ETHICAL MATTERS

Factors Associated with the Withdrawal of Life-SustainingTherapies in Patients with Severe Traumatic Brain Injury:A Multicenter Cohort Study

Nicolas Cote • Alexis F. Turgeon • Francois Lauzier • Lynne Moore •

Damon C. Scales • Francis Bernard • Ryan Zarychanski • Karen E. A. Burns •

Maureen O. Meade • David Zygun • Jean-Francois Simard • Amelie Boutin •

Jacques G. Brochu • Dean A. Fergusson

Published online: 26 October 2012

� Springer Science+Business Media New York 2012

Abstract

Purpose To identify factors associated with decisions to

withdraw life-sustaining therapies in patients with severe

traumatic brain injury (TBI).

Materials and Methods We conducted a 2-year multicenter

retrospective cohort study (2005–2006) in mechanically ven-

tilated patients aged 16 years and older admitted to the

intensive care units (ICUs) of six Canadian level I trauma

centers following severe TBI. One hundred and twenty charts

were randomly selected at each center (n = 720). Data on ICU

management strategies, patients’ clinical condition, surgical

procedures, diagnostic imaging, and decision to withdraw life-

sustaining therapies were collected. The association of factors

pertaining to the injury, interventions, and management strat-

egies with decisions to withdraw life-sustaining therapies was

evaluated among non-survivors.

Results Among the 228 non-survivors, 160 died following

withdrawal of life-sustaining therapies. Patients were pre-

dominantly male (69.7 %) with a mean age of 50.7 (±21.7)

years old. Brain herniation was more often reported in

patients who died following decisions to withdraw life-sus-

taining therapies (odds ratio [OR] 2.91, 95 % confidence

interval [CI] 1.16–7.30, p = 0.02) compared to those who

This study was performed in six level I trauma centers in Canada:

Quebec (Centre Hospitalier Affilie Universitaire de Quebec—Hopital

de l’Enfant-Jesus, Hopital du Sacre-Coeur de Montreal), Ontario

(Hamilton General Hospital, Sunnybrook Health Science Centre,

St-Michael’s Hospital); Alberta (Foothill Medical Centre).

N. Cote � A. F. Turgeon � F. Lauzier � J. G. Brochu

Division of Critical Care Medicine, Department of

Anesthesiology, Universite Laval, Quebec, QC, Canada

A. F. Turgeon (&) � F. Lauzier � L. Moore � J.-F. Simard �A. Boutin

Centre de Recherche FRQ-S du Centre Hospitalier Affilie

Universitaire de Quebec (Hopital de l’Enfant-Jesus),

Traumatologie—Urgence—Soins Intensifs (CHA-Research

Center [Enfant-Jesus Hospital], Trauma—Emergency—Critical

Care Medicine Unit), Universite Laval, 1401, 18e rue,

Local H-012a, Quebec, QC G1J 1Z4, Canada

e-mail: Alexis.Turgeon@fmed.ulaval.ca

F. Lauzier

Department of Medicine, Universite Laval,

Quebec, QC, Canada

L. Moore

Department of Social and Preventive Medicine,

Universite Laval, Quebec, QC, Canada

D. C. Scales

Interdepartmental Division of Critical Care Medicine,

University of Toronto, Toronto, ON, Canada

F. Bernard

Department of Internal Medicine, Universite de Montreal,

Montreal, QC, Canada

R. Zarychanski

Department of Internal Medicine, Section of Critical Care

Medicine, University of Manitoba, Winnipeg, MB, Canada

K. E. A. Burns

Department of Critical Care Medicine, St-Michael’s Hospital,

University of Toronto, Toronto, ON, Canada

M. O. Meade

Department of Critical Care Medicine, McMaster University,

Hamilton, ON, Canada

D. Zygun

Department of Critical Care Medicine, University of Calgary,

Calgary, AB, Canada

D. A. Fergusson

Clinical Epidemiology Unit, Ottawa Hospital Research Institute,

University of Ottawa, Ottawa, ON, Canada

123

Neurocrit Care (2013) 18:154–160

DOI 10.1007/s12028-012-9787-9

died due to other causes (e.g., cardiac arrest, shock, etc.).

