Laura M. Kinlin and Stephen B. Freedman 2012;129;e1211 DOI: 10.1542/peds.2011-2985

Evaluation of a Clinical Dehydration Scale in Children Requiring Intravenous
Rehydration
Laura M. Kinlin and Stephen B. Freedman
Pediatrics 2012;129;e1211; originally published online April 23, 2012;
DOI: 10.1542/peds.2011-2985
The online version of this article, along with updated information and services, is
located on the World Wide Web at:
http://pediatrics.aappublications.org/content/129/5/e1211.full.html
PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly
publication, it has been published continuously since 1948. PEDIATRICS is owned,
published, and trademarked by the American Academy of Pediatrics, 141 Northwest Point
Boulevard, Elk Grove Village, Illinois, 60007. Copyright © 2012 by the American Academy
of Pediatrics. All rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.
Downloaded from pediatrics.aappublications.org by guest on August 22, 2014
ARTICLE
Evaluation of a Clinical Dehydration Scale in Children
Requiring Intravenous Rehydration
AUTHORS: Laura M. Kinlin, BSc, MPH,a and Stephen B.
Freedman, MDCM, MSc, FRCPC, FAAPb
aFaculty of Medicine, Dalhousie University, Halifax, Nova Scotia,
Canada; and bDivisions of Pediatric Emergency Medicine and
Gastroenterology, Hepatology, and Nutrition, and Child Health
Evaluative Sciences, The Hospital for Sick Children, Departments
of Pediatrics, and Health Policy, Management and Evaluation,
University of Toronto, Faculty of Medicine, Toronto, Ontario,
Canada
KEY WORDS
dehydration, gastroenteritis, clinical score, severity of illness,
emergencies
ABBREVIATIONS
AUC—area under the curve
CDS—clinical dehydration scale
CI—confidence interval
ED—emergency department
IQR—interquartile range
ROC—receiver operating characteristic
Dr Freedman is the guarantor of this article and declares that
he participated in the design of this study and revisions of the
article; Dr Freedman also takes full responsibility for the work
and conduct of the study, the integrity of the data, and the
decision to publish. Ms Kinlin made substantial contribution to
analysis and interpretation of data, drafting of the article for
publication, and she has seen and approved the version to be
published.
This study is a secondary analysis of data collected as part of
a clinical trial, which was registered at www.clinicaltrials.gov
(identifier NCT00392145). The main content of the clinical trial
has been published by the British Medical Journal. Two other
manuscripts arising from the original clinical trial data set are
currently under review by other journals. The authors of all
manuscripts attest to the fact that all manuscripts contain
unique information and minimal if any overlapping content.
www.pediatrics.org/cgi/doi/10.1542/peds.2011-2985
doi:10.1542/peds.2011-2985
Accepted for publication Jan 12, 2012
Address correspondence to Stephen B. Freedman, MDCM, MSc,
FRCPC, FAAP, Division of Pediatric Emergency Medicine, The
Hospital for Sick Children, 555 University Ave, Toronto, ON, Canada
M5G 1X8. E-mail: [email protected]
(Continued on last page)
WHAT’S KNOWN ON THIS SUBJECT: Evaluating dehydration
severity is a challenging task. Clinical dehydration scores that
combine multiple clinical findings are promising. One clinical
dehydration scale score has been developed and subsequently
evaluated; however, few participants in the derivation and
validation studies were significantly dehydrated.
WHAT THIS STUDY ADDS: In children requiring intravenous
rehydration, the dehydration scale displayed moderate reliability
and weak associations with objective measures. Thus, although
the scale can assist in assessing dehydration, it should not be
used in isolation to dictate interventions (eg, intravenous
rehydration, hospitalization).
abstract
OBJECTIVES: To evaluate the reliability and validity of a previously derived clinical dehydration scale (CDS) in a cohort of children with gastroenteritis and evidence of dehydration.
METHODS: Participants were 226 children older than 3 months who
presented to a tertiary care emergency department and required intravenous rehydration. Reliability was assessed at treatment initiation,
by comparing the scores assigned independently by a trained research
nurse and a physician. Validity was assessed by using parameters reflective of disease severity: weight gain, baseline laboratory results,
willingness of the physician to discharge the patient, hospitalization,
and length of stay.
RESULTS: Interobserver reliability was moderate, with a weighted k of
0.52 (95% confidence interval [CI] 0.41, 0.63). There was no correlation
between CDS score and percent weight gain, a proxy measure of fluid
deficit (Spearman correlation coefficient = 20.03; 95% CI 20.18, 0.12).
There were, however, modest and statistically significant correlations
between CDS score and several other parameters, including serum
bicarbonate (Pearson correlation coefficient = 20.35; 95% CI 20.46,
20.22) and length of stay (Pearson correlation coefficient = 0.24; 95%
CI 0.11, 0.36). The scale’s discriminative ability was assessed for the
outcome of hospitalization, yielding an area under the receiver operating characteristic curve of 0.65 (95% CI 0.57, 0.73).
