Fever in Infants: Assessing Variability in Sepsis Evaluation

Background:Fever is a common presentation of infants resulting in frequent medical visits. Since fever may be the sole sign of invasive bacterial infection (IBI) in infants less than 3 months of age, invasive testing is often performed. Many physicians are guided by standardized criteria, which were created to aid in determining those at low risk of IBI. Though these criteria exist, there is limited guidance regarding appropriate testing in the first month of life and wide variability in practice during the first 90 days. An American Academy of Pediatrics national quality improvement collaboration, Reducing Excessive Variability in Infant Sepsis Evaluation, is standardizing management of these infants. This study evaluates current institutional practice in assessing febrile infants.

Methods:Retrospective chart review of well-appearing previously healthy term infants with no obvious source of fever on initial examination between the ages of 0-90 days presenting with documented or reported fever to either a tertiary emergency department or inpatient hospital, with specific International Classification of Diseases codes over 1 year period. The infants were then separated into three groups: 0-28, 29-60 and 61-90 days.

Results:Of 83 infants meeting criteria, 10% had IBI with 75% of these being urinary tract infection. Evaluation with complete blood count (CBC), blood culture, urinalysis (UA) and urine culture varied between groups from 84%, 87% and 29% respectively. Within this latter group, 75% were underimmunized. CBC results were abnormal in 64% of all infants with leukopenia the most common abnormality. Of those with bacterial infection and where CBC was obtained, 50% had leukopenia and 50% had normal white blood cell (WBC) count. UA collection differed between the groups from 88%, 87% and 68% and lumbar puncture attempts performed in 84%, 30% and 4%. CXR was obtained in 27% of infants and all were negative; 40% of these infants that underwent imaging were asymptomatic.

Conclusion:Most criteria rely on leukocytosis to identify high risk for IBI; infants with IBI in this study had leukopenia or normal WBC counts. Sepsis evaluation in febrile infants varies tremendously and an updated guideline for identifying IBI could minimize unnecessary imaging, laboratory testing and unwarranted antibiotic therapy.