Etiology and treatment of community-acquired pneumonia in ambulatory children

Objectives. To determine the etiology of community-acquired pneumonia in ambulatory children and to compare responses to treatment with azithromycin, amoxicillin-clavulanate or erythromycin estolate. Methods. Ambulatory patients with pneumonia were identified at the Children's Medical Center of Dallas, TX. Children age 6 months to 16 years with radiographic and clinical evidence of pneumonia were enrolled and randomized to receive either azithromycin suspension for 5 days or a 10-day course of amoxicillin- clavulanate for those <5 years or erythromycin estolate suspension for those ≥5 years. Blood culture was obtained in all patients and we obtained nasopharyngeal and pharyngeal swabs for culture and polymerase chain reaction (PCR) testing for Chlamydia pneumoniae and Mycoplasma pneumoniae and nasopharyngeal swabs for vital direct fluorescent antibody and culture. Acute and convalescent serum specimens were tested for antibodies to C. pneumoniae, M. pneumoniae and Streptococcus pneumoniae. Patients were evaluated 10 to 37 days later when repeat specimens for serology, PCR and culture were obtained. For comparative purposes healthy children attending the well-child clinic had nasopharyngeal and pharyngeal swabs obtained for PCR and culture for C. pneumoniae and M. pneumoniae. Results. Between February, 1996, and December, 1997, we enrolled 174 patients, 168 of whom fulfilled protocol criteria for evaluation. There were 55% males and 63% were <5 years of age. All blood cultures were sterile and there was no correlation between the white blood cell and differential counts and etiology of pneumonia. Etiologic agents were identified in 73 (43%) of 168 patients. Infection was attributed to M. pneumoniae in 7% (12 of 168), C. pneumoniae in 6% (10 of 168), S. pneumoniae in 27% (35 of 129) and viruses in 20% (31 of 157). None of the swab specimens from 75 healthy control children was positive for C. pneumoniae or M. pneumoniae. Clinical response to therapy was similar for the three antibiotic regimens evaluated, including those with infection attributed to bacterial agents. Conclusion. Although a possible microbial etiology was identified in 43% of the evaluable patients, clinical findings and results of blood cultures, chest radiographs and white blood cell and differential counts did not distinguish patients with a defined etiology from those without a known cause for pneumonia. There were no differences in the clinical responses of patients to the antimicrobial regimens studied.