Injury, Int. J.

Care Injured 44 (2013) 606–610

Contents lists available at SciVerse ScienceDirect

Injury
journal homepage: www.elsevier.com/locate/injury

Redefining the golden hour for severe head injury in an urban setting:
The effect of prehospital arrival times on patient outcomes
Michael M. Dinh a,*, Kendall Bein b, Susan Roncal c, Christopher M. Byrne a,
Jeffrey Petchell a, Jeffrey Brennan d
a
Royal Prince Alfred Hospital, Trauma Office level 10, Missenden Road, Camperdown, NSW 2050, Australia
b
Royal Prince Alfred Hospital, Emergency Department, Missenden Road, Camperdown, NSW 2050, Australia
c
Department of Trauma Services, Royal Prince Alfred Hospital, Missenden Road, Camperdown NSW 2050, Australia
d
Royal Prince Alfred Hospital, Missenden Road, Camperdown, NSW 2050, Australia

A R T I C L E I N F O A B S T R A C T

Article history: Background: In patients with severe head injuries, transportation to a trauma centre within the ‘‘golden
Accepted 13 January 2012 hour’’ are important markers of trauma system effectiveness but evidence regarding impacts on patient
outcomes is limited.
Keywords: Objective: To determine the effect of patient arrival within the golden hour on patient outcomes.
Head injury Methods: A retrospective cohort of adult patients with severe head injuries (head AIS  3) arriving
Prehospital within 24 h of injury was identified using the trauma registry from 2000 to 2011. Survival analysis was
Time
used to determine the effect of patient arrival time on overall mortality. Study outcomes were in hospital
Patient outcome
mortality and survival to hospital discharge without requiring transfer for ongoing rehabilitation or
nursing home care.
Results: There was a significant association with mortality with each incremental minute of patient
arrival (HR 1.002, 95%CI 1.001–1.004, p = 0.001). There was however no survival benefit observed for
patients arriving within 60 min of injury time (HR 0.77, 95%CI 0.50–1.18, p = 0.22) but an apparent
benefit for those presenting within 2 h of injury time (HR 0.31, 95%CI 0.15–0.66, p = 0.002). Patient
arrival within 60 min of injury time was associated with increased odds of survival to hospital discharge
without requiring ongoing rehabilitation (OR 1.78, 95%CI 1.14–2.79, p = 0.01).
Conclusion: A survival benefit exists in patients arriving earlier to hospital after severe head injury but
the benefit may extend beyond the golden hour. There was evidence of improved functional outcomes in
patients arriving within 60 min of injury time.
Crown Copyright ß 2012 Published by Elsevier Ltd. All rights reserved.

Introduction
Trauma systems are largely designed and implemented around
the capability of providing timely access to definitive care. This is
predicated on the principle that time is a critical factor in
determining the outcome of an injured patient. One of the most
well known principles in medicine is the ‘‘golden hour’’ of trauma,
which specifies that patient outcomes are improved when patient
transport to a designated trauma centre is completed within an
hour of injury.1
Although it would seem intuitive that treating patients more
rapidly results in reduced mortality and secondary injury, the
* Corresponding author. Tel.: +61 02 9515 6111.
E-mail addresses: Dinh.mm@gmail.com (M.M. Dinh), kendallbein@tpg.com.au
(K. Bein), roncal@email.cs.nsw.gov.au (S. Roncal), chrismbyrne@hotmail.com
(C.M. Byrne), jfpetchell@aapt.net.au (J. Petchell), jeffbrennan@westnet.com.au
(J. Brennan).
evidence for the golden hour in general and timing of craniotomies
for head injuries on patient outcomes is limited.2–6 A recent large
multicentre study in the US did not find any association between
prehospital transport time and mortality in all trauma patients.6
Some studies have even suggested that decreasing time to
craniotomy is associated with increased patient mortality.7,8 Such
studies may be limited by length and selection biases associated
with patients who survive to undergo delayed craniotomies. A
Canadian trauma centre study of 149 patients with acute traumatic
subdural haemorrhage found a trend to improved survival in
patients transported within 1 h of injury.9
Confirmation of such findings in more a more general group of
patients with severe head injuries, where timing is thought to be
crucial, may have important implications for trauma system
design and clinical approaches.10 The objective of the present
study was to determine the effect of prehospital time on patient
outcomes in a group of urban trauma patients with severe head
injuries.

0020–1383/$ – see front matter . Crown Copyright ß 2012 Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.injury.2012.01.011
 M.M. Dinh et al. / Injury, Int. J. Care Injured 44 (2013) 606–610
Methods
The study was conducted at a Major Trauma Centre in New
South Wales Australia treating around 3000 trauma admissions a
year, of which 250 are defined as major trauma (injury severity
score > 15). After institutional human research ethics approval,
the hospital trauma registry was queried for all adult (age 15 years
or older) major trauma admissions with severe head injury (head
abbreviated injury score  3) due to blunt trauma between January
2000 and June 2011. Exclusion criteria were patients transferred
from other health facilities, injuries occurring more than 24 h prior
to hospital presentation, patients who self presented or did not
come by ambulance and patients with no vital signs on arrival.
Patients with associated spinal injuries transferred to other spinal
trauma hospitals for ongoing care were also excluded.
Demographic and injury characteristics were obtained, includ-
ing the time of incident routinely recorded on ambulance sheets,
triage time from emergency department records and time of any
craniotomy from operation notes. Patients were classified as
having an urgent craniotomy if one was performed within 24 h of
injury time. The patient arrival time was defined as the number of
minutes from recorded incident time to triage time. Patients who
arrived to the emergency department within 60 min or less of
injury time were classified ‘‘Early’’ whilst the patients arriving after
60 min were classified as ‘‘Delayed’’.
For the purpose of multivariable analyses vital signs were
categorised into clinically relevant groups of Glasgow coma score
(GCS 3–8, 9–13 and 14–15) and systolic blood pressure (SBP < 90
and >90 mmHg). Vital signs were obtained from the first recorded
observations after arrival to the emergency department and GCS
calculated regardless of intubation status on arrival (Best verbal
response = 1 for intubated patients). Patients with large extradural,
subdural or intracerebral haemorrhages (both anterior and
posterior fossa) were defined by abbreviated injury score (AIS)
codes as those with depth greater than 1 cm on initial CT scan. All
head CT scans for trauma were reviewed and reported by a
Radiologist and entered prospectively into the trauma data
registry. Patient disposition (transfer to rehabilitation, other
hospital or nursing home or discharge for follow up), from the
hospital was also recorded.
The primary outcome was in-hospital mortality analysed as a
time dependent outcome with respect to observed length of stay.
The secondary outcome was survival to hospital discharge without
requiring transfer for ongoing rehabilitation or nursing home care.
Patient management
Patients were managed using standardised severe head injury
algorithms based on adult trauma life support principles.
Performance indicators that were routinely assessed as part of a
rigorous quality assurance programme included prehospital scene
time of 20 min or less, definitive airways management within
10 min of arrival, CT scanning within 1 h of arrival and urgent
craniotomies within 4 h of injury time, based on current American
College of Surgeon quality indicators.
Statistical analysis
Univariate analyses were used to describe crude differences in
baseline characteristics and outcomes in patients who did or did
not meet arrival benchmarks (less than 60 min from injury time)
and craniotomy time benchmarks (less than 4 h from injury time).
Means were described using standard deviations (SD) and medians
with interquartile range (IQR). Multivariable Cox proportional
hazards models were used to adjust for a priori confounders such
as age, GCS categories, injury severity scores (ISS), hypotension and
607
need for craniotomy within 24 h of injury. All other covariates
