Hypernatremia Is Associated with Increased Risk of Mortality in Pediatric Severe Traumatic Brain Injury

Abstract

Acquired hypernatremia in hospitalized patients is often associated with poorer outcomes. Our aim was to evaluate the relationship between acquired hypernatremia and outcome in children with severe traumatic brain injury (sTBI). We performed a retrospective cohort study of all severely injured trauma patients (Injury Severity Score ≥12) with sTBI (Glasgow Coma Scale [GCS] ≤8 and Maximum Abbreviated Injury Scale [MAIS] ≥4) admitted to a Pediatric Critical Care Unit ([PCCU]; 2000–2009). In a cohort of 165 patients, 76% had normonatremia (135–150 mmol/L), 18% had hypernatremia (151–160 mmol/L), and 6% had severe hypernatremia (>160 mmol/L). The groups were similar except for lower GCS (p=0.002) and increased incidence of fixed pupil(s) on admission in both hypernatremia groups (p<0.001). Mortality rate was four-fold and six-fold greater with hypernatremia and severe hypernatremia, respectively (p<0.001), and mortality rates were unchanged when patients with fixed pupils or those with central diabetes insipidus were excluded (p<0.001). Hypernatremic patients had fewer ventilator-free days (p<0.001). Survivors with hypernatremia had greater PCCU (p=0.001) and hospital (p=0.031) lengths of stays and were less frequently discharged home (p=0.008). Logistic regression analyses of patient characteristics and sTBI interventions demonstrated that hypernatremia was independently associated with the presence of fixed pupil(s) on admission (odds ratio [OR] 5.38; p=0.003); administration of thiopental (OR 8.64; p=0.014), and development of central diabetes insipidus (OR 5.66; p=0.005). Additional logistic regression analyses demonstrated a significant association between hypernatremia and mortality (OR 6.660; p=0.034). In summary, acquired hypernatremia appears to signal higher risk of mortality in pediatric sTBI and is associated with a higher discharge level of care in sTBI survivors.

Introduction

S evere traumatic brain injury (sTBI) is a major cause of death and disability in children.¹ Sodium dysregulation can complicate pediatric sTBI, manifesting as either hyponatremia or hypernatremia.^2
–4 While hyponatremia is generally associated with poor neurological outcome, hypernatremia is often considered benign or even a beneficial complication that might help control cerebral edema. The etiologies of hypernatremia in sTBI patients are diverse and include insensible water losses, inadequate provision of free water, excess sodium administration, the development of central diabetes insipidus (DI), and other coexisting predisposing neurological conditions.⁵

Hypernatremia associated with sTBI is often attributed to therapeutic interventions directed at controlling elevated intracranial pressure (ICP).^6,7 In particular, treatment regimens to lower ICP after sTBI include administering either hypertonic saline (HS, or 3% NaCl), which has the potential to raise the serum sodium above physiologic levels^8

–11 or mannitol, which promotes renal water loss.^12,13

In critically ill children, only a single retrospective study (non-English) has examined the relationship of between hypernatremia and clinical outcome, with the worst outcomes occurring in patients with hypernatremia and polyuria, likely representing DI.¹⁴ Thus, the objective of this study was to examine the relationship between acquired hypernatremia and outcome in a large cohort of children with sTBI.

Patients and Methods

This study was approved by the Western University, Health Sciences Research Ethics Board. Our retrospective cohort study included all severely injured (Injury Severity Score [ISS] ≥12) patients (<18 years) admitted to the Pediatric Critical Care Unit ([PCCU]; 2000–2009) after experiencing a severe TBI (defined as an initial Glasgow Coma Scale [GCS] ≤8 and a Maximum Abbreviated Injury Scale [MAIS] score of ≥4 for the head).¹⁵ The PCCU is located in the Children's Hospital, London Health Sciences Centre (LHSC), which is the regional Pediatric Level I trauma center for Southwestern Ontario, serving a geographical area of 19,000 square kilometers with a pediatric population more than 500,000. Patients admitted to the PCCU >12 h after the injury were excluded. Patients were identified from the prospectively collected written and electronic admission records of the PCCU and collated with the data of the LHSC Trauma Registry to ensure all eligible patients were captured.¹⁶