Epidural hematomas (OR 0.18, 95 % CI 0.06–0.56,

p < 0.01), craniotomies (OR 0.12, 95 % CI 0.02–0.68,

p = 0.02), and other non-neurosurgical procedures (OR

0.08, 95 % CI 0.02–0.43, p < 0.01) were less often asso-

ciated with death following withdrawal of life-sustaining

therapies than death from other causes.

Conclusions Death following decisions to withdraw life-

sustaining therapies is associated with specific patient and

clinical factors, and the intensity of care.

Keywords Withdrawal of life-sustaining therapy �Mortality � Traumatic brain injury � Risk factors �End-of-life

Introduction

Traumatic brain injury (TBI) is a major health and socio-

economic problem. Among individuals under 40 years of

age, TBI is the most common cause of death and disability

[1]. Despite the establishment of specialized intensive care

units (ICUs) and advances in treatment and monitoring,

TBI mortality remains high [2–5]. Previous studies have

found that 40–70 % of all deaths in ICUs follow a decision

to withdraw life-sustaining therapies [6–14]. Withdrawal of

life-sustaining therapies is, therefore, a major determinant

of death following TBI. In a previous study, we observed

significant variation in both mortality and incidence of

withdrawal of life-sustaining therapies in patients with

severe TBI among six level I trauma centers across Canada

[14].

Efforts have been made to understand factors contrib-

uting to decisions to withdraw life-sustaining therapies in

other critically ill populations [6–11, 15–21]. Several

determinants, including patient characteristics, the disease

process, therapeutic interventions, the acute care center, the

physician, and the surrogate decision-makers, have been

associated with decisions to withdraw or withhold life-

sustaining therapies in the critically ill [6–11, 15–21].

However, the overall critically ill populations studied to

date differ from the population of patients with severe TBI,

who are generally younger and without prior comorbidities

or explicit advanced care directives [12–14, 22]. To date,

little is known about the factors associated with decisions

to withdraw life-sustaining therapies in severe TBI patients

[12–14, 22]. We hypothesized that not only factors related

to the severity of injury, but also the use of certain inter-

ventions and management strategies, are associated with

decisions to withdraw life-sustaining therapies in critically

ill patients with severe TBI. The primary objective was to

identify the factors that were associated with decisions to

withdraw life-sustaining therapies among non-survivors.

Materials and Methods

We conducted a Canadian multicenter retrospective cohort

study of patients with severe TBI admitted to six level I

trauma centers in Canada (Hopital de l’Enfant-Jesus,

Hopital du Sacre-Coeur de Montreal, Hamilton General

Hospital, Sunnybrook Health Science Centre, St-Michael’s

Hospital and Foothill Medical Centre) from January 2005

to December 2006. Data on the variation in mortality and

the incidence of withdrawal of life-sustaining therapies, in

this cohort, were previously published [14]. The with-

drawal of life-sustaining therapies was defined as the

withdrawal of mechanical ventilation, of dialysis or con-

tinuous renal replacement therapy, or of vasopressors or

inotropes. We included mechanically ventilated patients,

aged 16 years and older who were admitted to an ICU

following a severe blunt TBI, defined as a Glasgow Coma

Scale (GCS) score B8, documented in the emergency room

or at ICU admission. Research Ethics Board approval was

obtained at each study site.

Patient Identification and Case Report Form

Development

Patients were identified using the International Classifica-

tion of Diseases 10 Codes for TBI (S06.0–S06.9). A

standardized case report form was developed by a group of

critical care physicians, clinical researchers, and research

nurses with experience in chart review and data abstraction.

The case report form was pre-tested in two of the partici-

pating institutions prior to the study initiation. A detailed

operations manual helped to unify the approach to data

collection across sites. Data were retrieved from the med-

ical record by trained abstractors having received a half-

day formation on the use of the case report form.

