Answers are grounded in this paper and cite their sources.
In North America lower respiratory tract infections, consisting mainly of pneumonia and bronchiolitis, are most common in the first year of life with an annual incidence of 240 episodes per 1000 infants compared with 134 episodes per 1000 children from ages 1 to 15 years.1 Worldwide 1 in 200 infants less than 1 year of age is hospitalized for lower respiratory tract infections annually, with a mortality rate of 1 to 5%.2 In addition to the consequences of acute infection, there is evidence that lower respiratory tract infections in early infancy may be linked to sequelae in the form of chronic pulmonary disease, recurrent wheezing, fibrosis and bronchiectasis.3, 4 Infants with pneumonia present with tachypnea, dyspnea, cough and nasal flaring or with nonspecific signs such as fever, lethargy and rigors.5 There is no noninvasive “gold standard” test for the diagnosis of pneumonia but the clinical diagnosis is usually confirmed with a chest radiogram showing consolidation or interstitial infiltrates.5 Pneumonias caused by bacteria cannot be differentiated from those caused by viruses by radiographic or clinical criteria alone but requires either cultures of lung aspirate or indirect ancillary tests.6 Bronchiolitis is a clinical syndrome of young infants characterized by the acute onset of wheezing associated with rhinorrhea, dyspnea, indrawing or cough. It is usually caused by viruses, most commonly respiratory syncytial virus (RSV).7, 8 Patients with the clinical syndrome of bronchiolitis may have normal chest radiograms or may have hyperinflation, perihilar infiltrates and occasionally atelactasis.7, 8 Infants with wheezing who have associated radiographic interstitial infiltrates or consolidation are considered to have pneumonia and are usually treated with antibiotics because of difficulty in distinguishing viral from bacterial pneumonia.6, 9-12 A recent cross-Canada study indicated that 57 to 81% of all infants admitted with RSV lower respiratory tract infections receive antibiotics.13 Current treatment of pneumonia in young infants is usually aimed at bacterial pathogens and consists of beta-lactam antibiotics.9 However, several recent studies on the etiology of pneumonia have suggested a role for organisms such as Chlamydia trachomatis and Ureaplasma urealyticum that are susceptible to macrolide antibiotics such as erythromycin, but not to beta-lactam antibiotics.14-23 Needle aspiration of lung tissue for culture is usually not done in developed countries because of the low mortality associated with infantile pneumonia and the risks associated with the procedure. Thus the etiologic diagnosis of pneumonia usually depends on blood culture, isolation of an organism from the upper respiratory tract, seroconversion or a high initial antibody titer to a specific microorganism. Because these diagnostic methods are indirect, a difference in prevalence of these agents in infants with pneumonia compared with a group of infants without pneumonia is necessary to implicate the pathogens in causing disease. This study was therefore conducted to compare the rates of C. trachomatis and U. urealyticum infection in infants with pneumonia and bronchiolitis. Methods.Patient population and enrollment. During the study period, The Hospital for Sick Children, Toronto, Ontario, was a 420-bed primary, secondary and tertiary care hospital serving a population base of over 3 million people. All infants younger than 6 months of age hospitalized there with an admitting diagnosis of pneumonia, pneumonitis, bronchiolitis or lower respiratory tract infection between April 1, 1991, and September 30, 1992, were assessed daily for eligibility. Infants with known immunodeficiency and those who had never been discharged from a hospital after birth were excluded. The study was approved by the Human Subjects Review Board at The Hospital for Sick Children and the infants were enrolled after obtaining parental consent. For the purposes of this study a diagnosis of pneumonia was made in infants admitted with compatible clinical signs and symptoms (dyspnea, cough, nasal flaring or nonspecific signs such as fever and lethargy) and presence of an interstitial infiltrate or consolidation on chest radiogram. Infants with wheezing were diagnosed to have bronchiolitis, and those who were admitted during a similar time frame (generally the same week) as infants admitted with pneumonia were enrolled as the comparative group. All such infants received chest radiograms. All of the radiograms were independently reviewed by two senior radiologists who were unaware of the admitting diagnosis. Because they are frequently treated with antibiotics, patients with bronchiolitis as an admitting diagnosis who had consolidation on their chest radiogram as determined by the two radiologists were considered to have pneumonia. Patients with an admission diagnosis of pneumonia and who had wheezing, but whose radiograms did not reveal any consolidation on review by the two radiologists, were considered to have bronchiolitis. Patients admitted with a diagnosis of pneumonia with normal admission radiograms were excluded because both radiologic and clinical pneumonia were needed to meet entry criteria. Interobserver and intraobserver reliability for the presence of consolidation on a subset of 40 randomly selected chest radiograms from the current study was demonstrated to be