Full text of DOI:10.1111/j.1553-2712.2001.tb01145.x

ACADEMIC EMERGENCY MEDICINE • September 2001, Volume 8, Number 9
859
BASIC INVESTIGATIONS
Utility of an Initial D-dimer Assay in Screening for
Traumatic or Spontaneous Intracranial Hemorrhage
MARK E. HOFFMANN, MD, O. JOHN MA, MD, GARY GADDIS, PHD, MD
Abstract. Objective: To evaluate the sensitivity of likely to have a positive D-dimer assay (chi-square =
a D-dimer assay as a screening tool for possible trau- 13.075, p < 0.001). The D-dimer assay had 21 true-
matic or spontaneous intracranial hemorrhage. If ad- positive and four false-negative tests, resulting in a
equately sensitive, the D-dimer assay may potentially sensitivity of 84.0% (95% CI = 63.7% to 95.5%) and a
permit omission of a more expensive computed to- specificity of 55.8% (95% CI = 55.5% to 55.9%). The
mography (CT) scan of the head when such hemor- four false-negative cases included one small intrapa-
rhage is clinically suspected. Methods: Prospective, renchymal hemorrhage, one small subarachnoid
consecutive, blinded study of patients (age > 16 years) hemorrhage, one moderate-sized intraparenchymal
requiring a CT scan of the head for suspected intra- hemorrhage with mid-line shift, and one large sub-
cranial hemorrhage over a five-month period at a uni- dural hematoma requiring emergent surgery. Con-
versity, Level I trauma center. All study patients had clusions: Due to the catastrophic nature of missing
a serum D-dimer assay obtained prior to their CT an intracranial hemorrhage in the emergency de-
scans. Sensitivity and specificity, with 95% confidence partment, the D-dimer assay is not adequately sen-
intervals (95% CIs), of the enzyme-linked immunosor- sitive or predictive to use as a screening tool to allow
bent assay (ELISA) D-dimer assay for the detection routine omission of head CT scanning. Key words:
of intracranial hemorrhage were calculated. Results: D-dimer assay; intracranial hemorrhage; computed
Of the 319 patients entered in the study, 25 (7.8%) tomography; head injury. ACADEMIC EMERGENCY
had a CT scan positive for intracranial hemorrhage. MEDICINE 2001; 8:859–865
Patients with intracranial hemorrhage were more
EAD INJURY with intracranial hemorrhage hemorrhage. If an effective, sensitive, reliable, and
H
is a factor in 50% of all trauma-related fa-1 less costly screening test for intracranial hemor-
talities and 75% of motor vehicle crash fatalities. rhage could be identified, the number of CT scans
Up to 20% of all cerebral vascular accidents are of the head that are negative for intracranial hem-
2
hemorrhagic. Suspected intracranial hemorrhage orrhage might be reduced.
is a common reason for emergency physicians
EPs) to obtain a computed tomography (CT) scan
A D-dimer assay represents a theoretically at-
tractive candidate for such a screening test. The
brain contains high concentrations of tissue factors
that, if released into the circulation due to disrup-
tion of the blood–brain barrier, trigger the activa-
tion of the extrinsic coagulation pathway. An ac-
cepted hypothesis states that the magnitude of
intravascular coagulation is proportional to the
(
of the head since it is medically unacceptable to
fail to diagnose these conditions. Radiation dosage,
patient throughput time in the emergency depart-
ment (ED), utilization of radiologic technicians,
and financial and other costs accrue from liberal
use of CT scanning. However, physician and pa-
tient expectations, as well as medicolegal concerns,
dictate that EPs actively screen for intracranial
3
,4
amount of brain tissue disrupted following brain
3,5–9
injury.
The prognosis for severe head injury has
been based on the results of CT scanning, assays
for coagulopathy, and the neurological examina-
tion. More recent studies have demonstrated the
D-dimer assay and other serum markers of coag-
ulopathy to be sensitive and reliable indicators for
From the Department of Emergency Medicine, Truman Medi-
cal Center (MEH, OJM), and the Department of Emergency
Medicine, St. Luke’s Hospital (GG), University of Missouri–
Kansas City School of Medicine, Kansas City, MO.
