The American Journal of Surgery (2011) 201, 587591

North Pacific Surgical Association: First Place Residents Competition

Trauma induces a hypercoagulable state that is resistant

to hypothermia as measured by thrombelastogram
Jerome A. Differding, M.P.H., Samantha J. Underwood, M.S., Philbert Y. Van, M.D.,
Rakan A. Khaki, Nicholas J. Spoerke, M.D., Martin A. Schreiber, M.D., F.A.C.S.*
Division of Trauma, Critical Care and Acute Care Surgery, Oregon Health & Science University, 3181 Southwest Sam
Jackson Park Road, Portland, OR 97239, USA
KEYWORDS:
Trauma;
Hypercoagulable;
Thromboelastography;
Hypothermia

Abstract
BACKGROUND: The aim of this study was to test the hypothesis that severely injured trauma
patients would be hypercoagulable compared with controls measured by thromboelastography and that
this hypercoagulability would persist over a broad range of temperatures.
METHODS: A prospective study evaluating the effects of temperature on coagulation in trauma
patients with Injury Severity Scores 15 and controls was completed. Thromboelastography was
performed 24 hours after admission at 4 temperatures ranging from 32C to 38C.
RESULTS: Ninety-two subjects (46 patients) were analyzed. Patients had a median Injury Severity
Score of 20 (interquartile range, 16 26). Time to clot formation increased (P .001) and fibrin
cross-linking decreased (P .01) in both groups as temperature decreased. Between groups, time to
clot formation, fibrin cross-linking, and clot strength were significantly different at each temperature
(P .01), with patients being more hypercoagulable. Time to clot formation and fibrin cross-linking
were more affected by temperature in controls compared with patients (P .02).
CONCLUSIONS: Severely injured patients are more hypercoagulable than controls throughout a
broad range of temperature. Decreasing temperature has a greater effect on coagulation in controls
compared with patients.
2011 Elsevier Inc. All rights reserved.

Hemorrhage continues to be the leading cause of preventable death in trauma patients.1 Studies have shown that
the triad of hypothermia, acidosis, and coagulopathy contribute significantly to the severity of hemorrhage and rate
of mortality.2 6 Current methods of estimating coagulopathy include the prothrombin time (PT) and partial thromboplastin time (PTT). Human studies by Gubler et al7 and
Reed et al8 showed prolonged PT and PTT at lower temperatures. PT and PTT are poor indicators of in vivo coag* Corresponding author. Tel.: 503-494-3500; fax: 503-494-6519.
E-mail address: schreibm@ohsu.edu
Manuscript received November 9, 2010; revised manuscript January
21, 2011

0002-9610/$ - see front matter 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.amjsurg.2011.01.012

ulation for several reasons. First, these assays are routinely

performed at 37C in the laboratory, underestimating the
effects of hypothermia on coagulation. Second, to perform
these assays, blood is centrifuged, excluding the contribution of cellular elements to in vivo coagulation. Finally, PT
and PTT are designed to analyze specific portions of the
clotting cascade, not overall clot performance.9 Previous
swine studies evaluating the use of thromboelastography
(TEG; Haemoscope Corporation, Niles, IL) to estimate coagulation at various temperatures have shown statistically
significant coagulopathy at lower temperatures. Additionally, a limited human study of hypothermic patients undergoing neurosurgical procedures demonstrated coagulopathy
at lower temperatures when measured by TEG.6 In a 1987

588

The American Journal of Surgery, Vol 201, No 5, May 2011

Figure 1

Thromboelastographic tracing comparing 32C with 38C in a control participant.

study of TEG in hypothermic trauma patients, body temperature 34C correlated with abnormal values on TEG.10
Confounding these findings is the fact that the more hypothermic patients had significantly greater Injury Severity
Scores (ISS) and received more fluid resuscitation than the
less hypothermic patients.11 No study has evaluated the
direct effects of temperature on coagulation in trauma patients as measured by TEG.