The following data were obtained from LHSC's electronic charting system, paper copy of the patients' hospital chart, and the Trauma Registry. Demographic and injury data included age, sex, time from injury to arrival to the trauma center, mechanism of injury, MAIS, and ISS. Clinical variables recorded include first GCS (scene, referring hospital, or on arrival at LHSC trauma center), pupillary response on admission, PCCU and hospital lengths of stay (LOS) in days, ventilator-free days (unventilated days in the first 28 days post-injury), episodes of either hypotension (defined as systolic blood pressure [SBP] <70 for infants, SBP <70+(2*age) for toddlers and children less 10 years old, and SBP <90 mm Hg for children ≥10 years old)¹⁷ or hypoxemia (PO₂<65 mm Hg, admission arterial blood gas), initial admission blood gas, placement of an ICP monitoring device,¹⁸ abnormal ICP measurements (any daily maximum reading >20 mm Hg), and treatment with 3% HS, mannitol, desamino-8-D-arginine vasopressin and/or arginine vasopressin (DDAVP/AVP), thiopental infusion, or the use of therapeutic hypothermia. The administration of DDAVP/AVP was used as a surrogate for diagnosis of central DI.¹⁶

In-hospital mortality was identified as the primary outcome measure. Secondary outcomes included ventilator-free days, and in survivors, PCCU and hospital LOS as well as discharge status (i.e., rehabilitation hospital, another acute care hospital, or “home,” which included home and foster care).

For the purposes of this study, patients were divided into three groups according to daily average serum sodium measurements. Daily average serum sodium was derived from the readings of serum sodium measurements over 24 h, with a minimum of two and a maximum of eight readings per day, for each of the first 10 days in the PCCU. According to the highest daily average sodium recorded during the PCCU stay, patients were classified into one of the three groups: normonatremia (135–150 mmol/L), hypernatremia (151–160 mmol/L), or severe hypernatremia (>160 mmol/L).¹²

All data were screened for normality, and skewed data points were presented as medians with interquartile range (IQR). The Pearson chi square and Fisher exact test were used to analyze categorical variables, and the non-parametric Kruskal-Wallis test with a Dunn post-hoc test was used to analyze continuous variables. A P value of <0.05 was considered statistically significant.

Multivariate logistic regression analysis was undertaken to determine the effects of patients demographics, TBI severity, and treatment interventions on outcome, either hypernatremia or mortality. Possible confounders were identified a priori. A backward elimination strategy was used to include all statistically and clinically significant variables.¹⁹ The final model was used to determine the estimated odds (odds ratio [OR]) of death among hypernatremic patients, adjusted for confounders, relative to the odds of death among normal sodium patients. All analyses were performed using Predictive Analytics SoftWare (PASW) Statistics 18 (SPSS Inc., Chicago, IL).

Results

A total of 798 trauma patients were screened, and 180 cases of sTBI were subsequently identified as meeting the study inclusion criteria of GCS ≤8 and MAIS ≥4. After the exclusion of patients who experienced delayed arrival at our trauma center (n=3), death in the first 12 h of admission (n=7), or in whom clinical data were incomplete (n=5), a total of 165 patients were included in the analyses. At the time of PCCU admission, two (1.2%) patients were hypernatremic (initial serum sodium >150 mmol/L). Subsequently, however, severe hypernatremia developed in 6% of patients (n=10), and 18% (n=30) manifested hypernatremia. The remaining 76% of patients (n=125) were normonatremic throughout their PCCU stay.

The demographic and injury data were similar across the three patient groups, except for lower GCS and increased presence of fixed pupil(s) on admission in the hypernatremia groups (Table 1). The ISS and the severity of head injury (MAIS) did not differ between the patient groups.

Table 1.

Patient Demographics and Admission Injury Data Across the Three Sodium Groups

Variable^a	Normonatremia n=125 (76%)	Hypernatremia n=30 (18%)	Severe hypernatremia n=10 (6%)	P value
Age (years)	15.0 (10.0)	12.0 (15.3)	13.5 (8.3)	0.65
Male (%)	75 (60)	21 (70)	80 (80)	0.30
Etiology (%)				0.54
MVC	93 (74)	22 (73)	8 (80)
Abuse/assault	12 (10)	5 (17)	0 (0)
Other	20 (16)	3 (10)	2 (20)
ISS	35 (17)	35 (20)	43 (3)	0.53
Injury profile
MAIS head	5 (0)	5 (0)	5 (0)	0.10
MAIS face	2 (1)	1 (1)	1 (0)	0.27
MAIS chest	3 (1)	4 (0)	4 (1)	0.81
MAIS abdomen	2 (2)	2 (1)	4 (2)	0.50
MAIS extremity	3 (1)	3 (0)	3 (0)	0.48
MAIS external	1 (0)	1 (0)	1 (0)	0.88
GCS	5 (4)	3 (1)	4 (3)	0.002^b
Fixed pupils (%)
Left	23 (20)	18 (64)	8 (80)	<0.001^b
Right	21 (18)	18 (64)	8 (80)	<0.001^b
Hypoxia (%)	12 (10)	4 (13)	2 (20)	0.669
Hypotension (%)	2 (20)	11 (36)	2 (20)	0.123