1
0–12
Received February 15, 2001; revision received May 16, 2001; head injury prognosis.
Enzyme-linked immu-
accepted May 21, 2001. Presented at the SAEM annual meet-
ing, San Francisco, CA, May 2000.
Address for correspondence and reprints: Mark E. Hoffmann,
MD, 7092 SE 43rd Avenue, St. Cloud, MN 56304. e-mail:
crazydiamond@aol.com
nosorbent assay (ELISA) assays are known to be
13
highly sensitive for the presence of D-dimer. Our
laboratory can complete D-dimer assays within 30
minutes of receipt of the sample.

860
D-DIMER ASSAY
Hoffmann et al. • UTILITY OF INITIAL D-DIMER ASSAY
T
ABLE 1. Inclusion Criteria
year emergency medicine resident, derived from
history, physical examination, and other available
data. The setting was an ED of a university Level
I trauma center, with an annual volume of 53,000
patients.
Injury or symptom onset < 24 hours
Loss of consciousness
Suspected skull fracture
Glasgow Coma Scale score < 13
Focal neurological deficit
Age > 60 years with mental status changes
Suspected subarachnoid hemorrhage
New or post-traumatic seizure
Worsening mental status
Study Protocol. Upon decision to screen for in-
tracranial hemorrhage by CT scan of the head,
blood was obtained from intravenous line place-
ment, then sent in a 3.2% sodium citrate (‘‘blue
top’’) test tube to our hospital’s laboratory for
prompt centrifugation and cooling to negative
TABLE 2. Exclusion Criteria
Age < 16 years
20ЊC. The D-dimer assay kits, donated by Instru-
Injury or symptoms > 24 hours
Known pre-existing deep venous thrombosis, pulmonary em-
bolism, cancer, or cirrhosis
Major surgery < 2 weeks
Penetrating eye injury
mentation Laboratory Company (Lexington, MA),
were an automated latex-enhanced immunoassay
for the quantitative determination of D-dimer in
human citrated plasma. D-dimer assay results
were reported in ng/mL; the upper limit of normal
as stated by our laboratory and the package insert
was 284 ng/mL and 278 ng/mL, respectively.
D-dimer assays were performed biweekly, with re-
sults recorded on data sheets separate from CT re-
sults. This ensured that both investigators and cli-
nicians were blinded to the results. The assay
package insert states that for up to eight weeks,
the D-dimer assay remains accurate when cooled
to a temperature of Ϫ20ЊC.
All CT scans were initially interpreted by EPs
and radiology residents, and then formally over-
read by staff radiologists within 24 hours. Patients
with intracranial hemorrhage were admitted to ei-
ther the neurosurgery or the internal medicine ser-
vice.
Pregnancy
Multisystem trauma to chest, abdomen, or extremities
Although an association between brain injury
and coagulopathy has been known for two decades,
to the best of our knowledge, no study has evalu-
ated the sensitivity of the D-dimer assay as a po-
tential screening tool for patients with suspected
intracranial hemorrhage. We hypothesized the
D-dimer assay would have a sufficiently high sen-
sitivity and negative predictive value for intracra-
nial hemorrhage to permit the D-dimer assay to
serve as an effective screen for intracranial hem-
orrhage. This implies that to safely reduce the
number of CT scans of the head, the sensitivity
and negative predictive value of a D-dimer assay
for intracranial hemorrhage must approach 100%.
The objective of this study was to prospectively as-
sess whether an initial D-dimer assay can safely
screen patients for clinically suspected traumatic
or spontaneous intracranial hemorrhage by achiev-
ing a sensitivity and negative predictive value of
A single investigator reviewed patient medical
records twice weekly in order to assess accuracy of
enrollment, CT scan results, patient disposition,
and collection of relevant data. The data collection
sheets recorded patient age and gender, indica-
tions for CT scan, time since injury (0–3 hours, 3–
1
2 hours, or 12–24 hours), time blood was drawn,
100%.
and associated illness or injury.
