J Neurosurg 117:324–333, 2012
Outcome after severe brain trauma due to acute
subdural hematoma
Clinical article
Johannes Leitgeb, M.D.,1 Walter Mauritz, M.D., Ph.D., 2,3
Alexandra Brazinova, M.D., Ph.D., M.P.H., 4,5 Ivan Janciak, Ph.D., 6
Marek Majdan, Ph.D., 5,7 Ingrid Wilbacher, M.Sc., Ph.D., 8
and Martin Rusnak, C.Sc., M.D., Ph.D. 5,9
1
Department of Traumatology, Medical University of Vienna; 2Department of Anesthesiology and Intensive
Care Medicine, Trauma Hospital “Lorenz Boehler”; 3Vice President, 4Executive Director, 6Information and
Data Manager, 7Statistician, 8Senior Research Fellow, and 9President, International Neurotrauma Research
Organization, Vienna, Austria; and 5Department of Public Health, Faculty of Health and Social Services,
Trnava University, Trnava, Slovakia
Object. In this paper, the authors’ goal was to identify factors contributing to outcomes after severe traumatic
brain injury (TBI) due to acute subdural hematoma (SDH).
Methods. Between February 2002 and April 2010, 17 Austrian centers prospectively enrolled 863 patients with
moderate and severe TBI into observational studies. Data regarding accident, treatment, and outcomes were collected. Data sets from patients who had severe TBI (Glasgow Coma Scale score < 9) and acute SDH were selected.
Six-month outcomes were classified as “favorable” if the Glasgow Outcome Scale (GOS) scores were 5 or 4, and they
were classified as “unfavorable” if GOS scores were 3 or less. The Rotterdam score was used to classify CT findings,
and the scores published by Hukkelhoven et al. were used to estimate the predicted rates of death and of unfavorable
outcomes. Univariate (Fisher exact test, t-test, chi-square test) and multivariate (logistic regression) statistics were
used to identify factors associated with hospital mortality and favorable outcome.
Results. Of the 738 patients with severe TBI, 360 (49%) had acute SDH. Of these, 168 (46.7%) died in the hospital, 67 (18.6%) survived with unfavorable outcome, and 116 (32.2%) survived with favorable outcome. Long-term
outcome was unknown in 9 survivors (2.5%). Mortality rates predicted by the Rotterdam CT score showed good correlation with observed mortality rates. According to the Hukkelhoven scores, observed/predicted ratios for mortality
and unfavorable outcome were 1.09 and 1.02, respectively.
Conclusions. Age, severity of TBI, and neurological status were the main factors influencing outcomes after
severe TBI due to acute SDH. Nonoperative management was associated with significantly higher mortality.
(http://thejns.org/doi/abs/10.3171/2012.4.JNS111448)
Key Words • traumatic brain injury • acute subdural hematoma •
outcome • prognostic score
T
raumatic brain injuries represent a serious public
health problem. This injury is the leading cause of
death in people younger than 45 years old in the
US, Europe,27 and most other countries.11 It has been estimated that each year 1 million admissions to European
hospitals are caused by TBI.11 Incidence, mortality, and
Abbreviations used in this paper: AIS = Abbreviated Injury
Score; GCS = Glasgow Coma Scale; GOS = Glasgow Outcome
Scale; ICP = intracranial pressure; INRO = International Neurotrauma Research Organization; ISS = Injury Severity Score; IVH
= intraventricular hemorrhage; PD = probability of hospital death; PU
= probability of unfavorable long-term outcome; SDH = subdural
hematoma; TBI = traumatic brain injury.
324
morbidity of TBI are even higher in developing countries.11
Acute SDH has been recognized as a devastating injury. In patients with severe TBI due to acute SDH, mortality rates of 60% and more, depending on GCS scores,
have been published.13 Some authors have developed
algorithms to estimate the probability of functional recovery29 or poor outcome.22 However, most of the studies
on acute SDH came from single centers, were performed
more than 10 years ago, and enrolled comparatively few
patients. In addition, most previous studies enrolled patients with mild, moderate, and severe TBI. The goal of
this study was to identify factors influencing outcomes
after severe TBI due to acute SDH using fairly recent
J Neurosurg / Volume 117 / August 2012
Outcome after severe TBI due to acute SDH
data from a large sample of patients enrolled in Austrian
trauma centers.
Methods
Between 2001 and 2010 the INRO (a nongovernmental research organization, founded in 1999 and based in
Vienna, Austria) coordinated 2 projects that focused on
Austrian patients with TBI. The first project analyzed epidemiology and hospital treatment of patients with severe
TBI as well as the effects of guideline-based treatment.23
This project was begun in March 2002, and 5 centers enrolled 415 patients until June 2005. The second project
focused on prehospital and early hospital management
of patients with moderate and severe TBI. It started in
March 2009, and 16 centers enrolled 448 patients until
April 2010. Both projects were conducted after approval
was obtained from the local ethics committees.
Databases developed by INRO were used to collect
data for these projects. Data from the 2 projects were
compared by examination for significant differences in
epidemiology, treatment, and outcomes. No relevant differences were found, and the 2 databases were merged.
The merged database now holds data from 863 patients.
Since both projects were purely observational, pediatric
and geriatric patients, patients with TBI with multiple
trauma, and patients with low GCS scores were enrolled;
these patients are usually excluded from industry-sponsored studies.
The data for this study were collected in 17 Austrian
centers. All centers were of tertiary care level and were
able to provide patient management according to the Brain
Trauma Foundation guidelines for the management of patients with severe TBI.1 The number of patients enrolled
by these centers (median 28 [IQR 21–65], range 3–150
patients) varied considerably, as 4 centers participated in
both projects, and some centers joined the second project
with just a few weeks remaining for patient inclusion.
