Original Studies
Risk Factors for Severe Community-aquired
Pneumonia Among Children Hospitalized With CAP
Younger Than 5 Years of Age
Wei Shan, MPH,* Ting Shi, MD,† Kaile Chen, MPH,* Jian Xue, MD,† Yin Wang, MPH,* Jia Yu, MPH,*
Genming Zhao, PhD,* Jianmei Tian, MD,† and Tao Zhang, PhD*
Downloaded from https://journals.lww.com/pidj by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3Nu6/x40YHMN2JQVJyaPZSsRRkRxOhW8cR8irnudzogXm+PXUUWCoSQ== on 02/16/2019
Background: Community-acquired pneumonia (CAP) causes great mor-
bidity and mortality as well as enormous economic burden worldwide. This
study intended to describe the clinical characteristics of CAP and explore
the risk factors of severe CAP among children in downtown Suzhou, China.
Methods: This was a retrospective study of childhood hospitalizations in Soo-
chow University Affiliated Children’s Hospital from January 1, 2010, to Decem-
ber 31, 2014. Children who were residents of downtown Suzhou, 29 days to
< 5 years of age, with discharge diagnosis codes J09 to J18 and J20 to J22
were included. Medical charts and chest radiograph reports were reviewed for
included children to collect clinical information. CAP with intensive care unit
(ICU) admission and poor clinical outcome were categorized as severe CAP.
Results: A total of 28,043 children were identified with CAP; 17,501
(62.4%) of these children were male, and 20,747 (74.0%) children were
less than 2 years of age. The common clinical symptoms at admission
were cough (94.8%), fever (52.9%), wheezing (37.7%) and respiratory dis-
tress (9.5%). In total, 21,898 (78.1%) children had radiologic evidence of
pneumonia, and 1,403 (5.0%) children developed at least 1 complication.
Multivariate regression analysis showed that younger age, congenital heart
disease and abnormal white blood cells, and C-reactive protein results were
independent risk factors for both ICU admission and poor clinical outcome
(odds ratio [OR] > 1 for all). Respiratory distress symptoms at admission
(OR = 12.10) greatly increased the risk for ICU admission, while ICU
admission (OR = 8.87) and complications (OR = 2.55) increased the risk of
poor outcome. However, cough was a protective factor for ICU admission,
so were wheezing, antibiotic and antiviral therapies for clinical failure.
Conclusion: Pediatric CAP hospitalizations of those of younger age, with
congenital heart diseases, respiratory distress symptoms/tachypnea, abnor-
mal white blood cells and C-reactive protein results as well as complications
were at higher risk for progressing to severe CAP.
Key Words: severe community-acquired pneumonia, risk factors, hospitali-
zation, children
(Pediatr Infect Dis J 2019;38:224–229)
Accepted for publication April 25, 2018.
From the *Department of Epidemiology, School of Public Health, Fudan Univer-
sity; Key Laboratory of Public Health Safety, Ministry of Education; Collabora-
tive Innovation Center of Social Risks Governance in Health, Shanghai, China;
and †Soochow University Affiliated Children’s Hospital, Suzhou, China.
Supported by Fudan University. Funding was provided by Pfizer Inc., Shanghai
Municipal Commission of Health and Family Planning [GWTD2015S05
and 15GWZK0101], Suzhou Municipal Commission of Health and Fam-
ily Planning [LCZX201508 and SZXK201508], the National Key Research
and Development Program of China [2017YFC0211700] and SINO-US
collaborative program on Emerging and Re-emerging Infectious Diseases
[5U2GGH000018]. The founders have no role in study design, data collec-
tion, data analysis, and draft of article.
The authors have no funding or conflicts of interest to disclose.
Address for correspondence: Tao Zhang, PhD, Department of Epidemiology,
School of Public Health, Fudan University, 138 Yi Xue Yuan Road, Shang-
hai 200032, China. E-mail: tzhang@shmu.edu.cn or Jianmei Tian, MD,
Soochow University Affiliated Children’s Hospital, No. 92 Zhong Nan
Street, Suzhou, 215025, China. E-mail: jian_meitian@163.com.
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.
ISSN: 0891-3668/19/3803-0224
DOI: 10.1097/INF.0000000000002098
224 | www.pidj.com The Pediatric Infectious Disease Journal  •  Volume 38, Number 3, March 2019
Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
C
ommunity-acquired pneumonia (CAP) is currently 1 of the
primary causes of morbidity and mortality in children world-
wide, especially in those under 5 years old.1,2 In 2015, the World
Health Organization estimated that CAP accounted for 920,136
deaths in children of all ages and 16% of deaths in children younger
than 5 years of age globally.3 In China, there are 21.1 million new
CAP episodes in children under 5 years old annually, only second
to that of India.4 CAP is also the top cause of pediatric hospitaliza-
tions, associated with enormous disease burden.5–7
With the start of two-child policy in 2016, a “birth boom” is
expected to take place in China, which will likely continue into the
next decade. Therefore, the control and prevention of pediatric dis-
eases such as CAP is of particular importance for public health. There
are many studies focused on the epidemiology, diagnosis and thera-
peutics of CAP in children,8–10 but there are limited data on severe
pediatric CAP patients, a population which requires intensive care unit
(ICU) admission and are more likely to suffer poor clinical outcomes.
It is reported that severe CAP accounts for 6% of ICU admissions,
and that these patients carry a mortality rate of more than 10%.11,12
Although guidelines concerning the management of pediatric CAP
do exist, the treatments for severe pediatric CAP varies.13–15 In this
respect, 1 of the most pressing challenges is to explore independent
factors for severe CAP, and to implement better clinical guidelines.
To better describe the clinical characteristics of CAP and
explore potential risk factors for severe CAP that require ICU
admission or with poor outcome in children, this retrospective
study was designed to collected data from Chinese children under
5 years of age hospitalized with CAP in Soochow University Affili-
ated Children’s Hospital (SCH) from 2010 to 2014. The study find-
ings will help public health practitioners and pediatricians better
recognize high-risk children at risk for severe CAP and provide
evidence for decision making.
METHODS
Study Site
This study was conducted at SCH, the only comprehensive
tertiary children’s hospital serving children in Suzhou, China. The
ward building complex of SCH has a capacity of approximately
1000 beds. There are about 560,000 outpatient visits (including
emergency visits) and 38,000 hospitalizations at SCH annually.
Suzhou is 1 of the important central cities in the Yangtze
River Delta, with over 300,000 inhabitants less than 5 years of age
(in 2014). It consists of 5 county-level cities and 5 municipal dis-
tricts. The 5 municipal districts are recognized as “Suzhou down-
town area” and were selected as our study districts. Previous studies
performed by our group have demonstrated that the admission at
SCH accounts for 67.7% of the total admissions in Suzhou down-
town among children < 5 years of age.16,17
Study Subjects and Data Collection
Pediatric hospitalization records at SCH were retrospec-
tively screened through the electronic hospital information system.
The Pediatric Infectious Disease Journal  •  Volume 38, Number 3, March 2019 Risk Factors for Severe CAP
Patients who fulfilled all the following criteria were considered
as eligible: (1) date of admission between January 1, 2010, and