Data Collection

We collected the following baseline characteristics: age,

gender, injury severity score, cause of trauma, associated

traumatic injuries (c-spine, abdomen/pelvic content, chest,

bones other than c-spine), and referral. We also collected

daily data on vital signs, neurological exam findings, use of

jugular venous saturation monitoring or intracranial pressure

monitoring, and other laboratory tests from the time of

arrival in the emergency room until day 14 while in the ICU.

Data were also collected daily on management, including the

use of induced hypothermia, osmotic agents (mannitol,

hypertonic saline), vasopressors, sedation, paralytic agents,

seizure prophylaxis, intravenous insulin infusion, and deep

venous thrombosis prophylaxis. CT scan data were collected

on admission according to the official radiological reports.

Subsequent improvement or worsening CT scan was

Neurocrit Care (2013) 18:154–160 155

123

recorded. Data on surgical procedures were collected up to

28 days. Decisions to withdraw life-sustaining therapies

were collected.

Sample Size

The study sample size of 720 patients was originally

determined to estimate mortality rates and the proportion of

deaths related to decisions to withdraw life-sustaining

therapies with a precision of 10 % [14]. In order to avoid

season and calendar year variability, 60 patients per year

per center (n = 120 per center) were randomly selected

from all eligible patients.

Statistical Analysis

In order to evaluate the impact of various factors with the

decision to withdraw life-sustaining therapies in patients at

risk of in-hospital death, we restricted our study to the

cohort of non-survivors within the entire cohort of patients.

We thus compared patients who died following withdrawal

of life-sustaining therapies to patients who died due to

other causes (not following a decision to withdraw life-

sustaining therapies). This approach allowed identifying

patients with the highest risk of death and how deaths

following a decision to withdraw life-sustaining therapies

are different to other deaths when considering the severity

of the injury, interventions provided, and management.

We built a multivariable logistic regression model that

included gender and three baseline factors (age, GCS motor

score, and pupillary reflex at ICU admission) known to be

associated with prognosis in patients with severe TBI [23–

25]. Fractional polynomials and multiple different cate-

gorical divisions strategies were used to transform age and

GCS motor score variables in categorical variables. Age

was dichotomized as lower or equal to 55 years or greater

than 55 years old. Categories of GCS motor score were 1,

2–3, and 4–5–6. Pupil reflex was dichotomized as present

(bilateral or unilateral) or absent.

First, factors related to the injury were added to this

baseline model. Due to the large number of potential

explanatory variables, we used a manual backward elimi-

nation technique to select variables with p < 0.2 [26]. In a

second step, variables related to interventions and man-

agement strategies (e.g., craniotomy) were also added to

the model and selected applying the same backward

elimination technique. The four baseline variables were

never considered for elimination.

Variables associated with a p < 0.05 were considered

to be statistically significant factors related to decisions to

withdraw life-sustaining therapies. Multivariate analyses

were adjusted for clustering by trauma center using a

random intercept. Data are presented as mean ± standard

deviation (SD) for continuous variables, as median with

interquartile range (IQR) for ordinal variables and as odds

ratios (ORs) with 95 % confidence interval (CI) for the

regression analysis. Associations between categorical

variables were evaluated using Pearson v2 test. Statistical

analysis was performed using SAS version 9.2 (SAS

Institute, Cary, North Carolina).

Results

Mortality in the entire study cohort was 31.7 % (228/720)

and more than two-thirds (n = 160) of deaths were asso-

ciated with a decision to withdraw life-sustaining therapies

(Table 1). No patient for whom decisions to withdraw life-

sustaining therapies was taken survived. Among non-sur-

vivors (n = 228), 225 patients were considered in the final

multivariate analysis, following removal of 3 patients with

missing data for motor GCS score. Among variables of

baseline characteristics, only the cause of trauma was

missing for four patients (1.8 %) in the withdrawal of life-

sustaining therapies group. The mean age of patients

included in the analysis was 50.7 (±21.7) years. Patients

were predominantly male (69.7 %) and most trauma

involved a motor vehicle collision (57.3 %). Median GCS

was 3 (IQR: 3–4). Timing of death was comparable in both

patients who died following a decision to withdraw life-

sustaining therapies (median 3.0 days; IQR: 1.0–7.0) and

patients who died due to other causes (median: 2.0 days;

IQR: 1.0–6.0).

