Diagnostic Accuracy of Clinical Signs and Symptoms for Serious Bacterial Infection in Young Febrile Children

Abstract & Commentary

By Hal B. Jenson, MD, FAAP, Professor of Pediatrics, Tufts University School of Medicine, and Chief Academic Officer, Baystate Medical Center, Springfield, MA, Dr. Jenson reports no financial relationships relevant to this field of study., is Associate Editor for Infectious Disease Alert.

Synopsis: Serious bacterial infections occur in about 7% of febrile children younger than five years of age presenting to the emergency department, and pose substantial diagnostic difficulties. In this study, antibiotics were prescribed acutely in only 66%-81% of cases of serious bacterial infections.

Source: Craig JC, et al. The accuracy of clinical symptoms and signs for the diagnosis of serious bacterial infection in young febrile children: Prospective cohort study of 15,781 febrile illnesses. BMJ. 2010;340:c1594.

A two-year prospective cohort study of children less than five years of age presenting with a febrile illness (≥ 38.0° C) to a single emergency department in Australia, used a standardized clinical evaluation with a mandatory entry for 40 clinical features between July 1, 2004, and June 30, 2006. Greater than 94% of the children with serious bacterial infections (defined as urinary tract infection, pneumonia, or bacteremia) had the appropriate tests (urine culture, chest radiograph, and blood culture). Follow-up data were available for 93% of the 15,781 cases of febrile illnesses.

Most febrile illnesses occurred in children under three years of age, with the peak between 1 and 2 years of age. There were 543 cases (3.4%) of urinary tract infection, 533 cases (3.4%) of pneumonia, and 64 cases (0.4%) of bacteremia, with a combined prevalence of 7.2%. Antibiotics were prescribed acutely in 66% (359/543) of children with urinary tract infection, 69% (366/533) with pneumonia, and 81% (52/64) with bacteremia. However, 20% (2686/13,557) of children without a serious bacterial infection were also prescribed antibiotics. Comparing data from the clinical evaluations with the confirmed diagnosis, physicians' clinical diagnosis of urinary tract infection, pneumonia, or bacteremia had low sensitivity (10%-50%) and high specificity (90%-100%).

The variable that was most predictive for presence of any serious bacterial infection was "appearing generally unwell," followed by increased temperature, no fluid intake in the previous 24 hours, increased capillary refill time, and pre-existing chronic disease. The presence of localizing symptoms and signs was significant for pneumonia and urinary tract infection but not for bacteremia.


Healthy children less than five years of age typically have 5-8 acute upper respiratory tract infections, 0-1 cases of acute otitis media, and 0-2 cases of enteritis each year. In most cases, these febrile illnesses are self-limited, presumably viral infections. It is often very difficult to clinically distinguish the non-serious viral infections from serious bacterial infections that require early antibiotic therapy. This is underscored in this study by the acute treatment of serious bacterial infections with antibiotics in only 66%-81%. This is not because the diagnoses are not considered (94% of the children in this study had the appropriate diagnostics tests) but because of the absence of objective information that identifies serious bacterial infection early, often with fever being the only finding. Difficulty interpreting and suboptimal sensitivity and specificity of the urinalysis for urinary tract infection and chest radiograph for pneumonia further complicate early recognition. Physicians may underestimate the risk of infection in an individual patient because it is relatively low (7.2% in this study). Management of serious bacterial infection in this age group is also confounded because many of these serious bacterial infections can resolve without antibiotic treatment, such as occult bacteremia that occurs in children less than three years of age, which may give some providers a false sense of security.

The clinical variable with the strongest diagnostic significance for all serious bacterial infections was "appearing generally unwell." This is consistent with many other studies, and underlies the "art" of pediatrics in managing these children. The clinical features demonstrated in this study to be sensitive and specific for diagnosis of serious bacterial infections include the overall appearance of the child, urinary symptoms for urinary tract infection, and cough for pneumonia.