December 31, 2014; (2) between 29 days to < 5 years of age; (3)
household registered residents of Suzhou downtown area; (4) with
discharge diagnosis ICD-10 codes containing J09-J18 (influenza
and pneumonia) and J20-J22 (other acute lower respiratory infec-
tions). Basic information about demography and hospitalization
could be directly exported from the hospital information system
database; these data included admission number, date of admission,
date of discharge, date of birth, gender, address, discharge diagno-
ses and ICD-10 codes, prognosis and medical insurance.
Review of individual medical charts was performed for these
eligible subjects by trained investigators applying a predesigned
case report form. During this procedure, the detailed informa-
tion, such as previous admission, gestational weeks, birth weight,
congenital heart disease, asthma, clinical symptoms at admission
including fever (axillary temperature ≥ 37.5°C), cough, wheeze,
tachypnea (< 2 m, > 60 breaths/min; 2- < 12 m, > 50 breaths/min;
1- < 5 y, > 40 breaths/min; ≥ 5 y, > 20 breaths/min), dyspnea and
chest indrawing, blood biochemical examination results (including
white blood cell count [WBC] and C-reactive protein [CRP]), chest
radiograph (CXR) results, complications (an unexpected illness,
symptoms or consequence of the current CAP episode), therapeu-
tics (including antibiotics, antivirals, supplemental oxygen and ICU
admission), and outcome were abstracted from the individual medi-
cal charts. The blood biochemical examination and CXRs within 72
hours before or after admission were assessed, and WBC < 5 or >
12 × 109/L, CRP > 8 mg/L, CXRs with consolidation, alveolar infil-
trates or pleural fluid were defined as abnormal results.
After the medical chart review, patients without previous
admissions for the same ICD-10 code within the last 30 days were
defined as CAP and enrolled into the final analysis. In this study,
cured and improved prognoses were considered as fine clinical
outcome while uncured and deceased prognoses as poor clinical
outcome. CAP patients with treatments in the ICU during hospi-
talization or poor clinical outcome at discharge were recognized
as severe CAP.
Statistical Analysis
Continuous variables are described as the mean with stand-
ard deviation or as the median with interquartile range, categorical
variables as numbers and percentages. χ2 test was used to compare
the demographic profile and clinical characteristics. χ2 trend test
was used to verify the age trend of these features. Univariate and
multivariate analyses were conducted to explore potential risk fac-
tors for ICU admission and poor clinical outcome among children
hospitalized with CAP. Gender, age, underlying conditions (includ-
ing congenital heart disease, asthma and prematurity), symptoms at
admission, blood biochemical examination and CXR results were
included into the logistic regression model to explore risk factors
for requiring ICU care during hospitalization. Variables above plus
complications and therapeutics were taken as independent variables
into the regression model to explore risk factors for poor outcome at
discharge. However, there existed some multicollinearity between
cough and antibiotics, and oxygen treatment and ICU admission,
so we excluded cough and oxygen treatment in the final regression
model. All statistical tests were 2-tailed, with a significance level of
0.05. All statistical analyses were conducted with SPSS software
(version 22.0, IBM, Armonk, NY).
Ethics Statement
As a retrospective study reviewing medical records, there
was no personal identifier and no patient contact involved, and an
informed consent was exempted. The study was approved by the
© 2018 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Institute Review Board at the School of Public Health, Fudan Uni-
versity and the Institute Review Board at SCH.
RESULTS
Study Subjects
A total of 185,750 records of pediatric patients at SCH from
January 1, 2010, to December 31, 2014, were screened. Based on
the exclusion criteria including diagnosis, residence and previous
admission within 30 days, 28,043 (15.1%) children were catego-
rized as cases of CAP and enrolled into this study.
Of the 28,043 children with CAP, 17,501 (62.4%) were
male, 20,747 (74.0%) children were < 24 months of age and
14,887 (53.1%) children were < 12 months of age. Age distribu-
tion between male and female children was significantly different
(χ2 = 242.6; P < 0.001). Overall, 13,160 children (46.9%) had med-
ical insurance, and female children had greater coverage (χ2 = 14.1;
P < 0.001). Prematurity, congenital heart diseases, low birth weight
and asthma were observed in 1,718 (6.3%), 1,668 (5.9%), 1,441
(5.5%) and 658 (2.3%) children, respectively. There were no sig-
nificant differences in prematurity, low birth weight and asthma
between male and female children while congenital heart disease
was more frequent in female children (χ2 =8.8; P = 0.003). In total,
671 (2.4%) children experienced ICU admission; male children
were more likely to be transferred to ICU than female children
(χ2 = 5.2; P = 0.023; Table 1).
Clinical Characteristics
Among the 28,043 CAP cases, the most frequent clinical
symptom at admission was cough, occurring in 26,427 (94.8%)
children, followed by fever (52.9%), wheezing (37.7%) and res-
piratory distress symptoms (9.5%) including tachypnea, dyspnea
and chest indrawing. Cough occurred in over 94.0% of children in
every age group. The incidence rate of fever significantly increased
while that of respiratory distress symptoms decreased continuously
with increasing age groups (Ptrend < 0.001 for both). Also, the per-
centage of wheezing tended to decrease with age groups, especially
in children ≥ 6 months of age (χ2trend = 518.7; P < 0.001; Table 2).
Overall, 21,898 (78.1%) children had abnormal CXR find-
ings (radiologic evidence of pneumonia), 9,387 (34.3%) children
were examined with abnormal WBC and 6,895 (25.3%) children
with abnormal CRP level. There was a tendency for abnormal
CXR findings to be more frequently observed in younger children
while abnormal CRP level in older children (Ptrend < 0.001 for all;
Table 2).
When considering all children enrolled, 1,403 (5.0%) devel-
oped at least 1 complication; convulsion or shock was the most
common complication (n = 530 [1.9%]). The incidence of compli-
cation was higher among children ≥ 12 months old when compared
with younger children. In addition, 1.0% of children developed
heart failure, 0.8% respiratory failure, 0.7% meningitis or encepha-
litis and septicemia, 0.4% other pulmonary diseases (including pul-
monary edema, empyema, atelectasis and emphysema) and 0.3%
hydrothorax. Except for septicemia and other pulmonary diseases,
the incidence of other complications was significantly different
across age groups. It was observed that younger children were
more likely to develop respiratory failure and heart failure while
hydrothorax and convulsion or shock were more frequent in older
children (Ptrend < 0.001 for all; Table 2).
Antibiotics were prescribed to 26,673 children (95.3%),
antivirals to 17,935 children (64.1%) and supplemental oxygen
was given to 1,861 children (6.7%). A total of 671 children (2.4%)
needed to be cared for in the ICU, and the majority were very young
children: 403 patients < 6 months old (60.1%) and 524 patients <
www.pidj.com | 225
Shan et al The Pediatric Infectious Disease Journal  •  Volume 38, Number 3, March 2019

