Bacteraemia in febrile children presenting to a pae3iatric emergency department.

OBJECTIVE: To determine the prevalence of bacteraemia in young febrile children presenting to a paediatric emergency department.

DESIGN: Prospective observational case study.

SETTING: Emergency Department of the Royal Children's Hospital, Melbourne, between May 1996 and May 1997.

PARTICIPANTS: Patients aged 3-36 months presenting to the Emergency Department with temperature > or = 39 degrees C and without specific viral illnesses (varicella, croup or herpes gingivostomatitis).

OUTCOME MEASURES: Bacteraemia (defined as presence of pathogenic bacteria in a blood culture), white blood cell count (WCC), McCarthy score, and final diagnosis based on clinical features and investigations.

RESULTS: Bacteraemia was identified in 18 of 534 patients (3.4%). Pathogens isolated were Streptococcus pneumoniae (15), Neisseria meningitidis (2) and Klebsiella pneumoniae (1). Increased WCC counts (P < 0.001) and brief duration of fever (P < 0.001) were associated with bacteraemia. Nevertheless, clinical features, including McCarthy scores, and high WCC counts (> or = 20 x 10(9)/L) had < 10% predictive accuracy for bacteraemia. Overall, final diagnoses in the 534 febrile children included non-specific viral infections (25%), upper respiratory tract infections (24%), otitis media (10%), gastroenteritis (9%), pneumonia (7%), and urinary tract infection (5%).

CONCLUSIONS: Most urban Australian children aged 3-36 months presenting to a paediatric emergency department with temperature > or = 39 degrees C without a clinical focus have a viral infection. However, 3%-4% have occult bacteraemia. Neither clinical features nor high WCC counts reliably identify these patients. As empiric antibiotics may contribute to increasing antibiotic resistance and have not been shown to prevent the rare complication of meningitis, we believe that close contact and regular review of these patients is preferable to empiric antibiotic therapy.