Pediatric Pulmonology 43:567–575 (2008)

Lung Function and Bronchial Responsiveness After

Mycoplasma pneumoniae Infection in Early Childhood
Birgitte B. Kjaer, MD, PhD,1* Jørgen S. Jensen, MD, DMSc,2 Kim G. Nielsen, MD, DMSc,
3

Anders Fomsgaard, MD, DMSc,4 Blenda Böttiger, MD, PhD,4

Birthe Dohn,2 and Hans Bisgaard, MD, DMSc5
Summary. Mycoplasma (M.) pneumoniae has been associated with exacerbation of symptoms in
asthmatic school children and adults; and an etiological role in asthma has been suggested. The
purpose of this study was to investigate whether infection with M. pneumoniae in early childhood
has a long-term influence on lung function and bronchial responsiveness. In a retrospective, clinical
cohort-study children younger than 5 years-of-age when PCR-tested for M. pneumoniae were
enrolled. Sixty-five children with clinical symptoms suggesting infection with M. pneumoniae during
an epidemic season completed a clinical follow-up examination including lung function testing
(28 PCR-positive and 37 PCR-negative). In addition to the PCR-test for M. pneumoniae all
respiratory tract specimens were additionally tested for other atypical bacteria and for viruses by
PCR. Lung function was measured as specific airway resistance by whole-body plethysmography
and bronchial hyperresponsiveness was assessed by cold, dry air hyperventilation. Neither
baseline lung function nor bronchial response to cold dry air hyperventilation differed between M.
pneumoniae-positive and -negative children: mean baseline lung function were 1.17 versus
1.21 (kPa sec), P ¼ 0.45; and mean change in specific resistance was 13% versus 9%, P ¼ 0.42. In
conclusion, M. pneumoniae infection in early childhood was not associated with long-term effects
on lung function and bronchial hyperresponsiveness 2 years after infection. Pediatr Pulmonol.
2008; 43:567–575. ß 2008 Wiley-Liss, Inc.

Key words: asthma, pathogenesis; plethysmography, whole body; respiratory tract

infections, bacteria and virus; polymerase chain reaction; cohort study;
bronchial hyperreactivity.

1
INTRODUCTION Department of Paediatrics, Copenhagen University Hospital, Glostrup,
Denmark.
Mycoplasma pneumoniae infections occur endemically
2
with epidemic outbreaks every 4–7 years. In school-age Department of Bacteriology, Mycology and Parasitology, Statens Serum
children it is a major cause of community-acquired Institut, Copenhagen, Denmark.
pneumonia, but in preschool children the bacterium is 3
Department of Paediatrics Pulmonary Service, 5003, Copenhagen
usually considered less important due to a low frequency University Hospital, Rigshospitalet, Copenhagen, Denmark.
of severe disease.1 However, evidence has begun to
accrue indicating that this view may be too simplified.2
4
Department of Virology, Statens Serum Institut, Copenhagen, Denmark.
Hospitalization due to community acquired pneumonia is 5
Danish Pediatric Asthma Center, Copenhagen University Hospital,
relatively frequent also in children under 5 years of age;3,4 Gentofte, Denmark.
and late complications have been demonstrated by histo-
pathological changes in high-resolution CT scans5 or Grant sponsor: Foundation for ‘‘Kong Christian IX og Dronning Louises
reduced pulmonary diffusion capacity.6 Jubilaeumslegat’’.
M. pneumoniae may precipitate wheezing in asthmatic
*Correspondence to: Birgitte B. Kjaer, MD, PhD, Department of
children and adults2,7–10 and a possible causal relation has Paediatrics, Glostrup Hospital, Nordre Ringvej, 2600 Glostrup, Denmark.
been suggested.10,11 E-mail: bbk@dadlnet.dk
We aimed to investigate a possible long-term effect of
M. pneumoniae infections on lung function and bronchial Received 16 April 2007; Revised 6 February 2008; Accepted 8 February
responsiveness in early childhood. In this study, we 2008.
assessed specific airway resistance (sRaw) by whole body DOI 10.1002/ppul.20813
plethysmography, and bronchial responsiveness to hyper- Published online in Wiley InterScience
ventilation with cold, dry air in 3- to 7-year-old children (www.interscience.wiley.com).

ß 2008 Wiley-Liss, Inc.

568 Kjaer et al.

previously PCR-tested for this infection during an and controls was less than 6 months, and controls were
epidemic. matched to be included with less than 2 months separation
in time whenever possible. It was not possible to match
MATERIALS AND METHODS patients according to clinical setting (hospitals or general
practitioners) to any reasonable degree. The matching in a
The study was approved by the local ethics committee
1–2 ratio was chosen in order to ensure a sufficient number
of Greater Copenhagen [(KF) 01-090/00].
of participants in the Mpn-Neg group, since a lower
acceptance rate to invitation was expected in this group.
Study Design
The children were scheduled for clinical examination
The study was a retrospective, clinical cohort study according to age starting with the oldest. A 4-week period
established by selecting patients having specimens free of any symptoms of respiratory tract infection prior to
submitted to a central laboratory in Copenhagen for the examination was required. The study design has been
diagnostic PCR-testing for M. pneumoniae during the illustrated in Figure 1.
epidemic winter season of 1998 and 1999. At the time
of the epidemic in 1998, the Mycoplasma laboratory at Clinical Examination
Statens Serum Institut was the only place in Denmark
(a) Medical history and physical examination focusing on
performing diagnostic testing for M. pneumoniae by PCR
atopic predisposition; past or present atopy-related
on a routine basis. Thus, when the epidemic was publicly
symptoms and infectious upper or lower respiratory
announced, the laboratory received a very large number of
symptoms.
specimens for diagnosis of M. pneumoniae by PCR. The
(b) Lung function was determined as specific airway
specimens were submitted from all types of health care
resistance (sRaw). Methods and equipment for sRaw
facilities around the country, but the majority came from
have previously been described in detail.12 sRaw was
GP’s. The PCR-testing was requested by the doctor
measured as the relationship between simultaneous
submitting the specimen. Practically all specimens from
variations of respiratory flow and variations of pressure
children sampled by GP’s were throat swabs. The smaller
in a constant-volume whole body plethysmograph, and
number of specimens submitted from hospitals were either
sRaw was calculated at the maximal changes in
throat-swabs or aspirates from the lower respiratory tract.
plethysmographic pressure during inspiration and
The specimens were tested within a few days after arrival
expiration.12 Cold air challenge was performed as a
in the laboratory, and results were reported back to the
single-step 4-min isocapnic hyperventilation test using
doctors when the test was finalized.
158C dry air mixed with 5% CO2 as described
After the acceptance of the study by the ethical
previously.13 The mean value of duplicate measure-
committee, the families of the study participants were
ments (a) before challenge was used as baseline
invited based on addresses obtained through the Central
and (b) 3–5 min after challenge was used as measure-
Office of Civil Registration.
ment of response. An increase of 20% or more
Clearly, this design has the disadvantage of loss of
in sRaw was considered indicative of bronchial hyper-
information about the acute disease phase. As a rough
responsiveness.13
measure of disease severity or disease burden, we obtained
(c) Skin prick testing was performed using the Soluprick
data from the Danish National Patients’ Registry to
test kit1 (ALK-Abello A/S, Hørsholm, Denmark) for
control for major differences between the two groups. This
allergens of the standard inhalation panel: birch, grass,
gave records of hospital admissions concerning infectious
mugwort, dog, cat, horse, two house dust mites, and two
respiratory symptoms, asthmatic symptoms, or atopy-
molds. Positive and negative controls were included.
related ICD-10 codes registered during the period from
The test was carried out only in children cooperating
infancy until November 2002 when data were collected,
with the lung function test.
giving a mean observation period of 46 months after PCR
testing for M. pneumoniae. Only finalized hospital
Microbiological Testing
contacts are registered; thus ambulatory contacts that
were still on-going by the end of the observation period are The respiratory tract specimens, predominantly throat
not included. swabs, were sent to Statens Serum Institut for routine M.
Children under the age of 5 at the time of M. pneumoniae PCR and were tested upon arrival using an
pneumoniae PCR with an address in the Greater inhibition controlled14 M. pneumoniae specific PCR
Copenhagen area were eligible for a clinical follow-up detecting the P1 adhesion gene.15 All suspected positive
examination. All eligible patients with a positive M. test results required confirmation with an independent
pneumoniae PCR (Mpn-Pos) were matched in a ratio of PCR test using primers detecting a different part of the P1
1–2 according to age and sex with children tested PCR- gene in order to out-rule PCR contamination. The
negative (Mpn-Neg). The age difference between cases specificity of the M. pneumoniae PCR is greater than
Pediatric Pulmonology