TABLE 1.  Population Profile of Children Hospitalized with CAP [n (%)]

Total Male Female

Characteristics (N = 28043) (N = 17501) (N = 10542)

n % n % n % χ2 P
Age
 29 d to < 6 mo 8922 31.8 5758 32.9 3164 30.0 242.6 < 0.001
 6 mo to < 12 mo 5965 21.3 4040 23.1 1925 18.3
 12 mo to < 24 mo 5860 20.9 3662 20.9 2198 20.8
 24 mo to < 60 mo 7296 26.0 4041 23.1 3255 30.9
Medical insurance
 Yes 13,160 46.9 8061 46.1 5099 48.4 14.1 < 0.001
 No 14,883 53.1 9440 53.9 5443 51.6
Underlying condition*
 Prematurity 1718 6.3 1107 6.5 611 6.0 3.3 0.069
 CHD 1668 5.9 984 5.6 684 6.5 8.8 0.003
 LBW 1441 5.5 871 5.3 570 5.8 2.2 0.134
 Asthma 658 2.3 414 2.4 244 2.3 0.1 0.785
Referral to ICU
 Yes 671 2.4 447 2.6 224 2.1 5.2 0.023
 No 27,981 97.6 17,014 97.4 10,296 97.9
*CHD means “congenital heart disease”; LBW means “low birth weight,” which was defined as birth weight < 2500 g.

