A clinical prediction rule to identify febrile infants 60 days and younger at low risk for serious bacterial infections
Kuppermann N, Dan PS, Levine DA, et al. [published ahead of print February 18, 2019] JAMA Pediatr. doi:10.1001/jamapediatrics.2018.5501.
SUMMARY:
The Pediatric Emergency Care Applied Research Network (PECARN) group sought to derive and validate an accurate prediction rule in a cohort of febrile infants <60 d old to identify those at low risk of serious bacterial infection (SBI), defined as bacterial meningitis, bacteremia, or urinary tract infection (UTI).
This multicenter prospective study included infants with a temperature ≥38°C in the ED, clinic, or home, and excluded those critically ill, those who had already received antibiotics, and those with a high-risk past medical history, with work-up at the discretion of treating physician.
The study enrolled 1,896 febrile infants (1,821 with analyzable procalcitonin data and complete assessments for SBI), who were randomly divided into derivation and validation groups (908 and 913).
SBIs were present in 170/1,821 infants (9.3%), including 26 (1.4%) with bacteremia, 151 (8.3%) with UTIs, and 10 (0.5%) with bacterial meningitis.
Among the infants, 76.8% had a lumbar puncture performed, including 68.8% of infants 29-60 d old.
Using a recursive partitioning analysis algorithm, the authors found low risk if there was a negative urinalysis, a serum procalcitonin level 1.71 ng/ml or less, and an absolute neutrophil count < 4.09 × 109/L.
In the validation cohort, the rule sensitivity was 97.7%, the specificity was 60.0%, the NPV was 99.6%, and the NLR was 0.04.
One infant with bacteremia and two infants with UTIs were misclassified. No patients with bacterial meningitis were missed by the rule.
Some minor limitations include that this was a convenience sample depending on research-assistant availability and that the authors did not examine other potential biomarkers. A greater limitation is that the authors did not include serious viral infections (eg, herpes simplex virus encephalitis) or overall viral testing (eg, respiratory syncytial virus).
EDITOR’S COMMENTARY: This is a very well conducted study on a large group of febrile infants less than 60 d old. The authors were able to identify a low-risk cohort for SBI by using 3 simple, objective variables: urinalysis, absolute neutrophil count, and serum procalcitonin. This procedure has the unique perks of not relying on subjective assessments or requiring lumbar puncture. Importantly, this rule must be subject to external validation, so it is not ready to use yet. You should still be conservative with these patients, particularly infants ≤28 d old and un/underimmunized people.
Copyright 2019 by Emergency Medical Abstracts – All Rights Reserved 09/19 - #01