CONCLUSIONS: In children administered intravenous rehydration, the
CDS was characterized by moderate interobserver reliability and weak
associations with objective measures of disease severity. These data do
not support its use as a tool to dictate the need for intravenous rehydration or to predict clinical course. Pediatrics 2012;129:e1211–e1219
PEDIATRICS Volume 129, Number 5, May 2012
Downloaded from pediatrics.aappublications.org by guest on August 22, 2014
e1211
In the United States, gastroenteritis accounts for .1.5 million outpatient visits
and 200 000 hospitalizations annually.1
Treatment guidelines use goal-directed
therapy, with recommendations based on
assessment of dehydration severity1–3;
however, this can be a challenge as
there is “a lack of compelling evidence to
support efforts to accurately distinguish
varying degrees of dehydration on the
basis of symptoms and signs.”2 Percent
decrease in body weight, recognized as
an objective “gold standard,”2 often
cannot be calculated because baseline
hydrated weights are rarely available at
acute-care visits.4 Consequently, traditional teaching has encouraged grouping children, based on several variables,
into 3 categories: mild, moderate, and
severe.5,6 Recent guidelines have acknowledged the limitations of this approach, and divide patients into minimal
(,3%–5%), some (ie, mild-moderate,
5%–10%), and severe dehydration
(.10%)1,3,7; however, physicians classify
children with “some” dehydration correctly only 33% of the time.8 Inaccurate
assessments have important ramifications: underestimation may result
in delayed therapy; overestimation may
lead to unnecessary interventions, complications and costs.9
Recently, a clinical dehydration scale
(CDS) has been developed10 and undergone validation11,12; however, these
reports have included few subjects
with significant dehydration. Moreover,
interobserver reliability has not been
evaluated, and the studies have been
underpowered to evaluate the correlation between CDS score and biochemical parameters. Before widespread
adoption of the CDS, further evaluation
is necessary to establish that it produces consistent results and correlates
with meaningful measures of disease
severity.13
We sought to evaluate the reliability and
validity of Friedman et al’s CDS10 in
a cohort of children who presented to
e1212
a pediatric emergency department (ED)
with gastroenteritis and dehydration.
Our primary objective was to evaluate
interobserver reliability by using independent, simultaneous, blinded assessments. The secondary objective was to
explore validity by examining the score’s
association with other clinical data.
METHODS
Setting and Participants
Between December 2006 and April 2010,
data were prospectively collected on
226 children $3 months of age who
presented to The Hospital for Sick
Children’s ED (Toronto, Ontario) with
gastroenteritis and dehydration requiring intravenous rehydration. The
study was approved by the hospital’s
research ethics board. Written informed
consent was obtained from the guardians of all participants.
Eligibility criteria were a diagnosis of
acute gastroenteritis, evidence of dehydration, and a decision to administer
intravenous rehydration therapy. The
diagnosis of gastroenteritis was at the
discretion of the supervising physician;
“acute” implied ,6 days of symptoms.
Dehydration was defined based on the
presence of one of the following features as assigned by a trained research nurse: (1) CDS $3 (Table 1)10,11;
(2) capillary refill time $2 seconds15;
(3) abnormal skin turgor, with prolonged retraction time and “tenting”15;
or (4) abnormal respiratory pattern
(.50 breaths per minute for children
aged ,12 months; .40 breaths per
minute for children aged $12 months).15
All potentially eligible children underwent a trial of oral rehydration therapy conducted with administration of
5 mL of a flavored oral rehydration solution through a syringe every 5 minutes,
with the rate increased based on tolerance and the child’s weight.1 For children
with persistent vomiting, ondansetron
was administered orally. Subsequently,
based on the success of oral rehydration,
the decision regarding the need for
intravenous rehydration was made by
the responsible physician, unaware of
the CDS score assigned by the nurse.
Exclusion criteria were weight ,5 kg;
presence of a significant underlying
disease that would affect hydration
assessment (eg, renal insufficiency,
diabetes mellitus); suspicion of previously undiagnosed cardiac or renal
disease; history of abdominal surgery
or presence of an acute surgical abdomen; history of significant head,
chest, or abdominal trauma within 7
days; evidence of hypotension, hypoglycemia (,2.8 mmol/L), or hyperglycemia (.11.0 mmol/L); and previous
study enrollment.
Procedures
Children were assessed for eligibility
once the treating physician had determined that intravenous rehydration
was required. Eligible children then
had CDS scores assigned by a research
nurse and the responsible physician,
who was blinded to the score recorded
by the nurse and unaware that interobserver scores would be calculated.
Thirty-one physicians, who received no
formal training in using the CDS but
were provided with standardized definitions (Table 1), assigned scores. All
children had the following investigations performed at the time of intravenous insertion (Advia IMS Integrated
Modular System [Bayer, Elkhart, IN], accuracy = 60.3 mmol/L): sodium, potassium, chloride, pH, bicarbonate, carbon
dioxide, glucose, blood urea nitrogen,
and creatinine. CDS scores were subsequently documented by the research
nurse every 30 minutes for 4 hours.
