Community-acquired pneumonia in adults is commonly caused by viruses, bacteria, Mycoplasma pneumoniae and Chlamydia pneumoniae, but defining the cause of pneumonia in children is difficult because procedures such as bronchoalveolar lavage and lung puncture for culture are invasive and are indicated only in patients with complicated pneumonia.1-3 Specimens for culture are difficult to obtain and results can be misleading. Culture and PCR testing for C. pneumoniae and M. pneumoniae are not generally available and take many days before results are confirmed. Acute and convalescent serum specimens for antibody titer determinations to selected pathogens can be obtained to confirm the etiology, but the results are not available during the acute illness. There are only fragmentary data on serum antibody titers in children in relation to culture, PCR, clinical course and therapy. Two large prospective studies performed by Block et al.4 and Harris et al.5 described the association between Mycoplasma and Chlamydia infections in children with pneumonia but viruses, Streptococcus pneumoniae serology and C. pneumoniae PCR were not evaluated.
Our study comprised previously healthy children who were diagnosed with pneumonia in the emergency clinic and was designed to compare results of culture, PCR and serum antibody titers before and after antibiotic therapy. The data provide additional information about community-acquired pneumonia in children.
MATERIALS AND METHODS
Study design. This study was a prospective, randomized, unblinded trial conducted at the Children's Medical Center of Dallas, TX. Children with community-acquired pneumonia were identified in the emergency clinic from February, 1996, through December, 1997, and written informed consent was obtained from all parents. The protocol was approved by the Institutional Review Board of the University of Texas South-western Medical Center, Dallas, TX. Using a list of randomized therapy assignments, a research pharmacist provided patients with azithromycin suspension, 10 mg/kg (maximum, 500 mg/day) on Day 1, followed by 5 mg/kg (maximum dosage, 250 mg/day) daily for 4 days or either amoxicillin-clavulanate, 40 mg/kg/day in three divided doses for those younger than 5 years, or erythromycin estolate suspension, 40 mg/kg/day (maximum dosage 1500 mg/day) in three divided doses for 10 days in patients older than 5 years.
Inclusion/exclusion criteria. Previously healthy male and female children 6 months to 16 years of age with tachypnea, fever, cough or rales and an abnormal chest roentgenogram consistent with pneumonia were considered to have community-acquired infection and included in the study. Chest radiographs were interpreted initially by the attending physician in the emergency clinic and a staff radiologist. All radiographs were later reviewed by one of two radiologists who were not familiar with the patients' clinical history or results of special studies. Exclusion criteria included hypersensitivity to macrolide or beta-lactam antibiotics, pregnant or lactating females, nosocomial acquired infections, hospitalization, systemic antibiotic within 72 h before enrollment, cefixime or ceftriaxone within the previous 7 days and chronic diseases. Patients were also excluded if they were receiving medications that had potential adverse interactions with erythromycin or azithromycin.
Patient enrollment and evaluation. Patients were evaluated at enrollment and again at 2 to 3 and 10 to 37 days after start of therapy. Initial samples at enrollment included blood culture, nasopharyngeal and pharyngeal swab for C. pneumoniae and M. pneumoniae culture and PCR, nasopharyngeal swab for viral direct fluorescent antibody (DFA) and culture and acute serum antibody titers for these bacteria and S. pneumoniae. A tuberculin intradermal test was performed and the study medication was dispensed. At Day 2 to 3 a telephone call was made to the caregiver to assess symptoms, interventions and adverse reactions. If the caregiver reported that the purified protein derivative site was red or indurated, the patient returned to the emergency room for evaluation. Patients were assessed at Weeks 2 to 5 for symptoms, adverse reactions and outcome. Compliance in taking the study medication was checked by measuring the volume of drug in the bottle at this visit. At the 2- to 5-week follow-up we obtained nasopharyngeal and pharyngeal swabs for C. pneumoniae and M. pneumoniae cultures and PCR and serum for convalescent antibody titers for these bacteria and for S. pneumoniae. A chest roentgenogram was repeated only if a patient had clinical signs of persistent or new infection.
Clinical response was defined as: cure, resolution of all signs and symptoms; improvement, incomplete resolution of all signs and symptoms; and failure, persistence or progression after 3 days of therapy, new clinical findings suggesting active infection or death related to pneumonia. Patients were considered nonevaluable when concomitant antibiotics had been given, when <80% of the total study medication had been ingested or when patients who, on review, did not meet inclusion or exclusion criteria. Adverse events were monitored throughout the study and, if present, reported to the institutional review board and to Pfizer, Inc.
