Screening for Serious Bacterial Infections in Young Febrile Infants

The recent article by Baker and Bell¹ does not provide data that support the conclusion that febrile infants younger than 1 month are different from febrile infants aged 1 to 2 months. Their current study retrospectively applied some of the criteria previously analyzed prospectively in infants aged 1 to 2 months.² In the original study, infants were judged to be "screening-positive" if the infant had evidence of a bacterial infection on examination, an infant observation score higher than 10, or a laboratory value outside the defined reference range. Their definition of reference laboratory values included the following: white blood cell count of peripheral blood, fewer than 15,000 cells per cubic millimeter; findings from urinalysis, fewer than 10 white blood cells per high-power field and few or no bacteria detected by bright-field microscopy (with spun specimens used for both tests); cerebrospinal fluid values, fewer than 8 white blood cells per cubic millimeter in a nonbloody specimen and negative findings from Gram stain; and no evidence of a discrete infiltrate on chest x-ray film as determined by the attending physician (and subsequently confirmed by an attending radiologist). Because 1 infant was not accurately screened with these criteria, the criteria were modified to include a band-neutrophil ratio less than 0.2. Although it is not apparent from either study that a stool analysis in infants with a history of diarrhea is part of the screening criteria, in the recent article,¹ stool specimens were considered abnormal if they were positive for blood. It is also not clear why chest radiography is part of the routine evaluation of these infants because previous studies have not shown the utility of conducting this examination in infants without respiratory symptoms.^3- 4