There are few data on risk factors, chosen therapy and healthcare utilization among US children with extended spectrum beta lactamase-positive urinary tract infection (ESBL UTI). We performed a multicenter case-control study on childhood ESBL UTI from Nov 2014-Feb 2017; herein we present preliminary data from a single Los Angeles County hospital.
We defined UTI per 2011 AAP guidelines and ESBL per CLSI specifications. ESBL (-) UTI controls were matched by sex and age. Descriptive and matched univariate analyses on medical record data (up to 6 months after index culture) were performed.
Among 893 urinary Enterobacteriaceae isolates, 28 were ESBL(+),of which 23 were included: 13 girls, 0-5yo; 4 girls, ≥6yo; and 6 boys, 0-5yo. Prior hospitalization (55 vs 78% for cases vs controls, respectively), prior receipt of systemic antibiotics (55 vs 38%), index hospitalization (39 vs 20%), mean length of stay (3.9 vs 3.6d), and medical comorbidity (44 vs 56%) did not differ significantly between groups. As well, several biosocial risk factors were similar in both groups, including: race, ethnicity, non-English-speaker, access to public benefits, international travel, non-US-birth, domestic violence/child abuse/neglect, and housing insecurity. Of cases and controls receiving any therapy, 16% and 96%, respectively, got empiric antibiotics to which the isolate was susceptible (p=0.001). After culture results were available, only 39% of cases and 96% of controls received effective agents (p=0.00002). Forty-two percent of cases had clinical improvement (within a mean of 2-3d), vs 43% of controls. Total treatment duration did not differ, and no deaths were recorded. In the 6 months after index UTI, groups did not differ in number of clinical encounters, proportion with documented follow-up, repeat urine tests, receipt of additional therapy, or prophylactic antibiotics. The proportions undergoing any GU-specific imaging were similar (62 vs 47%), but this imaging included modalities with ionizing radiation in 4 cases vs. none of the controls (p<0.05).
Our data suggest that clinical improvement occurs with initial (and potentially ineffective) empiric regimens, regardless of ESBL phenotype. The finding of more ionizing radiation exposure warrants additional study.
T. Van, None
S. Friedlander, None
S. H. Yeh, None