A control group of healthy children attending a well child clinic at Children's Medical Center of Dallas were selected to match study patients by age, sex and approximate time of enrollment for nasopharyngeal and pharyngeal swabs for culture and PCR for C. pneumoniae and M. pneumoniae.
Bacteriology and serology. A positive culture, DFA, PCR or serology was considered indicative of infection by that organism.
Nasopharyngeal and pharyngeal swabs were placed in SP4 transport media and stored at −70°C until testing was done. C. pneumoniae cultures were performed with cyclohexamine-treated Hep 2 vial cultures (Bartels, Inc., Issaquah, WA) and subsequently stained with C. pneumoniae-specific monoclonal antibody for identification. M. pneumoniae cultures were performed using biphasic SP4 broth and agar and confirmed by guinea pig hemolysis and PCR. Viral DFA (Bartels) and cultures were performed on nasopharyngeal swab samples. Respiratory syncytial virus, adenovirus, parainfluenza 1, 2, 3 and influenza A and B were included in the DFA panel. Cell lines used for viral culture included MRC5, monkey kidney, AF549, WI 38 and Hep 2 cells. Blood cultures were performed with the BacT/Alert system (Organon Teknika).
Polymerase chain reaction. Nasopharyngeal and pharyngeal swabs were placed in sucrose-phosphate-glutamic acid transport medium and stored at −70°C until tested. Polymerase chain reaction used OMP 2 primers specific for C. pneumoniae and ORF 6 primers for M. pneumoniae. When the initial set of primers was positive, a second set of primers with molecular masses of 76 and 65 kDa, respectively, was used for confirmation (Abbott Diagnostics, Abbott Park, IL).6, 7 To test the reliability of our PCR technique for C. pneumoniae, selected pharyngeal and nasopharyngeal swab specimens from children were sent to the laboratory of Thomas Quinn, M.D., The Johns Hopkins University, Baltimore, MD, to compare results from the two laboratories.8-11 Positive and negative PCR test results in our laboratory were confirmed by Dr. Quinn for all specimens compared.
Serology. Serum was stored at −70°C for later transport on dry ice to the reference laboratory. Micro-immunofluorescence for C. pneumoniae serum IgM and IgG titers was performed in the laboratory of Thomas Grayston, M.D., Seattle, WA.12 The antibody titer was considered to be negative when the IgG was ≤1:8, acutely positive when the IgM was ≥1:16 or there was a 4-fold rise in IgG titer and acutely high when the IgG was ≥1:512. Chronic antibody titers were defined as an IgG ≥ 1:8 and ≤ 1:256 and an IgM ≤ 1:8.13 Enzyme-linked immunosorbent assay for M. pneumoniae serum antibody titer was performed in the laboratory of Gail Cassell, Ph.D., Birmingham, AL. The antibody titer was considered negative when the IgM was <1:10 and positive when the IgM was ≥1:10 or there was a 4-fold rise in IgG titer.14 Antibodies to pneumococcal C-polysaccharide and pneumolysin and immune complexes specific to C-polysaccharide, pneumococcal surface protein and pneumolysin were measured in the laboratory of Maija Leinonen, M.D., Oulu, Finland.15-21 Results of these studies were considered positive if the immune complex-bound antibody value to C-polysaccharide, pneumococcal surface protein or pneumolysin was ≥150 (optical density) and suggestive if it was ≥100 OD. A ≥2-fold titer rise in serum antibody between paired sera was considered indicative of acute infection.19
Statistical analysis. Statistical analyses for comparison of etiology, antibiotic effectiveness and treatment outcome were done with Epistat using Fisher's exact test.
Demographics. There were 174 patients enrolled from February, 1996, through December, 1997. Forty-seven percent were 0 to 2 years of age, 16% were 3 to 4 years, 25% were 5 to 8 years and 12% 9 to 16 years old. Ethnic origins included 92 African American patients, 54 Latin American, 23 Caucasian, 2 Oriental, 2 Indian and 1 Pakistani patient.
Patient accountability. We enrolled 174 patients. Six were excluded because of normal chest roentgenograms, leaving 168 patients who were evaluable for etiology of pneumonia. All 168 patients had initial nasopharyngeal or oral swab specimens available; however, paired serum samples were available from only 136 patients, an initial specimen only in 34 patients and no serum specimens in 4 patients.
Etiologic agents. An etiology was established in 73 (43%) of the 168 patients; 15 patients had 2 organisms.
