Fluid Overload in Pediatric Severe Traumatic

Brain Injury
Casey Stulce, MD1; Andrew Reisner, MD2,3; Jason M. Kane, MD, MS1; H. Stella Shin, MD4;
Courtney McCracken, PhD4; Julie Williamson, DO2,4; Karen Walson, MD2; Matthew Paden, MD2,4

Objective: Pediatric traumatic brain injury is a major public health acute kidney injury, or increased PICU length of stay. (Pediatr Crit
problem in the United States. Hypertonic saline therapy is a well-es- Care Med 2019; XX:00–00)
tablished treatment in patients with severe traumatic brain injury Key Words: fluid overload; hypertonic saline; pediatric; traumatic
(Glasgow Coma Scale ≤ 8) who have intracranial hypertension. In brain injury
children, fluid overload is associated with increased mortality, ven-
tilator duration, and length of PICU stay, even when controlling for
severity of illness. This study reports prevalence of fluid overload in

P
pediatric patients with severe traumatic brain injury treated with 3% ediatric traumatic brain injury (TBI) is a major public
hypertonic saline and effect on clinical outcomes. health problem in the United States with almost half a
Design: Single-center retrospective chart review. million pediatric visits made to the emergency depart-
Setting: PICUs at two tertiary children’s hospitals. ment annually (1). Traumatic brain injuries encompass a spec-
Patients: One hundred thirty-eight patients with traumatic brain trum of disease ranging from mild to severe in both initial
injury with postresuscitation Glasgow Coma Scale less than or presentation and long-term neurologic outcomes.
equal to 8 who received hypertonic saline from September 1, Intracranial hypertension is a common sequela of se-
2010, to February 28, 2016, and intracranial pressure monitoring vere TBI. Evidence-based guidelines recommend treatment
and survived at least 24 hours from admission. of increased intracranial pressure (ICP) using a multisystem
Interventions: None. approach, including the use of hyperosmolar therapy (2).
Measurements and Main Results: We used fluid balance per- Hyperosmolar therapy creates an osmotic gradient that draws
centage greater than or equal to 10% as our definition of fluid water from the interstitium to the vascular space (3). Mannitol
overload. Ninety-one percent of patients less than 1 year old and hypertonic saline (HS) represent two of the most com-
had fluid overload on day 10 of admission compared with 47% monly used hyperosmolar therapies. Mannitol is a sugar al-
of patients greater than 1 year. Fluid overloaded patients did not cohol that works, among other mechanisms, as an osmotic
have increased mortality, acute kidney injury, PICU length of stay, diuretic to raise serum osmolarity (4). Although very effective
or ventilator days. Hypertonic saline was not the cause of fluid in lowering ICP, the brisk diuresis that occurs with mannitol
overload in these patients. in pediatric patients can lead to hypotension and an overall
Conclusions: Patients with severe traumatic brain injury do have decrease in cerebral perfusion pressure. In contrast, HS low-
high rates of fluid overload. However, fluid overload did not con- ers the ICP without the diuretic effect seen with mannitol and
tribute to mortality, longer days on the ventilator, increased risk of acts as an intravascular volume expander, thus supporting the
blood pressure. These hemodynamic differences have led to an
1
Department of Pediatrics, University of Chicago, Chicago, IL.
increased use of HS for the treatment of pediatric TBI.
2
Children’s Healthcare of Atlanta, Atlanta, GA.
Fluid overload is associated with increased morbidity and
3
Department of Neurosurgery, Emory University, Atlanta, GA.
mortality in the PICU (5–11). Critically ill patients with fluid
4
Department of Pediatrics, Emory University, Atlanta, GA.
overload have increased duration of mechanical ventilation
Dr. Williamson received $1,500 in internal grant support from the Depart- and longer ICU and hospital lengths of stay (LOSs), even when
ment of Anesthesiology at Emory University for an unrelated project. Dr. controlling for severity of illness (9). Critically ill children
Paden received multiple patents (both issued and pending) and multiple requiring mechanical ventilation are at high risk for early fluid
grants, all of which are not related to this project. The remaining authors
have disclosed that they do not have any potential conflicts of interest. overload due to the initial resuscitation fluid, secretion of anti-
For information regarding this article, E-mail: caseystulce@gmail.com diuretic hormone in response to positive-pressure ventilation,
Copyright © 2019 by the Society of Critical Care Medicine and the World increased capillary leak syndrome, and decreased mobility due
Federation of Pediatric Intensive and Critical Care Societies to sedation requirements (5, 12, 13). Patients with TBI are sus-
DOI: 10.1097/PCC.0000000000002134 ceptible to all of the above.

