Acta Pædiatrica ISSN 0803–5253

REGULAR ARTICLE

Radiographic follow-up of community-acquired pneumonia in children

Pål Surén (pal.suren@fhi.no)1 , Kirsti Try2 , Jan Eriksson2 , Behzad Khoshnewiszadeh2 , Karl-Olaf Wathne1
1.Children’s Department, Ullevål University Hospital, Oslo, Norway
2.Department of Radiology, Ullevål University Hospital, Oslo, Norway

Correspondence
Pål Surén, Norwegian Institute of Public Health,
P.O. Box 4404 Nydalen, N-0403 Oslo, Norway.
Tel: +47 23 40 81 26 | Fax: +47 23 40 82 52 |
Email: pal.suren@fhi.no
Received
8 May 2007; revised 20 September 2007;
accepted 28 September 2007.
DOI:10.1111/j.1651-2227.2007.00567.x
Abstract
Aim: To evaluate the value of radiographic follow-up of community-acquired pneumonia in children
who are previously healthy.
Methods: Patient records for the years 2003 and 2004 at the Ullevål University Hospital in Oslo were
reviewed, and a total of 245 children were selected for the study. Radiographs were evaluated by two
paediatric radiologists independently.
Results: One hundred and thirty-three patients had control radiographs, of which 106 were normal
and 27 were abnormal. Only three of 27 patients with abnormal findings had further clinical
problems that could be related to the pneumonia. Two of 106 with normal findings had further
clinical problems, despite the normal control radiograph. Of the 112 without radiographic follow-up,
10 had subsequent clinical problems, but most occurred within the first 4 weeks after discharge,
before controls would have been scheduled.
There were five patients who may have benefited from controls. One relapse could theoretically have
been prevented. Four patients were cases for whom the pneumonias were the first manifestations of
chronic lung disease. Such patients may have some benefit from control radiographs, but only in
terms of detecting the chronic disease at an earlier stage, not in altering the clinical course. Such
modest benefits must be weighed against the consequences of providing follow-up to a large
number of healthy children, and making lots of abnormal findings with no clinical significance.
Conclusion: Control radiographs are not very valuable in children who are otherwise healthy.

parents. In summary, the study found no subsequent ill-

INTRODUCTION
Community-acquired pneumonia has an annual incidence
of 34–40 cases per 1000 in children under the age of
five in Europe and North America (1). Children with
pneumonia usually respond to antibiotics treatment within
48–96 h (2) and make a rapid, uneventful recovery (3). Ra-
diographic changes normally persist much longer. In most
cases, those changes will disappear within 4–6 weeks, but
it often takes longer (2). Pneumococcal pneumonias often
require 1–3 months for complete radiographic resolution
(2). Viral pneumonias and pneumonias caused by legionella,
staphylococcus or gram-negative enteric bacteria, may take
several months to clear (2).
In paediatric practice, it is common to provide a follow-up
clinical exam and take a new chest radiograph some weeks
after discharge to check if the patient is recovering. How-
ever, there are no clear guidelines for the follow-up of pneu-
monia in paediatric literature. Very few studies have been
conducted to evaluate the benefit of control radiographs.
The largest and most rigorous study to date was conducted
in Finland by Virkki et al., and published in 2005 (4). The
study enrolled 196 children hospitalized with community-
acquired pneumonia from 1993 to 1995. Clinical exams and
chest radiographs were done 3–7 weeks after discharge.
Thirty percent had abnormal findings on follow-up radio-
graphs (infiltrate, atelectasis, enlarged lymph nodes etc.). In
2003, 8–10 years later, medical record reviews were done
for all the children, and questionnaires were sent to their
ness related to the initial pneumonia in any of the children.
Moreover, there was no association between abnormalities
on follow-up radiographs and future clinical problems. The
conclusion was that radiographic follow-up of pneumonia
in children is not necessary.
Results of three other studies have been published, but
those studies only included 41, 72 and 70 children, respec-
tively (5,6,7). They were all too small to provide any valuable
evidence.
MATERIAL AND METHODS
Approach
The goal of this study was to investigate the value of ra-
diographic follow-up of community-acquired pneumonia in
children who (a) were previously healthy, (b) had clini-
cal and radiographic signs of pneumonia and (c) had a
community-acquired disease. Children with medical condi-
tions predisposing for pneumonia were not included in the
study.
Study location
The study was conducted at the Children’s Department of
the Ullevål University Hospital in Oslo, which is the largest
children’s hospital in Norway. The department is responsible
for all hospital services and most outpatient services for chil-
dren living in Oslo, and is also a tertiary care centre for chil-
dren from southeast Norway. The department has not had