M
ETHODS
Data Analysis. Sample size estimation was
done. The critical point of this study, however, is
the sensitivity of the D-dimer assay, not the pres-
ence of a sample size adequate to find a significant
difference between populations with and without a
positive D-dimer assay.
Study Design. This was a prospective clinical ob-
servational study of patients suspected to have in-
tracranial hemorrhage. The study was approved by
the institution’s investigational review board.
Nonetheless, we knew from a retrospective re-
Study Setting and Population. This prospective view at our institution that of 1,110 patients with
study consecutively enrolled adult patients (age Ն a CT scan obtained to screen for intracranial hem-
1
1
6 years) between June 1, 1999, and November 30, orrhage, 46 scans were positive for intracranial
999. Specific inclusion and exclusion criteria are hemorrhage. Of these 46, 42 were known to have
detailed in Tables 1 and 2. Age of suspected intra- a positive D-dimer assay. D-dimer assays for the
cranial hemorrhage, based on history, was strati- other four patients were unavailable. Also, the
fied as 0–3 hours, 3–12 hours, or 12–24 hours. prevalence of positive and negative D-dimer assays
Intracranial hemorrhage suspicion was based on among patients with CT scans negative for intra-
the clinical opinion of the attending staff or third- cranial hemorrhage was unknown. These facts

ACADEMIC EMERGENCY MEDICINE • September 2001, Volume 8, Number 9
861
T
ABLE 4. Types of Intracranial Hemorrhage on Computed
precluded use of a retrospective survey to assess
D-dimer assay as a screen for intracranial hemor-
Tomography Scan of the Head
rhage. Some false-positive cases (positive D-dimer Subdural hematoma
7
7
6
5
assay with negative intracranial hemorrhage) Subarachnoid hemorrhage
Intraparenchymal hemorrhage
Intracerebral contusion
would be expected in this series due to the pres-
ence of concomitant injuries. We utilized a pro-
jected 50% rate for patients with CT scans nega-
tive for intracranial hemorrhage having elevated
T
ABLE 5. Sensitivity, Specificity, Positive Predictive Value,
D-dimer levels to estimate required sample size.
Although this rate was higher than expected, it
and Negative Predictive Value for the Initial D-dimer Assay
in Screening for Intracranial Hemorrhage
1
4
was chosen to ensure adequate power. With these
assumptions, a sample size of 266 patients or more
would achieve a power of at least 0.8.
CT Positive CT Negative Total
Positive D-dimer assay
21
4
130
164
151
168
Chi-square testing was done to evaluate for any Negative D-dimer assay
significant difference of proportion of positive
D-dimer levels between patients whose CT scans
TOTAL
25
294
319
were positive and negative. Sensitivity, specificity, CT = computed tomography. Sensitivity: 84.0% (95% CI =
negative predictive value, and positive predictive 63.7% to 95.5%); specificity: 55.8% (95% CI = 50.1% to 61.5%);
positive predictive value: 13.9% (95% CI = 8.8% to 19.7%); neg-
ative predictive value: 97.6% (95% CI = 94.1% to 99.3%).
value, along with their 95% confidence intervals
(95% CIs), were calculated. Because the ‘‘cutoff’’
hemorrhage, 21 had positive D-dimer assays (true
positives) and four had normal assays (false neg-
atives). The calculated test sensitivity was 84%
and the confidence interval for sensitivity did not
overlap 100%.
Table 5 demonstrates the sensitivity, specificity,
negative predictive value, and positive predictive
tracranial hemorrhage.
cranial hemorrhage. Patients with intracranial
hemorrhage were more likely to have elevated
D-dimer levels (chi-square = 13.075, df = 1, p <
RESULTS
0.001).
Overall, 455 patients were identified as meeting
Figure 1 demonstrates a ROC curve for the
criteria for inclusion in the study; 319 patients D-dimer assay, in recognition that our laboratory’s
were enrolled. Table 3 notes the characteristics of ‘‘cutoff’’ value for ‘‘normal’’ is somewhat arbitrary.
the enrolled patients.
