Whether ‘tis nobler in mind to accept
The risks and benefits of empiric treatment.
This is a guest post by Dr. Dennis Ren. Dennis is currently a pediatric emergency medicine fellow at Children’s National Hospital in Washington, DC. He has been a guest on podcasts such as Peds Admit and The Skeptics’ Guide to Emergency Medicine. His areas of interest include medical education, simulation, interprofessional communication and teamwork, disaster/emergency preparedness.
The PECARN Febrile Infant Study
Please forgive my modern adaptation of a classic Shakespearian soliloquy, but it does seem that we have been struggling with the conundrum of the febrile infant for a very long time. To miss a serious bacterial infection (SBI), defined as urinary tract infection, bacteremia, and bacterial meningitis, can have serious consequences. (Justin comment: I think there is some reasonable debate about whether UTI is truly a serious bacterial illness, but I want to find it in these very young infants. That being said, we need to be aware of the false positive rate – or high baseline asymptomatic bacteriuria – in diaper wearing infants). However, overtreatment leads to unnecessary lumbar punctures, exposure to empiric antibiotics, and prolonged hospitalization. Although some practitioners continue to use screening criteria from the 1990s, clinical decision tools are still being created to help identify infants at low risk for SBI. Despite these new tools and evolving medical knowledge, there is still a wide practice variation in the evaluation of febrile infants in the emergency department.
It is estimated that 8-13% of febrile infants presenting to the emergency department may have SBI. It is difficult to identify which infants have SBI based solely on clinical exam. Many of the original low risk criteria were created back in the 1990s. They include the Boston1, Philadelphia2, and Rochester3 criteria that risk stratified febrile infants based on a combination of white blood cell counts, bands, urinalysis, stool, chest X-ray, and cerebral spinal fluid findings. With updates to vaccination, these criteria may be out of date.
Kuppermann et al’s study4 offers a way to identify low risk febrile infants based on three objective lab criteria that include the urinalysis, absolute neutrophil count, and procalcitonin.
This study was part of a larger multicenter study through the Pediatric Emergency Care Applied Research Network (PECARN) looking at RNA microarray analysis and detection of bacterial infections. The study included febrile infants (defined as rectal temperature of at least 38°C in the ED, prior healthcare setting, or at home within 24 hours) who had blood cultures obtained. They excluded any infants who were deemed critically ill, received antibiotics in the past 48 hours, or who had a history of prematurity (less than or equal to 36 weeks), pre-existing medical conditions, indwelling devices, or soft-tissue infections.
Unlike previous prediction rules which used predetermined thresholds for lab values, this study divided infants into derivation and validation groups and employed recursive partitioning to identify the optimal thresholds for each predictor.
The final group included 1,821 febrile infants less than 60 days of age. The mean age was 36 days old and 42% were female. 9.3% were SBI positive with 7.7% from urinary tract infection alone.
The low risk prediction rule was based on three objective laboratory findings:
- Normal urinalysis
- Absolute neutrophil count ≤4,090/µL
- Serum procalcitonin ≤1.71 ng/ml
This yielded a sensitivity of 97.7%, specificity of 60%, negative predictive value of 99.6%, and negative likelihood ratio of 0.04.
This rule did not miss any infants with bacterial meningitis. There were three missed cases. Two infants in the validation data set had UTI (E. coli and Pseudomonas) without pyuria. (Justin comment: So possibly incidental, asymptomatic bacteriuria.) One infant in the derivation set had Enterobacter cloacae bacteremia but was admitted for observation with a negative repeat blood culture prior to antibiotics. All missed cases had uneventful courses.
Overall, this is a well-designed and executed study and proposes a clinical prediction tool for identifying febrile infants at low risk for SBI based on three objective lab criteria.
One key feature of this study is that the laboratory thresholds are derived and not predetermined. When the researchers rounded the ANC to 4,000/µL and decreased the procalcitonin threshold to 0.5 ng/ml, it yielded similar sensitivity and specificity decreased only slightly.
The researchers did lose a portion (41%) of their study population due to issues with procalcitonin so it is hard to determine whether or not this would have changed the results. Patients were only recruited when a research coordinator was available and only patients with blood cultures obtained were included which may lead to a selection bias. Unfortunately, this study was unable to address the utility of CRP which may be a more readily available lab test in comparison to procalcitonin at certain institutions.
A key limitation of this study is that it does not address concurrent viral infection. Previous studies suggest that concurrent viral infection may decrease, but not necessarily exclude SBI. There were four patients with HSV infections (three with HSV encephalitis, and one with HSV detected in the nasopharynx) in this study population.
This rule should NOT be applied to infants ≤28 days as they have higher risk of infection including HSV. Keep in mind that the procalcitonin threshold of ≤1.71 ng/mL should be applied only after the urinalysis and absolute neutrophil count are below cut-off. Finally, this clinical prediction rule is not a substitute for clinical judgement, but may be helpful in guiding workup and management.
(Justin comment: Given that at least some of these UTIs were probably not really UTIs, the numbers are probably even better than presented here.)
Other FOAMed on the Febrile Infant
Skeptics Guide to Emergency Medicine #296: She’s Got the Fever But Does She Need an LP, Antibiotics or an Admission?
Skeptics Guide to Emergency Medicine #171: Step-by-Step Approach to the Febrile Infant
Academic Life in Emergency Medicine: PECARN Infant Fever Rule Age 29-60 days
REBEL EM: A Clinical Prediction Rule for Febrile Infants ≤60 days at Low Risk for Serious Bacterial Infections.
Don’t Forget the Bubbles: Fever Under 60 Days of Age
Core EM: PECARN Febrile Neonate Decision Rule Derivation and Internal Validation
- Baskin MN, O’rourke EJ, Fleisher GR. Outpatient treatment of febrile infants 28 to 89 days of age with intramuscular administration of ceftriaxone. J Pediatr. 1992;120(1):22-7.
- Baker MD, Bell LM, Avner JR. Outpatient management without antibiotics of fever in selected infants. N Engl J Med. 1993;329(20):1437-41.
- Jaskiewicz JA, Mccarthy CA, Richardson AC, et al. Febrile infants at low risk for serious bacterial infection–an appraisal of the Rochester criteria and implications for management. Febrile Infant Collaborative Study Group. Pediatrics. 1994;94(3):390-6.
- Kuppermann N, Dayan PS, Levine DA, et al. A Clinical Prediction Rule to Identify Febrile Infants 60 Days and Younger at Low Risk for Serious Bacterial Infections. JAMA Pediatr. 2019;173(4):342-351.
Ren, D. To LP, or not to LP (the febrile infant): That is the question, First10EM, August 3, 2020. Available at: