The Journal of Emergency Medicine, Vol. -, No. -, pp.

1–10, 2017
Published by Elsevier Inc.
0736-4679/$ - see front matter

http://dx.doi.org/10.1016/j.jemermed.2017.03.014

Clinical
Review

EMERGENCY MEDICINE MYTHS: CEREBRAL EDEMA IN PEDIATRIC DIABETIC

KETOACIDOSIS AND INTRAVENOUS FLUIDS

Brit Long, MD* and Alex Koyfman, MD†

*Department of Emergency Medicine, San Antonio Military Medical Center, Fort Sam Houston, Texas and †Department of Emergency
Medicine, The University of Texas Southwestern Medical Center, Dallas, Texas
Reprint Address: Brit Long, MD, Department of Emergency Medicine, San Antonio Military Medical Center, 3841 Roger Brooke Dr., Fort Sam
Houston, TX 78234.

, Abstract—Background: Pediatric diabetic ketoacidosis
(DKA) is a disease associated with several complications
that can be severe. One complication includes cerebral edema
(CE), and patients may experience significant morbidity with
this disease. Objective: This review evaluates the myths
concerning CE in pediatric DKA including mechanism, pre-
sentation of edema, clinical assessment of dehydration, and
association with intravenous (i.v.) fluids. Discussion: Multi-
ple complications may occur in pediatric DKA. CE occurs
in < 1% of pediatric DKA cases, though morbidity and mor-
tality are severe without treatment. Several myths surround
this disease. Subclinical CE is likely present in many patients
with pediatric DKA, though severe disease is rare. A multi-
tude of mechanisms likely account for development of CE,
including vasogenic and cytotoxic causes. Clinical dehydra-
tion is difficult to assess. Literature has evaluated the associ-
ation of fluid infusion with the development of CE, but most
studies are retrospective, with no comparator groups. The
few studies with comparisons suggest fluid infusion is not
associated with DKA. Rather, the severity of DKA with
higher blood urea nitrogen and greater acidosis contribute
to CE. Multiple strategies for fluid replacement exist. A
bolus of 10 mL/kg of i.v. fluid is likely safe, which can be
repeated if hemodynamic status does not improve. Conclu-
sions: Pediatric CE in DKA is rare but severe. Multiple
mechanisms result in this disease, and many patients
This review does not reflect the views or opinions of the U.S.
government, Department of Defense, U.S. Army, U.S. Air
Force, or SAUSHEC EM Residency Program.
experience subclinical CE. Intravenous fluids are likely not
associated with development of CE, and 10-mL/kg or 20-
mL/kg i.v. bolus is safe. Published by Elsevier Inc.
, Keywords—cerebral edema; dehydration; diabetic
ketoacidosis; fluid infusion; pediatric
INTRODUCTION
Diabetes mellitus is a common chronic disease among
children, with increasing frequency in type 1 and 2 dia-
betes (1–5). One complication is diabetic ketoacidosis
(DKA), which has an incidence of 25% in known type
1 diabetics (3–8). Close to one-third of patients at the
time of initial diagnosis of diabetes have DKA, and chil-
dren younger than 5 years or age are at high risk for DKA
(3–7). Other risk factors for DKA at the time of diagnosis
include ethnic minority, smaller body mass index,
delayed treatment, infectious trigger, and lack of health
insurance (1,2,8). Though most commonly occurring in
type 1 diabetics, patients with type 2 diabetes can also
experience DKA, as 5–25% of patients with type 2
diabetes are in DKA at time of diagnosis (4,8,9). The
most common cause is insulin omission, though
infection is another common trigger (1,2).
Pediatric DKA is demonstrated by hyperglycemia
(serum glucose > 200 mg/dL), anion gap metabolic
acidosis, and ketonemia (1,10–14). Disease develops

RECEIVED: 25 February 2017;

ACCEPTED: 8 March 2017

1
2 B. Long and A. Koyfman
absolute or relative deficiency in insulin and excess
counterregulatory hormones, resulting in dehydration
and electrolyte abnormalities. The mainstay of
treatment for these patients includes rehydration with
fluids, insulin, and potentially potassium repletion (10–
14). One major complication is cerebral edema.
Cerebral Edema
Cerebral edema (CE) is clinically apparent and life threat-
ening in 0.5–1% of patients with DKA (2,14–17). Though
the mortality of DKA is < 1%, CE accounts for a
significant proportion of these deaths due to brain
herniation, which can occur prior to initiation of
treatment (15,17). Mortality ranges from 20–90%, with
one-fourth of survivors suffering permanent neurologic
deficits (1,2,14–16,18). However, CE may be
asymptomatic or subtle, with minor mental status
changes, which can appear in many cases of pediatric
DKA (19–22). This severe complication presents most
commonly within the first 7 h of treatment (66%), with
33% presenting 10–24 h after initiation of treatment in
type 1 and type 2 diabetics (15,16,23–25).
The diagnosis is clinical, as approximately 40% of pa-
tients with CE display normal neuroimaging
(1,2,22,26,27). However, diagnosis can be difficult,
depending on the presentation. Muir et al. published
several criteria for diagnosing cerebral edema, consisting
of diagnostic criteria, major criteria, and minor criteria;
this is demonstrated in Table 1 (26). A significant consid-
eration is close evaluation of patient neurologic status,
with frequent reassessments during management. If sus-
pected or diagnosed, mannitol is the most common first-
line therapy at 1 g/kg i.v., though hypertonic saline (3%)
is an option at 5–10 mL/kg i.v. (1,2,10–13).
This review will evaluate the literature concerning CE
in pediatric DKA, specifically the presentation, underlying
mechanism, and potential association with fluid infusion.
Authors conducted a search of Google Scholar, PubMed,
Table 1. DKA Diagnostic Criteria (26)
Pediatric DKA Cerebral Edema Diagnostic Criteria
Diagnostic Criteria: abnormal motor or verbal response to pain,
decorticate or decerebrate posture, cranial nerve palsy,
abnormal neurologic respiratory pattern
Major Criteria: altered mentation/fluctuating level of
consciousness, heart rate decelerations (more than 20 beats/
min) not improved with hydration or sleep, age-inappropriate
incontinence
Minor Criteria: vomiting, headache, lethargy or difficulty arousing
from sleep, diastolic blood pressure > 90 mm Hg, age < 5 years
Diagnosis: 1 diagnostic criterion, 2 major criteria, or 1 major and 2
minor criteria
*Sensitivity 92% and Specificity 96%
DKA = diabetic ketoacidosis.
Google FOAM, and Medline. We sought randomized tri-
als, case controls, case series, and chart reviews that
compared groups including patients with DKA and CE
and those in DKA with no CE. Much of the literature is
relegated to descriptive studies that lack comparator
groups. Few randomized studies are available, with the
majority including case control and case series (27,28).
DISCUSSION
A great deal of literature has examined risk factors for the
development of CE, with the goal of predicting patients at
greater risk for CE. Since the first description of this dis-
ease in 1936, a great deal of study has focused on eliciting
the cause of CE, means of prevention, and treatment (29).
However, CE in pediatric DKA is still a mysterious dis-
ease. To evaluate the association of i.v. fluid and CE, first
we must evaluate the clinical spectrum of CE and the pro-
posed mechanism, as well as the clinical assessment of
dehydration in these patients and management of DKA.
Myth: Cerebral Edema in Pediatric DKA is Rare and
Always Clinically Apparent
CE in pediatric DKA that is clinically overt with marked
neurologic change is infrequent (20–23,30). Subtle
edema occurs in the majority of patients with DKA, as
studies using neuroimaging (computed tomography [CT]
or magnetic resonance imaging [MRI]) in children with
DKA have demonstrated the presence of edema before
treatment is initiated and during therapy
(14,15,23,31,32). Patients who have abnormal mental
status during treatment are likelier to possess subtle CE,
defined by cerebral ventricle narrowing, than those with
normal neurologic status during treatment (33). Any
abnormal neurologic assessment, including abnormal
Glasgow Coma Scale score (GCS), is associated with
higher frequency of MRI changes (16,22,26,33). Krane
et al. found edema on head CT in 6 patients treated for
DKA, though none of these patients experienced
clinically evident signs of the disease (22). Cerebral edema
is not rare in pediatric DKA, but the severe form likely rep-
resents the extreme representation of a disease spectrum.
Bottom Line: Cerebral edema occurs along a spectrum
in pediatric DKA and is likely more common than origi-
nally thought. However, the form of edema that results in
herniation is likely rare.
Myth: The Mechanism of Cerebral Edema is
Predominantly Due to Rapid Osmotic Changes with
Treatment
Many published treatment recommendations for pediatric
DKA attempt to minimize the risk of CE, and providers
A Clinical Review Evaluating Myths Surrounding Cerebral Edema in Pediatric Diabetic Ketoacidosis
may associate rate of fluid infusion or size of bolus to the
development of CE (1,2,27,28). Multiple mechanisms for
development of CE have been hypothesized, including
vasogenic edema from blood-brain barrier destruction, os-
motic edema due to fluid therapy, and cytotoxic edema
from ischemia.
Vasogenic edema refers to damage of the cerebral
vascular endothelial layer resulting in increased blood-
brain barrier dysfunction, allowing abnormal diffusion
of fluid into the central nervous system (16,19,34,35).
This has been suggested from observational data based
on MRI showing abnormal diffusion of fluid into the
brain (16,19,34,35). Increased cerebral blood flow
(CBF) and oxygenation have also been found despite
hypocapnia, which suggests a vasogenic component to
cerebral edema (34–36). However, these studies are
small, with patients without CE. This may contribute,
but the data are not definitive.
Another mechanism commonly thought to result in
edema is serum osmotic changes. This is based on the
thought that osmolyte accumulation in the brain is due to
the presence of the hyperosmolar state in DKA. Fluids re-
sulting in a rapid decrease in the extracellular osmolar state
would cause brain swelling (17,37,38). This is one of the
predominant theories behind the recommendation for
slow fluid and electrolyte replacement over 48 h, rather
than in bolus form (39–42). Although this hypothesis is
simple and easy to comprehend, data are lacking to
support this (39–42). One study suggests that patients
receiving replacement over 12 h vs. 48 h demonstrate no
higher risk of edema (43). DKA is a hyperosmolar state,
and large osmotic shifts occur in all patients with DKA;
however, cerebral edema develops in a small percentage
of patients with DKA (1,2,15,16,43). This mechanism
may contribute, but it is not the sole component of edema.
The other thought is that the development of CE is due
to central nervous system hypoperfusion. During DKA,
CBF is decreased prior to treatment (15,19,30,44). This
hypoperfusion can result in cytotoxic edema due to
ischemia (15,19,30,44). Levels of cerebral lactate
increase and brain levels of high-energy phosphate
decrease when DKA is left untreated, suggesting poor
CBF (45). Thus, it may be that CE is a consequence of hy-
poperfusion during DKA. This is similar to cytotoxic
edema in ischemic stroke. One descriptive study found
no evidence of edema on neuroimaging at the initial
time of neurologic decline, though patients demonstrated
severe symptoms (26). Imaging completed several hours
to days later showed evidence of CE including hemor-
rhage or cerebral infarction in several patients (46–50).
Rather than osmotic changes, hypoperfusion during
severe DKA results in neurologic decompensation
(15,19,46–50). Once ischemic insult has occurred,
3
vasogenic edema through hyperglycemia may further
increase neurologic damage (1,2,16).
Bottom Line: The mechanism of cerebral edema is
complex and not associated with just one factor, and hy-
poperfusion during severe DKA is likely a contributor to
cerebral edema.
Myth: Cerebral Edema is Directly Linked to Greater
Rates of Intravenous Fluid Infusion or Larger Fluid
Boluses
Many have sought associations between DKA treatment
and CE development. However, a review of the literature
demonstrates a paucity of well-constructed, controlled
trials. A summary of this literature is demonstrated in
Table 2 (14,15,18,24,25,27,28,31,32,39,51–57). The
majority of studies are poorly controlled and
retrospective, with no control or comparison groups.
One of the first descriptions of CE and i.v. fluid
infusion was suggested in 1971, which was an
observational study with cerebrospinal fluid pressure
measured during treatment (58). No cases of CE were
diagnosed, but the authors state that treatment raised ce-
rebrospinal fluid pressure (58). Another example is a
study in 1988, which is a review of 42 cases of DKA
that found increased fluid administration rate to be asso-
ciated with decreased time to herniation (39). Though this
study utilized no controls, this is one of the primary
studies used for the argument that increased fluid infusion
causes CE and harm (39).
One of the first studies evaluating risk factors for cere-
bral edema with comparator groups was conducted by
Glaser et al. in 2001 (15). This retrospective study evalu-
ated patients under age 18 years with CE in 10 U.S. hos-
pitals, though investigators did not use specific criteria for
CE diagnosis. Investigators compared 61 cases of CE
with 174 control cases of DKA, defined by low pH or bi-
carbonate with ketonuria and serum glucose > 300 mg/
dL. They randomly selected patients with DKA but no ce-
rebral edema, and matched patients with regard to age,
onset of diabetes (new or established diagnosis), initial
serum glucose, and initial venous pH. Investigators
used logistic regression to compare three groups. Authors
evaluated i.v. fluid administration by the volume infused
per kilogram of weight per hour. The adjusted relative
risk (RR) for i.v. fluid was 1.1 (95% confidence interval
[CI] 0.4–3.0). Factors associated with CE include lower
initial partial pressures of arterial carbon dioxide (RR
3.4 for each decrease by 7.8 mm Hg, 95% CI 1.9–6.3),
higher blood urea nitrogen (BUN) (RR 1.7 for each in-
crease of 9 mg/dL or 3.2 mmol/L, 95% CI 1.2–25), and
bicarbonate therapy (RR 4.2, 95% CI 1.5–12.1)
(15,27,28).
4 B. Long and A. Koyfman

Table 2. Literature Summary of CE in Pediatric DKA

Study Year Cases Design Results

Rosenbloom 1980 17 cases of CE Case series CE not related to treatment

et al. (25)
Duck, Wyatt (39) 1988 42 cases of CE Case series CE may be related to fluids, with
edema in severe dehydration
Rosenbloom (24) 1990 69 cases of CE Case series CE is not related to treatment of
DKA including fluid
Bello & Sotos (54) 1990 11 cases of CE, 20 control cases Retrospective chart review CE may be related to osmolarity
change
Mel & Werther (51) 1995 6 cases of CE in 3134 DKA cases Retrospective evaluation of two CE not related to treatment
treatment groups
Hale et al. (52) 1997 4 cases of CE, 10 control cases Retrospective case-control CE may be related to decline in
serum sodium levels, not fluid
Mahoney et al. (18) 1999 9 cases of CE, 195 cases of DKA Retrospective chart review CE may be related to fluid over
50 mL/kg; CE related to
dehydration
Glaser et al. (15) 2001 61 cases of CE, 174 cases of DKA Retrospective case-control CE not related to treatment with
fluids or insulin
Edge et al. (14) 2001 34 cases of CE, 2940 DKA cases Descriptive Study CE can occur with DKA, though
fluid choice likely not
associated
Felner & White (55) 2001 0.3–0.5% CE in 520 DKA cases Retrospective evaluation of two CE risk similar between treatment
treatment groups groups, fluid likely not related
Marcin et al. (53) 2002 61 cases of CE Retrospective chart review CE not related to fluid infusion
Lawrence et al. (31) 2005 17 cases of CE, 28 control cases Retrospective case-control CE not associated with fluid
Edge et al. (32) 2006 43 cases of CE, 169 controls Retrospective case-control CE associated with more fluid, but
dehydrated patients received
more fluid
Hsia et al. (56) 2014 2.0-5.2% suspected cases of CE Retrospective cohort study Adverse outcomes not related to
in 163 DKA cases evaluating two treatment fluid infusion or osmolarity
groups change
Bakes et al. (57) 2015 50 cases of DKA randomized to Randomized control trial Higher volume infusion related to
separate infusion strategies (n = 25 in each arm) faster resolution of metabolic
normalization; no cases of CE

CE = cerebral edema; DKA = diabetic ketoacidosis.