Fig. 1. Study design: ageneral practitioners, bMycoplasma pneumoniae, cpolymerase chain
reaction, dGreater Copenhagen area, eM. pneumoniae PCR positive, fM. pneumoniae PCR
negative, gsupplementing tests on original specimens for relevant bacteria and viruses (see text).
99%, and the sensitivity is calculated to be 98.4%
when using a combination of cultural and serological
criteria.16
In addition to this, the specimens were tested in a
panel of PCRs for bacteria and viruses that were
considered clinically relevant as differential diagnoses:
Chlamydophila pneumoniae,17 Bordetella pertussis18,19
and B. parapertussis,20 Ureaplasma urealyticum and U.
parvum, RS virus,21,22 influenza A and B viruses,21,22
parainfluenza-1, -2 and -3 viruses,21,22 and adenovirus.23
For sample preparation for bacterial PCRs, DNA from
clinical material is released in Chelex 100 slurry (Biorad,
Hercules, CA) as described in detail previously.24 All
primary bacterial PCRs included an inhibition control14
and all positive results were confirmed by secondary,
confirmatory tests; either by an unrelated primer-pair
(PCRs for C. pneumoniae, B. pertussis, and Ureaplasma)
or by re-testing with the same primer-pair on material
from a new sample preparation of the original material to
out-rule contamination (B. parapertussis). All bacterial
PCRs were carried out as hot-start PCRs. The Ureaplasma
PCR is a TaqMan real-time PCR using primers detecting
the urease gene and probes distinguishing between the
two species Ureaplasma urealyticum and U. parvum
(previously biovars). All bacterial PCRs except for
the Ureaplasma PCR are accredited according to ISO
17025 as routine diagnostic tests.
The viral PCRs are used for routine clinical testing in
the Danish WHO Influenza virus reference laboratory
Lung Function After M. pneumoniae 569
according to ISO17025. Adenovirus DNA was identified
using a nested PCR23 with gel-electrophoresis detection.
All PCRs for RNA-viruses were performed as real-time
reverse transcriptase PCR (RT/PCR) specific for each of
the viruses listed above. Single-step RT/PCR was
performed for 1–2 virus per capillary tube using a
Light-Cycler instrument (Roche, Hvidovre, Denmark).
Sample RNA was extracted by QIAamp Viral RNA Mini
Kit (Qiagen, Ballerup, Denmark) and the RT/PCR was
performed using Master Hybridisation probes mix
(Roche) according to the manufacturer’s instructions.
Specificity and sensitivity was tested using the respective
cultured virus, clinical samples and international blinded
quality control panels.
Data From the Danish National Patients’ Registry
The Danish National Patients’ Registry includes data
from all hospitalizations and completed outpatient
registrations in Denmark based on ICD10-codes and
unique personal registration numbers.
Registrations concerning infectious respiratory or
atopy-related ICD10-codes were obtained for all children
of parents consenting to the study. These data were
obtained as a marker for differences in morbidity in the
Mpn-Pos versus the Mpn-Neg group, and thus to
help validate data. The registry data were obtained in the
fall of year 2002, well after completing the clinical
examinations.
Pediatric Pulmonology
570 Kjaer et al.
TABLE 1— Characteristics of Children With a Full Lung Function Test (LF), Numerical Data

Mpn-Pos Mpn-Neg

n Median Min–max n Median Min-max P-value Theta CIlow CIhigh

Age at PCR (months) 28 41 7–58 37 47 9–58 0.14 5.0 1.6 13.2

Age at LF (months) 28 71 37–85 37 74 47–84 0.20 3.2 1.9 10.0
PCR-LF interval (months)1 28 28 24–43 37 27 24–41 0.11 1.0 1.9 0.1
Weight (kg) 28 21 16–43 37 23 15–34 0.04 2.0 0.0 4.5
Height (cm) 28 117 92–135 37 119 100–134 0.11 4.0 1.0 9.0
Antibiotics, age-adjusted2 25 0.7 0–4.3 32 0.6 0–5.4 0.93 0.0 0.3 0.4

P-value, for the Mann–Whitney statistic; Theta, Hodges–Lehmann estimate for the difference between Mpn-Pos and Mpn-Neg; CIlow, exact lower
limit of the 95% CI for Theta; CIhigh, exact upper limit for the 95% CI for Theta.
1
Time interval between M. pneumoniae PCR and lung function test.
2
Number of antibiotic treatments (estimated by parents) per year of life.

Statistical Methods (43 Mpn-Pos and 56 Mpn-Neg) gave written informed

consent to the study protocol. Eighty-two showed up for
The SAS 8.2 for Windows25 (SAS Institute, Cary NC)
the clinical examination. The reasons for the missing
and Proc-StatXact 6 for SAS users26 (Cytel Software
examination of 17 children were: 8 did not show up as
Corporation, Cambridge MA) were used for the statistical
booked; 4 had frequent respiratory tract infections; and
analyses.
5 were cancelled for various reasons other than recurrent
Numerical data were analyzed by exact Mann–
respiratory tract problems.
Whitney tests supplemented with Hodges–Lehmann
Sixty-five children (79%) completed the lung function
estimates27 (with exact 95% confidence limits),26
test program; baseline values were achieved in further
Tables 1 and 3. The Hodges–Lehmann estimate of the
5 children, leaving 12 children without lung function
difference between the Mpn-Pos and the Mpn-Neg groups
measurements, but with a full medical history.
is the median of all possible differences between a value
Initially, a matched statistical analysis was intended;
from the Mpn-Pos and a value from the Mpn-Neg groups
however, due to the relatively small number of children
for each variable.
completing the study, a large proportion of the patients
Categorical/ordinal data were dichotomized and ana-
eventually seen had lost their match. Consequently, it was
lyzed by Fisher’s exact tests, supplemented with estimated
decided to use the statistical methods described in the
odds ratios with exact 95% confidence intervals,26 Table 2.
Materials and Methods Section.

