Fever in the Infant and Toddler Workup

Diagnostic studies in infants and toddlers with fever are based on their age groups. Febrile neonates (< 28d) and young infants (28-60d) may require a full sepsis workup. Young infants are generally assessed for urinary tract and respiratory infections as well as their risk for serious bacterial infections. Febrile children aged 3 months to 3 years are evaluated based on epidemiologic and focal findings revealed during the history taking and physical examination as well as whether or not these children are at low risk for serious bacterial infections.

A full sepsis evaluation is often recommended in febrile neonates and young infants. This includes a complete blood cell (CBC) count, blood culture, urinalysis, urine culture, and cerebrospinal fluid (CSF) analysis and culture. These patients should be hospitalized with intravenous antibiotics pending results of these cultures.

A study by Cruz et al analyzed the accuracy of individual complete blood cell count parameters to identify febrile infants with invasive bacterial infections. The study included 4313 infants, 1340 (31%) were aged 0 to 28 days, of which ninety-seven (2.2%) had an invasive bacterial infection. The study reported low sensitivities for common CBC parameter thresholds. WBC less than 5000/µL, was detected 10% of the time, white blood cell count ≥15 000/µL, 27%; absolute neutrophil count ≥10 000/µL, 18%; and platelets < 100 x103 /µL, 7%. ^[10]

For well-appearing febrile infants aged 8-21 days, clinical practice guidelines from the American Academy of Pediatrics (AAP) recommend a complete evaluation for invasive bacterial infection that includes urine, blood, and CSF for culture. ^{[11, 12]} This approach considers all infants in this age cohort as high risk. Urine samples for culture from these young infants should be obtained by catheterization or suprapubic aspiration.

According to the AAP guidelines, well-appearing, febrile infants aged 22-28 days are at intermediate risk. ^{[11, 12]} The recommendation for these infants is to obtain a urine specimen by catheterization or suprapubic aspiration for both urine analysis and culture. Blood culture, as well as inflammatory marker levels, should also be obtained from all infants in this age group. The AAP recommends considering a procalcitonin value of 0.5 ng/mL or higher as positive, which indicates that the infant is at greater risk for invasive bacterial infection and possibly should undergo an extensive workup. (See Guidelines for more detail.)

CSF studies

A lumbar puncture for CSF examination is recommended in all neonates younger than 28 days if empiric antibiotics are to be given or if the neonate had a seizure. CSF should be assessed for WBC count and differential, glucose level, protein level, Gram stain, and routine culture. CSF should be assayed for herpes simplex virus (HSV) using polymerase chain reaction (PCR) in all neonates in the first 28 days of life who appear ill, who have mucocutaneous lesions, or who have had a seizure.

Enterovirus PCR analysis should be performed on the CSF during the summer enteroviral season.

Urine and stool studies

Because the incidence of urinary tract infections (UTIs) is high in this age group, a urine specimen should be obtained for urinalysis and urine culture. A negative urine dipstick or urinalysis finding alone does not exclude the diagnosis of a UTI; only a negative urine culture finding can exclude this diagnosis. ^[13] A urine culture should be obtained via either a suprapubic aspiration or urethral catheterization, because bag urine specimens are frequently contaminated.

A study by Tzimenatos et al that included an analysis of data from 4147 febrile infants ≤60 days old reported that for the 289 infants with a UTI and colony counts ≥50 000 CFUs/mL, a positive urinalysis regardless of bacteremia showed sensitivities of 0.94; 1.00 with bacteremia; and 0.94 without bacteremia. Specificity in all groups was 0.91. ^[14]

A stool culture is recommended when blood, mucus, or both are present in the stool; when diarrhea is present; and when more than 5 white blood cells (WBCs) per high-power field (HPF) are noted on methylene blue stain of fresh stool.

Pulmonary studies

A chest radiograph should be considered for neonates with signs of respiratory illness such as cough, coryza, tachypnea, rales, rhonchi, retractions, grunting, nasal flaring, or wheezing. During respiratory viral season, an attempt should be made to identify a respiratory viral etiology using direct fluorescent antigen (DFA) detection or PCR and viral culture on nasal wash specimens.

In infants older than 28 days, low risk criteria are well defined. The reference range white blood cell (WBC) count is 5,000-15,000 cells/μL. The band count should be less than 1500 cells/μL. However, the WBC count alone has poor sensitivity and specificity for identifying young infants with bacteremia and meningitis; thus, the decision to perform a sepsis workup should not be based on the WBC count alone.

Urine and stool studies

Because the incidence of urinary tract infections (UTIs) is still high in this age group, obtain a urine specimen for urinalysis and urine culture. In one study, only 20% of febrile infants with the diagnosis of pyelonephritis had pyuria. A urine culture should be obtained by either a suprapubic aspiration or urethral catheterization, because bag urine specimens are frequently contaminated. 

A stool culture is recommended when blood, mucus, or both are present in the stool, when diarrhea is present, and when more than 5 WBCs per high-power field (HPF) are noted on methylene blue stain of fresh stool.