Lawrence et al. conducted a case-control study,
including patients younger than 16 years with DKA and
cerebral edema (31). Investigators evaluated 17 cases of
CE and 28 controls with DKA, defined by low pH or bi-
carbonate with ketonuria. Cerebral edema was found on
initial presentation of DKA in 19% of patients, though
no specific definition was utilized for cerebral edema. In-
vestigators evaluated i.v. fluid through the volume infused
per kilogram per hour. Cerebral edema was associated
with lower initial bicarbonate, higher serum BUN, and
higher initial glucose. Fluid infusion rates were not asso-
ciated with significant difference in cerebral edema,
though adjusted RR or odds ratio (OR) were not reported
(27,28,31).
Edge et al. utilized a case-control study to identify 43
cases of cerebral edema in 2940 patients with DKA,
defined by low pH or bicarbonate with ketonuria (32). In-
vestigators selected 169 patients as control subjects. CE
was defined by deterioration of mental status with associ-
ated signs of increased intracranial pressure (hyperten-
sion and bradycardia, blurred disc margin, abnormal
motor posturing, squinting, respiratory abnormalities).
Investigators evaluated the total amount of i.v. fluid,
which was divided into tertiles of total i.v. fluid adminis-
tered. Importantly, unlike the prior two studies by Glaser
et al. and Lawrence et al., this study measured infused
volume without correcting for patient weight (15,31).
The study discarded a significant number of cases and
control patients due to misclassification, which created
a large number of unmatched patients. Investigators
performed a conditional analysis (consisting of only
appropriately matched cases) restricted to a small
subset of patients with complete sets of data. With
inclusion of all patients, the results of this study
demonstrate an association of higher infusion volumes
with CE within the first 3 h of treatment (OR 7.3; 95%
CI 1.51–35.12) and within 4 h (OR 6.55; 95% CI
1.38–30.97). When patients with incomplete data were
not analyzed, results were similar, with much larger
CIs. However, authors did not utilize BUN in matching
or adjusting for their analysis. This is important, as
higher BUN is associated with greater dehydration, and
in this study, patients with severe dehydration were
given larger boluses of i.v. fluids (27,28,32).
Other studies have suggested fluid infusion is not
correlated with cerebral edema. Rosenbloom et al. did
not find an association between fluid infusion and edema,
and two episodes of CE occurred in patients with oral
A Clinical Review Evaluating Myths Surrounding Cerebral Edema in Pediatric Diabetic Ketoacidosis
rehydration only, and a second study by Rosenbloom
found similar results, with no correlation between fluid
and edema (24,25). Mel and Werther conducted a
20-year retrospective study evaluating two different fluid
rehydration protocols, with one rapidly correcting dehy-
dration within 6 h and the other using rehydration over
24 h (51). No difference in rates of cerebral edema
occurred (51). Hale et al. found no difference in volume
administered in patients with cerebral edema and those
without (52).
Marcin et al. retrospectively evaluated 61 patients un-
der the age of 18 years with DKA and cerebral edema
(53). Investigators utilized an ordinal logistic regression
analysis, finding that 17 patients died or survived in a
vegetative state, 8 were mildly to moderately disabled,
and 36 experienced no sequelae. Factors with poor
outcome in CE included neurologic depression at the
time of diagnosis of cerebral edema, high BUN, and intu-
bation with hyperventilation to PCO2 < 22 mm Hg (53).
Hoffman et al. evaluated serial head CTs in 9 patients
with DKA, finding no change in pretreatment CT and
CT 6–8 h after treatment (21). One study in 1997 found
that 6 of 7 patients undergoing CT demonstrated evidence
of edema before treatment was started (59). No evidence
of edema was found on head CT in patients treated for
DKA (59). A summary of the literature, with discussion
of the design and results, is shown in Table 2. This Table
includes the majority of studies evaluating DKA; howev-
er, case series with no cases of CE and observational
studies without comparison groups were excluded from
this Table.
Bottom Line: Intravenous fluid infusion, either size of
bolus or infusion rate, is likely not associated with devel-
opment of cerebral edema.
What is Associated with Development of Cerebral
Edema?
Several factors have been consistently associated with
development of CE in pediatric DKA, unlike fluid
Table 3. Potential Risk Factors of CE in Pediatric DKA
DKA-Induced Cerebral Edema Risk Factors
New onset
Younger age (<5 years)
Severe acidemia (pH < 7.1)
Severe hypocapnia (PCO2 < 20 mm Hg)
Insulin administration during first hour of resuscitation
High blood urea nitrogen
Lower initial bicarbonate
Treatment with bicarbonate
Slow increase in serum sodium concentration
during treatment
CE = cerebral edema; DKA = diabetic ketoacidosis.
5
infusion size or rate, which are discussed in Table 3
(15,17,18,24,31,32,37,53,60,61). Similar to the prior
studies discussed, these studies are mostly retrospective
and uncontrolled; however, the studies by Glaser et al.,
Lawrence et al., and Edge et al. are more rigorously
designed and include groups for comparison
(14,15,18,24,25,31,32,39,51,53,55). Bello and Sotos
found declining sodium to be an ominous sign (54). How-
ever, fluid rate was not associated (54). Glaser et al. sug-
gest that treatment with bicarbonate, higher serum BUN,
and lower partial pressures of arterial CO2 are related to
development of edema (15). These imply that patients
who were more toxic in the initial stages of DKA demon-
strate higher rates of CE. The same studies that suggest
that greater infusion rates are associated with CE also
include younger age, higher BUN, and new diagnosis of
diabetes, though as discussed, these studies did not utilize
controls (14,18,24,25,39,51,53,55). A large number of
patient factors may be associated with CE,
demonstrated in Table 3. Many of these factors are asso-
ciated with greater dehydration. Another component is
patient age, as the brains of younger patients may be
more susceptible to metabolic and vascular changes in
DKA (1,2,15,16). Ultimately, patients who are sicker
with greater dehydration upon initial presentation are at
higher risk for CE.
Part of the assessment for DKA severity revolves
around dehydration, as patients with 5–7% are classified
as moderate DKA, whereas those with 10–14% may be
classified as severe DKA (1,2,12,13,16). Classically,
examination findings such as reduced skin elasticity,
dry mucosal membranes, tachycardia, and hyperpnea
were thought to be associated with 5% dehydration,
whereas 10% dehydration was assumed if capillary
refill > 10 s was found with sunken eyes (16,62,63).
However, the degree of dehydration is often overesti-
mated (64–68). Physical examination findings such as
reduced skin turgor, capillary refill, dry membranes,
and sunken eyes are not reliable predictors of hydration
status in patients with DKA, as they may not be due to
Studies
Rosenbloom (24) & Duck, Wyatt (39)
Rosenbloom (24) & Duck, Wyatt (39)
Edge et al. (32), Lawrence et al. (31), Durr et al. (37), Mahoney et al. (18)
Glaser et al. (15), Mahoney et al. (18)
Edge et al. (32)
Glaser et al. (15), Lawrence et al. (31), Marcin et al. (53)
Lawrence et al. (31)
Glaser et al. (15), Chua et al. (60), Bureau et al. (61)
Glaser et al. (15), Harris et al. (17), Duck & Wyatt (39), Bello & Sotos (54)
6 B. Long and A. Koyfman

only fluid loss (64–67). Confounding factors include

Replace deficit + maintenance

tachypnea, resulting in dry mucous membranes, and

fluids over 48 h evenly

vasoconstriction, leading to cool extremities. Acidosis
can cause tachypnea, vasoconstriction, and dry mucous