Methods
This was a prospective, noninterventional, single-center
study evaluating the effects of temperature on coagulation
measured by TEG in normal control subjects and in severely
injured trauma patients. All trauma patients with ISS 15
were eligible to be enrolled in the study. Patients were
excluded from eligibility for the following reasons: current
use of therapeutic anticoagulation, preexisting coagulopathy, pregnancy, or inability to obtain consent from either the
patient or a legal authorized representative. Patients and
controls were consented before enrollment in the study.
Controls were age and gender matched to patients.

moves and clotting is initiated, the blood begins to adhere to

the pin, which in turn deflects the torsion wire. The deflection of the torsion wire increases as the clot strengthens. In
1 hour, the thromboelastographic tracing and its respective values give an interpretation of the entire coagulation
panel (Fig. 1).
Analysis of kaolin-activated thromboelastograms were
performed 24 hours after admission using venous whole
blood at 38C, 36C, 34C, and 32C. For subjects receiving heparin, thromboelastographic assays were run with and
without heparinase. This was done to negate the effects of
unfractionated heparin or lowmolecular weight heparin on
the assay. The primary end point was the effect of temperature on coagulation status. The institutional review board
of Oregon Health & Science University approved the study,
and written consent was obtained from subjects or their
legal representatives.

Statistical analysis
Statistical analyses were performed using SPSS version
18.0 (SPSS, Inc, Chicago, IL). For normally distributed
data, an analysis of variance and Students t test were used

TEG assay
Table 1

Citrated samples were collected and analyzed using

TEG, a point-of-care coagulation test that uses time-dependent viscoelastic properties of whole blood. One milliliter of
blood was placed in a vial containing kaolin, which accelerates clotting. A fixed quantity of the kaolin-activated
blood (340 L) and 20 L of calcium chloride were pipetted into a cup housed in a carriage assembly. A pin
affixed to a torsion wire was lowered into the cup, which
begins to oscillate at a fixed rate and amplitude. As the cup

Demographics

Variable

Patients

Controls

Age (y)
Men
Women
ISS

48 (3164)
23
23
20 (1626)

38 (3350)
23
23
No data

Data are expressed as median (interquartile range) or as numbers.

There were no statistical differences between any of the groups (P
.05).

J.A. Differding et al.

Hypercoagulability is resistant to hypothermia

589

Figure 2 Mean differences in time to clot formation (R-time). Ct

controls; Pt patients. *P .001 between 32C and other
temperatures; #P .01 between groups.

Figure 3 Mean differences in clot strength. Ct controls; Pt

patients. *P .001 between 32C and other temperatures; #P
.01 between groups.

to determine differences between groups. Mann-Whitney U

tests were used for nonparametric comparisons between
groups. Paired t tests and Wilcoxons tests were conducted
for normally distributed and nonparametric paired comparisons, respectively. P values .05 were considered statistically significant.

temperature (Figs. 2 and 3, Table 2). No difference was seen

in the percentage lysis between the groups at any temperature (Table 2). Decreases in temperature produced greater
increases in time to clot formation (Fig. 4) and greater
decreases in clot strength in controls compared with patients
(Fig. 5).

Results

Comments

Ninety-two subjects (46 patients) were included in the

analysis. Participant demographics are highlighted in Table
1. Both groups were similar with respect to age and gender,
and the trauma group was severely injured, as indicated by
a median ISS of 20. Time to clot formation increased (P
.001) in both groups as temperature decreased (Fig. 2).
Fibrin cross-linking significantly decreased (P .001) in
both groups with decreasing temperatures (Table 2). Clot
strength was unaffected by temperature in both groups, with
the exception that it was greater in the control group at 32C
compared with other temperatures. Patients were more hypercoagulable than controls as measured by time to clot
formation, clot strength, and fibrin cross-linking at each

Hemorrhage and coagulopathy continue to complicate

the care of critically injured patients. A number of factors
contribute to coagulopathy after trauma, including hemorrhage, shock, acidosis, fluid resuscitation, and hypothermia.
Hypothermia is common after trauma, with up to 66% of
trauma patients having temperatures 36C on arrival to
the emergency department. In a group of patients with ISS
25, the mortality rate increases from 10% to as high as
100% when body temperatures fall from 35C to 32C.
The exact contribution of hypothermia to coagulopathy has
not been elucidated.
The current standard for measuring coagulation status is
PT and activated PTT. These tests do not provide informa-