Mycoplasma pneumoniae. Twelve of the 168 (7%) patients had evidence of acute M. pneumoniae infection by serology (10 patients), culture (5) and PCR (5). All infections occurred between October and June of the 2 study years. Three percent of patients 0 to 2 years of age had Mycoplasma infection compared with 11% of older children (Table 1).
C. pneumoniae. There were 10 of 168 (6%) patients with acute C. pneumoniae infection identified by serology (10 patients), culture (2) and PCR (5). Infection in 9 of the 10 patients occurred in children older than 3 years (Table 1).
Nineteen (11%) children meeting enrollment criteria had serologic evidence indicating prior exposure to C. pneumoniae. Twenty percent of the children 5 years of age and older had evidence of prior infection. Children with evidence of acute or prior C. pneumoniae infection increased from 5% at 0 to 2 years of age to 43% in those older than 9 years (Table 1).
None of the nasopharyngeal and pharyngeal swab specimens from 75 healthy children was positive by culture and PCR for C. pneumoniae or M. pneumoniae.
Viral disease. Viral studies were performed in 157 patients. Thirty-one of 157 (20%) were positive by culture (24 patients), DFA (15) or both (12). All viral infections occurred between October and March with the exception of two parainfluenza 1 infections, one each in July and September. The viral agents were respiratory syncytial virus (in 13 patients), influenza A (4), parainfluenza 3 (4), adenovirus (3), parainfluenza 1 (3), influenza B (1), enterovirus (1), cytomegalovirus (1) and herpes simplex virus (1). All isolates occurred in patients younger than 8 years of age.
Streptococcus pneumoniae. There were 129 patients evaluated for acute and convalescent serologic responses to S. pneumoniae infection. Thirty-five patients (27%) had evidence of acute pneumococcal infection (Table 2). Seventy-one percent of the patients with positive serology were seen between January and June of each year.
Coinfection. Fifteen patients had coinfection. Of the 35 patients with serologic evidence for acute pneumococcal infection, 14 (40%) had coinfection with a virus (5 patients), C. pneumoniae (5) or M. pneumoniae (4). One patient had a 4-fold rise in serum IgM antibody to C. pneumoniae and a positive IgM for M. pneumoniae.
Bacteriologic responses. For M. pneumoniae, there was concordance between results of culture and PCR in the initial samples of 5 patients, and 3 of those had a positive serology. By contrast culture and PCR were negative in 7 of the 10 patients who had positive serologic responses. One patient had a positive follow-up culture and PCR for M. pneumoniae, and one had positive PCR only. Both received azithromycin therapy. Three patients positive in the initial culture and PCR for Mycoplasma received erythromycin and had negative cultures and PCR on follow-up. All patients with evidence of Mycoplasma pneumonia were classified as clinical cure after treatment.
Of 10 patients with serologic evidence of acute C. pneumoniae infection, 5 had positive PCR and 2 of these had positive cultures. An acute serologic response was found in the 5 patients with positive PCR and culture results. There were no patients with positive culture or PCR for C. pneumoniae on follow-up. All Chlamydia patients were classified as a clinical cure.
Laboratory evaluation. The two radiologists who interpreted all chest roentgenograms independently were unable to distinguish among patients with bacterial, viral or no etiologic agent identified. Blood cultures were sterile in all patients. There was no correlation between results of the peripheral white blood cells and differential counts at the time of enrollment and specific etiology.
Reactive airway disease. Sixty-eight patients meeting enrollment criteria had a history of reactive airway disease (RAD) defined as one or more past episodes of wheezing.22 With the exceptions of pneumococcal infection in nonwheezing patients with a prior history of RAD and of chronic C. pneumoniae infection in nonwheezing children with no prior history, there were no significant differences among patients with serologic evidence of acute infection with regard to whether there was a prior history of RAD and whether wheezing was present at the time of enrollment (Table 3).
Clinical evaluation. Of the 168 patients who were assessed for etiology of pneumonia, 21 were excluded from clinical evaluation; 10 failed to return for follow-up examination and 11 did not complete treatment (see below). There were 88 patients younger than 5 years of age, of whom 39 (44%) were randomized to azithromycin and 49 (56%) to amoxicillin-clavulanate. There were 59 (40%) evaluable patients older than 5 years of age with 30 (51%) treated with azithromycin and 29 (49%) treated with erythromycin estolate.
Of the 147 clinically evaluable patients, 143 were classified as clinical cure. There were no differences in effectiveness of the different therapies even when only the patients with a presumed bacterial etiology were assessed.