10 mL/kg 0.9% NS

5% of body weight
membranes from Kussmaul respirations (65,68). The

B2
hyperosmolarity can drastically affect intravascular

0.9% saline
volume, as diuresis can result in hypovolemia, whereas
hyperosmolarity increases intravascular volume through

None
osmosis (16,64–68).
Multiple prospective studies demonstrate that dehy-

Replace deficit + maintenance

dration cannot be adequately predicted clinically
(16,64,65,67,68). These studies are based on the percent

fluids over 48 h evenly

loss of body weight as a surrogate for dehydration,
finding a median degree of dehydration of 5–8%

10 mL/kg 0.9% NS

5% of body weight
(64,65,67,68). Close to 70% of patients were clinically

B1
assessed incorrectly, either overestimated or

0.45% saline
underestimated. In one study, 60% of patients were in
severe DKA based on laboratory and clinical criteria,

None
though the median dehydration was 5.4%. The
measured degree of dehydration was not significantly

deficit + maintenance fluids over 12 h,

altered between different severity groups (64,65,67,68).

then remaining deficit + maintenance

What Should the Emergency Physician Do for Treatment?

fluids over following 24 h

Fluids are vital in the initial stages of resuscitation for
rehydration, while also improving blood glucose levels. A2

10% of body weight

Inadequate resuscitation may worsen cerebral hypoperfu-

Replace half of fluid

10 mL/kg 0.9% NS
10 mL/kg 0.9% NS
sion. A number of regimens are advocated for resuscita-
tion, including 10-mL/kg bolus or 20-mL/kg bolus over

0.9% saline
1–2 h, as well as calculation of the fluid requirement
over the following 48 h (15,16,55–57,69–72).
However, as discussed, the optimal volume and resus-
citation rate are controversial, and dehydration is difficult
deficit + maintenance fluids over 12 h,

to assess. To date, few randomized trials have evaluated

then remaining deficit + maintenance

infusion rate or i.v. bolus amount, though studies have

PECARN = Pediatric Emergency Care Applied Research Network.

evaluated bag systems for treatment (69,70). Felner and

White evaluated different protocols for rehydration,
fluids over following 24 h

both utilizing an initial bolus (55). No difference in cere-

bral edema was found based on the initial bolus. After
A1

10% of body weight

bolus, one group received 1.5 times maintenance fluids

Replace half of fluid
10 mL/kg 0.9% NS
10 mL/kg 0.9% NS

plus fluids based on patient weight with 0.5 normal saline,

whereas the other group received 2.5 times maintenance
0.45% saline

with 0.75 normal saline. This second group demonstrated

Table 4. PECARN Study Groups (72)

faster correction of acidosis and was more cost effective

(55). A two-bag system consists of two bags of fluids
with similar electrolyte contents, but with different
Standard initial fluid bolus

glucose amounts that are provided into the same i.v.

line. Dehydration and patient needs affect the rate of
Assumed fluid deficit
Deficit replacement