RESULTS Clinical Examination

A total of 249 children were invited for the study The two groups of children (Mpn-Pos and Mpn-Neg)
(84 Mpn-Pos and 165 Mpn-Neg). Parents of 99 children were similar in terms of: sex; height; atopic predisposition

TABLE 2— Characteristics of Children With a Lung Function Test, Categorical Data

Mpn-Pos Mpn-Neg

Yes n % Yes n % P-value OR CI95% ORlow CI95% ORhigh

BHR 6 28 21 9 37 24 1.00 0.9 0.2 3.2
Sex (male) 11 28 39 19 37 51 0.45 0.6 0.2 1.9
Atopic predisp1 19 27 70 22 36 61 0.59 1.5 0.5 5.1
Atopic history2 8 28 29 15 37 41 0.43 0.6 0.2 1.9
Furry pets3 7 28 25 15 37 41 0.29 0.5 0.1 1.6
Tobacco4 10 28 36 8 37 22 0.27 2.0 0.6 7.0
Cough 7 28 25 14 35 40 0.28 0.5 0.1 1.7
Registry data5 8 28 29 7 37 19 0.39 1.7 0.5 6.5

P-value, for Fisher’s exact test; BHR, bronchial hyperresponsiveness; Cough, without concurrent signs of respiratory infections.
1
Atopic predisposition as past or present asthma, allergic rhinitis or eczema in first-degree relatives.
2
Atopic history as past or present eczema or recurrent asthmatic episodes or allergy. 7/84 (8.8%) children have asthma.
3
Present exposure to furry pets at home.
4
Present, daily exposure to tobacco smoke.
5
Registered hospital contacts at the Danish National Registry of Patients for atopy related illness or respiratory tract infections.

Pediatric Pulmonology

TABLE 3— Lung Function Data
Mpn-Pos
n Mean SD
sRaw (kPa sec) 28 1.17 0.28
delta sRaw (%) 28 13.4 24.5
P-value, for the Mann–Whitney statistic; Theta, Hodges–Lehmann estimate for the median difference between Mpn-Pos and Mpn-Neg; CIlow,
exact lower limit of the 95% CI for Theta; CIhigh, exact upper limit for the 95% CI for Theta; sRaw, specific airway resistance as measured by whole
body plethysmography; delta sRaw, change in sRaw before and after cold air challenge.
from first-degree relatives (parents and siblings); history
of atopy-related symptoms (eczema, allergy, recurrent
asthmatic episodes); present, daily exposure to tobacco or
furry pets; history of coughing in between respiratory
infections; past or present treatment with antibiotics
(Tables 1 and 2); and inhaled b2-agonists or steroids, or
ointments for eczema; atopic findings (eczema, positive
skin prick testing); history of lower or upper respiratory
tract infections (such as pneumonia, otitis media,
pharyngitis or recurrent colds) (data not shown).
The two groups were also similar regarding age at PCR-
testing, age at lung function test and cold air challenge,
and the interval between PCR-testing and lung function
testing (Table 1). Overall, 77% of the children were tested
within a time range of 24–29 months after infection (71%
of the Mpn-Pos and 81% of the Mpn-Neg children). All
statistical procedures were performed identically on the
whole population (n ¼ 65) and on the 77% examined
within the 5-month time range with similar findings in the
sub-population and the whole population.
No differences were found between the two groups in
baseline lung function, bronchial responsiveness to cold
air challenge as measured by sRaw or the degree of
responsiveness (Table 3 and Figs. 2 and 3). Also, the same
proportion of children in the two groups exhibited
Fig. 2. Baseline lung function: measured as specific airway
resistance (sRaw, kPa sec) with whole body plethysmography.
The figure shows the distribution of sRaw in the M. pneumoniae
positive and M. pneumoniae negative groups (each  represents
an individual).
Lung Function After M. pneumoniae 571
Mpn-Neg
n Mean SD P-value Theta CIlow CIhigh
37 1.21 0.26 0.45 0.1 0.1 0.2
37 9.0 18.4 0.49 3.0 10.2 5.6
bronchial hyperresponsiveness to cold air challenge
(Table 2).
Microbiological Examination
A total of 90 positive PCR results were found among all
children included. In 67 children at least one microbe was
found and 20 children had more than one positive PCR.
Details for the 65 children with full lung function tests are
given in Table 4. The most prevalent microbes, apart from
M. pneumoniae, were adenoviruses and RS virus. Other
microbes were detected only in very small numbers. These
findings did not allow for any statistical calculations.
Data From the Danish National Patients’ Registry
A total of 27 (13 Mpn-Pos and 14 Mpn-Neg) of the
99 children included in the study had a record at the Danish
National Patients’ Registry within the relevant ICD10
codes. Only six children (five Mpn-Pos and one Mpn-Neg)
had relevant hospital admissions related in time to the
Mpn-PCR test. The Mpn-Neg child, who was not seen, had
RS virus pneumonia with asthmatic bronchitis. The five
Mpn-Pos all had pneumonia; four of these were seen, and
two obtained full lung function measurements, which
showed no bronchial hyperresponsiveness (both had had
Fig. 3. Bronchial responsiveness to hyperventilation with cold,
dry air measured as change in specific airway resistance (DsRaw,
%) before and after a 4-min isocapnic cold air challenge at a
ventilation rate equivalent to body weight (l  min1). The figure
shows the distribution in the M. pneumoniae positive and M.
pneumoniae negative groups (each  represents an individual).
Pediatric Pulmonology
572 Kjaer et al.
TABLE 4— Microbiological Findings by PCR in 65 TABLE 5— Proportion of PCR Specimens From Hospitals
Respiratory Tract Specimens (97% Throat Swabs) Related to
M. pneumoniae PCR Results Mpn-Pos (%) Mpn-Neg (%)

