Predictors of Low-Risk Invasive Bacterial Infection in Febrile Infants

Mother taking temperature of infant
Infants have a low probability of invasive bacterial infection if they are ≤60 days old with a fever by history only and have a normal urinalysis result.

Infants have a low probability of invasive bacterial infection if they are age ≤60 days with a fever by history only, have a normal urinalysis result, and have an absolute neutrophil count <5185 cell/µL, according to study results published in Pediatrics.

Febrile infants age ≤60 days are commonly evaluated in primary care offices and emergency departments, but relatively few infants will harbor invasive bacterial infection (defined as bacteremia and/or bacterial meningitis). While clinicians use various risk-stratification criteria to classify infants as either low-risk or not-low-risk for infection, many of these criteria were developed more than 25 years ago or include procalcitonin, which is not readily available in office settings or some hospitals. Therefore, researchers aimed to derive and internally validate a prediction model for identifying non-ill-appearing febrile infants age ≤60 days at low probability of invasive bacterial infection.

Researchers conducted a case-control study of febrile infants who presented to the emergency departments of 11 hospitals across the United States. Investigators matched every infant with invasive bacterial infection (case group; n=181) with 2 infants without invasive bacterial infection (control group; n=362). Using multiple logistical regression and 10-fold cross validation, predictors of invasive bacterial infection were identified and an invasive bacterial infection score was calculated with a score range of 0 to 10 (predictor with the adjusted odds ratio of the smallest magnitude was assigned 1 point).

Four predictors were associated with presence of invasive bacterial infection:

  • Age <21 days (1 point)
  • Highest temperature recorded in the emergency department 38.0°C to 38.4°C (2 points) or ≥38.5°C (4 points)
  •  Absolute neutrophil count ≥5185 cell/µL (2 points)
  • Abnormal urinalysis (3 points)

The area under the curve was 0.83 (95% CI, 0.79-0.86) for the weighted model and 0.83 (95% CI, 0.78-0.87) on the 10-fold cross validation.

An invasive bacterial infection score ≥2 had a sensitivity of 98.8% and a low specificity at 31.3%. Using a moderate-risk IBI score of ≥3 or 4, the sensitivity was lower (92.9% and 88.2%, respectively) but the specificity was higher (52.0% and 57.3%, respectively).

Of note, all 26 infants with bacterial meningitis had an invasive bacterial infection score of ≥2.

Thus, febrile infants with an invasive bacterial infection score <2, who have a low probability of invasive bacterial infection and who may not require lumbar puncture, are infants with a fever by history only, a normal urinalysis, and absolute neutrophil count <5185 cell/µL.

Related Articles

However, because of its low specificity, “the [invasive bacterial infection] score alone should not be used to determine the need for lumbar puncture in infants with scores ≥2,” stressed the researchers. Further, “a prospective investigation is needed to externally validate this [invasive bacterial infection] score,” concluded the researchers.  


Aronson PL, Shabanova V, Shapiro ED, et al; Febrile Young Infant Research Collaborative. A prediction model to identify febrile infants ≤60 days at low risk of invasive bacterial infection [published on June 5, 2019]. Pediatrics. doi:10.1542/peds.2018-3604