Journal of Microbiology, Immunology and Infection (2015) 48, 153e159

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.e-jmii.com

ORIGINAL ARTICLE

A population-based analysis of children with

pneumonia among intensive care units in
Taiwan
Chien-Lun Hsu a, Yu-Sheng Lee b, Chun-Jen Chen a,
Ming-Luen Lee a, Chia-Feng Yang b,c, Wen-Jue Soong b,
Mei-Jy Jeng b,c,*, Keh-Gong Wu a,*

a
Division of Infectious Diseases, Department of Pediatrics, Taipei Veterans General Hospital and
National Yang-Ming University, Taipei, Taiwan
b
Division of General Pediatrics, Department of Pediatrics, Taipei Veterans General Hospital and
National Yang-Ming University, Taipei, Taiwan
c
Institute of Emergency and Critical Care Medicine, School of Medicine, National Yang-Ming University,
Taipei, Taiwan

Received 30 April 2013; received in revised form 27 June 2013; accepted 19 July 2013
Available online 21 September 2013

KEYWORDS Background: Pneumonia is a major diagnosis in children that requires intensive care and is a
Children; major cause of mortality in critically ill children. A survey on current epidemiology and case
Intensive care unit; fatalityeassociated conditions is crucial for the care of critically ill children with pneumonia
National Health in an intensive care unit (ICU).
Insurance Research Methods: The sex, age, seasonality of admission, area of distribution, and case fatality rate of
Database; children younger than 18 years who had pneumonia and were admitted to an ICU during the
Pneumonia period 2006e2010 were obtained from the National Health Insurance Research Database
(NHIRD) of Taiwan. The enrolled children were grouped by age (0e2 years, 3e5 years, 6e11
years, and 12e17 years). The need for invasive procedures such as endotracheal tube (ET)
insertion, mechanical ventilation (MV), tracheostomy, central venous catheter (CVC) insertion,
chest tube insertion/drainage, chest surgery, and extracorporeal membranous oxygenation
(ECMO) were analyzed to clarify their association with case fatality in critically ill children with
pneumonia.
Results: Of the 12,577 children enrolled, 7131 (56.7%) were boys and 5446 (43.3%) were girls.
The younger age groups had more cases of pneumonia, but less often required invasive proce-
dures. Children 0e2 years old (n Z 6083) accounted for approximately one-half (48.4%) of all

* Corresponding authors. Department of Pediatrics, Taipei Veterans General Hospital, 201, Section 2, Shih-Pai Road, Taipei 112, Taiwan.
E-mail addresses: [email protected] (M.-J. Jeng), [email protected] (K.-G. Wu).

1684-1182/$36 Copyright ª 2013, Taiwan Society of Microbiology. Published by Elsevier Taiwan LLC. All rights reserved.
http://dx.doi.org/10.1016/j.jmii.2013.07.007
154 C.-L. Hsu et al.

enrolled children. This group had the lowest case fatality rate (3.1%; 187/6083 children) and
lowest need for invasive procedures (31.1%; 1892/6083 children), whereas children in the 12
e17 year-old group had the highest case fatality rate (9.9%; 140/1417 children) and the highest
need for invasive procedures (59.8%; 847/1417 children) (p < 0.001). The percentage of pneu-
monia cases was highest in the spring (30.1%) and lowest in the summer (21.7%). The invasive
procedures associated with case fatality were ET/MV (OR, 14.31; p < 0.001), CVC insertion
(OR, 7.46; p < 0.001), ECMO intervention (OR, 4.59; p < 0.001), and chest tube insertion/
drainage (OR, 1.87; p < 0.001).
Conclusion: The number of cases of pneumonia that required ICU admission was greater among
younger children than among older children. Factors associated with the higher case fatality
rate included older age at presentation, the need for invasive procedures (e.g., ET/MV, CVC
insertion, chest tube insertion/drainage, and ECMO), underlying comorbidities and complica-
tions.
Copyright ª 2013, Taiwan Society of Microbiology. Published by Elsevier Taiwan LLC. All rights
reserved.

