Vous êtes sur la page 1sur 5

Retrospective Study

Use of thromboelastography in dogs with immune-

mediated hemolytic anemia: 39 cases (2000^2008)
Virginia B. Sinnott, DVM and Cynthia M. Otto, DVM, PhD, DACVECC
Objective To analyze thromboelastograms (TEGs) of naturally occurring cases of immune-mediated
hemolytic anemia (IMHA) in order to identify whether a hypercoagulable state was present and whether its
presence was associated with differences in survival.
Design Retrospective study spanning January 2000 to June 2008. Medical records of dogs were evaluated.
Endpoints were considered death or discharge from the hospital.
Setting Academic teaching hospital.
Animals Thirty-nine dogs with a diagnosis of IMHA and at least one TEG performed during hospitalization
were included.
Interventions None.
Measurements and Main Results Four values were evaluated from the TEG: the R time (R), K time (K),
alpha angle (a), and maximum amplitude. From these values, a coagulation index (CI) was calculated to
classify patients as normocoagulable, hypercoagulable, or hypocoagulable. Thirty-three of 39 patients were
hypercoagulable based on the CI. The 6 remaining dogs were normocoagulable. The patients with a
normocoagulable CI had an increased mortality rate (100%) when compared with the hypercoagulable
patients using Fishers exact test (P50.02). Additionally, prolongation of partial thromboplastin time did not
preclude hypercoagulable TEG values.
Conclusions The majority of dogs with IMHAwere hypercoagulable as measured by TEG. A normal CI was
associated with a worse outcome in this patient population. TEG may provide additional and complementary
information to prothrombin time and partial thromboplastin time relating to coagulation status in dogs with
IMHA and may help predict prognosis and potentially guide clinical decisions to utilize anticoagulant drugs.
(J Vet Emerg Crit Care 2009; 19(5): 484488) doi: 10.1111/j.1476-4431.2009.00455.x
Keywords: coagulation, hemostasis, thromboembolism
Immune-mediated hemolytic anemia (IMHA) is a dis-
ease associated with mortality rates of 2770% despite
the availability of immunosuppressive drugs to treat
the primary condition.
Thrombosis is a complication
that has been documented at necropsy in 6080% of
dogs with IMHA.
Thrombosis complicates the
course of treatment and may inuence outcome in
Hypercoagulability may result from different
mechanisms in individual dogs (ie, increased procoag-
ulant activity, decreased anticoagulant activity, or in-
creased platelet activation); these differences would
necessitate patient-specic anticoagulant therapy. Such
heterogeneity may explain the variable effectiveness of
antithrombotic therapies reported in the current liter-
In order to more effectively target preventa-
tive therapy, rigorous clinicopathologic interrogation of
the hemostatic system in individual IMHA patients is
required. Unfortunately, currently available commercial
coagulation testing cannot readily identify a hyperco-
agulable state making differentiation of populations
difcult. Recent studies using thromboelastography
(TEG) have identied and quantied hypercoagulable
states in dogs for several disease processes including
parvoviral enteritis, neoplasia, and disseminated intra-
vascular coagulation.
TEG is a point-of-care coagu-
Presented in abstract form at the International Veterinary Emergency and
Critical Care Symposium, Phoenix, AZ, September 2008.
Present address: Virginia Sinnott, Cape Cod Veterinary Specialists, 11
Bourne Bridge Approach, Buzzards Bay, MA 02532, USA.
Address correspondence and reprint requests to
Dr. Virginia B. Sinnott, Department of Emergency and Critical Care, The
University of Pennsylvania School of Veterinary Medicine, Matthew J. Ryan
Veterinary Hospital, 3900 Delancey St, Philadelphia, PA 19104-6010, USA.
Email: virginiasinnott@hotmail.com
From the Department of Emergency and Critical Care, The University of
Pennsylvania School of Veterinary Medicine, Matthew J. Ryan Veterinary
Hospital, Philadelphia, PA
The authors declare no conicts of interest.
Journal of Veterinary Emergency and Critical Care 19(5) 2009, pp 484488
& Veterinary Emergency and Critical Care Society 2009 484
lation device that assesses the viscoelastic properties of
a thrombus developing in a reaction cup. As such, it is
considered a global assessment of coagulation as the
overall interaction between plasma-based clotting fac-
tors leading to the formation of brin and their inter-
action with platelets is evaluated. Specically, a pin is
placed in a reaction cuvette (cup) containing blood. The
cup begins to oscillate and at rst it transmits no move-
ment to the pin. Over time, the forming thrombus cre-
ates an increasingly rigid connection between the cup
and the pin causing the pin to move. The motion of the
pin is recorded electronically and becomes the TEG
tracing that is evaluated clinically.
Furthermore, the
assay may be modied by use of activators to speed
clotting or by the use of heparinase-containing cups to
evaluate the native coagulation status of the patient
in the face of heparin therapy. The purpose of this
study was to analyze the TEGs of dogs with naturally
occurring IMHA in order to identify whether a hyper-
coagulable state was present, and whether hyper-
coagulability was associated with survival. Our
hypothesis was that dogs with IMHA would be hyper-
coagulable and that a hypercoagulable TEG would be
associated with a poor outcome.
Materials and Methods
Medical records between January 2000 and July 2008
were reviewed to identify all dogs with a diagnosis of
IMHA and that had TEG performed at least once dur-
ing the hospital stay. A diagnosis of IMHAwas made if
the dog had anemia (PCVo34%) and at least one of the
following: persistent autoagglutination, positive Coo-
mbs test, or cytologic evidence of erythrocyte or
erythrocyte precursor targeting in a cytologic prepara-
tion of a bone marrow aspirate.
Cases were excluded
if the TEG tracing was deformed upon visual inspec-
tion, or if the patient was on heparin but a heparinase-
containing cup was not used to perform the TEG assay.
All patients in this study had received at least 1 im-
munosuppressive dose of corticosteroids. Data collected
included signalment, concurrent (to TEG) prothrombin
time (PT), partial thromboplastin time (PTT), platelet
count, lactate at presentation, medications administered
at the time of TEG, and survival to discharge. When
more than 1 TEG analysis was performed on a dog
during hospitalization, the most hypercoagulable tracing
was used to calculate survival statistics. TEG was run at
the discretion of the primary clinician and because of the
retrospective nature of this study, the reason for TEG
interrogation in each case is unknown.
TEGs were produced by a computerized TEG
recalcied blood after a 30-minute incubation in a glass
vial containing 1 part 3.2% buffered sodium citrate to 9
parts whole blood.
No additional activators were used
to accelerate clotting. If the blood sample was obtained
via a sampling catheter or if the patient was undergoing
heparin therapy of any kind, paired TEG was per-
formed in normal cups and heparinase-containing
TEG cups.
Four TEG parameters were generated from the trac-
ing. The rst value, R, is the reaction time, expressed in
minutes and represents the time from activation of the
blood sample with calcium until the arms of the TEG
tracing are 2 mm apart. The second value obtained is
the growth angle, which represents the angle created by
the tangent to the TEG tracing at a width of 2 mm. The
third value is the clotting time (K) expressed in minutes
and represents the time it takes the 2 arms of the tracing
to diverge by 20 mm. The nal value, maximum am-
plitude (MA) represents the maximum width between
the 2 arms of the TEG tracing, expressed in millimeters.
The coagulation index (CI) was calculated using a pre-
viously published dog-specic equation with normal
being a CI between 4 and 4.
The CI is an equation
that relates all 4 TEG values to produce an overall
number representing coagulation status. Reference in-
tervals used were established on normal dogs in our
Data were inspected visually. Non-normally distrib-
uted data were reported as the median value and range.
Statistical analyses to compare survival with the pres-
ence of hypercoagulable TEG values were performed
using either the w
-test or Fisher exact test when ap-
propriate to compare differences between groups. The
Fisher exact test was used to compare traditional coag-
ulation values (PT, PTT, platelet count), lactate, and
TEG values, as well as to compare differences in
survival status and traditional coagulation values. A
P-value 0.05 was considered signicant.