Table 2

Thromboelastographic data

Variable
Fibrin cross-linking
Patients
Controls
Lysis
Patients
Controls

32C

34C

36C

38C

62.7 (57.466.4)*
54.1 (48.457.5)*

63.4 (59.267.3)*
57.0 (52.360.9)*

67.1 (63.470.6)*
57.7 (53.261.7)*

67.8 (62.569.9)*
61.6 (55.965.3)*

.05 (.00.50)
.15 (.001.50)

.30 (.001.03)
.20 (.00.75)

.95 (.201.60)
.50 (.101.30)

Data are expressed as median (interquartile range).

*There was significantly less fibrin cross-linking as temperature decreased within group (P .001).
There was significantly less fibrin cross-linking in controls versus patients at all temperatures (P .001).

1.20 (.502.65)
1.75 (.703.30)

590

Figure 4 Median changes in time to clot formation (R). CT

controls; PT patients. *Significantly lower time between groups
(P .001). **Significantly lower time between groups (P .02).

tion about the entire coagulation and fibrinolytic cascade,

are poor indicators of in vivo coagulation, and are not
adequately sensitive to detect clinically significant coagulopathy. These measurements are performed by drawing
and centrifuging the patients blood, collecting the plasma,
and then adding activating factors to the plasma at 37C.
They represent a poor estimation of in vivo clotting ability
because they neglect the cellular contributions to coagulation. In addition, running the samples at 37C ignores the
temperature-dependent depression of enzymatic function.
Additionally, PT and activated PTT have been shown to be
unable to detect derangements in coagulation in hypothermic injured swine compared with TEG.
TEG is an excellent test for assessing overall coagulation.
TEG rapidly assesses the interaction of platelets with the protein coagulation cascade from the initial platelet-fibrin interaction, through platelet aggregation, clot strengthening, and fibrin
cross-linking, to eventual clot lysis. In 30 minutes, a thromboelastogram can provide information on clotting factor activity, platelet function, and any clinically significant fibrinolytic
process.

The American Journal of Surgery, Vol 201, No 5, May 2011

TEG has been shown to be effective in the evaluation of
overall coagulation in trauma patients. Coagulopathies have
been associated with the severity of injury and temperature.
Minimally injured patients have normal results on TEG,
whereas moderately injured patients are hypercoagulable by
TEG on admission. Severely injured patients are frequently
hypocoagulable initially, placing them at greatest risk for
bleeding. The effect of temperature on TEG in this patient
group has not been evaluated.
This study allowed us to make several important observations. A reduction in temperature resulted in increased
time to clot formation and decreased fibrin cross-linking in
both groups. Clot strength was relatively unaffected by
reductions in temperature. These findings suggest that clotting factor function and fibrin cross-linking are more affected by the temperature ranges studied than platelet function. We also found that severely injured trauma patients are
more hypercoagulable than a comparable healthy population throughout the range of temperatures studied. Interestingly, reductions in temperatures had a greater effect on
coagulation in control patients compared with trauma patients, suggesting that hypercoagulability observed after severe trauma is resistant to hypothermia.
There were several limitations associated with this study.
The effect of temperature on coagulation was measured in
vitro using TEG, and these measurements may not accurately reflect temperature effects in vivo. We chose to perform coagulation assessments 24 hours after patients were
admitted to the hospital to obtain a more uniform population
of patients. Whereas hypocoagulability has been shown to
increase with severity of injury at the time admission,12 the
incidence of hypercoagulability after resuscitation and stabilization of trauma patients is very high.13

Conclusions
Hypothermia is associated with increased time to clotting
and decreased fibrin cross-linking in both severely injured
trauma patients and controls. Stabilized trauma patients are
more hypercoagulable than controls with respect to all aspects of clotting as measured by TEG regardless of temperature. Hypercoagulability in severely injured trauma patients is relatively resistant to hypothermia compared with
controls.

