Rochester Criteria for Febrile Infant Risk Assessment

About

The Rochester Criteria were developed by Ron Dagan, MD, and colleagues in 1985 at the University of Rochester School of Medicine to identify febrile infants (≤90 days) at low risk for serious bacterial infection (SBI). The criteria combine historical, physical examination, and laboratory parameters including peripheral WBC count, absolute band count, urinalysis, and stool microscopy. Infants meeting all criteria have a <1% risk of SBI and can be safely managed as outpatients without empiric antibiotic therapy or lumbar puncture. The criteria have been validated in multiple prospective studies and remain a cornerstone of febrile infant management protocols worldwide, endorsed by the American Academy of Pediatrics (AAP) in their clinical practice guidelines for febrile infants.

Formula

Low risk = ALL criteria met: previously healthy + no focal infection + WBC 5-15K/mL + bands <1.5K/mL + normal urinalysis (<10 WBC/hpf) + normal stool (<5 WBC/hpf, if diarrhea)

The Rochester Criteria require ALL of the following to be met for an infant to be classified as low risk for SBI. The six criteria are: (1) Previously healthy — term infant (≥37 weeks gestation), no perinatal complications (no maternal fever, no chorioamnionitis, no prolonged rupture of membranes >24h), no prior antibiotic treatment, no unexplained hyperbilirubinemia, and no prior hospitalization; (2) No focal bacterial infection on physical examination — absence of erythema, tenderness, discharge, or swelling indicating ear infection (otitis media), skin/soft tissue infection (cellulitis, abscess), bone/joint infection (septic arthritis, osteomyelitis), or lymphadenitis; (3) WBC count between 5,000 and 15,000/mL (5.0-15.0 ×10³/mm³) — values outside this range are associated with increased SBI risk, with WBC <5K suggesting possible overwhelming infection (leukopenia) and WBC >15K suggesting systemic inflammation; (4) Absolute band count (immature neutrophils) <1,500/mL (<1.5 ×10³/mm³) — elevated bands indicate a left shift and are associated with bacterial infection; (5) Urinalysis with <10 WBC per high-power field (hpf) on microscopy — >10 WBC/hpf suggests urinary tract infection, the most common SBI in this age group; (6) If diarrhea is present, stool microscopy with <5 WBC/hpf — >5 WBC/hpf suggests bacterial gastroenteritis. Important caveats: The Rochester Criteria do NOT apply to ill-appearing infants, those with known immunodeficiency, or those who have received antibiotics within 48 hours. The criteria have a negative predictive value of 98.9% (95% CI: 97.4-99.6%) for SBI in low-risk infants.

Score Interpretation

The Rochester Criteria address a critical clinical dilemma in pediatric emergency medicine: febrile infants ≤90 days of age have a 5-15% prevalence of serious bacterial infection (SBI), but performing full septic workups (blood culture, urine culture, lumbar puncture) on all febrile infants is invasive, costly, and exposes many infants to unnecessary procedures and empiric antibiotics. Before the Rochester Criteria were introduced, practice patterns varied widely, with hospitalization rates of 30-70% and LP rates of 40-80% for febrile infants. The Rochester Criteria were the first validated clinical prediction rule to identify a low-risk group (<1% SBI risk) who could be safely managed without invasive testing. In validation studies, the criteria demonstrated a sensitivity of 92-100% for detecting SBI with a negative predictive value of 98.9% (95% CI: 97.4-99.6%). The criteria have been incorporated into the American Academy of Pediatrics (AAP) clinical practice guideline for the management of febrile infants (2012, reaffirmed 2016 and 2021). Key clinical impact: (1) Reduction in unnecessary lumbar punctures by 50-70% in low-risk infants; (2) Reduction in empiric antibiotic use and associated adverse effects; (3) Decreased hospital length of stay and healthcare costs; (4) Identification of the need for urine culture in all febrile infants (subsequent data showed that UTI is the most common SBI, accounting for 50-70% of all SBIs in febrile infants). The Philadelphia Criteria and Boston Criteria are alternative prediction rules developed subsequently with slightly different cutoff values and variable inclusion of chest radiography and CRP. The Rochester Criteria remain the most widely studied and most conservative (fewest LPs avoided but safest). Contemporary management of febrile infants incorporates the Rochester Criteria as part of a structured approach that may also include procalcitonin measurement, CRP, and newer viral diagnostics (RSV, influenza, enterovirus PCR) that further refine risk stratification.

Reference Ranges