Pediatric Critical Care Medicine www.pccmjournal.org 1

Copyright © 2019 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited
 Stulce et al

We hypothesized that children with severe TBI who receive in previous studies (6, 14, 15). Admission weight was recorded
HS to treat intracranial hypertension would have high rates of as first measured weight in the CHOA system. If there was no
fluid overload and that HS volume used would contribute di- measured weight within the first 24 hours, then the estimated
rectly to fluid overload. The purpose of this study was to assess weight from the emergency department was used. Estimated
the frequency of fluid overload in pediatric patients with se- weights were calculated via the Broselow tape in the emergency
vere TBI and to evaluate the impact of fluid overload on clin- department.
ical outcomes including mechanical ventilation, PICU LOS,
and acute kidney injury (AKI). Assessment of AKI
AKI was defined according to the Kidney Disease Improving
Global Outcomes Guidelines: an increase in serum creatinine
MATERIALS AND METHODS
by 0.3 mg/dL within 48 hours or an increase in serum creati-
Design, Settings, and Patients nine by greater than or equal to 50% from baseline (16). Given
This retrospective study was conducted on patients admit- that infants have low baseline serum creatinine, we required
ted to the PICUs of Children’s Healthcare of Atlanta (CHOA) that serum creatinine reach a minimum of greater than or
between September 2010 and February 2016. The two multi- equal to 0.5 mg/dL prior to diagnosing AKI. The minimum
disciplinary PICUs at CHOA are located on separate medical creatinine was used in order to prevent small variations in
campuses. CHOA at Egleston is a level 1 trauma center with 36 serum creatinine from overestimating the true frequency of
PICU beds. CHOA at Scottish Rite is a level 2 trauma center AKI. Baseline creatinine was defined as the serum creatinine
with an active neurosurgery service and 38 PICU beds. Both at the time of hospital admission. We corrected serum creati-
centers use the same standardized TBI protocol for the care nine for fluid balance based on previously published data using
of these patients that includes the management of mechanical the following formula: corrected creatinine = measured creati-
ventilation, the use of sedating and paralytic agents, and the nine × (1 +[accumulated net fluid balance / total body water]),
use of osmolar therapy. Inclusion criteria were TBI and re- where total body water = 0.6 × weight (kg) (17).
ceipt of 3% HS within the first 10 days of admission to the
PICU. Exclusion criteria were postresuscitation Glasgow Coma Statistical Analysis
Scale (GCS) greater than 8, lack of ICP monitoring, and death The primary outcome studied was frequency of fluid over-
within the first 24 hours of admission. The Institutional Re- load. The secondary outcomes included effect of fluid over-
view Board (IRB) of CHOA approved the study protocol and load status on clinical outcomes and contribution of HS to
waived patient consent (IRB 15–111). the rate of fluid overload. Descriptive statistics were calcu-
lated for all variables of interest and included means and sds,
Data Collection median and interquartile ranges, or counts and percentages,
Data collection included the daily volume of HS received and where appropriate. Clinical characteristics, demographics,
total daily fluid intake and output (in mL) beginning at pre- and outcomes were compared among patients with and
sentation to the emergency department and continuing for without fluid overload on day 10. We used chi-square tests
up to 10 days. We recorded whether the HS was received as for categorical variables and Wilcoxon rank-sum test for con-
a bolus dose or a continuous infusion. Additional data col- tinuous variables.
lected included patient demographics, admission diagnoses, Because of differences in demographic and clinical charac-
radiographic diagnoses, markers for AKI (daily serum creati- teristics of patients with and without fluid overload, adjusted
nine values), and clinical outcomes including 28-day mortality, and unadjusted analyses were performed for the outcomes of
ventilator-free days, and PICU LOS. Ventilator-free days were interest. The adjusted models contained the primary predictor
defined as days alive and free from mechanical ventilation at fluid overload status (fluid balance ≥ 10% vs fluid balance <
day 28 from the day of admission. 10%) and controlled for diagnoses, presence of subdural he-
matoma, presence of epidural hematoma, use of furosemide,
Assessment of Fluid Overload age, and GCS score at admission. For the categorical outcomes
Fluid intake included all enterally and parenterally adminis- of death and AKI, unadjusted and adjusted outcomes were
tered fluids, including blood products and IV medications. obtained using logistic regression; the effect of fluid overload
Fluid output included urine, stool, nasogastric output, external on these outcomes is presented as odds ratios with 95% CIs. For
ventricular drain output, or any other output from surgical continuous outcomes, multivariable linear regression models
drains. The fluid balance on each day of admission for each pa- were used and contained the same variables described above.
tient was recorded based on the previously reported formula: Because PICU LOS, duration of mechanical ventilation, and
fluid balance % = (total amount of fluid intake [L] – total ventilator-free days were not normally distributed, data were
amount of fluid output [L])/ admission weight (kg) x 100% (8). log-transformed prior to statistical modeling. Resulting model
Fluid overload was defined as a fluid balance % greater than or estimates were on the log-scale but were back-transformed via
equal to 10% at day 10 of admission or death, whichever came exponentiation for ease of interpretation. Statistical analyses
first. The cutoff of greater than or equal to 10% for fluid over- were performed using SAS v. 9.4 (SAS Institute, Cary, NC), and
load was used given its association with poor clinical outcomes statistical significance was assessed at the 0.05 level.