Four patients were considered failures to therapy. One patient <5 years of age given azithromycin had serologic evidence of acute pneumococcal infection and was classified as a failure because at the 2-week follow-up visit, new symptoms and an infiltrate on chest radiography were identified. This patient had symptomatically recovered from the initial illness by Day 7. A second patient with documented respiratory syncytial virus infection was hospitalized on the 11th day after enrollment because of increasing respiratory distress. Chest radiograph revealed consolidation and a pleural effusion, from which grew S. pneumoniae resistant to penicillin (MIC 2 μg/ml), cefotaxime (3.0 μg/ml) and erythromycin (8 μg/ml). The patient was originally assigned to receive amoxicillin-clavulanate but was documented to receive only one-half of the prescribed amount of drug. The remaining two patients classified as failure, one from the amoxicillin-clavulanate group and one from the erythromycin group, did not have an etiology identified and failed to improve during treatment.
Adverse events. Eleven patients did not complete the prescribed therapy. A 5-year-old child was hospitalized 10 h after enrollment and after only one dose of erythromycin because of emesis and diarrhea thought to be caused by a viral illness. A second child, 4 years of age, was hospitalized 20 h after enrollment with worsening respiratory distress and increasing pulmonary infiltrates on chest film. This patient had received 2 doses of erythromycin. A third child, 8 years old, was hospitalized on the second day after enrollment, having received five doses of erythromycin therapy, because of emesis and dehydration thought to be associated with macrolide therapy. A fourth child, 6 years of age, was hospitalized at 24 h after enrollment for an acute exacerbation of reactive airways disease. Only one dose of azithromycin had been taken. An etiologic agent was not identified in any of these four patients. The remaining seven children stopped their medications at various times for inexplicable reasons. By telephone interviews they improved uneventfully and did not return for follow-up examinations.
A 5-year-old patient had a positive intradermal tuberculin test of 15 mm induration. The patient received a 5-day course of azithromycin therapy and became asymptomatic, at which time gastric aspirates were obtained which were negative on stained smears and culture. Repeat chest roentgenogram was considered normal and isoniazid was administered for 6 months.
Adverse events most likely associated with the study drugs were recorded in 33 (67%) of 49 patients who received amoxicillin-clavulanate: diarrhea (20 patients); genital candidiasis (6); rash (4); abdominal pain (2) and vomiting (1). There were 8 (25%) adverse events in the 29 subjects given erythromycin: diarrhea (2); headache (2); vomiting (1); rash (1); nausea (1); and swollen face (1). For the 69 patients who were treated with azithromycin; 10 (14%) adverse effects were recorded: diarrhea (3); rash (3); abdominal pain (2); vomiting (1); and oral thrush (1).
We were able to identify the presumed etiology of community-acquired pneumonia in 73 (43%) of 168 Dallas children. Culture was as sensitive as PCR for identifying M. pneumoniae infection, but neither was as good as serology. Culture for C. pneumoniae was less sensitive than PCR or serology. These data suggest that measurement of antibody response in paired sera still represents the best means for determining etiology of pneumonia caused by these two organisms in ambulatory children. It is possible that PCR for bacterial agents of pneumonia will become more sensitive and specific for diagnosis when using various body fluids, including plasma and buffy coat.
Because there is no standard for diagnosing nonbacteremic pneumococcal pneumonia, it is difficult to interpret results of serologic studies for S. pneumoniae infection. An analysis of pneumococcal serologic findings in children with acute otitis media or pneumonia compared with results in healthy controls suggests that determination of immune complexes and antibody titers can be useful for identifying S. pneumoniae as an etiologic agent.19 Furthermore elevated titers of pneumolysin antibody were demonstrated in only 3 of 186 healthy children23 and of pneumolysin and C-polysaccharide antibodies in 1 of 200 young adults with uncomplicated viral upper respiratory infections.24 Moreover in children with acute viral laryngitis, these antibody assays for pneumococcal infection were negative.25 These data suggest that the specificity of these pneumococcal serologic studies is high. The sensitivity of these tests, however, is unknown although in one recent study 22 of 25 patients with bacteremic pneumococcal pneumonia had at least one positive serologic test for S. pneumoniae, yielding a sensitivity of 88%.26 Our finding of serologic evidence of pneumococcal infection in 27% of ambulatory children with pneumonia was similar to the 25% of Finnish children studied by Heiskanen-Kosma et al.27
The percentage of cases with infection attributed to M. pneumoniae and C. pneumoniae in our study was low compared with the data of Block et al.4 and Harris et al.5 One difference is that our study was conducted in one community at a different time compared with many geographic locations 3 to 5 years earlier. Season, yearly patterns and age of the population studied can all influence the epidemiology of these two infections. Additionally those investigators often found M. pneumoniae and C. pneumoniae in nasopharyngeal culture or PCR in the absence of a serologic response. This contrasts with our results of greater sensitivity of serology for these organisms and could be explained by differences in carrier rates among the three studies and prevalence rates of Mycoplasma and Chlamydia infections in those communities. In the Dallas patients 31% of children 5 years of age and older had evidence of acute or chronic C. pneumoniae infections compared with 9% of younger subjects.