TABLE 2.  Clinical Characteristics of Children Hospitalized with CAP by Age Group [n(%)]

29 d to < 6 mo 6 mo to < 12 mo 12 mo to < 24 mo 24 mo to < 60 mo

Characteristics Total* (n = 8922) (n = 5965) (n = 5860) (n = 7296) χ2trend P

Symptoms at admission†
 Cough 26,427 (94.8) 8415 (94.6) 5691 (95.8) 5465 (94.0) 6856 (94.5) 1.832 0.176
 Fever 14,659 (52.9) 1886 (21.5) 3201 (54.3) 3973 (68.6) 5599 (77.4) 5378.9 < 0.001
 Wheezing 10,514 (37.7) 3456 (38.9) 3106 (52.3) 2323 (40.0) 1629 (22.5) 518.7 < 0.001
 Respiratory distress 2649 (9.5) 1291 (14.5) 516 (8.7) 426 (7.3) 416 (5.7) 372.5 < 0.001
Diagnostic tests‡
 Abnormal CXR 21,898 (78.1) 7331 (82.2) 4692 (78.7) 4458 (76.1) 5417 (74.2) 161.9 < 0.001
 Abnormal WBC 9387 (34.3) 2668 (30.5) 2320 (39.8) 1957 (34.3) 2442 (34.4) 14.1 < 0.001
 CRP (> 8  mg/L) 6895 (25.3) 1156 (13.3) 1227 (21.2) 1592 (28.0) 2920 (41.0) 1631.0 < 0.001
Complications§ 1403 (5.0) 392 (4.4) 218 (3.7) 354 (6.0) 439 (6.0) 37.1 < 0.001
 Convulsion or shock 530 (1.9) 62 (0.7) 86 (1.4) 174 (3.0) 208 (2.9) 132.2 < 0.001
 Heart failure 279 (1.0) 168 (1.9) 46 (0.8) 37 (0.6) 28 (0.4) 93.1 < 0.001
 Respiratory failure 219 (0.8) 123 (1.4) 36 (0.6) 31 (0.5) 29 (0.4) 50.6 < 0.001
 Meningitis or encephalitis 202 (0.7) 43 (0.5) 30 (0.5) 64 (1.1) 65 (0.9) 16.5 < 0.001
 Septicemia 191 (0.7) 69 (0.8) 39 (0.7) 38 (0.6) 45 (0.6) 1.4 0.231
 Other pulmonary diseases 111 (0.4) 34 (0.4) 20 (0.3) 24 (0.4) 33 (0.5) 0.7 0.405
 Hydrothorax 78 (0.3) 9 (0.1) 5 (0.1) 14 (0.2) 50 (0.7) 48.9 < 0.001
Therapeutics¶
 Antibiotics 26,673 (95.3) 8468 (95.1) 5677 (95.3) 5523 (94.5) 7005 (96.3) 6.5 0.011
 Antivirals 17,935 (64.1) 5345 (60.0) 3876 (65.1) 3868 (66.2) 4846 (66.6) 79.2 < 0.001
 Oxygen treatment 1861 (6.7) 1087 (12.2) 300 (5.0) 254 (4.3) 220 (3.0) 551.5 < 0.001
 ICU admission 671 (2.4) 403 (4.5) 121 (2.0) 77 (1.3) 70 (1.0) 232.08 < 0.001
Prognosis
 Cured 1076 (3.8) 300 (3.4) 231 (3.9) 272 (4.6) 273 (3.7) 21.3 < 0.001
 Improved 26,608 (94.9) 8467 (94.9) 5651 (94.7) 5525 (94.3) 6965 (95.5)
 Uncured 338 (1.2) 146 (1.6) 76 (1.3) 60 (1.0) 56 (0.8)
 Deceased 21 (0.1) 9 (0.1) 7 (0.1) 3 (0.1) 2 (0.0)
*The information of certain variables of some cases is missing, so the detailed information about sample sizes is as follows, fever: 27,691 (98.7%), other symptoms at admission:
27,900 (99.5%), WBC: 27,364 (97.6%), CRP: 27,264 (97.2%), antivirals: 27,976 (99.8%), other therapeutics: 27,981 (99.8%), others: 28,043 (100.0%).
†Respiratory distress symptoms including tachypnea, dyspnea and chest indrawing.
‡CXR: chest radiograph; WBC: white blood cell count; CRP: C-reactive protein. Abnormal CXR was defined as the presence of consolidation, alveolar infiltrates and/or pleural
fluid; abnormal WBC means that WBC was < 5 or > 12 × 109/L.
§Other pulmonary diseases including pulmonary edema, empyema, atelectasis and emphysema.
¶Oxygen treatment here included nasal catheter oxygen inhalation, supplemental oxygen through face mask and extracorporeal membrane oxygenation but no oxygen atomization.

