Extended-Spectrum Beta-Lactamase Surveillance in Children: To Screen or Not to Screen?

Background:

We have noted an increasing number of children at our institution with infection due to gram-negative bacteria producing extended-spectrum beta-lactamases (ESBLs).  Limited information is available on surveillance, isolation, and clearance protocols for these children.

Methods:

We performed prospective admission surveillance rectal swabs in our pediatric intensive care unit (PICU) and hematology/ oncology (HO) patients from June, 2013 – March, 2016. Discontinuation of isolation required negative results on at least 3 consecutive cultures (original site and stool/ rectal swab) taken > 1 week  apart, obtained after antibiotics effective against ESBL organisms had been discontinued for >48 hours.

Results:

There were 386 ESBL positive patients, of whom 210 (54%) were identified by surveillance culture only, resulting in a substantial increase in ESBL identified patients at our institution, of whom 37/386 (10%) resided in long-term care facilities. Medical comorbidities (MCM) were present in 152 (39%), HO diagnoses in 113 (29%), and 121 (31%) were healthy (H).  Most HO and MCM patients were identified as colonized, and most H patients as infected (Figure 1). Most colonized patients were identified from rectal screening (210/243=86%). Of HO patients, 51/113 (45%) were cleared per protocol; 26/51 (51%) had only a single surveillance culture positive, and 7/51 (14%) had recurrence (Figure 2).  Two of these resulted in bloodstream infection. Of MCM patients, 6/152 cleared per protocol and 1/6 recurred.  There were 12 patients with known previous rectal colonization who developed infection ≥3days later (4 from long-term care facilities and 7 HO patients).  There were 67/386 (17%) ESBL patients colonized with another multi-drug resistant organism (MDRO), of which 35 had a second ESBL organism (Figure3).

Conclusion:

Admission surveillance rectal swabs in our PICU and HO patients identified a large number of colonized patients.  A substantial proportion of HO patients had follow-up cultures, and were cleared per our protocol. A paucity of colonized patients progressed to infection, rendering broad admission screening of questionable value.  Recurrence was not uncommon, but rarely led to infection.  Colonization with another MDRO, usually another ESBL organism, occurred frequently.