More Related Content
Similar to Cedars-Sinai Publication
Similar to Cedars-Sinai Publication (20)
Cedars-Sinai Publication
- 1. Prehospital hypertension is predictive of traumatic brain injury
and is associated with higher mortality
Galinos Barmparas, MD, Douglas Z. Liou, MD, Alexander W. Lamb, BS, Alexandra Gangi, MD,
Mike Chin, BS, Eric J. Ley, MD, Ali Salim, MD, and Marko Bukur, MD, Delray Beach, Florida
BACKGROUND: The purpose of the current study was to investigate the effect of early adrenergic hyperactivity as manifested by prehospital
(emergency medical service [EMS]) hypertension on outcomes of traumatic brain injury (TBI) patients and to develop a
prognostic model of the presence of TBI based on EMS and admission (emergency department [ED]) hypertension.
METHODS: This study is a retrospective review of the 2007 to 2008 National Trauma Data Bank including blunt trauma patients 15 years or
older with available EMS and ED vital signs. Patients with head Abbreviated Injury Scale (AIS) score of 3 or greater were
selected, and mortality was examined within EMS systolic blood pressure (SBP) groups: lower than 100 mm Hg, 110 mm Hg
to 150 mm Hg, 160 mm Hg to 180 mm Hg, and 190 mm Hg to 230 mm Hg. A forward logistic regression model including the
EMS heart rate, EMS SBP, EMS Glasgow Coma Scale (GCS) score, ED heart rate, and ED SBP was used to identify predictors
of a TBI in patients with ED GCS score of less than or equal to 8, 9 to 13, and 14 to 15.
RESULTS: For the 5-year study period, 315,242 patients met inclusion criteria. Adjusted odds for mortality increased in a stepwise fashion
with increasing EMS SBP compared with patients with normal EMS SBP (adjusted odds ratio [95% confidence interval], 1.33
[1.22Y1.44], p G 0.001, for EMS SBP of 160Y180 mm Hg and 1.97 [1.76Y2.21], p G 0.001, for EMS SBP of 190Y230 mm Hg).
A 7-point scoring system was developed for each ED GCS score group to predict the presence of a TBI. EMS SBP of greater
than 150 mm Hg and ED SBP of greater than 150 mm Hg were both predictive of the presence of a TBI in patients with ED
GCS score of 8 or less and in patients with ED GCS score of 9 to 13 or 14 to 15, respectively.
CONCLUSION: Prehospital hypertension in TBI is associated with a higher mortality risk. Early hypertension in the prehospital setting and at
admission can be used to predict the presence of such injuries. These findings may have important early triage and treatment
implications. (J Trauma Acute Care Surg. 2014;77: 592Y598. Copyright * 2014 by Lippincott Williams & Wilkins)
LEVEL OF EVIDENCE: Prognostic study, level III.
KEY WORDS: Traumatic brain injury; hypertension; adrenergic hyperactivity; prediction; score.
Traumatic brain injury (TBI) is the leading cause of death
and disability worldwide.1
In the United States alone, there
are at least 1.7 million injuries secondary to TBI yearly,
resulting in nearly one third of injury-related deaths.2
Most of
the deaths are linked to the secondary injury processes that are
a consequence of additional insults such as hypoxia, hypoten-
sion,3
and a cascade of complex neurophysiologic interactions.4
Advances in neurocritical care have resulted from a better un-
derstanding of the neurobiology behind TBI and have lead to
reductions in mortality after TBI.5
One complex neurobiologic interaction, known as par-
oxysmal sympathetic hyperactivity or sympathetic storm, is the
dysregulation of the sympathetic nervous system that has been
shown to occur after severe TBI and that is cathecholamine
mediated.6Y8
Symptoms of paroxysmal sympathetic hyperac-
tivity include hypertension, tachycardia, fever, and tachypnea9
and have been shown to occur in up to one third of TBI patients
admitted to the neurologic intensive care unit.10
Elevated levels
of serum catecholamines and repetitive episodes of paroxysmal
sympathetic hyperactivity are predictive of negative outcomes
after severe TBI.11,12
Identifying patients at risk for TBI early after admission
is crucial, as it can expedite treatment to avoid secondary
neurologic insult. The diagnosis of TBI is largely based upon
clinical symptoms and incorporates the level of consciousness,
as measured by the Glasgow Coma Scale (GCS) score, and
pupillary examination, which typically prompt neuroimaging
with computed tomography. However, these clinical findings
can easily be confounded by neurodepressants such as illicit
drugs, alcohol, or patient hemodynamic instability.13
While
there have been several prognostic models examining out-
comes after TBI,14,15
few have examined a model to predict
TBI in patients presenting to the emergency department (ED).
The purpose of the current study was twofold: to investi-
gate the effect of early adrenergic hyperactivity as manifested by
prehospital (emergency medical service [EMS]) hypertension on
outcomes of TBI patients and to develop a preliminary prog-
nostic model of the presence of TBI based on EMS and ad-
mission (ED) vital signs.
