Does the height of fever predict the probability of a Serious Bacterial Infection (SBI) in a child?

Does Fever Height Predict Serious Bacterial Infection in Children?

Fever height shows a modest association with serious bacterial infection (SBI) risk in febrile children, but it cannot reliably exclude or confirm SBI when used alone and should not drive clinical decision-making in isolation.

The Evidence on Fever Height and SBI Risk

The relationship between fever height and SBI is complex and age-dependent:

Pre-Vaccine Era Data

• In the pre-pneumococcal vaccine era, children aged 3-36 months with very high fever (≥40°C/104°F) had approximately 12% risk of bacteremia, compared to 7% with fever ≥38°C (100.4°F) in infants ≤90 days 1
• The combination of fever ≥39.5°C (103°F) and elevated WBC count (≥15×10⁹/L) also conferred approximately 12% bacteremia risk 1

Contemporary Evidence

• A large multicenter study (n=4,821) found that fever height was significantly associated with SBI risk, with an odds ratio of 1.5 (95% CI 1.2-1.8), meaning higher temperatures modestly increase SBI probability 2
• However, the American College of Emergency Physicians guideline analyzing 429 infants aged 57-180 days explicitly states that "height of fever, sex, and age were not associated with increased risk of SBI" in their cohort 3
• Neither fever height nor duration can reliably exclude SBI when used in isolation 3

What Actually Matters More Than Fever Height

Clinical Appearance

• Only 58% of infants with bacteremia or bacterial meningitis appear clinically ill, meaning clinical appearance alone cannot exclude SBI 1
• Infants with Yale Observation Scale scores ≥21 ("very ill-appearing") had 40% SBI rate versus 10% in those with scores <10 1

Duration of Fever

• Fever duration is more predictive than fever height: infants aged 2-6 months with SBI had significantly longer mean fever duration (approximately 26 hours) compared to those without SBI (approximately 19 hours) 3
• Children with fever ≥5 days had higher SBI incidence (8.4%) compared to those with fever <5 days (5.7%) 4

Other Critical Risk Factors

• Uncircumcised male infants have substantially higher rates of bacteremia and urinary tract infection 1
• Absence of viral source: Children with positive viral testing (RSV, influenza, etc.) had lower SBI rates (4.9%) compared to those with negative viral testing (13.5%) 1
• Age-specific risk: Neonates 0-28 days have 13% SBI incidence, infants 29-56 days have 9% incidence 1

Practical Clinical Algorithm

For Infants <60 Days

• Comprehensive evaluation and hospitalization with empirical antibiotics are required regardless of fever height 3
• This includes blood culture, urine culture (catheterized specimen), and lumbar puncture 1

For Infants 2-6 Months

• Consider urinalysis and urine culture for fever ≥38°C (100.4°F), especially with risk factors: female <12 months, uncircumcised male, fever duration >24 hours, higher fever (≥39°C), negative respiratory pathogen testing 1
• Use validated risk stratification tools (Rochester or Philadelphia criteria) rather than fever height alone 3
• Add inflammatory markers (CRP, procalcitonin) to clinical assessment: CRP <20 mg/L has negative likelihood ratio of 0.16 for ruling out SBI 4, 5

For All Febrile Children

• Warning signs with high specificity (>0.90) for SBI include: oxygen saturation <94%, tachypnea, chest wall retractions, prolonged capillary refill time >3 seconds 6
• However, absence of warning signs does not reliably exclude SBI (sensitivity 0.92, negative LR 0.34) 4

Critical Pitfalls to Avoid

• Do not use fever height as a standalone decision tool: The modest association (OR 1.5) means many children with high fever will not have SBI, and many with lower fever will 2
• Do not assume antipyretic use excludes serious infection: Antipyretic use in the previous 4 hours may result in normal or lower temperature at presentation 1
• Do not assume viral infection excludes bacterial infection: The presence of one viral infection does not preclude coexisting bacterial infection 1
• Consider non-infectious causes with prolonged fever: Kawasaki disease (fever ≥5 days), malignancies, and inflammatory conditions must be considered 4

The Bottom Line

Fever height provides limited discriminatory value for SBI risk and should be integrated with clinical appearance, fever duration, age-specific risk factors, viral testing results, and inflammatory markers rather than used as an isolated predictor 3, 2, 4.