If you’ve been around long enough, you’ve watched the limbo bar for lumbar punctures (LP) in well-appearing febrile infants slowly drop.
Once upon a time, we used to perform lumbar punctures for any febrile baby ≤3-months-old.
When I started training, that bar had already moved down. Now babies ≤28 days with fevers get an LP.
Then, newer guidance from the American Academy of Pediatrics nudged the “automatic LP” group back down to <22 days, with a bit more nuance about whether to perform an LP based on inflammatory markers for infants 22–28 days old.
It’s always been a careful balancing act. We don’t want to perform unnecessary LPs, give antibiotics, or hospitalise babies who don’t need it. At the same time, we don’t want to miss any babies with invasive bacterial infections (IBI) like bacteremia and especially bacterial meningitis because that could be catastrophic.
The PECARN febrile infant rule for ≤60-day-olds uses three lab values to identify babies who were “low risk” for serious bacterial infection (SBI):
Negative urinalysis/dipstick
Serum Procalcitonin ≤ 0.5 ng/mL
Blood ANC ≤ 4000/mm³
The original study looked at infants up to 60 days old and combined UTIs, bacteremia, and bacterial meningitis into the SBI category. UTIs are the most common of the SBIs and are much more common than the infections we care about most in febrile infants: bacteremia and bacterial meningitis.
Research question
In well-appearing febrile infants ≤28-days-old, how accurately does the PECARN prediction rule identify infants at low risk of bacteremia and/or bacterial meningitis?
Burstein B, Waterfield T, Umana E, Xie J, Kuppermann N. Prediction of Bacteremia and Bacterial Meningitis Among Febrile Infants Aged 28 Days or Younger. JAMA. Published online December 08, 2025. doi:10.1001/jama.2025.21454
What did they do?
This was a pooled analysis of four prospective cohort studies conducted in pediatric EDs across six countries (Canada, Spain, Italy, Switzerland, the UK/Ireland, and the U.S.).
They included previously healthy, term, non–ill-appearing infants aged 0–28 days presenting with fevers ≥38.0°C and were seen in pediatric EDs. All had blood, urine, and procalcitonin. LP was performed at the clinician’s discretion. However, those who did not have LPs performed were followed up.
They looked to see whether these babies met PECARN low-risk criteria and whether they had IBIs.
The primary analysis included only the four external cohorts (i.e., non-PECARN). The secondary analysis added the original PECARN derivation + validation cohorts.
What did they find?
The four external cohorts included 1537 infants aged 0–28 days.
- 69 (4.5%) had IBIs (bacteremia and/or bacterial meningitis)
- 11 (0.7%) had bacterial meningitis
For IBI in the primary analysis, the PECARN rule had:
- Sensitivity: 94.2% (95% CI 85.6-97.8)
- Specificity: 41.6% (95% CI 36.7-46.7)
- Negative predictive value (NPV): 99.4% (95% CI 98.1-99.8)
- Positive predictive value (PPV): 6.9% (95% CI 4.8-9.9)
In the secondary analysis (including the PECARN cohorts), rule performance was similar.
Five infants in the full pooled sample were categorised as low-risk but had bacteremia (no bacterial meningitis).
Across 1,537 infants in the primary analysis and 2,531 infants in the full primary + secondary pool, there were 11 and 22 infants with bacterial meningitis, respectively, and the PECARN Rule missed ZERO cases of bacterial meningitis.
What does this mean for clinical practice? Can I stop performing LPs in febrile infants 0-28 days who meet PECARN Low-Risk Criteria?
Yes, for some.
This study provides evidence suggesting that selective performance of LPs can be considered an option for low-risk febrile infants in the first month of life and provides precise evidence to support conversations and shared decision-making with families. Consider all three pillars of EBM:
Strengths
- International collaboration, including six countries
- Sample size sufficient to compare clinical outcomes
- Use of biomarkers that are readily available in most countries included in the study
A Few Words of Caution
- These decision rules/tools do not supersede clinical judgement. Do not apply this or any prediction rule to infants who are ill-appearing.
- If the 0–28-day infant does NOT meet low-risk criteria, nothing changes. Continue with the LP.
- Be wary of Herpes Simplex Virus (HSV). Just because the infant may be low-risk for IBI, it does not mean they do not have risk for HSV. Look for risk factors for HSV.
- Infants in these studies mainly came from higher-resourced countries. Be careful in applying this finding to lower-resourced environments.
- Currently, the use of this rule depends on the availability of procalcitonin, which may not be available at all sites. There is no substitute currently with other inflammatory markers like CRP.
Comments from the authors:
“We finally have sufficient numbers of febrile infants in the first month of life to confidently say which infants can safely avoid LP.
This has the potential to change decades of practice and provides the evidence needed for guideline reconsideration.
Families can now hear:
“Based on this testing, your baby’s risk of meningitis is likely lower than 1 in two thousand.”
That conversation simply wasn’t possible with older analyses.
This study is practice-informing but not yet guideline-replacing. We believe this level of data is what clinicians need to inform the discussions of risks and benefits with families and consider approaches without LPs for subsets of these patients. The authors anticipate changes to guidelines based on these data.
