In 2021, the American Academy of Pediatrics (AAP) published clinical practice guidelines for the evaluation and management of well-appearing febrile infants 8 to 60 days old (of which one of us [N.K.] is an author).1 These guidelines were much needed for 2 fundamental reasons: (1) hundreds of thousands of febrile infants are evaluated in US emergency departments and other outpatient settings annually and (2) substantial variation in the management of this vulnerable patient population suggests inefficient and suboptimal care.
The AAP Guidelines and Inflammatory Markers
The AAP guidelines reviewed a considerable body of research performed over the past several decades on this topic to summarize the evidence and present simplified algorithms for clinicians to apply at the bedside. These guidelines represent a substantial step forward not only in synthesizing the evidence but also in providing the evidence in a form geared toward translating into practice. Nonetheless, trying to simplify such a comprehensive body of research (a task of more than a decade for the committee) inevitably has led to some challenges in applying the guidelines at the bedside. Comments and inquiries about the guidelines since publication have reflected several issues pertaining to inflammatory markers. One of the main goals of the AAP guidelines was to identify which inflammatory markers best stratify the risk of “important bacterial infections” (here defined as urinary tract infections, bacteremia, and bacterial meningitis) in young febrile infants. In addition, the guidelines attempted to define the threshold values at which clinicians should act on these inflammatory markers with more invasive testing (lumbar punctures [LPs]), empirical antibiotics, and hospital admission.
In the algorithms in the AAP guidelines, 4 different inflammatory markers with “high risk” thresholds are provided (T >38.5°C, procalcitonin (PCT) >0.5 ng/mL, C-reactive protein (CRP) >20 mg/L, and absolute neutrophil count (ANC) either >4000 or >5200/mm3). These appear in the footnotes of the figures of the algorithms in the AAP guidelines for the management of febrile infants 8 to 21 days old (Fig 1), 22 to 28 days old (Fig 2), and 29 to 60 days old (Fig 3).1 How to use those inflammatory markers in combination, however, and at which thresholds, needs further clarification. In addition, the algorithm footnotes state that “If procalcitonin is unavailable, both ANC and CRP should be obtained, and a temperature >38.5°C is considered abnormal.”1 In this commentary, we discuss the use of 3 of these inflammatory markers: fever, the ANC, and PCT. We will not address CRP because there is insufficient evidence on how best to use CRP and at what thresholds, as a component of prediction rules. Robust multicenter data are needed to understand the exact role CRP could play in circumstances when PCT is unavailable.
Temperature and ANC Thresholds With and Without PCT
Fever is the hallmark clinical finding for identifying infants at risk for important bacterial infections. Earlier guidelines2 identified temperatures ≥38.0°C or 38.2°C as indicating risk of important bacterial infections and suggesting the need for a comprehensive laboratory evaluation, usually including LPs. Other studies identified higher temperatures3 or a temperature >38.5°C4 as markers of higher risk. As a consequence, the AAP guidelines suggest that a temperature >38.5°C should be considered an inflammatory marker and that this should trigger a more comprehensive evaluation.1
An important issue that needs clarification, however, is the consideration of whether other clinical and laboratory factors are available and how they affect the importance of height of temperature. The above-referenced articles citing higher temperatures as a risk factor for important bacterial infections did not include and therefore adjust for serum PCT, a biomarker that has the best test characteristics (compared with the ANC and CRP) for identifying febrile infants with bacteremia and/or bacterial meningitis in this age group.5 In fact, the authors of 2 large prospective multicenter studies who aimed to generate prediction rules to identify young, febrile infants with these important bacterial infections did not find that height of temperature was a significant predictor after adjusting for other variables, including serum PCT.6,7 Therefore, for clinicians who have access to routine laboratory testing that includes PCT, height of temperature should not be the driver of more aggressive management, including LPs, empirical antibiotics, or hospitalization in the face of normal biomarker values, including PCT.6,7 Although adding height of temperature to prediction algorithms that include PCT may slightly increase the sensitivity of the prediction rules for identifying infants with important bacterial infections, this will decrease specificity and likely lead to unnecessary LPs, empirical antibiotics, and hospitalizations.
In contrast, in settings in which clinicians do not have access to PCT testing, height of temperature should be considered an inflammatory marker2,3,4 and should be incorporated as a risk factor for bacterial infections.
Of the routine measures included in the complete blood count, the ANC is the most accurate for identifying febrile infants at high and low risk for invasive bacterial infections2,5,7 and enhances the test accuracy beyond serum PCT.6,7 As a result, the AAP guidelines appropriately include the ANC as an inflammatory marker that, if elevated, should trigger a more comprehensive evaluation of the infant. Given the existing literature, the AAP guidelines reference 3 different ANC cutoffs from 3 different studies that attempt to identify febrile infants at high and low risk of important bacterial infections. The Step-by-Step study6 uses the ANC at a predetermined cutoff of 10 000 cells/mm3, as well as serum PCT and CRP as blood biomarkers. The febrile infant prediction rule derived and validated by our group in the Pediatric Emergency Care Applied Research Network identified an ANC threshold of 4100 cells/mm3 using recursive partitioning (although an ANC threshold of 4000 cells/mm3 was nearly identical in accuracy, is easier to use, and is, therefore, appropriately recommended in the AAP guidelines).7 Finally, a prediction rule derived in a multicenter retrospective study to identify young febrile infants at risk for invasive bacterial infections, which did not include either PCT or CRP in the model, identified an ANC threshold of 5200 cells/mm3 for determining high risk. In the algorithms provided by the AAP guidelines, both the 4000/mm3 and 5200/mm3 ANC thresholds are mentioned in the footnotes. However, to be clear, the 4000/mm3 threshold should be used in conjunction with serum PCT and the urinalysis, and the 5200/mm3 threshold should be used with height of temperature and the urinalysis when PCT is unavailable. This distinction is essential for consideration of the importance of fever and the appropriate application of the ANC in the prediction rules.
In summary, the AAP guidelines on the management of well-appearing young febrile infants 8 to 60 days of age at risk for important bacterial infections have synthesized years of research and provided actionable data that can be translated into clinical practice. Nonetheless, clinicians must be mindful that even these guidelines need careful interpretation, espescially based on the availability or non-availabilty of PCT. Appropriate use of inflammatory markers with evidence-based thresholds will optimize the sensitivity and specificity of the risk stratification approaches. The height of temperature is an important predictor when PCT is unavailable but not when PCT is available, and the ANC threshold determining high risk differs based on whether PCT is included in the algorithm.
Dr Kuppermann drafted the manuscript, Drs Mahajan and Dayan reviewed and made critical revisions to the manuscript, and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
The guidelines/recommendations in this article are not American Academy of Pediatrics policy, and publication herein does not imply endorsement.
CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have no potential conflicts of interest relevant to this article to disclose.
Comments