fluids, whereas serum glucose and rate of glucose decline

Replacement fluid
Protocol

affect dextrose infusion (0% vs. 10%). At first, no glucose

Additional fluid

is provided, but as hyperglycemia improves, dextrose is

titrated to allow for control of glucose decrease. Another
study released in 2013 utilized a three-bag system, with
two bags of rehydration fluids (one containing glucose),
A Clinical Review Evaluating Myths Surrounding Cerebral Edema in Pediatric Diabetic Ketoacidosis
and the third bag with insulin (71). Fluids were
administered at 2–2.5 times maintenance. This system
demonstrates improved flexibility and timeliness when
compared with a single-bag system. The separate bag
with glucose provides the ability to quickly respond to
changing serum glucose levels (71).
A recent study evaluated the volume infusion on meta-
bolic normalization, length of stay, and adverse out-
comes, including patients 0 to 18 years of age with type
1 diabetes mellitus and DKA (57). Investigators random-
ized patients to low volume (10-mL/kg bolus with 1.25
times maintenance rate) vs. high volume (20-mL/kg
bolus with 1.5 maintenance rate). The study suggests
that higher volume infusion rates improved normalization
of pH, but made no difference in length of stay or time to
discharge. No adverse events, including signs/symptoms
of cerebral edema, occurred in either group (57).
The Pediatric Emergency Care Applied Research
Network (PECARN) network is currently completing a
trial evaluating fluid infusion in pediatric DKA, including
patients under the age of 18 years (Table 4) (72). This trial
is a prospective study utilizing four different fluid proto-
cols, shown in Table 2. Of note, all patients receive an
initial 10-mL/kg bolus of 0.9% saline. The primary study
outcome is abnormal GCS (< 14) during treatment. Those
with initial GCS < 14 are excluded from evaluating the
primary outcome, but not secondary outcomes including
clinically overt edema, forward and backward digit span
recall tests, and memory tests 3 months after recovery
from DKA (72). Results of this trial are pending.
At this time, 10 mL/kg is safe in clinically dehydrated
patients with poor capillary refill, dry mucosal mem-
branes, or sunken eyes, and this is recommended by the
International Diabetes Foundation (73). Hemodynamic
status must be carefully assessed, and any signs of hemo-
dynamic instability warrant i.v. fluid bolus. If the patient
is severely dehydrated and hemodynamic status does not
improve after the first i.v. bolus, a second bolus may be
provided. Maintenance fluid with further replacement
can then be calculated for replacement over 48 h (73).
If the patient is not severely dehydrated, maintenance
fluids can be provided with correction over 48 h (73).
Dehydration and hypoperfusion in the setting of DKA
are likely the factors most strongly associated with CE.
CONCLUSIONS
Pediatric DKA and CE are diseases that can cause signif-
icant morbidity and mortality. CE is a clinical diagnosis
and is rare in its severe form. A variety of myths concern-
ing CE in pediatric DKA are present. CE may occur sub-
clinically in many patients in DKA. A variety of
mechanisms likely account for CE, including vasogenic
and cytotoxic causes. Patients with DKA are often dehy-
7
drated, but dehydration can be difficult to assess clini-
cally. The literature evaluating fluid infusion in DKA
suffers from low sample sizes, lack of comparator groups,
and retrospective nature. However, several studies with
comparator groups suggest that i.v. fluids are not associ-
ated with CE, though severity of DKA may be associated,
including the degree of acidosis and dehydration. A fluid
bolus of 10–20 mL/kg is likely safe and may be repeated
if needed.
REFERENCES
1. Olivieri L, Chasm R. Diabetic ketoacidosis in the pediatric emer-
gency department. Emerg Med Clin North Am 2013;31:755–73.
2. Rosenbloom AL. The management of diabetic ketoacidosis in chil-
dren. Diabetes Ther 2010;1:103–20.
3. Klingensmith GJ, Tamborlane WV, Wood J, et al. Diabetic ketoaci-
dosis at diabetes onset: still an all too common threat in youth. J Pe-
diatr 2013;162:330–3341.
4. Rewers A, Klingensmith G, Davis C, et al. Presence of diabetic ke-
toacidosis at diagnosis of diabetes mellitus in youth: the Search for
Diabetes in Youth Study. Pediatrics 2008;121:e1258–66.
5. Levy-Marchal C, Patterson CC, Green A, et al. Geographical vari-
ation of presentation at diagnosis of type 1 diabetes in children: the
EURODIAB study. Diabetologia 2001;44:B75–80.
6. Usher-Smith JA, Thompson M, Ercole A, et al. Variation between
countries in the frequency of diabetic ketoacidosis at first presenta-
tion of type 1 diabetes in children: a systematic review. Diabetologia
2012;55:2878–94.
7. Usher-Smith JA, Thompson MJ, Sharp SJ, et al. Factors associated
with the presence of diabetic ketoacidosis at diagnosis of diabetes in
children and young adults: a systematic review. BMJ 2011;343:
d4092.
8. Rewers A, Chase HP, Mackenzie T, et al. Predictors of acute com-
plications in children with type 1 diabetes. JAMA 2002;287:2511–
8.
9. Type 2 diabetes in children and adolescents. American Diabetes As-
sociation. Diabetes Care 2000;23:381–9.
10. Kitabchi AE, Umpierrez GE, Murphy MB, et al., American Dia-
betes Association. Hyperglycemic crises in diabetes. Diabetes
Care 2004;27(suppl 1):S94–102.
11. Dunger DB, Sperling MA, Acerini CL, et al. European Society for
Pediatric Endocrinology / Lawson Wilkins Pediatric Endocrine So-
ciety consensus statement on diabetic ketoacidosis in children and
adolescents. Pediatrics 2004;113:e133–40.
12. Wolfsdorf J, Glaser N, Sperling MA. Diabetic ketoacidosis in in-
fants, children, and adolescents. A consensus statement from the
American Diabetes Association. Diabetes Care 2006;29:1150–9.
13. Wolfsdorf J, Craig ME, Daneman D, et al. ISPAD Clinical Practice
Consensus Guidelines 2009 Compendium. Diabetic ketoacidosis in
children and adolescents with diabetes. Pediatr Diabetes 2009;
10(suppl 12):118–33.
14. Edge J, Hawkins M, Winter D, Dunger D. The risk and outcome of
cerebral oedema developing during diabetic ketoacidosis. Arch Dis
Child 2001;85:16–22.
15. Glaser N, Barnett P, McCaslin I, et al. Risk factors for cerebral
edema in children with diabetic ketoacidosis. N Engl J Med 2001;
344:264–9.
16. Watts W, Edge JA. How can cerebral edema during treatment of dia-
betic ketoacidosis be avoided? Pediatr Diabetes 2014;15:271–6.
17. Harris GD, Fiordalisi I, Harris WL, et al. Minimizing the risk of
brain herniation during treatment of diabetic ketoacidemia: a retro-
spective and prospective study. J Pediatr 1990;117:22–31.
18. Mahoney CP, Vleck BW, Del Aguila M. Risk factors for developing
brain herniation during diabetic ketoacidosis. Pediatr Neurol 1999;
21:721–7.
8 B. Long and A. Koyfman
19. Glaser N, Gorges S, Marcin J, et al. Mechanism of cerebral edema in
children with diabetic ketoacidosis. J Pediatr 2004;145:164–71.
20. Glaser N, Wooton-Gorges S, Buonocore M, et al. Frequency of sub-
clinical cerebral edema in children with diabetic ketoacidosis. Pe-
diatr Diabetes 2006;7:75–80.
21. Hoffman W, Steinhart C, El Gammal T, Steele S, Cuadrado A,
Morse P. Cranial CT in children and adolescents with diabetic ke-
toacidosis. AJNR Am J Neuroradiol 1988;9:733–9.
22. Krane E, Rockoff M, Wallman J, Wolfsdorf J. Subclinical brain
swelling in children during treatment of diabetic ketoacidosis. N
Engl J Med 1985;312:1147–51.
23. Edge JA. Cerebral oedema during treatment of diabetic ketoacido-
sis: are we any nearer finding a cause? Diabetes Metab Res Rev
2000;16:316–24.
24. Rosenbloom AL. Intracerebral crises during treatment of diabetic
ketoacidosis. Diabetes Care 1990;13:22–33.
25. Rosenbloom AL, Riley WJ, Weber FT, et al. Cerebral edema
complicating diabetic ketoacidosis in childhood. J Pediatr 1980;
96:357–61.
26. Muir AB, Quisling RG, Yang MCK, Rosenbloom AL. Cerebral
edema in childhood diabetic ketoacidosis. Natural history, radio-
graphic findings, and early identification. Diabetes Care 2004;27:
1541–6.
27. Brown TB. Cerebral oedema in childhood diabetic ketoacidosis: is
treatment a factor? Emerg Med J 2004;21:141–4.
28. Hom J, Sinert R. Is fluid therapy associated with cerebral edema in
children with diabetic ketoacidosis? Ann Emerg Med 2008;52:69–751.
29. Dillon ES, Riggs HE, Dyer WW. Cerebral lesions in uncomplicated
fatal diabetic ketoacidosis. Am J Med Sci 1936;192:360–5.
30. Lam T, Anderson S, Glaser N, O’Donnell M. Bumetanide reduces
cerebral edema formation in rats with diabetic ketoacidosis. Dia-
betes 2005;54:510–6.
31. Lawrence SE, Cummings EA, Gaboury I, Daneman D. Population-
based study of incidence and risk factors for cerebral edema in pe-
diatric diabetic ketoacidosis. J Pediatr 2005;146:688–92.
32. Edge JA, Jakes RW, Roy Y, et al. The UK case-control study of ce-
rebral oedema complicating diabetic ketoacidosis in children. Dia-
betologia 2006;49:2002–9.
33. Glaser N, Marcin J, Wooton-Gorges S, et al. Correlation of clinical
and biochemical findings with DKA-related cerebral edema in chil-
dren using magnetic resonance diffusion weighted imaging. J Pe-
diatr 2008;153:541–6.
34. Figueroa RE, Hoffman WH, Momin Z, Pancholy A,
Passmore GG, Allison J. Study of subclinical cerebral edema in
diabetic ketoacidosis by magnetic resonance imaging T2 relaxom-
etry and apparent diffusion coefficient maps. Endocr Res 2005;31:
345–55.
35. Vavilala MS, Richards TL, Roberts JS, et al. Change in blood-brain
barrier permeability during pediatric diabetic ketoacidosis treat-
ment. Pediatr Crit Care Med 2010;11:332–8.
36. Roberts JS, Vavilala MS, Schenkman KA, Shaw D, Martin LD,
Lam AM. Cerebral hyperemia and impaired cerebral autoregulation
associated with diabetic ketoacidosis in critically ill children. Crit
Care Med 2006;34:2217–23.
37. Durr JA, Hoffman WH, Sklar AH, el Gammal T, Steinhart CM. Cor-
relates of brain edema in uncontrolled IDDM. Diabetes 1992;41:
627–32.
38. Silver SM, Clark EC, Schroeder BM, Sterns RH. Pathogenesis of
cerebral edema after treatment of diabetic ketoacidosis. Kidney
Int 1997;51:1237–44.
39. Duck S, Wyatt D. Factors associated with brain herniation in the
treatment of diabetic ketoacidosis. J Pediatr 1988;113:10–4.
40. Harris G, Fiordalisi I, Finberg L. Safe management of diabetic ke-
toacidemia. J Pediatr 1988;113:65–7.
41. Arieff A, Kleeman C. Cerebral edema in diabetic comas. II. Effects
of hyperosmolality, hyperglycemia and insulin in diabetic rabbits. J
Clin Endocrinol Metab 1974;38:1057–67.
42. Prockop L. Hyperglycemia, polyol accumulation, and increased
intracranial pressure. Arch Neurol 1971;25:126–40.
43. Fiordalisi I, Novotny WE, Holbert D, Finberg L, Harris GD, Critical
Care Management Group. 18 yr prospective study of paediatric dia-
betic ketoacidosis: an approach to minimizing the risk of brain her-
niation during treatment. Pediatr Diabetes 2007;8:142–9.
44. Yuen N, Anderson S, Glaser N, O’Donnell M. Cerebral blood flow
and cerebral edema in rats with diabetic ketoacidosis. Diabetes
2008;57:2588–94.
45. Glaser N, Yuen N, Anderson S, Tancredi D, O’Donnell M. Cerebral
metabolic alterations in rats with diabetic ketoacidosis: effects of
treatment with insulin and intravenous fluids and effects of bumeta-
nide. Diabetes 2010;59:702–9.
46. Wootton-Gorges SL, Buonocore MH, Kuppermann N, et al. Cere-
bral proton magnetic resonance spectroscopy in children with dia-
betic ketoacidosis. AJNR Am J Neuroradiol 2007;28:895–9.
47. Wootton-Gorges SL, Buonocore MH, Kuppermann N, et al. Detec-
tion of cerebral {beta}-hydroxybutyrate, acetoacetate, and lactate
on proton MR spectroscopy in children with diabetic ketoacidosis.
AJNR Am J Neuroradiol 2005;26:1286–91.
48. Hoffman WH, Andjelkovic AV, Zhang W, Passmore GG, Sima AA.
Insulin and IGF-1 receptors, nitrotyrosin and cerebral neuronal def-
icits in two young patients with diabetic ketoacidosis and fatal brain
edema. Brain Res 2010;1343:168–77.
49. Hoffman WH, Artlett CM, Zhang W, et al. Receptor for advanced
glycation end products and neuronal deficit in the fatal brain edema
of diabetic ketoacidosis. Brain Res 2008;1238:154–62.
50. Hoffman WH, Stamatovic SM, Andjelkovic AV. Inflammatory me-
diators and blood brain barrier disruption in fatal brain edema of
diabetic ketoacidosis. Brain Res 2009;1254:138–48.
51. Mel JM, Werther GA. Incidence and outcome of diabetic cerebral
oedema in childhood: are there predictors? J Pediatr Child Health
1995;31:17–20.
52. Hale PM, Rezvani I, Braunstein AW, et al. Factors predicting cere-
bral edema in young children with diabetic ketoacidosis and new
onset type I diabetes. Acta Paediatr 1997;86:626–31.
53. Marcin JP, Glaser N, Barnett P, et al. Factors associated with adverse
outcomes in children with diabetic ketoacidosis-related cerebral
edema. J Pediatr 2002;141:793–7.
54. Bello FA, Sotos JF. Cerebral oedema in diabetic ketoacidosis in
children. Lancet 1990;336:64.
55. Felner EI, White PC. Improving management of diabetic ketoacido-
sis in children. Pediatrics 2001;108:735–40.
56. Hsia DS, Tarai SG, Alimi A, et al. Fluid management in pediatric
patients with DKA and rates of suspected clinical cerebral edema.
Pediatr Diabetes 2015;16:338–44.
57. Bakes K, Haukoos JS, Deakyne SJ, et al. Effect of volume of fluid
resuscitation on metabolic normalization in children presenting in
diabetic ketoacidosis: a randomized controlled trial. J Emerg Med
2016;50:551–9.
58. Clements RS Jr, Blumenthal SA, Morrison AD, Winegrad AI.
Increased cerebrospinal- fluid pressure during treatment of diabetic
ketoacidosis. Lancet 1971;2:671–5.
59. Smedman L, Escobar R, Hesser U, et al. Subclinical cerebral
oedema does not occur regularly during treatment for diabetic ke-
toacidosis. Acta Pediatr 1997;86:1172–6.
60. Chua HR, Schneider A, Bellomo R. Bicarbonate in diabetic ketoa-
cidosis – a systematic review. Ann Intensive Care 2011;1:23.
61. Bureau MA, Bégin R, Berthiaume Y, Shapcott D, Khoury K,
Gagnon N. Cerebral hypoxia from bicarbonate infusion in diabetic
acidosis. J Pediatr 1980;96:968–73.
62. Rosenbloom AL. Diabetic ketoacidosis: treatment guidelines. Clin
Pediatr (Phila) 1996;35:261–6.
63. Finberg L. Appropriate therapy can prevent cerebral swelling
in diabetic ketoacidosis. J Clin Endocrinol Metab 2000;85:
507–8.
64. Koves I, Neutze J, Donath S, et al. The accuracy of clinical assess-
ment of dehydration during diabetic ketoacidosis in childhood. Dia-
betes Care 2004;27:2485–7.
65. Ugale J, Mata A, Meert K, Samaik A. Measured degree of dehydra-
tion in children and adolescents with type 1 diabetic ketoacidosis.
Pediatr Crit Care Med 2012;13:e103–7.
66. Smith L, Rotta A. Accuracy of clinical estimates of dehydration in
pediatric patients with diabetic ketoacidosis. Pediatr Emerg Care
2002;18:395–6.
A Clinical Review Evaluating Myths Surrounding Cerebral Edema in Pediatric Diabetic Ketoacidosis
67. Sottosanti M, Morrison G, Singh R, et al. Dehydration in children
with diabetic ketoacidosis: a prospective study. Arch Dis Child
2012;97:96–100.
68. Fagan MJ, Avner J, Khine H. Initial fluid resuscitation for patients
with diabetic ketoacidosis: how dry are they? Clin Pediatr 2008;
47:851–5.
69. Grimberg A, Cerri RW, Satin-Smith M, et al. The ‘‘two bag system’’
for variable intravenous dextrose and fluid management: benefits in
diabetic ketoacidosis management. J Pediatr 1999;134:376–8.
70. Poirier MP, Greer D, Satin-Smith M. A prospective study of the
‘‘two-bag system’’ in diabetic ketoacidosis management. Clin Pe-
diatr (Phila) 2004;43:809–13.
9
71. White PC, Dickson BA. Low morbidity and mortality in children
with diabetic ketoacidosis treated with isotonic fluids. J Pediatr
2013;163:761–6.
72. Glaser NS, Ghetti S, Casper TC, Dean JM, Kuppermann N, Pediat-
ric Emergency Care Applied Research Network (PECARN) DKA
FLUID Study Group. Pediatric diabetic ketoacidosis, fluid therapy,
and cerebral injury: the design of a factorial randomized controlled
trial. Pediatr Diabetes 2013;14:435–46.
73. International Diabetes Federation. ISPAD guideline for diabetes in
childhood and adolescence. Available at: http://www.idf.org/sites/
default/files/Diabetes-in-Childhood-and-Adolescence-Guidelines.
pdf. Accessed February 20, 2017.
10 B. Long and A. Koyfman

ARTICLE SUMMARY
1. Why is this topic important?
Cerebral edema is a severe complication of pediatric
diabetic ketoacidosis (DKA), though it is rare. Many
studies have evaluated risk factors associated with the
development of cerebral edema.
2. What does this review attempt to show?
This review evaluates the literature concerning cerebral
edema (CE) in pediatric DKA including mechanism, pre-
sentation of edema, clinical assessment of dehydration,
and association with intravenous (i.v.) fluids.
3. What are the key findings?
CE occurs in <1% of DKA cases. The literature evalu-
ating CE consists of mainly retrospective observational
studies. A multitude of mechanisms account for the devel-
opment of CE, including vasogenic and cytotoxic edema.
Signs commonly used to evaluate for dehydration are not
reliable, but hemodynamic status should be carefully as-
sessed. Subclinical CE is likely common in DKA. Much
of the literature suggesting an association with i.v. fluids
and CE are retrospective, with no comparator groups.
Several well-constructed case-control studies suggest
i.v. fluid infusion may not be associated with CE. A bolus
of 10 mL/kg of fluids is safe in patients with severe dehy-
dration or hemodynamic concerns.
4. How is patient care impacted?
CE in pediatric DKA is rare, and hemodynamic status
should be carefully assessed. If hypotensive or hemody-
namically stable, i.v. fluid bolus should be given and the
patient reassessed. If no evidence of instability is present,
fluids may still be provided.