Thirty-nine dogs were included in this study. The me-
dian age was 8 years (range 0.514 y). Breeds repre-
sented included: mixed breed (n 514), Cocker Spaniel
(n 53), Doberman Pinscher (n 52), Labrador Retriever
(n 52), Lhasa Apso (n 52), and 1 each of 16 other
breeds. Forty-six percent of the dogs were females (16
spayed, 2 intact) and 54% were males (18 castrated, 3
intact). At the time of blood sampling for TEG analysis,
6 dogs (15%) were receiving either low-molecular-
weight heparin or unfractionated heparin using dosing
to achieve either target anti-Xa activity or prolongation
of the PTT. An additional 6 dogs (15%) were receiving
aspirin (0.5 mg/kg, PO, q 24 h). Thirty-three of the 39
patients TEG tracings were hypercoagulable based on
the CI. Six dogs had a normal CI and none of the dogs
& Veterinary Emergency and Critical Care Society 2009, doi: 10.1111/j.1476-4431.2009.00455.x 485
Thromboelastography in canine IMHA
were hypocoagulable. A summary of the TEG values
obtained and their classication as normo-, hyper-, or
hypocoagulable with respect to the reference intervals
is shown in Table 1. Platelet counts performed at initial
presentation were available for 27 dogs. The median
platelet count was 167 10
/L (range, 10.6675 10
reference interval, 177398 10
/L). One patient had a
platelet count low enough to be consistent with
Evans syndrome. Three additional patients had CBCs
submitted, however, the platelet count was reported as
An association was found between a normal CI and
decreased survival. Specically, no dog with a normal
CI (n 56) survived while 48% of all patients and 56%
of hypercoagulable patients survived to discharge.
This difference was statistically signicant between
normocoagulable and hypercoagulable dogs (P50.02)
as well as all patients (P50.03). No statistically signi-
cant differences were found between the survival status
and any individual TEG parameter.
In the subset of dogs in which concurrent TEG and
traditional coagulation data were available (n 515 for
PT, n514 for PTT, n 512 for platelet count), an asso-
ciation was found between a normal PTT and a hyper-
coagulable MA, K, and a values (P50.04) while there
was an association between a prolonged PT and a
normocoagulable or hypocoagulable R value (P50.03).
In other words, all patients with a normal PTT had
hypercoagulable MA, K, and a values while no patients
with a prolonged PT had a hypercoagulable R value.
Conversely, patients with a prolonged PTT or a normal
PT had variable TEG values (hypo-, normo-, or hyper-
coagulable). No statistically signicant differences were
found between the concurrent platelet count (catego-
rized as low, normal, or elevated) or lactate value (cat-
egorized as normal or elevated, n 529) and coagulation
status as measured by TEG.
The majority of deaths in this study population were
due to euthanasia (18 euthanized/2 died) and all 6 of
the normocoagulable dogs were euthanized. None of
these 6 dogs were necropsied, therefore the reason for
euthanasia was obtained from the medical record and
the referral letter. All 6 dogs had either hypoxemia
documented in the medical record or respiratory dis-
tress listed as a reason for euthanasia by the attending
clinician. Additionally, none of the dogs were critically
anemic at the time of death with the lowest hematocrit
measured to be 18%. The reasons listed for euthanasia
were much more varied in the hypercoagulable group;
however, suspicion of thromboembolic disease was also
common. Of the 12 dogs that were euthanized, 10 were
euthanized due to progressive disease. Four of those
dogs were dyspneic at the time of euthanasia. Two eu-
thanized patients had necropsy-documented throm-
botic disease.
In 1856, Virchow rst theorized that the triad of
venous stasis, hypercoagulability, and endothelial
damage would interact synergistically to result in
clinically relevant thrombosis.
Additionally, throm-
bosis has been well documented in IMHA patients
since 1989.
