Discussion
A recent American Academy of Pediatrics consensus statement[4] on pediatric UTI excluded infants 0–2 months of age from consideration due to insufficient published data characterizing this group. Prior analyses estimated the prevalence of UTI in febrile young infants within the context of larger age groups: ages 0–2 months (10–13.6%)[7,9,10] and ages 0–3 months (7–9%).[11,12] Among febrile infants ages 0–3 months, 7.5% of girls had UTI; whereas UTI was present in 2.4% of circumcised and up to 20% of uncircumcised boys.[11]
Few prior studies examining outpatient-evaluated febrile neonates specifically analyzed UTI; several series examining +SBI rates reported UTI prevalence < 5%.[1,2] By contrast, we found a much greater overall UTI rate of 15.4% in these very young febrile infants (approximately 1 in 6 cases). This higher rate may be due in part to more comprehensive criteria we used to diagnose UTI. No ideal microbiologic standard has been devised to accurately define UTI in this age group. Although nearly 80% of positive urine cultures in our series grew a uropathogen at a concentration ≥50,000 CFU/mL, we sought to maximize inclusion of all "treatment-indicated" cases by diagnosing the infection whenever a febrile neonate with pyuria had a positive UC for a single uropathogen at colony counts ≥10,000 CFU/mL.[5–7]
A prior prospective study[7] applying similar microbiologic criteria for diagnosing UTI showed 11% of febrile outpatient-evaluated neonates had UTI. Being an offshoot of a bronchiolitis study, UTI sampling only occurred during peak respiratory season (between October through March), which precluded an accurate assessment of prevalence; and limited their conclusions due to skewed proportions analyzed. A relative strength of our analysis is that hospital protocol dictated a stereotyped evaluation of each patient (complete sepsis evaluation) which captured for analysis 97% of all febrile neonates who presented to our ED during the 10-year study period; providing a more comprehensive annual prevalence rate.
In accord with other series[7,13–16] we documented, E. coli was the predominant uropathogen (71%) and UTI was more common in males (73%) vs. females.[9,15,17] Similarly consistent with prior series of febrile young infants with +SBI, higher degrees of fever (>=102°F) were absent in nearly two-thirds of febrile neonates with UTI.[1,2,16]
Prior studies[9,13–15] documented prevalence rates of associated bacteremia in infants with UTI ranging as high as 16–31%. By contrast, our results showed only 4% of neonates had urosepsis (none had bacterial meningitis), which accords with other studies documenting relatively lower rates ranging from 0 to 6%.[18–20]
Outpatient evaluation variables were relatively insensitive for identifying UTI risk. Only 39% of patients had peripheral blood leukocytosis (CBC total WBC count ≥15,000/mm3). A urine dipstick test positive for leukocyte esterase or nitrite was only 79% sensitive, lower than previously reported for relatively older children (up to 90–100%).[4,21] Reliance on urine dipstick test results to determine whether to perform a urine culture would have resulted in missed diagnosis of 21% of UTI cases in our cohort. Similarly, microscopic urinalysis was relatively insensitive at identifying those with underlying UTI. Our laboratory uses standard technique for urinalysis; perhaps usage of enhanced urinalysis technique would have improved detection of pyuria in these patients.[21] These findings reiterate that outpatient tools to identify UTI risk in febrile young infants are imperfect[7] and emphasizes the need to culture the urine in all cases.[22]
There is little published data establishing specific rates of RUS abnormalities in febrile neonates with UTI.[14,20,23–26] Such abnormalities in relatively older children ages 2–24 months with a first UTI are identified in approximately 12–15% of cases.[4,24] Some prior neonatal UTI series performing renal imaging were limited by surveying only those in a neonatal intensive care unit,[16,26] examining a select (nonconsecutive) group of neonates with UTI[15,20] or reporting neonatal statistics as a subset within the context of wider age ranges surveyed.[20,23,24]
Several studies specifically reported RUS findings in outpatient-evaluated young infants with UTI. One[15] examining 45 selected males aged 0–8 weeks with UTI showed hydronephrosis in 24%. Another[20] examining 64 selected neonates with UTI found 20% with abnormal RUS findings and 20% with VCUG-diagnosed VUR.
By contrast, our study of consecutive outpatient-evaluated febrile neonates (95% had RUS performed) found nearly half with UTI had a renal anatomic abnormality; specifically, pelviectasis (20%) and hydronephrosis (27%). Although the finding of isolated pelviectasis usually requires no intervention and has a high rate of spontaneous resolution,[27] recognizing this underlying abnormality identifies those who require close urologic monitoring and evaluation for UTI with repeat febrile illnesses. Of patients with hydronephrosis who had a VCUG performed, nearly one-fourth had VUR; of microbiologic interest is that in 80% with VUR, the uropathogen was a non-E. coli Gram-negative organism (Klebsiella or Enterobacter).
Pediatr Infect Dis J. 2014;33(4):342-344. © 2014 Lippincott Williams & Wilkins