ORIGINAL ARTICLE
J Trauma Acute Care Surg
Volume 77, Number 4592
Submitted: March 31, 2014, Revised: May 4, 2014, Accepted: May 20, 2014.
From the Department of Surgery (G.B., D.Z.L., A.W.L., A.G., M.C., E.J.L.), Di-
vision of Acute Care Surgery and Surgical Critical Care, Cedars-Sinai Medical
Center, Los Angeles, California; Department of Surgery (A.S.), Division of
Acute Care Surgery and Surgical Critical Care, Brigham and Women’s Hospital,
Boston, Massachusetts; and Department of Surgery (M.B.), Division of Trauma
Surgery and Surgical Critical Care, Delray Medical and Broward Health
Medical Centers, Fort Lauderdale, Florida.
Address for reprints: Marko Bukur, MD, Delray and Broward Health Medical
Centers, Delray Trauma Offices, Fair Oaks Pavilion, 5352 Linton Blvd, Delray
Beach, FL 33484; email: mbukur@browardhealth.org.
DOI: 10.1097/TA.0000000000000382
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
- 2. PATIENTS AND METHODS
For the purposes of this study, data from the National
Trauma Data Bank (NTDB) Research Data Sets 2007 and 2008
were used. Patients presenting with a blunt injury, 15 years or
older, and with available EMS and ED vital signs were included.
Patients with missing demographics, vital signs, nonsurvivable
injuries defined as any body region Abbreviated Injury Scale
(AIS) score of 6,orhospitaldisposition were excluded.A dataset
was created that incorporated the variables age, sex, EMS heart
rate (HR), EMS systolic blood pressure (SBP), ED HR, ED SBP,
ED GCS score, body region AIS score, Injury Severity Score
(ISS), and ED disposition. Outcomes included overall mortality
and the predicted presence of TBI. For the purposes of the
analysis, clinically relevant cut points were created and included
age older than 55 years, ISS of less than or equal to 16, ISS of
17 to 25, ISS of greater than 25, and isolated head injuries, de-
fined as head AIS score of 3 or greater with all other body AIS
scores of less than 3.
Defining Normal Vital Signs
Using the outlier labeling rule,16,17
patients with outlying
values for EMS HR, EMS SBP, ED HR, and ED SBP were
excluded. To achieve this, the values for the lowest (25th or Q1)
and the highest (75th or Q2) quartile for each variable were
identified, and the difference between the two was multiplied by
the factor g (2.2) to identify the upper and the lower limit for
each variable according to the following formulas: Q3 + 1.5 Â
(Q3 j Q1) and Q3 + 1.5 Â (Q3 j Q1), respectively. Values
higher and lower than the resultant values, respectively, were
excluded. Subsequently, these same variables were examined
as increments of 10. Mortality for each EMS HR and SBP
increment was plotted, and those with a lower crude mortality,
compared with the overall crude mortality of the examined
population, were considered to represent normal vital signs.
Effect of Early Hypertension on Outcomes
of TBI Patients
To examine the effect of early hypertension on outcomes
of TBI patients, patients with head AIS score of 3 or greater
were selected from the entire population, and mortality was
examined within EMS SBP groups: lower than 100 mm Hg,
110 mm Hg to 150 mm Hg, 160 mm Hg to 180 mm Hg, and
190 mm Hg to 230 mm Hg. The four groups were compared
using analysis of variance. Logistic regression model was used
to adjust for confounding factors between the groups. Patients
with normal EMS SBP (100Y150 mm Hg) were considered the
reference group.
Development of a Predictive Model for TBI
Based on EMS and ED Vital Signs
For this part of the analysis, patients with available ED
GCS score were selected and were randomly divided into two
groups: the study cohort and the validation cohort. Each one of
these groups was subdivided into three groups based on ad-
mission GCS score: less than or equal to 8, 9 to 13, and 14 to
15. The study cohort was then selected, and a forward logistic
regression model including, among other variables, the EMS
HR, EMS SBP, EMS GCS score, ED HR, and ED SBP was
used for each one of these subgroups to identify predictors of a
TBI. For each predictor, the adjusted odds ratio (AOR) was
multiplied by the coefficient, and the result was rounded. This
rounded number was considered as a score number for each
variable. Using this scoring system, each patient received a
score. The incidence of TBI based on this score was compared
between the study and the validation cohort.
All statistical analyses were performed using the IBM
SPSS Statistics for Windows, Version 20.0 (IBM Corp., Armonk,
NY). This study received an institutional review board exemption.
RESULTS
Defining Normal Vital Signs
During the 2-year study period, of a total of 1,134,946
eligible patients, 315,242 (27.8%) met the inclusion criteria.
