The European Brain Injury Consortium Survey of Head Injuries
G. D. Murray1, G. M. Teasdale2, R. Braakman3, F. Cohadon4, M. Dearden5, F. Iannotti6, A. Karimi7,
F. Lapierre8, A. Maas9, J. Ohman10, L. Persson11, F. Servadei12, N. Stocchetti13, T. Trojanowski14,
and A. Unterberg15 on behalf of the European Brain Injury Consortium
1 Medical Statistics Unit, University of Edinburgh Medical School, Edinburgh, U.K.
2 University Department of Neurosurgery, Institute of Neurological Sciences, Southern General Hospital, Glasgow, U.K.
3 Berkel Enschot, The Netherlands
4 Department of Neurosurgery, Hospital Pellegrin, Bordeaux, France
5 Department of Anaesthetics, Leeds General In®rmary, Leeds, U.K.
6 Department of Clinical Neurosciences, Southampton General Hospital, Southampton, U.K.
7 Neurochirurgische Klinik, Universita
È
tKo
È
ln, Ko
È
ln, Federal Republic of Germany
8 Centre Hospitalier Universitaire de Poitiers, Service de Neurochirurgie, Poitiers, France
9 Department of Neurosurgery, University Hospital Rotterdam, Rotterdam, The Netherlands
10 Department of Neurosurgery, Helsinki University Central Hospital, Helsinki, Finland
11 Department of Neurosurgery, Akademiska Hospital, Uppsala, Sweden
12 Divisione di Neurochirurgica, Ospedale M. Bufalini, Cesena, Italy
13 Terapia Intensiva Neurochirurgica Servizio Anestesia e Rianimazione, Ospedale Maggiore Policlinico, IRCCS Milano, Milano, Italy
14 Department of Neurosurgery, University Medical School, Lublin, Poland
15 Department of Neurosurgery, Virchow-Klinikum, Medizinische Fakulta
È
t, Humboldt Universita
È
t Berlin, Berlin, Federal Republic of
Germany
Summary
To provide a picture of contemporary practice, a survey was car-
ried out of severely and moderately head injured patients admitted to
67 `neuro' centres in 12 European countries. 1,005 adult head injuries
were recruited over a three month period. Sixty items of information
on demography, clinical features, investigations, management and
early complications were captured on a simple, two-page question-
naire and, information on outcome at six months on a third page.
The median age of the subjects was 38 years, 74% were male and
51% injured in road tra½c accidents; 57% of patients were trans-
ferred to the `neuro' centre from another hospital. Assessment of
clinical responsiveness was limited by the use of sedation and in-
tubation and information from four early time points (pre-hospital,
arrival at the Accident and Emergency department, post-resuscita-
tion, and arrival at the `neuro' unit) was combined to stratify the
subjects as severe (58%), moderate (17%) or intermediate (19%). In
48% of patients classi®ed the CT scan showed features of a `mass le-
sion' and in 40% showed a subarachnoid haemorrhage. Fifty-®ve
centres provided the data on outcome for 94% of the cases recruited
in these centres six months after injury. 31% died, 3% were vegeta-
tive, 16% severely disabled, 20% moderately disabled and 31% had
made a good recovery. Comparison of the data from di¨erent parts
of Europe showed di¨erences in the frequency of secondary transfer,
cause of injury, occurrence of major extracranial injury, CT scan
®ndings, intracranial operation, clinical severity of injury and uti-
lisation of the components of intensive care and the occurrence of a
favourable outcome, although the latter di¨erence was not statisti-
cally signi®cant when variations in the initial severity of injury were
taken into account.
The ®ndings in the present survey are compared with newly ana-
lysed information for three previous large series: the International
Data Bank involving the UK, the Netherlands and the USA, the
North American Traumatic Coma Data Bank, and data from four
centres in the UK. The comparisons showed substantial similarities
and also di¨erences that may re¯ect variations in policy for admis-
sion of the head injury to `neuro' units, and evolution in methods of
assessment, investigation and management. The e¨ects of these dif-
ferences on outcome requires further, rigorous prospective study.
Keywords: Head injury; European survey; management; outcome.
Introduction
The European Brain Injury Consortium (EBIC) is a
network of European units, experienced in the care of
head injured patients, and was formally constituted in
1995 [38]. The Consortium promotes international,
multicentre, interdisciplinary research aimed at im-
proving the outcome of patients who have su¨ered
a head injury or other kind of acute brain damage.
