Enhanced urinalysis improves identification of febrile infants ages 60 days and younger at low risk for serious bacterial illness

Objective. Investigators have sought to establish "low-risk" criteria to identify febrile young infants who can be observed safely without antibiotics. Previous studies have used criteria for standard urinalysis to identify suspected urinary tract infection; however, cases of urinary tract infection have been missed. Enhanced urinalysis, using hemocytometer cell count and Gram stain performed on uncentrifuged urine, has been shown to have greater sensitivity and negative predictive value than standard urinalysis. The objective of this study was to evaluate the ability of criteria that incorporate enhanced urinalysis to identify febrile young infants who are at low risk for serious bacterial illness (SBI). Methods. Institutional guidelines were established in 1999 to evaluate in a retrospective cohort study infants who were less than or equal to 60 days of age with temperature greater than or equal to 38.0 degreesC. "Low-risk" criteria included 1) well appearance without focal infection (excluding otitis media); 2) no history of prematurity, illness, or previous antibiotics; 3) peripheral white blood cell count (WBC) between 5 and 15 000/mm(3); 4) absolute band count less than or equal to 1500/mm(3); 5) cerebrospinal fluid WBC less than or equal to5/mm(3) with a negative Gram stain; 6) enhanced urinalysis with WBC less than or equal to9/mm(3) with a negative Gram stain; 7) stool WBC <5/high power field in infants with diarrhea; and 8) chest radiograph without lobar infiltrate(s) in infants with respiratory signs or symptoms. SBI was defined as a lobar infiltrate on chest radiograph or presence of a bacterial pathogen in blood, urine, cerebrospinal fluid, stool, or culture obtained from the soft tissue. The hospital records of all infants who presented to the emergency department for evaluation of fever after January 1999, including those who did not meet low-risk criteria, were reviewed; data were collected regarding history, physical examination, laboratory test results, treatment, and clinical course. Results. During the study period, 434 infants presented to the emergency department for evaluation of fever. Thirty patients were excluded from additional analysis because of incomplete data; 60 patients were identified immediately as "not low risk" on the basis of history or physical examination. Of the 344 remaining infants, 127 were identified as "low risk" on the basis of laboratory criteria; 83 (65.4%) were observed without antibiotics. None of the "low-risk" infants had an SBI. A total of 217 well-appearing infants were classified as "not low risk" on the basis of laboratory criteria; 28 (12.9%) had an SBI. The overall incidence of SBI in infants with complete data was 10.1%, whereas the incidence of SBI in all "not low-risk" infants was 14.8%. The negative predictive value for the "Pittsburgh" criteria was 100% (95% confidence interval: 96.7%-100%); the sensitivity was 100% (95% confidence interval: 89.7%-100%). Conclusions. The application of low-risk criteria using enhanced urinalysis improves identification of infants who are at low risk for SBI.