NATURAL HISTORY OF URINARY TRACT INFECTIONS IN FEBRILE INFANTS 0 TO 3 MONTHS OLD: INFERENCES FROM THE PROS FEBRILE INFANT STUDY TB Newman, JA Bernzweig, JI Takayama, S Finch, GM Spitalny, RH Pantell. Epidemiology and Biostatistics, UCSF, San Francisco, CA; Pediatrics, UCSF, San Francisco, CA; School of Public Health, University of California, Berkeley, CA; and Pediatric Research in Office Settings (PROS), Center for Child Health Research, American Academy of Pediatrics, Elk Grove Village, IL. Presented as a platform presentation at the 2000 Pediatric Academic Societies? Annual Meeting.

BACKGROUND: The outcome of infants with UTIs not diagnosed at the initial visit of a febrile infant is relevant for decisions about whether all such infants should have urine testing at their initial visit.

OBJECTIVE: To draw inferences about the natural history of urinary tract infections (UTIs) in febrile infants.

DESIGN/METHODS: Prospective cohort study in 219 practices in the American Academy of Pediatrics' Pediatric Research in Office Settings (PROS) research network. The subjects were 3066 infants 0-3 months with T 38°C in the office or in the previous 24 hours at home, managed according to the clinical judgment of the practitioner caring for the child. The main outcome variable was delayed diagnosis of UTI or UTI with bacteremia

RESULTS: Of the 1660 infants (54%) who had urine tested at the initial visit, 143 (8.6%) had UTIs. These infants formed the basis for a logistic model predicting the probability of UTI in infants not initially tested. The average predicted probability of UTI among the 1406 infants not initially tested was 6.9%, reflecting their slightly lower prevalence of risk factors for UTI than those initially tested (whose probability of UTI was 8.6%). Follow-up to resolution of illness was obtained on 1330 of the 1406 infants (95%). Of these, 811 were initially managed without antibiotics. The average predicted probability of UTI among these 811 was 6.7%. Thus, about 811*.067 = 54 UTIs would have been expected in these 811 infants, had they been tested at the initial visit. Only 2 infants were subsequently diagnosed with UTIs. Neither infant was bacteremic; both were treated and recovered uneventfully. The remaining 809 infants recovered without a UTI being diagnosed.

CONCLUSION: Acute symptoms of many UTIs in febrile infants 0-3 months old appear to resolve spontaneously. Although long term effects of these undiagnosed UTIs are not known, this study provides some support for many PROS practitioners' current practice of ordering urine tests selectively. Testing should focus on the youngest and sickest infants, and on those most likely to have a UTI.