Genomic classification and antimicrobial resistance profiling of Streptococcus pneumoniae and Haemophilus influenzae isolates associated with paediatric otitis media and upper respiratory infection

Acute otitis media (AOM) is the most common childhood bacterial infectious disease requiring antimicrobial therapy. Most cases of AOM are caused by translocation of Streptococcus pneumoniae or Haemophilus influenzae from the nasopharynx to the middle ear during an upper respiratory tract infection (URI). Ongoing genomic surveillance of these pathogens is important for vaccine design and tracking of emerging variants, as well as for monitoring patterns of antibiotic resistance to inform treatment strategies and stewardship. In this work, we examined the ability of a genomics-based workflow to determine microbiological and clinically relevant information from cultured bacterial isolates obtained from patients with AOM or an URI. We performed whole genome sequencing (WGS) and analysis of 148 bacterial isolates cultured from the nasopharynx ( N = 124, 94 AOM and 30 URI) and ear ( N = 24, all AOM) of 101 children aged 6–35 months presenting with AOM or an URI. We then performed WGS-based sequence typing and antimicrobial resistance profiling of each strain and compared results to those obtained from traditional microbiological phenotyping. WGS of clinical isolates resulted in 71 S. pneumoniae genomes and 76 H. influenzae genomes. Multilocus sequencing typing (MSLT) identified 33 sequence types for S. pneumoniae and 19 predicted serotypes including the most frequent serotypes 35B and 3. Genome analysis predicted 30% of S. pneumoniae isolates to have complete or intermediate penicillin resistance. AMR predictions for S. pneumoniae isolates had strong agreement with clinical susceptibility testing results for beta-lactam and non beta-lactam antibiotics, with a mean sensitivity of 93% (86–100%) and a mean specificity of 98% (94–100%). MLST identified 29 H. influenzae sequence types. Genome analysis identified beta-lactamase genes in 30% of H. influenzae strains, which was 100% in agreement with clinical beta-lactamase testing. We also identified a divergent highly antibiotic-resistant strain of S. pneumoniae , and found its closest sequenced strains, also isolated from nasopharyngeal samples from over 15 years ago. Ultimately, our work provides the groundwork for clinical WGS-based workflows to aid in detection and analysis of H. influenzae and S. pneumoniae isolates.