Factors Associated with Withdrawal of Life-Sustaining

Therapy

Injury Related Factors

In multivariate analyses, only two variables related to the

injury were significantly associated with decisions to

withdraw life-sustaining therapies: herniation on the initial

CT scan was associated with increased odds of decisions to

withdraw life-sustaining therapies (OR 2.91, 95 % CI

1.16–7.30, p = 0.02), whereas the presence of an epidural

hematoma on the initial CT scan was associated with a

reduction in the odds of withdrawal of life-sustaining

therapies in non-surviving patients (OR 0.18, 95 % CI

0.06–0.56, p < 0.01) (Table 2).

Interventions and Management Related Factors

In multivariate analyses (Table 2), only craniotomy (OR

0.12, 95 % CI 0.02–0.68, p = 0.02) and other surgical

procedures, excluding tracheotomy or a percutaneous

endoscopic gastrostomy tube insertion (ex. laparotomy,

156 Neurocrit Care (2013) 18:154–160

123

Table 1 Patient’s characteristics

among non-survivors

GCS Glasgow Coma Scale,

MVC motor vehicle collision,

DVT deep venous thrombosis,

PEG percutaneous endoscopic

gastrostomya Above ischemic range

(threshold value varying

according to centers)b Use of vasopressor at any

moment for at least 1 h

Variables Withdrawal of

life-sustaining

therapies

No withdrawal

of life-sustaining

therapies

p value

n = 160 (%) n = 68 (%)

Demographics

Male 110 (68.8) 50 (73.5) 0.51

Age >55 years 88 (55.0) 23 (33.8) <0.01

Referred from peripheral hospital 94 (58.8) 36 (52.9) 0.42

Cause of trauma

MVC 67 (41.9) 40 (58.8) <0.01

Struck by/against (assault) 5 (3.1) 7 (10.3)

Fall 74 (46.3) 20 (29.4)

Other 10 (6.3) 1 (1.5)

Unknown 4 (2.5) 0

Injury

Clinical

Non-reactive pupils 84 (52.5) 46 (67.6) 0.03

GCS motor score

1 107 (66.9) 58 (85.3) 0.03

2 or 3 21 (13.1) 4 (5.9)

4, 5, or 6 29 (18.1) 6 (8.8)