Treatment in the field was provided by emergency
physicians. All patients underwent rapid examination,
which included documentation of vital signs (GCS score,
pupil status, blood pressure, heart rate, and oxygen saturation). Rapid-sequence intubation facilitated by hypnotics
and relaxants, ventilation, treatment of hemorrhage, and
fluid resuscitation were done as appropriate. After admission, each patient was examined by a trauma team (anesthesiologists, trauma surgeons, and/or neurosurgeons,
radiologists, and nurses), and a CT scan was obtained.
The patients then underwent surgery as appropriate and/
or were admitted to the ICU. Neurosurgery was provided
by neurosurgeons (6 centers) or by trauma surgeons (11
centers) who had the option of consulting neurosurgeons
for more difficult cases. Intensive care was provided by
anesthesiologists in cooperation with neurosurgeons or
trauma surgeons.
Basic patient demographic data, cause and location of
trauma, prehospital status and treatment, mechanism and
severity of trauma (AIS and ISS), CT scanning findings,
laboratory test results, and data on surgical procedures
and outcomes were recorded prospectively. Prehospital
data were documented by the local paramedics, and these
J Neurosurg / Volume 117 / August 2012
data were then transferred into the databases. Computed
tomography scans were interpreted by neurosurgeons,
trauma surgeons, and radiologists, and the summarized
findings were entered into a separate page in the databases. This page collected detailed data on the appearance of
the basal cisterns (open, compressed, or absent), midline
shift, and main findings (such as edema, hematoma, and
contusions). No central review of CT scans was done in
the first project, as no actual images were uploaded into
the databases. Central review of the CT scans was done in
the second project; CT scans were collected on compact
discs, and a radiologist and a trauma surgeon checked
the accuracy of the data entered into the database. Data
regarding duration of various treatments, complications,
and outcome were collected at discharge from the ICU
and at hospital discharge. Information on status and location was recorded at 6 months after injury. This was
done by phone calls to the patients and/or their relatives;
in some cases the GOS score was recorded at the patients’
follow-up visits to the centers. In all centers, data were
collected by local research fellows; the data quality was
monitored by the INRO project manager (A.B.). Missing
or implausible data were reported to local investigators
who then submitted missing or corrected values. Personal
data protection was observed, and the identifiers were
kept separately from the data.
All patients for whom an acute SDH was noted on
their first CT scan were selected for this analysis. Only
patients with severe TBI (defined as a GCS score of 8
or less in the field and/or following resuscitation) and
who survived at least until admission to the ICU were included. Data on trauma mechanism, trauma severity, CT
findings, treatment, and outcomes were retrieved for each
patient. The 6-point Rotterdam CT score15 was used to
classify CT findings, and the PD according to this score
was calculated using the formula given in the paper. In
addition, the Marshall classification19 was also used to
classify CT findings. The prognostic scores developed by
Hukkelhoven et al.9 were used to estimate the PD and the
PU. To describe long-term outcomes, the GOS12 was used.
“Favorable outcome” was defined as a GOS score of 5
or 4, and “unfavorable outcome” was defined as a GOS
score of 3 or less at 6 months after trauma. According
to their hospital outcomes, patients were assigned to the
groups “alive” and “dead.” The differences between these
2 groups of patients were analyzed.
The free software provided by P. Wessa (Free Statistics Software version 1.1.23-r6, http://www.wessa.net)
was used for calculations. The 2-tailed t-test (for comparisons of mean values), Fisher exact test (for analysis of 2 ×
2 tables), and chi-square test (for analysis of N × 2 contingency tables) were done as appropriate to identify differences between the groups. To check for associations with
outcomes, 2 logistic regression models were constructed:
1 each for hospital survival and favorable long-term outcome. The parameters age, injury mechanism, sex, ISS,
head AIS, direct transfer, prehospital hypotension, prehospital hypoxia, pupil reactivity, GCS score, presence of
IVH, Rotterdam CT score, and treatment strategy (operative vs nonoperative treatment) were used as confounding variables, with backward exclusion of nonsignificant
325
J. Leitgeb et al.
parameters. A p value < 0.05 was considered statistically
significant.
Results
There were 863 data sets in the database. Of these,
outcome was not recorded in 18 patients, 9 patients died
prior to ICU admission, and 98 patients had only moderate TBI (GCS scores of 9–12 after admission); this left
738 patients for analysis. Of these, 360 (48.8%) had an
acute SDH on their CT scan and were selected for this
analysis.
Demographic data for the 2 groups of patients are
given in Table 1. About 70% of the patients were male,
and there was no significant difference between the
groups. Sex had no effect on outcome, although female
patients were significantly older than males. There was
a significant effect of age; patients who survived were
significantly younger. Fifty-six percent of the patients
who died (94 of 168 patients) were older than 60 years
of age. There was a significant increase in mortality rates
(from 25% to 63%) with increasing age. With regard to
trauma mechanism, there were no significant differences,
although the rate of low-level falls was higher in nonsurvivors. Type of trauma (blunt vs penetrating trauma) had
no effect on mortality.
