Letters

RESEARCH LETTER
Pneumonia in Children Presenting to the Emergency
Department With an Asthma Exacerbation
Asthma and pneumonia are 2 of the most common reasons
children visit the emergency department.1 These conditions
have overlapping features and can occur concomitantly. Sub-
sequently, physicians often obtain chest radiographs to diag-
nose pneumonia in children with asthma, although less than
5% of children have pneumonia that can be radiographically
confirmed (hereafter referred to as radiographic pneumonia).2
The consequences of the overuse of radiography include in-
creased time in the hospital, unnecessary radiation expo-
sure, increased costs, and inappropriate antibiotic use due to
equivocal imaging findings.3 Given the high rate of normal
chest radiographs and the consequences of unnecessary

Table 1. Factors Associated With Radiographic Pneumonia Among Children Presenting

to the Emergency Department With Asthma Exacerbation

Children, Children, No. (%)a

Total No. (%) No Pneumonia Pneumonia
Factor (N = 4708) (n = 4428) (n = 280) P Valuea OR (95% CI)b
Female sex 1965 (41.7) 1830 (41.3) 135 (48.2) .02 1.29 (1.01-1.64)
Age ≥5 y 2573 (54.7) 2385 (53.9) 188 (67.1) <.001 1.91 (1.47-2.48)
Racec
White 2058 (43.8) 1931 (43.7) 127 (45.4)
African American 2051 (43.7) 1927 (43.6) 124 (44.3) .53
Other 587 (12.5) 558 (12.6) 29 (10.4)
Hispanic ethnicity 270 (5.7) 258 (5.8) 12 (4.3) .28
Insurance
Private 1607 (34.1) 1512 (34.1) 95 (33.9)
Government 2982 (63.3) 2804 (63.3) 178 (63.6) >.99
Self-pay 119 (2.5) 112 (2.5) 7 (2.5)
Medical history 488 (10.4) 449 (10.1) 39 (13.9) .04 1.32 (0.92-1.89)
of pneumonia
Taking antibiotics 233 (4.9) 215 (4.9) 18 (6.4) .24
prior to ED visit
Tachycardiad
At triage 2993 (63.6) 2798 (63.2) 195 (69.6) .03 1.16 (0.87-1.54)
During ED visitd 3888 (82.6) 3646 (82.3) 242 (86.4) .08
Tachypnead
At triage 3169 (67.3) 2964 (66.9) 205 (73.2) .03 1.18 (0.88-1.59)
During ED visit 3838 (81.5) 3603 (81.4) 235 (83.9) .30
Oxygen saturation at triagee
Abbreviations: ED, emergency
Normal (≥94%) 3804 (80.8) 3595 (81.2) 209 (74.6) department; OR, odds ratio.
Mild hypoxia (92%-94%) 508 (10.8) 471 (10.6) 37 (13.2) a
Univariable analyses (children with
Moderate hypoxia 212 (4.5) 195 (4.4) 17 (6.1) .06 radiographic pneumonia are those
(90%-92%) with pneumonia that can be
Severe hypoxia (<90%) 165 (3.5) 150 (3.4) 15 (5.4) radiographically confirmed).
b
Lowest oxygen saturation Multivariable analyses.
during ED stay c
A total of 4696 children (4416
Normal (≥94%) 3078 (65.4) 2920 (65.9) 158 (56.4) 1 [Reference] without pneumonia and 280 with
Mild hypoxia (92%-94%) 815 (17.3) 761 (17.2) 54 (19.3) 1.28 (0.93-1.78) pneumonia).
d
Moderate hypoxia 388 (8.2) 361 (8.2) 27 (9.6) <.001 1.27 (0.82-1.95) Tachycardia and tachypnea for age
(90%-92%) were defined as having a heart rate
Severe hypoxia (<90%) 408 (8.7) 369 (8.3) 39 (13.9) 1.88 (1.28-2.74) or respiratory rate in the 99th
percentile or greater for age.5
Fever e
To avoid collinearity, we only
During ED stay 1240 (26.3) 1139 (25.7) 101 (36.1) <.001 1.65 (1.27-2.15) included 1 variable for tachycardia,
(temperature ≥38°C)
oxygen saturation, and fever in the
At triage (temperature 974 (20.7.1) 893 (20.2) 81 (28.9) .001 multivariable model based on the
≥38°C)e best model fit.

jamapediatrics.com (Reprinted) JAMA Pediatrics August 2016 Volume 170, Number 8 803

Copyright 2016 American Medical Association. All rights reserved.