During the formal establishment of EBIC, it was de-
Acta Neurochirurgica
> Springer-Verlag 1999
Printed in Austria
Acta Neurochir (Wien) (1999) 141: 223±236
cided to conduct a survey of head injured patients in
the interested centres.
The survey had three purposes. First, the exercise
would test if it was possible to collect comprehensive,
credible data through an organisation with strong
commitment but only modest resources. Second, the
results would be of considerable intrinsic interest and
importance: existing comprehensive databases on
severe head injury are over a decade old [9, 15, 23, 28] ±
and more recent reports concern only selected pop-
ulations entered into clinical trials. The survey there-
fore would provide a unique picture of contemporary
practice in di¨erent parts of Europe and how the ®nd-
ings related to previous data. Third, the results would
be invaluable for conducting `what if ?' evaluations of
potential inclusion/exclusion criteria for formal clini-
cal studies and trials, for example, the proportion of
severe head injuries who are admitted to a neuro-
surgical unit within di¨erent times of injury, or with
di¨erent clinical states and how they are currently
managed, and how these factors in¯uence the outcome
expected with `conventional' treatment.
In this paper we describe the features of the 1005
adult patients, considered to have a severe or moderate
head injury, reported to the European Core Data
Bank, and compare the ®ndings in di¨erent groups of
subjects and in di¨erent parts of Europe. The results of
the present series are compared to previous reports of
multicentre series collected prospectively in routine
clinical practice. Problems, identi®ed in the current
series in de®ning clinical severity of the injury, are ad-
dressed in relation to previous experience. The ®ndings
in the more selected populations customarily recruited
to trials of pharmacological agents are considered in a
separate paper [20].
Methods
A two page questionnaire was designed to capture 60 items
of information on demography, clinical features, investigations,
management, complications and early outcome. The ®rst page
covered the ®rst day following the injury, and included age; sex;
cause of injury; mode of admission to the neurosurgical hospital
(direct or transfer); timing of injury, admission to ®rst hospital and
admission to the neurosurgical hospital; details of any extracranial
injuries; clinical evidence of severity of injury was assessed by the
Glasgow Coma Scale (GCS) [39] and pupil response to light. Data
were recorded at four stages: 1. pre-hospital (ie the ®rst reliable ob-
servation made by a `paramedic' or medical sta¨ ); 2. arrival at the
Accident and Emergency Department of the hospital where the
patient was ®rst taken; 3. post resuscitation (ie the state after initial
resuscitation); 4. neuro unit (ie the point at which the patient comes
under specialist, usually neurosurgical, care but also neurointensive
and neurological). Features such as early complications (hypoxia,
hypotension or hypothermia); the results of an admission CT scan;
details of any intracranial operation within the ®rst 24 hours and of
any emergency extracranial operations were also covered. The sec-
ond page covered the hospital care up to discharge from the neuro-
surgical hospital and included details of management and monitor-
ing (intracranial pressure monitoring, ventilation, jugular SVO
2
monitoring, invasive blood pressure monitoring); the results of a
®nal CT scan; details of any intracranial complications which
required treatment (delayed haematoma, raised ICP, meningitis/
ventriculitis, seizures); details of any life threatening systemic com-
plications (respiratory, cardiovascular, metabolic, infection); timing
and mode of discharge; and cause of death where applicable.
Patient Inclusion
The data collection exercise ran from 1st of February 1995 to 30th
of April 1995. All centres that had, at that time, expressed interest in
EBIC were invited to participate and were asked to return details of
all moderate or severe adult (> 16 years) head injuries admitted to
their care within 24 hours of injury. Patients were to be included if
their Glasgow Coma Score [40] was 12 or less at any of the four
stages described above, this corresponds to previous de®nitions of
severe (GCS 3 to 8 [9, 23]) and moderate (GCS 9±12 [34]) head
injuries.
