This noninferiority trial involved children aged 2 months to 10 years diagnosed with symptomatic UTI that exhibited clinical improvement after the first 5 days of antimicrobials. From there, they received either placebo (short-course group) or more antimicrobials (standard-course group). The standard-course group was associated with lower rates of treatment failure, asymptomatic bacteriuria, and positive urine culture after days 11-14. However, treatment failure rate was still low in the short-course group, and rates of UTI after day 11 to 14 were similar between groups, suggesting that differences in bacteriuria or positive urine culture did not contribute to subsequent UTI development. Children with fever at the time of diagnosis did not show a significant difference in treatment failure between study arms. Furthermore, most children whose therapy failed were afebrile and not at risk for kidney scarring. Given that approximately 67 children need to be treated with standard-course therapy to prevent 1 febrile UTI and that scarring occurs in approximately 1 in 7 children with febrile UTI, approximately 469 children would need to be treated to prevent 1 child from developing kidney scarring.

This trial has several strengths: a large, diverse population of children; enrollment of a sizeable proportion of children with fever at presentation (38%); use of stringent diagnostic criteria for UTI; enrollment of children treated with various antimicrobials;monitoring for targeted antimicrobial resistance in stool commensals; andmodest attrition. Limitations stem from slight imbalance between treatment groups in number of children excluded from the primary analysis, reduced power for subgroup analyses, assessment of emergence of antimicrobial resistance only for E coli and K pneumoniae strains, lack of detailed data on the societal costs and benefits of each treatment strategy, absence of data on adherence to the originally prescribed antimicrobials on days 1 to 5, lack of strain-level data on recovered uropathogens, and lack of data on outcomes, such as kidney scarring.