Introduction unit (ICU). Thus, a detailed investigation of the post-

pneumococcal vaccination period in critically ill children is
Pneumonia is a common infectious disease worldwide that crucial in Taiwan.
can occur at any age. It causes severe complications and A previous investigation in Taiwan indicated that
has a poor prognosis, and can result in death. In pediatric approximately 20% of hospitalized children diagnosed with
patients, it is also a major infectious disease that results in pneumonia require ICU admission. Despite this interven-
children being diagnosed as critical, which requires inten- tion, some children still have a poor outcome, especially
sive care. The incidence of community-acquired pneumonia children with pneumococcal pneumonia.13 Based on disease
is reportedly 36e40 episodes per 1000 children per year in severity and complications that are present, these children
children younger than 5 years and 11e16 episodes per 1000 may need invasive procedures for therapeutic purposes. In
children per year in children 5e14 years.1 In Europe alone, such circumstances, a detailed survey on current epidemi-
2.5 million cases of pneumonia occur annually.1 The etiol- ology and related factors associated with case fatality may
ogy of community-acquired pneumonia has been analyzed help guide clinicians in the care of critically ill children with
in many previous studies to allow adjustments in the pneumonia.
administration of antibiotics.1e7 Streptococcus pneumonia The purpose of this study was to investigate the epide-
is the leading bacterial etiology of pediatric pneumonia, miology and case fatality-associated conditions in critically
particularly among children younger than 5 years.4,8 This ill children with pneumonia who are admitted to an ICU.
bacterial species accounts for an estimated 17e44% of pe- This study was conducted by using 5 years (2006e2010) of
diatric pneumonia admissions.7,9 data from the nationwide National Health Insurance
Since 2000, a pneumococcal conjugated vaccine has Research Database (NHIRD) of Taiwan.
been used nationwide in the United States to prevent
invasive pneumococcal diseases; this has led to a 39%
decline in the admission rate of children younger than 2 Methods
years old with all-cause pneumonia.9 There are neverthe-
less many other pathogens that can cause pneumonia or Data sources
similar severe infections.10 Hospital-acquired pneumonia is
also a major subset of pediatric pneumonia because it In 1995, a universal compulsory national health insurance was
causes severe complications that require invasive pro- introduced in Taiwan; in 2011, it provided coverage for 99.6%
cedures as interventions. A previously published report also of the population living in Taiwan.14 The National Health
demonstrates that the implementation of routine vaccina- Research Institute (NHI) of Taiwan maintains a large
tion with the conjugated pneumococcal vaccine resulted in computerized administrative database from the National
an 84% reduction of Streptococcus pneumoniae bacteremia Health Insurance program that includes data on complete
and 67% reduction in overall bacteremia in United States.10 outpatient visits, hospital admissions, prescriptions, disease,
Previous studies in 1997e2004 in Taiwan have shown higher and vital status for this population. Therefore, the National
incidences and case fatality rates in hospitalized children Health Insurance Research Database (NHIRD) of Taiwan is one
with pneumonia who were younger than 5 years old.11,12 A of the largest and most comprehensive nationwide
heptavalent pneumococcal conjugated vaccine was intro- population-based databases available in the world.
duced in Taiwan in 2005. The age distribution, disease All information from the NHIRD was double encrypted to
severity, and the case fatality rate after the introduction of exclude any possibility of identifying individual informa-
the vaccine may differ from these factors prior to the use of tion.15 The Institutional Review Board of Taipei Veterans
pneumococcal conjugated vaccine in critically ill children General Hospital approved this study (VGHIRB; No. 2012-06-
with pneumonia who were admitted to an intensive care 006A).
Analysis of children with pneumonia in ICU 155

