ABSTRACT
BACKGROUND: Febrile neonates undergo lumbar puncture (LP), empiric antibiotic administration, and admission for increased risk of invasive bacterial infection (IBI), defined as bacteremia and meningitis. OBJECTIVE: Measure IBI prevalence in febrile neonates, and operating characteristics of Rochester Criteria (RC), Yale Observation Scale (YOS) score, and demographics as a low-risk screening tool. METHODS: Secondary analysis of healthy febrile infants < 60 days old presenting to any of 26 emergency departments in the Pediatric Emergency Care Applied Research Network between December 2008 and May 2013. Of 7334 infants, 1524 met our inclusion criteria of age ≤ 28 days. All had fevers and underwent evaluation for IBI. Receiver operator characteristic (ROC) curve and transparent decision tree analysis were used to determine the applicability of reassuring RC, YOS, and age parameters as an IBI low-risk screening tool. RESULTS: Of 1524 neonates, 2.9% had bacteremia and 1.5% had meningitis. After applying RC and YOS, 15 neonates were incorrectly identified as low risk for IBI (10 bacteremia, 4 meningitis, 1 bacteremia, and meningitis). Age ≤ 18 days was a statistically significant variable ROC (area under curve 0.63, p < 0.05). Incorporating age > 18 days as low-risk criteria with reassuring RC and YOS misclassified 7 IBI patients (6 bacteremia, 1 meningitis). CONCLUSION: Thirty percent of febrile neonates met low-risk criteria, age > 18 days, reassuring RC and YOS, and could avoid LP and empiric antibiotics. Our low-risk guidelines may improve patient safety and reduce health care costs by decreasing lab testing for cerebrospinal fluid, empiric antibiotic administration, and prolonged hospitalization. These results are hypothesis-generating and should be verified with a randomized prospective study.