46 
C 2007 The Author(s)/Journal Compilation 
C 2007 Foundation Acta Pædiatrica/Acta Pædiatrica 2008 97, pp. 46–50
Surén et al. Radiographic follow-up of community-acquired pneumonia

clear guidelines for the follow-up of patients with pneumo- 245

nia. Control radiographs have been taken rather randomly,
Included in study
in the sense that each doctor has decided for him- or herself
whether or not to take them. Clinical follow-up exams have
usually not been provided.
133 112
Selection of cases Control radiograph No control
Ullevål University Hospital has computerized patient taken radiograph
records. A search was conducted for children diagnosed with
pneumonia during the years of 2003 and 2004. Records for
all patients with a diagnosis of pneumonia (ICD-10 codes 10 102
J12-J18) were reviewed, and study subjects were selected ac-
Clinical No clinical
cording to the following criteria: problems problems
106 27

Inclusion criteria:
1. Admission from home (i.e. the pneumonia should be
community-acquired and not hospital-acquired).
2. Temperature > 38.0◦ C, and clinical evidence of pneumo-
nia such as cough, fast breathing, flaring of the alae of
the nose, intercostal and subcostal retractions, grunting,
dullness to percussion or crackles on lung auscultation.
3. Previously unknown infiltrate on chest radiograph at ad-
mission or within 24 h after admission. A positive find-
ing requires an infiltrate outside of the perihilar area or
a consolidation or an empyema.
Exclusion criteria:
1. Underlying condition predisposing to pneumonia (im-
munosuppression, heart disease, etc.), or suspicion that
infiltrate was caused by another disease (tuberculosis,
malignancy, etc.). Asthma was not considered an exclu-
sion criterion.
2. Neutropenia (<0.5 × 109 /L)
3. Immunosuppression (cancer treatment, systemic
steroids, HIV infection etc.)
Evaluation of radiographs
Two paediatric radiologists examined the chest radiographs
taken upon admission and decided whether the radiograph
was consistent with pneumonia (i.e. meeting inclusion cri-
terion no. 3). Subsequently, all control radiographs were ex-
amined and described by the same radiologists. They worked
independently of each other. In cases of disagreement be-
tween the two radiologists, a third paediatric radiologist had
the final word.
Follow-up
For each child, the computerized patient record was re-
viewed for a period of one year after discharge from hospital.
The review had two purposes
1. To evaluate if there were any differences in outcomes be-
tween children who had control radiographs and those
who did not.
2. To find out if there were complications that were pre-
vented, or could have been prevented, by taking control
radiographs.
Normal Abnormal
2 104 3 24
Clinical No clinical Clinical No clinical
problems problems problems problems
Figure 1 Distribution of patients.
RESULTS
Throughout 2003 and 2004, 457 patients were given a diag-
nosis of pneumonia. After the medical record review, 41 were
excluded because the diagnosis was wrong, or the record
contained too little information to judge whether the diag-
nosis was correct. One hundred and twenty-three patients
were excluded on clinical grounds. Most of those had con-
ditions predisposing to pneumonia, such as chronic heart
disease, cerebral palsy, immunosuppression due to cancer
treatment and so on. Some were excluded because the
pneumonia was acquired in hospital, and not community-
acquired. Of the 293 remaining children, 245 had posi-
tive radiographs on admission. These 245 were selected
for the study. One hundred and thirty-three of the 245
had control radiographs, of which 106 were normal and
27 abnormal. One hundred and twelve did not have con-
trol radiographs. The distribution of patients is depicted in
Figure 1.
Children receiving radiographic follow-up
Twenty-seven children had abnormal findings on control ra-
diographs. The abnormal findings were mostly residual in-
filtrates. Only three of the 27 had further clinical problems,
as specified in Table 1. The remaining 24 children had no
clinical problems, despite of the abnormal findings. In those
24 children, the findings were residual infiltrates (20 pa-
tients), pleural thickening (two patients who had pneumonia
with empyema), new infiltrate in another location (1 patient)
and sparse atelectasis (1 patient).
One hundred and six children had normal findings on con-
trol radiographs. Two of those still had further clinical prob-
lems, as specified in Table 2.


C 2007 The Author(s)/Journal Compilation 
C 2007 Foundation Acta Pædiatrica/Acta Pædiatrica 2008 97, pp. 46–50 47
Radiographic follow-up of community-acquired pneumonia Surén et al.

Table 1 Clinical problems in patients with abnormal control radiographs

Sex Age Finding on control radiograph Clinical problem Potential benefit of control

Girl 21/2 years Residual infiltrate and Several admissions due to obstructive airways. Early discovery
atelectasis Atelectasis persisted. Clinical course
suggestive of chronic lung disease, but no
diagnosis found.
Girl 10 months Sparse residual infiltrate Control was taken 13 days after discharge. No None. Regular control radiograph 4 weeks after
information available about why the control discharge would have occurred too late to
was taken early. No action until patient was prevent relapse.
readmitted with relapse of pneumonia
23 days after discharge.
Girl 16 months Sparse residual infiltrate New pneumonia at same location 2 months Abnormal finding was not acted upon. Patient
and atelectasis later may have benefited if atelectasis had been
treated.