Study population interventions involving an invasive procedure (p Z 0.324)

or in the case fatality rate (p Z 0.299).
The inpatient data file of the Taiwan NHIRD was searched to When the different age groups were explored, the 0e2
identify pediatric patients younger than 18 years old who year-old group (n Z 6083, 48.4%) had the largest number of
had a recorded ICU stay during 2006e2010 and a diagnosis ICU admissions for pneumonia, followed by the 3e5 year-
of pneumonia, based on the diagnosis codes (480.x, 481, old group (n Z 2906, 23.1%). The smallest number of ICU
482.xx, 483.x, 484.x, 485, 486, and 487.0) of the Interna- admissions for pneumonia was in the 12e17 year-old group
tional Classification of Diseases, 9th Revision, Clinical (n Z 1417; 11.3%; Table 1 and Fig. 1A). In patients requiring
Modification (ICD-9-CM). These data were extracted from invasive procedures, more than one-third of patients
the database and analyzed. (37.4%; 4702/12517 children) received one or more pro-
cedures. The number of pneumonia cases was greater in
younger childrendespecially in the 0e2 year-old group
Data analysis (40.2%; 1982/4702)dthan in the older children (Fig. 1B).
However, the trend of the need for invasive procedures and
The demographic characteristics (e.g., age, sex, season- the case fatality rate was higher for the older age groups,
ality of admission, and area distributions) of the enrolled and highest for the 12e17 year-old group (invasive proce-
population were collected for analysis. The enrolled chil- dure, 59.8%; case fatality rate, 9.9%; p < 0.001; Table 1 and
dren were divided into four age groups for further com- Fig. 1C).
parisons: 0e2 years, 3e5 years, 6e11 years, and 12e17 The distribution of patients across the four geographical
years. Seasonality was divided into spring (March, April, and regions of Taiwan showed that east Taiwan had the fewest
May); summer (June, July, and August); autumn children enrolled. The need for invasive procedures was
(September, October, and November); and winter highest for children from north Taiwan (47.6%) and lowest
(December, January, and February).16 The area distribu- for children from the middle of Taiwan (30.7%; p < 0.001).
tions were divided into north, middle, south, and east of However, there was no significant difference in the case
Taiwan. fatality rate among the different geographical areas.
Invasive procedures such as endotracheal tube (ET) In regard to the seasonality of ICU admissions for pneu-
insertion/mechanical ventilation (MV), tracheostomy, cen- monia, admissions were highest in the spring (30.1%;
tral venous catheter (CVC) insertion, chest tube insertion/ especially in April) and lowest in summer (21.7%; especially
drainage, chest surgery, and extracorporeal membranous in July; Table 1 and Fig. 2). The need for administering
oxygenation (ECMO) intervention were analyzed to clarify invasive procedures was lowest in the spring (34.4%;
their association with the case fatality rate. Other di- p < 0.001). There was no significant difference in the case
agnoses aside from pneumonia were also retrieved and the fatality rates among the different seasons (p Z 0.082).
top ten diagnoses associated with pneumonia were An analysis of the case fatality-associated conditions
analyzed. showed that age was a significant factor. There were
significantly higher ORs for older children, specifically the
6e11 years (OR, 1.99) and 12e17 years groups (OR, 3.46;
Statistical analysis p < 0.001). The examination of the use of invasive pro-
cedures showed that case fatality was significant for chil-
The dataset from the NHIRD was retrieved by using the dren who had undergone ET/MV (OR, 14.31; p < 0.001),
software (Microsoft SQL Server 2008 R2, Redmond, Wash- CVC insertion (OR, 7.46; p < 0.001), ECMO intervention
ington, USA). The Chi-square test was used to compare (OR, 4.59; p < 0.001), and chest tube insertion/drainage
differences in the case numbers and case fatality rates with (OR, 1.87; p < 0.001; Table 2).
respect to age, sex, seasonality of admission, and area Table 3 lists the top ten diagnoses associated with
distribution. Logistic regression analysis was used to iden- pneumonia and the ORs for case fatality among the enrolled
tify case fatality-associated conditions of pediatric pneu- critically ill children. Respiratory failure was the most
monia that required ICU admission. Statistical significance common diagnosis, followed by cerebral palsy and epilepsy.
was set at p < 0.05. The odds ratios (ORs) and their 95% Positively significant diagnoses included respiratory failure,
confidence intervals (CIs) were also calculated to deter- cerebral palsy, epilepsy, septicemia, congenital heart dis-
mine the association factors of case fatality. ease, and gastrointestinal hemorrhage (p < 0.05) Inversely
significant diagnoses included asthma, pleural effusion, and
volume depletion (p < 0.05). Convulsions were also a
Results common diagnosis, but there was no significant influence on
case fatality.
A total of 12,577 critically ill children younger than 18 years
old who were diagnosed with pneumonia and admitted to
an ICU were identified. Pneumonia accounted for 23.7% of Discussion
all records of pediatric patients with an ICU admission.
Table 1 shows the demographic characteristics of the The present study demonstrates that there are a greater
patients. There were 7131 (56.7%) boys and 5446 (43.3%) number of children of younger age among critically ill
girls. There were 522 (4.2%) mortalities, which comprised children with pneumonia who are admitted to an ICU. The
308 (59.0%) boys and 214 (41.0%) girls. There was no sta- number of patients requiring invasive procedure in-
tistical difference between the male and female patients in terventions were also higher in the younger age groups.
156 C.-L. Hsu et al.