After applying the outlier labeling rule, it was determined that
patients with EMS SBP of higher than 216 mm Hg or lower than
54 mm Hg, patients with EMS HR of higher than 148 beats
per minute (bpm) or lower than 30 bpm, patients with ED SBP
of higher than 225 mm Hg or lower than 53 mm Hg, and, lastly,
patients with ED HR of higher than 153 bpm or lower than
23 bpm were all outliers. This left 305,503 patients for analysis.
EMS SBP and HR were then divided into increments of
10, and mortality was plotted on the basis of these increments
(Fig. 1). The overall mortality for the entire population was
3.2%. Increments with lower, compared with the overall, mor-
tality were selected to represent normal vital signs, that is, EMS
SBP of higher than 100 mm Hg, EMS SBP of 150 mm Hg or
lower,EMSHRofgreater than70bpm,andEMSHRof100bpm
or less. These same normal vital signs were applied for the ED
SBP and HR.
Effect of Early Hypertension on Outcomes
of TBI Patients
Patients with head AIS score of 3 or greater were then
selected (n = 45,732) and analyzed separately. The mean (SD)
agewas48.5(22.6)years(median,46.0years),69.6%weremale,
and 28.2% had isolated TBI. These patients were then divided
into three groups, based on EMS SBP: lower than 100 mm Hg
(n = 2,913 or 6.4%), 100 mm Hg to 150 mm Hg (n = 32,505 or
71.1%), 160 mm Hg to 180 mm Hg (n = 7,826 or 16.5%), and
190 mm Hg to 230 mm Hg (n = 1,821 or 4.0%). Isolated TBI was
more likely to occur in patients with higher EMS SBP: 18.4%
(535 of 2,913) for EMS SBP of lower than 100 mm Hg, 28.4%
(9,222 of 32,505) for EMS SBP of 100 mm Hg to 150 mm Hg,
30.2% (2,360 of 7,826) for EMS SBP of 160 mm Hg to
180 mm Hg, and 30.7% (765 of 2,488) for EMS SBP of
190 mm Hg to 230 mm Hg (p G 0.001). The overall mortality
was 13.9%; 28.5%, 10.7%, 17.2%, and 27.1%, respectively, for
each group (p G 0.001). After adjusting for differences between
the four groups, including age, sex, ISS, EMS GCS score, AIS
score for all body regions, and isolated TBI, patients who were
hypotensive at the scene (EMS SBP G 100 mm Hg) had a sig-
nificantly higher risk for mortality compared with patients who
were normotensive (EMS SBP, 100Y150 mm Hg) (Fig. 2). In
addition, adjusted odds for mortality increased in a stepwise
fashion with increasing EMS SBP compared with patients with
normal EMS SBP (Fig. 2). This pattern of increase in mortality
J Trauma Acute Care Surg
Volume 77, Number 4 Barmparas et al.
* 2014 Lippincott Williams & Wilkins 593
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
- 3. with increasing EMS SBP from normal was also found within
ISS groups (e16, 17Y25, and 925), age groups (e55 and
955 years), and isolated TBI group.
Development of a Predictive Model for TBI
Based on EMS and ED Vital Signs
Of 305,503 patients available for analysis, a total of
269,645 had available ED GCS score. This cohort was then
randomly divided into the study and the validation cohort and
subdivided into three groups based on admission (ED) GCS
score (Fig. 3). In the study cohort, there was an increasing
incidence of TBI with decreasing ED GCS score; however,
only 58.7% of patients in the groups with GCS score of 8 or less
had a TBI. Multiple forward logistic regressions were used for
each one of the ED GCS score groups to identify predictors of
a TBI. Each of these regressions included age, sex, EMS HR
(G70 bpm, 70Y100 bpm, 9100 bpm), EMS SBP (G100 mm Hg,
100Y150 mm Hg, 9150 mm Hg), EMS GCS score (e8, 9Y13,
Figure 1. Mortality according to increments of 10 of prehospital SBP (EMS SBP) (A) and prehospital HR (EMS HR) (B). Increments
with lower mortality, compared with the overall mortality of 3.2%, are highlighted in black. These increments, that is, EMS SBP
of higher than 110 and lower than 160 mm Hg and EMS HR of greater than 70 and less than 110 bpm, were considered
normal prehospital vital signs.
Figure 2. AOR and 95% confidence interval (CI) for mortality in patients with AIS score of 3 or greater (n = 45,732), based on
prehospital (EMS) SBP groups. Patients with EMS SBP of 100 mm Hg to 150 mm Hg were considered the reference group.
Variables in the model: age (e55 vs. 955 years), sex (male vs. female), ISS groups (e16 vs. 17Y25 vs. 925), EMS GCS score groups
(e8 vs. 9Y13 vs. 14Y15), chest AIS score (Q4 vs. G4), abdomen AIS score (Q4 vs. G4), extremity AIS score (Q4 vs. G4), and isolated
TBI, defined as head AIS score of 3 or greater with all other AIS scores of less than 3 (yes vs. no).
J Trauma Acute Care Surg
Volume 77, Number 4Barmparas et al.