Patient Assessment and Data
Collection
Data elements collected included basic
demographics, historical variables
(eg, number of vomiting episodes in
the preceding 24 hours), and clinical
KINLIN and FREEDMAN
Downloaded from pediatrics.aappublications.org by guest on August 22, 2014
ARTICLE
TABLE 1 Friedman et al’s10 CDSa
Characteristic
Score of 0
Score of 1
Score of 2
General appearance
Normalb
Eyes
Mucous membranesc
Tears
Normal
Moist
Tears
Thirsty, restless, or lethargic
but irritable when touched
Slightly sunken
Sticky
Decreased tears
Drowsy, limp, cold, sweaty;
comatose or not
Very sunken
Dry
Absent tears
Scores for the individual items are summed.
a Higher scores indicate more severe dehydration. Scores range from 0 to 8. A score of 0 correlates with ,3% dehydration
(positive likelihood ratio 2.2; 95% CI 0.9–5.3), scores of 1–4 correlate with some (3%–6%) dehydration (positive likelihood
ratio 1.3, 95% CI 0.9, 1.7), and 5–8 correlates with moderate to severe ($6%) dehydration (positive likelihood ratio 5.2; 95% CI
2.1, 12.8).14
b “Normal” includes children who may be sleeping but are easily aroused to a normal level of consciousness. This
assessment takes into account the time of day and the child’s usual pattern as described by the child’s guardian.
c This is assessed on the buccal mucosa and tongue, and not the lips.
examination features (eg, heart rate).
The respiratory rate was measured
for 60 seconds by observing chest wall
movements with the child quiet and
comfortable.15 Capillary refill was assessed at the fingertip by using a
standardized technique.16 Skin turgor
was assessed by pinching a small skin
fold between the thumb and index finger on the lateral abdominal wall at the
level of the umbilicus.15
Weight gain was recorded at the time
of disposition determination or at discharge for admitted children. Percent
weight gain was calculated as the difference between discharge and triage
weights, divided by the triage weight.
Other outcome data recorded included
physician willingness to discharge the
child at time 2 hours measured by using
a Likert scale, the need for hospitalization, and total length of stay (ie, from
initiation of intravenous rehydration to
ED or hospital discharge). Data were
entered into a secure database and
double-checked for accuracy by the
senior investigator (S.B.F.).
real-world model. The secondary outcome was validity. Criterion validity,
which is based on a scale’s correlation
with an accepted “gold standard,” was
assessed by using weight change as
a proxy measure for percent dehydration. Construct validity, which reflects
the scale’s correlation with other measures predicted by theory,13 was evaluated by using parameters thought to be
reflective of dehydration severity: (1)
number of episodes of vomiting and
diarrhea before presentation to the
ED,17 (2) respiratory rate,15 (3) capillary
refill time,15 (4) serum bicarbonate,8,15
(5) serum pH,11 (6) physician assessment of readiness for discharge documented by using a Likert scale, and (7)
length of stay.11 Discriminative validity,
the ability to discriminate between
individuals with a given characteristic/
outcome and those without, was explored for the outcome of hospitalization.
The scale’s responsiveness to change
was evaluated through comparison
of scores before and after intravenous rehydration, a treatment of known
efficacy.
Outcomes
Our primary outcome was the interobserver agreement between 2 independent raters’ CDS scores (ie, reliability).
We sought to determine the agreement
of (1) individuals with extensive experience and training using the score
with (2) individuals with limited training; the latter group representing the
Sample Size
This study used data collected as part of
a clinical trial18 and, as such, is a secondary data analysis. It was estimated
post hoc that enrollment of 226 participants would provide 80% power to detect an interobserver correlation as low
as 0.35 when the k coefficient under the
null hypothesis is 0.20 with significance set at .05 (PASS 2008, NCSS, LLC,
Kaysville UT).
Analysis
Characteristics of the study sample
were described with frequency counts
and percentages for categorical variables; means and SDs or medians and
interquartile ranges (IQRs) were used
for continuous variables. Dichotomous
variables were compared by using the
x2 test or Fisher’s exact test. Continuous variables were compared by using
the t test or Wilcoxon rank-sum test, as
appropriate, depending on data normality. CDS score refers to the total
score assigned by the research nurse at
treatment initiation; this value was used
in all analyses unless otherwise specified. A P value ,.05 was considered
statistically significant.
Interobserver reliability was evaluated via calculation of the quadratic
weighted k statistic (k). Unlike Cohen’s
unweighted k, the weighted coefficient
considers partial agreement (ie, penalizes disagreement based on degree of
divergence) and is recommended when
calculating interobserver agreement
for ordinal scales.19 Quadratic weighting creates mathematical equivalence
with the intraclass correlation coefficient for agreement by using a 2-way
random-effects model single-measure
reliability.20 Results are interpreted
according to the criteria of Landis and
Koch,21 in which k values of 0.21 to 0.40,
0.41 to 0.60, 0.61 to 0.80, and .0.80
signify fair, moderate, substantial, and
almost perfect agreement, respectively.