Table 1 Demographic characteristics, percentage of patients undergoing invasive procedures, and the case fatality rate
among critically ill children with pneumonia who were admitted to an ICU in Taiwan (in 2006e2010)
Variable No. of patients Invasive procedures, p (c2 test) Case fatality p (c2 test)
Total 12,577 (100) 4702 (37.4) 522 (4.2)
Sex
Male 7131 (56.7) 2639 (37.0) 0.324 308 (4.3) 0.299
Female 5446 (43.3) 2063 (37.9) 214 (3.9)
Age (y)
0e2 6083 (48.4) 1892 (31.1) <0.001* 187 (3.1) <0.001*
3e5 2906 (23.1) 1047 (36.0) 66 (2.3)
6e11 2171 (17.3) 916 (42.2) 129 (5.9)
12e17 1417 (11.3) 847 (59.8) 140 (9.9)
Area of Taiwan
North 4589 (36.5) 2185 (47.6) <0.001* 218 (4.8) 0.082
Middle 3548 (28.2) 1088 (30.7) 138 (3.9)
South 4128 (32.8) 1302 (31.5) 155 (3.8)
East 312 (2.5) 127 (40.7) 11 (3.5)
Season
Spring 3786 (30.1) 1301 (34.4) <0.001* 136 (3.6) 0.231
Summer 2725 (21.7) 1043 (38.3) 121 (4.4)
Autumn 2924 (23.2) 1155 (39.5) 129 (4.4)
Winter 3142 (25.0) 1203 (38.3) 136 (4.3)
Data are presented as n (%).
* Indicates a statistical significance (p < 0.05).
ICU Z intensive care unit.