Unknown 3 (1.9) 0

Other traumatic injury associated 74 (46.3) 36 (52.9) 0.35

CT scan

Epidural hematoma/hemorrhage 12 (7.5) 9 (13.2) 0.17

Subdural hematoma/hemorrhage 93 (58.1) 35 (51.5) 0.35

Intracerebral contusion/hemorrhage 91 (56.9) 33 (48.5) 0.25

Intraventricular hemorrhage 51 (31.9) 21 (30.9) 0.88

Subarachnoid hemorrhage 101 (63.1) 35 (51.5) 0.10

Cerebral edema 52 (32.5) 22 (32.4) 0.98

Diffuse brain injury 23 (14.4) 10 (14.7) 0.95

Herniation 42 (26.3) 12 (17.6) 0.16

Shift of the median line C5 mm 46 (28.8) 21 (30.9) 0.75

Lab

SpO2B90 % 29 (18.1) 12 (17.6) 0.93

Troponinsa 30 (18.8) 22 (32.4) 0.03

Interventions

Medical

Vasopressor infusionb 91 (56.9) 51 (75.0) 0.01

Sedation: propofol 82 (51.3) 30 (44.1) 0.32

Sedation: barbituric 5 (3.1) 2 (2.9) N/A

Osmotic agent: mannitol 76 (47.5) 27 (39.7) 0.28

Osmotic agent: hypertonic saline 33 (20.6) 15 (22.1) 0.81

Muscle paralyzing agent 19 (11.9) 13 (19.1) 0.15

Insulin intravenous infusion 113 (70.6) 33 (48.5) <0.01

DVT prophylaxis 17 (10.6) 8 (11.8) 0.80

Induced hypothermia 24 (15.0) 9 (13.2) 0.73

Surgical

Intracranial pressure monitoring 58 (36.3) 21 (30.9) 0.44

Craniotomy/craniectomy 45 (28.1) 14 (20.6) 0.23

Tracheotomy/tracheostomy OR PEG tube insertion 5 (3.1) 1 (1.5) 0.48

Other 97 (60.6) 52 (76.5) 0.02

Surgical procedure within the first 48 h 152 (95.0) 66 (97.1) 0.49

Neurocrit Care (2013) 18:154–160 157

123

thoracotomy, sternotomy, open reduction of internal frac-

ture) (OR 0.08, 95 % CI 0.02–0.43, p < 0.01), were

associated with reduced odds of death following a decision

to withdraw life-sustaining therapies compared to death

due to other causes (Table 2).

Discussion

We identified four factors associated with decisions to

withdraw life-sustaining therapies in patients with severe

TBI. Among patients who ultimately died, evidence of

brain herniation on CT scan was associated with decisions

to withdraw life-sustaining therapies. Conversely, epidural

hematoma on CT scan and the performance of surgical

interventions (e.g., craniotomy) were associated with

decreased odds of death following a decision to withdraw

life-sustaining therapies than death not following the

withdrawal of life-sustaining therapies. These associations

cannot be considered causal, but rather provide information

on the processes and determinants surrounding death and

end-of-life decision-making for critically ill patients with

severe TBI at high risk of death.

Discussions about goals of care are common and part of

quality of care in the ICU. Since the majority of deaths in the

ICU occur following decisions to withdraw life-sustaining

therapies, it is vital to understand the determinants of these

decisions [6–11, 15–21]. Many studies have sought to

identify factors related to decisions to withdraw life-sus-

taining therapies in the overall population of critically ill

patients, but very few studies have focused on patients with

TBI [12–14, 22].

Findings on CT scans have previously been reported to

be associated with decisions to withdraw life-sustaining

therapies in patients with TBI. In our study, we observed

that some, but not all CT scan features, were associated the

decisions to withdraw life-sustaining therapies following

severe TBI. Evidence of brain herniation was associated

with decisions to withdraw life-sustaining therapies, while

the presence of an epidural hematoma was associated with

death due to other causes. The fact that an epidural

hematoma is associated with a better outcome when

addressed by surgery [27–29] while brain herniation is the

most common mechanism leading to brain death, may

explain these observations. Brain herniation could, there-

fore, be interpreted as a benchmark for severe neurological

suffering and become a trigger for decisions to withdraw

life-sustaining therapies. Unlike a prior study conducted in

a general trauma population, we did not observe a rela-

tionship between the presence of a midline shift of the

brain on CT scan and subsequent decisions to withdraw

life-sustaining therapies among non-survivors [7].