Continuous variables reported as median (interquartile range); categorical data reported as n (%).

Significant differences between normonatremia and both hypernatremia and severe hypernatremia groups only.

MVC=motor vehicle collision; ISS=injury severity score; MAIS=Maximum Abbreviated Injury Scale; GCS=Glasgow Coma Scale.

The mortality rate differed significantly across the three groups exhibiting greater than four- and six-fold increases in the hypernatremia and severe hypernatremia groups, respectively, compared with the normonatremia group (Table 2). When all patients who received DDAVP/AVP, used as a surrogate for central DI, or those patients with fixed pupil(s) on admission were excluded, the mortality rate was still significantly higher in the hypernatremic patients (P<0.001).

Table 2.

Patient Outcomes Across the Three Sodium Groups

All patients^a	Normonatremia n=125 (76%)	Hypernatremia n=30 (18%)	Severe hypernatremia n=10 (6%)	P value
Mortality rate (%)	18 (14)	19 (63)	9 (90)	<0.001^b
Ventilator-free days	12 (22)	0 (18)	0 (0)	<0.001^b

Survivors only^a	Normonatremia n=107	All hypernatremian n=12^c
PCCU LOS^b	5 (9)	16 (12)	0.001
Hospital LOS^b	21 (27)	35 (33)	0.031
Discharged to (%)^b,c			0.008
Home	74 (69)	3 (25)
Acute care	10 (9)	2 (17)
Rehabilitation	23 (22)	7 (58)

Continuous variables reported as median (interquartile range); categorical data reported as n (%).

Significant differences between normonatremia and both hypernatremia and severe hypernatremia groups only.

Hypernatremia and severe hypernatremia categories were collapsed for calculation of a valid p value, because only one patient survived in the severe hypernatremia group.

PCCU=Pediatric Critical Care Unit; LOS=length of stay.

Secondary outcome variables are also listed in Table 2. Ventilator-free days were significantly reduced in hypernatremic patients. In hypernatremic survivors, PCCU and hospital LOS were significantly elevated. Finally, more survivors in the normonatremic group were discharged home compared with the hypernatremic groups. In the hypernatremic group, a greater proportion of patients were discharged to a higher level of care, including either an acute care or a rehabilitation institution.

While only 33% (n=55/165) of sTBI patients had placement of ICP monitors, the number of patients who received an ICP monitor was similar among the three sodium groups (Table 3), as were the number of patients with abnormal ICP readings (daily maximum ICP >20 mm Hg; normonatremia 89%, hypernatremia 86%, severe hypernatremia 100%; p=0.634).

Table 3.

Traumatic Brain Injury-Related Interventions Administered Across the Three Sodium Groups

Intervention^a	Normonatremia n=125 (76%)	Hypernatremia n=30 (18%)	Severe hypernatremia n=10 (6%)	P value
ICP monitor	36 (29)	13 (43)	6 (60)	0.058
Mannitol^b	43 (34)	17 (57)	6 (60)	0.034^c
Hypertonic saline^b	31 (25)	11 (37)	6 (60)	0.037^d
Thiopental infusion	5 (4)	8 (27)	6 (60)	<0.001^e
Therapeutic hypothermia	10 (8)	2 (7)	0 (0)	0.654
Decompressive craniectomy	12 (10)	9 (30)	2 (20)	0.014^c
DDAVP/AVP	10 (8)	14 (47)	8 (80)	<0.001^f

ICP=intracranial pressure; DDAVP=desamino-8-D-arginine vasopressin; AVP=arginine vasopressin.

Categorical data reported as n (%).

The median volume per patient of mannitol and hypertonic saline in mL/kg were similar between groups (p=0.905 and p=0.995, respectively).

Significant difference between normonatremia and hypernatremia.

Significant difference between normonatremia and severe hypernatremia.