Trauma severity was significantly lower in patients
who survived (Table 2); mean ISS, head AIS, GCS score,
the number of patients with unreactive pupils, and the
number with hypotension and hypoxia in the field were
TABLE 1: Age, sex, and trauma mechanism in 360 patients who suffered a TBI
Value*
Parameter
patients
female
male
total
mean age (yrs)
female
male
total
age group (yrs)
1–15
16–30
31–45
46–60
61–75
>76
total
trauma mechanism
blunt assault
gunshot
fall <3 m
fall >3 m
traffic related
bicycle
sports related
work related
unknown
other
total
type of trauma
blunt
penetrating
total
Alive
Dead
Total
% Mortality
p Value†
59 (30.7)
133 (69.3)
192 (100.0)
55 (32.7)
113 (67.3)
168 (100.0)
114 (31.7)
246 (68.3)
360 (100.0)
48.2
45.9
46.7
0.73
63.7 ± 18.6
49.0 ± 19.9
53.5 ± 20.6
68.5 ± 19.8
56.7 ± 21.4
60.6 ± 21.6
66.0 ± 19.3
52.6 ± 20.9
56.8 ± 21.3
3 (1.6)
34 (17.7)
33 (17.2)
36 (18.8)
53 (27.6)
33 (17.2)
192 (100.0)
1 (0.6)
20 (11.9)
21 (12.5)
32 (19.0)
39 (23.2)
55 (32.7)
168 (100.0)
4 (1.1)
54 (15.0)
54 (15.0)
68 (18.9)
92 (25.6)
88 (24.4)
360 (100.0)
25.0
37.0
38.9
47.1
42.4
62.5
46.7
4 (2.1)
0 (0.0)
69 (35.9)
23 (12.0)
46 (24.0)
12 (6.2)
12 (6.2)
8 (4.2)
6 (3.1)
12 (6.2)
192 (100.0)
1 (0.6)
4 (2.4)
79 (47.0)
16 (9.5)
34 (20.2)
11 (6.6)
4 (2.4)
5 (3.0)
2 (1.2)
12 (7.1)
168 (100.0)
5 (1.4)
4 (1.1)
148 (41.1)
39 (10.8)
80 (22.2)
23 (6.6)
16 (4.4)
13 (3.6)
8 (2.2)
24 (6.7)
360 (100.0)
20.0
100.0
53.4
41.0
42.5
25.0
38.5
25.0
50.0
46.7
0.065
177 (92.2)
15 (7.8)
192 (100.0)
156 (92.9)
12 (7.1)
168 (100.0)
333 (92.5)
27 (7.5)
360 (100.0)
46.8
44.4
46.7
0.84
<0.001
<0.001
0.018
* Mean values are presented as the mean ± SD. All other values are the number of patients (%).
† The p values refer to differences between surviving versus nonsurviving patients.
326
J Neurosurg / Volume 117 / August 2012
Outcome after severe TBI due to acute SDH
TABLE 2: Parameters relevant to severity of brain injury and overall severity of injury, and incidence of field
hypotension and hypoxia
Value*
Parameter
additional regions w/ AIS >2
0
1
2
>2
total
GCS score
3
4
5
6
7
8
total
mean ISS
mean head AIS
mean enrollment GCS score
pupils
0 reactive
1 reactive
2 reactive
unknown
total
prehospital hypotension
no
yes
total
prehospital hypoxia
no
yes
total
Alive
Dead
Total
% Mortality
p Value†
119 (62.0)
41 (21.4)
24 (12.5)
8 (4.2)
192 (100.0)
113 (67.3)
32 (19.0)
16 (9.5)
7 (4.2)
168 (100.0)
232 (64.4)
73 (20.3)
40 (11.1)
15 (4.2)
360 (100.0)
48.7
43.8
40.0
46.7
46.7
0.72
70 (36.5)
11 (5.7)
23 (12.0)
34 (17.7)
25 (13.0)
29 (15.1)
192 (100.0)
24.1 ± 10.5
4.1 ± 0.5
5.1 ± 1.9
96 (57.1)
23 (13.7)
19 (11.3)
19 (11.3)
6 (3.6)
5 (3.0)
168 (100.0)
38.7 ± 23.3
4.8 ± 0.9
4.0 ± 1.4
166 (46.1)
34 (9.4)
42 (11.7)
53 (14.7)
31 (8.6)
34 (9.4)
360 (100.0)
31.0 ± 19.1
4.4 ± 0.8
4.6 ± 1.8
57.8
67.6
45.2
35.8
19.4
14.7
46.7
11 (5.7)
1 (0.5)
143 (74.5)
37 (19.3)
192 (100.0)
49 (29.2)
2 (1.2)
86 (51.2)
31 (18.5)
168 (100.0)
60 (16.7)
3 (0.8)
229 (63.6)
68 (18.9)
360 (100.0)
81.7
66.7
37.6
45.6
46.7
<0.001
183 (95.3)
9 (4.7)
192 (100.0)
147 (87.5)
21 (12.5)
168 (100.0)
330 (91.7)
30 (8.3)
360 (100.0)
44.5
70.0
46.7
0.012
174 (90.6)
18 (9.4)
192 (100.0)
140 (83.3)
28 (16.7)
168 (100.0)
314 (87.2)
46 (12.8)
360 (100.0)
44.6
60.9
46.7
0 .041
<0.001
<0.001
<0.001
<0.001
* Mean values are presented as the mean ± SD. All other values are the number of patients (%).
† The p values refer to differences between surviving versus nonsurviving patients.
significantly lower. Fifty-five percent of the patients had
GCS scores of 3 or 4; more than 70% of the patients
who died, but only 42% of those who survived, had GCS
scores of 3 or 4. Nearly two-thirds of the patients from
both groups had isolated TBI. Associated injuries with an
AIS greater than 2 were seen more frequently in survivors, but this difference was not significant. The number
of additionally injured regions had no effect on mortality.
With regard to CT findings (Table 3), increasing
values of the Rotterdam CT score were significantly associated with increasing mortality rates. No patient had
the best score of 1. There was a significant correlation
between the PD predicted by the score and the observed
mortality rates (y = 1. 2281x + 6.11; R2 = 0.933; p = 0.007
J Neurosurg / Volume 117 / August 2012
for scores 2–6). According to the Marshall classification,
198 patients (55%) had an evacuated mass lesion, and 162
(45%) had a nonevacuated mass lesion. Mortality was
lower in the patients with evacuated mass lesions (42.9%
vs 51.2%), but this difference was not significant (p =
0.14). With the exception of IVH, additional CT findings
were not associated with significantly increased mortality.