Table 2 Factors associated

with the withdrawal of life-

sustaining therapies in non-

survivors

GCS Glasgow Coma Scale,

DVT deep venous thrombosisa Above ischemic range

(threshold value varying

according to centers)b Use of vasopressor at any

moment for at least 1 h

Variables Univariate Multivariate

OR 95 % CI p value OR 95 % CI p value

Demographics

Male 0.68 0.35–1.32 0.30 0.66 0.31–1.42 0.29

Age >55 years 2.45 1.33–4.52 <0.01 1.69 0.82–3.47 0.16

Injury

Clinical

Non-reactive pupils 0.54 0.29–1.02 0.06 0.49 0.23–1.05 0.07

GCS motor score

1 0.43 0.16–1.14 0.09 0.63 0.20–1.97 0.43

2 or 3 1.47 0.34–6.31 0.61 1.51 0.29–7.89 0.63

4, 5, or 6 1.00 – – 1.00 – –

CT scan

Epidural hematoma/hemorrhage 0.40 0.15–1.05 0.07 0.18 0.06–0.56 <0.01

Herniation 1.52 0.71–3.25 0.28 2.91 1.16–7.30 0.02

Troponinsa 0.45 0.22–0.91 0.03 0.50 0.22–1.17 0.11

Interventions

Medical

Vasopressor infusionb 0.41 0.21–0.80 0.01 0.50 0.22–1.11 0.09

Insulin intravenous infusion 2.25 1.17–4.32 0.02 2.13 0.99–4.62 0.06

DVT prophylaxis 0.86 0.34–2.16 0.75 0.33 0.11–1.03 0.06

Surgical

Craniotomy/craniectomy 1.41 0.70–2.85 0.34 0.12 0.02–0.68 0.02

Other 0.42 0.22–0.83 0.01 0.08 0.02–0.43 <0.01

158 Neurocrit Care (2013) 18:154–160

123

Age and GCS are prognostic factors in patients with

severe TBI that have been shown to be associated with

decisions to withdraw life-sustaining therapies in general

ICU and trauma populations [7, 11, 17, 18, 21]. However,

we found no evidence that these factors were associated

with decisions to withdraw life-sustaining therapies in TBI

populations. No medical interventions were significantly

associated with decisions to withdraw life-sustaining ther-

apies. This is not surprising since it might be expected that

patients will be treated according to guidelines until a

decision to withdraw life-sustaining therapies is taken. On

the other hand, surprisingly few patients who died in our

cohort actually had an intracranial pressure monitor inser-

ted (34.6 %).

Strengths and Limitations

Our study has several strengths. First, it involved highly

standardized data abstraction across centers using trained

data abstractors and an operation manual to minimize

measurement bias. Second, our analyses were adjusted for

the most important known factors associated with prog-

nosis in severe TBI patients (age, GCS motor score, and

pupillary reflex) [23–25], thereby controlling for potential

variation in the risk of death due to the severity of brain

injury. Third, we restricted our cohort to non-survivors,

also reducing the potential for confounding by disease

severity.

Our study also has limitations. Due to the retrospective

design of our study, the observed data were limited to the

information detailed in the patients’ charts, including the

decision to withdraw life-sustaining therapies. Poor docu-

mentation may have resulted in underestimation of the

incidence of withdrawal of life-sustaining therapies [30].

However, considering the medico-legal importance of this

information, it is unlikely that it would be missing in a

significant number of charts. In addition, we used available

information from CT scan reports, and thus did not have

information on scoring system (i.e., Marshall score), which

could have been a good way to categorize the severity of

the brain injury on CT scan. Second, since our model was

based on non-survivors, it does not represent the entire

population of severe TBI patients at risk of withdrawal of

life-sustaining therapies who survived. This may also

explain some of the divergences of our results with those of

previous studies [16, 17]. Third, our study was not

designed to understand whether deaths following the

withdrawal of life-sustaining therapies happened following

a thoughtful process or not, but to aim understanding what

factors were associated with these decisions to withdraw

life-sustaining therapies, notwithstanding if it was due to a

self-fulfilling prophecy or not [31]. Finally, our study

aimed to identify factors related to death following

decisions to withdraw life-sustaining therapies, but was not

designed to evaluate all factors potentially related to the

determinants of decisions to withdraw life-sustaining

therapies in critically ill patients with severe TBI, for

example estimations of patients’ preferences and values.

Although we agree that reasons and processes leading to a

decision to withdraw life-sustaining therapies are complex,

variable, and likely multifactorial [7, 8, 18, 32], we focused

on specific factors related to the injury and therapeutic

interventions.

Conclusion

Evidence of epidural hematoma or brain herniation on CT

scan and specific surgical interventions were associated

with decisions to withdraw life-sustaining therapies among

non-survivors of severe TBI. Considering the high rate of

decisions to withdraw life-sustaining therapies among

patients with severe TBI, further research is required to

understand the processes and determinants associated with

decisions to withdraw life-sustaining therapies in this

population.

