The article by Hellerstein¹ in this issue focuses on risk factors for urinary tract infection in children and reviews methods of prevention. Leaving aside the issue of circumcision and abnormal voiding patterns, this commentary will focus on the diagnosis of urinary tract infection in children and the selection of imaging procedures after diagnosis.

Infant girls are at a particularly high risk of urinary tract infection. The incidence of urinary tract infection is higher during the first year of life than at any other age in childhood.² Our initial study investigated the prevalence of urinary tract infections in febrile infants, with and without an apparent source of fever, who presented to the emergency department. Urine cultures obtained by catheter were positive in 5.3 percent of 945 infants with fever. The prevalence did not vary with age but was higher in girls than in boys (8.8 percent versus 2.5 percent; P< 0.0001) and was higher in white infants than in black infants (6.6 percent versus 3.6 percent; P< 0.05). When sex, race and degree of temperature were combined as risk factors, white females whose highest temperature had been at least 102.2°F (39°C) were found to be at particularly high risk of urinary tract infection (prevalence of 17 percent).³

Diagnosis of urinary tract infection is a major issue, and it should be based on both the results of urinalysis and the results of culture of a properly collected urine specimen. In older, toilet-trained children, a midstream clean-catch specimen is appropriate. Suprapubic aspiration may be necessary for pre-term infants, girls who cannot be catheterized because of labial adhesions, uncircumcised boys with tight foreskins and children with various anatomic abnormalities. In most other situations, a catheterized specimen is usually obtained.

Bacteria grown from a urine culture may arise from the following sources: (1) contamination outside of the urinary tract, (2) colonization of the distal urethra (contamination from within the urinary tract), (3) asymptomatic colonization of the bladder urine or (4) true urinary tract infection. The “enhanced urinalysis” has been a useful screening test to distinguish these conditions.

During the past years, we have evaluated uncentrifuged urine specimens obtained by catheter from young febrile children, and enumerated white blood cells per mm³ using a Neubauer hemocytometer. Of 3,257 cultures of urine, most (2,983 or 92 percent) showed no growth. Specimens with counts of bacteria between one and 49,000 colony-forming units (CFU) per mL were more likely to yield nonpathogens or mixed organisms than single pathogens. Specimens with counts of at least 50,000 CFU were most likely to yield single pathogens. Of all of the specimens in which there was a sterile culture, 97 percent had fewer than 10 white blood cells per mm³, and only 89 specimens or 3 percent had more than 10 white blood cells per mm³. In contrast, 90 percent of the urine specimens with more than 50,000 CFU per mL had at least 10 white blood cells per mm³.

Therefore, urinary tract infection is best defined by the presence of at least 10 white blood cells per mm³ on hemocytometer and cultures with growth of at least 50,000 CFU per mL on specimens obtained by catheter from young febrile children.⁴ For midstream clean-voided specimens, at least 10⁵ CFU per mL of a single urinary pathogen is usually regarded as a positive urine culture. These definitions will almost always discriminate true urinary tract infections from bacteriuria associated with contamination or colonization of the urinary tract (asymptomatic bacteriuria). It is important to recognize that, occasionally, low colony counts of bacteria in urine may be significant.

When urine specimens are obtained by suprapubic aspiration, growth of urinary pathogens in any number is considered significant (with the exception of 20,000 to 30,000 CFU per mL of coagulase-negative staphylococci). Even with catheterized specimens, the repeated recovery of a single species of bacteria in specimens with colony counts lower than 50,000 CFU per mL (especially in a symptomatic patient) should be regarded as representative of true infection. Colony counts of bacteria may fall below the range that is characteristic of infection when one of the following factors is present: (1) a bacteriostatic agent is present in the urine; (2) there is a rapid rate of urine flow with reduced incubation time; (3) there is an obstruction of the ureter that interferes with discharge of bacteria into the bladder; or (4) the infection is limited to areas of the kidney not directly accessible to renal tubules.

Imaging studies are part of the standard care after diagnosis of a first urinary tract infection in young children. As part of a recently completed clinical trial,⁵ 179 children aged one to 24 months with fever (higher than 100.9°F [38.2°C]) had a renal ultrasound examination and a renal scan with dimercaptosuccinic acid labeled with technetium 99m (DMSA scan) performed within 48 hours of diagnosis, a voiding cystourethrogram performed one month following diagnosis, and a repeat DMSA scan performed six months later. Results of renal ultrasound examination and DMSA scan at the time of presentation with urinary tract infection did not modify management in any patient.

Current widespread use of prenatal ultrasonography leads to identification of obstruction of the urinary tract in utero. Accordingly, selective rather than routine performance of ultrasound is recommended (i.e., in patients with persistent fever or abdominal findings and in those who have not had a prenatal ultrasound examination). In the study described above,⁵ a voiding cystourethrogram at one month and a DMSA scan six months later were useful for identifying patients with vesicoureteral reflux who required prophylactic antimicrobial therapy and patients with renal scarring. Patients with renal scarring may benefit from the early performance of urine culture during subsequent episodes of fever.