12 months old (78.1%). With increasing age, the rate of antiviral
use gradually increased, while oxygen treatment and ICU admis-
sion decreased (Ptrend < 0.001 for all; Table 2).
At discharge, the great majority of the children were
improved or cured (98.7%), while 338 children (1.2%) were still
not cured, and 21 children (0.1%) were deceased. The differences
across age groups were statistically significant (P < 0.001; Table 2).
Exploring Potential Risk Factors
For this study, referral to the ICU was considered as a
measure of severe CAP. Univariate analysis revealed that male
children and younger children were more likely to be referred
to the ICU. Also, children with congenital heart disease, prema-
turity, respiratory distress symptoms at admission and abnor-
mal WBC and CRP results were at higher risk for more severe

226 | www.pidj.com © 2018 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
The Pediatric Infectious Disease Journal  •  Volume 38, Number 3, March 2019 Risk Factors for Severe CAP

illness. Multivariable logistic regression (controlling for other
variables) revealed that children with CAP younger than 12
months of age more frequently developed severe CAP compared
with those ≥ 24 months of age (OR = 4.19). Congenital heart
disease (OR = 3.03), prematurity (OR = 1.99), abnormal WBC
(OR = 1.27) and abnormal CRP level (OR = 2.20), and respira-
tory distress at admission (OR = 12.10) increased the risk of
developing severe CAP. However, children with cough at admis-
sion had lower risk to be cared for in ICU (OR = 0.28). And there
was no relationship between gender, asthma, fever or wheezing at
admission, abnormal CXR findings and the requirement of ICU
admission (Table 3).
We classified the prognoses of those patients who were not
cured or who were deceased at discharge into “poor clinical out-
come”; other prognoses were categorized into “fine clinical out-
come.” We found from multivariate analysis that children under 2
years of age were at higher risk to suffer poor clinical outcome than
others, so were children with congenital heart diseases (OR = 2.59),
abnormal WBC (OR = 1.68), abnormal CRP level (OR = 1.49), any
complications (OR = 2.55) and ICU admission(OR = 8.87). How-
ever, children with wheezing at admission (OR = 0.61), abnormal
CXR findings (OR = 0.75), receiving antibiotics (OR = 0.42) and
antivirals (OR = 0.72) therapeutics were more likely to be more
healthy at discharge (Table 4).

TABLE 3.  Potential Risk Factors for ICU Admission Among Children Hospitalized with CAP

Non-ICU ICU
Admission Admission
COR AOR
Factors n % n % (95% CI) (95% CI) P

Gender (male) 17,014 97.4 447 2.6 1.21 (1.03–1.42) 1.18 (0.99–1.41) 0.069
Age
 24 mo to < 60 mo 7207 99.0 70 1.0 1.00 1.00
 12 mo to < 24 mo 5770 98.7 77 1.3 1.37 (0.99–1.90) 1.41 (1.00–2.00) 0.053
 6 mo to < 12 mo 5833 98.0 121 2.0 2.14 (1.59–2.87) 1.93 (1.38–2.69) < 0.001
 29 d to < 6 mo 8500 95.5 403 4.5 4.88 (3.78–6.30) 4.19 (3.10–5.66) < 0.001
Underlying condition
 Prematurity 1625 95.0 85 5.0 2.41 (1.91–3.04) 1.99 (1.55–2.57) < 0.001
 CHD 1527 91.8 137 8.2 4.33 (3.57–5.26) 3.03 (2.43–3.79) < 0.001
 Asthma 646 98.3 11 1.7 0.69 (0.38–1.26) 1.29 (0.68–2.44) 0.439
Symptoms at admission
 Cough 25,829 97.9 556 2.1 0.26 (0.21–0.32) 0.28 (0.21–0.36) < 0.001
 Fever 14,381 98.2 257 1.8 0.56 (0.47–0.65) 0.74 (0.60–0.91) 0.054
 Wheezing 10,233 97.4 268 2.6 1.10 (0.95–1.30) 1.25 (1.04–1.51) 0.017
 Respiratory distress 2290 86.6 355 13.4 12.33 (10.53–14.44) 12.10 (8.96–16.36) < 0.001
Diagnostic tests
 Abnormal CXR 21,326 97.5 543 2.5 1.19 (0.98–1.45) 1.24 (1.00–1.55) 0.056
 Abnormal WBC 9098 97.0 278 3.0 1.39 (1.19–1.62) 1.27 (1.07–1.51) 0.006
 CRP (> 8  mg/L) 6672 96.9 213 3.1 1.44 (1.22–1.70) 2.20 (1.79–2.70) < 0.001

TABLE 4.  Potential Risk Factors for Poor Clinical Outcome Among Children Hospitalized with CAP