594 * 2014 Lippincott Williams & Wilkins
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
- 4. 14Y15), ED HR (G70 bpm, 70Y100 bpm, 9100 bpm), and ED
SBP (G100 mm Hg, 100Y150 mm Hg, 9150 mm Hg). Of note,
ED SBP of higher than 150 mm Hg was one of the predictors
of TBI for patients admitted with a GCS score of 8 or less, as
was EMS SBP of higher than 150 mm Hg for patients admitted
with a GCS score of 9 to 13 or 14 to 15. For each ED GCS
score group, each predictor received a score based on the AOR
multiplied by the coefficient (Table 1). All of these scores were
Figure 3. Development of the predictive score for TBI.
TABLE 1. Independent Predictors for Severe Head Injury by Admission (ED) GCS Score With Proposed Score for Each Variable
Based on the Coefficient and the AOR
Group Variable Adjusted OR (95% CI) Adjusted p Coefficient Coefficient  Adjusted OR Assigned Score
ED GCS score e 8* Age 9 55 y 1.61 (1.44Y1.81) G0.001 0.48 0.77 1
EMS HR G 70 bpm 1.83 (1.58Y2.12) G0.001 0.61 1.11 2
EMS GCS score e 8 1.92 (1.72Y2.16) G0.001 0.66 1.26 2
ED SBP 9 150 mm Hg 1.35 (1.20Y1.51) G0.001 0.30 0.41 1
ED HR G 70 bpm 1.54 (1.32Y1.79) G0.001 0.43 0.66 1
ED GCS score 9Y13** Age 9 55 y 1.73 (1.54Y1.95) G0.001 0.55 0.95 1
Male 1.22 (1.09Y1.37) 0.001 0.20 0.25 1
EMS HR G 70 bpm 1.68 (1.39Y2.04) G0.001 0.52 0.88 1
EMS SBP 9 150 mm Hg 1.40 (1.21Y1.62) G0.000 0.33 0.47 1
EMS GCS score e 8 1.53 (1.32Y1.77) G0.001 0.42 0.65 1
ED GCS score 14Y15† Age 9 55 y 1.73 (1.66Y1.80) G0.001 0.55 0.94 1
Male 1.48 (1.42Y1.54) G0.001 0.39 0.58 1
EMS HR G 70 bpm 1.18 (1.12Y1.25) G0.001 0.17 0.20 1
EMS SBP 9 150 mm Hg 1.53 (1.41Y1.66) G0.000 0.42 0.65 1
EMS GCS score e 8 1.27 (1.08Y1.49) 0.004 0.24 0.30 1
*Variables in the equation: older than 55 years (yes vs. no), sex (male vs. female), EMS HR of less than 70 bpm (yes vs. no), EMS HR of 70 bpm to 100 bpm (yes vs. no), EMS HR of
greater than 100 bpm (yes vs. no), EMS SBP of lower than 100 mm Hg (yes vs. no), EMS SBP of 100 mm Hg to 150 mm Hg (yes vs. no), EMS SBP of higher than 150 mm Hg (yes vs. no),
EMS GCS score of 8 or less (yes vs. no), EMS GCS score of 9 to 13 (yes vs. no), EMS GCS score of 14 to 15 (yes vs. no), ED HR of less than 70 bpm (yes vs. no), ED HR of 70 bpm to
100 bpm (yes vs. no), ED HR of greater than 100 bpm (yes vs. no), ED SBP of lower than 100 mm Hg (yes vs. no), ED SBP of 100 mm Hg to 150 mm Hg (yes vs. no), ED SBP of
higher than 150 mm Hg (yes vs. no); area under the curve for the model, 0.593.
**Variables in the equation: older than 55 years (yes vs. no), sex (male vs. female), EMS HR of less than 70 bpm (yes vs. no), EMS HR of 70 bpm to 100 bpm (yes vs. no), EMS HR of
greater than 100 bpm (yes vs. no), EMS SBP of lower than 100 mm Hg (yes vs. no), EMS SBP of 100 mm Hg to 150 mm Hg (yes vs. no), EMS SBP of higher than 150 mm Hg (yes vs. no),
EMS GCS score of 8 or less (yes vs. no), EMS GCS score of 9 to 13 (yes vs. no), EMS GCS score of 14 to 15 (yes vs. no), ED HR of less than 70 bpm (yes vs. no), ED HR of 70 bpm
to 100 bpm (yes vs. no), ED HR of greater than 100 bpm (yes vs. no), ED SBP of lower than 100 mm Hg (yes vs. no), ED SBP of 100 mm Hg to 150 mm Hg (yes vs. no), ED SBP of
higher than 150 mm Hg (yes vs. no); area under the curve for the model, 0.632.