Downloaded From: on 10/21/2018

 Letters

radiographs, more accurate criteria are needed to identify chil-

Table 2. Odds of Radiographic Pneumonia for Combinations of
dren with asthma who are at increased risk for pneumonia. This Significant Factors From Multivariable Analysis
study aims to identify risk factors for pneumonia in children
Significant Factors Odds Ratio (95% CI)a
with asthma exacerbations.
Female sex and fever 2.12 (1.47-3.07)
Female sex and severe hypoxia (<90%) 2.39 (1.51-3.79)
Methods | This is a retrospective cohort study of children 2 to
Age ≥5 y and female sex 2.47 (1.72-3.54)
18 years of age who presented to the emergency department
Severe hypoxia (<90%) and fever 3.08 (1.92-4.34)
of the Cincinnati Children’s Hospital Medical Center between
January 1, 2010, and December 31, 2013, with an asthma ex- Age ≥5 y and fever 3.17 (2.14-4.69)

acerbation. Children were identified using a validated algo- Age ≥5 y and severe hypoxia (<90%) 3.57 (2.23-5.72)

rithm of an International Classification of Diseases, Ninth
Revision, Clinical Modification diagnosis code for asthma (code
493.X) in the first 3 diagnosis positions and receipt of 1 or more
doses of albuterol sulfate in the emergency department.4 We
excluded children who were younger than 2 years of age to
minimize including infants with bronchiolitis. The Cincin-
nati Children’s Hospital Medical Center institutional review
board approved this study. Informed consent was not
required because the data were deidentified.
Data were abstracted from the electronic health record
(Epic Systems Corporation). Potential factors associated with
pneumonia included demographics, medical history, medi-
cations, and vital signs (Table 1). The primary outcome was
radiographic pneumonia as determined by 2 investigators
who manually reviewed the final radiologist reports using a
coding manual. Meetings to clarify disagreements occurred
weekly. Abnormal radiographs were classified as definite
atelectasis, favoring atelectasis, atelectasis vs pneumonia,
favoring pneumonia, and definite pneumonia. Radiographic
pneumonia was present if the radiologist favored or defini-
tively diagnosed pneumonia.2 Less than 1.9% of data were
missing from any variable, so a complete case analysis was
performed.
The t test and the χ2 test were performed to examine the
relationship between each potential risk factor and radio-
graphic pneumonia. Variables with P ≤ .20 in the univariable
analysis were included in a multivariable logistic regression
model to determine the association of risk factors with radio-
graphic pneumonia.
Results | There were 14 007 emergency department visits for
asthma exacerbation; chest radiography was performed dur-
ing 4708 of 14 007 emergency department visits (33.6%).
Radiographic pneumonia was present in 280 of the 4708 chil-
dren who underwent chest radiography (5.9%). Factors inde-
pendently associated with pneumonia included an age of 5
years or older, female sex, fever, and an oxygen saturation of
less than 90% (Table 1 and Table 2).
Discussion | Less than 2% of patients with an asthma exacerba-
tion had radiographic pneumonia, yet one-third underwent
chest radiography. Being 5 years of age or older, female, fe-
Female sex, severe hypoxia (<90%), and fever 3.96 (2.31-6.78)
Age ≥5 y, female sex, and fever 4.07 (2.55-6.50)
Age ≥5 y, female sex, and severe hypoxia (<90%) 4.59 (2.69-7.84)
Age ≥5 y, severe hypoxia (<90%), and fever 5.90 (3.37-10.30)
Severe hypoxia (<90%), female sex, age ≥5 y, 7.59 (4.09-14.10)
and fever
a
Interpreted as the odds of having radiographic pneumonia (ie, pneumonia that
can be radiographically confirmed) with all factors listed compared with
having none of those factors.
Our study is limited by its retrospective nature; however,
prior studies found that many of the variables that we did not
include (eg, auscultatory findings) were not associated with
pneumonia in wheezing children.2 The validated algorithm that
we used to identify children may lead to misclassification be-
cause we are excluding children with asthma but no active
wheezing by requiring albuterol use.4 This will bias our re-
sults toward the null because we are underestimating the total
burden of asthma, thus likely underestimating the magni-
tude of effect. In the final models, we only included children
who underwent radiography of the chest, which may result in
an overestimation of pneumonia. The use of radiographic read-
ings from clinical care may introduce variability and bias, but
knowledge of clinical history represents real-world applica-
tion and does not influence a radiologist’s interpretation.6 Our
results suggest that routine use of radiography for younger chil-
dren with asthma who are not febrile or hypoxic should be
discouraged.
Todd A. Florin, MD, MSCE
Hannah Carron, BS
Guixia Huang, MS
Samir S. Shah, MD, MSCE
Richard Ruddy, MD
Lilliam Ambroggio, PhD, MPH
Author Affiliations: Division of Emergency Medicine, Cincinnati Children’s
Hospital Medical Center, Cincinnati, Ohio (Florin, Ruddy); Department of
Pediatrics, University of Cincinnati College of Medicine, Cincinnati, Ohio (Florin,
Carron, Shah, Ruddy, Ambroggio); Division of Biostatistics and Epidemiology,
Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio (Huang,
Ambroggio); Division of Hospital Medicine, Cincinnati Children’s Hospital
Medical Center, Cincinnati, Ohio (Shah, Ambroggio); Division of Infectious
Diseases, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio (Shah).