With regard to treatment, indirect transfer was associated with significantly lower mortality (Table 4). The
mode of transport, rates of prehospital intubation, and
timing of CT scanning and surgery were not significantly
different. Conservative management without ICP monitoring was associated with the highest mortality (64.2%);
327
J. Leitgeb et al.
TABLE 3: Rotterdam CT scores and various additional CT findings*
No. of Patients (%)
Parameter
Rotterdam CT score
2
3
4
5
6
ND
total
additional EDH
no
yes
total
additional SAH
no
yes
total
additional IVH
no
yes
total
additional contusion
no
yes
total
Alive
Dead
Total
% Mortality
p Value†
42 (21.9)
91 (47.4)
42 (21.9)
14 (7.3)
2 (1.0)
1 (0.5)
192 (100.0)
22 (13.1)
55 (32.7)
44 (26.2)
35 (20.8)
8 (4.8)
4 (2.4)
168 (100.0)
64 (17.8)
146 (40.6)
86 (23.9)
49 (13.6)
10 (2.8)
5 (1.4)
360 (100.0)
34.4
37.7
51.2
71.4
80.0
80.0
46.7
<0.001
169 (88.0)
23 (12.0)
192 (100.0)
142 (84.5)
26 (15.5)
168 (100.0)
311 (86.4)
49 (13.6)
360 (100.0)
45.7
53.1
46.7
0.36
93 (48.4)
99 (51.6)
192 (100.0)
70 (41.7)
98 (58.3)
168 (100.0)
163 (45.3)
197 (54.7)
360 (100.0)
42.9
49.7
46.7
0.21
180 (93.8)
12 (6.3)
192 (100.0)
143 (85.1)
25 (14.9)
168 (10 0.0)
323 (89.7)
37 (10.3)
360 (100.0)
44.3
67.6
46.7
0.009
104 (54.2)
88 (45.8)
192 (100.0)
93 (55.4)
75 (44.6)
168 (100.0)
197 (54.7)
163 (45.3)
360 (100.0)
47.2
46.0
46.7
0.83
* EDH = epidural hematoma; ND = no data; SAH = subarachnoid hemorrhage.
† The p values refer to differences between surviving versus nonsurviving patients.
mortality rates were significantly lower (p = 0.01) with
conservative management with ICP monitoring (38.3%)
or with operative management (42.9%). Primary craniectomy was associated with lower mortality than craniotomy (39% vs 46%), but the difference was not significant
(p = 0.32). The use of ICP monitoring was associated with
significantly lower mortality (42% vs 54%, p = 0.037).
The observed hospital mortality rate was 46.7%
(168 of 360 patients), and the PD predicted by the Hukkelhoven score was 42.7% ± 22.3%. The observed versus
predicted ratio for mortality was 1.09; this ratio equals
32 unexpected deaths but is within the predicted range.
The PD predicted by the Rotterdam CT score was 52.6 ±
18; the observed versus predicted ratio for mortality for
this score was 0.89; this equals 39 unexpected survivors
and is also within the predicted range. Most of the nonsurvivors died in the ICU (146 [87%]), and 22 died on
intermediate or regular wards. Most patients died of brain
death (113 [67.3%] of 168), cardiovascular problems (24
[14.3%] of 168), hemorrhage (7 [4.2%] of 168), or multiple
organ failure (6 [3.6%] of 168). Four patients (2.4%) died
of other causes, and the cause of death was unknown in
the remaining 14 patients (8.4%). Logistic regression re328
vealed that age, GCS score, pupil reactivity, head AIS,
prehospital hypoxia, and nonoperative treatment were
significantly associated with hospital mortality (Table 5).
Sex, injury mechanism, ISS, indirect transfer, prehospital
hypotension, Rotterdam CT score, and presence of IVH
had no significant influence on hospital mortality.
Of the 192 hospital survivors 116 (60.4%) had favorable outcomes, 67 (34.9%) had unfavorable outcomes, and
long-term outcome was unknown in 9 patients (4.7%).
Of all patients in this study, 116 (32.2%) had favorable
outcomes, 235 (65.3%) had unfavorable outcomes (168
patients who died in the hospital plus 67 patients who
survived with unfavorable outcomes), and long-term outcome was unknown in 9 patients (2.5% of all patients).
