Retrospective evaluation of carbapenemase-producing Pseudomonas aeruginosa carriage and a proposed follow-up strategy: insights from a blaVIM-dominant setting

This is the authors' abstract. We don't add key points for this paper.

BACKGROUND: Follow-up strategies to determine the loss of carbapenemase-producing Pseudomonas aeruginosa (CPPA) carriage vary considerably, with a lack of evidence-based guidance to direct discontinuation of infection prevention measures. AIM: The aim of this study was to determine the duration of carriage (DOC), risk factors for longer carriage, positive follow-up screenings and the number and timing of negative screenings needed to identify loss of CPPA carriage. METHODS: This retrospective cohort study included adult CPPA carriers under active follow-up from 2010 to 2023. Active follow-up entailed a minimum of six bi-monthly screenings, with six consecutive negative screenings indicating loss of CPPA carriage. Cox regression models were used to estimate the median DOC and to identify risk factors for longer carriage; a logistic mixed-effect regression model was used to identify risk factors associated with positive follow-up screenings. FINDINGS: Sixty-eight patients were included, with 52 achieving loss of CPPA carriage. Analysis of maximum DOC results generated a median value of 606 days (95% confidence interval: 529-813 days) and a recurrence rate of 11.5%. Risk factors for longer carriage were first detection in a clinical sample and a positive first follow-up screening. A policy of five or six consecutive negative screenings generated false-negative rates of 13.5% and 11.5%, respectively. CONCLUSIONS: Postponement of the start of follow-up, for up to one year, should be considered for patients presenting with risk factors for longer CPPA carriage. This would facilitate a more economical approach to CPPA screening while minimizing the burden for patients. Follow-up may begin immediately for patients without any risk factors, and five consecutive negative screenings might be sufficient.