Selective strategy for urethral catheterization in febrile young girls to confirm urinary tract infection diagnosis.

BACKGROUND: Urinary tract infections (UTIs) are a common source of bacterial infections in childhood. Making a proper diagnosis is important but requires invasive urine collection techniques. We aimed to derive a clinical decision rule to identify non-toilet-trained febrile girls at high risk for UTIs to restrict urethral catheterizations (UCs) to this high-risk group of patients.

METHODS: We included all non-toilet-trained girls with a positive microscopic urinalysis from urine collected by sterile bag in a prospective cohort study to derive a model to predict UTI assessed by urine culture from UC.

RESULTS: Thirty-seven patients were included. Absence of another source of fever on examination and the child's unusual behaviour were found to be independent predictors of UTI. The corresponding model offered an 85% sensitivity [95% confidence interval (CI): 56-96], with a 59% specificity (95% CI: 30-83) for UTI. The internal cross-validation by bootstrap led to an 85% sensitivity (95% CI: 68-100), and a 59% specificity (95% CI: 35-83).

CONCLUSION: We derived a clinical decision model to selectively identify young febrile girls at high risk for UTI with a positive microscopic analysis and propose UC with an 85% sensitivity, which would avoid approximately 60% of unnecessary UCs; although further validation is necessary before daily clinical use.