Mpn-PCR result Children eligible, n ¼ 865 7 9

Children invited 7 6
Positive, Negative, Sum, Children included 12 2
n ¼ 28 n ¼ 37 n ¼ 65 Children seen 11 0
Children with full sRaw þ CACh 7 0
C. pneumoniae 0 0 0
B. parapertussis 0 0 0 Information about the health care setting from which the PCR-testing
U. urealyticum 0 0 0 was requested (hospital or GP) was used as an indication of disease
U. parvum 0 0 0 severity and hence to evaluate our findings with regards to participation
Influenza A virus 0 2 2 bias: At invitation an equal amount of specimens derived from hospitals.
Influenza B virus 0 1 1 At inclusion the Mpn-Pos had a higher representation from hospitals,
Parainfluenza 1,2,3 viruses 0 0 0 but this difference diminished at full sRaw and CaCh. This indicates a
RS virus 4 6 10 participation bias toward a higher attendance in the Mpn-Pos group with
Adenovirus 4 7 11 increasing illness and lower in the Mpn-Neg group. This kind of bias
Adenovirus þ B. pertussis 0 1 1 would favor a result that demonstrated a difference in bronchial
Adenovirus þ RS virus 1 1 2 hyperresponsiveness between groups.
Sum 9 18 27

Adenovirus was found either alone or in combination with M.
pneumoniae, B. pertussis, and RS virus.
atelectases from their M. pneumoniae pneumonia and one
also asthmatic bronchitis). Of the 65 children with full
lung function measurements, only one in each group had
recurrent contacts and both included asthmatic symptoms.
All together, the only apparent difference between the
Mpn-Pos and Mpn-Neg groups was the frequency of
pneumonias in the Mpn-Pos group (Table 2).
To estimate potential participation bias, that is, if results
might be influenced by a higher proportion of more severe
disease in the participating Mpn-Neg children, data from
the Danish National Patients’ Registry were used in
combination with laboratory data: Four of the five children
hospitalized with M. pneumoniae pneumonia had
specimens submitted from hospitals and one sample was
from a general practitioner. The Mpn-Neg child with
pneumonia was tested by a general practitioner and later
hospitalized. Hence, information about the health care
setting (specimen from hospital or general practitioner)
was used to estimate participation bias with regards to
disease severity. No support was found for participation
bias as a confounder of our conclusions on bronchial
hyperresponsiveness as judged by the health care setting
(Table 5).
DISCUSSION
This study found no long-term influence on lung
function or bronchial responsiveness from M. pneumoniae
infection in preschool children (ages 0–5 years).
M. pneumoniae is a well known trigger of symptoms in
asthma patients,2 and has also been associated with onset
of wheezing.28 Findings of elevated total serum IgE, IgE
specific to M. pneumoniae and IgE to common allergens
during the infection has associated the infection with the
Pediatric Pulmonology
asthmatic reactions,29 although this might also indicate
boosting in already sensitized individuals.30 Increased
cytokine levels (IL-4 and IL-4/IFN-g ratio) in BAL fluid
have suggested a TH2-like cytokine response favoring IgE
production.31 An etiological role in the pathophysiology
of asthma has also been suggested based on findings of
colonization or chronic infection in adults with chronic
asthma.32,33
Our findings do not support a role of early infection with
M. pneumoniae in the development of recurrent wheeze or
asthma. No differences between the M. pneumoniae
infected and the control group were detected with regard
to baseline lung function or bronchial responsiveness to
cold air (Table 3, Figs. 2 and 3). This clearly suggests that a
previous infection with M. pneumoniae in preschool
children had not influenced lung function 2–3 years later.
We determined the specific airway resistance (sRaw) by
whole-body plethysmography before and after cold air
challenge to evaluate lung function and bronchial
responsiveness. This method has been shown to be a valid
measure of lung function in young children and it has a
high discriminatory power between asthmatic and non-
asthmatic responses.34 In this study full measurements
were achieved in 79% of the participants seen for
examination.
The lung function test was performed several months
after the infection, and our timing can obviously be
questioned. A short-term difference between groups
would have been overlooked because the children were
examined on average 2 years after the infection. On the
other hand, it could be argued that any potential difference
in lung function that has disappeared after 2–3 years
cannot be considered a true long-term effect, which was
the main focus of the present study.
The number of children studied is small, and thus, it
could be argued that the inability to demonstrate a
difference between groups could be due to a type II error.
 Lung Function After M. pneumoniae 573