Table 2 Clinical problems in patients with normal control radiographs

Sex Age Finding on control Clinical problem Potential benefit

radiograph of control

Boy 14 years Normal Patient had asthma. Several pneumonias during the following year, all infiltrates at the None.
same site. HRCT thorax showed pulmonary infiltrates, but no other abnormal findings.
Boy 11/2 years Normal Several pneumonias during the following year. Diagnostic work-up of immune function None.
and test for cystic fibrosis. All test results normal.

Children with no radiographic follow-up
Ten of 112 children without radiographic follow-up had vari-
ous kinds of clinical problems after discharge. The lung prob-
lems in these 10 children are listed in Table 3.
DISCUSSION
Among the 133 children who had control radiographs taken,
there were none who directly benefited from it, but there
were two who could have benefited if proper action had
been taken. The first was a 21/2 -year-old girl who had a
residual infiltrate and an atelectasis on the control radio-
graph. Her clinical course through the following year was
suggestive of chronic lung disease. The diagnostic work-up
for chronic lung disease did not start until several months
after the first hospitalization. However, if the first abnor-
mal finding had been acted upon, diagnostics and treatment
could have started earlier.
The other patient who could have benefited, was a
16-month-old girl with a sparse residual infiltrate and atelec-
tasis. Nothing was done about this finding, but the girl was
readmitted to hospital with a new pneumonia at the same
site 2 months after first discharge. It is possible that she could
have avoided the relapse if the atelectasis had been treated,
but that is of course a matter of speculation.
Two of the children with abnormal findings were children
who had pneumonia with empyema. The abnormality was
pleural thickening, which is to be expected. Although these
patients had no subsequent complications, empyema is a
high-risk complication, and a control radiograph is always
warranted in such cases.
Among the 112 children without radiographic follow-up,
there were 10 who had clinical problems of some kind af-
ter discharge (as depicted in Table 3). It is unlikely that any
complications could have been prevented by a control ra-
diograph. The ones who might have had any benefit were
three children who were later suspected of having chronic
lung disease – again, the disease could have been detected
earlier by taking a control radiograph.
The number of children with subsequent clinical problems
was higher in the group who did not have control radio-
graphs. However, this cannot be interpreted as a higher risk
of unfavourable outcomes in this group. Some children with-
out control radiographs were probably meant to have them,
but relapses occurred within 4 weeks after discharge, be-
fore controls would normally be scheduled. Since we do not
know who were intended to have control radiographs (the
intention to treat), we cannot make reliable statistical analy-
ses comparing outcomes in the two groups. This also points
to the major weakness of this study, which is the retrospec-
tive design. The ideal study design for this purpose would
be a prospective study, in which patients were randomized
to have either control radiographs or no follow-up, and out-
comes in the two groups were compared.
In summary, there were five patients of 245 (2%) for which
radiographic follow-up could have had any benefit. Four
were children who had, or most likely had, chronic lung dis-
ease. The ability to detect chronic lung disease at an earlier
stage appears to be the most prominent benefit of control
radiographs. However, it is unlikely that the clinical course
of chronic lung disease would be significantly altered by de-
tecting it a little earlier. Moreover, if chronic lung disease is

48 
C 2007 The Author(s)/Journal Compilation 
C 2007 Foundation Acta Pædiatrica/Acta Pædiatrica 2008 97, pp. 46–50
Surén et al. Radiographic follow-up of community-acquired pneumonia