were entered into the model using stepwise selection algorithm
with an entry criteria p value of 0.10. To investigate the effect of
increasing patient arrival time, the same model was repeated using
dummy variables for different patient arrival to injury times (30,
60, 90, 120 and 180 min) as well as arrival time on a continuous
scale. A Kaplan–Meier estimator plot was used to compare survival
functions and Schoenfields residuals were used to test the
proportional hazards assumption. Analysis was performed on
SAS 9.2 (SAS Institute, Cary NC, USA) and significance was defined
as a two tailed p value of less than 0.05.
Results
Between 2000 and 2011, 1550 adult patients were identified
with major trauma and severe head injury (AIS  3) from the
hospital trauma registry. Of these 1034 patients fulfilled inclusion
criteria, being patients transferred directly from scene within 24 h
of injury. A further 51 records were excluded due to missing values
for incident date or time. Therefore 983 cases were available for
analysis of patient arrival time.
Of the 983 cases with severe head injury, 192 patients (20%)
arrived after 60 min and 40 patients (4%) arrived after 120 min.
Males comprised 72% of the study population and the mean age
was 51 years (SD 23 years). The most common mechanisms of
injury were falls (50%), pedestrians (16%), assaults (12%) and motor
vehicle incidents (11%). The mean patient arrival time was 60 min
(SD 100 min) (median 43 min IQR 33–56) from injury. Airway
intubation was required in 47% of cases. The median ISS was 21
(IQR 17–26) and 21% had a large subdural or extradural
haemorrhage. One hundred and fifty-five patients (16%) had a
craniotomy performed within 24 h of injury.
Overall in-hospital mortality was 15% with 61% of patients
requiring intensive care admission. Survival to hospital discharge
without requiring transfer for rehabilitation or nursing home care
occurred in 202 (21%) patients. Table 1 compares baseline
characteristics and outcomes in ‘‘early’’ and ‘‘delayed’’ patient
arrival to the emergency department. Patients in the delayed group
were older and associated with more falls as the mechanism of
injury. Patients in the early group were associated with lower GCS
scores and were more likely to require airway intubation.
Fig. 1 shows the Kaplan–Meier overall unadjusted survival
curves for the two study groups. In patients arriving within 60 min
unadjusted 30-day survival was higher (81%, 95%CI 78–85%)
compared to those arriving after 60 min (74%, 95%CI 60–84%) but
did not reach statistical significance (log rank p = 0.61).
Using multivariable Cox proportional hazards model across
strata of GCS (GCS covariate did not meet the proportional hazards
assumption), there was a rise in mortality with each incremental
increase in patient arrival time in minutes (HR 1.002, 95%CI 1.001–
1.004, p = 0.001). This was adjusted using a stepwise selection
algorithm which included into the final model age, presence of
hypotension, GCS categories, airway intubation, large intracranial
haemorrhage and whether a craniotomy was performed or not.
When divided into relevant time intervals using the same model,
there was no apparent association between mortality and the early
patient arrival group (arriving within 60 min of injury) compared
to delayed arrival (HR 0.77 95%CI 0.50–1.18 p = 0.22) (see Tables 2a
and 2b). There was no association between arrival within 30 min
and mortality (HR 1.15, 95%CI 0.75–1.77, p = 0.51). There appeared
to be a survival benefit for patients arriving within 90 min (HR
0.35, 95%CI, 0.18–0.65 p = 0.001) and 120 min of injury (HR 0.30,
95%CI 0.16–0.64, p = 0.002).
Using the same covariates above in a multivariable logistic
regression model, the odds of survival to hospital discharge
without requiring transfer for ongoing rehabilitation or nursing
608 M.M. Dinh et al. / Injury, Int. J. Care Injured 44 (2013) 606–610