Indeed, with a 24% prevalence of bronchial hyper-
responsiveness in the MpnNeg group, the prevalence
should have been doubled (48%) in the MpnPos group to
reach statistical significance. However, the fact that the
prevalence of bronchial hyperresponsiveness was 24% in
the MpnNeg group as compared to 21% in the MpnPos
group makes it highly unlikely that an enlargement of the
groups would have enabled us to demonstrate a significant
adverse effect of M. pneumoniae infection in early
childhood in relation to asthma (Fig. 3).
A true adverse effect of a M. pneumoniae infection
could be obscured if the M. pneumoniae negative children
had a higher predisposition to hyperresponsiveness due to
unbalanced participation bias between groups. No such
differences were found, and the prevalence of atopic
predisposition from first-degree relatives was comparable
to recent data from a Danish birth-cohort of 562 children
(51%).35 The participants of our study were recruited
retrospectively among children having a respiratory tract
specimen examined for M. pneumoniae by PCR. Thus, all
the children had been symptomatic. Only 99 consented of
249 invited. This probably reflects a general recruitment
bias to the study but does not affect the comparison
between the Mpn-Pos and -Neg groups. The retrospective
design of the present study did not enable collection of
clinical data on acute disease other than data of hospital
registrations. This might be considered critical in
controlling for differences between groups or as basis
for sub-group analyses. The cohort design of selecting by
infectious agent and not by disease presentation was,
however, essential in addressing the issue of etiology.
We considered the lack of acute disease data justified,
since the design allowed us to address a relatively
narrow age group, which was laboratory tested during an
epidemic season with a broader epidemiological repre-
sentation than during endemic seasons.
Our study design has the disadvantage of missing
clinical data about the acute disease phase. As a rough
measure of disease severity or disease burden, we obtained
data from the Danish National Patients’ Registry to
control for major differences between the two groups.
Only six children (five Mpn. positive and one Mpn.
negative) had admissions related in time to the M.
pneumoniae PCR test.
The design benefits from the broad representation
of children as opposed to a hospital setting, but only
prospective, population based epidemiological cohort
studies can address the issue of etiology in a conclusive
manner.
Based on earlier studies5,6,36 one would have expected
to find an influence of an earlier M. pneumoniae infection.
We found no difference in bronchial hyperresponsiveness
between groups, but more hyperresponsive children than
expected from a previous study.13 However, the popula-
tion in that study was not representative, and the true
prevalence of bronchial hyperresponsiveness in the back-
ground population is not known for this age group. On the
other hand, the proportion of children with asthma (8–9%)

Fig. 4. M. pneumoniae prevalence in symptomatic children under 7 years of age as shown by

the monthly number of M. pneumoniae PCR tests performed over a 9-year period at Statens Serum
Institut. Full line (positive) indicates positive test results and dotted line (negative) indicates
negative test results. For the epidemic winter of 98/99 the values not shown are: 781 negative PCR
tests in November and 1836 in December (truncated Y-axis). Note that the graph represents
specimens tested from all Danish children less than 7 years of age, whereas patients recruited for
the study were those younger than 5 years of age living in the Greater Copenhagen area.