Significant difference between hypernatremia and both normonatremia and severe hypernatremia.

Significant difference between normonatremia and both hypernatremia and severe hypernatremia.

Acquired hypernatremia was associated with a greater number of patients who received mannitol, HS, thiopental infusion, decompressive craniectomy, and DDAVP/AVP (Table 3). Despite the fact that a higher proportion of patients in the hypernatremia groups were administered HS and mannitol compared with patients with normonatremia, the median volume of HS and mannitol administered to individual patients in mL/kg was similar in each group (HS, p=0.995; mannitol, p=0.905).

To further examine the association of patient-specific variables and sTBI interventions on the occurrence of hypernatremia and to control for the effects of possible confounding variables, multivariate logistic regression modeling was undertaken with hypernatremia as the outcome variable (Table 4). Hypernatremia was significantly associated with fixed pupil(s) on admission (p=0.003), administration of thiopental (p=0.014) and development of central DI (p=0.005). The adjusted R² suggested a good fit, with more than half of the variability in hypernatremia explained by the chosen variables in this model.

Table 4.

Logistic Regression Modeling with Hypernatremia As Outcome

Variable	B	SE	OR	P value
Thiopental	2.156	0.880	8.64	0.014
Central diabetes insipidus	1.733	0.613	5.66	0.005
Fixed pupil(s)	1.682	0.571	5.38	0.003
Decompressive craniectomy	0.899	0.740	2.46	0.225
Therapeutic hypothermia	−1.047	1.024	0.35	0.307
Injury Severity Score	−0.018	0.024	0.98	0.434
3% saline	−0.619	0.857	0.54	0.470
Mannitol	0.445	0.662	1.56	0.501
ICP monitor	0.290	0.826	1.34	0.726
Constant	−2.323	0.883
Adjusted R²	0.501

SE=standard error; OR=odds ratio; ICP=intracranial pressure.

Finally, to determine the association of sTBI patient variables on mortality and to control for the effects of possible confounding variables, multivariate logistic regression modeling was undertaken with mortality as the outcome variable (Table 5). Mortality was associated with development of central DI (p=0.003), the presence of fixed pupil(s) on admission (p=0.001), hypernatremia (p=0.034), hypotension (p=0.043), and low GCS (p=0.030). The relationship between serum sodium and the odds of death was evident, with hypernatremic patients 6.6 times more likely to die than for patients with normonatremia, controlling for other variables. The adjusted R² suggested a very good fit, with more than 80% of the variability in mortality explained by the chosen variables in this model.

Table 5.

Logistic Regression Modeling with Mortality As Outcome

Variable	B	SE	OR	P value
Central diabetes Insipidus	3.256	1.089	25.957	0.003
Fixed pupil(s)	3.029	0.859	20.668	0.001
Hypernatremia	1.896	0.897	6.660	0.034
Hypotension	1.873	0.925	6.510	0.043
Glasgow Coma Scale	−0.683	0.314	0.505	0.030
Age	0.024	0.070	1.025	0.729
Injury Severity Score	−0.001	0.041	0.999	0.977
Constant	−1.886	2.106
Adjusted R²	0.824

SE=standard error; OR=odds ratio.

Discussion

Our study suggests an association exists between hypernatremia and mortality in pediatric sTBI patients, a finding consistent with previous adults studies, which showed that hypernatremia increased the risk of mortality in intensive care,^13,20 neurocritical care,¹² cardiac intensive care,²⁰ and in sTBI patients.²² In critically ill children, only a single retrospective study (non-English) has examined the relationship of hypernatremia and outcome, in which the worst outcomes occurred in patients with hypernatremia and polyuria, likely representing DI.¹⁴

Despite being a single center study, our data are likely to be representative of all pediatric sTBI patients, as we employed well-accepted inclusion and exclusion criteria to identify our sTBI subjects,²³ and our population mortality rate (22%) was similar to previously reported mortality rates for sTBI in children (24%–30%).^24,25 Only two (1.2%) of our patients presented to the hospital with hypernatremia, while 38 (23%) patients acquired hypernatremia after admission to the PCCU. Therefore, 95% of our hypernatremia cases were acquired during the patients' PCCU stay, an incidence higher than previously reported for hospital-acquired hypernatremia in the general pediatric population (56%–60%).^2,3 The increased proportion of acquired hypernatremia in our sTBI patients was likely because of our specific focus on the brain injured.