2 www.pccmjournal.org XXX 2019 • Volume XX • Number XXX

Copyright © 2019 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited
 Neurocritical Care

TABLE 1. Comparison of Patient Characteristics Stratified by Fluid Overload at Day 10

Not Overloaded Fluid Overloaded
Characteristic (< 10%) n = 64 (47%) (≥ 10%) n = 74 (53%) p

All patients
Sex, n (%) 0.197
 Male 44 (68.8) 43 (58.1)
 Female 20 (31.3) 31 (41.9)
Age (yr), median (25th–75th) 9.0 (4.3–13.3) 3.4 (0.9–8.1) < 0.001
Age group, n (%) < 0.001
  Infant (< 1 yr old) 2 (3.1) 20 (27.0)
  ≥ 1 yr old 62 (96.9) 54 (73.0)
Weight at admission, median (25th–75th) 32.0 (18.5–55.6) 14.5 (10.0–30.0) < 0.001
Diagnosis, n (%) 0.014
  Nonaccidental trauma 8 (12.5) 25 (33.8)
  Traumatic brain injury
  Blunt 52 (81.3) 46 (62.2)
  Gunshot wound 4 (6.3) 3 (4.1)
Glasgow Coma Scale at presentation, median (25th–75th) 5 (3–7) 5 (3–7) 0.809