Pediatric Pulmonology
574 Kjaer et al.
and atopic predisposition corresponds well with preva-
lences for both in the Danish population.
Obviously, any effect of other microbiological agents
(e.g., RS virus) causing infection in the M. pneumoniae
negative children could theoretically have interfered with
our results if these microbes had a significant influence on
lung function. However, the baseline lung function in both
study groups was comparable with that observed in other
studies.37 Furthermore, the long follow-up period would
tend to minimize such an effect, since most of the children
would have experienced several viral infections during the
follow-up period, whereas the risk of an M. pneumoniae
infection during the intervening period was much less
likely due to the low incidence of M. pneumoniae
infections after epidemics,38 including this epidemic
(Fig. 4). We are aware of the discussion of chronic
infection or a possible carrier state with M. pneumoniae.
However, when diagnostic tests specific for M. pneumo-
niae have been used, the carrier state has only been
described as a slow clearance rate within months after
symptomatic infections or after verified exposure to the
microbe.39–41 Indeed, we tested samples from two other
cohorts for the presence of M. pneumoniae or C.
pneumoniae and found a very low prevalence of both
pathogens, even when specimen collection was performed
in an epidemic period.42
Analysis for a selection of other relevant microbio-
logical causes of respiratory tract infection in these
children revealed a microbe, mainly viruses, in another
18 of the patients in addition to the 28 that were positive
for M. pneumoniae but no relation was seen between any
microbe or combinations of microbes and lung function.
In conclusion, we found that M. pneumoniae lung
infection in young children of 12 –5 years, is not associated
with long-term influence on lung function and bronchial
hyperresponsiveness 2–3 years after the infection. Further
investigations of M. pneumoniae in asthma etiology
should be undertaken as population cohort studies rather
than case–control studies to allow for a full description of
various disease patterns and potential complications.
ACKNOWLEDGMENTS
Anders Mørup Jensen, head of Biostatistics Unit,
Statens Serum Institut is thanked for statistical assistance;
Gerda Wigh Jensen for technical assistance and handling
of viral PCRs; Merete Scuzzarella for proofreading of
database. Part of this work was funded by a grant from the
Foundation for ‘‘Kong Christian IX og Dronning Louises
Jubilaeumslegat’’.
REFERENCES
1. Cherry JD. Mycoplasma and ureaplasma infections. In: Cherry
JD, Feigin RD, editors. Textbook of pediatric infectious diseases.
W.B. Saunders, Philadelphia; 1998; pp 2259–2286.
Pediatric Pulmonology
2. Waites KB. New concepts of Mycoplasma pneumoniae infections
in children. Pediatr Pulmonol 2003;36:267–278.
3. Principi N, Esposito S. Emerging role of Mycoplasma pneumo-
niae and Chlamydia pneumoniae in paediatric respiratory-tract
infections. Lancet Infect Dis 2001;1:334–344.
4. Sakurai N, Nagayama Y, Honda A, Makuta M, Yamamoto K,
Kojima S. Mycoplasma pneumoniae and other pathogens in the
aetiology of lower respiratory tract infections among Japanese
children. J Infect 1988;16:253–261.
5. Kim CK, Chung CY, Kim JS, Kim WS, Park Y, Koh YY. Late
abnormal findings on high-resolution computed tomography after
Mycoplasma pneumonia. Pediatrics 2000;105:372–378.
6. Marc E, Chaussain M, Moulin F, Iniguez JL, Kalifa G, Raymond
J, Gendrel D. Reduced lung diffusion capacity after Mycoplasma
pneumoniae pneumonia. Pediatr Infect Dis J 2000;19:706–710.
7. Berkovich S, Millian SJ, Snyder RD. The association of viral and
mycoplasma infections with recurrence of wheezing in the
asthmatic child. Ann Allergy 1970;28:43–49.
8. Johnston SL. Is Chlamydia pneumoniae important in asthma? The
first controlled trial of therapy leaves the question unanswered.
Am J Respir Crit Care Med 2001;164:513–514.
9. Seggev JS, Lis I, Siman Tov R, Gutman R, Abu Samara H,
Schey G, Naot Y. Mycoplasma pneumoniae is a frequent cause of
exacerbation of bronchial asthma in adults. Ann Allergy 1986;
57:263–265.
10. Yano T, Ichikawa Y, Komatu S, Arai S, Oizumi K. Association of
Mycoplasma pneumoniae antigen with initial onset of bronchial
asthma. Am J Respir Crit Care Med 1994;149:1348–1353.
11. Johnston SL, Martin RJ. Chlamydophila pneumoniae and
Mycoplasma pneumoniae: a role in asthma pathogenesis? Am J
Respir Crit Care Med 2005;172:1078–1089.
12. Bisgaard H, Klug B. Lung function measurement in awake young
children. Eur Respir J 1995;8:2067–2075.
13. Nielsen KG, Bisgaard H. Lung function response to cold air
challenge in asthmatic and healthy children of 2–5 years of age.
Am J Respir Crit Care Med 2000;161:1805–1809.
14. Jensen JS, Borre MB, Dohn B. Detection of Mycoplasma
genitalium by PCR amplification of the 16S rRNA gene. J Clin
Microbiol 2003;41:261–266.
15. Jensen JS, Sondergard-Andersen J, Uldum SA, Lind K. Detection
of Mycoplasma pneumoniae in simulated clinical samples by poly-
merase chain reaction. Brief report. APMIS 1989;97:1046–1048.
16. Jensen JS. Direct detection of Mycoplasma pneumoniae in
clinical samples. An acute phase diagnostic test. Faculty of
Medicine, University of Copenhagen, Thesis; 1992.
17. Pollard DR, Tyler SD, Ng CW, Rozee KR. A polymerase chain
reaction (PCR) protocol for the specific detection of Chlamydia
spp. Mol Cell Probes 1989;3:383–389.
18. Birkebaek NH, Heron I, Skjodt K. Bordetella pertussis diagnosed
by polymerase chain reaction. APMIS 1994;102:291–294.
19. Glare EM, Paton JC, Premier RR, Lawrence AJ, Nisbet IT.
Analysis of a repetitive DNA sequence from Bordetella pertussis
and its application to the diagnosis of pertussis using the
polymerase chain reaction. J Clin Microbiol 1990;28:1982–1987.
20. van der Zee A, Agterberg C, van Agterveld M, Peeters M, Mooi
FR. Characterization of IS1001, an insertion sequence element of
Bordetella parapertussis. J Bacteriol 1993;175:141–147.
21. Grondahl B, Puppe W, Hoppe A, Kuhne I, Weigl JA, Schmitt HJ.
Rapid identification of nine microorganisms causing acute respira-
tory tract infections by single-tube multiplex reverse transcription-
PCR: feasibility study. J Clin Microbiol 1999;37:1–7.
22. Osiowy C. Direct detection of respiratory syncytial virus,
parainfluenza virus, and adenovirus in clinical respiratory
specimens by a multiplex reverse transcription-PCR assay. J Clin
Microbiol 1998;36:3149–3154.