Acknowledgments The authors would like to thank Tran Cong

Dung MD, MSc; Mohana Ratnapalan HBSc, Stephanie Todd BSc,

MBT, John Harlock MD, Frederic Morin RN and David Simonyan

MD MSc, for their help in data acquisition; Valerie Murat MSc for

her participation in the data cleaning process; Mrs. Valerie Boucher

for her secretarial assistance. This work was presented in part at the

Intensive Care and Emergency Medicine Symposium (Brussels,

Belgium, March 2011). The work was supported in part by the

Fondation de l’Hopital de l’Enfant-Jesus (Enfant-Jesus Hospital

Foundation) who had no role in any part of conduct of the study or

preparation of the manuscript. Drs. Turgeon and Lauzier are recipi-

ents of a Research Career Award from the Fonds de RechercheQuebec-Sante (FRQ-S). Drs. Moore, Scales, and Fergusson are

recipients of New Investigator Awards from the Canadian Institutesfor Health Research (CIHR). Dr. Zarychanski is a recipient of a CIHR

RCT Mentorship Award. Dr. Burns holds a Clinician Scientist Phase

2 Award from the CIHR.

References

1. National Scientific Advisory Committee. Identifying priorities for

research and capacity development in injury as a multi-institute

strategic initiative within the Canadian Institutes of Health

Research. In: Listening for direction on injury: final report

Instituts canadiens de recherche en sante; 2004.

2. Coronado VG, Xu L, Basavaraju SV, et al. Surveillance for

traumatic brain injury-related deaths—United States, 1997–2007.

MMWR Surveill Summ. 2011;60:1–32.

3. Masson F, Thicoipe M, Aye P, et al. Epidemiology of severe

brain injuries: a prospective population-based study. J Trauma.

2001;51:481–9.

4. Murray GD, Teasdale GM, Braakman R, et al. The European

Brain Injury Consortium survey of head injuries. Acta Neurochir

(Wien). 1999;141:223–36.

Neurocrit Care (2013) 18:154–160 159

123

5. Myburgh JA, Cooper DJ, Finfer SR, et al. Epidemiology and

12-month outcomes from traumatic brain injury in Australia and

New Zealand. J Trauma. 2008;64:854–62.

6. Spronk PE, Kuiper AV, Rommes JH, Korevaar JC, Schultz MJ.

The practice of and documentation on withholding and with-

drawing life support: a retrospective study in two Dutch intensive

care units. Anesth Analg. 2009;109:841–6.

7. Cooper Z, Rivara FP, Wang J, MacKenzie EJ, Jurkovich GJ.

Withdrawal of life-sustaining therapy in injured patients: varia-

tions between trauma centers and nontrauma centers. J Trauma.

2009;66:1327–35.

8. Azoulay E, Metnitz B, Sprung CL, et al. End-of-life practices in

282 intensive care units: data from the SAPS 3 database. Inten-

sive Care Med. 2009;35:623–30.

9. Zahuranec DB, Morgenstern LB, Sanchez BN, Resnicow K, White

DB, Hemphill JC III. Do-not-resuscitate orders and predictive

models after intracerebral hemorrhage. Neurology. 2010;75:

626–33.

10. Rocker G, Cook D, Sjokvist P, et al. Clinician predictions of

intensive care unit mortality. Crit Care Med. 2004;32:1149–54.

11. Becker KJ, Baxter AB, Cohen WA, et al. Withdrawal of support

in intracerebral hemorrhage may lead to self-fulfilling prophecies.

Neurology. 2001;56:766–72.

12. Rocker GM, Cook DJ, Shemie SD. Brief review: practice vari-

ation in end of life care in the ICU: implications for patients with

severe brain injury. Can J Anaesth. 2006;53:814–9.

13. O’Callahan JG, Fink C, Pitts LH, Luce JM. Withholding and

withdrawing of life support from patients with severe head injury.

