Impact of COVID-19 isolation measures on ICU microbial resistance dynamics: simulation-based statistical modeling analysis.

BACKGROUND

The transmission of antibiotic-resistant bacteria in intensive care units (ICUs) poses a significant challenge to infection control and patient safety. While direct patient-to-patient transmission is well documented, the relative contributions of endogenous bacterial selection and cross-transmission remain uncertain.

METHODS

This retrospective study analyzed microbiological data from two ICUs at St. Anne's University Hospital in Brno, Czech Republic, between 2018 and 2021. Machine learning algorithms and random simulation models were employed to evaluate clustering patterns of resistant bacterial detections and to distinguish between exogenous cross-transmission and endogenous bacterial acquisition. Bacterial findings were compared across three epidemiologically distinct periods-precovid, covid, and intercovid-characterized by differing hygiene protocols and patient populations. The study assumes that the historically unprecedented hygiene measures during the COVID-19 pandemic substantially reduced horizontal cross-transmission, thereby providing a unique opportunity to estimate the relative contributions of exogenous transmission and endogenous acquisition under routine ICU conditions.

RESULTS

The prevalence of Pseudomonas aeruginosa (PSAE) was four times higher during the covid period than precovid and remained elevated in the intercovid period. Stenotrophomonas maltophilia detections tripled during covid, while Klebsiella pneumoniae and Escherichia coli resistant to cefotaxime doubled. The proportion of first bacterial detections occurring after 48 h of ICU admission was significantly higher during covid. Clustering analysis revealed no significant deviation from random distribution for most bacteria, except for PSAE, which exhibited non-random clustering, particularly in the intercovid period. Stenotrophomonas maltophilia showed a highly uneven distribution between the two ICUs, suggesting long-term environmental persistence.

CONCLUSION

Our findings suggest that antibiotic selection pressure is the primary driver of resistant bacteria acquisition in ICUs, while direct cross-transmission appears to play a limited role. However, environmental persistence may contribute to the recurrent detection of Stenotrophomonas maltophilia, emphasizing the need for enhanced decontamination strategies.