Is urine culture necessary to rule out urinary tract infection in young febrile children?

OBJECTIVE: To determine whether the absence of pyuria on the enhanced urinalysis can be used to eliminate the diagnosis of urinary tract infection, avoiding the need for urine culture and sparing large health care expenditures.

DESIGN: Results of an enhanced urinalysis (hemocytometer counts and interpretation of Gram-stained smears) performed on uncentrifuged urine specimens obtained by catheter were correlated with urine cultures in young febrile children at the Children's Hospital of Pittsburgh Emergency Department. In a group of 4253 children (95% febrile) less than 2 years of age, pyuria was defined as > or = 10 white blood cells/mm3, bacteriuria as any bacteria on any of 10 oil immersion fields in a Gram-stained smear and a positive culture as > or = 50,000 colony-forming units/ml. A subgroup of 153 children with their first diagnosed urinary tract infection were enrolled in a separate treatment trial, acute phase reactants (peripheral white blood cell count, erythrocyte sedimentation rate and C-reactive protein) were obtained and 99Tc-dimercaptosuccinic acid renal scans were performed.

RESULTS: The presence of either pyuria or bacteriuria and the presence of both pyuria and bacteriuria have the highest sensitivity (95%) and positive predictive value (85%), respectively, for identifying positive urine cultures. Because a white blood cell count in a hemocytometer is the technically simpler component of the enhanced urinalysis, we chose to analyze the false negative results and achievable cost savings of using pyuria alone as the sole criterion for omitting urine cultures. If in this study urine cultures had been performed only on specimens from children who had pyuria or were managed presumptively with antibiotics, cultures of 2600 (61%) specimens would have been avoided. Twenty-two of 212 patients with positive urine cultures would not have been identified initially. However, based on interpretation of acute phase reactants, initial 99Tc-dimercaptosuccinic acid scan results, response to management and incidence of renal scarring 6 months later, 14 of the 22 patients most likely had asymptomatic bacteriuria and fever from another cause. The remaining 8 patients probably had early urinary tract infection.

CONCLUSIONS: The analysis of urine samples obtained by catheter for the presence of significant pyuria (> or = 10 white blood cells/mm3) can be used to guide decisions regarding the need for urine culture in young febrile children.