Management of well-appearing febrile young infants aged ≤90 days

Correspondence: Canadian Paediatric Society, 100–2305 St Laurent Blvd, Ottawa, Ontario K1G 4J8, Canada. www.cps.ca.

Paediatrics & Child Health, Volume 29, Issue 1, February 2024, Pages 50–57, https://doi.org/10.1093/pch/pxad085

Cite

Brett Burstein, Marie-Pier Lirette, Carolyn Beck, Laurel Chauvin-Kimoff, Kevin Chan, Management of well-appearing febrile young infants aged ≤90 days, Paediatrics & Child Health, Volume 29, Issue 1, February 2024, Pages 50–57, https://doi.org/10.1093/pch/pxad085

Abstract

The evaluation and management of young infants presenting with fever remains an area of significant practice variation. While most well-appearing febrile young infants have a viral illness, identifying those at risk for invasive bacterial infections, specifically bacteremia and bacterial meningitis, is critical. This statement considers infants aged ≤90 days who present with a rectal temperature ≥38.0°C but appear well otherwise. Applying recent risk-stratification criteria to guide management and incorporating diagnostic testing with procalcitonin are advised. Management decisions for infants meeting low-risk criteria should reflect the probability of disease, consider the balance of risks and potential harm, and include parents/caregivers in shared decision-making when options exist. Optimal management may also be influenced by pragmatic considerations, such as access to diagnostic investigations, observation units, tertiary care, and follow-up. Special considerations such as temperature measurement, risk for invasive herpes simplex infection, and post-immunization fever are also discussed.

• Fever among infants aged ≤90 days is any rectal temperature ≥38.0°C, measured at home or in a clinical setting.
• Serious bacterial infections (SBIs) include urinary tract infections (UTIs), bacteremia, and bacterial meningitis.
• Invasive bacterial infections (IBIs) include bacteremia and bacterial meningitis.

BACKGROUND

Approximately 2% of healthy, term newborns are brought to medical attention for fever within their first 3 months of life ( 1–4). While most of these infants have self-limited viral illness, between 10% and 13% harbour a serious bacterial infection (SBI) ( 5). SBIs include urinary tract infections (UTIs), bacteremia, and bacterial meningitis. The prevalence of invasive bacterial infections (IBIs), specifically bacteremia and bacterial meningitis, is greatest in the first month post-birth and decreases with age ( 6, 7). When evaluating febrile young infants, health care providers (HCPs) must balance the risks of infection against the harms of over-investigation and over-treatment ( 8, 9). No single management strategy has been universally adopted, which has led to wide variations in care in Canada and elsewhere ( 5, 10, 11).

Enhanced clinical care for febrile infants ≤90 days of age has been the focus of several large-scale quality improvement ( 12) and knowledge translation initiatives ( 13). Several risk-stratification criteria exist to identify infants at low risk for SBI ( 14–16). However, older criteria provide conflicting recommendations, and all use subjective clinical findings and pre-determined laboratory values rather than statistically derived thresholds. They also pre-date widespread use of pneumococcal conjugate and Haemophilus influenzae type b vaccines and intrapartum group B streptococcus (GBS) prophylaxis ( 17, 18), and do not incorporate newer diagnostic tests ( 19–21). Moreover, older strategies were developed to identify SBIs, of which approximately 80% are UTIs ( 22, 23), which in turn distorts prediction for IBIs. Historic strategies lack specificity for IBI, such that thousands of infants routinely undergo invasive cerebrospinal fluid (CSF) testing, prolonged hospitalization, and broad-spectrum antibiotic therapy to prevent rare cases of bacterial meningitis from being missed ( 5).

IDENTIFYING IBIs

A shift toward patient-centred outcomes and value of care in recent decades has placed new emphasis on reducing unnecessary tests and treatments. Newer research has focused on developing statistically derived clinical prediction rules to identify infants with specific risk for an IBI rather than SBI. Several large multicentre studies have derived and validated approaches to identify low-risk infants who can be managed without lumbar puncture (LP), antibiotics, or hospitalization, notably the Pediatric Emergency Care Applied Research Network (PECARN) prediction rule ( 1), the Step-by-Step method ( 2), and the Aronson rule ( 3) ( Table 1). These approaches favour newer biomarkers such as procalcitonin (PCT) and C-reactive protein (CRP) because, as predictors of IBI, they outperform absolute neutrophil counts (ANC) and white blood cell (WBC) count ( 19, 21, 24). While PCT is the most sensitive and specific biomarker for IBI, test availability and turnaround times vary widely ( 25, 26). PCT is the diagnostic test of choice, when available, and clinicians should use either the PECARN prediction rule or the Step-by-Step method to stratify risk. The Aronson rule does not use PCT, making it reasonable to maintain a lower threshold for LP and hospitalization, even for low-risk infants, when applying this rule ( 27).

Risk-stratification for febrile young infants

ANC absolute neutrophil count; CRP C-reactive protein; ED emergency department; IBI invasive bacterial infection; NPV negative predictive value; PCT procalcitonin; PECARN Pediatric Emergency Care Applied Research Network; WBC/hpf white blood cells per high-powered field

Risk-stratification for febrile young infants

ANC absolute neutrophil count; CRP C-reactive protein; ED emergency department; IBI invasive bacterial infection; NPV negative predictive value; PCT procalcitonin; PECARN Pediatric Emergency Care Applied Research Network; WBC/hpf white blood cells per high-powered field

For infants identified as low risk using the Aronson, Step-by-Step, or PECARN strategies, the specific risk for IBI decreases to 0.6% (95% CI 0.2 to 1.2), 0.7% (95% CI 0.3 to 1.5), or 0% (95% CI 0.0 to 0.8), respectively. Because bacteremia is approximately fourfold more common than bacterial meningitis in infants diagnosed with IBI ( 5, 6), the number needed to test (NNT) by LP to exclude one case of bacterial meningitis among low-risk infants ranges between 333 to 2000 (Aronson), 267 to 1333 (Step-by-Step), or 500 to undefinably high (PECARN).

Based on current best evidence and expert opinion, these guidelines are intended to be pragmatic and applicable for clinicians in a broad range of settings (e.g., emergency department [ED], clinic, urban, rural), always recognizing that regional differences exist regarding access to tests, inpatient observation units, follow-up, and tertiary care. The recommendations apply to well-appearing infants ≤90 days old with a documented history of fever, defined as any rectal temperature ≥38.0°C, taken by a caregiver or HCP at home or in a clinical setting ( 28, 29). Importantly, even well-appearing febrile young infants can deteriorate relatively rapidly, and clinicians should maintain a high index of suspicion for infection in this age group.

Studies of febrile young infants generally exclude infants with commonly accepted risk factors ( Table 2). Therefore, given the paucity of evidence regarding these risk factors, individualized care and consideration of conservative management are warranted. Ill-appearing infants should be assessed immediately using a structured assessment tool (e.g., the ‘Pediatric Assessment Triangle’ [ 30], or similar). The priority for all ill-appearing infants involves supporting the airway, breathing, and circulation. In all such cases, obtain blood and urine cultures and consider deferring LP testing until the infant is hemodynamically stable. Treatment with empiric antibiotics should not be delayed.

Commonly accepted risk factors