brile, and/or hypoxic was statistically associated with radio-
graphic pneumonia. A previous study reported that 4.9% of
children with wheezing (many <2 years of age) had pneumo-
nia; fever and hypoxia were the strongest pneumonia
predictors.2 Our study specifically examines pneumonia in
older children with an asthma exacerbation.
Corresponding Author: Todd A. Florin, MD, MSCE, Division of Emergency
Medicine, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Ave, ML
2008, Cincinnati, OH 45229 (todd.florin@cchmc.org).
Published Online: June 6, 2016. doi:10.1001/jamapediatrics.2016.0310.
Author Contributions: Dr Florin had full access to all of the data in the study
and takes responsibility for the integrity of the data and the accuracy of the data
analysis.

804 JAMA Pediatrics August 2016 Volume 170, Number 8 (Reprinted) jamapediatrics.com

Copyright 2016 American Medical Association. All rights reserved.

Downloaded From: on 10/21/2018


Study concept and design: Florin, Carron, Ruddy, Ambroggio.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Florin.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Florin, Carron, Huang, Ambroggio.
Obtained funding: Ambroggio.
Administrative, technical, or material support: Ambroggio.
Study supervision: Florin, Shah, Ruddy, Ambroggio.
Conflict of Interest Disclosures: None reported.
Funding/Support: Dr Florin received support from the National Center for
Research Resources and the National Center for Advancing Translational
Sciences/National Institutes of Health (grant 5KL2TR000078-05). This study
was also funded in part by National Institutes of Health grant 1T35HL113229-01
(Ms Carron).
Role of the Funder/Sponsor: The funding agencies played no role in the design
and conduct of the study; collection, management, analysis, or interpretation of
the data; preparation, review, or approval of the manuscript; and decision to
submit the manuscript for publication.
Disclaimer: Dr Shah is the JAMA Pediatrics Clinical Challenge Editor but was not
involved in the editorial review or decision to accept the manuscript for
publication.
Additional Contributions: We acknowledge John G. Filcik, BS, for assistance in
the review of the radiograph reports and Mona Ho, MS, for statistical assistance.
No compensation from a funding sponsor was received by either.
1. Alpern ER, Stanley RM, Gorelick MH, et al; Pediatric Emergency Care Applied
Research Network. Epidemiology of a pediatric emergency medicine research
network: the PECARN Core Data Project. Pediatr Emerg Care. 2006;22(10):
689-699.
2. Mathews B, Shah S, Cleveland RH, Lee EY, Bachur RG, Neuman MI. Clinical
predictors of pneumonia among children with wheezing. Pediatrics. 2009;124
(1):e29-e36.
3. Schuh S, Lalani A, Allen U, et al. Evaluation of the utility of radiography in
acute bronchiolitis. J Pediatr. 2007;150(4):429-433.
4. Sharifi M, Krishanswami S, McPheeters ML. A systematic review of validated
methods to capture acute bronchospasm using administrative or claims data.
Vaccine. 2013;31(suppl 10):K12-K20.
5. Fleming S, Thompson M, Stevens R, et al. Normal ranges of heart rate and
respiratory rate in children from birth to 18 years of age: a systematic review of
observational studies. Lancet. 2011;377(9770):1011-1018.
6. Test M, Shah SS, Monuteaux M, et al. Impact of clinical history on chest
radiograph interpretation. J Hosp Med. 2013;8(7):359-364.
Trends in Breastfeeding Initiation and Duration
by Birth Weight Among US Children, 1999-2012
In the United States, breastfeeding initiation rates have risen
to 80%.1 We report secular trends of breastfeeding initiation
and duration by birth weight using nationally representative
data from the National Health and Nutrition Examination
Survey (NHANES).
Methods | The NHANES is a complex, stratified, multistage prob-
ability sample of the US civilian noninstitutionalized popula-
tion, conducted by the National Center for Health Statistics,