In assessing validity, associations were
quantified by using the Pearson correlation coefficient (r) or Spearman rank
correlation coefficient (r), depending
on the distribution of the data. Significant outliers, defined as values greater
than the 75th percentile plus 1.5 3 the
IQR, or less than the 25th percentile
minus 1.5 3 the IQR,22 were identified in
PEDIATRICS Volume 129, Number 5, May 2012
Downloaded from pediatrics.aappublications.org by guest on August 22, 2014
e1213
the analysis of criterion validity; hence,
this analysis was repeated excluding
such values. We also evaluated the CDS’s
validity as a dichotomous variable, by
using cut points recommended in previous validation studies: scores of ,5
indicate none or some dehydration;
those $5 reflect moderate to severe
dehydration.11,14 In assessing responsiveness to change, CDS scores at time
0, 2, and 4 hours were compared by
using the Wilcoxon rank-sum test.
Discriminative validity was assessed as
the area under the receiver operating
characteristic (ROC) curve. Sensitivity,
specificity, positive likelihood ratio, positive predictive value, and negative predictive value were quantified by using
a prespecified CDS cutoff of $5.11
Because the CDS was developed in a
cohort of children ,36 months of age,
we analyzed the scale’s performance
separately for study participants ,36
and $36 months of age. The CDS’s
performance was also evaluated in the
subgroup of children with baseline
scores $5. All analyses were conducted by using Stata version 9.2 (Stata
Corp, College Station, TX).
RESULTS
A total of 226 children were enrolled
(Fig 1); 92 (41%) had moderate/severe
dehydration (CDS $5) and 134 (59%)
had none/some dehydration (CDS ,5)
as assessed by the research nurse.
Complete interobserver assessments
were performed for 208 participants
(92%). Participants who had interobserver assessments completed were
similar to those who did not. All partially
complete physician scores (n = 7) were
lacking a value for the tear category.
These patients were older than the remainder of the study sample (4.36 years
vs 2.18 years; P = .03).
FIGURE 1
Enrollment as it relates to the recording of baseline CDS scores.
nurses and physicians (Fig 2). Interobserver agreement, at the level of
the individual patient, between trained
nurses and the physicians was moderate (k = 0.52; 95% confidence interval
[CI] 0.41, 0.63; Fig 3). The k coefficient
values for the individual elements of
the CDS were as follows: eyes = 0.32
(95% CI 0.18, 0.46), mucous membranes =
0.38 (95% CI 0.26, 0.50), tears = 0.40
(95% CI 0.27, 0.51), and general appearance = 0.49 (95% CI 0.35, 0.60).
Reliability was similar for children
,36 and those $36 months of age
(k for the total score = 0.51 [95% CI
0.40, 0.65] and 0.53 [95% CI 0.27, 0.68],
CDS Reliability
The overall CDS distribution was different (P = .002) between research
e1214
FIGURE 2
CDS scores in the study population, as assigned by research nurses and physicians (n = 208). The
distribution of scores was different between the 2 groups of raters (P = .002).
KINLIN and FREEDMAN
Downloaded from pediatrics.aappublications.org by guest on August 22, 2014
ARTICLE
Compared with those with none/some
dehydration (CDS ,5), participants
with moderate/severe dehydration (CDS
$5) had a greater number of diarrhea
episodes before ED presentation, lower
serum bicarbonate and pH values at
baseline, increased capillary refill time,
and increased length of stay in hospital
(Table 3). The median physician dischargeability Likert score was higher
in the moderate/severe category, indicating a decreased willingness to discharge the children at 2 hours after
initiation of intravenous rehydration.
FIGURE 3
Scatter-plot diagram comparing research nurse–assigned and physician-assigned CDS scores (n =
208). Marker size is proportional to the number of observations. Solid line represents best fit through
the data points.
95% CI –0.18, 0.12). CDS did not correlate with percent weight gain in the
subgroup of participants with baseline
scores $5 (r = –0.06; 95% CI –0.29,
0.18). Of the variables evaluated to assess construct validity, serum bicarbonate exhibited the strongest correlation
with the CDS score (r = –0.35; 95% CI
–0.46, –0.23). Subanalysis in participants ,36 months old and in those
with moderate/severe dehydration (CDS
$5) did not reveal any additional strong
correlations.
respectively). Reliability in the subgroup with moderate/severe dehydration
(CDS $5) was fair (k = 0.30; 95% CI 0.13,
0.45)
CDS Validity
CDS did not correlate with percent
weight gain in the 180 participants for
whom discharge weight was available
(r = –0.04; 95% CI –0.19, 0.10) (Table 2).
The exclusion of 9 values deemed to be
outliers did not significantly alter the
correlation coefficient (Fig 4, r = –0.03;
TABLE 2 Correlation Between the Baseline CDS Score and Clinical Markers Reflective of Disease
Severity
Correlation Coefficient (95% CI)a
Variable
Percent weight gain
Vomiting episodesb
Diarrhea episodesb
Heart rate
Respiratory rate
Capillary refill time
Serum bicarbonate
Serum pH
Length of stay
Physician dischargeability
Likert scorec
All Participants (n = 226)
Age ,36 mo (n = 140)
Age $36 mo (n = 86)
20.04 (20.19, 0.10)
0.04 (20.09, 0.17)
0.14 (0.01, 0.27)
0.09 (20.04, 0.22)
0.18 (0.05, 0.30)
0.23 (0.10, 0.35)
20.35 (20.46, 20.22)
20.23 (20.36, 20.10)
0.24 (0.11, 0.36)
0.18 (0.05, 0.31)
20.12 (20.30, 0.07)
0.17 (0.01, 0.33)
0.10 (20.07, 0.26)
0.07 (20.10, 0.23)
0.17 (0.00, 0.32)
0.34 (0.19, 0.48)
20.31 (20.46, 20.15)
20.25 (20.40, 20.08)
0.20 (20.04, 0.36)
0.20 (0.03, 0.35)
0.05 (20.19, 0.28)
20.16 (20.36, 20.06)
0.16 (20.05, 0.36)
20.01 (20.22, 0.20)
0.13 (20.08, 0.34)
0.04 (20.18, 0.25)
20.36 (20.54, 20.16)
20.17 (20.38, 0.06)
0.25 (0.04, 0.44)
0.14 (20.07, 0.34)
a Pearson correlation coefficient (serum bicarbonate and serum pH) or Spearman rank correlation coefficient (percent
weight gain, vomiting and diarrhea episodes, heart rate, respiratory rate, capillary refill time, length of stay, and physician
dischargeability Likert score).