Viruses were the principal organisms present in nasopharyngeal secretions of children younger than 8 years of age. Respiratory syncytial virus, influenza A and parainfluenza 3 were the most common viruses identified in our study. Identifying a virus in the nasopharynx of a patient with roentgenographic evidence of pneumonia is suggestive of a causal relationship although dual viral-bacterial infection is possible. Coinfection occurred in 15 (9%) patients, 14 of whom had serologic evidence of S. pneumoniae coupled with evidence of viral, C. pneumoniae or M. pneumoniae infection. This rate of dual infection was smaller than that observed in children by other investigators.4, 5, 27
There have been repeated attempts to correlate M. pneumoniae and C. pneumoniae infection with exacerbation of reactive airway disease in children.28-32 Although our study was not designed to evaluate RAD and its relationship to etiology, we did not find any correlations among children with or without a history of RAD or the presence of wheezing on initial examination and acute infection with M. pneumoniae and C. pneumoniae.
In our ambulatory children with pneumonia, chest roentgenograms, blood cultures and total white blood cells and differential counts were not helpful in distinguishing one etiologic agent from another. There was no difference in effectiveness of the antibiotics used in our study, even among those with infections attributed to M. pneumoniae, C. pneumoniae and S. pneumoniae. Until additional information is available the selection of an antimicrobial agent for therapy of ambulatory children with pneumonia should be based on clinical judgment.
We thank the following people for their participation in this study: Kurt Olsen, Sharon Shelton, David Bakken, Hasan Jafri, Fanny Brito and Jean Hoyt. Special thanks to Thomas Quinn and Charlotte Gaydos for verification of the C. pneumoniae PCR.
The study was supported by a grant from Pfizer, Inc.
1. Bartlett JG, Breiman RM, Mandell LA, File TM. Community-acquired pneumonia in adults: guidelines for management. J Infect Dis 1998;26:811-38.
2. Klein JO. Diagnostic lung puncture in the pneumonias of infants and children. Pediatrics 1969;44:486-92.
3. Hughes JR, Dinesh SP, Cooper MR, Shah KV, Bose SK. Lung tap in childhood. Pediatrics 1969;44:477-85.
4. Block S, Hedrick J, Hammerschlag M, et al. Mycoplasma pneumoniae
and Chlamydia pneumoniae
in pediatric community-acquired pneumonia: comparative efficacy and safety of clarithromycin vs.
erythromycin ethylsuccinate. Pediatr Infect Dis J 1995;14:471-7.
5. Harris JS, Kolokathis A, Campbell M, et al. Safety and efficacy of azithromycin in the treatment
of community acquired pneumonia in children. Pediatr Infect Dis J 1998;17:865-71.
6. Ruland K, Himmelreich R, Herrmann R. Sequence divergence in the ORF6
gene of Mycoplasma pneumoniae.
J Bacteriol 1994;176:5202-9.
7. Wagels G, Rasmussen S, Timms P. Comparison of Chlamydia pneumoniae
isolates by Western blot (immunoblot) analysis and DNA sequencing of the omp 2
gene. J Clin Microbiol 1994;32:2820-3.
8. Gaydos CA, Quinn TC, Eiden JJ. Identification of Chlamydia pneumoniae
by DNA amplification of the 16S rRNA gene. J Clin Microbiol 1992;30:796-800.
9. Gaydos CA, Eiden JJ, Oldach D, et al. Diagnosis of Chlamydia pneumoniae
infection in patients with community-acquired pneumonia by polymerase chain reaction enzyme immunoassay. Clin Infect Dis 1994;19:157-60.
10. Gaydos CA, Fowler CL, Gill VJ, et al. Detection of Chlamydia pneumoniae
by polymerase chain reaction-enzyme immunoassay in an immunocompromised population. Clin Infect Dis 1993;17:718-23.
11. Gaydos CA, Palmer L, Quinn TC, et al. Phylogenetic relationship of Chlamydia pneumoniae
to Chlamydia psittaci
and Chlamydia trachomatis
as determined by analysis of 16S ribosomal DNA sequences. Int J Syst Bacteriol 1993;43:610-2.
12. Wang S, Grayston T. Immunologic relationship between genital TRIC, lymphogranuloma venereum, and related organisms in a new microtiter indirect immunofluorescence test. Am J Ophthalmol 1970;70:367-74.
13. Grayston T. Infections caused by Chlamydia pneumoniae
strain TWAR. Clin Infect Dis 1992;15:757-63.
14. Cassell GH, Gambill G, Duffy L. ELISA in respiratory infections of humans. In: Tully JG, Razin S, eds. Molecular and diagnostic procedures in mycoplasmology. vol 2. San Diego, CA: Academic Press, 1996:123-36.
15. Koskela M, Leinonen M, Luotonen J. Serum antibody response to pneumococcal otitis media. Pediatr Infect Dis J 1982;1:245-52.
16. Koskela M. Serum antibodies to pneumococcal C-polysaccharide in children: a response to acute pneumococcal otitis media or to vaccination. Pediatr Infect Dis J 1987;6:519-26.
17. Jalonen E, Paton JC, Koskela M, et al. Measurement of antibody responses to pneumolysin: a promising method for the presumptive etiological diagnosis of pneumococcal pneumonia. J Infect 1989;19:127-34.
18. Jalonen E, Taira S, Paton JC, et al. Pneumolysin produced by Bacillus subtilis
as antigen for measurement of pneumococcal antibodies by enzyme immunoassay. Serodiagn Immunother Infect Dis 1990;4:451-8.
19. Leinonen M. Syrjala H, Jalonen E, et al. Demonstration of pneumolysin antibodies in circulating immune complexes: a new diagnostic method for pneumococcal pneumonia. Serodiagn Immunother Infect Dis 1990;78:1020-7.
20. Korppi M, Leinonen M. Pneumococcal immune complexes in the diagnosis of lower respiratory infections in children. Pediatr Infect Dis J 1998;17:992-5.
21. Holloway Y, Snijder J, Boersma W. Demonstration of circulating pneumococcal immunoglobulin immune complexes by means of an enzyme-linked immunosorbent assay. J Clin Microbiol 1993;31:3247-54.
22. Emre U, Roblin PM, Gelling M, et al. The association of Chlamydia pneumoniae
infection and reactive airway disease in children. Arch Pediatr Adolesc Med 1994;148:727-32.
23. Korppi M, Leinonen M. Pneumococcal pneumonia in children: new data from circulating immune complexes. Eur J Pediatr 1997;156:341-2.
24. Korppi M, Heiskanen-Kosma M, Jalonen E, et al. Etiology of community-acquired pneumonia in children treated in hospital. Eur J Pediatr 1993;152:24-30.
25. Korppi M, Leinonen M, Mäkelä PH, Launiala K. Bacterial involvement in laryngeal infections in children. Acta Paediatr Scand 1990;79:564-5.
26. Porath A, Schlaeffer F, Pick N, et al. Pneumococcal community-acquired pneumonia in 148 hospitalized adult patients. Eur J Clin Microbiol Infect Dis 1997;16:863-70.
27. Heiskanen-Kosma T, Korppi M, Jokinen C, et al. Etiology of childhood pneumonia: serological results of a prospective population-based study. Pediatr Infect Dis J 1998;17:986-91.
28. Berkovich S, Millian SJ, Snyder RD. The association of viral and Mycoplasma
infections with recurrence of wheezing in the asthmatic child. Ann Allergy 1970;28:43-9.
29. Drews A, Atmar R, Glezen WP, et al. Dual respiratory virus infections. Clin Infect Dis 1997;25:1421-9.
30. Gil JC, Cedillo RL, Maygoitia BG, Paz MD. Isolation of Mycoplasma pneumoniae
from asthmatic patients. Ann Allergy 1993;70:23-5.
31. Horn M, Reed S, Taylor P. Role of viruses and bacteria in acute wheezy bronchitis in childhood: a study of sputum. Arch Dis Child 1979;54:587-92.
32. Huhti E, Mokka T, Nikoskelainen J, Halonen P. Association of viral and Mycoplasma
infections with exacerbations of asthma. Ann Allergy 1974;33:145-9.