High-risk extended-spectrum β-lactamase-producing Klebsiella pneumoniae ST307 in a neonatal sepsis outbreak in Zambia: a secondary genomic analysis.

Neonatal sepsis caused by multidrug-resistant Klebsiella pneumoniae contributes substantially to neonatal morbidity and mortality, especially where treatment options and genomic surveillance are limited. Phenotypic microbiology defines susceptibility, whereas Illumina short-read whole-genome sequencing (WGS) links phenotypes to lineage, capsule, resistome, virulome and mobilome contexts. However, complete characterization of mobile genetic element (MGE) structure remains inferential. This secondary analysis re-examined 411 public K. pneumoniae genomes from the University Teaching Hospital neonatal outbreak in Zambia. The population was dominated by ST307/KL102/OL2α.2, a comparatively recent but globally expanding high-risk clone, accounting for 284 out of 411 genomes (69.1%). Minority genomes included established resistance or virulence-associated high-risk backgrounds, including carbapenemase-positive KL15/OL4 lineages and yersiniabactin-positive ST147/ST985. Extended-spectrum β-lactamase (ESBL) genotypes were widespread (396 out of 411; 96.4%) and mainly driven by blaCTX-M-15 (406 out of 411; 98.8%), whereas carbapenemase carriage was restricted to blaNDM-5-positive ST5856/ST340 (11 out of 411; 2.7%). Plasmids were the principal acquired resistance layer, as most AMR genes were plasmid-associated, and similarity analysis revealed a cohesive ST307-linked IncFIB/IncFII AA277 ESBL background, in contrast to distinct minority ESBL and IncX3 blaNDM-5 architectures. This pattern supports clonal maintenance of a stable ESBL plasmid background within ST307, while identifying IncX3 blaNDM-5 plasmids in minority lineages as a key convergence risk. Resistance genes were further structured by integron cassettes, insertion-sequence modules and plasmid cargo containing metal-, biocide- or stress-response genes. Virulence was limited and mainly ICEKp/ybt-associated, and widespread predicted prophages did not carry priority determinants. These findings show that WGS can add mechanistic resolution to routine phenotypic surveillance by identifying clone-MGE configurations that warrant targeted monitoring in neonatal K. pneumoniae outbreaks in Zambia.