Fine Poor
Outcome Outcome

Factors n % n % COR (95% CI) AOR (95% CI) P

Gender (male) 17,280 98.7 221 1.3 0.96 (0.78–1.19) 0.96 (0.77–1.28) 0.760
Age
 24 mo to < 60 mo 7238 99.2 58 0.8 1.00 1.00
 12 mo to < 24 mo 5797 98.9 63 1.1 1.36 (0.95–1.94) 1.65 (1.08–2.51) 0.021
 6 mo to < 12 mo 5882 98.6 83 1.4 1.76 (1.26–2.47) 1.96 (1.29–2.97) 0.001
 29 d to < 6 mo 8767 98.3 155 1.7 2.21 (1.63–2.99) 1.59 (1.06–2.40) 0.025
Underlying condition
 Prematurity 1694 98.6 24 1.4 1.23 (0.81–1.87) 0.87 (0.54–1.42) 0.579
 CHD 1594 95.6 74 4.4 4.25 (3.30–5.52) 2.59 (1.86–3.62) < 0.001
 Asthma 657 99.8 1 0.2 0.12 (0.02–0.82) 0.27 (0.04–1.96) 0.196
Symptoms at admission
 Fever 14,502 98.9 157 1.1 0.72 (0.59–0.90) 0.75 (0.56–1.01) 0.056
 Wheezing 10,422 99.1 92 0.9 0.58 (0.46–0.74) 0.61 (0.46–0.81) 0.001
 Respiratory distress 2576 97.2 73 2.8 2.55 (1.96–3.31) 1.04 (0.72–1.51) 0.834
Diagnostic tests
 Abnormal CXR 21647 98.9 251 1.1 0.65 (0.52–0.81) 0.75 (0.57–1.00) 0.049
 Abnormal WBC 9225 98.3 162 1.7 1.77 (1.43–2.19) 1.68 (1.31–2.16) < 0.001
 CRP (> 8  mg/L) 6788 98.4 107 1.6 1.44 (1.14–1.81) 1.49 (1.12–1.99) 0.006
Having complications 1315 93.7 88 6.3 6.51 (5.09–8.33) 2.55 (1.80–3.61) < 0.001
Therapeutics
 Antibiotics 26,357 98.8 316 1.2 0.38 (0.27–0.53) 0.42 (0.27–0.65) < 0.001
 Antivirals 17,760 99.0 175 1.0 0.54 (0.44–0.66) 0.72 (0.56–0.92) 0.009
 ICU admission 582 86.7 89 13.3 15.49 (12.02–19.96) 8.87 (6.07–12.98) < 0.001

© 2018 Wolters Kluwer Health, Inc. All rights reserved. www.pidj.com | 227

Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Shan et al The Pediatric Infectious Disease Journal  •  Volume 38, Number 3, March 2019
DISCUSSION
In this retrospective study, we reviewed the medical charts
of 28,043 subjects for pediatric CAP, described the clinical char-
acteristics of CAP and explored the risk factors for severe CAP
in children. We found that the most common clinical symptoms at
admission were cough (94.8%), fever (52.9%), wheezing (37.7%)
and respiratory distress (9.5%). These results are similar to the
results from a retrospective study on severe acute respiratory infec-
tion in children under 5 years of age in China (cough: 94.3%;
wheeze: 28.1%; tachypnea: 5.1%; dyspnea: 2.1%).18 However, our
rates are slightly lower than that seen in a similar analysis in Taiwan
through a retrospective review of medical records (fever: 99.2%;
cough: 96.9%; tachypnea: 22.8%),19 and greatly different from
that observed in Norway via prospective and observational study
(fever ≥ 38.5°C: 80.8%; cough: 89.4%; tachypnea: 75.8%; wheeze:
18.1%).20 The discrepancy may be partly due to the different admis-
sion standards of pediatric CAP patients between China and other
regions.21 In addition, with the changing medical environment in
China, pediatricians were inclined to enroll patients for early diag-
nosis and treatment even though they presented with mild symp-
toms, a phenomenon which is also recommended by many interna-
tional studies.22,23 Meanwhile, a decreasing tendency for respiratory
distress symptoms at admission and abnormal CXR findings was
seen with increasing age. This result was not surprising as younger
children possess less mature respiratory tracts than older children,
and are more susceptible to respiratory infections.24 In addition,
older children with CAP more frequently had fever compared with
younger children. The reason for this observation may be the early
intervention for young children with respiratory or systemic symp-
toms in China. Complications are common among children with
CAP and have a marked impact on the selection of therapies and
clinical prognosis. Complications were observed in 1,403 children
(5.0%), and most of them were seen in children ≥ 12 months old.
While these results are unlike those in other studies,25–27 we also
included some complications that were not focused on in other
studies, such as convulsions or shocks, which were frequently seen
and occurred more frequently in older children.
In addition, 671 children (2.4%) were referred to the ICU,
and the majority were younger than 12 months old. The mortality
rate increased 20–50% among CAP patients requiring ICU care,
and delayed ICU transfers may be associated with increased mor-
tality.11,28 It is important and necessary to conduct an elaborate eval-
uation of risk factors for referral to the ICU among children with
CAP. Multivariate logistic regression revealed that younger ages
(< 12 months), congenital heart disease, prematurity, respiratory
distress symptoms at admission, abnormal WBC and CRP results
were predictive of ICU admission, while cough was an independent
protective factor for ICU admission. This may be because children
with cough were treated more effectively through symptomatic
treatment than those with respiratory distress symptoms, which
were easily to relapse and are difficult to be completely cured.29,30
Also, cough is an automatic protective response to clear respiratory
tracts and remove foreign bodies.31 Thus, cough can be considered
as a clinical symptom of respiratory infection, but also a marker of
being in the process of self-rehabilitation.
Although a great majority of the children with CAP were
cured and improved, it was not negligible that 359 (1.3%) children
were not cured or deceased, and were considered as poor clinical
outcome as well as a criterion of severe CAP. A number of fac-
tors were taken into account in the logistic regression analysis, and
the results demonstrated that younger ages (< 24 months) remained
an independent risk factor of poor outcome, and congenital heart
disease, abnormal WBC and CRP results and complications and
ICU admission were highly associated with increased risk of
228 | www.pidj.com
Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
clinical failure. Children with wheezing at admission, antibiotics
and antivirals treatments seemed to be at lower risk of clinical fail-
ure. These high-risk factors for poor outcome were similar to that
for ICU admission. It is understandable that patients who had poor
clinical outcomes were usually diagnosed with severe illnesses or
complications and experienced ICU treatment. Additionally, anti-
biotic and antiviral treatments seemed prognostic for a better out-
come. CAP is commonly associated with infections with various
pathogens (bacteria, viruses or co-infection), but it is not easy to
identify disease etiology early in the course. In clinical practice,
antibiotic therapy is prescribed for most children with confirmed
CAP diagnosis as early empirical treatment, and antiviral therapy is
also frequently used, especially in children under 5 years of age.32
Therefore, rational antibiotic and antiviral therapies may help pedi-
atric CAP patients effectively avoid deteriorating into severe CAP.
Some limitations to this study should be acknowledged before
drawing further conclusions. First, this retrospective study relied on
existing data from medical records, and it was inevitable that certain
specific variables were absent. Fortunately, most required informa-
tion in this study included routine items in the medical records. Sec-
ond, the interpretation of the results may be contentious, because this
study included pediatric patients not only with pneumonia (J12-J18),
but also influenza (J09-J11) and acute lower respiratory tract infec-
tion (ALRI) (J20-J22), which were usually diagnosed with pneumo-
nia. The heterogeneity of the study population is very low, because
only 377 (1.3%) children of all the subjects were diagnosed with
influenza or ALRI, and 74 children of them were also with pneumo-
nia. Third, the rate of radiologically confirmed pneumonia (78.5%)
was high, as we only had access to the CXR reports rather than
CXRs, and there existed discrepancies between the interpretation
rules that World Health Organization recommended and the inter-
pretations that study hospitals applied. Fourth, the findings that anti-
biotic and antiviral therapies were predictive of fine clinical outcome
may be challenging. It was reported that failure of initial empirical
treatment was an independent risk factor for increased mortality of
CAP,33,34 but we only collected general information about therapeu-
tics and data on detailed therapeutics and microbiology was unavail-
able in this study. Thus, further information needs to be collected, and
prospective studies are needed to analyze the association between the
specific treatments and outcomes in children with CAP.