RESULTS Outcomes
Overall mortality rate for these patients was 22% (30/138).
Patient Characteristics
AKI was present in 33% of patients (Table 3). No patient re-
Demographic information is presented in Table 1. There were
quired continuous renal replacement therapy. There was no
200 patients who met inclusion criteria. Sixty-two patients were
significant difference in clinical outcomes for the fluid over-
excluded due to death within 24 hours, lack of ICP monitor-
loaded group when compared with the group without fluid
ing, or postresuscitation GCS greater than 8. This brought the
overload (Tables 3 and 4). After adjusting for differences in
total number of patients analyzed to 138. The majority of the
clinical and demographic variables, there was no increase in
patients were male (87/138, 63%), and the median age (inter-
odds of mortality (adjusted odds ratio [AOR], 1.30; 95% CI,
quartile range) was 5.8 years (1.7–11.4 yr). These patients had
0.47–3.59) when comparing fluid overloaded versus children
severe TBI with a median GCS at admission of 5. The admit-
without fluid overload. There was also no significant increase
ting diagnosis was nonaccidental trauma (NAT) in the setting
in odds of AKI (AOR, 1.71; 95% CI, 0.74–3.95). There was no
of child abuse in 24% of patients. The median weight of these
change in the mean PICU LOS (8.5 vs 9.3 d in nonfluid over-
patients was 20 kg. Twenty-two percent of these patients had
load vs fluid overloaded children, respectively). There was also
estimated weights from the emergency department because no
no change in the mean ventilator-free days (14.6 vs 13.0 d in
weight was recorded in the PICU during the first 24 hours.
nonfluid overload vs fluid overloaded children, respectively).
Fluid Overload
Overall, 53% of patients were fluid overloaded in this study DISCUSSION
(Table 1). Patients less than 1 year old were more likely to be Most previous studies have examined fluid overload in ge-
fluid overloaded. Ninety-one percent of infants under 1 year neral PICU populations. This study is the first to explore fluid
old were fluid overloaded. Patients with NAT were more likely overload specifically in the pediatric severe TBI population in
to be fluid overloaded than patients with other types of TBI. an attempt to better characterize if and how they differ from
Fluid overloaded patients received more furosemide dur- the broader PICU patients. The majority of the patients were
ing their first 10 days of admission (Table 2). Both peak and fluid overloaded in this study. Although more recent studies
average daily serum sodium values were higher in the fluid have looked at early fluid overload as a contributor to poor
overloaded group (Table 2). Patients with fluid overload re- outcomes (6, 7, 9), there is literature to support that fluid over-
ceived a larger volume of HS (55 mL/kg) when compared with load at any day of PICU admission is pathologic and greater
patients without fluid overload (38 mL/kg), although this dif- than or equal to 10% should be the threshold on which to
ference was not statistically significant in a univariate analysis initiate treatment (5, 8, 18–24). We chose to look at positive
(Table 2). fluid balances on day 10 of PICU admission to account for

Pediatric Critical Care Medicine www.pccmjournal.org 3

Copyright © 2019 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited
 Stulce et al

TABLE 2. Interventions Received and Sodium Measurements

Intervention/ Not Overloaded Fluid Overloaded
Laboratory Result (< 10%) n = 64 (47%) (≥ 10%) n = 74 (53%) p

Vasopressin use for diabetes insipidus (n = 37), n (%) 15 (23.4) 22 (29.7) 0.405
Furosemide use in first 10 d (n = 47), n (%) 17 (26.6) 30 (40.5) 0.084
Furosemide received (mg/kg), median (25th–75th) 0.71 (0.41–1.11) 3.5 (1.50–5.83) 0.002
Mannitol use in first 10 d (n = 32), n (%) 18 (28.1) 14 (18.9) 0.201
Mannitol received (g/kg), median (25th–75th) 0.50 (0.49–1.16) 0.75 (0.50–1.0) 0.834
Pentobarbital coma use (n = 43), n (%) 18 (28.1) 25 (34.3) 0.441
Decompressive craniectomy (n = 55), n (%) 24 (37.5) 31 (41.9) 0.599
HS continuous infusion received (n = 54), n (%) 30 (46.9) 24 (32.4) 0.083
Percentage of 3% HS relative to total fluids day 10, median (25th–75th) 6.2 (3.0–12.5) 6.1 (3.0–10.2) 0.659
Total volume of HS received (L/kg), median (25th–75th) 0.038 (0.016–0.074) 0.055 (0.025–0.119) 0.054
Average daily serum sodium (mmol/L), median, (25th–75th) 144 (141–148) 147 (143–152) 0.019
Peak serum sodium (mmol/L), median (25th–75th) 153 (146–161) 156 (151–163) 0.032
HS = hypertonic saline.