It is not known which of the 3 components
of Virchows triad contributes most (or at all) to
thrombosis in these animals. The majority of the dogs
in this study had TEG values that were hypercoagul-
able when compared with our laboratorys established
reference intervals. In fact, 76% of all TEG variables in
this study were hypercoagulable. This nding sug-
gests that a hypercoagulable state is at least in part
responsible for the thrombosis reported in dogs with
An unexpected nding was that no dog with a nor-
mal CI survived. A normal CI in this patient population
may represent consumption of coagulation factors and
platelets due to either macroscopic thrombosis (ie, pul-
monary thromboembolism) or disseminated intravas-
cular coagulation. Thus, patients that would normally
be expected to be hypercoaguable, may have a normal
CI (or relative hypocoagulability) due to consumption.
This nding is in concert with a recent veterinary study
Table1: Results of thromboelastography (TEG) in 39 dogs with immune-mediated hemolytic anemia
TEG parameter Median Range Reference interval
Number of patients in whom value was:
Hypercoagulable Normocoagulable Hypocoagulable
R (min) 4.9 1.515 57 20 11 8
K (min) 1.4 0.86 3 30 5 3
A (s1) 71.4 3080.7 4 32 4 3
MA (mm) 74.6 36.192.2 4456 31 4 4
CI 6.2 0.29.8 4 to 4 33 6 0
R, the R time; K, the K time; A, a or the alpha angle; MA, maximum abundance; CI, coagulation index, calculated from the following equation:
CI 50.1227 R10.0092 K10.1655 (MA) 0.0241(a) 5.0220.
& Veterinary Emergency and Critical Care Society 2009, doi: 10.1111/j.1476-4431.2009.00455.x 486
V.B. Sinnott & C.M. Otto
of disseminated intravascular coagulation in dogs
where the most common coagulation abnormality was
hypercoagulability (44% of patients) based on the TEG
G value (a mathematical transformation of the MA
value), however, 34% of the patients were normocoag-
ulable and 22% were hypocoagulable.
the hypocoagulable patients had a statistically signi-
cantly increased mortality when compared with the
hypercoagulable patients.
It is not known whether a
normocoagulable TEG in IMHA represents levels of
consumption that would be associated with a hypocoag-
ulable TEG in another disease state. Future studies
evaluating clinical evidence of thrombosis, necropsy ev-
idence of thrombosis, D-dimers, and other laboratory
markers of thrombosis are warranted.
TEG represents global coagulation and may provide
information not available in the routine plasma-based
endpoint tests (PT, PTT). When evaluating TEG in con-
cert with traditional coagulation data (PT, PTT) there
appears to be some variability as to whether a patient
with a prolonged PTTwill have a hypercoagulable MA,
K, and a. As prolongation of the PTT is a common
nding present in 4547% of dogs with IMHA, TEG
may provide useful information about the remaining
coagulant capacity in this patient population and may
help guide clinical decisions about whether an antico-
agulant is still required for these patients.
dogs with a normal PT had R values in all 3 categories
of normo-, hyper-, or hypocoagulability. As such, TEG
may provide information about early depletion of co-
agulation factors. Recent studies in human medicine
suggest that TEG may correlate more closely with the
incidence of presumed thrombotic myocardial infarc-
tion and deep vein thrombosis after surgery than tra-
ditional coagulation tests.
The promise that TEG
may predict thrombotic events in certain human dis-
ease states warrants its further investigation in veter-
inary diseases with high thrombotic potential such as
IMHA. Ultimately, such research may help target
thromboprophylaxis to the patients who would most
benet from this therapy. Further study of the relation-
ship between TEG, PT, PTT, and clinical evidence of
thrombosis in IMHA is clearly indicated.
TEG in this study was performed at the discretion of
the primary clinician and all patients had commenced
treatment for IMHA with corticosteroids, or other im-
munosuppressive drugs, or both. Corticosteroid ther-
apy alone has been associated with a hypercoagulable
state and may have contributed to hypercoagulability
in this study.
Additionally, early consumptive co-
agulopathy may be missed by TEG as the hypercoagul-
ability induced by steroid use may have masked early
consumption of coagulation factors. Additionally, ane-
mia may affect TEG values and because patients in this
study had variable hematocrits at the time TEG was
performed, the contribution of anemia alone to changes
in TEG parameters could not be controlled in this ret-
rospective study.