Table 1
Comparison of baseline characteristics in early and delayed patient arrivals (>60 min). IQR, interquartile range.

Early patient arrival (60 min) N = 791 Delayed patient arrival (>60 min) N = 192 Significance

Age mean (sd) years 49.8 (22.6) 56.9 (23.4) <0.001

15–34 years 268 (34) 52 (27)
35–64 years 276 (35) 54 (28)
65 years 247 (31) 86 (45) 0.001
Male (%) 576 (73) 131 (68) 0.20
Falls (%) 364 (46) 124 (64) <0.001
Systolic blood pressure mean mmHg (sd) 134 (29) 139 (35) 0.03
Glasgow coma scale
14–15 463 (59) 150 (78)
9–13 111 (14) 17 (9)
3–8 216 (27) 25 (13) <0.001
Injury severity score (median IQR) 21 (17–26) 21 (17–25) 0.08
<25 531 (67) 145 (76)
25–50 229 (29) 45 (23)
>50 31 (4) 2 (1) 0.03
Airway intubation (%) 319 (49) 68 (35) 0.001
Intensive care admission (%) 516 (65) 98 (51) <0.001
Craniotomya 152 (19) 32 (17) 0.42
Large extradural haemorrhage 48 (6) 5 (3) 0.06
Large subdural haemorrhage (%) 131 (17) 32 (17) 0.97
Large intracranial haemorrhage (%) 33 (4) 7 (4) 0.74
Death (%) 117 (15) 30 (16) 0.77
Survival without rehabilitation or nursing home (%) 172 (22) 30 (16) 0.06
a
Craniotomy performed within 24 h of injury time.

Table 2a
Univariable and multivariable models for in-hospital mortality. Multivariable derived from stepwise selection including patient arrival time, age, injury severity score (ISS)
across strata of Glasgow coma scale (GCS) which did not meet proportional hazards assumption. ICH, intracerebral haemorrhage, Covariates excluded during stepwise
selection include sex, mechanism of injury, extradural haemorrhage > 1 cm and subdural haemorrhage > 1 cm.

Covariate Patients N = 983 Deaths n = 147 Univariable analysis Multivariable analysis

HR (95%CI) p-Value HR (95%CI) p-Value

Patient arrival time

>60 min 192 30 1 [Reference] 1 [Reference]
60 min 791 117 0.91 (0.61–1.35) 0.63 0.77 (0.50–1.18) 0.22
Age
15–34 years 319 34 1 [Reference] 1 [Reference]
35–65 years 330 36 0.95 (0.60–1.52) 0.84 1.19 (0.73–1.92) 0.48
65 years 333 76 2.17 (1.45–3.26) <0.001 4.96 (3.18–7.74) <0.001
SBP
90 mmHg 927 117 1 [Reference] 1 [Reference]
<90 mmHg) 56 30 5.84 (3.90–8.73) <0.001 3.89 (2.47–6.07) <0.001
GCS
14–15 613 34 1 [Reference] NA
9–13 128 18 2.26 (1.28–4.01) 0.005 NA
3–8 241 95 7.36 (4.96–10.92) <0.001 NA
ISS
16–25 676 46 1 [Reference] 1 [Reference]
25–50 274 79 3.88 (2.69–5.61) <0.001 1.88 (1.29–2.76) 0.001
>50 33 22 13.58 (8.16–22.62) <0.001 4.90 (2.84–8.52) <0.001
Airway intubation
No 524 19 1 [Reference] 1 [Reference]
Yes 459 128 7.02 (4.32–11.39) <0.001 2.56 (1.45–4.54) 0.001
ICH
1 cm 943 126 1 [Reference] 1 [Reference]
>1 cm 21 40 4.14 (2.60–6.59) <0.001 1.84 (1.13–3.00) 0.01
Craniotomy < 24 h
No 828 106 1 [Reference] 1 [Reference]
Yes 155 41 1.83 (1.27–2.63) 0.001 0.98 (0.66–1.44) 0.90

Table 2b
Adjusted HR for mortality comparing patients arriving within 60 min across relevant strata of Glasgow coma scale (GCS) NB one missing value for GCS.

GCS n HR (early arrival) 95%CI p-Value

9 741 0.87 0.46–1.66 0.67

3–8 241 0.80 0.44–1.45 0.47
 M.M. Dinh et al. / Injury, Int. J. Care Injured 44 (2013) 606–610
Fig. 1. Patient survival curves by arrival time within 60 min. Unadjusted log rank test p = 0.61.
home care were 1.78 times higher in patients arriving within
60 min of injury (OR 1.78, 95%CI 1.14–2.79, p = 0.01).
Discussion
This was a retrospective study of major trauma patients with
severe head injury who arrived directly to this trauma centre
within 24 h of injury. The results confirm a survival benefit for
patients arriving earlier from time of injury with each minute of
delay resulting in around 0.2% increase in hazard of mortality.
Interestingly this survival benefit did not become statistically
significant within 60 min of arrival, even when the analysis was
repeated across clinically important strata of GCS. Rather it appears
that the benefit likely extends beyond the golden hour with
decreased mortality within 90 and 120 min demonstrated. There
was also a potential benefit in functional outcomes in patients
being transported within 60 min of injury time, with increased
odds of survival to hospital discharge without requiring transfer
from hospital for ongoing rehabilitation.
The finding of improved survival with decreased prehospital
transport time is consistent with similar studies examining
isolated acute subdural haemorrhages.9 The study does therefore
reaffirm the general principle of the golden hour, at least for severe
head injury. This has potential implications for prehospital and
trauma system organisation, particularly in rural areas where
patients frequently required longer transport times.
It is unclear from the present study how earlier patient arrival
improves mortality. One possible explanation could be that the
patients in the delayed group were older with under-recognised
injuries and comorbidities that contributed to poorer outcomes.
The study did adjust for age and injury severity scores but the
residual confounding is possible. The survival benefit of earlier
transfer may not be related to patients having craniotomies within
24 h of injury as this was not significantly related to mortality in
this multivariable model. This is in contrast to studies that have
shown survival benefits for patients undergoing neurosurgical
intervention within 24 h of injury.11 Proposed pathways for the
survival benefit include earlier stabilisation, airway management
and resuscitation in the prehospital phase, emergency depart-
ments and intensive care units that may help minimise the impact
of secondary brain injury. Recent studies examining airway
609
intubation in the field have however not demonstrated improved
mortality with this specific prehospital intervention.12
There are several acknowledged limitations to this study. This
was a single centre trauma registry study with relatively low
volumes of major trauma. Nevertheless the overall mortality for
this population of patients with severe head injuries is comparable
if not lower than those quoted in other studies.13–15 The sample
size was imbalanced and as the number of patient arriving beyond
120 min was small (n = 40) conclusions regarding the survival
benefit in this group are limited without a confirmatory study. It is
also likely that the study was under powered to detect a significant
adjusted HR of 0.7 observed for the patients arriving within 60 min
of injury time. More reliable results would be obtained from state-
wide data using the same adjusting methods as described above.
Importantly we were unable to control for evidence of raised
intracranial pressure, other than depressed GCS (GCS < 9). Raised
intracranial pressure is an important predictor of outcome in head
injury.16 We used craniotomy within 24 h as a proxy for clinical
urgency which would include evidence of raised intracranial
pressure. However, we did not adjust for changes in GCS
preoperatively which would also be an important indicator of
herniation and would have influenced the decision for an urgent
craniotomy. The use of head AIS  3 as the criteria for severe head
injury was consistent with previous studies11 and included many
patients with small subdural haemorrhages and other intracranial
bleeds that were managed non-operatively. A further study is
required to analyse the effect of craniotomy timing on patients
who required an urgent neurosurgical operation.
We also did not examine or control for the effects of
anticoagulation, medical comorbidities, presence of seizures and
open skull fractures and performance of craniectomy as opposed to
craniotomy, all of which can affect prognosis. The cause of death
(neurological or other) was not used in this study due to the
potential for recall and misclassification bias. An interesting line of
further inquiry may be to investigate the effect of operative timing
