What defines fever and what temperature thresholds indicate fever for adults, hospital‑acquired infection screening, neutropenic patients, elderly in long‑term care, and children, and how should core temperature be measured to confirm fever?

Definition of Fever and Temperature Thresholds

Fever is defined as a single temperature measurement ≥38.3°C (101°F) in adult ICU patients, but the specific threshold varies by patient population and clinical context. 1, 2

Population-Specific Fever Definitions

Adult ICU Patients

• Single temperature ≥38.3°C (101°F) is the standard definition endorsed by the Society of Critical Care Medicine and Infectious Diseases Society of America 1, 2
• This threshold should guide diagnostic workup in critically ill adults 1

Hospital-Acquired Infection Screening

• Temperature >38.0°C (100.4°F) per CDC criteria 2
• This lower threshold increases sensitivity for detecting nosocomial infections 2

Elderly Patients in Long-Term Care Facilities

The definition is more nuanced for this population due to blunted fever responses:

• Single oral temperature >37.8°C (100°F) (70% sensitive, 90% specific for infection) 1, 3
• OR repeated oral temperatures >37.2°C (99°F) 1
• OR rectal temperatures >37.5°C (99.5°F) 1
• OR increase from baseline >1.1°C 1

Neutropenic/Immunocompromised Patients

These patients require aggressive evaluation at lower thresholds:

• Single oral temperature ≥38.3°C (101°F) 1, 2
• OR temperature ≥38.0°C (100.4°F) sustained over 1 hour 1, 3

Pediatric Patients

• Rectal temperature ≥38.0°C (100.4°F) for children younger than 3 years 2, 3
• Rectal measurement is preferred in neonates and young children 2

Optimal Temperature Measurement Methods

Hierarchy of Accuracy (Most to Least Reliable)

When accurate temperature is critical to diagnosis and management, central monitoring is strongly preferred: 1

1. Pulmonary artery catheter thermistors (gold standard) 1, 3
2. Bladder catheter thermistors (essentially identical to intravascular readings, less invasive, provide continuous monitoring regardless of urine flow) 1, 3
3. Esophageal balloon thermistors (comparable accuracy but uncomfortable in alert patients, difficult to confirm placement) 1, 3
4. Rectal thermometers (a few tenths of a degree higher than core, not predictably consistent but acceptable when central monitoring unavailable) 1
5. Oral thermometers (safe and convenient for alert, cooperative patients but distorted by mouth breathing, hot/cold fluids, or intubation) 1, 3

Avoid these unreliable methods: 1, 3

• Axillary measurements
• Tympanic membrane/infrared ear thermometers (often 1-2 degrees different from core temperature)
• Temporal artery thermometers
• Chemical dot thermometers

Practical Recommendation

• Use central thermometers when devices are already in place or when accurate measurement will influence management 1
• For patients without central monitoring, use oral or rectal temperatures 1, 2
• Never rely on tympanic, axillary, or temporal artery measurements in critical care settings 1, 3

Critical Clinical Pitfalls

Absence of Fever Does Not Exclude Serious Infection

A substantial proportion of infected patients are not febrile, and absence of fever is associated with worse outcomes. 1, 2, 4

High-risk populations for absent/blunted fever response include:

• Elderly patients 1, 4
• Patients with open abdominal wounds or large burns 1, 3, 4
• Patients receiving extracorporeal membrane oxygenation or continuous renal replacement therapy 1, 3, 4
• Patients with congestive heart failure, end-stage liver disease, or chronic renal failure 1, 4
• Patients taking anti-inflammatory drugs, corticosteroids, or antipyretics 1, 4
• Immunosuppressed patients 4

Alternative Infection Indicators When Fever is Absent

Evaluate for these signs even without temperature elevation: 1, 3

• Unexplained hypotension, tachycardia, or tachypnea 1, 3
• New confusion or altered mental status 1, 3
• Rigors or new skin lesions 1, 3
• Oliguria or lactic acidosis 1, 3
• Leukocytosis, leukopenia, or ≥10% immature neutrophils (bands) 1, 3
• Thrombocytopenia 1

Measurement Inconsistency

Using different measurement methods can yield significantly different readings and lead to missed diagnoses or inappropriate interventions. 2, 4

• Establish a consistent measurement protocol for each patient 2
• Document the measurement site with each temperature reading 1
• Be aware that normal body temperature has been decreasing by 0.03°C per birth decade over the last 157 years 2, 4

Age-Specific Considerations

Failure to use age-appropriate fever definitions is particularly problematic in elderly patients who have lower baseline temperatures. 2, 3

• In older adults, suspect infection even with temperatures that would be "normal" in younger patients 1, 2
• Consider functional decline (new confusion, incontinence, falls, reduced mobility, decreased food intake) as infection indicators in the elderly 2