Table 3 Clinical problems in patients without radiographic follow-up

Sex Age Clinical problem Potential benefit of control

Girl 11/2 years Readmitted with relapse of pneumonia at same site 33 days after Probably none. Control would have been scheduled
discharge. No symptoms in the meantime. around the time of relapse.
Girl 5 years Readmitted with persistent pulmonary infiltrate and development of None. Readmitted before control would have been
empyema 15 days after discharge. done.
Boy 3 years Had an infiltrate with atelectasis upon admission, thought to be of viral Probably none. Control would have been scheduled
origin and not treated with antibiotics. Did not fully recover after around the time of relapse.
discharge. Readmitted after 27 days because of worsening of
symptoms.
Boy 2 months Frequent admissions for airway infections and obstructive airways during Early discovery.
the following year. Second admission 10 days after first discharge.
Later diagnosed with -mannosidosis, a lysosomal storage disease.
Girl 4 years Readmitted with pneumonia 23 days after discharge, but infiltrate at None. Readmitted before control would have been
another site. No symptoms in the meantime. scheduled.
Boy 3 years Readmitted with relapse of pneumonia 24 days after discharge. Infiltrate None. Readmitted before control would have been
at same site, and also new infiltrates. scheduled.
Boy 10 months Several admissions with airway infections during the following year. Early discovery.
Second admission several months after first discharge. Diagnostic
work-up for potential chronic lung disease – suspected ciliary
dyskinesia, but diagnosis not confirmed.
Girl 13 months Readmitted 35 days after discharge with asthma triggered by viral None.
infection. No pneumonia then, no antibiotics given.
Boy 11/2 years Several admissions during the following year because of recurrent airway Early discovery
infections and chronic cough. Second admission 2 months after first
discharge. Atelectasis and bronchiectasis developed in both lungs,
demonstrated on chest CT. Diagnostic work-up for chronic lung
disease, but no diagnosis found. Eventually spontaneous clinical
improvement. Not re-examined with another chest CT to evaluate
whether atelectasis and bronchiectasis had subsided.
Boy 14 months Readmitted with another pneumonia 2 months after first discharge. None.
Infiltrate at another site than upon first admission.

present, the child will most likely be in touch with a doctor
again quite soon, and a control radiograph would not make
much of a difference in terms of saving time.
Most abnormal findings on control radiographs do not
have any clinical significance. If control radiographs are
taken of all patients with pneumonia, abnormal findings will
be made in a lot of children who are actually healthy, and
those findings will require further follow-up. Such unneces-
sary follow-up is a burden to the patients and their parents.
Another observation from our study was that abnormal
findings on control radiographs were often not acted upon.
If controls are to have any value, hospitals must have good
working routines for the follow-up of abnormal findings.
Virkki et al. (4) included all children with pneumonia,
regardless of any predisposing condition. In this study, we
chose to only include previously healthy children, in order
to be able to make general recommendations on the basis
of our findings. In our view, it only makes sense to make
general recommendations for the follow-up of healthy chil-
dren. For children with underlying conditions predisposing
to pneumonia, follow-up must be decided according to pa-
tients’ individual needs.
In our study sample, most control radiographs were taken
around 4 weeks after discharge. Given the fact that pneu-
monic infiltrates often take more time to resolve, it might
make sense to conduct controls at a later time. On the other
hand, delaying controls would reduce the benefit of detect-
ing complications early.
A critical question is whether this study was able to cap-
ture all complications that may have occurred among the
patients. When patient record review is the only mode of
follow-up, there is a risk that not all complications are reg-
istered. However, given the way child health care is orga-
nized in Oslo, underreporting should not be too much of a
problem. The Children’s Department is the only hospital for
children living in the city, and all advanced diagnostics for
children take place there. Unless the child moves out of the
city shortly after discharge, all significant complications will
be registered and treated at the Children’s Department.
Our conclusion is that radiographic follow-up after
community-acquired pneumonia benefits only a small pro-
portion of children – one in 50 in our study sample – and
the benefit is not great. Control radiographs generate a lot of
abnormal findings without any clinical significance. In our
opinion, it is safe to drop control radiographs for children
who are previously healthy.
ACKNOWLEDGEMENT
This study was part of a project receiving support from
Wyeth, Norway.


C 2007 The Author(s)/Journal Compilation 
C 2007 Foundation Acta Pædiatrica/Acta Pædiatrica 2008 97, pp. 46–50 49
Radiographic follow-up of community-acquired pneumonia
References
1. McIntosh K. Community-acquired pneumonia in children. N
Engl J Med 2002; 346: 429–37.
2. Behrman RE, Kleigman RM, Jenson HB. Nelson textbook of
pediatrics 17th ed. Philadelphia: Saunders, 2004.
3. Juvén T, Mertsola J, Waris M, Leinonen M, Ruuskanen O.
Clinical response to antibiotic therapy for community-
acquired pneumonia. Eur J Pediatr 2004; 163:
140–4.
50 
C 2007 The Author(s)/Journal Compilation 
Surén et al.
4. Virkki R, Juvén T, Mertsola J, Ruuskanen O. Radiographic
follow-up of pneumonia in children. Pediatr Pulmonol 2005;
40: 223–7.
5. Heaton P, Arthur K. The utility of chest radiography in the
follow-up of pneumonia. N Z Med J 1998; 111: 315–7.
6. Gibson NA, Hollman AS, Paton JY. Value of radiological
follow-up of childhood pneumonia. BMJ 1993; 307: 1117.
7. Grossman LK, Wald ER, Nair P, Papiez J. Roentgenographic
Follow-Up of Acute Pneumonia in Children. Pediatrics 1979;
63: 30–1.
C 2007 Foundation Acta Pædiatrica/Acta Pædiatrica 2008 97, pp. 46–50