Longitudinal whole-genome based comparison of carriage and infection associated Staphylococcus aureus in northern Australian dialysis clinics

BACKGROUND: The study objective was to reveal reservoirs potentially leading to Staphylococcus aureus infections in haemodialysis clinic clients in the tropical north of the Australian Northern Territory (NT). This client population are primarily Aboriginal Australians who have a greater burden of ill health than other Australians. Reservoir identification will enhance infection control in this client group, including informing potential S. aureus decolonisation strategies. METHODS AND FINDINGS: The study participants were 83 clients of four haemodialysis clinics in the Darwin region of the NT, and 46 clinical staff and researchers who had contact with the clinic clients. The study design was longitudinal, encompassing swabbing of anatomical sites at two month intervals to yield carriage isolates, and also progressive collection of infection isolates. Swab sampling was performed for all participants, and infection isolates collected for dialysis clients only. Analysis was based on the comparison of 139 carriage isolates and 27 infection isolates using whole genome sequencing. Genome comparisons were based on of 20,651 genome-wide orthologous SNPs, presence/absence of the mecA and pvl genes, and inferred multilocus sequence type and clonal complex. Pairs of genomes meeting the definition of “not discriminated” were classed as defining potential transmission events. The primary outcome was instances of potential transmission between a carriage site other than a skin lesion and an infection site, in the same individual. Three such instances were identified. Two involved ST762 (CC1) PVL- MRSA, and one instance ST121 PVL+ MSSA. Three additional instances were identified where the carriage strains were derived from skin lesions. Also identified were six instances of potential transmission of a carriage strains between participants, including transmission of strains between dialysis clients and staff/researchers, and one potential transmission of a clinical strain between participants. There were frequent occurrences of longitudinal persistence of carriage strains in individual participants, and two examples of the same strain causing infection in the same participants at different times. Strains associated with infections and skin lesions were enriched for PVL and mecA in comparison to strains associated with long term carriage. CONCLUSIONS: This study indicated that strains differ with respect to propensity to stably colonise sites such as the nose, and cause skin infections. PVL+ strains were associated with infection and skin lesions and were almost absent from the carriage sites. PVL- MRSA (mainly CC1) strains were associated with infection and also with potential transmission events involving carriage sites, while PVL- MSSA were frequently observed to stably colonise individuals without causing infection, and to be rarely transmitted. Current clinical guidelines for dialysis patients suggest MRSA decolonisation. Implementation in this client group may impact infections by PVL- MRSA, but may have little effect on infection by PVL+ strains. In this study, the PVL+ strains were predominant causes of infection but rarely colonised typical carriage sites such as the nose, and in the case of ST121, were MSSA. The important reservoirs for infection by PVL+ strains appeared to be prior infections.