, et al
Lack of accuracy of body temperature for detecting serious bacterial infection in febrile episodes
Ped Infect Dis J
; doi:

Researchers at the University of Sydney, Australia, sought to determine whether the height of fever in young children is associated with serious bacterial infection (SBI). They also ascertained whether the duration of the fever, the age of the child, or the child’s appearance at presentation modified this association.

Between 2004 and 2006, consecutive children aged <5 years who presented with fever to the emergency department (ED) at the Children’s Hospital at Westmead in Sydney were enrolled in the study. Fever was defined as a measured axillary temperature of ≥38.0° C in the ED, parent report of temperature of ≥38.0°C, or subjective fever (“felt hot”) within 24 hours preceding the visit. Researchers also collected the highest temperature recorded by the caregiver in the previous 24 hours, clinical appearance of the patient (well/mildly unwell, moderately unwell, very unwell), and duration of fever. SBI was defined as either a positive blood or urine culture or a chest radiograph interpreted by a radiologist as indicative of pneumonia. Patients were followed until the diagnosis of SBI was made or fever had been resolved for ≥24 hours. Results were analyzed using receiver operating characteristics (ROC) curves with associated area under the curve (AUC) and logistic regression to test for interaction between temperature and age, clinical appearance, and illness duration.

Among 15,781 febrile episodes, 1,120 (7.1%) SBIs were diagnosed (3.4% urinary tract infection, 3.4% pneumonia, and 0.4% bacteremia). The AUC for temperature was a modest 0.60 (95% CI, 0.58–0.62). Although more maximum temperatures of ≥39°C were reported by caregivers than were recorded at time of presentation (42.5% vs 23.5%; P < .001), caregiver-reported temperatures did not perform better than ED temperatures in identifying children with SBI. Clinician assessment of general appearance did not alter the test characteristics of temperature in detecting SBI. Younger children and those with a longer duration of illness were more likely to have SBI. Both age and duration of illness were associated with small but significant changes in the AUC for temperature. Despite some increases in specificity of temperature for predicting serious bacterial illness when factoring in age and duration of illness, the overall sensitivity and specificity of temperature for indicating SBI remained poor.

The authors conclude that in febrile children presenting to an ED, temperature is an inaccurate indicator of SBI.

Dr Springer has disclosed no financial relationship relevant to this commentary. This commentary does not contain a discussion of an unapproved/investigative use of a commercial product/device.

When faced with a febrile child, the pediatrician must choose an appropriate diagnostic evaluation, and is appropriately concerned with missing the occasional febrile child who truly is seriously ill. Multiple clinical guidelines have been published around the world1–3  and offer variable recommendations for management based on...

You do not currently have access to this content.