†Variables in the equation: older than 55 years (yes vs. no), sex (male vs. female), EMS HR of less than 70 bpm (yes vs. no), EMS HR of 70 bpm to 100 bpm (yes vs. no), EMS HR of
greater than 100 bpm (yes vs. no), EMS SBP of lower than 100 mm Hg (yes vs. no), EMS SBP of 100 mm Hg to 150 mm Hg (yes vs. no), EMS SBP of higher than 150 mm Hg (yes vs. no),
EMS GCS score of 8 or less (yes vs. no), EMS GCS score of 9 to 13 (yes vs. no), EMS GCS score of 14 to 15 (yes vs. no), ED HR of less than 70 bpm (yes vs. no), ED HR of 70 bpm to
100 bpm (yes vs. no), ED HR of greater than 100 bpm (yes vs. no), ED SBP of lower than 100 mm Hg (yes vs. no), ED SBP of 100 mm Hg to 150 mm Hg (yes vs. no), ED SBP of
higher than 150 mm Hg (yes vs. no); area under the curve for the model, 0.622.
CI, confidence interval; OR, odds ratio.
J Trauma Acute Care Surg
Volume 77, Number 4 Barmparas et al.
* 2014 Lippincott Williams & Wilkins 595
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
- 5. used to develop a predictive model for TBI based only on sex,
prehospital and admission HR, SBP, and GCS score (Table 2).
With higher score, there was increasing incidence of TBI,
reaching 80.7% in patients with admission GCS score of 8 or
less and a maximum score of 5 (Table 3). The scoring system
was then implemented on the validation cohort, and the inci-
dence of TBI per group and with increasing score was almost
identical to that found in the study cohort (Table 3).
DISCUSSION
Despite the modern advances in neurocritical care, severe
TBI continues to be a significant cause of mortality after trauma,
accounting for one third to one half of all posttraumatic deaths,18
as well as a heavy financial burden on our health care system,
with rehabilitation costs of $60 billion annually in the United
States alone.19
Early aggressive care based upon evidence-based
guidelines has resulted in improved outcomes when instituted
early after injury.20Y22
As the diagnosis of TBI is clinical and relies heavily upon
the patients’ level of consciousness, this can present challenges
to the treating medical team. The GCS, introduced by Teasdale
and Jennett23
in the 1970s, is the scale most frequently used to
assess a patient’s level of consciousness; however, it can be
affected by ethanol and other neurodepressants.13,24,25
An all
too common scenario is the intoxicated patient with concom-
itant TBI, which occurs in 35% to 50% of these injuries26
and
can be nearly impossible to distinguish on clinical diagnosis
alone. While the GCS score has been shown to improve in
intoxicated patients with and without TBI as ethanol is me-
tabolized,27
rarely is an injured patient observed before diag-
nostic workup. In this NTDB series, we have created a simple
7-point model that can be rapidly calculated using prehospital
and admission parameters to estimate the incidence of asso-
ciated TBI in patients presenting following a blunt traumatic
injury. By demonstrating an association between elevated EMS
and ED blood pressure as well as documented intracranial
injuries, it is suggested that early markers of sympathetic ac-
tivity may aid in a more expeditious diagnosis of TBI. The use
of these markers in the early prediction of TBI has not been
previously explored. Perhaps the patients most likely to benefit
from these findings are those who are labeled as ‘‘found down,’’
as they may be initially triaged for medical evaluation and often
have a high incidence of occult injury28
and significant in-
hospital mortality. Previous investigators have demonstrated
that trauma team involvement leads to more expeditious workup
and diagnosis compared with those solely evaluated by non-
surgical teams initially.29
Early diagnosis of TBI is vitally im-
portant for optimizing patient morbidity and mortality, and
earlier intervention and appropriate neurosurgical consulta-
tion may lead to improved outcomes.
It has been well established that, after TBI, an increase in
adrenergic activity is associated with negative outcomes.30
The
paroxysmal sympathetic hyperactivity that occurs after trauma
exacerbates the mechanical damage done by the initial injury
and begins almost immediately after insult, contributing to
secondary brain injury. Although the full mechanism behind
sympathetic hyperactivity has been difficult to pinpoint, cate-
cholamines after TBI are a likely contributor to the injury
cascade. This secondary injury can manifest itself as increased
intracranial pressure, changes in cerebral blood flow and is-
chemia, and cerebral immune dysfunction.31
We have previ-
ously demonstrated that TBI patients presenting in the ED with
an SBP of greater than 160 mm Hg are at a higher risk for
delayed complications.32
In addition, it has been demonstrated
that prehospital hypertension of greater than 160 mm Hg in
TBI patients is an indicator for in-hospital mortality,33
and a
deviation from admission HR of 70 bpm to 89 bpm is asso-
ciated with increased mortality.32
Our current study appears to
be congruent with these findings and may be suggestive that
early treatment of these abnormalities may improve outcomes
in patients with TBI.
One potential therapeutic intervention for the treatment of
paroxysmal sympathetic hypertension is the use of A-blockers.