b Number of episodes in the 24 hours before ED presentation, as reported by parents/guardians.
c Physician willingness to discharge patient at 2 hours after initiation of intravenous rehydration, ranging from 1 (strongly
agree that patient is ready for discharge, based on clinical appearance and vital signs) to 5 (strongly disagree that patient is
ready for discharge).
CDS scores decreased over time after
the initiation of intravenous rehydration therapy (Fig 5). Median scores at 2
hours and 4 hours were significantly
changed from baseline (P , .001 at
both time points), indicating responsiveness of the scale to change.
For the outcome of hospitalization, the
area under the ROC curve was 0.65 (95%
CI 0.57, 0.73) (Fig 6). Test characteristics
were optimal by using a baseline CDS
cut point of $5, which yielded a sensitivity of 62% (95% CI 48, 75), specificity
of 66% (95% CI 58, 73), positive likelihood ratio of 1.8 (95% CI 1.3, 2.4), and
negative likelihood ratio of 0.59 (95% CI
0.41, 0.84). Positive predictive and negative predictive values were 35% (95%
CI 25, 45) and 85% (95% CI 78, 91), respectively. The CDS scores assigned at
2 and 4 hours after treatment initiation
had areas under the curve (AUC) of 0.70
(95% CI 0.62, 0.78) and 0.72 (95% CI 0.64,
0.80) respectively. For participants with
baseline scores $5, the area under the
ROC curve was 0.53 (95% CI: 0.41, 0.65).
Discriminative validity of the baseline
CDS did not differ significantly between
participants ,36 months of age (AUC =
0.66; 95% CI 0.56, 0.76) and those $36
months of age (AUC = 0.60; 95% CI 0.46,
0.75), P = .52.
DISCUSSION
In this study, the reliability and validity
of the CDS10 was evaluated in a cohort
PEDIATRICS Volume 129, Number 5, May 2012
Downloaded from pediatrics.aappublications.org by guest on August 22, 2014
e1215
12
11
10
9
8
7
6
5
4
Percent weight gain 3
2
1
0
-1
-2
-3
-4
-5
-6
1
2
3
4
5
6
7
8
CDS score
FIGURE 4
Scatter-plot diagram of percent weight gain at the time of ED discharge (y-axis) by CDS score at treatment
initiation (x-axis). Black squares indicate median percent weight gain values (n = 171). CDS did not
correlate with percent weight gain (r = –0.03; 95% CI –0.18, 0.12).
of children with gastroenteritis and
dehydration. Interobserver reliability
was found to be moderate.21 Although
responsiveness to change was detected,
no statistically significant correlation
was identified between CDS scores and
percent weight gain, a proxy measure
of fluid deficit. There was a modest
correlation with other parameters reflective of dehydration severity: episodes of diarrhea, serum bicarbonate
and pH, length of stay, and physician
willingness to discharge 2 hours after
initiation of intravenous rehydration.
Similar associations were observed
when the CDS’s validity was evaluated
as a dichotomous variable by using a
cut point of $5.
A previous evaluation in children 1
month to 5 years of age found the CDS
to be useful in predicting the need for
intravenous rehydration and length of
stay.11 These findings were replicated
at a different tertiary care center, where
an association between CDS score and
use of laboratory tests was documented12;
however, abnormal results were not
significantly different across CDS severity assignments. Nonetheless, the
authors concluded that the scale is
externally valid, and might be useful
in guiding therapy (eg, initiation of
intravenous rehydration in triage for
CDS $5). Their findings, however, may
simply indicate that physicians use the
clinical findings in the CDS to make
treatment decisions; not that children
with higher CDS scores require that
such treatment decisions be made.
Hence, it may be inappropriate to conclude that those clinical findings are
necessarily associated with dehydration. In addition, because the assigned
CDS score was correlated with the use
of intravenous rehydration, it is to be
expected that length of stay is prolonged in such patients.23 Both validation studies were further limited
by inclusion of few participants with
moderate/severe dehydration. Thus,
the performance of the CDS could not
be evaluated in such children. In the
current study, eligibility was defined
based on evidence of dehydration sufficient to require intravenous rehydration. Consequently, we were able to
evaluate the CDS in a relatively large
number of patients with moderate to
severe dehydration. Including only children receiving intravenous rehydration
eliminates the potential confounding
effect of treatment on length of stay.