A further limitation of this retro-
spective study is that the timing of TEG analyses was
not standardized. As a result, an individual TEG result
may represent data from before or after a thrombotic
event. Future studies evaluating serial TEG during hos-
pitalization could identify trends in coagulability that
herald thrombosis and elucidate the contribution of
corticosteroids to hypercoagulability in IMHA. Addi-
tionally, dogs on low-dose aspirin were not excluded
and a possibility exists that this therapy may have
affected their TEG tracings. However, a previous study
showed that TEG in dogs was not signicantly affected
by doses of aspirin 20 higher (5 mg/kg, PO, q 12 h)
than the dose (0.5 mg/kg, PO, q 24 h) routinely used at
our facility for thromboprophylaxis.
Additionally, the
relatively small patient numbers, especially with re-
spect to the evaluation of PT, PTT, lactate, and platelet
count may have allowed for type 2 error in that no
difference was seen when in fact one existed. Finally,
correlation of CI and other established markers of
thrombosis such as D-dimers was not possible as too
few patients had D-dimers analyzed at the same time as
their TEG.
In this retrospective study, a normal CI (between 4
and 4) identied a patient population with a 100%
mortality rate, however, because of the small number of
dogs in this group, it is possible that a dog with IMHA
and a normocoagulable TEG tracing may survive to
discharge. Additionally, it cannot be known whether
the TEG changes play any causal role in the death of
these patients or whether they simply act as a marker
for more severe cases of IMHA. TEG may provide ad-
ditional, complementary information to the PT and
PTT. The ability of TEG to identify a hypercoagulable
state in IMHA, even in the face of normal PT or abnor-
mal PTT results, may provide an opportunity to tailor
anticoagulant therapy to patients that require thrombo-
prophylaxis and limit its use in patients with an in-
creased risk of bleeding. Future prospective studies are
needed to evaluate the role of TEG in predicting throm-
bosis and outcome so that new or more targeted ther-
apies may be used to improve survival rates in this
often-fatal disease.
The authors would like to thank Dr. Amberly M.
Sokoloff for her assistance with database layout and
& Veterinary Emergency and Critical Care Society 2009, doi: 10.1111/j.1476-4431.2009.00455.x 487
Thromboelastography in canine IMHA
TEG3000 or 5000 Hemostasis Analyzer, Haemoscope Corporation, Niles, IL.
2.7 mL sodium citrate (0.3 mL; 0.109 M) Vacutainer, BD, Franklin Lakes,
Rose L, Bedard C, Dunn M. Effects of prednisone administration on
thromboelastography parameters in healthy beagles. J Vet Intern Med
Vilar P, Hansell J, Westendorf N et al. Effects of hematocrit on
thromboelastography tracings in dogs. J Vet Intern Med 2008;22(3):774.
1. Weinkle TK, Center SA, Randolph JF, et al. Evaluation of prog-
nostic factors, survival rates, and treatment protocols for immune-
mediated hemolytic anemia in dogs: 151 cases (19932002). J Am
Vet Med Assoc 2005; 226(11):18691880.
2. Piek CJ, Junius G, Dekker E, et al. Idiopathic immune-mediated
hemolytic anemia: treatment outcome and prognostic factors in
149 dogs. J Vet Intern Med 2008; 22(2):366373.
3. Carr AP, Panciera DL, Kidd L. Prognostic factors for mortality and
thromboembolism in canine immune-mediated hemolytic anemia: a
retrospective study of 72 dogs. J Vet Intern Med 2002; 16(5):504509.
4. Klein MK, Dow SW, Rosychuck RAW. Pulmonary thromboembo-
lism associated with immune-mediated hemolytic anemia in dogs:
ten cases (19821987). J Am Vet Med Assoc 1989; 195(2):246249.
5. Grundy SA, Barton C. Inuence of drug treatment on survival of
dogs with immune-mediated hemolytic anemia: 88 cases (1989
1999). J Am Vet Med Assoc 2001; 218(4):543546.
