Prediction Rule Identifies Low Infection Risk in Febrile Infants

Heidi Splete

June 17, 2021

A clinical prediction rule combining procalcitonin, absolute neutrophil count, and urinalysis effectively identified most febrile infants at low risk for serious bacterial infections, based on data from 702 individuals

The clinical prediction rule (CPR) described in 2019 in JAMA Pediatrics was developed by the Febrile Infant Working Group of the Pediatric Emergency Care Applied Research Network (PECARN) to identify febrile infants at low risk for serious bacterial infections in order to reduce unnecessary procedures, antibiotics use, and hospitalization, according to April Clawson, MD, of Arkansas Children's Hospital, Little Rock, and colleagues.

In a poster presented at the Pediatric Academic Societies annual meeting, the researchers conducted an external validation of the rule via a retrospective, observational study of febrile infants aged 60 days and younger who presented to an urban pediatric ED between October 2014 and June 2019. The study population included 702 infants with an average age of 36 days. Approximately 45% were female, and 60% were White. Fever was defined as 38°C or greater. Exclusion criteria were prematurity, receipt of antibiotics in the past 48 hours, presence of an indwelling medical device, and evidence of focal infection (not including otitis media); those who were critically ill at presentation or had a previous medical condition were excluded as well, the researchers said. A serious bacterial infection (SBI) was defined as a urinary tract infection (UTI), bacteremia, or bacterial meningitis.

Based on the CPR, a patient is considered low risk for an SBI if all the following criteria are met: normal urinalysis (defined as absence of leukocyte esterase, nitrite, and 5 or less white blood cells per high power field); an absolute neutrophil count of 4,090/mL or less; and procalcitonin of 1.71 ng/mL or less.

Overall, 62 infants (8.8%) were diagnosed with an SBI, similar to the 9.3% seen in the parent study of the CPR, Clawson said.

Of these, 42 had a UTI only (6%), 10 had bacteremia only (1.4%), and 1 had meningitis only (0.1%). Another five infants had UTI with bacteremia (0.7%), and four had bacteremia and meningitis (0.6%).

According to the CPR, 432 infants met criteria for low risk and 270 were considered high risk. A total of five infants who were classified as low risk had SBIs, including two with UTIs, two with bacteremia, and one with meningitis.

"The CPR derived and validated by Kupperman et al. had a decreased sensitivity for the patients in our study and missed some SBIs," Clawson noted. "However, it had a strong negative predictive value, so it may still be a useful CPR."

The sensitivity for the CPR in the parent study and the current study was 97.7 and 91.9, respectively; specificity was 60 and 66.7, respectively. The negative predictive values for the parent and current studies were 99.6 and 98.8, respectively, and the positive predictive values were 20.7 and 21.1.

The results support the potential of the CPR, but more external validation is needed, they said.

PECARN Rule Keeps It Simple

"It has always been a challenge to identify infants with fever with serious bacterial infections when they are well-appearing," Yashas Nathani, MD, of Oklahoma University, Oklahoma City, said in an interview. "The clinical prediction rule offers a simple, step-by-step approach for pediatricians and emergency medicine physicians to stratify infants in high or low risk categories for SBIs. However, as with everything, validation of protocols, guidelines and decision-making algorithms is extremely important, especially as more clinicians start to employ this CPR to their daily practice. This study objectively puts the CPR to the test and offers an independent external validation.

"Although this study had a lower sensitivity in identifying infants with SBI using the clinical prediction rule as compared to the original study, the robust validation of negative predictive value is extremely important and not surprising," said Nathani. "The goal of this CPR is to identify infants with low-risk for SBI and the stated NPV helps clinicians in doing just that."

Overall, "the clinical prediction rule is a fantastic resource for physicians to identify potentially sick infants with fever, especially the ones that appear well on initial evaluation," said Nathani. However, "it is important to acknowledge that this is merely a guideline, and not an absolute rule. Clinicians also must remain cautious, as this rule does not incorporate the presence of viral pathogens as a factor.

"It is important to continue the scientific quest to refine our approach in identifying infants with serious bacterial infections when fever is the only presentation," Nathani noted. "Additional research is needed to continue fine-tuning this CPR and the thresholds for procalcitonin and absolute neutrophil counts to improve the sensitivity and specificity." Research also is needed to explore whether this CPR can be extended to incorporate viral testing, "as a large number of infants with fever have viral pathogens as the primary etiology," he concluded.

The study received no outside funding. The researchers had no financial conflicts to disclose. Nathani had no financial conflicts to disclose.

This article originally appeared on, part of the Medscape Professional Network.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.