TABLE 3 Characteristics of Study Participants
Age, median (IQR), y
Male, n (%)
Triage weight, median (IQR), kg
CTAS score, median (IQR)
Vomiting, median (IQR), episodesb
Diarrhea, median (IQR), episodesb
Heart rate, mean 6 SD, beats per min
Respiratory rate, median (IQR), breaths per min
Capillary refill time, median (IQR), s
Serum pH, mean 6 SD
Serum bicarbonate, mean 6 SD, mmol/L
Physician dischargeability Likert score, median (IQR)c
Length of stay, median (IQR), h
Percent weight gain, median (IQR), %
All Participants
(n = 226)
Participants With Moderate/Severe
Dehydration (CDS $5) (n = 92)
Participants With None/Some
Dehydration (CDS ,5) (n = 134)
P Valuea
2.21 (1.36, 3.96)
117 (51.8)
12.6 (9.98, 16.2)
3 (2.5, 3)
8 (3, 14)
4 (0, 10)
127 6 20
28 (24, 30)
0.70 (0.56, 1.04)
7.36 6 0.06
18.0 6 3.65
4 (2, 4)
5.41 (4.25, 14.6)
2.24 (0.67, 4.67)
1.96 (1.30, 3.54)
44 (47.8)
11.9 (9.65, 15.7)
3 (2, 3)
8 (4, 14.8)
5 (1.25, 10.8)
126 6 21
28 (24, 32)
0.78 (0.61, 1.20)
7.35 6 0.06
16.3 6 3.34
4 (2, 4)
7.07 (4.67, 25.5)
2.19 (0.65, 4.04)
2.44 (1.41, 4.24)
73 (54.5)
13.1 (10.1, 16.7)
3 (3, 3)
7.5 (3, 13)
3 (0, 8)
130 6 17
28 (24, 30)
0.67 (0.55, 0.88)
7.37 6 0.05
19.2 6 3.39
3.5 (2, 4)
4.88 (4.00, 8.03)
2.27 (0.68, 4.96)
.13
.33
.10
,.001
.83
.05
.15
.02
.001
,.001
,.001
.02
,.001
.53
CTAS, Canadian Triage and Acuity Scale.
a P values are for comparison between participants with moderate/severe dehydration versus those with none/some dehydration.
b Number of episodes in the 24 hours before ED presentation, as reported by parents/guardians.
c Physician willingness to discharge patient at 2 hours after initiation of intravenous rehydration, ranging from 1 (strongly agree that patient is ready for discharge, based on clinical
appearance and vital signs) to 5 (strongly disagree that patient is ready for discharge).
e1216
KINLIN and FREEDMAN
Downloaded from pediatrics.aappublications.org by guest on August 22, 2014
ARTICLE
significantly influenced by the treatments provided. A child who receives
a large volume of intravenous fluids
will gain a significant amount of
weight before discharge even if he
or she was euvolemic at presentation.
In an ideal world, the score would have
been derived by using serial weights
after discharge. 25
FIGURE 5
Scores on the CDS at 30-minute intervals. Time 0 represents the initiation of intravenous rehydration
therapy. As in a standard box plot, sample medians are depicted by the center lines of the boxes; the upper
and lower limits represent the 75th and 25th percentiles, respectively. Means are superimposed as black
squares.
Although interobserver agreement was
evaluated in the derivation study, it included only 58 paired scores assigned
exclusivelybyresearchteam members.10
We evaluated the real-world performance of the CDS by including individuals without extensive training in use
of the score. Our finding of moderate
interobserver agreement, whereas the
derivation study found substantial
agreement (k = 0.77), suggests that the
CDS may produce inconsistent results
when used by those without formal
training. We found, moreover, that even
the baseline CDS score assigned by
trained nurses was not a clinically
useful tool for predicting hospitalization. The AUC did improve, however,
when the time 2- and 4-hour CDS
scores were used, indicating that the
score may have some utility in assessing improvement or predicting
hospitalization once treatment is completed.24
The CDS was derived by using weight
change as a measure of “true” dehydration status,10 with the assumption that
ED discharge weight reflects preillness
weight. This assumption is questionable, as the ED discharge weight is
FIGURE 6
ROC curve for the outcome of hospitalization, based on CDS score assigned at treatment initiation by the
research nurse (solid line), 2 hours after treatment initiation (dashed line), and 4 hours after treatment
initiation (dotted line); n = 226. AUC is 0.65 (95% CI 0.57, 0.73), 0.70 (95% CI 0.62, 0.78), and 0.72 (95% CI
0.64, 0.80), respectively.
Despite these concerns, the derivation10
and validation11,12 studies do show that
the CDS can accurately identify children
with none/minimal dehydration. Our
data caution against the use of the
CDS to predict outcomes in children
with evidence of dehydration and
does not support the authors’ suggestion that laboratory tests and intravenous rehydration be initiated in
all children with a CDS .4.12 Such
a recommendation would likely result
in increased intravenous rehydration
use, which is already excessive,26,27 as
most children with moderate dehydration can be rehydrated orally.1 Triage
nurses should focus on ensuring the
success of oral rehydration therapy,
as very few children in developed
countries have severe dehydration.
