Comparison of Phenotypic and Whole-Genome Sequencing-Derived Antimicrobial Resistance Profiles of Legionella pneumophila Isolated in England and Wales from 2020 to 2023

<b>Background:</b> Antimicrobial resistance (AMR) in <i>Legionella pneumophila</i> is emerging as a concern, particularly with resistance to macrolides and fluoroquinolones. Although clinically significant resistance in <i>Legionella pneumophila</i> remains uncommon, systematic genomic surveillance using whole-genome sequencing (WGS) is needed to anticipate treatment failure as metagenomic diagnostics move toward routine use. <b>Objectives:</b> We assessed the UK Health Security Agency AMR pipeline for predicting resistance in <i>L. pneumophila</i> by analysing 522 <i>L. pneumophila</i> isolates from England and Wales (2020-2023) together with nine database sequences that carry confirmed 23S rRNA mutations conferring high-level azithromycin resistance. The objective of the present study was to examine the presence of antimicrobial resistance genes (ARGs) in <i>L. pneumophila</i> isolates and to determine whether they exhibited phenotypic resistance through minimum inhibitory concentration (MIC) testing. <b>Methods:</b> Serogroups (sgs) were determined using an in-house qPCR assay, and <i>L. pneumophila</i> non-sg1 isolates were serogrouped using the Dresden monoclonal antibody (mAb) typing method. Sequence types were determined using the standard sequence-based typing method by Sanger sequencing. WGS reads were screened against standard AMR databases to identify resistance genes and resistance-mediating mutations. Agar dilution measured MICs for azithromycin, erythromycin, ampicillin, levofloxacin, tetracycline and spectinomycin in isolates possessing the <i>bla_OXA-29</i>, <i>lpeAB</i> or <i>aph(9)-Ia</i> gene. <b>Results:</b> AMR screening detected <i>lpeAB</i>, two allelic β-lactamase variants (<i>bla_OXA-29</i> and <i>bla_LoxA</i>) and <i>aph(9)-Ia</i> in 165 of the 522 <i>L. pneumophila</i> isolates, while all high-azithromycin MIC reference sequences contained the expected 23S mutation. Only <i>lpeAB</i> was associated with a significant twofold elevation in macrolide MICs. Neither β-lactamase variant increased ampicillin MICs, and <i>aph(9)-Ia</i> carriage did not correlate with higher spectinomycin MICs. <b>Conclusions:</b> Advanced genomic analytics can now deliver timely therapeutic guidance, yet database-flagged genes may not translate into phenotypic resistance. Continuous pairing of curated mutation catalogues with confirmatory testing remains essential for distinguishing clinically actionable determinants such as 23S mutations and <i>lpeAB</i> from silent markers like <i>bla_OXA-29</i> and <i>aph (9)-Ia</i>.