TABLE 3. Comparison of Outcomes Stratified by Fluid Overload

Not Overloaded Fluid Overloaded
Outcome (< 10%) n = 64 (47%) (≥ 10%) n = 74 (53%) p

Mortality, n (%) 12 (18.8) 18 (24.3) 0.429

PICU length of stay, median (25th–75th) 11.7 (7.4–17.3) 12.7 (8.7–16.8) 0.492
Ventilator-free days, median (25th–75th) 18.0 (5.5–21.0) 15.0 (0–19.0) 0.089
Acute kidney injury, n (%) 18 (28.1) 28 (37.8) 0.227

TABLE 4. Adjusted Outcomes Analysis

Outcome Unadjusted p Adjusted p

Mortality, OR (95% CI)

  Fluid overload, yes vs no 1.39 (0.61–3.17) 0.430 1.30 (0.47–3.59) 0.618
PICU length of stay, mean (95% CI)
  No fluid overload (reference) 10.7 (9.0–12.7) 8.5 (6.6–10.9)
  Fluid overload 11.2 (9.6–13.2) 0.670 9.3 (7.2–11.8) 0.464
Ventilator-free days, mean (95% CI)
  No fluid overload (reference) 14.3 (12.1–16.4) 14.6 (11.4–17.7)
  Fluid overload 12.0 (10.0–14.0) 0.135 13.0 (9.9–16.1) 0.300
Acute kidney injury, OR (95% CI)
  Fluid overload, yes vs no 1.56 (0.76–3.19) 0.229 1.71 (0.74–3.95) 0.208
OR = odds ratio.

the majority of doses of HS received. The fluid overloaded co- overloaded group received an even a greater amount of other
hort did receive a larger volume of HS than the nonfluid over- fluids not evaluated in this retrospective review. The fluid over-
loaded cohort. Interestingly, however, the percentage of fluid load seen in these patients was not the result of HS, given that
intake that was attributed to HS was relatively low and even HS accounted for less than 10% of the total fluids in the great
less in the fluid overloaded cohort. This indicates that the fluid majority of patients. It is important to note that the severe TBI

4 www.pccmjournal.org XXX 2019 • Volume XX • Number XXX

Copyright © 2019 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited
 Neurocritical Care

