Background: Nosocomial infections cause 4%–56% mortality in newborns. Several epidemiological studies have shown that transmission of opportunistic pathogens from the sink to the patient, including Pseudomonas aeruginosa, Stenotrophomonas maltophilia, and Serratia marcescens are associated with nosocomial infections in neonatal intensive care units (NICUs). In this project, we aimed to develop fast, accurate, and high-throughput multilocus sequence typing assays (HiMLST-Illumina) to detect opportunistic pathogens to assess their distribution in the sink environment of NICUs and their transfer to patients. Methods: Genome sequences of P. aeruginosa (n = 45), S. maltophilia (n = 23) and S. marcescens (n = 34) strains were retrieved from public genome databases to build their pangenomes, using the open-source PGAdb-builder server. The core genome was identified for each opportunistic pathogen and was searched for genes displaying the highest polymorphism. The minimal number of loci to include in a HiMLST-Illumina assay was determined by comparing topology of phylogenetic trees of concatenated loci based on genome similarity, computed as the average nucleotide identity (ANI) score. The primers used for HiMLST-Illumina schemes were designed in silico on a conserved domain and were tested on reference strains of each species. Results: Bioinformatics analyses showed that 3–4 loci (<300 base pairs per locus) distinguished strains with the same performances than ANI scores. The assays were tested using opportunistic pathogen isolates and environmental DNA originating from NICU sinks. The HiMLST-Illumina analysis of environmental DNA revealed the presence of at least 1 of the 3 studied opportunistic pathogens in 50% of sampled drains (n = 20). In a previous sampling, P. aeruginosa was isolated on selective culture media before and 48 hours after disinfection of a sink drain with chlorine. S. marcescens was also isolated from another sink 2 weeks after disinfection. Identification of the isolates was confirmed by HiMLST-Illumina analyses and will be typed to compare with clinical isolates. Conclusions: Initial in silico tests predict a high discriminating power of the HiMLST-Illumina method, suggesting that it would be possible to quickly identify strains of interest in a large number of samples. The power of this method is also in the possibility for molecular typing without a need for cultivation. Preliminary results suggest that sinks are readily colonized by opportunistic pathogens. This HiMLST-Illumina scheme will be applied in a 2-year intensive survey of NICUs in 3 hospitals in Montreal to evaluate the performance of new sink designs in limiting bioaerosol production and transmission of opportunistic pathogens to patients.
Funding: None
Disclosures: None