However, the percentage of invasive procedures and the The use of invasive procedures in critically ill children
case fatality rate in each age group were higher among the with pneumonia in an ICU is much higher in younger chil-
older age groups. The need for invasive procedures was dren than in older children. This is consistent with the total
furthermore associated with a higher ratio of case fatality case numbers of each age group (Table 1). Of the enrolled
in critically ill children. patients, 37.4% (4702/12,577 children) of the patients
The present study focuses on critically ill children who required one or more invasive procedures during their ICU
were diagnosed with pneumonia and admitted to an ICU stay. Furthermore, 2939 children (62.5% of all enrolled
during the period 2006e2010. This is the first report to use children) who were aged 0e5 years required invasive pro-
the nationwide population-based NHI database to investi- cedures. This was much higher than the total number of
gate critically ill pediatric patients admitted to an ICU for pneumonia cases among children aged 6e17 years old (1763
pneumonia. The present study reveals that the incidence of patients; 37.5% of all enrolled cases). However, the ratio of
pneumonia in ICU-admitted children is higher in boys than children requiring invasive procedures and the case fatality
in girls and is higher in younger children (especially children rate were higher in the 6e17 year-old children than in the
who are 0e2 years old) than in older children (Fig. 1). These 0e5 year-old children. These results are compatible with
findings are consistent with a previous report by Lin et al17 previous investigations11,17,21 and imply that, even though
that was performed in a single center in the middle part of fewer older critically ill children are admitted to an ICU for
Taiwan and covered the period of 2002e2005, and reports pneumonia, their disease severity may be worse in com-
by Wu et al,11,12 who used the NHI database that covered parison to the disease severity in younger children.
the period of 1997e2004. However, these three studies The critical pneumonia-related case fatality rate is
investigated all hospitalized children and did not focus on highest in the children aged 12e17 years. In a report of
critically ill patients. A European study examining the Wu et al11 on children hospitalized with pneumonia in the
period of 1981e1982 also reported that the incidence of period of 1997e2004, the population-based mortality
pneumonia was significantly higher in children younger than rate was much higher in children younger than 5 years
5 years.18 Thus, young children are the most vulnerable than in older children. The biphasic pneumonia-related
population for pneumonia. This remains true for critically case fatality rate was also markedly high in children
ill children who are admitted to an ICU. aged 12e17 years.11 Therefore, the high pneumonia-
The most common causes of pneumoniadother than related case fatality rate in 12e17 year-old children
virusesdin children younger than 5 years is bacterial was consistent prior to and after 2005. Clinicians should
pneumonia.1,2,4,6 Streptococcus pneumoniae, Haemophilus pay more attention to older children who are admitted to
influenzae, Streptococcus pyogenes, and Staphylococcus an ICU for pneumonia, especially children with critically
aureus are the primary causes of pneumonia.1,4 In older ill presentations.
children and adolescents, Mycoplasma pneumoniae is the It is generally believed that younger children are more
most frequent cause.5,19,20 likely to get sick and are more vulnerable to more severe
Analysis of children with pneumonia in ICU 157

diseases. Parents and clinicians have been generally

warned about this situation. As a result, parents often bring
their sick children for medical help as soon as possible, and
do so in Taiwan partly because of a convenient and nearly
full-coverage national health insurance program. Further-
more, because of a pediatrician’s concerns, the admission
criteria to a general ward or to an ICU for young children
may not be as stringent as the criteria for older children. By
contrast, older children may be brought to a hospital only
when they are very sick because the parents or the
children themselves think they are stronger than younger
kids. They do not want to interrupt their schooling or other
daily activities. However, the aforementioned are purely
conjecture.
Since 2005 in Taiwan, the heptavalent pneumococcal
conjugated vaccine was administered to young children.
For younger children (0e5 years), it provides protection
from fatal pneumococcal pneumonia and invasive pneu-
mococcal diseases.22 The study period was 2006e2010;
therefore, the influence of the heptavalent pneumococcal
conjugated vaccine may play an important role in the
finding that the severity of the disease among the younger
children was not severe, although the number pneumonia
cases was higher in younger children (0e5 years) than in
older children. This may also explain the findings on the
severity and the case fatality rate in critically ill children
with pneumonia in which the younger children had a
greater number of cases and the older children had a higher
frequency of undergoing invasive procedures and a higher
case fatality rate.
The use of invasive procedures was highest in northern
Taiwan, although the case fatality ratio was not signifi-
cantly different among the different sites (Table 1). Our
Figure 1. Total number of cases, patients requiring invasive hypothesis to explain this issue is that the number of ter-
procedures, and case fatalities in the different age groups of tiary medical centers is highest in northern Taiwan, and
critically ill children with pneumonia who were admitted to an possibly so are the number of pediatric intensivists who are
intensive care unit. (A) Total number of enrolled cases. (B) available to perform invasive procedures. To clarify this
Number of children requiring invasive procedure interventions issue, a further study to investigate the differences in
and number of fatalities. (C) Percentage of patients per age medical care behaviors of different medical centers and/or
group who required invasive procedure intervention and fa- pediatric intensivists may be helpful in the future.
talities (i.e., case fatality rate). The seasonal distribution of pediatric pneumonia ad-
missions to the ICU shows that the highest rate of admis-
sions occurs in spring and the lowest rate occurs in summer.
These findings are compatible with a previous study,
although with less variation.4 The difference among the
four seasons was not very apparent. One possible expla-
nation is that the seasons in Taiwan have fewer variations
because of its subtropical geographic location. The lower
seasonal climate variations may reduce the variation in the
incidence of pneumonia with complications requiring ICU
admission. A further study to evaluate the causal organisms
of critical pneumonia in children may help to elucidate the
seasonal differences in Taiwan.
The odds ratios of the case fatality rate of critically ill
children with pneumonia are significantly higher in the
6e11 year-old group and the 12e17 year-old group
(p < 0.001), and significantly higher in children who receive
any of the following invasive procedures: ET/MV, CVC
insertion, chest tube insertion/chest tube drainage, and
ECMO (p < 0.001). These findings suggest that an older age
Figure 2. Seasonal distribution of intensive care unit and higher degree of disease severity may be associated
admissions of critically ill children with pneumonia. with a higher risk of mortality. This is the most important
158 C.-L. Hsu et al.

Of the underlying comorbidities (Table 3), cerebral

Table 2 Analysis of case fatality-associated conditions in
palsy, epilepsy, and congenital heart disease are the most
critically ill children who were admitted to an ICU with
common underlying diseases and are significantly corre-
pneumonia
lated with high mortality in critically ill children with
Variable Odds ratio 95% CI p pneumonia (p < 0.05). Physicians should focus more
Sex attention on these patients in regard to their pulmonary
Male 1.10 0.92e1.32 0.278 condition. The highest odds ratios of the most common
Female 1.00 e e complications correlated with case fatality were for respi-
Age (y) ratory failure (OR, 7.04; CI, 5.87e8.46) and septicemia (OR,
0e2 1.00 e e 7.07; CI, 5.88e8.50; p < 0.001). These are also often the
3e5 0.73 0.55e0.97 0.032* most common reasons that patients require intensive care.
6e11 1.99 1.58e2.51 <0.001* The other significant complication correlated with case
12e17 3.46 2.75e4.34 <0.001* fatality is gastrointestinal hemorrhage, which is a common
Area of Taiwan complication in critical patients. Pediatric intensive care
North 1.23 0.99e1.53 0.060 specialists should remember the potential risk of mortality
Middle 1.00 e e when these underlying comorbidities and complications are
South 0.96 0.76e1.22 0.759 present in critically ill children with pneumonia.
East 0.90 0.48e1.69 0.749 A limitation of the present study is that it is a retro-
Season spective analysis based on the encrypted NHIRD. Thus,
Spring 0.82 0.65e1.05 0.117 there was no detailed laboratory information on the or-
Summer 1.03 0.80e1.32 0.835 ganism causing the pneumonia. The causal organism may
Autumn 1.02 0.80e1.31 0.874 have influenced the severity of the disease, the complica-
Winter 1.00 e e tion rate, the need for invasive procedures, and the case
Invasive procedure fatality rate. A further multicenter hospital-based investi-
ET/MV 14.31 11.27e18.18 <0.001* gation that collects detailed laboratory data would be
CVC 7.46 6.22e8.95 <0.001* necessary to understand the influence of these factors.
ECMO 4.59 3.29e6.41 <0.001* In conclusion, in children diagnosed with pneumonia
Chest tube 1.87 1.48e2.36 <0.001* that requires an ICU admission, age is the most important
insertion/drainage factor in regard to incidence, requirement for invasive
Chest surgery 1.20 0.49e2.97 0.686 procedures, and survival. Children younger than 3 years had
Tracheostomy 0.64 0.24e1.73 0.378 the highest number of episodes. Children aged 12e17 years
old had the lowest number of episodes, but had the highest
* Indicates a significant difference (p < 0.05). ratio of the need for invasive procedures and highest case
CI Z confidence interval; CVC Z central venous catheter;
fatality rate, compared to other age groups. Factors asso-
ECMO Z extra-corporeal membranous oxygenation;
ciated with the higher case fatality rate include an older
ET Z endotracheal tube; ICU Z intensive care unit;
MV Z mechanical ventilation; OR Z odds ratio. presenting age (especially in children aged 12e17 years
old), undergoing invasive procedures (e.g., ET/MV, CVC
insertion, chest tube insertion/drainage, and ECMO), hav-
ing underlying comorbidities (e.g., cerebral palsy, epilepsy,
finding of the present study: greater attention should be and congenital heart disease), and developing complica-
focused on the clinical condition of older children with tions (e.g., respiratory failure, septicemia, and gastroin-
complications who are admitted to an ICU for pneumonia. testinal hemorrhage).