Furthermore, because vomiting is
a major driving force in the use of
intravenous rehydration,28 antiemetic agents should be used selectively in children with elevated CDS
scores. 29
Several limitations should be recognized. First, this study represents a
secondary analysis of data collected as
part of an intervention trial.30 Nevertheless, these data were collected
prospectively under supervised and
standardized conditions. There should
not have been bias to the study question, as participants were randomly
assigned to treatment groups, thereby
balancing the CDS distribution between groups. Second, the study used
an untrained individual who was unaware that interobserver agreement
was being evaluated. Although this
PEDIATRICS Volume 129, Number 5, May 2012
Downloaded from pediatrics.aappublications.org by guest on August 22, 2014
e1217
pragmatic approach is a strength of
the study, it may have resulted in
physicians assigning scores without
paying close attention to the actual
values assigned. Finally, we used percent weight gain at discharge as a
proxy measure of fluid deficit even
though this may not accurately reflect
dehydration severity; however, as this
method was used in the derivation
study,10 we expected to find a similar
correlation.
CONCLUSIONS
In a cohort of children administered
intravenous rehydration secondary
to gastroenteritis, the CDS exhibited
moderate interobserver reliability and
limited criterion, construct, and discriminatory validity. Although previous research has found that it can differentiate
children with none/some and moderate/
severe dehydration, our data do not
support its use in isolation to make
treatment determinations (ie, need for
intravenous rehydration) among children with evidence of dehydration.
ACKNOWLEDGMENTS
We are indebted to the ED nurses, administrative staff, and physicians at
The Hospital for Sick Children for their
assistance with patient recruitment
and adherence to the protocol and to
the research nurses for their instrumental role in patient enrollment.
REFERENCES
1. King CK, Glass R, Bresee JS, Duggan C;
Centers for Disease Control and Prevention. Managing acute gastroenteritis among
children: oral rehydration, maintenance, and
nutritional therapy. MMWR Recomm Rep.
2003;52(RR-16):1–16
2. National Collaborating Centre for Women’s
and Children’s Health. Diarrhoea and vomiting caused by gastroenteritis: diagnosis,
assessment and management in children
younger than 5 years: Commissioned by the
National Institute for Health and Clinical
Excellence; April 2009; London, UK. London,
UK: RCOG Press; 2009
3. WHO/UNICEF. Joint Statement: Clinical Management of Acute Diarrhoea (WHO/FCH/
CAH/04.07). Geneva, New York: World Health
Organization, Department of Child and
Adolescent Health and Development, and
United Nations Children’s Fund, Programme
Division; 2004
4. Liebelt EL. Clinical and laboratory evaluation and management of children with
vomiting, diarrhea, and dehydration. Curr
Opin Pediatr. 1998;10(5):461–469
5. Practice parameter: the management of
acute gastroenteritis in young children.
American Academy of Pediatrics, Provisional
Committee on Quality Improvement, Subcommittee on Acute Gastroenteritis. Pediatrics. 1996;97(3):424–435
6. Kim MS, Somers MJG. Fluid and electrolyte
physiology and therapy. In: McMillan JA,
Feigin RD, DeAngelis CD, Jones MD, eds.
Oski’s Pediatrics: Principles & Practice. 4th
ed. Philadelphia, PA: Lippincott Williams &
Wilkins; 2006:54–70
7. Sandhu BK; European Society of Pediatric
Gastroenterology, Hepatology and Nutrition Working Group on Acute Diarrhoea.
Practical guidelines for the management
e1218
8.
9.
10.
11.
12.
13.
14.
15.
16.
of gastroenteritis in children. J Pediatr
Gastroenterol Nutr. 2001;33(suppl 2):S36–
S39
Vega RM, Avner JR. A prospective study of
the usefulness of clinical and laboratory
parameters for predicting percentage of
dehydration in children. Pediatr Emerg
Care. 1997;13(3):179–182
Freedman SB, Thull-Freedman JD. Pediatric
dehydration assessment and oral rehydration therapy. Pediatr Emerg Med Rep. 2008;
13:13–28
Friedman JN, Goldman RD, Srivastava R,
Parkin PC. Development of a clinical dehydration scale for use in children between
1 and 36 months of age. J Pediatr. 2004;145
(2):201–207
Goldman RD, Friedman JN, Parkin PC. Validation of the clinical dehydration scale for
children with acute gastroenteritis. Pediatrics. 2008;122(3):545–549
Bailey B, Gravel J, Goldman RD, Friedman
JN, Parkin PC. External validation of the
clinical dehydration scale for children with
acute gastroenteritis. Acad Emerg Med.