substantial using the average quadratic weighted kappa statistic (kappa for intraobserver = 0.92 and for interobserver = 0.91, overall agreement for interpretation of a radiogram showing air space disease of 97.5%).24 Clinical specimens. A nasopharyngeal swab for viral diagnosis and chlamydial culture and a nasopharyngeal aspirate for Bordetella pertussis and U. urealyticum culture were obtained from all patients. However, there was not always enough specimen available to perform all studies on all patients. Urine specimens were obtained for bacterial antigen detection. Blood cultures were obtained from febrile patients (temperature ≥38.5°C). Acute serum was obtained within 24 h of admission and convalescent serum was obtained 2 to 4 weeks later. Results from these diagnostic tests were available to clinicians managing the children. Laboratory methods. For direct immunofluorescence microscopy of viruses, nasopharyngeal swabs were delivered to the laboratory in transport medium containing phosphate-buffered saline and 0.5% gelatin. The cells present on the swab were collected by centrifugation, spotted and fixed on multiple wells of a masked slide which were reacted with monoclonal antibodies to RSV; parainfluenza 1, 2 and 3 viruses; and influenza A and B viruses (Chemitron, Temecula, CA; Institute Armand Frappier, Montreal, Canada).25 The wells were then reacted with a fluorescein-conjugated antibody to murine IgG and examined by fluorescence microscopy for the presence of cells with characteristic fluorescence staining. The remainder of each specimen was inoculated into HeLa and primary African green monkey kidney cell cultures which were incubated for 7 to 14 days for virus isolation. Nasopharyngeal swabs for C. trachomatis were delivered to the laboratory in transport medium containing phosphate-buffered saline, glutamine and sucrose. The specimen was inoculated into HeLa cells in shell vials under centrifugal inoculation and the cultures followed by immunofluorescence microscopy.26 Sera were tested for total antibody to C. trachomatis by a commercial indirect immunofluorescence test (Electronucleonics Inc., Columbia, MD) with infected cells fixed on glass slides as antigen. Sera with titers ≥1:32 were further tested for the presence of IgM antibody to C. trachomatis.27 Nasopharyngeal aspirates were cultured for U. urealyticum on trypticase soy broth and agar plates and identified by the indirect urease test.28 Hayflick broth and agar plates supplemented with arginine were used to culture for large colony mycoplasmas. Acute and convalescent sera were tested for antibody to Mycoplasma pneumoniae by an enzyme-linked immunoassay as described previously.29 Blood samples (0.5 to 3 ml) were inoculated into vials of PEDS Plus medium for processing by the Bactec NR-660 (Becton Dickinson Diagnostic Instrument Systems, Mississauga, Ontario, Canada) according to the manufacturer's instructions. Specimens were kept for 7 days and subcultured onto 5% horse blood agar and chocolate agar plates for incubation at 37°C in 5% CO2 if a significant growth index was generated (growth value, >10 U on Day 1 or >20 U on Days 2 to 7 or an increase in growth index of >5 units between readings). Fresh urine samples were used for detection of Haemophilus influenzae type b and Streptococcus pneumoniae antigens by latex agglutination (Directigen Meningitis Combo Test; Becton Dickinson Microbiology Systems, Cockeysville, MD). Portions of nasopharyngeal aspirates were subjected to direct fluorescent antibody testing,30 for Bordetella pertussis cultured using a Bordet Gengou medium as well as a charcoal-cephalexin medium31 and incubated at 35°C for 5 days in a moist chamber. Characteristic colonies were identified by Gram's stain, immunofluorescence and biochemical testing. Interpretation of laboratory results. Patients were diagnosed as having an infection if an organism was isolated or antigen was detected by one of the immunospecific detection methods. Infection was also diagnosed if seroconversion to C. trachomatis occurred or if C. trachomatis IgM was detected. Statistical analysis. The primary outcome was the comparison between pneumonia and bronchiolitis groups for isolation of C. trachomatis and U. urealyticum. The sample size was based on: (1) the desire to determine the frequency of infection with these erythromycin-treatable agents in infants admitted with pneumonia at our institution, assuming a 50% incidence with 90% confidence limits ±10%; and (2) ability to show a relative risk of colonization of 3.0 in patients with pneumonia compared with those with bronchiolitis with an 80% power if the incidence in bronchiolitis patients was 20%. A sample size of 69 patients/group satisfied both of these requirements.32 Differences in group proportions were compared using chi square with Yates correction or Fisher's exact test. Unconditional logistic regression modeling was used to evaluate the association between each of C. trachomatis, U. urealyticum, cytomegalovirus and respiratory virus infection (predictor variables) and the outcome of pneumonia or bronchiolitis.32, 33 Univariate screening of each predictor variable (C. trachomatis, U. urealyticum or virus) was performed, followed by continuous adjustment for age (in weeks) as a confounding variable. Odds ratios and 95% confidence intervals were calculated from coefficients obtained from the univariate and age-adjusted equations.32 For all analyses a two-tailed P value of 0.05 was considered to indicate statistical significance. Results. During the study period 80 infants (≤6 months of age) were admitted with an initial diagnosis of pneumonia. Seventy-four infants of the same age group admitted with bronchiolitis were enrolled as controls. After review of all the chest radiograms by the 2 study radiologists, 13 (17.5%) of the 74 bronchiolitis patients were considered to have radiologic pneumonia and 16 (20%) of the 80 pneumonia patients with normal radiograms but with wheezing were considered to have bronchiolitis. Six patients admitted with pneumonia were excluded from the analysis because of normal chest radiograms. Thus for the final analysis, there were 71 infants with pneumonia and 77 with bronchiolitis. Before admission 13 (18.3%) of the 71 pneumonia patients were receiving oral antibiotics compared with 24 (31.2%) of the 77 study bronchiolitis patients, P = 0.07. The demographic characteristics of the pneumonia and bronchiolitis patients are shown in Table 1. There were more female, younger and black patients with pneumonia. Wheezing and coryza were more common in patients with bronchiolitis, and apnea was more common in those with pneumonia (Table 1). Five of the 10 pneumonia patients with apnea had RSV, 1 had parainfluenza 1 and 3 and no organism was identified in the remaining 4. Two of the 3 bronchiolitis infants with apnea had RSV. The mean number of days of illness before presentation (5.6 days) in the infants with pneumonia was not significantly different from that of those with bronchiolitis (6.3 days), P = 0.64. Figure 1 shows the seasonal distribution of patients based on diagnosis. The majority (>80%) of the bronchiolitis controls were enrolled within the same time frame (generally 1 week) at the cases. The identified infectious pathogens are shown in Table 2. A pathogen was identified in 35 (49.3%) of pneumonia patients and in 51 (66.2%) of bronchiolitis patients. More patients with bronchiolitis had a viral agent identified than did patients with pneumonia. The most commonly isolated virus in both groups was RSV, accounting for 15 (65%) of the 23 viruses in infants diagnosed with pneumonia and 39 (85%) of 46 viruses identified in those with bronchiolitis (P = 0.07). Eight (35%) of the viruses identified in pneumonia patients were parainfluenza 1 (2 patients) or parainfluenza 3 (6 patients), whereas only 2 (4.4%) parainfluenza viruses (1 parainfluenza 1 and 1 parainfluenza 3) were isolated in bronchiolitis patients. One pneumonia patient was coinfected with parainfluenzas 1 and 3. Influenza viruses were identified in 2 (9%) of the 23 patients with pneumonia and 5 (10.8%) of the 46 infants with bronchiolitis in whom a virus was identified. Both of the influenza viruses in pneumonia infants were influenza A, whereas 2 bronchiolitis patients had influenza A and 3 had influenza B. Of the 12 infants diagnosed with chlamydial infection, 8 had elevated serum IgM titers, 1 had a positive culture of nasopharyngeal swab and the remaining 3 were positive by both tests. Six patients had dual infections. One infant with pneumonia was infected with both C. trachomatis and U. urealyticum, another was infected with C. trachomatis and parainfluenza virus 3 and a third was infected with U. urealyticum and parainfluenza virus 3. Two patients with bronchiolitis who were positive for U. urealyticum also had an RSV infection, and a third was infected with both C. trachomatis and RSV. Of 54 pneumonia patients assessed for C. trachomatis, it was identified in 6 of 19 (31.6%) who presented without wheezing compared with 3 of 35 (8.6%) with wheezing, P = 0.05 (Table 3). All 9 patients with C. trachomatis-associated pneumonia were afebrile. Pertussis was not recovered from any of 46 infants with pneumonia or 58 with bronchiolitis who were tested. Acute serum IgG and IgM was negative in all of 59 infants with bronchiolitis and all 33 infants with pneumonia tested for M. pneumoniae infection. Convalescent serology (IgG) remained negative in all 30 bronchiolitis and 13 pneumonia patients who were tested. One (1.5%) of 68 pneumonia patients tested had a positive blood culture for S. pneumoniae, whereas none of 52 bronchiolitis patients had a positive blood culture. All urine latex agglutination tests for bacterial antigens were negative. Thus among pneumonia patients C. trachomatis and U. urealyticum were more commonly identified than conventional bacterial pathogens (P < 0.001). On univariate analysis C. trachomatis was significantly associated with pneumonia whereas viral infection was associated with bronchiolitis (Table 4). After adjusting for age, the odds ratio of C. trachomatis association with pneumonia was 3.3 (P = 0.09), whereas the odds ratio for virus association with pneumonia was 0.30 (P = <0.01). U. urealyticum was not significantly associated with pneumonia when compared with bronchiolitis. Discussion. In this comparative study young infants admitted with pneumonia were infected with macrolide antibiotic treatable organisms (C. trachomatis and U. urealyticum) at more than double the frequency of infants admitted with bronchiolitis. Because all the diagnostic tests were indirect, the addition of a comparison group of infants was important for interpretation of organisms identified. This study is unique in that it is the first one to use expert pediatric radiologists to categorize infants as having