Serious bacterial infections in febrile infants in the post-pneumococcal conjugate vaccine era.

OBJECTIVES: The objective was to identify the epidemiology of serious bacterial infections (SBI) and the current utility of obtaining routine complete blood counts (CBC) and blood cultures to stratify infants at risk of SBI, in the study population of febrile infants in the post-heptavalent pneumococcal conjugate vaccine (PCV7) era.

METHODS: A cohort study with nested case-controls was undertaken at a tertiary care military hospital emergency department (ED) from December 2002 through December 2003. Irrespective of clinical findings at the initial encounter, patients were included if they were under 3 months of age and had a home or ED temperature of >or=100.4 degrees F or if they were between 3 and 24 months of age with a temperature of >or=102.3 degrees F. Data abstracted included age, temperature, peripheral white blood cell (WBC) count, and discharge diagnosis. Culture (blood, urine, and cerebrospinal fluid [CSF]) and chest radiograph (CXR) results were obtained through review of the electronic hospital archives. SBI was defined as pneumonia, urinary tract infection (UTI), meningitis, or bacteremia.

RESULTS: A total of 985 children aged 0 to 24 months were enrolled. Fifty-five percent were male, the median age was 12 months (interquartile range = 8-17 months), and 79% had received at least one PCV7. A total of 132 cases of SBI were identified in 129 infants (13.1%): 82 pneumonias, 45 UTI, five bacteremias, and no cases of bacterial meningitis. The frequency of bacteremia was 0.7%. No statistical difference was detected in the WBC count between the SBI and non-SBI groups (13.8 +/- 5.8 and 11.7 +/- 5.6, respectively; p = 0.055). No readily available WBC cutoff on the receiver operating characteristic (ROC) curve proved to be an accurate predictor of SBI. No statistical difference was detected in mean temperature between the SBI and non-SBI groups (103.3 +/- 1.2 and 103.2 +/- 1.2 degrees F, respectively; p = 0.26), nor was there a difference noted when groups were broken down by age or height of fever.

CONCLUSIONS: The WBC count and height of fever were not found to be accurate predictors of SBI in infants age 3 to 24 months. UTI and pneumonias made up the vast majority of SBI in this population of infants. The overall bacteremia frequency was well below 1%. This calls into question the continued utility of obtaining routine complete cell counts and blood cultures in the febrile infant in the post-PCV7 era.