Only one-third of our sTBI patients had placement of an ICP monitor. The indications used for placement of ICP monitors in our sTBI population were difficult to determine retrospectively, but the procedure is undertaken at the discretion of the clinical team and is aided by information related to the mechanism of injury, clinical assessment, and abnormalities on brain imaging. Hence, sTBI patients deemed at the highest risk for development of cerebral edema and elevated ICP are selected for ICP monitoring, and one would expect that those patients with an ICP monitor would have received more ICP lowering therapies. Our logistic regression modeling, however, failed to show a relationship between ICP monitoring and subsequent development of hypernatremia. In those sTBI patients with ICP monitors placed, the number of abnormal ICP readings was also similar between normonatremic and hypernatremic groups.

Therapies to control elevated ICP after sTBI have the potential to raise serum sodium levels.²⁶ Our logistic regression analysis failed to show a relationship between use of either HS or mannitol and the subsequent development of hypernatremia. Our logistic regression analysis was supported by the findings: (1) that HS and mannitol were administered primarily to children with ICP monitors and the proportions of patients with elevated ICP were similar across all three groups, and (2) that the volumes administered of either HS or mannitol in mL/kg were similar across all three groups.

Our logistic regression analysis suggested that thiopental administration was associated with development of hypernatremia in sTBI patients. Thiopental infusion is known to cause electrolyte disturbances, but these have usually been related to abnormdities in serum potassium.^27,28 To our knowledge, only one study reported development of hypernatremia in a mixed adult neurocritical care population as a consequence of thiopental infusion.²⁹ The mechanisms whereby thiopental may cause hypernatremia are unclear, but may be related to the preparation (provided as a sodium salt) and/or transient renal insufficiency.³⁰

Because of data complexity, our analyses were insufficient to identify concrete cause-effect relationships, and we concede that our analyses also lack the sensitivities to determine the additive effects of combined ICP therapies on the development of hypernatremia. A scale for examining the intensity of individual ICP therapies has been proposed, but also has inherent limitations.³¹ Beyond ICP therapies, other potential contributions to the observed incidence of hypernatremia may include intrinsic sodium dysregulation, increased insensible losses associated with critical illness, and/or inadequate free fluid administration.

Despite these data limitations, our study does suggest an association between hypernatremia and poor outcome in children with sTBI, including increased mortality, reduced ventilator-free days, increased PCCU and hospital LOS, as well as greater discharge morbidities. The significant difference in mortality associated with acquired hypernatremia persisted even when we controlled for GCS and patients with fixed pupil(s) on admission. Exclusion of all patients treated with DDAVP/AVP during their PCCU stay also did not change the statistically significance of the difference in mortality. Mortality rates of 69–86% have been reported on both children and adults with central DI after sTBI.^32,33

Hypernatremia appears to be a significant mortality signal in children after sTBI and, thus, requires attention.³⁴ The increased odds of death (OR=6.6) for the hypernatremic groups, after controlling for other variables such as severity of head and overall injury, supports a significant association between hypernatremia and mortality in children with sTBI. While this study was not designed to conclusively identify the causes of acquired hypernatremia, the magnitude of the increase in risk of mortality in hypernatremic patients is substantial and indicates the need for a greater clinical attention to hypernatremia after sTBI. A re-examination of water and salt homeostasis in sTBI and prospective evaluation of elevated ICP therapies are warranted.

Our study has a number of limitations. First, ours was a retrospective study based on the experience at a single center. Despite this, our patient cohort is likely representative of other sTBI populations based on strict inclusion criteria and, therefore, our results should be generalizable. Second, there was a significant difference in the admission GCS and fixed pupil(s) between the groups; however, we used logistic regression analysis to control for the potential confounding effect of GCS and fixed pupil(s) on mortality, and also excluded patients with fixed pupils in other mortality analyses. Third, while thiopental administration was the only sTBI intervention associated with acquired hypernatremia in our analyses, the sTBI patients were also exposed to other therapies that may have had subtle influence on sodium and water balances. Finally, we were unable to adequately control for fluid balances in this retrospective study, again emphasizing the need for prospective studies.

Conclusion

This study suggests that there is a significant association between acquired hypernatremia and mortality in children with sTBI. In sTBI survivors, additional poor outcomes were identified including reduced ventilator-free days, increased PCCU and hospital LOS, and a higher level of care needed on discharge. Acquired hypernatremia could be attributed to a variety of factors, including severity of brain injury based on admission fixed pupil(s), thiopental administration, and development of central DI.