protocol practiced in this study does not address other aspects with severe TBI. Trauma patients, unlike sepsis or acute respira-
of fluid management. Infants had the highest rate of fluid over- tory distress syndrome patients with vasodilatory shock, often
load in this study. Other studies have also shown that younger have blood loss and have intravascular volume depletion from
patients are more likely to be fluid overloaded (10). The reason hemorrhage on presentation. Therefore, the initial weight used
behind the high prevalence of fluid overload in infants is un- to calculate fluid overload may be falsely low. Also, in patients
clear. One possible explanation is that the infant cohort has a with severe neurologic injury, weaning from mechanical ven-
high rate of NAT as a cause of their TBI. Previous studies have tilation may be due more to impaired airway protection and
shown that victims of NAT have delayed time to presentation neurologic status than impaired gas exchange or pulmonary
and may present dehydrated from vomiting or poor feeding mechanics (36–38). Therefore, even though fluid overload has
(25–27). Their initial weight, therefore, may be falsely low. In been shown to be associated with impaired gas exchange, in this
addition, younger patients are smaller and errors in fluid bal- study population, it may be the patients’ neurologic injury that is
ance calculation would have a greater impact on fluid overload the limiting factor in their successful extubation (5, 39).
percent. In a study of fluid overload in postoperative cardiac Although not associated with fluid overload, patients in this
surgery infants, it was found that only the weight-based (rather study had high rates of AKI. A large, international, multicenter
than fluid balance based) method of determining fluid over- study looking at the epidemiology of AKI in PICU patients
load correlated with poor outcome (28). Given that infants showed an AKI prevalence of 27% (40), whereas the preva-
may have higher insensible losses than older children perhaps lence of AKI in this study was 33%. AKI has been associated
the weight-based method of determining fluid overload is with the use of HS and resulting in higher serum sodium, but
more accurate in this population (29). this was seen with sustained sodium greater than 170 mmol/L
Positive fluid balance in PICU patients has been shown to (41). The patients in the present study did not have this degree
be associated with worse clinical outcomes (7, 11), but to date, of hypernatremia, so it is unclear what other factors may have
there has not been specific scrutiny of this finding in pediatric contributed to the rate of AKI.
trauma patients. Historically, there was a preference to use There are a number of limitations to this study common
diuretics to promote hypovolemia and decrease brain edema to those of a single-center, retrospective observational design.
in patients with severe TBI (30), but this is no longer current First, because it is an observational study, we cannot make
practice. A study by Clifton et al (31) looked retrospectively statements regarding causal relationships among fluid overload
at adult patients with severe TBI and found that a total fluid status and outcomes observed. Although there are two different
balance more negative than 594 mL was associated with an campuses represented in this study, there is a single “severe TBI
increased risk of poor outcome. These patients with the most committee” that makes recommendations for the entire system.
negative fluid balance also received more therapy for ICP con- This study also covers 5 years of patient data, during which
trol than other patients, including increased mannitol doses. other aspects of management may have changed. Another lim-
Mannitol is a potent osmotic diuretic and can lead to contrac- itation is the use of estimated weights if a measured weight was
tion of the intravascular volume and resultant hypotension not entered in the record within the first 24 hours. Estimated
(32). The poor outcome, therefore, could be from more severe weights are known to have variable accuracy in the emergency
brain injury that necessitated increased doses of mannitol or department and are less accurate in older and heavier children
it could be due to a direct effect of the negative fluid balance, (42, 43). This may have led to incorrect calculations of fluid
such as reduced cerebral perfusion pressure. balance % in these patients. The lack of measured weights at
Positive fluid balance in adult patients with severe TBI has admission is concerning for patient care and has led us to im-
been associated with pulmonary edema and longer hospital plement more education regarding the importance of accurate
LOS (33); however, positive fluid balance was not shown to weights for all patients. Another limitation is the lack of data
influence refractory intracranial hypertension nor mortality regarding ICP measurements in these patients. Although the in-
in this population. Given that non-neurologic organ dysfunc- itial GCS scores were similar among the cohorts, one group may
tion has been associated with worse outcomes in the severe have had worse intracranial hypertension, which could have
TBI population (34), it is important to try to limit pulmonary influenced outcomes. Finally, stratification of patient outcomes
complications when possible. Current practice guidelines rec- by the use of validated trauma, TBI, and pediatric severity of ill-
ommend euvolemia for patients with severe TBI (2, 35) but ness scoring systems was not able to be performed in our cohort
otherwise offer no recommendations on fluid management. due to lack of routine collection of the necessary data fields but
The rate of fluid overload in critically ill pediatric patients should be included in future work addressing this topic.
in this study was higher than previously reported data (6, 7). In
spite of this, the fluid overload did not appear to have a negative CONCLUSIONS
effect on the clinical outcomes studied. This finding contrasts The rate of fluid overload in the pediatric severe TBI popula-
prior data that have demonstrated an association between fluid tion is high, especially in infants under 1 year old. However,
overload and negative clinical outcomes, including longer dura- unlike other PICU cohorts, fluid overload does not appear to
tion of mechanical ventilation, AKI, PICU LOS, and mortality be associated with worse clinical outcomes in the severe TBI
(5–9, 14). The reasons behind this discrepancy are not clear but patients. HS does not appear to be a contributing factor to the
may be due to the fact that all patients in this study were patients high rate of fluid overload in severe TBI.

Pediatric Critical Care Medicine www.pccmjournal.org 5

Copyright © 2019 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited
 Stulce et al

