Pantell RH, Roberts KB, Adams WG, Dreyer BP, Kuppermann N, O’Leary ST, Okechukwu K, Woods CR. Evaluation and Management of Well-Appearing Febrile Infants 8 to 60 Days Old. Pediatrics 2021;148 (2) e2021052228; DOI: 10.1542/peds.2021-052228
Earlier this month, The American Academy of Pediatrics (AAP) published new guidelines on the evaluation and management of well-appearing febrile infants 8-60 days old. Members of the working group included representation from general paediatrics, pediatric emergency medicine, pediatric infectious disease, hospital medicine, and family medicine.
They came up with:
- 21 key action statements.
- 3 algorithms based on age.
Why is this important?
We have been trying to decipher the riddle of the young febrile infant for more than four decades. When I started my paediatric training, I was told, “babies are liars.” This is not to say that infants are purposely facetious, but rather it is sometimes difficult to distinguish a sick versus not sick febrile infant based on physical exam alone.
While we do not want to miss urinary tract infections, bacteremia, or bacterial meningitis in these young infants, we also do not want to expose this population to unnecessary procedures (lumbar puncture), antimicrobial therapy, or hospitalisation.
Currently, there still exists much practice variation among clinicians in the care of young febrile infants.
Back in my day…
Initial risk stratification strategies and prediction tools for identifying febrile infants at risk for serious bacterial infection (SBI) had arbitrarily determined laboratory criteria and thresholds for white blood cell (WBC) count, absolute neutrophil count (ANC), immature to total neutrophil ratio, etc. Unfortunately, these previous tools do not hold up well in the current landscape and misclassified almost a third of infants with bacterial meningitis.
This may be due to the changing of SBIs in the era of routine Group B streptococcus screening for pregnant women, conjugate pneumococcal vaccination, and more robust food safety. Overall, there seems to be a shift from Gram-positive to Gram-negative organisms causing these infections.
The new AAP guidelines
Inclusion for consideration in the guideline:
- Well appearing
- Fever ≥38°C (100.4°F) at home in the past 24 hours or in the clinical setting
- Gestation between ≥37 and <42 weeks
- 8 to 60 days of age
- Home after discharge from newborn nursery or born at home
- Preterm infants <37 weeks gestation
- Infants < 2 weeks old with perinatal course of maternal fever, infection, or antimicrobial use
- High suspicion for herpes simplex virus (HSV)
- Focal bacterial infection, including cellulitis, omphalitis, septic arthritis, osteomyelitis
- Clinical bronchiolitis
- Neonatal course complicated by surgery or infection
- Congenital or chromosomal abnormality
- Medically fragile infants
- Infants who received immunisations within the last 48 hours
Four key components of evaluation
- Blood culture
- Inflammatory Markers (IM)
- Procalcitonin>0.5 ng/mL*
- ANC > 4000 mm3 (PECARN cutoff which included PCT); >5200 mm3 † (Febrile Young Infant Research Collaborative which did not include PCT)
- CRP >20 mg/L
- Temperature > 38.5°C
*Procalcitonin is more specific for invasive bacterial infections (bacteremia and bacterial meningitis) and most accurate for risk stratification.
†You read that correctly. There are two ANC cutoffs included based on the PECARN study and the Febrile Young Infant Research Collaborative study.
- Cerebrospinal Fluid (CSF) from lumbar puncture (LP)
Special caution: HSV
What if there’s a positive respiratory viral test?
Respiratory viral polymerase chain reaction (PCR) testing has become more widely available. Some studies suggest the presence of a positive respiratory viral infection may decrease the risk of invasive bacterial infection. However, it still remains unclear how a positive respiratory viral PCR influences evaluation and management in the first month of life. Therefore, the presence of a positive viral respiratory PCR does not exclude an infant from the algorithms.
Key takeaway points
There is a lot of information and nuance in this clinical practice guideline that is difficult to unpack in a blog post, including weighing the risks and benefits of each key action statement and discussion of future directions and limitations. Have a look at the official document for full details. However, here are some highlights:
- Stop using the term “serious bacterial infection” or “SBI.” We need to be specific about what we are talking about. Is it urinary tract infections, bacteremia, or bacterial meningitis?
- Continue to be wary of possible HSV infection, especially in the first three weeks of life.
- WBC count is NOT as reliable as other inflammatory markers for risk stratification. Stop using it. That is, not all IMs are created equal!
- Inflammatory markers can be used for risk stratification, but no single marker is perfectly reliable by itself.
- Procalcitonin with ANC or CRP is preferred.
- If procalcitonin is not available, use both ANC and CRP and temperature >38.5
- The age cutoff of 21 days is new compared to most previous studies with age cutoffs of 28 days.
- If hospitalised, monitor cultures for 24 to 36 hours.
- This clinical practice guideline does NOT include infants within the first week of life or those ill-appearing.
Ultimately, this is a guideline and does not supplant clinician judgement and each individual’s risk tolerance and aversion. Clinician experience, in addition to parent or caretaker values, should be considered whenever possible in a shared decision-making process.
Nate Kuppermann’s comments:
The AAP Clinical Practice Guideline on the Evaluation and Management of Well-appearing Febrile Infants 8 to 60 Days Old represent a quantum advance in the approach to this conundrum for clinicians everywhere. It reflects approximately 15 years of work by a group of investigators and clinicians representing several specialities and was reviewed and edited by many committees of the AAP.
As a member of the team and writing group and an investigator of this topic for 25 years, I would like to add just a few comments, some of which are already alluded to in this well-considered DFTB commentary.
- No guideline or prediction rule is meant to replace clinician judgment. These evidence-based tools are meant to empower the clinician with evidence to be used, along with their judgment and other considerations, to make the best clinical decisions for the patient and family.
- The older guidelines and prediction rules dating back to the mid-1980s did not use statistically derived thresholds for laboratory cutoffs and frequently used lumbar punctures (LPs) in the evaluation of these infants. As a consequence, although the sensitivity of these previous algorithms was generally high, their specificities were not. And many infants received unnecessary invasive testing, empirical antibiotics and hospitalizations.
- Other fundamental limitations of older algorithms included:
- Enrollment of hundreds rather than thousands of infants in the studies that derived these algorithms. Therefore, the resulting point estimates of risk were not precise.
- Different temperature thresholds for inclusion.
- Different age cutoffs for inclusion.
- Different WBC thresholds for defining high and low-risk
- Different methods for obtaining and different cutoffs for UA WBC
- Differential inclusion of stool and LP WBC counts included in the protocol
- The Yale Observational Scale score has proven not to be accurate for identifying infants with invasive bacterial infections (IBI; bacteremia, bacterial meningitis)
- The WBC count has poor accuracy for identifying IBIs in these febrile infants; the absolute neutrophil count (ANC) is significantly better
- Neither the WBC nor ANC are as accurate as the C-reactive protein (CRP) or serum procalcitonin (PCT) for identifying infants with IBIs
- However, the PCT is substantially better than the CRP – serum PCT has become a fundamental biomarker for highly accurate rules that do not use the LP to stratify risk and help maintain high sensitivity and specificity. While misclassifying a few patients with IBIs, prediction rules using PCT also help limit the use of LPs, empirical antimicrobials and hospitalizations when not indicated.
- The two rules that have emerged as most accurate (sensitive and specific) as they include the use of PCT are the PECARN prediction rule (rounded to thresholds of ANC of 4K cells / mm3 and PCT of 0.5 ng/ml and validated, making it safer, easier, and more generalizable while maintaining its accuracy) and the Step-by-Step rule. Neither rely on LPs, and neither have reported misclassified patients with bacterial meningitis.
- The role of respiratory viral testing and how it impacts the two above prediction rules are under investigation. The PECARN group reported results at the 2021 PAS and SAEM meetings.
- Machine learning (ML) / artificial intelligence algorithms are being investigated by some (including the PECARN group) to enhance the accuracy of current algorithms to identify young febrile infants with IBIs. Although these algorithms may enhance accuracy by increasing specificity, there must be essentially no risk of missed bacterial meningitis. ML-based rules are frequently difficult to understand and require computerized decision support to implement.
- Genomic analysis (RNA transcriptional analyses, whole genome sequencing, etc) may soon prove to be substantially better than blood cultures for identifying pathogens. In the future, as these new technologies become more accurate and efficient, they may obviate “screening algorithms” and become both the screening test and the definitive test at the same time!
- HSV infections remain of critical importance in this age range. Given that most of these infections are in the first 3 weeks of life, LPs remain necessary in these early weeks, and the PECARN and Step-by-Step rules should not be applied to infants 3 weeks and younger.
- Finally, as no research will be impactful without robust implementation, we must think about how to get the evidence provided in these AAP guidelines to the clinician at the bedside to effect change. The use of computerized decision support and evidence-enhanced clinician/parent decision support will be needed and are being studied by the PECARN group.
Why not head to the SGEM to listen to Dennis and Ken Milne discuss the guidelines in detail?