Table 3 The top 10 associated diagnoses and odds ratios for case fatality among children with pneumonia who were admitted
to an ICU in Taiwan (2006e2010)
Diagnosis ICD-9-CM code Total cases (n) Case fatality (n) OR 95% CI p
Respiratory failure 518.8x 2296 304 7.04 5.87e8.46 <0.001
Cerebral palsy 343.x 2075 91 2.37 1.86e3.01 <0.001
Epilepsy 345.xx 1756 74 1.49 1.16e1.93 0.002
Septicemia 038.xx 1348 248 7.07 5.88e8.50 <0.001
Asthma 493.xx 1107 11 0.17 0.09e0.31 <0.001
Convulsions 780.3x 830 38 0.88 0.63e1.24 0.478
Pleural effusion 511.x 812 32 0.67 0.47e0.97 0.033
Congenital heart disease 745.xx 756 43 1.46 1.05e2.02 0.023
Volume depletion 276.5x 745 10 0.23 0.12e0.42 <0.001
Gastrointestinal hemorrhage 578.9 607 61 2.96 2.23e3.95 <0.001
CI Z confidence interval; ICD-9-CM Z International Classification of Diseases, 9th Revision, Clinical Modification; ICU Z intensive care
unit; OR Z odds ratio.
Analysis of children with pneumonia in ICU 159