Footnotes

Acknowledgments

We thank Mr. Jamie Seabrook of the Children's Health Research Institute for statistical advice and critical review.

Author Disclosure Statement

No competing financial interests exist.

References

Adelson

P.D.

, Bratton

S.L.

, Carney

N.A.

, Chesnut

R.M.

, du Coudray

H.E.

, Goldstein

, Kochanek

P.M.

, Miller

H.C.

, Partington

M.D.

, Selden

N.R.

, Warden

C.R.

, Wright

. 2003. Guidelines for the acute medical management of severe traumatic brain injury in infants, children, and adolescents. Chapter 1: Introduction. Pediatr. Crit. Care Med., 4,Suppl. 3:S2–S4.

Dunn

, Butt

1997. Extreme sodium derangement in a paediatric inpatient population. J. Paediatr. Child Health, 33:26–30.

Moritz

M.L.

, Ayus

J.C.

1999. The changing pattern of hypernatremia in hospitalized children. Pediatrics, 104:435–439.

Moro

, Katayama

, Igarashi

, Mori

, Kawamata

, Kojima

2007. Hyponatremia in patients with traumatic brain injury: incidence, mechanism, and response to sodium supplementation or retention therapy with hydrocortisone. Surg. Neurol., 68:387–393.

Tisdall

, Crocker

, Watkiss

, Smith

2006. Disturbances of sodium in critically ill adult neurologic patients: a clinical review. J. Neurosurg. Anesthesiol., 18:57–63.

Chan

K.H.

, Yue

C.P.

, Mann

K.S.

1990. The risk of intracranial complications in pediatric head injury. Results of multivariate analysis. Childs Nerv. Syst., 6:27–29.

Tepas

J.J.

3rd , DiScala

, Ramenofsky

M.L.

, Barlow

1990. Mortality and head injury: the pediatric perspective. J. Pediatr. Surg., 25:92–96.

Fisher

L.A.

, Ko

, Miss

, Tung

P.P.

, Kopelnik

, Banki

N.M.

, Gardner

, Smith

W.S.

, Lawton

M.T.

, Zaroff

J.G.

2006. Hypernatremia predicts adverse cardiovascular and neurological outcomes after SAH. Neurocrit. Care, 5:180–185.

Khanna

, Davis

, Peterson

, Fisher

, Tung

, O'Quigley

, Deutsch

2000. Use of hypertonic saline in the treatment of severe refractory posttraumatic intracranial hypertension in pediatric traumatic brain injury. Crit. Care Med., 28:1144–1151.

10.

Peterson

, Khanna

, Fisher

, Marshall

2000. Prolonged hypernatremia controls elevated intracranial pressure in head-injured pediatric patients. Crit. Care Med., 28:1136–1143.

11.

Simma

, Burger

, Falk

, Sacher

, Fanconi

1998. A prospective, randomized, and controlled study of fluid management in children with severe head injury: lactated Ringer's solution versus hypertonic saline. Crit. Care Med., 26:1265–1270.

12.

Aiyagari

, Deibert

, Diringer

M.N.

2006. Hypernatremia in the neurologic intensive care unit: how high is too high? J. Crit. Care., 21:163–172.

13.

Hoorn

E.J.

, Betjes

M.G.

, Weigel

, Zietse

2008. Hypernatraemia in critically ill patients: too little water and too much salt. Nephrol. Dial. Transplant., 23:1562–1568.

14.

Lekmanov

A.U.

, Azovskii

D.K.

, Piliutik

S.F.

, Gegueva

E.N.

, Abramova

V.M.

, Chernova

A.S.

2010. [Hypernatremia as a predictor of poor outcomes in children with severe brain injury] (Rus) Anesteziol. Reanimatol., 1:41–44.

15.

The Abbreviated Injury Scale Manual. 1995. Association for the Advancement of Automotive Medicine (AAAM): Des Plaines, IL.

16.

Alharfi

I.M.

, Charyk. Stewart

, Foster

, Morrison

G.C.

, Fraser

D.D.

2013. Central diabetes insipidus in pediatric severe traumatic brain injury. Pediatr. Crit. Care Med., Jan 9[Epub ahead of print.]

17.

Ralston

American Heart Association and American Academy of Pediatrics. 2006. PALS Provider Manual. American Heart Association: Dallas, Tex.

18.

Ngo

Q.N.