Centers for Disease Control and Prevention. The NHANES
was approved by the National Center for Health Statistics
Research Ethics Review Board.
Using data from 13 859 children ages 0 to 6 years who
participated in the NHANES from 1999 to 2014, we con-
structed birth year cohorts (1997-2000, 2001-2004, 2005-
2008, and 2009-2012) and birth weight categories (<2500 g
[low birth weight (LBW)], ≥2500 g). Birth weight and breast-
feeding history (ever breastfed or received breast milk) were
obtained by proxy interview, most commonly by a parent.2
jamapediatrics.com
Copyright 2016 American Medical Association. All rights reserved.
Downloaded From: on 10/21/2018
Letters
The unweighted response rate ranged from 77% to 93%.3
Written parental consent was obtained for those younger
than 18 years.
We estimated the proportion of infants ever breastfed (ini-
tiated) and those reporting any breastfeeding at 1 month, 4
months, and 6 months by birth weight categories and birth year
cohorts. Infants were excluded if they had not attained the age
at which breastfeeding was being estimated. Significance was
set at α < .05 for adjusted Wald method and orthogonal con-
trast matrices for linear tests of trends over time. Taylor se-
ries linearization was used for variance estimation and 95% CIs
were computed by the Wald method. Kaplan-Meier curves and
log-rank tests assessed differences in the duration of breast-
feeding by birth weight cohort from 2009 to 2012. LOESS
smoothed curves and 95% CIs from Kaplan-Meier curves are
presented. Infants were censored at the time of interview if
they were currently receiving breast milk (12% of respon-
dents, unweighted n = 324). Overall, 8% of infants were LBW
(unweighted n = 1279).
SAS version 9.3 (SAS Institute) and SUDAAN version 11.0
(RTI International) were used. Survey design variables and in-
terview sample weights, which account for differential prob-
abilities of selection, nonresponse, noncoverage, and sample
design, were used to obtain estimates representative of the
civilian noninstitutionalized US population.
Results | Among all infants, breastfeeding initiation rates
increased from 65.7% (95% CI, 62.4%-69.1%) in 1997-2000
to 76.7% (95% CI, 73.5%-79.8%) (P < .001) in 2009-2012.
Over the same period, breastfeeding initiation rates
increased from 48.8% (95% CI, 39.6%-57.9%) to 65.4% (95%
CI, 55.4%-75.3%) (P = .01) for LBW infants and from 67.4%
(95% CI, 64.0%-70.8%) to 77.5% (95% CI, 74.3%-80.8%)
(P < .001) for infants weighing 2500 g or more at birth
(Table). An increasing secular trend in the percentage of
infants still breastfeeding at 1 month (59.7% [95% CI, 56.2%-
63.2%] to 68.7% [95% CI, 65.0-72.4%]; P < .001), 4 months
(42.6% [95% CI, 39.3%-46.0%] to 49.6% [95% CI, 45.6%-
53.7%]; P = .005), and 6 months (36.2% [95% CI, 33.0%-
39.5%] to 43.1% [95% CI, 39.3%-47.0%]; P = .002) was
observed for infants weighing 2500 g or more. For LBW
infants, the increase in the percentage still breastfeeding
over time was not significant at any time after birth. This is
likely owing to insufficient power arising from the small
sample of LBW infants.
Within each birth cohort, the percentage of infants initi-
ating breastfeeding and breastfeeding at 1, 4, and 6 months was
significantly smaller for LBW infants compared with their
heavier counterparts. For example, in 2009-2012, 43.1% of in-
fants (95% CI, 39.3%-47.0%) weighing 2500 g or more were
breastfeeding at 6 months compared with 27.7% of LBW in-
fants (95% CI, 18.2%-37.2%) (P < .001). The only exception was
among those still breastfeeding at 1 month in 2001-2004; this
difference was not statistically significant.
The Figure presents the smoothed Kaplan-Meier curves
for the overall duration of breastfeeding in the first year of
life by birth weight categories. Infants who never breastfed
were censored at time zero. Among infants born in
(Reprinted) JAMA Pediatrics August 2016 Volume 170, Number 8 805