The PU predicted by the Hukkelhoven score was 63.8% ±
21.9%; the observed/predicted ratio for unfavorable outcome was 1.02. This ratio equaled 7 patients with unexpected unfavorable outcome and was within the range of
the predicted PU. If all patients with unknown long-term
outcome were classified as “unfavorable” the observed/
predicted ratio for unfavorable outcome would be 1.06
and would still be within the predicted range. Logistic regression analysis revealed that age, GCS score, head AIS,
J Neurosurg / Volume 117 / August 2012
Outcome after severe TBI due to acute SDH
TABLE 4: Selected parameters of prehospital and hospital treatment*
No. of Patients (%)
Parameter
indirect transfer
no
yes
total
mode of transport
air
ground
total
prehospital intubation
no
yes
total
time btwn admission & CT scan (mins)
0–30
31–60
61–90
>90
unknown
total
time btwn CT scan & start of op (mins)
0–20
21–40
41–60
61–120
>120
total
neurosurgery
no surgery
ICP monitoring
ASDH evacuation
ASDH + EDH evacuation
ASDH + ICH evacuation
ASDH + EDH + ICH evacuation
decompressive
total
surgical technique
craniectomy
craniotomy
decompressive
total
ICP monitoring
no
yes
total
Alive
Dead
Total
% Mortality
p Value†
147 (76.6)
45 (23.4)
192 (100.0)
148 (88.1)
20 (11.9)
168 (100.0)
295 (81.9)
65 (18.1)
360 (100.0)
50.2
30.8
46.7
0.006
70 (36.5)
122 (63.5)
192 (100.0)
54 (32.1)
114 (67.9)
168 (100.0)
124 (34.4)
236 (65.6)
360 (100.0)
43.5
48.3
46.7
0.44
67 (34.9)
125 (65.1)
192 (100.0)
50 (29.8)
118 (70.2)
168 (100.0)
117 (32.5)
243 (67.5)
360 (100.0)
42.7
48.6
46.7
0.31
116 (60.4)
48 (25.0)
13 (6.8)
14 (7.3)
1 (0.5)
192 (100.0)
97 (57.7)
44 (26.2)
7 (4.2)
14 (8.3)
6 (3.6)
168 (100.0)
213 (59.2)
92 (25.6)
20 (5.6)
28 (7.8)
7 (1.9)
360 (100.0)
45.5
47.8
35.0
50.0
85.7
46.7
0.73
31 (19.6)
45 (28.5)
16 (10.1)
28 (17.7)
38 (24.1)
158 (100.0)
25 (22.1)
32 (28.3)
18 (15.9)
23 (20.4)
15 (13.3)
113 (100.0)
56 (20.7)
77 (28.4)
34 (12.5)
51 (18.8)
53 (19.6)
271 (100.0)
44.6
41.6
52.9
45.1
28.3
41.7
0.19
29 (15.1)
50 (26.0)
93 (48.4)
16 (8.3)
4 (2.1)
0 (0.0)
0 (0.0)
192 (100.0)
52 (31.0)
31 (18.5)
72 (42.9)
9 (5.4)
2 (1.2)
1 (0.6)
1 (0.6)
168 (100.0)
81 (22.5)
81 (22.5)
165 (45.8)
25 (6.9)
6 (1.7)
1 (0.3)
1 (0.3)
360 (100.0)
64.2
38.3
43.6
36.0
33.3
100.0
100.0
46.7
0.01
55 (48.7)
58 (51.3)
0 (0.0)
113 (100.0)
35 (41.2
49 (57.6)
1 (1.2)
85 (100.0)
90 (45.5)
107 (54.0)
1 (0.5)
198 (100.0)
38.9
45.8
100.0
42.9
0.32
61 (31.8)
131 (68.2)
192 (100.0)
72 (42.9)
96 (57.1)
168 (100.0)
133 (36.9)
227 (63.1)
360 (100.0)
54.1
42.3
46.7
0.037
* ASDH = acute SDH; ICH = intracerebral hematoma.
† The p values refer to differences between surviving versus nonsurviving patients.
J Neurosurg / Volume 117 / August 2012
329
J. Leitgeb et al.
TABLE 5: Coefficients of logistic regressions for hospital outcome and 6-month outcome*
Variable
hospital outcome†
intercept
age
GCS score
pupils
head AIS score
field hypoxia
op treatment
sex
mechanism
ISS
field hypotension
transfer
IVH
Rotterdam score
6-mo outcome‡
intercept
age
GCS score
head AIS score
Rotterdam score
pupils
field hypoxia
op treatment
sex
mechanism
ISS
field hypotension
transfer
IVH
Parameter
Standard Error
t Statistic
2-Sided p Value
Significant
1.4391
−0.0304
0.3019
1.2261
−1.0352
1.1149
1.3882
−0.3046
−0.2381
−0.0221
−0.2802
0.3403
0.1991
0.2034
2.2788
0.0077
0.0815
0.4089
0.2730
0.3621
0.4023
0.2997
0.1771
0.0129
0.5455
0.2081
0.2816
0.2957
0.6315
−3.9777
3.7072
2.9985
−3.792
2.5499
3.4508
−1.0161
−1.3443
−1.7096
−0.5138
1.6348
0.7072
0.6881
0.5281
0.0008
0.0002
0.0029
0.0002
0.0032
0.0006
0.3102
0.1797
0.0882
0.6077
0.1029
0.4799
0.4918
yes
yes
yes
yes
yes
yes
no
no
no
no
no
no
no
2.7676
−0.0333
0.2049
−1.1776
0.9242
0.8248
0.1825
0.0203
−0.2438
−0.2298
−0.0069
−0.3723
0.4029
0.2480
2.3698
0.0075
0.0762
0.2803
0.4689
0.4088
0.4446
0.3020
0.3009
0.1716
0.01415
0.5868
0.3583
0.2817
1.1678
−4.3893
2.6893
−4.2019
1.9709
2.9714
0.4106
0.0673
−0.8102
−1.3389
−0.4939
−0.6344
1.1245
0.8805
0.2437
0.0002
0.0075
0.0003
0.0596
0.1665
0.6817
0.9463
0.4183
0.1815
0.6217
0.5262
0.2616
0.3792
yes
yes
yes
yes
no
no
no
no
no
no
no
no
no
* All parameters that were included in the model are presented, including those that had no significant influence.
† Survival = 1.
‡ Favorable = 1.
and the Rotterdam CT score were significantly associated
with long-term outcomes (Table 5). All other variables
had no significant influence on long-term outcome.
Discussion
This study presents a large sample of patients from 17
Austrian centers who had severe TBI due to acute SDH.
Our findings show that acute SDH is the mass lesion with
the highest incidence (49%) in patients with severe TBI.
They also show that acute SDH is associated with high
mortality (47%) and a low rate of favorable long-term
outcomes (32%) in this group of patients. These poor results, however, were well within the ranges predicted by
the scores developed by Hukkelhoven et al.9 and by Maas
et al.15 Our study also shows that age, neurological status,
330
and severity of trauma are the main factors determining
outcomes after severe TBI due to acute SDH.