6. McManus PM, Craig LE. Correlation between leukocytosis and nec-
ropsy ndings in dogs with immune-mediated hemolytic anemia: 34
cases (19941999). J Am Vet Med Assoc 2001; 218(8):13081313.
7. Thomson MF, Scott-Moncrieff JC, Brooks MB. Effect of a single
plasma transfusion on thromboembolism in 13 dogs with primary
immune-mediated hemolytic anemia. J Am Anim Hosp Assoc
2004; 40(6):446454.
8. Palmer KG, King LG, Van Winkle TJ. Clinical manifestations and
associated disease syndromes in dogs with cranial venal cava
thrombosis: 17 cases (19891996). J Am Vet Med Assoc 1998;
9. Wiinberg B, Jensen AL, Johansson PI, et al. Thromboelastographic
evaluation of hemostatic function in dogs with disseminated in-
travascular coagulation. J Vet Intern Med 2008; 22(2):357365.
10. Kristensen AT, Wiinberg B, Jessen LR, et al. Evaluation of human
recombinant tissue factor-activated thromboelastography in 49
dogs with neoplasia. J Vet Intern Med 2008; 22(1):140147.
11. Otto CM, Reiser TM, Brooks MB, et al. Evidence of hypercoagul-
ability in dogs with parvoviral enteritis. J Am Vet Med Assoc 2000;
12. Ganter MT, Hofer CK. Coagulation monitoring: current techniques
and clinical use of viscoelastic point-of-care coagulation devices.
Anesth Analg 2008; 106(5):13661375.
13. Weiss DJ. Bone marrow pathology in dogs and cats with non-
regenerative immune-mediated haemolytic anaemia and pure red
cell aplasia. J Comp Pathol 2008; 138(1):4653.
14. Donahue SM, Otto CM. Thromboelastography: a tool for measur-
ing hypercoagulability, hypocoagulability, and brinolysis. J Vet
Emerg Crit Care 2005; 15(1):916.
15. Virchow RLK. Thrombose und embolie. Gefassentzu ndung und
septische infektion, In: Gesammelte Abhandlungen zur wissensc-
haftlichen Medicin. Frankfurt am Main: Von Meidinger & Sohn.
Reprint edition: Virchow R 1998; (AC Matzdorff, WR Bell transl,
Canton, MA: Science History Publications); 1856, pp. 219732.
16. Scott-Moncrieff JC, Treadwell NG, McCullough SM, et al. Hemo-
static abnormalities in dogs with primary immune-mediated he-
molytic anemia. J Am Anim Hosp Assoc 2001; 37:220227.
17. McCrath DJ, Cerboni EC, Frumento RJ, et al. Thromboelasto-
graphy maximum amplitude predicts postoperative thrombotic
complications including myocardial infarction. Anesth Analg
2005; 100(6):15761583.
18. Traverso C, Arcelus J, Gomez E, et al. Prospective assessment of
the risk of deep vein thrombosis in elective abdominal surgery.
Predictive role of thromboelastography. Thromb Haemorrhagic
Disord 1993; 71:915.
19. Patrassi GM, Sartori MT, Livi U, et al. Impairment of brinolytic
potential in long-term steroid treatment after heart transplanta-
tion. Transplantation 1997; 64(11):16101614.
20. Brainard BM, Meredith CP, Callan MB. Changes in platelet
function, hemostasis, and prostaglandin expression after treat-
ment with nonsteroidal anti-inammatory drugs with various
cyclooxygenase selectivities in dogs. Am J Vet Res 2007; 68(3):
21. Nelson OL, Andreasen C. The utility of plasma D-dimer to iden-
tify thromboembolic disease in dogs. J Vet Intern Med 2003;
22. Stokol T, Brooks M, Erb HN, et al. D-dimer concentrations in
healthy dogs and dogs with disseminated intravascular coagula-
tion. Am J Vet Res 2000; 61(4):393398.
& Veterinary Emergency and Critical Care Society 2009, doi: 10.1111/j.1476-4431.2009.00455.x 488
V.B. Sinnott & C.M. Otto