A ‘‘cutoff’’ of 131 ng/mL would have permitted a
Twenty-five (7.8%) of the 319 study patients 100% sensitivity, but would have markedly in-
had intracranial hemorrhage identified on CT scan creased the number of false-positive test results.
(Table 4). Of these 25 patients with intracranial The failure of the ROC curve to approach the up-
per left corner of the graph shows that alternative
T
ABLE 3. Characteristics of the Subjects
‘‘cutoff’’ values for a positive test would not permit
D-dimer assay to be highly predictive of hemor-
rhage.
Gender
Male
Female
204/319
115/319
The time interval between injury and obtaining
blood from the four false-negative patients ranged
between three hours and 24 hours. The four false-
negative patients had the following outcomes: pa-
tient 1 was a 26-year-old woman who had a spon-
taneous, atraumatic intracerebral hemorrhage
that required no surgical intervention. During the
patient’s hospitalization, the D-dimer assay was
initially 131 ng/mL and remained negative (161
ng/mL) on repeat testing 12 hours after presenta-
tion. Patient 2 was a 19-year-old man who com-
plained of a headache and loss of consciousness
after a fall and presented within three hours
Age
41.14 Ϯ 17.4 (16–99)
Type of injury
Nontrauma
Trauma
131/319
188/319
Mechanisms of traumatic injuries (in
descending order of prevalence)
Motor vehicle crash
Blunt object to head
Fall from a height
Continuous variables expressed as mean Ϯ SD (range). Cate-
gorical variables expressed as proportions.

862
D-DIMER ASSAY
Hoffmann et al. • UTILITY OF INITIAL D-DIMER ASSAY
Figure 1. Receiver operating characteristic (ROC) curve for D-dimer assay data as a predictor of the presence or
absence of intracranial hemorrhage. CT = computed tomography.
of symptom onset. He was diagnosed as having ter initial presentation. This patient had a right
a small subarachnoid hemorrhage. The patient hemiparesis and was discharged to a long-term
required no surgical intervention. The initial care facility for rehabilitation. Patient 4 was a 41-
D-dimer level was 210 ng/mL. This patient was year-old man who fell, experienced a seizure, and
discharged home without a repeat D-dimer assay presented less than three hours after symptom on-
having been obtained during the hospital course. set. He was diagnosed as having a large subdural
Patient 3 was a 40-year-old man who presented hematoma that necessitated an emergent crani-
less than three hours after developing altered otomy. The initial D-dimer level was 227 ng/mL.
mental status and was diagnosed as having a spon- This patient had a positive D-dimer assay (788
taneous, atraumatic, moderate-sized intraparen- ng/mL) drawn the morning following emergency
chymal hemorrhage with mid-line shift. The ini- craniotomy. The patient was discharged five
tially normal D-dimer level of 258 ng/mL turned days postoperatively without any neurological se-
positive (416 ng/mL) on repeat testing 12 hours af- quelae.

ACADEMIC EMERGENCY MEDICINE • September 2001, Volume 8, Number 9
863
Some investigators have compared the coagu-
lopathies between patients with isolated head in-
jury and trauma patients without head injury.
Nanzaki and Kemmotsu reported the coagulofibri-
nolytic changes after isolated head injury were not
different from those in trauma patients without
head injury. This was a small study consisting of
five patients with isolated head injury and 11
D
ISCUSSION
In this study, the sensitivity of the D-dimer assay
as a screen for suspected intracranial hemorrhage
was 84% (95% CI = 63.7% to 95.5%). This sensitiv-
ity is inadequate for us to advocate the use of a
D-dimer assay as a screening tool for intracranial
hemorrhage even though the patients with intra-
cranial hemorrhage were statistically more likely
to have an elevated D-dimer level.