2010;17(6):583–588
Streiner DL, Norman GR. Health Measurement Scales: A Practical Guide to Their
Development and Use. 3rd ed. New York:
Oxford University Press Inc.; 2003
Parkin PC, Macarthur C, Khambalia A,
Goldman RD, Friedman JN. Clinical and
laboratory assessment of dehydration
severity in children with acute gastroenteritis. Clin Pediatr (Phila). 2010;49(3):
235–239
Steiner MJ, DeWalt DA, Byerley JS. Is this
child dehydrated? JAMA. 2004;291(22):2746–
2754
Gorelick MH, Shaw KN, Baker MD. Effect of
ambient temperature on capillary refill in
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
healthy children. Pediatrics. 1993;92(5):
699–702
Freedman SB, Powell E, Seshadri R. Predictors of outcomes in pediatric enteritis:
a prospective cohort study. Pediatrics. 2009;
123(1). Available at: www.pediatrics.org/cgi/
content/full/123/1/e9
Freedman SB, Parkin PC, Willan AR, Schuh
S. Rapid versus standard intravenous rehydration in paediatric gastroenteritis: pragmatic blinded randomised clinical trial. BMJ.
2011;343:d6976
Terwee CB, Bot SD, de Boer MR, et al.
Quality criteria were proposed for measurement properties of health status questionnaires. J Clin Epidemiol. 2007;60(1):
34–42
Fleiss JL, Cohen J. The equivalence of
weighted kappa and the intraclass correlation coefficient as measures of reliability.
Educ Psychol Meas. 1973;33:613–619
Landis JR, Koch GG. The measurement of
observer agreement for categorical data.
Biometrics. 1977;33(1):159–174
Tukey JW. Exploratory Data Analysis. Reading, MA: Addison-Wesley; 1977
Bender BJ, Ozuah PO. Intravenous rehydration for gastroenteritis: how long does
it really take? Pediatr Emerg Care. 2004;20
(4):215–218
Fan J, Upadhye S, Worster A. Understanding
receiver operating characteristic (ROC)
curves. CJEM. 2006;8(1):19–20
Gorelick MH, Shaw KN, Murphy KO. Validity and reliability of clinical signs in
the diagnosis of dehydration in children. Pediatrics. 1997;99(5). Available at:
www.pediatrics.org/cgi/content/full/99/
5/E6
Conners GP, Barker WH, Mushlin AI,
Goepp JG. Oral versus intravenous:
KINLIN and FREEDMAN
Downloaded from pediatrics.aappublications.org by guest on August 22, 2014
ARTICLE
rehydration preferences of pediatric
emergency medicine fellowship directors. Pediatr Emerg Care. 2000;16(5):
335–338
27. Santosham M. Oral rehydration therapy:
reverse transfer of technology. Arch Pediatr
Adolesc Med. 2002;156(12):1177–1179
28. Ozuah PO, Avner JR, Stein RE. Oral rehydration, emergency physicians, and practice
parameters: a national survey. Pediatrics.
2002;109(2):259–261
29. DeCamp LR, Byerley JS, Doshi N, Steiner MJ.
Use of antiemetic agents in acute gastroenteritis: a systematic review and meta-analysis.
Arch Pediatr Adolesc Med. 2008;162(9):858–
865
30. Freedman SB, Parkin PC, Willan AR, Schuh
S. Rapid versus standard intravenous rehydration in paediatric gastroenteritis: pragmatic blinded randomised clinical trial. BMJ.
2011;343:d6976
(Continued from first page)
PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).
Copyright © 2012 by the American Academy of Pediatrics
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose. Dr Freedman previously served as
a consultant for Baxter Healthcare.
FUNDING: Ms Kinlin received funding from the SickKids Summer Research Program. The original study on which the present article is based was funded by a grant
from The Physicians’ Services Incorporated Foundation.
PEDIATRICS Volume 129, Number 5, May 2012
Downloaded from pediatrics.aappublications.org by guest on August 22, 2014
e1219
Evaluation of a Clinical Dehydration Scale in Children Requiring Intravenous
Rehydration
Laura M. Kinlin and Stephen B. Freedman
Pediatrics 2012;129;e1211; originally published online April 23, 2012;
DOI: 10.1542/peds.2011-2985
Updated Information &
Services
including high resolution figures, can be found at:
http://pediatrics.aappublications.org/content/129/5/e1211.full.
html
References
This article cites 25 articles, 8 of which can be accessed free
at:
http://pediatrics.aappublications.org/content/129/5/e1211.full.
html#ref-list-1
Citations
This article has been cited by 1 HighWire-hosted articles:
http://pediatrics.aappublications.org/content/129/5/e1211.full.
html#related-urls
Post-Publication
Peer Reviews (P3Rs)
One P3R has been posted to this article:
http://pediatrics.aappublications.org/cgi/eletters/129/5/e1211
Subspecialty Collections
This article, along with others on similar topics, appears in
the following collection(s):
Emergency Medicine
http://pediatrics.aappublications.org/cgi/collection/emergency
_medicine_sub
Permissions & Licensing
Information about reproducing this article in parts (figures,
tables) or in its entirety can be found online at:
http://pediatrics.aappublications.org/site/misc/Permissions.xh
tml
Reprints
Information about ordering reprints can be found online:
http://pediatrics.aappublications.org/site/misc/reprints.xhtml
PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly
publication, it has been published continuously since 1948. PEDIATRICS is owned, published,
and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk
Grove Village, Illinois, 60007. Copyright © 2012 by the American Academy of Pediatrics. All
rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.
Downloaded from pediatrics.aappublications.org by guest on August 22, 2014