, Ranger

, Singh

R.N.

, Kornecki

, Seabrook

J.A.

, Fraser

D.D.

2009. External ventricular drains in pediatric patients. Pediatr. Crit. Care Med., 10:346–351.

19.

Hosmer

D.W.

, Lemeshow

1989. Applied Logistic Regression: Wiley: New York.

20.

Lindner

, Funk

G.C.

, Schwarz

, Kneidinger

, Kaider

, Schneeweiss

, Kramer

, Druml

2007. Hypernatremia in the critically ill is an independent risk factor for mortality. Am. J. Kidney Dis., 50:952–957.

21.

Lindner

, Funk

G.C.

, Lassnigg

, Mouhieddine

, Ahmad

S.A.

, Schwarz

, Hiesmayr

2010. Intensive care-acquired hypernatremia after major cardiothoracic surgery is associated with increased mortality. Intensive Care Med., 36:1718–1723.

22.

Maggiore

, Picetti

, Antonucci

, Parenti

, Regolisti

, Mergoni

, Vezzani

, Cabassi

, Fiaccadori

2009. The relation between the incidence of hypernatremia and mortality in patients with severe traumatic brain injury. Crit. Care, 13:R110.

23.

Javouhey

, Guérin

A.C.

, Amoros

, Haddak

, Ndiaye

, Floret

, Chiron

2006. Severe outcome of children following trauma resulting from road accidents. Eur. J. Pediatr., 165:519–525.

24.

Bahloul

, Ben Hamida

, Chelly

, Chaari

, Kallel

, Dammak

, Rekik

, Bahloul

, Ben Mahfoudh

, Hachicha

, Bouaziz

2009. Severe head injury among children: prognostic factors and outcome. Injury, 40:535–540.

25.

Tude Melo

J.R.

, Di Rocco

, Blanot

, Oliveira-Filho

, Roujeau

, Sainte-Rose

, Duracher

, Vecchione

, Meyer

, Zerah

2010. Mortality in children with severe head trauma: predictive factors and proposal for a new predictive scale. Neurosurgery, 67:1542–1547.

26.

Adelson

P.D.

, Bratton

S.L.

, Carney

N.A.

, Chesnut

R.M.

, du Coudray

H.E.

, Goldstein

, Kochanek

P.M.

, Miller

H.C.

, Partington

M.D.

, Selden

N.R.

, Warden

C.W.

, Wright

D.W.

2003. Guidelines for the acute medical management of severe traumatic brain injury in infants, children, and adolescents. Chapter 17. Critical pathway for the treatment of established intracranial hypertension in pediatric traumatic brain injury. Pediatr. Crit. Care Med., 4,Suppl. 3:S65–S67.

27.

Schalen

, Messeter

, Nordstrom

C.H.

1992. Complications and side effects during thiopentone therapy in patients with severe head injuries. Acta anaesthesiologica Scandinavica, 36:369–377.

28.

Yanay

, Brogan

T.V.

, Martin

L.D.

2004. Continuous pentobarbital infusion in children is associated with high rates of complications. J. Crit. Care., 19:174–178.

29.

H.S.

, Cho

B.N.

, Kang

J.H.

, Kim

M.K.

, Oh

S.M.

, Park

S.H.

2010. Efficacy of low dose barbituate coma therapy for the patients with intractable intrcranial hypertension using the bispectral index monitoring. J. Korean Neurosurg. Soc., 47:252–257.

30.

Huynh

, Mabasa

V.H.

, Ensom

M.H.

2009. A critical review: does thiopental continuous infusion warrant therapeutic drug monitoring in the critical care population? Ther. Drug Monit., 31:153–169.

31.

Shore

P.M.

, Hand

L.L.

, Roy

, Trivedi

, Kochanek

P.M.

, Adelson

P.D.

2006. Reliability and validity of the Pediatric Intensity Level of Therapy (PILOT) scale: a measure of the use of intracranial pressure-directed therapies. Crit. Care Med., 34:1981–1987.

32.

Barzilay

, Somekh

1988. Diabetes insipidus in severely brain damaged children. J. Med., 19:47–64.

33.

Boughey

J.C.

, Yost

M.J.

, Bynoe

R.P.

2004. Diabetes insipidus in the head-injured patient. Am. Surg., 70:500–503.

34.

Zygun

D.A.

2009. Sodium and brain injury: do we know what we are doing? Crit. Care., 13:184.