LAUSR

OBJECTIVES In recent years, an increasing prevalence of macrolide resistance among pneumococci in Bangladesh has been observed. However, the scenario remains incomplete, as few isolates (<1%) are available from pneumonia cases and most pneumococcal meningitis cases (>80%) are culture-negative. This study optimised a triplex PCR method to detect macrolide resistance genes (MRGs) (mefA and ermB) and cpsA from culture-negative pneumococcal cases to predict the prevalence and level of macrolide resistance. METHODS The presence of MRGs among pneumococcal strains (n=153) with a wide range of erythromycin MICs (<0.5 to ≥256mg/L) was determined by PCR. Triplex PCR was validated by simultaneous detection of MRG(s) and cpsA in culture-negative clinical specimens and corresponding isolates. The known impact of the presence of specific MRG(s) on MICs of strains was used to predict the MICs of non-culturable strains based on the presence/absence of MRG(s) in the specimens. RESULTS None of the erythromycin-susceptible isolates possessed any of the MRGs, and all non-susceptible strains had ≥1 MRG. MICs were 2-16mg/L and ≥256mg/L for 93% of strains with mefA and ermB, respectively, whereas 100% of isolates with both genes had MICs≥256mg/L. PCR for body fluids showed 100% concordance with corresponding isolates when tested for MRG(s) in parallel. CONCLUSIONS Erythromycin MICs can be predicted for non-culturable strains with 93-100% precision based on detection of ermB and/or mefA. This method will be useful for establishing comprehensive surveillance for macrolide resistance among pneumococci, specifically in the population with prior antibiotic use.