REFERENCES 21. Hassinger AB, Valentine SL: Self-reported management of IV flu-

1. Faul M XL, Wald MM, Coronado VG: Traumatic Brain Injury in the ids and fluid accumulation in children with acute respiratory failure.
United States: Emergency Department Visits, Hospitalizations and Pediatr Crit Care Med 2018; 19:e551–e554
Deaths 2002–2006. Atlanta (GA), Centers for Disease Control 22. Brierley J, Carcillo JA, Choong K et al: Clinical practice parameters for
and Prevention, National Center for Injury Prevention and Control, hemodynamic support of pediatric and neonatal septic shock: 2007
2010 update from the American college of critical care medicine. Crit Care
2. Kochanek PM, Carney N, Adelson PD et al: Guidelines for the acute Med 2009; 37:666–688.
medical management of severe traumatic brain injury in infants, chil- 23. Selewski DT, Goldstein SL: The role of fluid overload in the prediction
dren, and adolescents--second edition. Pediatr Crit Care Med 2012; of outcome in acute kidney injury. Pediatr Nephrol 2018; 33:13–24
13(Suppl 1):S1–82 24. Cortina G, McRae R, Hoq M et al: Mortality of critically ill children
3. Qureshi AI, Suarez JI: Use of hypertonic saline solutions in treat- requiring continuous renal replacement therapy: Effect of fluid over-
ment of cerebral edema and intracranial hypertension. Crit Care Med load, underlying disease, and timing of initiation. Pediatr Crit Care
2000; 28:3301–3313 Med 2018; 20:314–322
4. Kamel H, Navi BB, Nakagawa K, et al: Hypertonic saline versus 25. Miller Ferguson N, Sarnaik A, Miles D, et al; Investigators of the
mannitol for the treatment of elevated intracranial pressure: A Approaches and Decisions in Acute Pediatric Traumatic Brain Injury
meta-analysis of randomized clinical trials. Crit Care Med 2011; (ADAPT) Trial: Abusive head trauma and mortality-an analysis from an
39:554–559 international comparative effectiveness Study of children with severe
5. Arikan AA, Zappitelli M, Goldstein SL, et al: Fluid overload is associ- traumatic brain injury. Crit Care Med 2017; 45:1398–1407
ated with impaired oxygenation and morbidity in critically ill children. 26. Hettler J, Greenes DS: Can the initial history predict whether a child
Pediatr Crit Care Med 2012; 13:253–258 with a head injury has been abused? Pediatrics 2003; 111:602–607
6. Bhaskar P, Dhar AV, Thompson M, et al: Early fluid accumulation in 27. Jenny C, Hymel KP, Ritzen A, et al: Analysis of missed cases of abu-
children with shock and ICU mortality: A matched case-control study. sive head trauma. JAMA 1999; 281:621–626
Intensive Care Med 2015; 41:1445–1453 28. Hazle MA, Gajarski RJ, Yu S, et al: Fluid overload in infants following
7. Chen J, Li X, Bai Z, et al: Association of fluid accumulation with clin- congenital heart surgery. Pediatr Crit Care Med 2013; 14:44–49
ical outcomes in critically ill children with severe sepsis. PLoS One 29. Heeley AM, Talbot NB: Insensible water losses per day by hospital-
2016; 11:e0160093 ized infants and children. AMA Am J Dis Child 1955; 90:251–255
8. Goldstein SL, Currier H, Graf Cd, et al: Outcome in children re- 30. Eker C, Asgeirsson B, Grände PO, et al: Improved outcome after
ceiving continuous venovenous hemofiltration. Pediatrics 2001; severe head injury with a new therapy based on principles for brain
107:1309–1312 volume regulation and preserved microcirculation. Crit Care Med
9. Li Y, Wang J, Bai Z, et al: Early fluid overload is associated with acute 1998; 26:1881–1886
kidney injury and PICU mortality in critically ill children. Eur J Pediatr 31. Clifton GL, Miller ER, Choi SC, et al: Fluid thresholds and outcome
2016; 175:39–48 from severe brain injury. Crit Care Med 2002; 30:739–745
10. Sinitsky L, Walls D, Nadel S, et al: Fluid overload at 48 hours is 32. Fink ME: Osmotherapy for intracranial hypertension: Mannitol versus
associated with respiratory morbidity but not mortality in a general hypertonic saline. Continuum (Minneap Minn) 2012; 18:640–654
PICU: Retrospective cohort study. Pediatr Crit Care Med 2015; 33. Fletcher JJ, Bergman K, Carlson G, et al: Continuous renal replace-
16:205–209 ment therapy for refractory intracranial hypertension? J Trauma 2010;
11. Alobaidi R, Morgan C, Basu RK, et al: Association between fluid bal- 68:1506–1509
ance and outcomes in critically ill children: A systematic review and 34. Zygun DA, Kortbeek JB, Fick GH, et al: Non-neurologic organ dys-
meta-analysis. JAMA Pediatr 2018; 172:257–268 function in severe traumatic brain injury. Crit Care Med 2005;
12. Farge D, De la Coussaye JE, Beloucif S, et al: Interactions between 33:654–660
hemodynamic and hormonal modifications during PEEP-induced 35. Dutton RP, McCunn M: Traumatic brain injury. Curr Opin Crit Care
antidiuresis and antinatriuresis. Chest 1995; 107:1095–1100 2003; 9:503–509
13. Soni N, Williams P: Positive pressure ventilation: What is the real 36. Kutchak FM, Debesaitys AM, Rieder Mde M, et al: Reflex cough PEF
cost? Br J Anaesth 2008; 101:446–457 as a predictor of successful extubation in neurological patients. J Bras
14. Foland JA, Fortenberry JD, Warshaw BL, et al: Fluid overload before Pneumol 2015; 41:358–364
continuous hemofiltration and survival in critically ill children: A retro- 37. Namen AM, Ely EW, Tatter SB, et al: Predictors of successful extu-
spective analysis. Crit Care Med 2004; 32:1771–1776 bation in neurosurgical patients. Am J Respir Crit Care Med 2001;
15. Lane PH, Mauer SM, Blazar BR, et al: Outcome of dialysis for acute 163:658–664
renal failure in pediatric bone marrow transplant patients. Bone 38. Teixeira SN, Osaku EF, Costa CR, et al: Comparison of proportional
Marrow Transplant 1994; 13:613–617 assist ventilation plus, T-tube ventilation, and pressure support venti-
16. Khwaja A: KDIGO clinical practice guidelines for acute kidney injury. lation as spontaneous breathing trials for extubation: A randomized
Nephron Clin Pract 2012; 120:c179–c184 Study. Respir Care 2015; 60:1527–1535
17. Basu RK, Andrews A, Krawczeski C, et al: Acute kidney injury based 39. Upadya A, Tilluckdharry L, Muralidharan V, et al: Fluid balance and
on corrected serum creatinine is associated with increased morbidity weaning outcomes. Intensive Care Med 2005; 31:1643–1647
in children following the arterial switch operation. Pediatr Crit Care 40. Kaddourah A, Basu RK, Bagshaw SM, et al; AWARE Investigators:
Med 2013; 14:e218–e224 Epidemiology of acute kidney injury in critically ill children and young
18. Soler YA, Nieves-Plaza M, Prieto M, et al: Pediatric risk, injury, failure, adults. N Engl J Med 2017; 376:11–20
loss, end-stage renal disease score identifies acute kidney injury and 41. Gonda DD, Meltzer HS, Crawford JR, et al: Complications associated
predicts mortality in critically ill children: A prospective study. Pediatr with prolonged hypertonic saline therapy in children with elevated in-
Crit Care Med 2013; 14:e189–e195 tracranial pressure. Pediatr Crit Care Med 2013; 14:610–620
19. Gillespie RS, Seidel K, Symons JM: Effect of fluid overload and dose 42. So TY, Farrington E, Absher RK: Evaluation of the accuracy of dif-
of replacement fluid on survival in hemofiltration. Pediatr Nephrol ferent methods used to estimate weights in the pediatric population.
2004; 19:1394–1399 Pediatrics 2009; 123:e1045–e1051
20. Goldstein SL, Somers MJ, Baum MA, et al: Pediatric patients with 43. Wells M, Goldstein LN, Bentley A, et al: The accuracy of the Broselow
multi-organ dysfunction syndrome receiving continuous renal replace- tape as a weight estimation tool and a drug-dosing guide - A system-
ment therapy. Kidney Int 2005; 67:653–658 atic review and meta-analysis. Resuscitation 2017; 121:9–33

6 www.pccmjournal.org XXX 2019 • Volume XX • Number XXX

Copyright © 2019 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited