Fever is the most common chief complaint of children presenting to the ED, accounting for ~30% of pediatric outpatient visits. It is critical to differentiate mildly ill from seriously ill children with fever, especially in the neonate and infant. This challenge is compounded by the nonspecific symptoms and lack of a focus of infection in most children with fever. Many factors influence evaluation and management, including clinical assessment, physical examination findings, patient age, immunization status, and height of the fever.

This chapter focuses on the management of a neonate, infant, or child with acute fever at risk for serious bacterial illness, because morbidity and mortality are high if not properly treated. Neonates are infants <1 month old. For preterm neonates, the age should be calculated from the date of term birth, rather than from the actual preterm birth date. The significance of age groups is discussed in the subsequent sections.

Any elevation in temperature above normal is considered a fever, but the threshold for clinically important fever varies with the age group and is related to the ability of signs and symptoms to identify the underlying cause of fever. In the neonate or infant <2 to 3 months of age, the threshold for concerning fever is 38°C (100.4°F); in infants and children 3 to 36 months old, the threshold has traditionally been 39°C (102.2°F).¹ In children >36 months old, the definition of significant fever is not fixed because concern for serious bacterial illness in this age group should be directed by other signs or symptoms of the underlying cause. In children with developmental delay, with limited ability to demonstrate specific signs and symptoms, the cause of fever may be difficult to determine, and more testing is often necessary.

Axillary temperatures are 0.6°C (1°F) lower than oral temperatures, which are 0.6°C (1°F) lower than rectal temperatures. Temperatures taken with infrared thermometers that scan the tympanic membrane are of variable reliability and reproducibility.²

Fever is treated with acetaminophen or ibuprofen. The dosage of acetaminophen is 15 milligrams/kg/dose (maximum daily dose, 80 milligrams/kg) every 4 to 6 hours, up to five times per day. Acetaminophen can be given PO or PR. The dosage of ibuprofen is 10 milligrams/kg/dose (maximum daily dose, 40 milligrams/kg) every 6 to 8 hours. Ibuprofen can be given PO or IV and is recommended for children older than 1 year of age.

Infants ≤3 months of age, and especially neonates, are relatively immunodeficient. Neonates and young infants demonstrate decreased opsonin activity, decreased macrophage and neutrophil function, and bone marrow insufficiency.³ Infants and children demonstrate a poor immunoglobulin G antibody response to encapsulated bacteria until 24 months of age. Immune development is a continuum and improves as the child matures. Therefore, the age of the patient and the virulence of the bacteria are considerations for the evaluation of fever in children and the identification of serious bacterial illness. The most common manifestations of serious bacterial illness in children are discussed: urinary tract infection (UTI), bacteremia and sepsis, pneumonia and sinusitis, and meningitis. Of note, the following discussion applies primarily to Western countries.

URINARY TRACT INFECTION

Pathophysiology

Overall, the most common serious bacterial illness is UTI with or without pyelonephritis (see chapter 132, Urinary Tract Infection in Infants and Children). Among young children presenting to EDs with fever and no obvious source of infection, between 3% and 8% have UTI.⁴ The overall incidence of UTI is 5% in children between 2 months and 2 years old.⁵ Uncircumcised boys have a rate of UTI 5 to 20 times greater than circumcised boys. The presence of a fever of 39°C (102.2°F) and a urine suggestive of infection indicate renal parenchymal involvement, or pyelonephritis. Additional risk factors in boys include history of previous UTI, ill appearance, age less than 12 months, fever for at least 2 days, absence of another source for fever, and nonblack race.⁶ After 2 years of age, UTI remains a frequent bacterial cause of fever in girls but is more commonly associated with urinary symptoms. UTI is uncommon in boys older than 1 to 2 years of age, unless underlying risk factors exist.

Escherichia coli and other gram-negative bacteria are the most common causative organisms, although gram-positive organisms comprise a significant minority in older boys and in children with underlying medical conditions such as neurogenic bladder. UTIs may not produce symptoms other than fever, so routinely obtain a urinalysis and culture in the evaluation of the febrile neonate or infant without other source.

Diagnosis

The ideally obtained urine specimen for a child in diapers has traditionally been by urethral catheterization or suprapubic aspiration. In children with labial adhesions or phimosis, a bag collection specimen may be preferred as a screening test. However, if the urinalysis is positive from a bagged specimen, obtain a urine specimen for culture by suprapubic aspiration or clean-catch midstream method before initiating antibiotic therapy, because bag collection methods produce frequent false-positive cultures (up to 88%) from skin contamination.⁵

The initial diagnosis of UTI is made with chemical strip testing or a microscope urinalysis (see Table 132-4). Chemical testing detects leukocyte esterase or nitrites. A positive test for leukocyte esterase has a sensitivity of 67% to 85% for UTI, whereas a positive test for nitrites has a specificity of 95% to 99% for UTI. A positive test for leukocytes on microscope urinalysis testing is a urine WBC count of 5 to 10/high-power field and has a sensitivity of 51% to 91%. The identification of bacteria on Gram stain has a sensitivity and specificity of 80% to 97% and 87% to 99%, respectively. When Gram stain is not readily available, chemical and microscopy testing compares favorably with gram stain.

Before beginning antibiotic treatment, obtain an appropriate urine sample for culture and susceptibility testing. Consider blood and cerebrospinal fluid testing in young infants suspected to have a UTI. Approximately 5% to 10% of febrile infants with UTI will have bacteremia.^7,8 UTIs can be associated with bacteremia in up to 30% of infants between 4 and 8 weeks of age.⁹

One study reported that 13% of infants (15 of 117 infants) <3 months of age with a febrile UTI admitted to the hospital had a sterile pleocytosis of the cerebrospinal fluid thought to be due to systemic release of inflammatory mediators or low bacterial virulence in the subarachnoid space.^{7,8,9,10,11,12} Less than 1% of febrile infants with UTI will have a bacterial meningitis, but concomitant infection of the urine and cerebrospinal fluid has been reported.

A recent American Academy of Pediatrics practice parameter on UTI suggested that patients with a negative microscope urinalysis or chemical strip testing in association with a positive urine culture are likely to have asymptomatic bacteriuria, rather than a true UTI.⁵ Based on this reasoning, they suggested that negative urine microscope urinalysis or chemical strip testing does not warrant ordering urine culture. However, the recommendations are Level C, based on previous studies not directly relevant to the infant or child presenting to the emergency department with fever.^13,14 Therefore, if a patient is at risk for UTI, obtain a urine culture even if the initial urinalysis is negative.

BACTEREMIA AND SEPSIS

Most studies of febrile infants ≤3 months old cite a bacteremia/sepsis incidence of 2% to 3%. The most common causes of bacteremia and meningitis in this age group are E. coli, group B Streptococcus, and Listeria monocytogenes. Ill-appearing neonates or those identified at high risk because of laboratory testing have an incidence of serious bacterial illness of 13% to 21%.¹⁵ Overall, however, viral infections are the most frequent cause of fever in infants.

Before the widespread use of the pneumococcal conjugate vaccine, in febrile infants and children between 3 and 36 months old, high fever, WBC >15,000/mm³, and absolute neutrophil count >10,000/mm³ were independent predictors of occult bacteremia. The presence of any of these factors increased the incidence of bacteremia to 8% to 17%.

Administration of the Haemophilus influenzae type b vaccine and the heptavalent pneumococcal conjugate vaccine has decreased the occult bacteremia rate of well-appearing, febrile children 3 to 36 months of age from approximately 2% to 3% to 0.5% to 0.7%.¹⁶ The Centers for Disease Control and Prevention reports a 76% reduction in invasive infections from Streptococcus pneumoniae when comparing 2005 with 1998 data in the United States. The incidence of serious bacterial illness in children 2 to 6 months old, who were incompletely or not immunized, decreased because the widespread use of the vaccine resulted in herd immunity.¹⁷ In 2009, a decavalent pneumococcal conjugate vaccine was released in Europe, and in 2010, the 13-valent pneumococcal conjugate vaccine was introduced in the United States, both of which are expected to further decrease the incidence of pneumococcal disease.¹⁸ Ongoing Centers for Disease Control and Prevention Active Bacterial Core Surveillance cites declining national estimates of invasive disease.¹⁹ Given these declines and the fact that 80% of pneumococcal bacteremia resolves spontaneously, the traditional standards for routine evaluation of the febrile infant 3 to 36 months old will be changing as the prevalence of occult bacteremia decreases.²⁰

PNEUMONIA AND SINUSITIS

Pneumonia and sinusitis are common bacterial infections of childhood, frequently associated with or following upper respiratory tract symptoms (see chapter 120, Nose and Sinus Disorders in Infants and Children; chapter 121, Mouth and Throat Disorders in Infants and Children; and chapter 125, Pneumonia in Infants and Children). Pneumonia occurs in all age groups, with the most common causative agents being the same as those for bacteremia or meningitis in each age group. The incidence of pneumococcal pneumonia in all ages has decreased since the introduction of the pneumococcal conjugate vaccine.^21,22,23 Sinusitis is uncommon in children <3 years of age because sinus formation is incomplete.

Plain chest radiographs remain the gold standard for diagnosis of pneumonia. In neonates and young infants, routine chest radiographs are not necessary unless the patient has specific physical examination findings suggestive of pneumonia, such as respiratory distress, rales, grunting, significant tachypnea, or hypoxemia.^7,24 In older children with chronic medical problems, such as cystic fibrosis, congenital heart disease, or malignancy, consider pneumonia in the differential diagnosis of fever and upper respiratory tract symptoms, even if there are no signs of lower tract infection. In one study performed before the widespread use of the pneumococcal conjugate vaccine, a WBC count of 20,000/mm³ was associated with occult pneumonia in 19% of patients without focal findings.²⁵ Without predisposing conditions or abnormal test results, the decision to obtain a chest radiograph can otherwise be made clinically. Pneumonia in a febrile but otherwise asymptomatic child is unlikely.

MENINGITIS

In Western countries, most studies of febrile infants <3 months old cite a bacterial meningitis incidence of 1%. The most common organisms are the same as those for bacteremia/sepsis: E. coli, group B streptococci, and L. monocytogenes. For children >3 months old, the most common organisms are S. pneumoniae, Neisseria meningitidis, and Staphylococcus aureus, with a lower incidence of S. pneumoniae meningitis since routine vaccinations with the conjugate vaccine.

Outside of North America, the epidemiology of meningitis is more complex, depending on the region of the world in which the patient has been living or traveling. N. meningitidis and Mycobacterium tuberculosis are the more common causes. Although N. meningitidis rarely occurs in North America (groups B, C, W135, X, and Y), different serogroups exist elsewhere, especially group A in sub-Saharan Africa. In patients with symptoms of meningitis and recent travel to Africa, meningococcal meningitis should be considered as a diagnosis. Treatment is similar to other causes of bacterial meningitis. Tuberculosis meningitis is discussed in chapter 174, Central Nervous System and Spinal Infections.

The diagnosis of meningitis is made by obtaining cerebrospinal fluid by lumbar puncture (see later section, Procedures in Children: Lumbar Puncture). It is often difficult to distinguish between viral and bacterial meningitis because there is a wide overlap in cerebrospinal fluid and peripheral blood findings. Cerebrospinal fluid WBC >30 cells/mm³ in the neonate and >10 cells/mm³ in children >1 month old have been traditional markers to suggest meningitis. Risk factors for bacterial meningitis in children 29 days to 18 years old are listed in Table 116-1.²⁶ Each risk factor counts as a single point added toward the bacterial meningitis score. A negative bacterial meningitis score does not exclude some treaTable causes of meningitis/encephalitis such as herpes virus or Lyme disease.

Because of the serious morbidity of missed meningitis, it is best to admit any ill-appearing patient or patients <2 months old with any degree of pleocytosis and administer appropriate antibiotics in the ED.

The pathogenesis of bacterial meningitis suggests that steroids may attenuate the inflammatory response associated with meningitis. However, the use of steroids has been associated with decreased bactericidal activity by some antibiotics and decreased antimicrobial penetration into the cerebrospinal fluid. Studies are conflicting. However, a recent Cochrane review found no difference in mortality among studies of children (162/1229 in the steroid group vs. 166/1202 in the placebo group), but did suggest a positive effect in the prevention of hearing loss among pediatric patients (relative risk, 0.74; 95% confidence interval, 0.62 to 0.89) with the strongest effect seen in cases of H. influenzae. Subgroup analysis showed possible survival benefit from steroids in meningitis caused by S. pneumoniae when adults and children were combined.²⁷ If given, administer steroids before or during antibiotic administration.

Infants with aseptic meningitis generally should be hospitalized and ensured adequate long-term follow-up because they are at greater risk for dehydration and subsequent neurologic and learning disabilities.

For those with cerebrospinal fluid pleocytosis and likelihood of viral meningitis, if the child is to be discharged from the ED, it is wise to administer a long-acting parenteral antibiotic (ceftriaxone, 100 milligrams/kg IM or IV) and ensure follow-up in 24 hours.²⁶

The clinical challenge is to distinguish the cause of fever: a benign viral infection, serious bacterial illness, or a noninfectious illness. Most causes are due to viral infections, but bacterial infections are not infrequent. The significance of fever depends on multiple factors. If the physical examination identifies the source of infection, evaluation, testing, and treatment are dictated by the presumptive diagnosis. If the physical examination does not identify a source of infection causing fever, decision making is based first on age and then by height of fever. There are no absolute rules in the evaluation and management of fever, but the guidelines in Table 116-2 are suggested for the management of neonates, infants, and children who are well appearing, have had all relevant immunizations, and have no obvious cause for the fever. Again, this discussion should be applied judiciously to non-Western countries or patients recently emigrated from non-Western countries, because the epidemiology of fever may be more diverse in international settings and is beyond the scope of this chapter. Detailed discussion of evidence-based information is provided later in the section, Decision Rules for Assessment of Fever in Neonates and Young Infants. Any ill-appearing infant or child should have a complete sepsis evaluation performed and should be admitted for parenteral antibiotic therapy.