The D-dimer, a fragment of degradation prod-
ucts of cross-linked fibrin, is direct evidence of fi-
brin formation and its dissolution by plasmin. Cur-
rently, the D-dimer assay is the most sensitive test
1
7
trauma patients without head injury. Another
study found the D-dimer assay to be overly sensi-
tive in predicting head injury outcome in trauma
patients. This study also compared the coagulop-
athies between trauma patients and elective neu-
rosurgical patients. The D-dimer levels of the neu-
rosurgical patients were normal or near normal.
The authors proposed an additional factor, such
as catecholamine release, might be involved in
trauma-related brain injury by amplifying the pro-
1
2
for identifying this process. ELISA kits of various
manufacturers are known to be more sensitive
than semiquantitative latex agglutination stud-
1
3
ies. Assay systems for D-dimer antigen include
manual immunoagglutination assays, immunofil-
tration assays, microtiter plate ELISA assays, au-
tomated ELISA systems, and latex-enhanced pho-
tometric immunoassays. There are at least 13
different assays commercially available. According
to clinical studies, the characteristics of the differ-
ent reagents and methods that are used in various
laboratories yield a large variation in sensitivi-
8
cess of disseminated intravascular coagulopathy.
Recent studies in head injury have evaluated
the protein S-100. This protein is contained within
the astroglial cells of the central nervous system.
Injury to these cells releases S-100 into the circu-
lation and can be quantified serologically. Most
studies have found this protein to be a reliable in-
1
8–23
dicator of head injury prognosis.
One of the
1
5
studies found the S-100 protein had a potential
role as a screening test. The investigators calcu-
lated the negative predictive value of the S-100
protein to be 99% and an undetectable serum
S-100 level predicted normal intracranial findings
on CT scan. They concluded the S-100 marker
ties.
Previous studies have evaluated fibrinolytic pa-
rameters in the prognosis of head injury.
9
–11
A
strong association between coagulopathy at admis-
sion and poor neurological outcome was confirmed
by data obtained from the National Traumatic
Coma Data Bank. One study concluded that the
fibrinolytic parameters found on admission could
be reliable indicators of head injury outcome.
When the D-dimer assay was elevated two- to
three-fold, 92% of patients died regardless of the
level of consciousness at admission. These pa-
tients were evaluated within two hours of their in-
juries; no data were available for evaluating the
D-dimer assay as a screening tool for intracranial
hemorrhage.
24
might be used to select patients for CT scanning.
The S-100 serum marker, however, is unavailable
in most hospital laboratories.
Increase in blood–brain barrier permeability
after subarachnoid hemorrhage may be a time-de-
pendent event. Few studies have examined the re-
lationship between subarachnoid hemorrhage and
blood–brain barrier permeability. There have been
conflicting reports due to limitations in experi-
mental probes adopted and in timing of observa-
1
6
It is possible that the D-dimer assay failed to tion. One study by Germano et al. found significant
adequately screen intracranial hemorrhage in our permeability changes beginning at 36 hours after
25
study because inadequate time may have elapsed subarachnoid hemorrhage ; a study by Troja-
between hemorrhage and the crossing of the dam- nowski found extravasation and tissue staining
26
aged blood–brain barrier by fibrin degradation only four hours after bleeding. Johshita et al.
products into the circulation. Three of the four pa- suggested that barrier disturbances associated
tients with false-negative results in our study pre- with subarachnoid hemorrhage may be multifac-
27
sented within three hours after symptom onset or torial in time course and location.
injury. Currently, to the best of our knowledge,
there are no studies that investigate the time of du-
ration required from initial brain injury to the time
LIMITATIONS AND
F
UTURE
Q
UESTIONS
when the presence of fibrin degradation products in There are several limitations to this study. First,
the systemic circulation can be adequately mea- specific, widely accepted criteria do not exist re-
sured. Another possibility is that not all patients garding which patient requires a CT scan of the
with intracranial hemorrhage and disruption of the head in the ED for evaluation of intracranial hem-
blood–brain barrier develop coagulopathy.