Conflicts of interest the heptavalent-conjugated pneumococcal vaccine. Pediatr

Infect Dis J 2006;25:293e300.
11. Wu PS, Chang IS, Tsai FY, Hsieh YC, Shao PL, Chang LY, et al.
The authors declare that they have no financial or nonfi- Epidemiology and impacts of children hospitalized with pneu-
nancial conflicts of interest related to the subject matter or monia from 1997 to 2004 in Taiwan. Pediatr Pulmonol 2009;44:
materials discussed in the manuscript. 162e6.
12. Wu PS, Huang LM, Chang IS, Lu CY, Shao PL, Tsai FY, et al. The
epidemiology of hospitalized children with pneumo-
References coccal/lobar pneumonia and empyema from 1997 to 2004 in
Taiwan. Eur J Pediatr 2010;169:861e6.
1. Juven T, Mertsola J, Waris M, Leinonen M, Meurman O, 13. Huang YC, Ho YH, Hsieh YC, Lin HC, Hwang KP, Chang LY, et al.
Roivainen M, et al. Etiology of community-acquired pneumonia A 6-year retrospective epidemiologic study of pediatric pneu-
in 254 hospitalized children. Pediatr Infect Dis J 2000;19:293e8. mococcal pneumonia in Taiwan. J Formos Med Assoc 2008;107:
2. Drummond P, Clark J, Wheeler J, Galloway A, Freeman R, 945e51.
Cant A. Community acquired pneumonia: a prospective UK 14. Bureau of National Health Insurance. Universal health coverage
study. Arch Dis Child 2000;83:408e12. in Taiwan. http://www.nhi.gov.tw/Resource/webdata/21717_
3. British Thoracic Society Standards of Care Committee. British 1_20120808UniversalHealthCoverage.pdf [accessed 19.04.13].
Thoracic Society guidelines for the management of community 15. National Health Research Institutes. National health insurance
acquired pneumonia in childhood. Thorax 2002;57:i1e24. research database. http://www.nhri.org.tw/nhird/date_01.
4. Michelow IC, Olsen K, Lozano J, Rollins NK, Duffy LB, Zieqler T, html#_edn1; 2003 [accessed 19.04.13].
et al. Epidemiology and clinical characteristics of community- 16. Tourism Bureau, Republic of China (Taiwan). http://eng.
acquired pneumonia in hospitalized children. Pediatrics 2004; taiwan.net.tw/m1.aspx?sNoZ0002005 [accessed 19.04.13].
113:701e7. 17. Lin CJ, Chen PY, Huang FL, Lee T, Chi CS, Lin CY. Radiographic,
5. Sung CC, Chi H, Chiu NC, Huang DT, Weng LC, Wang NY, et al. clinical, and prognostic features of complicated and uncom-
Viral etiology of acute lower respiratory tract infections in plicated community-acquired lobar pneumonia in children. J
hospitalized young children in Northern Taiwan. J Microbiol Microbiol Immunol Infect 2006;39:489e95.
Immunol Infect 2011;44:184e90. 18. Jokinen C, Heiskanen L, Juvonen H, Kallinen S, Karkola K,
6. Chen CJ, Lin PY, Tsai MH, Huang CG, Tsao KC, Wong KS, et al. Korppi M, et al. Incidence of community-acquired pneumonia
Etiology of community-acquired pneumonia in hospitalized in the population of four municipalities in eastern Finland. Am
children in northern Taiwan. Pediatr Infect Dis J 2012;31: J Epidemiol 1993;137:977e88.
e196e201. 19. Murphy TF, Henderson FW, Clyde Jr WA, Collier AM, Denny FW.
7. Heiskanen-Kosma T, Korppi M, Jokinen C, Kurki S, Heiskanen L, Pneumonia: an 11-year study in a pediatric practice. Am J
Juvonen H, et al. Etiology of childhood pneumonia: serologic Epidemiol 1981;113:12e21.
results of a prospective, population-based study. Pediatr Infect 20. Tsolia MN, Psarras S, Bossios A, Audi H, Paldanius M,
Dis J 1998;17:986e91. Gourgiotis D, et al. Etiology of community-acquired pneumonia
8. Tan TQ. Update on pneumococcal infections of the respiratory in hospitalized school-age children: evidence for high preva-
tract. Semin Respir Infect 2002;17:3e9. lence of viral infections. Clin Infect Dis 2004;39:681e6.
9. Grijalva CG, Nuorti JP, Arbogast PG, Martin SW, Edwards KM, 21. Lee GE, Lorch SA, Sheffler-Collins S, Kronman MP, Shah SS.
Griffin MR. Decline in pneumonia admissions after routine National hospitalization trends for pediatric pneumonia and
childhood immunization with pneumococcal conjugate vaccine associated complications. Pediatrics 2010;126:204e13.
in the USA: a time-series analysis. Lancet 2007;369:1179e86. 22. Hsieh YC, Lin PY, Chiu CH, Huang YC, Chang KY, Liao CH, et al.
10. Herz AM, Greenhow TL, Alcantara J, Hansen J, Baxter RP, National survey of invasive pneumococcal diseases in Taiwan
Black SB, et al. Changing epidemiology of outpatient bacter- under partial PCV7 vaccination in 2007: emergence of serotype
emia in 3- to 36-month-old children after the introduction of 19A with high invasive potential. Vaccine 2009;27:5513e8.