In conclusion, younger age, congenital heart disease, res-
piratory distress symptoms (tachypnea, dyspnea and chest indraw-
ing) at admission, abnormal WBC and CRP results and complica-
tions were independent risk factors for severe CAP, while antibiotic
and antiviral therapies helped reduce the risk for progressing into
severe CAP requiring ICU admission or suffering poor clinical out-
come. This study provides reliable evidence to guide the clinical
practice for pediatric CAP hospitalizations. Research on the pedi-
atric CAP severity score and the management of pediatric severe
CAP are required.
REFERENCES
1. Haq IJ, Battersby AC, Eastham K, et al. Community acquired pneumonia in
children. BMJ. 2017;356:j686.
2. Rudan I, O’Brien KL, Nair H, et al.; Child Health Epidemiology Reference
Group (CHERG). Epidemiology and etiology of childhood pneumonia
in 2010: estimates of incidence, severe morbidity, mortality, underlying
risk factors and causative pathogens for 192 countries. J Glob Health.
2013;3:010401.
3. World Health Organization. Pneumonia. 2016. Available at: http://www.
who.int/mediacentre/factsheets/fs331/en/. Accessed September 20, 2017.
4. Rudan I, Boschi-Pinto C, Biloglav Z, et al. Epidemiology and etiology of
childhood pneumonia. Bull World Health Organ. 2008;86:408–416.
5. Liu L, Oza S, Hogan D, et al. Global, regional, and national causes of child
mortality in 2000-13, with projections to inform post-2015 priorities: an
updated systematic analysis. Lancet. 2015;385:430–440.
© 2018 Wolters Kluwer Health, Inc. All rights reserved.
The Pediatric Infectious Disease Journal  •  Volume 38, Number 3, March 2019 Risk Factors for Severe CAP
6. Zhang Q, Guo Z, MacDonald NE. Vaccine preventable community-acquired
pneumonia in hospitalized children in Northwest China. Pediatr Infect Dis
J. 2011;30:7–10.
7. Rudan I, Chan KY, Zhang JS, et al. Causes of deaths in children younger
than 5 years in China in 2008. Lancet. 2010;9720:1083–1089.
8. Wang XF, Liu JP, Shen KL, et al. A cross-sectional study of the clinical char-
acteristics of hospitalized children with community-acquired pneumonia in
eight eastern cities in China. BMC Complement Altern Med. 2013;13:367.
9. Michelow IC, Olsen K, Lozano J, et al. Epidemiology and clinical char-
acteristics of community-acquired pneumonia in hospitalized children.
Pediatrics. 2004;113:701–707.
10. Chiang WC, Teoh OH, Chong CY, et al. Epidemiology, clinical characteris-
tics and antimicrobial resistance patterns of community-acquired pneumo-
nia in 1702 hospitalized children in Singapore. Respirology. 2007;12:254–
261.
11. Liapikou A, Rosales-Mayor E, Torres A. The management of severe com-
munity acquired pneumonia in the intensive care unit. Expert Rev Respir
Med. 2014;8:293–303.
12. Koh JWJC, Wong JJ, Sultana R, et al. Risk factors for mortality in chil-
dren with pneumonia admitted to the pediatric intensive care unit. Pediatr
Pulmonol. 2017;52:1076–1084.
13. Harris M, Clark J, Coote N, et al.; British Thoracic Society Standards of
Care Committee. British Thoracic Society guidelines for the management
of community acquired pneumonia in children: update 2011. Thorax.
2011;66:ii1–i23.
14. Bradley JS, Byington CL, Shah SS, et al.; Pediatric Infectious Diseases
Society and the Infectious Diseases Society of America. The management of
community-acquired pneumonia in infants and children older than 3 months
of age: clinical practice guidelines by the Pediatric Infectious Diseases
Society and the Infectious Diseases Society of America. Clin Infect Dis.
2011;53:e25–e76.
15. World Health Organization. Revised WHO classification and treatment of
pneumonia in children at health facilities: evidence summaries. 2014.
16. Zhang T, Zhu Q, Zhang X, et al. The clinical characteristics and direct medi-
cal cost of influenza in hospitalized children: a five-year retrospective study
in Suzhou, China. PLoS One. 2012;7:e44391.
17. Zhang T, Zhu Q, Zhang X, et al. Clinical characteristics and direct medi-
cal cost of respiratory syncytial virus infection in children hospitalized in
Suzhou, China. Pediatr Infect Dis J. 2014;33:337–341.
18. Jiang Y, Zhu Q, Wu J, et al. Viral etiology and clinical features of severe
acute respiratory infection (SARI) in children under 5 years admitted to the
Children’s Hospital in Suzhou, China. Fudan Univ J Med Sci. 2013;40:407–
12 [in Chinese].
© 2018 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
19. Ma YJ, Wang SM, Cho YH, et al.; Taiwan Pediatric Infectious Disease
Alliance. Clinical and epidemiological characteristics in children with com-
munity-acquired mycoplasma pneumonia in Taiwan: a nationwide surveil-
lance. J Microbiol Immunol Infect. 2015;48:632–638.
20. Berg AS, Inchley CS, Fjaerli HO, et al. Clinical features and inflamma-
tory markers in pediatric pneumonia: a prospective study. Eur J Pediatr.
2017;176:629–638.
21. Prina E, Ranzani OT, Torres A. Community-acquired pneumonia. Lancet.
2015;386:1097–1108.
22. Mundy LM, Leet TL, Darst K, et al. Early mobilization of patients hospital-
ized with community-acquired pneumonia. Chest. 2003;124:883–889.
23. Stolbrink M, McGowan L, Saman H, et al. The early mobility bundle: a sim-
ple enhancement of therapy which may reduce incidence of hospital-acquired
pneumonia and length of hospital stay. J Hosp Infect. 2014;88:34–39.
24. Crawford D. Understanding childhood asthma and the development of the
respiratory tract. Nurs Child Young People. 2011;23:25–34.
25. Madhi SA, De Wals P, Grijalva CG, et al. The burden of childhood pneumo-
nia in the developed world: a review of the literature. Pediatr Infect Dis J.
2013;32:e119–e127.
26. Wang LJ, Mu SC, Lin CH, et al. Fatal community-acquired pneumonia: 18
years in a medical center. Pediatr Neonatol. 2013;54:22–27.
27. Welte T. Managing CAP patients at risk of clinical failure. Respir Med.
2015;109:157–169.
28. Vigg A. Severe Community Acquired Pneumonia (SCAP). Apollo Medicine.
2016;13:17–19.
29. Ribeiro CF, Ferrari GF, Fioretto JR. Antibiotic treatment schemes for very
severe community-acquired pneumonia in children: a randomized clinical
study. Rev Panam Salud Publica. 2011;29:444–450.
30. Clark JE, Hammal D, Spencer D, et al. Children with pneumonia: how do
they present and how are they managed? Arch Dis Child. 2007;92:394–398.
31. Davenport PW. Clinical cough I: the urge-to-cough: a respiratory sensation.
Handb Exp Pharmacol. 2009:263–276.
32. Principi N, Esposito S. Management of severe community-acquired
pneumonia of children in developing and developed countries. Thorax.
2011;66:815–822.
33. Rosón B, Carratalà J, Fernández-Sabé N, et al. Causes and factors associated
with early failure in hospitalized patients with community-acquired pneu-
monia. Arch Intern Med. 2004;164:502–508.
34. Kothe H, Bauer T, Marre R, et al.; Competence Network for Community-
Acquired Pneumonia study group. Outcome of community-acquired pneu-
monia: influence of age, residence status and antimicrobial treatment. Eur
Respir J. 2008;32:139–146.
www.pidj.com | 229