pneumonia based on radiographic findings. Previous studies of infantile pneumonia have relied on a single clinician or radiologist report of “pulmonary infiltrates” to diagnose pneumonia, without further definition. Although conditions such as atelectasis may mimic consolidation, we felt that agreement by two experienced radiologists on the presence of consolidation, with exclusion of any patient felt by either or both to have atelectasis alone would minimize this problem. Furthermore such a classification is relevant because in many institutions infants with a clinical diagnosis of bronchiolitis and pulmonary consolidation on their radiograms are usually treated with antibiotics.13, 34-36 The routine use of ampicillin in such infants has been shown in at least one open label randomized trial not to be of benefit,36 but there are no studies of the efficacy of macrolide agents. The age-adjusted odds ratio of association of macrolide-treatable organisms with pneumonia vs. bronchiolitis of 2.2, although not statistically significant, is consistent with the association noted by previous authors.14-23 The study sample size was limited by incomplete testing in the patients with pneumonia, but the relative stability of the odds ratio after adjustment for age suggests that the noted association is real. The highest frequency of isolation of C. trachomatis was in afebrile pneumonia patients who did not have wheezing, and these infants may represent a group that warrants special consideration for a trial of erythromycin therapy. The diagnosis of a pathogen in about one-half of the infants is similar to the detection rates noted from other studies of lower respiratory infection in infants and children.17, 19, 21, 22 The low incidence of U. urealyticum infection in patients with pneumonia may have been inferred from the lack of serologic diagnosis or may indicate that it is present in pneumonia patients only as an innocent bystander. This further emphasizes the importance of having a comparison patient population. The fact that the control population consisted of infants with bronchiolitis, another respiratory infection may have reduced the differences observed with the pneumonia group. A completely healthy comparison population would have been preferred but was not feasible given the diagnostic procedures called for in the study. The possibility of higher rates of infection with conventional bacteria in the infants with pneumonia cannot be ruled out because culture of lung aspirates could not be performed at our institution and almost one-fifth of the patients with pneumonia had received antibiotics before admission. However, previous studies37-39 have suggested that urinary antigen tests should be >90% sensitive in detecting infections with H. influenzae type b and about 50% sensitive for S. pneumoniae, the two most common bacterial pathogens of pneumonia in lung aspirate studies. Even if 50% of cases of pneumococci were missed by blood cultures and urine latex testing, the failure of detection of any bacteria with our urinary antigen testing combined with only one positive blood culture for S. pneumoniae suggests that usual bacterial pathogens were not a significant cause of pneumonia in our population and is consistent with other North American studies of pneumonia in infancy.14-18 One previous study of pediatric acute lower respiratory infections used serum enzyme antibody immunoassays to demonstrate infection with S. pneumoniae and H. influenzae and Branhamella catarrhalis, but the proportion of infected patients who were younger than 6 months of age could not be deduced from that report.40 Nonetheless the proportion of infants younger than 1 year of age diagnosed with bacterial infections even with the serologic techniques was less than 5%. Prior antibiotics may prevent recovery of more bacterial agents in bronchiolitis patients, but macrolide antibiotics are rarely used in this age group and are therefore unlikely to explain the differential recovery of C. trachomatis from pneumonia patients. Failure of diagnosis of M. pneumoniae is consistent with previous observations that young infants do not contract this infection.41 As observed previously14-22 respiratory viruses were the most frequent organisms identified in the pneumonia and bronchiolitis patients, with significantly higher rates of isolation of viruses in bronchiolitis patients. Only one patient had evidence of infection with a bacterial pathogen which would be treated by beta-lactam ± aminoglycosides antibiotic regimens often used to treat pneumonia in young infants. Erythromycin and other macrolide antibiotics are effective against these organisms as well as against C. trachomatis. Thus prospective trials of erythromycin or other macrolides are indicated given the lack of efficacy of conventional antibiotics for the most commonly identified pathogens in this patient population. Acknowledgments. The study was funded by Grant 04108 from the Ontario Ministry of Health. HDD was funded by the Ontario Ministry of Health Research Personnel Development Fellowship Programme award and the Canadian Fellowship Programme The for the serologic tests for M. H. M. C. M. of of Microbiology Hospital for Sick Children of and Hospital The of Toronto, Ontario, distribution of pneumonia and bronchiolitis cases.
No takes yet. Share an insight, caveat, or question.
Davies et al. (1996) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: