Michaël Racodon, Anne Véronique Davin, Valérie Leroux
Probiotic cleaning was associated with a lower crude incidence rate of microbiologically documented HAI episodes. Given the non-randomised design, potential residual confounding, and absence of environmental sampling, causal and mechanistic conclusions cannot be drawn. Controlled multicentre studies are needed to confirm these findings.
BACKGROUND: The hospital environment contributes to the transmission of healthcare-associated infections (HAIs). Probiotic-based cleaning represents an ecological approach aimed at modulating the surface microbiota through bacterial competition rather than eliminate it.
OBJECTIVE: To assess whether probiotic cleaning was associated with a change in the incidence density of microbiologically documented healthcare-associated infection (HAI) episodes compared with conventional biocide cleaning.
METHODS: A single-centre, quasi-experimental before-and-after study compared two consecutive 12-month cleaning periods. Potential HAI episodes were identified through monthly retrospective extraction of routine laboratory data and adjudicated from medical records using European Centre for Disease Prevention and Control (ECDC) definitions. Incidence densities were expressed per 1,000 patient-days and compared using an exact Poisson rate test.
RESULTS: The study included 5,753 hospitalisations, representing 37,511 patient-days during the biocide period and 32,453 during the probiotic period. Among the 475 laboratory-identified potential HAI episodes, 188 episodes during the conventional biocide control period and 117 episodes during the probiotic intervention period met the relevant ECDC criteria for an HAI. Incidence density was lower during the probiotic period (3.61 versus 5.01 episodes per 1,000 patient-days; crude incidence rate ratio, 0.72; 95% confidence interval, 0.57-0.91; p = 0.005), corresponding to an approximately 28% lower crude rate.
CONCLUSION: Probiotic cleaning was associated with a lower crude incidence rate of microbiologically documented HAI episodes. Given the non-randomised design, potential residual confounding, and absence of environmental sampling, causal and mechanistic conclusions cannot be drawn. Controlled multicentre studies are needed to confirm these findings.