23. Wadell G, Allard A, Hierholzer JC. Adenoviruses. In: Murray PR,
Baron EJ, Pfaller MA, Tenover FC, Yolken RH, editors. Manual
of clinical microbiology. Washington, DC: American Society for
Microbiology; 1999. pp 970–982.
24. Tarp B, Jensen JS, Ostergaard L, Andersen PL. Search for agents
causing atypical pneumonia in HIV-positive patients by inhibitor-
controlled PCR assays. Eur Respir J 1999;13:175–179.
25. SAS Institute Inc. SAS Language Reference, Version 8. Cary NC:
SAS Institute, Inc., 1999.
26. Cytel Software Corporation. Proc-StatXact 6 for SAS users, user
manual. Cambridge MA: Cytel Software Corporation; 2004.
27. Hollander M, Wolfe DA. Nonparametric statistical methods. New
York: Wiley; 1973.
28. Biscardi S, Lorrot M, Marc E, Moulin F, Boutonnat-Faucher B,
Heilbronner C, Iniguez JL, Chaussain M, Nicaud E, Raymond J,
Gendrel D. Mycoplasma pneumoniae and asthma in children. Clin
Infect Dis 2004;38:1341–1346.
29. Seggev JS, Sedmak GV, Kurup VP. Isotype-specific antibody
responses to acute Mycoplasma pneumoniae infection. Ann
Allergy Asthma Immunol 1996;77:67–73.
30. Nagayama Y, Sakurai N, Kojima S, Funabashi S. Total and
specific IgE responses in the acute and recovery phases
of respiratory infections in children. J Asthma 1987;24:159–
166.
31. Koh YY, Park Y, Lee HJ, Kim CK. Levels of interleukin-2,
interferon-gamma, and interleukin-4 in bronchoalveolar lavage
fluid from patients with Mycoplasma pneumonia: implication of
tendency toward increased immunoglobulin E production.
Pediatrics 2001;107:E39.
32. Kraft M, Cassell GH, Henson JE, Watson H, Williamson J,
Marmion BP, Gaydos CA, Martin RJ. Detection of Mycoplasma
pneumoniae in the airways of adults with chronic asthma. Am J
Respir Crit Care Med 1998;158:998–1001.
33. Martin RJ, Kraft M, Chu HW, Berns EA, Cassell GH. A link
between chronic asthma and chronic infection. J Allergy Clin
Immunol 2001;107:595–601.
Lung Function After M. pneumoniae 575
34. Nielsen KG, Bisgaard H. Discriminative capacity of bronchodi-
lator response measured with three different lung function
techniques in asthmatic and healthy children aged 2 to 5 years.
Am J Respir Crit Care Med 2001;164:554–559.
35. Jøhnke H. Environmental factors and atopic predisposition as
predictors for the development of atopic eczema in childhood.
Faculty of Health Sciences, University of Southern Denmark;
2003.
36. Mok JY, Waugh PR, Simpson H. Mycoplasma pneuminia
infection. A follow-up study of 50 children with respiratory
illness. Arch Dis Child 1979;54:506–511.
37. Klug B, Bisgaard H. Specific airway resistance, interrupter
resistance, and respiratory impedance in healthy children aged 2–
7 years. Pediatr Pulmonol 1998;25:322–331.
38. Lind K, Benzon MW, Jensen JS, Clyde WA, Jr. A seroepidemio-
logical study of Mycoplasma pneumoniae infections in Denmark
over the 50-year period 1946–1995. Eur J Epidemiol 1997;13:
581–586.
39. Dorigo-Zetsma JW, Zaat SA, Vriesema AJ, Dankert J. Demon-
stration by a nested PCR for Mycoplasma pneumoniae that M.
pneumoniae load in the throat is higher in patients hospitalised for
M. pneumoniae infection than in non-hospitalised subjects. J Med
Microbiol 1999;48:1115–1122.
40. Dorigo-Zetsma JW, Wilbrink B, van der NH, Bartelds AI, Heijnen
ML, Dankert J. Results of molecular detection of Mycoplasma
pneumoniae among patients with acute respiratory infection and
in their household contacts reveals children as human reservoirs. J
Infect Dis 2001;183:675–678.
41. Foy HM. Infections caused by Mycoplasma pneumoniae and
possible carrier state in different populations of patients. Clin
Infect Dis 1993;17:S37–S46.
42. Kjaer BB, Kaltoft MS, Storgaard M, Jensen JS. Mycoplasma
pneumoniae is rarely detected by PCR in healthy children.
22nd Annual meeting of the European Society for Pediatric
Infectious Diseases, Finland, Book of Abstracts, www.kenes.
com/espid2004/program/abstracts/213.doc. 2004.
Pediatric Pulmonology