Urinary tract infections in febrile children: Changing spectra of pathogenic bacteria and antibiotic susceptibilities?

Aim To compare the spectra of pathogens causing febrile urinary tract infections (UTI) in children, treatment and antimicrobial susceptibility between 2004-2006 and 2007-2009. Methods UTI were identified from a cohort study of febrile children younger than 5 years presenting to a large tertiary children's hospital's emergency department with febrile illnesses. We compared pathogenic profiles, antibiotic choices and susceptibilities between 2004-2006 and 2007-2009 and tested for differences using χ2 and Fisher's exact tests. Antibiotic choice was compared with national therapeutic guideline recommendations for UTI in children (oral cotrimoxazole, cephalexin or amoxycillin-clavulanate or intravenous gentamicin plus ampicillin). Results There were 539 (2.71%) confirmed UTI from 19 889 febrile illnesses in 2004-2006 and 654 (2.99%) confirmed UTI from 21 846 febrile illnesses in 2007-2009. There was no difference in the frequency of the isolated pathogens by period: Escherichia coli (69.2 vs. 69.7%, P = 0.85), Proteus mirabilis (7.9 vs. 7.2%, P = 0.66) and Klebsiella species (6.2 vs. 4.7%, P = 0.25). National therapeutic guideline recommendations were followed in 277 of 539 (51.4%) versus 318 of 654 (48.6%) (P = 0.34). Oral antibiotics were given in 20.6 versus 18.9%. There was no difference in extended spectrum beta lactamase (1.5 vs. 1.7%, P = 0.82) or other antibiotic susceptibilities (e.g. E. coli: cotrimoxazole = 75.9 vs. 75.2%, P = 0.8). Conclusions Overall, approximately 3% of febrile illnesses were due to UTI, but we found no change in the spectrum of pathogens or antibiotic susceptibility patterns, including extended-spectrum beta-lactamase, with time. In both time periods, treatment followed therapeutic guidelines approximately half the time, and most pathogens were susceptible to oral antibiotics, but they were infrequently used.