Data relating to the ®rst 24 hours following injury were to be re-
turned to the EBIC Co-ordination Centre by mail or fax within one
week of admission, and the discharge forms were sent in batches at
the end of each month. After the collection of the initial data was
completed, centres were contacted to ask for details of each patient's
outcome six months following their injury. For this, a third, one page
questionnaire was designed which gathered information on whether
the patient was alive at six months and, if so, the status on the
Glasgow Outcome Scale [14]. The GOS is generally accepted as a
valid measure of outcome after head injury, with adequate observer
reliability [21]. General de®nitions on outcome categories were pro-
vided to centres, but assignment of subjects was not based on a
structured interview as has recently been proposed by Wilson et al.
[44]. An individual form was prepared for each patient in the survey,
and these forms were sent to centres for completion.
The entire survey was designed to be conducted on a very limited
budget. In particular, there were no resources for additional research
assistants, or for site visits to check data against source records. The
monitoring was limited to checking the forms as they were received
at the EBIC Co-ordinating Centre, and any inconsistencies in the
data were queried with the relevant centres. On completion of the
data collection, a report was generated which was tailored for
each individual centre. This reported detailed results for that centre,
together with results for the relevant country and for the entire series,
and gave an opportunity for the centres themselves to raise queries
with the data.
Results
Response Rate
Core Data forms were sent to 104 centres, and of
these 67 (64%) in twelve countries returned data on a
total of 1005 adult head injuries. Forty seven (5%) of
the cases were injured just outside the three month
window set for the survey but are included and this
report gives results for all 1005 cases. Table 1 gives the
224 G. D. Murray et al.
numbers of centres and the number of cases per coun-
try, with the countries ordered by number of cases.
Demographics
Di¨erences in patient demographics, treatments,
complications and outcomes between various regions
were analyzed with the chi-squared test.
The patterns of age (mean 42, median 38 years), sex
(74% male) and cause of injury (51% some type of road
tra½c accident) are shown in Table 2.
Referral to ``Neuro'' Unit
Only 43% of patients were admitted directly to a
hospital with neurosurgical facilities, the remaining
57% were transferred secondarily from another hospi-
tal. Patients admitted directly to the hospital contain-
ing the neurotrauma unit took rather longer to reach
hospital (median 45 minutes) compared to those
admitted to another hospital for assessment before
transfer (median 35 minutes). On the other hand,
direct admission to a hospital with a neurotrauma unit
was associated with a shorter time from injury to the
patient coming under specialist care (median 1 hour) in
contrast with a median of four hours for patients
transferred from another hospital. Such direct admis-
sion was the rule in Spain and the Benelux countries,
whereas secondary transfer was the rule in the UK,
France and Scandinavia, with Italy and Germany
occupying intermediate positions.
Clinical Assessment
Assessment of the components of the Glasgow
Coma Scale was limited by widespread use of sedation
and intubation. Table 3 summarises the proportion of
cases where GCS was recorded at di¨erent time points
(including situations where the GCS was recorded as
`untestable'), and those where GCS could be assessed.
Table 1. Summary of Cases Reported
Country Number of centres Number of cases
Germany 19 241
United Kingdom 15 219
Italy 10 184
France 4 95
Spain 3 90
The Netherlands 4 58
Sweden 4 46
Finland 1 19
Switzerland 2 18
Denmark 2 15
Yugoslavia 1 11
Belgium 2 9
Total 67 1005
Table 2. Features of Patients Reported to the European Brain Injury
Consortium Core Data Survey
Period 1995 February ± April
Criteria
age adult (> 16 years)
GCS severe U8 or moderate (9±
12) at pre Hospital, A&E,
post-resuscitation or admis-
sion to NSU
other admitted to NSU within 24
hours of injury
Total sample size 1005
Direct admissions 422/989 (43%)
age
mean 42
SD 21
median 38
range 2 to 92 (23 aged U14)
interquartile range 24 to 59
Male 738/1000 (74%)
Type of injury number percent
motor vehicle occupant 295 (30%)
pedestrian 126 (13%)
RTA other (or unknown) 87 (9%)
work 63 (6%)
assault 53 (5%)
domestic 122 (12%)
sport 30 (3%)
fall under in¯uence of alcohol 121 (12%)
other 99 (10%)
996
Major extracranial injury 354/982 (36%)
Initial CT classi®cation
di¨use I 121 (12%)
di¨use II 273 (28%)
di¨use III 101 (10%)
di¨use IV 21 (2%)
mass 467 (48%)
983
Subarachnoid haemorrhage on CT 385/953 (40%)
Intracranial operation (within ®rst
24 hours)
no 569 (57%)
burr hole for ICP alone 85 (8%)
other 346 (35%)
1000
Ventilated 736/948 (78%)
ICP monitored 346/945 (37%)
Jugular SVO
2
monitored 173/939 (18%)
Arterial pressure monitored 631/933 (68%)
European Brain Injury Consortium
225
Twenty-four percent of cases were recorded as obeying
commands according to the GCS motor score on at
least one of the four assessments.
At time of admission to the Neurosurgical Unit,
GCS was U8 in 329 subjects, 9±12 in 162 subjects and
13±15 in 75 subjects, the GCS was untestable in 371
subjects and not recorded in 68 subjects.
Admission CT Findings
The appearances on the ®rst CT scan after admis-
sion were classi®ed according to the Traumatic Coma
Data Bank (TCDB) categories [26]. Twelve percent
were class I (normal), 28% class II (di¨use injury), 10%
class III (di¨use injury with swelling), 2% class IV
(di¨use injury with shift), and 48% were classi®ed as
having a `mass' lesion. Subarachnoid haemorrhage
was identi®ed in 40% of cases and intraventricular
haemorrhage was identi®ed in 14% of cases. In total
897 patients had data on a further `®nal' or `worst' CT
scan as well as their admission scan, and these data are
being presented fully in a separate report.
Early Complications
Twenty percent of patients were recorded as having
minor extracranial injuries, and 36% were recorded as
having major extracranial injuries, de®ned as an injury
which in itself would have required hospital admission.
Fourteen percent of all cases underwent an emergency
extracranial operation. Early complications were re-
corded as hypoxia (27%), hypotension (22%) and hy-
pothermia (6%), and 35% of patients underwent an
intracranial operation (other than the placement of an
ICP catheter or transducer) within the ®rst 24 hours
following injury.
Management and Monitoring
Ventilation was used in 78% of patients, ICP, jugu-
lar SVO
2
and invasive blood pressure monitoring were
used respectively in 37%, 18% and 68% of patients.
Delayed Complications
Intracranial infection was reported in 8 patients
(1%), and other intracranial complications of delayed
haematoma (after 24 hours), raised ICP and seizures
were reported respectively in 11%, 28% and 7% of
patients. Life threatening systemic complications clas-
si®ed as respiratory, cardiovascular, metabolic and
infection were reported respectively for 24%, 12%, 5%
and 13% of subjects.
Outcome at six Months
Fifty ®ve centres provided data on Glasgow Out-
come Score (GOS) at six months for 796 subjects.
These were 94% of the 847 cases initially reported from
these centres. One hundred and ®fty eight (76%) of the
remaining 209 cases without data on six month GOS
came from 12 centres which were unable to supply any
data on six month outcome in any patient. Therefore,
in the 55 centres able to provide information on six
month outcome, the data were 94% complete. The
features of the cases in the 12 centres that did not
provide outcome data were broadly similar to those in
the 55 centres reporting outcome. Furthermore, in the
latter centres, the initial features of the cases with and
without outcome data were very similar.
Of the 796 patients whose GOS was available at six
months, 244 (31%) were dead, 20 (3%) vegetative, 124
(6%) severely disabled, 159 (20%) moderately disabled,
and 249 (31%) were considered to have made a good
recovery. Thus, the combination of the last two groups
into a `favourable' outcome occurred in 51%.
Severity Subsets
The criteria for inclusion of patients included
patients with `moderate' (GCS 9±12) as well as severe
(GCS U8) head injuries. Identi®cation of severe cases
was complicated by the variability in data available at
the various initial time points, in particular data being
`missing' because of intubation, paralysis and ventila-
tion. For comparisons within this survey and with
previous series we identi®ed three subsets of patients:
Table 3. Availability of Glasgow Coma Scale at Di¨erent Stages in the Early Triage and Management of Head Injured Patients
Pre-hospital A&E department `Post-resuscitation' Admission to neurosurgery
GCS motor score recorded 65% 89% 76% 94%
GCS motor score testable 62% 82% 62% 72%
Full GCS recorded 65% 89% 75% 93%
Full GCS testable 61% 77% 49% 56%
226 G. D. Murray et al.
Severe cases n 583 were composed of: a) all
those with GCS 3±8 on admission to the neurosurgical
unit (NSU) n 329, b) those other patients whose
GCS at admission to NSU was untestable or not re-
corded and who had at least one previous observation
of a GCS 3±8 and none of a higher GCS n 254.