The results presented here, however, are actually better than those published by some other authors. Koç et
al.13 reported 60% mortality and 38% favorable outcome
for patients with acute SDH who needed surgery; that
study included patients with moderate TBI, whereas our
study included only patients with severe TBI. Hospital
mortality for patients who had acute SDH evacuation was
43% in our study. It seems that nearly all survivors in the
study by Koç et al.13 had favorable outcomes, and more
than one-third of the survivors had unfavorable outcomes
in our study. Hatashita et al.5 found an overall mortality of 55% and favorable outcome of 30% in surgically
treated patients with acute SDH; their study also included
patients with moderate TBI. These results are somewhat
closer to our own results than those of Koç et al.13 In
J Neurosurg / Volume 117 / August 2012
Outcome after severe TBI due to acute SDH
another study, Servadei et al.25 reported a 35.4% favorable outcome in 65 patients with severe TBI due to acute
SDH, and they managed 20% of the patients conservatively. These results are in accordance with our own experiences. According to Dent et al.2 only 24% of patients
with severe TBI due to acute SDH who required urgent
surgery had a favorable outcome, which is the worst result
published so far. However, it has to be noted that none of
the other authors provided data on the expected mortality of their patients, because no score had been available
at the times of their studies. Trauma severity and, thus,
expected mortality might have been different from that
seen in our study.
With regard to factors influencing outcomes, the
impact of GCS scores has been studied most frequently.
The most detailed analysis of the effects of GCS scores
on outcomes after severe TBI was done in the IMPACT
(International Mission for Prognosis and Analsysis of
Clinical Trials in TBI) study.16 It was shown that the
GCS score at hospital admission was strongly related to
the GOS score at 6 months after trauma (OR 1.7–7.5).
Hatashita et al.5 also found a strong correlation between
GCS scores and mortality: nearly all patients with acute
SDH and GCS scores of 3 died (93% mortality), patients
with GCS scores of 4–6 had a mortality of 45%–67%,
and all patients with GCS scores of 7 or higher survived.
These results are confirmed by the study by Koç et al.13
Gennarelli et al.3 published mortality rates of 74% for patients with acute SDH and GCS scores of 3–5, and 36%
for those with GCS scores of 6–8. This significant association between GCS scores and outcomes was also
found in our study.
Age is one of the most important factors influencing
survival as well as recovery after severe TBI, as demonstrated in the large study done by Hukkelhoven et al.10
They estimated that the odds for poor outcome increased
by 40%–50% per 10 years of age. With regard to patients
with acute SDH, Howard et al.7 reported that overall mortality was more than 4 times higher in patients older than
65 years than in those 18–40 years. In the study by Koç et
al.,13 age did not significantly influence outcome, although
their findings demonstrated that an older age was associated with increased mortality. The significant influence of
age on outcomes was also found in our study.
Koç et al.13 reported that patients with acute SDH
who presented with bilateral or unilateral unreactive pupils had mortality rates of 97% and 81%, respectively. In
the study by Marmarou et al.,16 one or both unreactive
pupils were significantly associated with poor outcome
(OR 2.71–7.31). With regard to hospital outcome, this association was also observed in our study; however, pupil
reactivity had no significant effect on long-term outcome.
Lefering et al.14 analyzed data from patients with
head injury from the German Trauma Registry and reported that patients with concomitant injuries with an
AIS greater than 3 had a 5% higher mortality rate than
patients without concomitant injuries. Heinzelmann et
al.6 came to different conclusions. They found that extracranial injuries had no significant effect on outcomes of
patients with epidural hematoma. In our study, additional
injuries also had no effect on outcomes.
J Neurosurg / Volume 117 / August 2012
With regard to CT findings, our study showed that
the PD predicted by the Rotterdam CT score correlated
significantly with the observed mortality rates. Unfortunately, this correlation could only be calculated for scores
2–6, as our study had no patient with a score of 1. The
Rotterdam CT score was much better suited for our group
of patients than the Marshall score, since our patients fell
into 2 of the 6 possible groups. The Rotterdam CT score
had significant effect on long-term outcome but not on
hospital mortality.
With regard to surgical management, Seelig et al.24
reported a mortality of 30% for comatose patients treated
less than 4 hours after injury versus 90% mortality for
those treated after 4 hours. Other authors, however, came
to different conclusions. Wilberger et al.28 found that mortality in patients with admission GCS scores less than
8 who underwent surgery within 4 hours of injury was
59% versus 69% for those who underwent surgery after 4
hours. They were not able to confirm a significant influence of timing of surgery on outcome. These findings are
supported by the studies done by Stone et al., 26 Hatashita
et al., 5 and Koç et al.13 Timing of surgery also had no
effect on outcome in our study. Craniotomy and craniectomy were associated with comparable mortality rates
in our study. This confirms earlier results from Paci et
al.21 However, these authors found higher mortality rates
with craniectomy, which is in contrast to our findings.
The multivariate analysis showed that conservative management of patients with acute SDH had significant effect on hospital mortality but not on long-term outcome.
However, this might reflect a conscious decision by the
medical team and the patients’ families not to treat the
most severely injured patients. The conservatively managed patients had the highest PD (0.47) of all patients with
acute SDH, while the PD of all surgically treated patients
was 0.41.
Another treatment aspect that may influence outcomes is direct transfer to a specialized neurosurgical
center. Stone et al.26 found that in patients with acute
SDH, mortality was significantly lower for those who had
been transported directly to their center (50% vs 76%).
