Legionnaires' disease is an important cause of hospital-acquired pneumonia and is caused by infection with the bacterium, Legionella. Since current typing methods often fail to resolve the infection source in possible nosocomial cases, we aimed to determine whether whole genome sequencing (WGS) could be used to support or refute suspected links between cases and hospitals. We focused on cases involving a major nosocomial-associated strain, L. pneumophila sequence type (ST) 1.
WGS data from 229 L. pneumophila ST1 isolates were analysed. These include 99 isolates from the water systems of 17 hospitals, 42 clinical isolates from patients with confirmed or suspected hospital-acquired infections, as well as isolates obtained from or associated with community-acquired sources of Legionnaires' disease.
Phylogenetic analysis demonstrated that all hospitals from which multiple isolates were obtained have been colonised by one or more distinct ST1 populations. However, deep sampling of one hospital also revealed the existence of substantial diversity and ward-specific microevolution within the population. Across all hospitals, suspected links with cases were supported with WGS, although the degree of support was dependent on the depth of environmental sampling and available contextual information. Finally, phylogeographic analysis revealed that hospitals have been seeded with L. pneumophila via both local and international spread of ST1.
WGS can be used to support or refute suspected links between hospitals and Legionnaires' disease cases. However, deep hospital sampling is frequently required due to the potential co-existence of multiple populations, existence of substantial diversity, and similarity of hospital isolates to local populations.