Although the treatment of febrile infants younger than two months has been the subject of extensive investigation in the past, no studies have exclusively focused on those two to six months of age. Hsaio and colleagues aimed to determine the frequency of serious bacterial infections in febrile infants in this age group and the implications for management.

Infants 57 to 180 days of age who presented to a medical center with rectal temperatures higher than 100.2°F (37.9°C) were prospectively enrolled in the study. Each infant was evaluated using the Yale Observation Scale (YOS) and blood count with differential, blood cultures, urinalysis, and nasal swabs testing for viral infections. Other evaluations were performed on a case-by-case basis. Any apparent source for the fever was documented.

Of 448 eligible infants, 429 (96 percent) were enrolled. Forty-four (10.3 percent) had positive blood cultures with or without positive urine cultures. Although 58 infants had spinal taps, no cases of meningitis were found. Nasal swabs were positive for virus in 163 (39.5 percent) of the 413 swabs obtained. Of those who tested positive for viral illness, 4.9 percent had concurrent bacterial infection. Comparatively, 13.5 percent of infants who tested negative for the virus had bacterial infection.

An obvious source of illness was identified in 264 infants (61.5 percent). The rate of infection in those with an obvious source of fever was 6.1 versus 18.1 percent (P < .001) in those with no obvious source. In terms of laboratory tests, a higher mean white blood cell count correlated with higher incidence of bacterial infection. C-reactive protein was higher in infants with an infection than in those without; however, there was a wide range of values in those without. Febrile infants with and without bacterial infection differed by a YOS score of 1.4. Although infants in the group labeled “very ill appearing” had the highest rate of infection, 34 of 44 infants with bacterial infection were labeled “well appearing.” Of the 178 infants who had a record of circumcision status, 36 percent of uncircumcised males had bacteriuria compared with 1.6 percent of circumcised males. Fever duration, but not height of fever correlated with likelihood of infection.

Incidence of invasive bacterial disease was low, occurring in only 0.97 percent of infants. However, three of the four infants identified with bacteremia were well appearing. Also, one in 10 well-appearing infants had an infection, indicating that clinical appearance is not a reliable indicator. Circumcision was associated with a lower risk of bacteriuria, and there was a 4.9 percent concurrence rate between positive viral swabs and bacteriuria. The ranges of the acute-phase reactants were too wide to be reliable predictors of infection.

The authors conclude that although there are few clues to determine which infants have an infection, they recommend evaluating for bacteriuria in infants two to six months of age because it is common even in infants who appear well.