Whole-genome sequencing establishes persistence of biofilm-associated Pseudomonas aeruginosa detected from microbiological surveillance of gastrointestinal endoscopes

Background: An increased incidence of Pseudomonas aeruginosa in microbiological surveillance (MS) cultures from gastrointestinal endoscopes was detected between March 2020 and March 2023 in Tan Tock Seng Hospital Singapore. Aim: To describe the use of whole-genome sequencing (WGS) in this investigation. Methods: WGS was performed for all P. aeruginosa isolates with pairwise comparison of isolates to assess for genomic linkage. Comprehensive review of reprocessing practices and environmental sampling was performed. Findings: Twenty-two P. aeruginosa isolates were detected from endoscopic MS cultures. Fifteen (68%) isolates were available for WGS. Eighteen pairwise comparisons of isolates were made, of which 10 were found to be genomically linked. One endoscope had P. aeruginosa repeatedly cultured from subsequent MS that were genomically linked and persistent despite repeat endoscopic reprocessing, establishing the persistence of biofilm that could not be eradicated with routine reprocessing. All P. aeruginosa isolates cultured from other different endoscopes were genetically distinct. Investigation into reprocessing practices revealed the use of air/water valves connected to endoscopes during clinical use. Inspection of these valves revealed the presences of cracks and tears. All other environmental samples were negative. Conclusion: The WGS findings helped to deprioritize common source contamination and supported the hypothesis of biofilm build-up within endoscopes, leading to repeatedly positive MS cultures that were genomically linked. This was possibly related to incomplete reprocessing of the damaged air/water valves, resulting in biofilm build-up. All faulty valves were changed and subsequently cleaned separately with ultrasonic cleaning followed by sterilization which resolved this incident. (c) 2024 The Healthcare Infection Society. Published by Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.