformation (1) and neonatal isoerythrolysis (1). Of the 17 septic foals, 14 had a positive blood culture and the following microorganisms were isolated: Escherichia coli (7), Actinobacillus spp. (4), Enterococcus spp. (2), Staphylococcus spp. (2), Clostridium perfringens (1) and Proteus spp. (1). The presence of at least one of the following foci of infection was recorded in 15 septic foals: septic arthritis (7), pneumonia (4), enteritis (5), omphalitis (2), cellulitis (2) and septic physitis (1). Three foals in the septic group and two foals in the sick non-septic group were euthanased because of a poor prognosis. Sepsis scores for the healthy group ranged from 0 to 3 points (median 1 point), the sick non-septic group ranged from 0 to 8 points (median 6 points) and the septic group ranged from 12 to 21 (median 14 points). No signs of bleeding or petechiation were seen in any foal at the time of sampling.
Results of TEG variables are shown in Table 1. Intra-assay, inter-individual and intra-individual (at 12 h, 24 h, and 7 days after birth) variations in clinically healthy foals expressed as CV (%) are shown in Table 2. Statistical differences (α, MA and G; all P < 0.05) were observed in the healthy foals that were tested over time (Figure 1).
Table 1. Thromboelastographic values (mean ± SD) in healthy, sick non-septic and septic newborn foals
aSignificantly different from septic group; bsignificantly different from non-septic group; csignificantly different from healthy group.
TEG values from septic foals were significantly different from those in sick non-septic foals and had greater α (P < 0.05), MA (P < 0.001) and G (P < 0.001) values, changes that were consistent with hypercoagulability. Septic foals had greater MA (P < 0.001) and G (P < 0.001) values when compared with healthy foals. No statistical differences were observed between the healthy and sick non-septic groups. Haematological values (PCV, TP, WBC and PLT), IgG results and coagulation profiles (PT, aPTT, Fib and AT) in all foals (mean ± SD) are presented in Table 3. The septic group had longer aPTT than the healthy group (P < 0.01), and the septic and sick non-septic groups had higher Fib than those in the healthy group (P < 0.001 and P < 0.05, respectively). The sick non-septic group had higher PCV than the septic group (P < 0.01). The remaining variables were not significantly different among the three groups.
Table 3. Haematological values and haemostatic profiles (mean ± SD) of healthy, sick non-septic and septic newborn foals
aSignificantly different from septic group; bsignificantly different from non-septic group; csignificantly different from healthy group.
Regression analysis showed a weak positive relationship between Fib and MA (P < 0.001, r2= 0.28) (Figure 2), and between Fib and G (P < 0.001, r2 =0.39) (Figure 3). PLT was also found to have a weak positive relationship with MA (P = 0.02, r2= 0.11), and with G (P = 0.04, r2= 0.09). A weak negative relationship was found between PCV and MA (P < 0.001, r2= 0.29), as well as with G (P < 0.001, r2= 0.39). No other significant correlations were found between the haematological values, the haemostatic profiles and the TEG variables.
TEG values obtained from healthy neonatal foals in this study are slightly different from those reported for adult horses.12 The thromboelastograph was very precise, as demonstrated by the low intra-assay CV, which was lower than that reported for adult horses and dogs.13,22 MA was the TEG variable with the lowest CV in our study, showing that MA could be the most consistent and precise TEG variable for evaluating coagulation in equine neonates. High intra-individual variation was detected in the healthy foals that were tested over time (12 h, 24 h and 7 days after birth), suggesting coagulation changes in neonatal foals during the first days of life, as previously reported.11 An explanation for this finding could be the combination of decreased liver synthesis of coagulation factors, normal synthesis with reduced functional activity or more efficient clearance of factors at birth, as suggested in fetal lambs and human infants.23 Although the number of animals tested over time in our study was low and more studies are needed, physiological haemostatic changes in neonatal foals should be considered when monitoring treatments using TEG. Ideally, coagulation laboratories should establish reference intervals for very specific age groups of equine neonates. In addition, the high inter-individual variability seen in this study could be attributed to the haemostatic changes that occur in foals during the first month of life,11 although breed variability also could have influenced the results. The inter-individual variation could not be explained by sex, as no differences between the sexes were observed. Moreover, the operator influence should be considered minimal, as the same operator performed all tests. The use of recombinant human tissue factor as an activator of clot formation has been recommended to minimise TEG variability in adult horses12 and although it may mask true hypocoagulability, its use should be considered in future studies involving neonatal foals.
Statistically significant differences between the septic and sick non-septic foals were found for some of the TEG variables. The greater α, MA and G values observed in the septic group could suggest that blood samples from this group were able to generate stronger clots (or a trend toward hypercoagulability), because MA represents the strength of the clot and α represents the rapidity of fibrin build-up and cross-linking. These findings are consistent with a recent study of D-dimer concentration (a fibrin-linked degradation product from fibrinolysis) in septic foals, which revealed a marked activation of coagulation in foals with septicaemia.8 Furthermore, septic foals in our study had the highest Fib concentration, which was positively correlated with MA and G in the present study, and with MA in a previous study.12 In addition, some TEG values (MA and G) from septic foals were statistically different from healthy foals, also suggesting hypercoagulability in the septic group. These findings should be interpreted carefully because of the different blood collection techniques used (venipuncture in the healthy group vs catheter in the septic group). Age could explain the differences between these healthy and septic foals, because the septic foals were older and changes towards a wider α and greater MA and G were detected in the small group of healthy foals that was evaluated over time. Despite the different techniques used in the healthy and sick non-septic groups to draw blood samples, no statistical differences were found between them. In general, the correlations detected between TEG variables and the haemostatic and haematological values were modest and less than we had hypothesised.
The hypercoagulability shown on TEG testing in septic foals contrasts with the hypocoagulability demonstrated by traditional coagulation testing in this group (prolonged aPTT). Similar findings (prolonged clotting times) and increased concentration of fibrin degradation products in septic equine neonates have been reported by several investigators.7–9 Because DIC is typically characterised by an early hypercoagulable state leading to depletion of coagulation factors, it is possible that the different methods of testing are demonstrating the hypercoagulable and hypocoagulable processes that are occurring at the same time. Hypercoagulation has also been suggested by D-dimer concentration, which is significantly higher in septic foals 24 h after hospitalisation than at admission.8 The results of our study suggest that TEG may be detecting hypercoagulation earlier in the disease progression (on admission), which would enable faster intervention by clinicians. Unfortunately, D-dimer concentrations were not available for comparison for the foals in our study. Repeated TEG analyses at 24 and 48 h from presentation were not performed in this study; however, serial TEG may be useful in cases of suspected DIC in order to monitor the progression of the condition and therapy efficacy. In this study, no comparison was attempted between survivors and non-survivors because of the low number of animals that died or were euthanased. One study using a viscoelastic analyser (Sonoclot) in septic foals showed that animals with decreased clot rate were more likely to be euthanased or die.15
In conclusion, TEG could be used to provide additional information about the haemostatic system in equine neonates. The authors recommend that reference intervals for healthy neonatal foals be obtained independently from adult horses and that the technique be standardised to reduce its variability. In addition, age-specific reference ranges may need to be developed for neonates because there is significant variation in several TEG variables during the first week of life. Further studies including a larger number of clinical cases are needed to assess if TEG has any value for diagnosing sepsis in foals.