Treatment with A-adrenergic blockers after TBI has been asso-
ciated with lower mortality in several previous retrospective
studies.34,35
Although the exact drug, dose, and timing of
therapeutic administration are yet unknown, the nonselective
A-adrenergic receptor antagonist propranolol is promising,
based upon several animal studies.36Y38
Propranolol is highly
lipophilic and thus able to pass through the blood-brain barrier
and may alleviate sympathetic hyperactivity due to TBI by pre-
venting the overstimulation of the adrenergic receptors. There is
currently an ongoing prospective randomized clinical trial using
propranolol as a pharmacologic therapeutic in moderate-
to-severe TBI.39
Further prospective studies examining early
TABLE 2. Proposed Score for Prediction of TBI, Based on
Admission GCS Score
ED GCS Score
Score Variables e8 9Y13 14Y15
Age 9 55 y 1 1 1
Male 0 1 1
EMS HR G 70 bpm 2 1 1
EMS SBP 9 150 mm Hg 0 1 1
EMS GCS score e 8 2 1 1
ED SBP 9 150 mm Hg 1 0 0
Maximum score 7 5 5
TABLE 3. Incidence of TBI Based on the Proposed Score
in the Study and Validation Cohorts
Severe HI
ED GCS Score Score Study Cohort Validation Cohort
e8 0Y1 0.0% 0.0%
2Y4 63.3% 63.3%
Q5 80.7% 82.0%
9Y13 0Y1 28.9% 31.1%
2Y4 45.3% 45.6%
Q5 100.0% 100.0%
14Y15 0Y1 7.9% 8.0%
2Y4 12.8% 12.9%
Q5 60.0% 50.0%
HI, head injury.
J Trauma Acute Care Surg
Volume 77, Number 4Barmparas et al.
596 * 2014 Lippincott Williams & Wilkins
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
- 6. treatment aimed at sympathetic hyperactivity after TBI may have
important management implications.
Our study is not without its inherent limitations. As this
was a retrospective study of NTDB data, it is only a conve-
nience sample that demonstrates an association between
sympathetic hyperactivity outcome and diagnosis of TBI in-
stead of true cause and effect. A high percentage of patients
were excluded because of unavailability of variables of interest,
including prehospital and admission vital signs, potentially
affecting our results. However, the high number of subjects
included in the data set allows for risk-adjusted analyses. Other
variables of interest that may have been confounders in our
study but were not recorded in the NTDB include ethanol levels
or other central nervous system depressants or stimulants,
pupillary examination, episodes of hypoxemia, or patient co-
morbidities and medications. In addition, because TBI is a dis-
ease with heterogeneity, it would have been useful to have
documentation of the types of bleed(s) patients sustained in
addition to the head AIS score. The transfer time, which was
missing for most patients, was not accounted for. Lastly, blood
pressure is a changing parameter, and having only two time
points (EMS/ED) is intrinsically limiting because of potential
medications that could have been given to alter this parameter
(i.e., antihypertensives, sedatives, or pressors) and therapeutic
interventions aimed at controlling intracranial pressure or cere-
bral edema.
Despite these limitations, we demonstrate that EMS and
ED hypertension in patients presenting with altered mental
status may serve as a marker of TBI and propose a simple
model that correlates highly with incidence of computed to-
mography abnormalities. Our study should further encourage
prospective assessment of EMS and ED hypertension as markers
of TBI in the ED as well as stimulate research examining the
potential of early treatments directed at the catecholamine-
mediated damage.
AUTHORSHIP
M.B. and G.B. contributed to study conception and design. M.B., G.B.,
A.W.L., and A.G. conducted the literature review. M.B., G.B., A.G., and
M.C. acquired the data. M.B., G.B., and D.L. performed analysis and
interpretation of data. M.B., G.B., D.L., and A.W.L. drafted the manu-
script. M.B., E.J.L., and A.S. contributed to critical revision.
DISCLOSURE
The authors declare no conflicts of interest.
REFERENCES
1. Maas AI, Murray GD, Roozenbeek B, Lingsma HF, Butcher I, McHugh
GS, Weir J, Lu J, Steyerberg EW; International Mission on Prognosis
Analysis of Clinical Trials in Traumatic Brain Injury (IMPACT) Study
Group. Advancing care for traumatic brain injury: findings from the
IMPACT studies and perspectives on future research. Lancet Neurol.
2013;12(12):1200Y1210.
2. CDC. TBI in the US reportVtraumatic brain injuryVInjury Center.
Available at: http://www.cdc.gov/traumaticbraininjury/tbi_ed.html. Accessed
March 8, 2014.
3. Stocchetti N, Furlan A, Volta F. Hypoxemia and arterial hypotension at the
accident scene in head injury. J Trauma. 1996;40(5):764Y767. Available at:
http://www.ncbi.nlm.nih.gov/pubmed/8614077. Accessed March 1, 2014.
4. Algattas H, Huang JH. Traumatic brain injury pathophysiology and
treatments: early, intermediate, and late phases post-injury. Int J Mol Sci.
2013;15(1):309Y341.