More recently Härtl et al.4 reported that indirect transfer was associated with a 50% increase in mortality for
patients who had suffered severe TBI. We were unable
to demonstrate an effect of direct transfer on outcomes
in our study; mortality was actually lower in the patients
with indirect transfer. This, however, was not significant
in the logistic regression analysis. There are 2 possible
reasons for this surprising result. First, only patients who
were admitted to the ICU were enrolled into the first
study; patients who did not survive until ICU admission
might have been lost. Second, there could be a “selection
bias;” it might be that only patients who were considered
to have a chance of survival were transferred to the study
centers, and the patients in the worst condition might have
died at the primary hospital.
Limitations of the Study
The scores used to estimate PD and PU have not been
validated for our study population. These scores were
created from the international and North American data
331
J. Leitgeb et al.
from the tirilazad trial8,18 and were validated against the
core data set of the EBIC (European Brain Injury Consortium) survey20 and data from the Traumatic Coma Data
Bank.17 It is quite likely that our patients are comparable
to those from the EBIC centers and the international arm
of the tirilazad trial. There could, however, be subtle differences, and the observed/expected ratios for mortality
and unfavorable outcome estimated for our groups of patients may be incorrect. The use of these scores seemed
justified, however, because interpretation of our results
would be difficult without comparison with the predictions.
Conclusions
Acute SDH was found in almost half of the patients
with severe TBI. Forty-seven percent of the patients
died in the hospital, 19% survived with unfavorable outcomes, and 32% had favorable outcomes. These results
were within the predicted ranges. Age, severity of TBI,
and neurological status were the main factors influencing morbidity and mortality. Nonoperative management
was associated with significantly higher mortality. The
mortality predictions made by the Rotterdam CT score
correlated well with the observed mortality rates.
Disclosure
The data used for this study were collected for a project
funded by the Austrian Workers’ Compensation Board (AUVA; FK
33/2003) and by the “Jubilee Fund” of the Austrian National Bank
(Project 8987), and for a project funded by the Ministry of Health
(Contract October 15, 2008) and the AUVA (FK 11/2008 and FK
11/2010). INRO is supported by an annual grant from Mrs. Ala
Auersperg-Isham and Mr. Ralph Isham, and by small donations from
various sources.
Author contributions to the study and manuscript preparation
include the following. Conception and design: Leitgeb, Mauritz.
Acquisition of data: Leitgeb, Brazinova. Analysis and interpretation of data: Leitgeb, Janciak. Drafting the article: Leitgeb, Majdan.
Critically revising the article: Leitgeb, Mauritz, Brazinova, Majdan,
Wilbacher, Rusnak. Reviewed submitted version of manuscript:
Leitgeb, Mauritz, Brazinova, Majdan, Wilbacher, Rusnak. Approved
the final version of the manuscript on behalf of all authors: Leitgeb.
Administrative/technical/material support: Wilbacher. Study supervision: Rusnak.
Acknowledgments
The authors are very grateful to the investigators from the
participating centers: H. Artmann, M.D. (Schwarzach), N. Bauer,
M.D. (Linz UKH), F. Botha, M.D. (Linz WJ), F. Chmeliczek, M.D.
(Salzburg LKA), G. Clarici, M.D. (Graz Uni), D. Csomor, M.D.
(Wr. Neustadt), R. Folie, M.D. (Feldkirch), R. Germann, M.D.,
Ph.D. (Feldkirch), F. Gruber, M.D. (Linz AKH), H.-D. Gulle, M.D.
(Klagenfurt), T. Haidacher, M.D. (Graz UKH), G. Herzer, M.D.
(Wr. Neustadt), P. Hohenauer, M.D. (Salzburg LKA), A. Hüblauer,
M.D. (Horn), J. Lanner, M.D. (Salzburg UKH), V. Lorenz, M.D.
(Wien UKH XII), C. Mirth, M.D. (St. Pölten), W. Mitterndorfer,
M.D. (Linz AKH), W. Moser, M.D. (Klagenfurt), H. Schmied,
M.D. (Amstetten), K.-H. Stadlbauer, M.D., Ph.D. (Innsbruck), H.
Steltzer, M.D., Ph.D. (Wien UKH XII), Ernst Trampitsch, M.D.
(Klagenfurt), A. Waltensdorfer, M.D. (Graz Uni), and A. Zechner,
M.D. (Klagenfurt).
332
References
1. Bullock R, Chesnut RM, Clifton G, Ghajar J, Marion DW, Narayan RK, et al: Guidelines for the management of severe head
injury. Eur J Emerg Med 3:109–127, 1996
2. Dent DL, Croce MA, Menke PG, Young BH, Hinson MS,
Kudsk KA, et al: Prognostic factors after acute subdural hematoma. J Trauma 39:36–43, 1995
3. Gennarelli TA, Spielman GM, Langfitt TW, Gildenberg PL,
Harrington T, Jane JA, et al: Influence of the type of intracranial lesion on outcome from severe head injury. J Neurosurg
56:26–32, 1982
4. Härtl R, Gerber LM, Iacono L, Ni Q, Lyons K, Ghajar J: Direct transport within an organized state trauma system reduces mortality in patients with severe traumatic brain injury. J
Trauma 60:1250–1256, 2006
5. Hatashita S, Koga N, Hosaka Y, Takagi S: Acute subdural hematoma: severity of injury, surgical intervention, and mortality. Neurol Med Chir (Tokyo) 33:13–18, 1993
6. Heinzelmann M, Platz A, Imhof HG: Outcome after acute extradural haematoma, influence of additional injuries and neurological complications in the ICU. Injury 27:345–349, 1996
7. Howard MA III, Gross AS, Dacey RG Jr, Winn HR: Acute
subdural hematomas: an age-dependent clinical entity. J Neu
rosurg 71:858–863, 1989
8. Hukkelhoven CW, Steyerberg EW, Farace E, Habbema JD,
Marshall LF, Maas AI: Regional differences in patient characteristics, case management, and outcomes in traumatic brain
injury: experience from the tirilazad trials. J Neurosurg 97:
549–557, 2002
9. Hukkelhoven CW, Steyerberg EW, Habbema JD, Farace E,
Marmarou A, Murray GD, et al: Predicting outcome after
traumatic brain injury: development and validation of a prognostic score based on admission characteristics. J Neurotrau
ma 22:1025–1039, 2005
10. Hukkelhoven CW, Steyerberg EW, Rampen AJ, Farace E,
Habbema JD, Marshall LF, et al: Patient age and outcome following severe traumatic brain injury: an analysis of 5600 pa
tients. J Neurosurg 99:666–673, 2003
11. Hyder AA, Wunderlich CA, Puvanachandra P, Gururaj G, Kobusingye OC: The impact of traumatic brain injuries: a global
perspective. NeuroRehabilitation 22:341–353, 2007
12. Jennett B, Bond M: Assessment of outcome after severe brain
damage. Lancet 1:480–484, 1975
13. Koç RK, Akdemir H, Oktem IS, Meral M, Menkü A: Acute
subdural hematoma: outcome and outcome prediction. Neu
rosurg Rev 20:239–244, 1997
14. Lefering R, Paffrath T, Linker R, Bouillon B, Neugebauer EA:
Head injury and outcome—what influence do concomitant injuries have? J Trauma 65:1036–1044, 2008
15. Maas AI, Hukkelhoven CW, Marshall LF, Steyerberg EW:
Prediction of outcome in traumatic brain injury with computed tomographic characteristics: a comparison between
the computed tomographic classification and combinations of
computed tomographic predictors. Neurosurgery 57:1173–
1182, 2005
16. Marmarou A, Lu J, Butcher I, McHugh GS, Murray GD, Steyerberg EW, et al: Prognostic value of the Glasgow Coma Scale
and pupil reactivity in traumatic brain injury assessed prehospital and on enrollment: an IMPACT analysis. J Neuro
trauma 24:270–280, 2007
17. Marshall LF, Becker DP, Bowers SA, Cayard C, Eisenberg H,
Gross CR, et al: The National Traumatic Coma Data Bank.
Part 1: Design, purpose, goals, and results. J Neurosurg 59:
276–284, 1983
18. Marshall LF, Maas AI, Marshall SB, Bricolo A, Fearnside M,
Iannotti F, et al: A multicenter trial on the efficacy of using
tirilazad mesylate in cases of head injury. J Neurosurg 89:
519–525, 1998
J Neurosurg / Volume 117 / August 2012
Outcome after severe TBI due to acute SDH
19. Marshall LF, Marshall SB, Klauber MR, van Berkum Clark
M, Eisenberg HM, Jane JA, et al: A new classification of head
injury based on computerized tomography. J Neurosurg 75:
S14–S20, 1991
20. Murray GD, Teasdale GM, Braakman R, Cohadon F, Dearden
M, Iannotti F, et al: The European Brain Injury Consortium
survey of head injuries. Acta Neurochir (Wien) 141:223–
236, 1999
21. Paci GM, Sise MJ, Sise CB, Sack DI, Shackford SR, Kureshi
SA, et al: Preemptive craniectomy with craniotomy: what role
in the management of severe traumatic brain injury? J Trau
ma 67:531–536, 2009
22. Phuenpathom N, Choomuang M, Ratanalert S: Outcome and
outcome prediction in acute subdural hematoma. Surg Neu
rol 40:22–25, 1993
23. Rusnak M, Janciak I, Majdan M, Wilbacher I, Mauritz W: Severe traumatic brain injury in Austria VI: effects of guideline-based management. Wien Klin Wochenschr 119:64–71,
2007
24. Seelig JM, Becker DP, Miller JD, Greenberg RP, Ward JD,
Choi SC: Traumatic acute subdural hematoma: major mortality reduction in comatose patients treated within four hours. N
Engl J Med 304:1511–1518, 1981
25. Servadei F, Nasi MT, Cremonini AM, Giuliani G, Cenni P,
Nanni A: Importance of a reliable admission Glasgow Coma
Scale score for determining the need for evacuation of post-
J Neurosurg / Volume 117 / August 2012
traumatic subdural hematomas: a prospective study of 65 patients. J Trauma 44:868–873, 1998
26. Stone JL, Lowe RJ, Jonasson O, Baker RJ, Barrett J, Oldershaw JB, et al: Acute subdural hematoma: direct admission to
a trauma center yields improved results. J Trauma 26:445–
450, 1986
27. Tagliaferri F, Compagnone C, Korsic M, Servadei F, Kraus J:
A systematic review of brain injury epidemiology in Europe.
Acta Neurochir (Wien) 148:255–268, 2006
28. Wilberger JE Jr, Harris M, Diamond DL: Acute subdural hematoma: morbidity, mortality, and operative timing. J Neuro
surg 74:212–218, 1991
29. Yanaka K, Kamezaki T, Yamada T, Takano S, Meguro K,
Nose T: Acute subdural hematoma—prediction of outcome
with a linear discriminant function. Neurol Med Chir (To
kyo) 33:552–558, 1993
Manuscript submitted August 26, 2011.
Accepted April 23, 2012.
Please include this information when citing this paper: published online May 25, 2012; DOI: 10.3171/2012.4.JNS111448.
Address correspondence to: Johannes Leitgeb, M.D., Department of Traumatology, Medical University of Vienna, Währinger
Gürtel 18-20, A-1090 Vienna, Austria. email: johannes.leitgeb@
meduniwien.ac.at.
333