The moderate group had a GCS of V9 on admission
to NSU and no other recording of a GCS of < 9
n 171.
A third group of `intermediate' severity cases
n 192 did not have a GCS of 3±8 at admission
to NSU but had, on other occasions at least one
GCS of 3±8 and at least one of V9. In 59 patients there
were insu½cient data on GCS to make any sensible
classi®cation.
Patients classi®ed as severe in this way, in compari-
son with the pooled intermediate and moderate
groups, were younger (median age 34 years versus 42
years), more often a vehicle occupant (35% versus
21%) more often admitted directly to a hospital with a
neuro unit (45% versus 40%), had a higher frequency
of major extracranial injury (41% versus 28%), of an
intracranial operation in the ®rst 24 hours (37% versus
30%), and their CT scans were less often normal (10%
versus 16%) and more often showed di¨use swelling
(13% versus 6%) or traumatic subarachnoid haemor-
rhage (47% versus 32%). They were more often venti-
lated (92% versus 56%), had invasive monitoring of
blood pressure (80% versus 48%), or intracranial pres-
sure (43% versus 24%). Thirty six percent of the severe
group had bilateral non-reacting pupils at admission
to NSU, in comparison with 7% for the intermediate/
moderate groups. The outcomes in these groups are
shown in Table 4. The proportions with a `favourable'
outcome at six months were: severe, 40%; intermedi-
ate, 63%; moderate, 77%; and unclassi®ed, 57%.
Comparison of Data from Di¨erent Areas of Europe
There were data from a su½cient number of subjects
for a country based analysis in only some cases (ie
Germany, UK, Italy, France and Spain), and other
countries were grouped arbitrarily by region of Europe
(Scandinavia and Benelux). There were no major
di¨erences in the data from these areas for subjects'
ages or sex distribution but several di¨erences in dis-
tribution that were signi®cantly di¨erent were noted
(Table 5).
1. The frequency of secondary transfer to the hospital
with the neuro unit ranged from 35% in the Benelux
countries to 75% in the UK p < 0:001.
2. There were substantial di¨erences amongst coun-
tries in the cause of injury p < 0:001. The pro-
portion who were injured as a vehicle occupant
ranged from 11% in the UK to 48% in the Benelux
countries, and of those who fell under the in¯uence
of alcohol from 1% in Spain to 33% in Scandinavia.
3. The proportion of the subjects with a major extra-
cranial injury ranged from 24% in Scandinavia to
53% in the Benelux countries p < 0:001.
4. There were marked di¨erences in admission CT
®ndings amongst the countries p < 0:001. The
proportion of patients with a normal CT scan
ranged from 4% in Scandinavia to 10% in France,
and those with swelling from 7% in Spain to 18% in
Italy. The proportions with subarachnoid haemor-
rhage ranged from 33% in the UK to 57% in Spain
p < 0:001.
5. The proportion of the subjects who had an in-
tracranial operation in the ®rst 24 hours ranged
from 18% in the Benelux to 53% in Germany
p < 0:001.
6. The frequency of the use of ventilation ranged from
53% of subjects in France to 88% in Germany, of
invasive blood pressure monitoring from 31% in
France to 89% in Scandinavia and of intracranial
pressure monitoring from 5% in France to 53% in
Spain (p < 0:001 in each case).
7. The proportion of subjects with a severe injury (as
de®ned above from the data available on GCS at
the four early time points) ranged from 42% in the
Table 4. Outcome at Six Months in Subjects Grouped by `Overall' Initial Severity. Severity Subsets were De®ned from Data at all 4 Early Points
(see Text). Figures are Numbers of Subjects (%)
Severity subset Total Dead Vegetative Severe disability Moderate disability Good recovery
Severe 481 192 (40%) 17 (4%) 78 (16%) 93 (19%) 101 (21%)
Intermediate 145 32 (22%) 2 (1%) 20 (14%) 27 (19%) 64 (44%)
Moderate 128 11 (9%) 0 (0%) 18 (14%) 31 (24%) 68 (53%)
Unclassi®ed 42 9 (21%) 1 (2%) 8 (19%) 8 (19%) 16 (38%)
European Brain Injury Consortium
227
