497 Risk Factors for Infection and Colonization with Antibiotic Resistant Gram Negative Bacilli in Hospitalized Children

Sunday, April 3, 2011
Trinity Ballroom (Hilton Anatole)
Sameer J. Patel, MD , Columbia University, New York, NY
Mandar Apte, MD, MPH , Columbia University, New York, NY
Elaine Larson, RN, PhD , Columbia University, New York, NY
Background: Infections with antibiotic resistant gram negative bacilli (GNB) are an emerging threat in pediatrics.  During hospitalizations children may be infected or colonized with multiple GNBs with varying antibiotic susceptibilities.

Objective: To describe the susceptibility patterns of GNB causing infection or colonization in pediatric patients with prolonged hospital stays across 1 or more admissions. To determine risk factors for acquisition of organism(s) with collective non-susceptibility to ≥ 3 antibiotic classes measured by clinical laboratory testing.

Methods: A case control study was performed involving patients £ 18 years admitted at a tertiary care children’s hospital from 2006-2008 who had ≥50 days of hospitalization over 1 or more admissions.   Cases were defined as patients with clinical cultures from any body site positive for GNB that collectively over the study period demonstrated non-susceptibility to ≥3 antibiotic classes.  Control patients were patients with no cultures, negative cultures, or cultures that demonstrated non-susceptibility to <3 antibiotic classes.  Antibiotic classes were beta-lactams (excluding carbapenems), aminoglycosides, fluoroquinolones, and carbapenems.  An organism was considered non-susceptible if all reported antibiotics within a class were resistant or intermediate. Multivariable logistic regression was used in a forward selection manner. The covariates of interest were age, gender, total length of stay (LOS), total admissions, transfer from another facility, and admission to the neonatal intensive care unit; presence of malignancy, cystic fibrosis, and GI tract abnormalities; number of admits with ICU stay and surgery; and receipt of antibiotics by class.

Results: 889 children were identified with total LOS ≥50 days; mean age in years, number of admissions, and total LOS were 3.2, 4.3, and105 days, respectively. The number of children with GNB non-susceptible to beta-lactams, aminoglycosides, fluoroquinolones, and carbapenems, were 55 (6.2%), 44 (4.9%), 87 (9.8%), and 51 (5.7%), respectively.  Thirty-four children (3.8%) had collective non-susceptibility to ≥3 classes, of which independent risk factors for positive cultures for collective ≥3 class non-susceptibility were total LOS (p<.0001; OR=1.011, 95% CI [1.006, 1.016]), number admissions with ICU stay (p<.0001; OR=1.629, 95%CI [1.32 , 2.00]), and receipt of aminoglycoside antibiotics (p=0.097; OR=1.62, 95%CI [1.12, 2.33]).

Conclusions: While relatively few patients had positive cultures for GNB with ≥3 class collective non-susceptibility, independent associations were found with total length of stay, ICU admissions, and receipt or aminoglycoside antibiotics.  Understanding patterns of GNB acquisition over time is important as colonizing GNB limit options for future empiric therapy.