5. Lu J, Marmarou A, Choi S, Maas A, Murray G, Steyerberg EW. Mortality
from traumatic brain injury. Acta Neurochir Suppl. 2005;95:281Y285.
Available at: http://www.ncbi.nlm.nih.gov/pubmed/16463866. Accessed
March 8, 2014.
6. Woolf PD, McDonald JV, Feliciano DV, Kelly MM, Nichols D, Cox C. The
catecholamine response to multisystem trauma. Arch Surg. 1992;127(8):
899Y903. Available at: http://www.ncbi.nlm.nih.gov/pubmed/1642533.
Accessed March 8, 2014.
7. Frayn KN, Little RA, Maycock PF, Stoner HB. The relationship of plasma
catecholamines to acute metabolic and hormonal responses to injury in
man. Circ Shock. 1985;16(3):229Y240. Available at: http://www.ncbi.nlm.
nih.gov/pubmed/3902270. Accessed March 8, 2014.
8. Kemp CD, Johnson JC, Riordan WP, Cotton BA. How we die: the impact
of nonneurologic organ dysfunction after severe traumatic brain injury.
Am Surg. 2008;74(9):866Y72. Available at: http://www.ncbi.nlm.nih.gov/
pubmed/18807680. Accessed March 8, 2014.
9. Choi HA, Jeon S-B, Samuel S, Allison T, Lee K. Paroxysmal sympa-
thetic hyperactivity after acute brain injury. Curr Neurol Neurosci Rep.
2013;13(8):1Y10.
10. Rabinstein AA. Paroxysmal sympathetic hyperactivity in the neurological
intensive care unit. Neurol Res. 2007;29(7):680Y682.
11. Yang SY, Zhang S, Wang ML. Clinical significance of admission hyper-
glycemia and factors related to it in patients with acute severe head injury.
Surg Neurol. 1995;44(4):373Y377.
12. Hamill RW, Woolf PD, McDonald JV, Lee LA, Kelly M. Catecholamines
predict outcome in traumatic brain injury. Ann Neurol. 1987;21(5):438Y443.
13. Golan JD, Marcoux J, Golan E, Schapiro R, Johnston KM, Maleki M,
Khetarpal S, Jacques L. Traumatic brain injury in intoxicated patients.
J Trauma. 2007;63(2):365Y369.
14. MRC CRASH Trial Collaborators, Perel P, Arango M, Clayton T,
Edwards P, Komolafe E, Poccock S, Roberts I, Shakur H, Steyerberg E,
Yutthakasemsunt S.Predicting outcomeafter traumaticbrain injury:practical
prognostic models based on large cohort of international patients. BMJ.
2008;336(7641):425Y429.
15. Steyerberg EW, Mushkudiani N, Perel P, Butcher I, Lu J, McHugh GS,
Murray GD, Marmarou A, Roberts I, Habbema JD, Maas AI. Predicting
outcome after traumatic brain injury: development and international val-
idation of prognostic scores based on admission characteristics. PLoS Med.
2008;5(8):e165; discussion e165.
16. Hoaglin DC, Iglewicz B, Tukey JW. Performance of some resistant rules
for outlier labeling. J Am Stat Assoc. 1986;81(396):991Y999.
17. Hoaglin DC, Iglewicz B. Fine-tuning some resistant rules for outlier la-
beling. J Am Stat Assoc. 1987;82(400):1147Y1149. doi:10.1080/01621459.
1987.10478551.
18. Sosin DM, Sniezek JE, Waxweiler RJ. Trends in death associated with
traumatic brain injury, 1979 through 1992. Success and failure. JAMA.
1995;273(22):1778Y1780. Available at: http://www.ncbi.nlm.nih.gov/
pubmed/7769773. Accessed March 8, 2014.
19. Langlois JA, Rutland-Brown W, Wald MM. The epidemiology and impact
of traumatic brain injury: a brief overview. J Head Trauma Rehabil. 2006;
21(5):375Y378. Available at: http://www.ncbi.nlm.nih.gov/pubmed/16983222.
Accessed March 1, 2014.
20. Gerber LM, Chiu Y-L, Carney N, Ha¨rtl R, Ghajar J. Marked reduction in
mortality in patients with severe traumatic brain injury. J Neurosurg.
2013;119(6):1583Y1590.
21. Hesdorffer DC, Ghajar J. Marked improvement in adherence to traumatic
brain injury guidelines in United States trauma centers. J Trauma. 2007;
63(4):841Y847; discussion 847Y848.
22. Barmparas G, Singer M, Ley E, Chung R, Malinoski D, Margulies D,
Salim A, Bukur M. Decreased intracranial pressure monitor use at Level II
trauma centers is associated with increased mortality. Am Surg. 2012;
78(10):1166Y1171. Available at: http://www.ncbi.nlm.nih.gov/pubmed/
23025964. Accessed March 8, 2014.
23. Teasdale G, Jennett B. Assessment and prognosis of coma after head injury.
Acta Neurochir (Wien). 1976;34(1Y4):45Y55. Available at: http://www.ncbi.
nlm.nih.gov/pubmed/961490. Accessed March 8, 2014.
