P2.151: Association of chest radiograph with antibiotic usage in febrile infants 2 to 6 months of age

Background: Chest radiograph (CXR) use in febrile infants varies considerably. Our aim was to examine the association between CXR and antibiotic use in the Emergency Department (ED).

Methods: We performed a five-center retrospective cross-sectional study of previously healthy 2 to 6-month-old infants who presented to the ED with a temperature ≥38.0C. Using generalized estimating equations to account for clustering by site, we measured the association between CXR and antibiotic prescription (in hospital or after discharge) with adjustment for clinical severity (age, peak temperature, heart rate, respiratory rate, oxygen saturation). For infants with an available CXR report, we classified as definite, possible, or no pneumonia and evaluated proportions receiving antibiotics by CXR result. For infants without pneumonia, we defined a suspected bacterial infection based on presence of pyuria or by discharge diagnosis codes.

Results: Of 21,150 febrile infants, 4,212 (19.9%) had a CXR (range by site 15.8 to 22.7%) and 5,603 (26.5%) received an antibiotic (range 20.5% to 46.7%). After adjustment, CXR was associated with antibiotic use (adjusted odds ratio 1.79, 95% confidence interval 1.49, 2.13). For the 2,871 children with an available CXR report, 120 (4.2%) had definite, 328 (11.4%) probable and 2,423 (84.4%) no pneumonia. Of the 2423 without pneumonia, 891 (36.7%) were prescribed antibiotics of whom 594 (66.7%) did not have a suspected bacterial infection.

Conclusion: Most febrile infants 2 to 6 months of age who had a CXR performed did not have radiographic pneumonia. Obtaining a CXR was associated with antibiotic use, even in the absence of radiographic pneumonia or other suspected bacterial infection. Evidence-based criteria for CXR utilization in febrile infants are needed.