Background: Despite pneumococcal antibiotic resistance rates in excess of 25%, macrolides remain first-line agents for the treatment of community-acquired pneumonia.
Methods: Prospective, population-based surveillance was conducted to identify cases of pneumococcal bacteremia in Toronto and Peel, Canada, between 2000 and 2004. "Macrolide failures" were defined as cases of bacteremia occurring during outpatient treatment with macrolide antibiotics or within 2 days after treatment. Macrolide susceptibility was determined according to Clinical Laboratory Standards Institute guidelines; common macrolide resistance mechanisms were determined by genotyping.
Results: During the 5 years of surveillance, there were 1696 episodes of pneumococcal bacteremia (8.5 cases/100,000 population/year), of which 60 (3.5%) were failures of outpatient macrolide therapy. Resistant isolates were more common among cases of bacteremia after failure of macrolide therapy (37 [64%] of 58 cases) than among cases of bacteremia after failure of nonmacrolide antibiotics (16 [22%] of 74 cases; P<.001) or cases of bacteremia that occurred without prior antibiotic therapy (193 [12%] of 1569 cases; P<.001). Macrolide failures were significantly more common among cases of pneumococcal bacteremia with isolates exhibiting an erythromycin MIC of 1 microg/mL than among isolates exhibiting MICs < or = 0.25 microg/mL (3 [38%] of 8 cases vs. 21 [1.5%] of 1394 cases of bacteremia; P<.001). Increases in the MIC to 1 microg/mL were not associated with further increases in the likelihood of macrolide failure. Low-level resistance conferred by mefA and high-level resistance conferred by ermB were equally overrepresented among macrolide failures.
Conclusions: Macrolide resistance contributes to an increased risk of macrolide failure, irrespective of the underlying resistance mechanism or of the degree of elevation in erythromycin MIC.