Background: Increasing macrolide resistance among middle ear isolates complicates treatment of otitis media in children. When macrolide resistance is mediated via an efflux pump (M phenotype), the MICs of erythromycin, clarithromycin and azithromycin are usually below 32 microg/ml, and the pneumococcus remains susceptible to clindamycin. The association of prior specific macrolide therapy with the isolation of a macrolide resistant strain has not been reported.
Objectives: To determine the mechanism of macrolide resistance in Streptococcus pneumoniae recovered from the middle ears of children with otitis media and their association, if any, with ethnicity, age, serogroup/serotype and prior antibiotic therapy.
Methods: Middle ear isolates collected by members of the United States Pediatric Multicenter Pneumococcal Surveillance Group during a 6-year period from September 1994 through August 2000 were studied. Antibiotic susceptibility to penicillin and ceftriaxone was determined by microbroth dilution. Disc diffusion susceptibility to erythromycin and clindamycin was performed to categorize macrolide resistance mechanisms. The medical record was reviewed to determine demographics and history of previous antibiotic therapy. Isolates were serogrouped or serotyped by the capsular swelling method.
Results: Of the 1088 isolates available for testing, 51% were nonsusceptible to penicillin and 37% were nonsusceptible to erythromycin. Erythromycin resistance increased form 15% in 1994 through 1995 to 56% in 1999 through 2000. Seventy-five percent of macrolide-resistant strains were M phenotype. Macrolide resistance was less likely in isolates recovered from African-Americans and more likely in isolates obtained from children <3 years of age and from isolates obtained at time of tympanostomy tube placement. Neither erythromycin, nor clarithromycin nor azithromycin prescribed in the 30 days before infection was more likely than another to be associated with increased macrolide resistance. However, any macrolide alone or in combination with another antimicrobial taken before infection was associated with increased macrolide resistance among the S. pneumoniae organisms isolated from the middle ear.
Conclusions: Macrolide resistance among middle ear isolates of S. pneumoniae increased during the 6-year study. The proportion of M phenotype remained constant at 75%, meaning that these isolates remain susceptible to clindamycin. Continued surveillance to document potential changes is essential.