An interventional quasi-experimental study to evaluate the impact of a rapid screening strategy in improving control of nosocomial extended-spectrum beta-lactamase-producing Enterobacterales and carbapenemase-producing organisms in critically ill patients

Rapid molecular tests could accelerate the control of extended-spectrum beta-lactamase-producing Enterobacterales (ESBL-PE) and carbapenemase-producing organisms (CPO) in intensive care units (ICUs).This interventional 12-month cohort study compared a loop-mediated isothermal amplification (LAMP) assay performed directly on rectal swabs with culturing methods (control period, 6 months), during routine ICU screening. Contact precautions (CP) were implemented for CPO or non-E. coli ESBL-producing Enterobacterales (nEcESBL-PE) carriers. Using survival analysis, we compared the time intervals from admission to discontinuation of unnecessary preemptive CP among patients at-risk and the time intervals from screening to implementation of CP among newly identified carriers. We also compared diagnostic performances, and nEcESBL-PE/CPO acquisition rates. This study is registered, ISRCTN 23588440.We included 1043 patients. During the intervention and control phases, 92/147 (62.6%) and 47/86 (54.7%) of patients at-risk screened at admission were candidates for early discontinuation of preemptive CP. The LAMP assay had a positive predictive value (PPV) of 44.0% and a negative predictive value (NPV) of 99.9% for CPO, and 55.6% PPV and 98.2% NPV for nEcESBL-PE. Due to result notification and interpretation challenges, the median time from admission to discontinuation of preemptive CP increased during the interventional period from 80.5 (95% CI 71.5-132.1) to 88.3 (95% CI 57.7-103.7) hours (p = 0.47). Due to the poor PPV, we had to stop using the LAMP assay to implement CP. No difference was observed regarding the incidence of nEcESBL-PE and CPO acquisition.A rapid screening strategy with LAMP assays performed directly on rectal swabs had no benefit for infection control in a low-endemicity setting.