on the rate of rebleeding and reoperation, as well as the effect of
changes in size of intracranial lesions and after hour’s patient
arrival on craniotomy timing, the success of non-operative
management and patient outcomes.17 Finally, the outcome of
survival without requiring rehabilitation was a crude measure of
functional outcome and was not adjusted for premorbid functional
state or nursing home care. We are presently unaware of other
610 M.M. Dinh et al. / Injury, Int. J. Care Injured 44 (2013) 606–610
studies that have demonstrated this relationship in patients with
severe head injury. Further studies could utilise more accurate
measures of health status and neurological state.18,19
In conclusion we found that in a group of patients with severe
head injuries, there was a survival benefit for patients arriving
earlier from time of injury. This survival benefit may extend beyond
the golden hour as no difference in mortality was observed for arrival
within 60 min of injury time. There was however a benefit with
respect to survival without requiring ongoing rehabilitation
observed in patients arriving within 60 min of injury time.
Funding
No funding was obtained for this study.
Conflict of interest statement
No conflict of interest was declared.
Author contributions
MD was responsible for study design ethics and data analysis
and manuscript preparation.
SR contributed towards data abstraction.
KB performed data coding and analysis.
CB, JP and JB were responsible for manuscript preparation and
review.
References
1. Trunkey DD. Trauma. Sci Am 1983;249:28–35.
2. Stelfox HT, Bobranska-Artiuch B, Nathens A, Straus S. Quality indicators for
evaluating trauma care. Arch Surg 2010;145(3):286–95.
3. Lerner EB, Moscati RM. The golden hour. Scientific fact or medical urban legend.
Acad Emerg Med 2001;8:758–60.
4. Evans C, Howes D, Pickett W, Dagnone L. Audit filters for improving processes of
care and clinical outcomes in trauma systems. Cochrane Datab Syst Rev
2009;(4):CD007590.
5. Willis CD, Stoelwinder JU, Cameron PA. Interpreting process indicators in
trauma. Construct validity versus confounding by indication. Int J Qual Health
Care 2008;20(5):331–8.
6. Newgard CD, Schmicker RH, Hedges JR, Trickett JP, Davis DP, Bulger EM, et al.
Emergency medical service intervals and survival in trauma: assessment of the
golden hour in a North American prospective cohort. Ann Emerg Med
2009;55(3):235–46.
7. Dent DL, Croce MA, Menke PG. Prognostic factors after acute subdural haema-
toma. J Trauma 1995;39(1):36–42.
8. Tian HL, Chen SW, Xu T. Risk factors related to hospital mortality in patients
with isolated traumatic acute subdural haematoma. Analysis of 308 patients
undergone surgery. Chin Med J 2008;20:1080–4.
9. Tien HCN, Jung V, Pinto R, Mainprize T, Scales DC, Rizoli SB. Reducing time to
treatment decreases mortality of trauma patients with acute subdural haema-
toma. Ann Surg 2011;253:1178–83.
10. Esposito TJ, Reed L, Gamelli RL, Luchette FA. Neurosurgical coverage. Essential
desired or irrelevant for good patient care and trauma centre status. Ann Surg
2005;242(3):364–74.
11. Hedges JR, Newgard CD, Veum-Stone J, Selden NR, Adams AL, Diggs BS, et al.
Early neurosurgical procedures enhance survival in blunt head injury. Propen-
sity score analysis. J Emerg Med 2009;37(2):115–23.
12. Davis D, Hoyt D, Ochs M. The effect of paramedic rapid sequence intubation on
outcome in patients with severe traumatic brain injury. J Trauma 2003;54:
444–53.
13. Alkhoury F, Courtney J. Outcomes after severe head injury. A national trauma
databank based comparison of level I and level II trauma centres. Am Surg
2011;77(3):11–280.
14. Utomo WK, Gabbe BJ, Simpson PM, Cameron PA. Predictors of in-hospital
mortality and 6 month functional outcomes in older adults with severe
traumatic brain injury. Injury 2009;40(9):973–7.
15. Farahvar A, Gerber LM, Chiu YL, Hartl R, Carney N, Ghajar J. Response
to intracranial hypertension treatment as a predictor of death in patients
with severe traumatic brain injury. J Neurosurg 2011;114(May (5)):
1471–8.
16. Todd Miller M, Pasquale M, Kurek S, White J, Partin P, Bannon K. Initial head
computed tomography scan characteristics have a linear relationship with
initial intracranial pressure. J Trauma 2004;56:967–73.
17. De Souza M, Moncure M, Lansford T, Albaugh A, Tarnoff M, Goodman M. Non
operative management of epidural haematomas and subdural haematomas:
is it safe in lesions measuring 1 cm or less? J Trauma 2007;63:370–2.
18. Andelic N, Hammergren N, Bautz-Holter E, Sveen U, Brunborg C, Roe C.
Functional outcome and health related quality of life 10 years after moderate
to severe traumatic brain injury. Acta Neurol Scand 2009;120:16–23.
19. Cremer OL, Moons KG, Van Dijk GW, van Balen P, Kalkman CJ. Prognosis
following severe head injury: development and validation of a model for
prediction of death, disability and functional recovery. J Trauma 2006;61:
1484–91.