orrhage. Second, the D-dimer assay is not a stan-

864
D-DIMER ASSAY
Hoffmann et al. • UTILITY OF INITIAL D-DIMER ASSAY
dardized test between various facilities. We eval- References
uated just one of several types of ELISA D-dimer
assays. Various D-dimer assays, however, have
similar sensitivities and specificities for coagulop-
1
. Kalsbeek WD, McLaurin RL, Harris BS III, et al. The Na-
tional Head and Spinal Cord Injury Survey: major findings. J
Neurosurg. 1980; 53:19–31.
athy when evaluating for deep venous thrombo- 2. Brott TG, Thalinger K, Hertzberg V. Hypertension as a risk
13
for spontaneous intracerebral hemorrhage. Stroke. 1986; 17:
sis. Third, 136 (30%) potential study patients
were excluded due to inadequate data collection
7
3
8–83.
. Hulka F, Mullins R, Frank E. Blunt brain injury activates
sheet completion by the clinician (18), misplaced the coagulation process. Arch Surg. 1996; 131:923–8.
D-dimer assay or inappropriately stored specimen
by ancillary services (76), or hemolyzed specimen
4
. Bredbacka S, Edner G. Soluble fibrin and D-dimer as de-
tectors of hypercoagulability in patients with isolated brain
trauma. J Neurosurg Anesthesiol. 1994; 6(2):75–82.
(
42). Fourth, not all patients admitted with the di- 5. Scherer RU, Spangenberg P. Procoagulant activity in pa-
tients with isolated severe head trauma. Crit Care Med. 1998;
agnosis of intracranial hemorrhage had repeat D-
dimer assays by the admitting services. Fifth, pa-
2
6
6(1):149–56.
. Becker S, Schneider W, Kreuz W, Jacobi G, Scharrer I,
tients who were discharged from the ED with Nowak-Gottl U. Post-trauma coagulation and fibrinolysis in
normal CT scans and false-positive D-dimer assays
did not have sufficient follow-up with repeat CT
scanning to fully assess for delayed injury. Sixth, significance of hemostatic abnormalities in patients with head
children suffering from severe cerebro-cranial trauma. Eur J
Pediatr. 1999; (158 suppl 3):S197–S202.
7. Olson JD, Kaufman HH, Moake J, et al. The incidence and
injuries. Neurosurgery. 1989; 24:825–32.
the D-dimer assay was an ELISA assay that,
8
. Kearney TJ, Bentt L, Grode M, Lee S, Hiatt JR, Shabot
though more sensitive than the semiquantitative
latex agglutination tests, does take longer to per- J Trauma. 1992; 32:608–12.
MM. Coagulopathy and catecholamines in severe head injury.
1
3
9
. Stein SC, Young GS, Talucci RC, Greenbaum BH, Ross SE.
form than a latex agglutination study. Finally, no
validated criteria of which we are aware assess a
Delayed brain injury after head trauma: significance of coag-
ulopathy. Neurosurgery. 1992; 30:160–5.
pretest probability of intracranial hemorrhage in 10. Raabe A, Grolms C, Keller M, Dohnert J, Sorge O, Seifert
patients suspected to have intracranial hemor- V. Correlation of CT findings and serum brain damage markers
rhage. We did not attempt to create such criteria
in this study. Patients were eligible for study entry
if the patient’s physician had clinical suspicion of ison of clinical, radiologic, and serum marker as prognostic fac-
an intracranial hemorrhage as noted by the inclu-
sion criteria. The factors that led to the decision to
perform CT scanning were not quantified.
Future studies should address the time-depen-
dent issue of D-dimer assays with the onset of se-
following severe head injury. Acta Neurochir. 1998; 140:
7
1
87–91.
1. Woertgen C, Rothoerl RD, Metz C, Brawanski A. Compar-
tors after severe head injury. J Trauma. 1999; 47:1126–30.
1
2. Bick RL, Baker WF. Diagnostic efficacy of the D-dimer as-
say in disseminated intra-vascular coagulation (DIC). Thromb
Res. 1992; 15:785–90.
1
3. van der Graaf F, van den Borne H, van der Kolk M, et al.
Exclusion of deep venous thrombosis with D-dimer testing:
comparison of 13 D-dimer methods in 99 outpatients suspected
vere headache or traumatic event. Also, future in- of deep venous thrombosis using venography as a reference
standard. Thromb Haemost. 2000; 83:191–8.
vestigations should prospectively review patient
outcomes after head injury with the results of se-
14. Glantz SA. Primer of Biostatistics: The Program. New
York: McGraw-Hill, 1992.
rial D-dimer assay testing. In addition, develop- 15. DeMaat MP, Meijer P, Nieuwenhuizen W, et al. Per-
formance of semiquantitative and quantitative D-dimer assays
ment of a scale utilizing history and physical ex-
amination data to assess pretest probability of
in the ECAT external quality assessment program. Semin
Thromb Hemost. 2000; 26:625–30.
hemorrhage would provide a focus for subsequent 16. Takahashi H, Urano T, Takada Y, Nagai N, Takada A. Fi-
brinolytic parameters as an admission prognostic marker of
study.
head injury in patients who talk and deteriorate. J Neurosurg.
1
997; 86:768–72.
1
7. Nanzaki S, Kemmotsu O. Coagulofibrinolytic changes af-
CONCLUSIONS
ter isolated head injury are not different from those in trauma
patients without head injury. J Trauma. 1999; 46:1070–6.
Patients with intracranial hemorrhage are more 18. Rothoerl RD, Woertgen C, Holzachuh M, Metz C, Brawan-
ski A. S-100 serum levels after minor and major head injury.
likely than not to have an elevated D-dimer assay.
Due to the catastrophic nature of missing an in-
J Trauma. 1998; 45:765–7.
19. Raabe A, Grolms C, Seifert V. Serum markers of brain
tracranial hemorrhage, however, the D-dimer as- damage and outcome prediction in patients after severe head
injury. Br J Neurosurg. 1999; 13(1):56–9.
say is not adequately sensitive or predictive to use
as a screening tool. The D-dimer assay cannot be
20. Ingebrigtsen T, Waterloo K, Jacobsen EA, Lanbakk B,
Romner B. Traumatic brain damage in minor head injury: re-
relied upon as a substitute for an emergent CT lation of serum S-100 protein measurements to MRI and neu-
robehavioral outcome. Neurosurgery. 1999; 45:468–76.
scan of the head when intracranial hemorrhage is
clinically suspected.
2
1. Raabe A, Grolms C, Sorge O, Zimmermann M, Seifert V.
Serum S-100B protein in severe head injury. Neurosurgery.
1
999; 45:477–83.
2
2. Woertgen C, Rothoerl RD, Metz C, Brawanski A. Compar-
ison of clinical, radiologic, and serum marker as prognostic fac-
The authors thank Jamie Patel, MD, for her assistance during
the data collection phase of this study. They also thank Instru-
tors after severe head injury. J Trauma. 1999; 47:1126–30.
2
3. Hermann M, Jost S, Kutz S, et al. Temporal profile of re-
mentation Laboratory Company (Lexington, MA) for donating lease of neurobiochemical markers of brain damage after trau-
the D-dimer assay kits at no expense to the patient. matic brain injury is associated with intracranial pathology as

ACADEMIC EMERGENCY MEDICINE • September 2001, Volume 8, Number 9
865
demonstrated in cranial computerized tomography. J Neuro- subarachnoid hemorrhage. Acta Neurochir. 2000; 142:575–80.
trauma. 2000; 17:113–22. 26. Trojanowski T. Blood–brain barrier changes after experi-
2
4. Romner B, Ingesbrigtsen T, Kongstad P, Borgesen SE. mental subarachnoid hemorrhage. Acta Neurochir. 1982; 60:
Traumatic brain damage: S-100 protein measurements related 45–54.
to neuroradiological findings. J Neurotrauma. 2000; 17:641–7. 27. Johshita H, Kassell NF, Sasaki T, et al. Blood–brain bar-
2
5. Germano A, d’Avella D, Imperatore C, et al. Time-course rier disturbance following subarachnoid hemorrhage in rab-
of blood brain barrier permeability changes after experimental bits. Stroke. 1990; 21:1051–8.
ⅷ