J Trauma Acute Care Surg
Volume 77, Number 4 Barmparas et al.
* 2014 Lippincott Williams & Wilkins 597
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
- 7. 24. Morris GF, Marshall LF. Traumatic brain injury. West J Med. 1997;
166(6):415Y416. Available at: http://www.pubmedcentral.nih.gov/
articlerender.fcgi?artid=1304325&tool=pmcentrez&rendertype=abstract.
Accessed March 8, 2014.
25. Sperry JL, Gentilello LM, Minei JP, Diaz-Arrastia RR, Friese RS, Shafi S.
Waiting for the patient to ‘‘sober up’’: effect of alcohol intoxication on
Glasgow Coma Scale score of brain injured patients. J Trauma. 2006;
61(6):1305Y1311.
26. Sloan AE, Arnold SM, St Clair WH, Regine WF. Brain injury: current
management and investigations. Semin Radiat Oncol. 2003;13(3):309Y321.
Available at: http://www.ncbi.nlm.nih.gov/pubmed/12903019. Accessed
March 8, 2014.
27. Shahin H, Gopinath SP, Robertson CS. Influence of alcohol on early
Glasgow Coma Scale in head-injured patients. J Trauma. 2010;69(5):
1176Y1181; discussion 1181.
28. Kornblith LZ, Kutcher ME, Evans AE, Redick BJ, Privette A, Schecter WP,
Cohen MJ. The ‘‘found down’’ patient: a diagnostic dilemma. J Trauma
Acute Care Surg. 2013;74(6):1548Y1552.
29. Rados A, Tiruta C, Xiao Z, Kortbeek JB, Tourigny P, Ball CG, Kirkpatrick
AW. Does trauma team activation associate with the time to CT scan
for those suspected of serious head injuries? World J Emerg Surg. 2013;
8(1):48.
30. Hinson HE, Sheth KN. Manifestations of the hyperadrenergic state after
acute brain injury. Curr Opin Crit Care. 2012;18(2):139Y145.
31. Ley EJ, Clond MA, Singer MB, Shouhed D, Salim A. IL6 deficiency
affects function after traumatic brain injury. J Surg Res. 2011;170(2):
253Y256.
32. Ley EJ, Singer MB, Clond MA, Ley HC, Mirocha J, Bukur M, Margulies
DR, Salim A. Admission heart rate is a predictor of mortality. J Trauma
Acute Care Surg. 2012;72(4):943Y947.
33. Sellmann T, Miersch D, Kienbaum P, Flohe´ S, Schneppendahl J, Lefering
R; DGU Trauma Registry. The impact of arterial hypertension on poly-
trauma and traumatic brain injury. Dtsch Arztebl Int. 2012;109(49):
849Y856.
34. Arbabi S, Campion EM, Hemmila MR, Barker M, Dimo M, Ahrns KS,
Niederbichler AD, Ipaktchi K, Wahl WL. Beta-blocker use is associated
with improved outcomes in adult trauma patients. J Trauma Inj Infect Crit
Care. 2007;62(1):56Y62.
35. Alali AS, McCredie VA, Golan E, Shah PS, Nathens AB. Beta blockers
for acute traumatic brain injury: a systematic review and meta-analysis.
Neurocrit Care. 2014;20:514Y23.
36. Ley EJ, Clond MA, Bukur M, Park R, Chervonski M, Dagliyan G,
Margulies DR, Lyden PD, Conti PS, Salim A. A-adrenergic receptor inhi-
bition affects cerebral glucose metabolism, motor performance, and in-
flammatory response after traumatic brain injury. J Trauma Acute Care
Surg. 2012;73(1):33Y40.
37. Ley EJ, Scehnet J, Park R, Schroff S, Dagliyan G, Conti PS, Margulies DR,
Salim A. The in vivo effect of propranolol on cerebral perfusion and
hypoxia after traumatic brain injury. J Trauma. 2009;66(1):154Y159;
discussion 159Y161.
38. Ley EJ, Park R, Dagliyan G, Palestrant D, Miller CM, Conti PS, Margulies
DR, Salim A. In vivo effect of propranolol dose and timing on cerebral
perfusion after traumatic brain injury. J Trauma. 2010;68(2):353Y356.
39. Patel MB, McKenna JW, Alvarez JM, Sugiura A, Jenkins JM,
Guillamondegui OD, Pandharipande PP. Decreasing adrenergic or sym-
pathetic hyperactivity after severe traumatic brain injury using propranolol
and clonidine (DASH After TBI Study): study protocol for a randomized
controlled trial. Trials. 2012;13:177.
J Trauma Acute Care Surg
Volume 77, Number 4Barmparas et al.
598 * 2014 Lippincott Williams & Wilkins
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.