Admission Screening for ESBL-producing Enterobacteriaceae (ESBL-E): Impact on Nosocomial Transmission Over a 6 Year Period

Background: ESBL-E infections are associated with increased morbidity, mortality and length of stay, are increasing in prevalence and may be contributing to the emergence of carbapenem resistant Enterobacteriaceae.  Patients with ESBL-E are more likely colonized than infected. Thus, admission screening may be an important strategy to reduce nosocomial ESBL-E transmission.

Objective: To examine the impact of risk factor based admission screening for ESBL-E colonization at a tertiary care hospital in Toronto, Canada in controlling nosocomial transmission of ESBL-E. 

Methods: We retrospectively examined all inpatients with clinical isolates or admission screens positive for ESBL producing E. coli, K. pneumoniae and K. oxytoca identified at Mount Sinai Hospital between 2004 - 2009.  Rectal swabs were plated on MacConkey agar with cefpodoxime (2μg/mL) and clinical isolates were screened with cefpodoxime (VITEK2).  Disk diffusion phenotypic confirmation (ceftriaxone, ceftazidime and aztreonam, plus/minus clavulanic acid) was performed on Mueller-Hinton agar.  Risk factor based admission rectal screening was initiated in 2004 but not performed routinely until late 2005.  All patients admitted to ICU or general medicine and surgical patients with risk factors for ESBL-E colonization (i.e. transfer from an acute or long term care facility, hospitalization within 6 months or a history of colonization with an antibiotic resistant organism) were screened.  In 2007, screening was extended to obstetrical patients and those admitted to hospital within 2 years.  All patients with ESBL-E were managed in contact precautions.  Nosocomial acquisition was defined as a clinical isolate positive for ESBL-E >72 hours after admission where admission screening was negative or not performed. 

Results: Between 2004 - 2009, 347 ESBL-E were identified from clinical specimens (79% E. coli, 12% K. pneumoniae, 9% K. oxytoca), of which 32% were nosocomial.  Total clinical ESBL-E incidence rose 64% (0.42 to 0.69/1000 pt days) from 2004 - 9. Nosocomial incidence of clinical isolates decreased 52% (0.27 to 0.13/1000 pt days) over the same period (Fig 1).  Changes in screening policy led to increased admission screening, from 3450 swabs in 2004 to 10675 in 2009.  Only 2.5% of patients with a positive admission screen had a subsequent ESBL-E clinical culture.  12 episodes of ESBL-E bacteremia with onset >72h after admission were identified: admission screening was positive in 2, negative in 8, and not performed in 2.

Conclusions:   We demonstrated a reduction in nosocomial ESBL-E over time coincident with the implementation of ESBL-E screening at our hospital despite an increase in total ESBL-E patients identified, suggesting that screening contributed to a reduction in nosocomial transmission of ESBL-E.  Admission screening failed to identify most patients that subsequently developed ESBL-E bacteremia.