Crit Care Med. 1995;23:1567–75.

14. Turgeon AF, Lauzier F, Simard JF, et al. Mortality associated

with withdrawal of life-sustaining therapy for patients with severe

traumatic brain injury: a Canadian multicentre cohort study.

CMAJ. 2011;183:1581–8.

15. Varelas PN, Abdelhak T, Hacein-Bey L. Withdrawal of life-

sustaining therapies and brain death in the intensive care unit.

Semin Neurol. 2008;28:726–35.

16. Cook D, Rocker G, Marshall J, et al. Withdrawal of mechanical

ventilation in anticipation of death in the intensive care unit.

N Engl J Med. 2003;349:1123–32.

17. Esteban A, Gordo F, Solsona JF, et al. Withdrawing and with-

holding life support in the intensive care unit: a Spanish

prospective multi-centre observational study. Intensive Care Med.

2001;27:1744–9.

18. Diringer MN, Edwards DF, Aiyagari V, Hollingsworth H. Factors

associated with withdrawal of mechanical ventilation in a neu-

rology/neurosurgery intensive care unit. Crit Care Med. 2001;29:

1792–7.

19. Ferrand E, Robert R, Ingrand P, Lemaire F. Withholding and

withdrawal of life support in intensive-care units in France: a

prospective survey. French LATAREA Group. Lancet. 2001;357:

9–14.

20. Mayer SA, Kossoff SB. Withdrawal of life support in the neu-

rological intensive care unit. Neurology. 1999;52:1602–9.

21. Keenan SP, Busche KD, Chen LM, McCarthy L, Inman KJ,

Sibbald WJ. A retrospective review of a large cohort of patients

undergoing the process of withholding or withdrawal of life

support. Crit Care Med. 1997;25:1324–31.

22. Thompson HJ, Rivara FP, Jurkovich GJ, Wang J, Nathens AB,

MacKenzie EJ. Evaluation of the effect of intensity of care on

mortality after traumatic brain injury. Crit Care Med. 2008;36:

282–90.

23. Perel P, Arango M, Clayton T, et al. Predicting outcome after

traumatic brain injury: practical prognostic models based on large

cohort of international patients. BMJ. 2008;336:425–9.

24. Murray GD, Butcher I, McHugh GS, et al. Multivariable prog-

nostic analysis in traumatic brain injury: results from the

IMPACT study. J Neurotrauma. 2007;24:329–37.

25. Steyerberg EW, Mushkudiani N, Perel P, et al. Predicting out-

come after traumatic brain injury: development and international

validation of prognostic scores based on admission characteris-

tics. PLoS Med. 2008;5:e165. discussion e165;1251–61.

26. Harrell FE Jr, Lee KL, Califf RM, Pryor DB, Rosati RA.

Regression modelling strategies for improved prognostic pre-

diction. Stat Med. 1984;3:143–52.

27. Brain Trauma Foundation. Guidelines for the Surgical Manage-

ment of Traumatic Brain Injury. Neurosurgery. 2006;58:S2-1–3.

28. Brain Trauma Foundation. Early indicators of prognosis in severe

traumatic brain injury. In: Management and prognosis of severe

traumatic brain injury. American Association of Neurological

Surgeons; 2000.

29. Brain Trauma Foundation. Guidelines for the Management of

Severe Traumatic Brain Injury. 3rd ed. New York: Brain Trauma

Foundation; 2007.

30. Ratnapalan M, Cooper AB, Scales DC, Pinto R. Documentation

of best interest by intensivists: a retrospective study in an Ontario

Critical Care Unit. BMC Med Ethics. 2010;11:1–7.

31. Wilkinson D. The self-fulfilling prophecy in intensive care. Theor

Med Bioeth. 2009;30:401–10.

32. Ho KM, Liang J. Withholding and withdrawal of therapy in New

Zealand intensive care units (ICUs): a survey of clinical directors.

Anaesth Intensive Care. 2004;32:781–6.

160 Neurocrit Care (2013) 18:154–160

123