References

Figure 5 Median changes in clot strength (maximum amplitude

[MA]). CT controls; PT patients. *Significantly less clot
strength (P .05).

1. Hess J, Brohi K, Dutton R, et al. The coagulopathy of trauma: a review

of mechanisms. J Trauma 2008;65:748 54.
2. Watts DD, Trask A, Soeken K, et al. Hypothermic coagulopathy in
trauma. J Trauma 1998;44:846 54.
3. Douning LK, Ramsay MA, Swygert TH, et al. Temperature corrected
thrombelastography in hypothermic patients. Anesth Analg 1995;81:
608 11.

J.A. Differding et al.

Hypercoagulability is resistant to hypothermia

4. Martini WZ, Pusateri AE, Uscilowicz JM, et al. Independent contributions of hypothermia and acidosis to coagulopathy in swine.
J Trauma 2005;58:100210.
5. Kettner SC, Kozek SA, Groetzner JP, et al. Effects of hypothermia on
thrombelastography in patients undergoing cardiopulmonary bypass.
Br J Anaesth 1998;80:3137.
6. Kettner SC, Sitzwohl C, Zimpfer M, et al. The effect of graded
hypothermia on hemostasis in anesthetized patients without surgical
trauma. Anesth Analg. 2003;96:1772 6.
7. Gubler KD, Gentilello LM, Hassantash A, et al. The impact of hypothermia on dilutional coagulopathy. J Trauma 1994;36:84751.
8. Reed RL, Johnston TD, Hudson, JD, et al. The disparity between hypothermic coagulopathy and clotting studies. J Trauma 1992;33:46570.
9. Felfernig M, Blaicher A, Kettner SC, et al. Effects of temperature on
partial thromboplastin time in heparinized plasma in vitro. Eur J
Anaesth 2001;18:46770.
10. Luna GK, Maier RV, Pavlin EG, et al. Incidence and effect of hypothermia in seriously injured patients. J Trauma 1987;27:1014 8.
11. Jurkovich GJ, Grieser WB, Luterman A, et al. Hypothermia in
trauma victims; an ominous predictor of survival. J Trauma 1987;
27:1019 24.
12. Kaufman CR, Dwyer KM, Crews JD, et al. Usefulness of thrombelastogram with common coagulation tests. Thromb Haemast 1981;46:752 6.
13. Schreiber MA, Differding J, Thorborg P, et al. Hypercoagulability is
most prevalent early after injury and in female patients. J Trauma
2005;58:475 80.

Discussion
Robert M. Rush, M.D. (Tacoma, WA): When a bleeding and sick trauma patient enters the resuscitation bay, the
last thing one thinks of is hypercoagulation. Usually it is the
opposite: Lets get this bleeding stopped! The authors

591

describe a hypercoagulable state associated with severe

trauma that is temperature independent and can only be
measured by thromboelastogram, specifically the Haemoscope Corp TEG machine in their study. Thromboelastography (TEG) was first developed in the late 1940s but is still
not guiding our clinical decision making for trauma patients. Studies by Kaufman and colleagues (Kaufmann CF,
et al. J Trauma 1997;42:716 722), as well as other studies
by this group, have shown that many trauma patients are
initially hypercoagulable or become hypercoagulable
shortly after initial stabilization in the first 24 to 48 hours.
This study adds a very important basic science piece to the
characterization of TEG with respect to other coagulation
parameters.
So when is the crossover? When do we start stopping the
stop the bleeding mentality and start thinking about the
hypercoagulable state that the authors have elucidated, not
only in this paper, but in their prior endeavors? I believe that
we always have to think about it and thus the current
recommendation according to the Joint Theater Trauma
System Clinical Practice Guideline is to begin DVT prophylaxis therapy as soon as coagulopathy is corrected in
patients not otherwise at increased risk of bleeding. It is
quite possible that the thromboelastogram could be the
study to help determine this crossover point, but certainly
not in a vacuum. The patients clinical picture, other coagulation studies, and potentially their fibrinogen levels are
also important in determining the safety of initiating higher
doses for chemoprophylaxis of thromboembolic events in
the severely injured.