Pulmonary studies

Chest radiography should be considered for infants with signs of respiratory illness, such as cough, coryza, tachypnea, rales, rhonchi, retractions, grunting, nasal flaring, or wheezing. During respiratory viral season, an attempt should be made to identify a respiratory viral etiology using direct fluorescent antigen (DFA) detection or polymerase chain reaction (PCR) and viral culture on nasal wash specimens.

AAP guidance

For well-appearing, febrile infants aged 29-60 days, a urine evaluation is recommended. The AAP suggests a two-step approach may be used: (1) obtaining a urine analysis by a noninvasive method and (2) obtaining a culture only if the urinalysis results are positive. However, only catheter or suprapubic specimens are appropriate for urine culture. ^{[11, 12]}

The AAP recommends that clinicians obtain blood cultures for all infants aged 29-60 days. Inflammatory markers should also be assessed, because lumbar puncture for CSF culture may be avoided in these infants if the tests for inflammatory markers are negative. If CSF is obtained, enterovirus testing is recommended.

Criteria for low risk of occult bacteremia

Although some experts consider a lumbar puncture optional in well-appearing infants with low-grade fever who are older than 28 days, a diagnosis of meningitis carries a significant risk of morbidity and mortality, and acceptable risk must be determined for each individual patient.

Various criteria have been developed in an attempt to identify the infant older than 28 days at low risk for a serious bacterial infection. The incidence of a serious bacterial infection in infants categorized as low risk after a full evaluation is small (0.5% in studies that included a lumbar puncture and 1.1% in studies without a lumbar puncture).

These studies have led to the development of the Rochester, Boston, and Philadelphia protocols, all of which were conducted in urban emergency departments. The Rochester and Boston criteria exclude patients with ear infections from low-risk groups, and the Philadelphia criteria considers low-risk infants those with an "unremarkable physical examination."

However, subsequent studies have shown that, when these criteria are applied to the neonatal population (age, 1-28 d), an increased number of infants with serious bacterial illnesses are misidentified as low risk when compared with infants aged 1-3 months. Neonates should be considered high risk and a complete sepsis evaluation should be performed. All febrile neonates should be hospitalized and should receive empiric antibiotic therapy.

The following criteria have not been universally adopted by community practitioners. The importance of reliable follow-up cannot be overstressed if the decision is made not to perform invasive studies in a febrile, well-appearing infant; no guidelines for the minimal evaluation of a febrile, well-appearing infant are recognized. If empiric antibiotics are given, a lumbar puncture should always be performed.

Rochester criteria

The Rochester criteria are used to assess febrile (temperature > 38°C) infants aged 28-60 days. ^[15] The risk for occult bacteremia in well-appearing febrile infants is 7-9%; if all Rochester criteria are present, the risk is less than 1%. Infants at high risk were hospitalized with empiric antibiotics, and infants at low risk were discharged with follow-up in 24 hours. ^[15]

The Rochester low-risk criteria for occult bacteremia include the following ^[15] :

• Infant must appear generally well

• Infant has been previously healthy: Born at term (≥ 37 weeks' gestation), no perinatal antimicrobial therapy, no treatment for unexplained hyperbilirubinemia, no previous antimicrobial therapy, no previous hospitalization, no chronic or underlying illness, not hospitalized longer than the mother

• Infant has no evidence of skin, soft-tissue, bone, joint, or ear infection

• Infant has the following laboratory values: WBC count of 5,000-15,000 cells/μL, absolute band count of 1,500 bands/μL or less, less than 10 WBCs/HPF on microscopic examination of the urine, less than 5 WBCs/HPF on microscopic examination of stool in an infant with diarrhea

Boston criteria

The Boston criteria are used to assess febrile (temperature > 38°C) infants aged 28-89 days. ^[16] Infants who met these criteria were managed as outpatients with 50 mg/kg ceftriaxone intramuscularly at the time of discharge. The scheduled follow-up visit was in 24 hours; 5.4% of patients had a serious bacterial infection at follow-up.

The Boston low-risk criteria for occult bacteremia are as follows ^[16] :

• No immunizations or antimicrobials within the preceding 48 hours

• No evidence of dehydration or ear, soft-tissue, or bone infection

• Well appearing

• Caretaker available by telephone

• Infant has the following laboratory values: WBC count less than 20,000 cells/μL, CSF with WBC count less than 10 cells/μL, ^[17] urinalysis with less than 10 WBCs/HPF on microscopic examination, no infiltrate on chest radiograph (if one was obtained)

Philadelphia criteria

The Philadelphia criteria were used to assess febrile infants aged 29-60 days with fever (> 38.2°C). All high-risk patients were hospitalized and treated with empiric antibiotics. Low-risk patients were not treated with antibiotics, with follow-up in 24 hours. Sensitivity for identifying patients with a serious bacterial infection was 98%, specificity was 42%, positive predictive value was 14%, and the negative predictive value was 99.7%.

The Philadelphia low-risk criteria for occult bacteremia included the following:

• Well appearing

• WBC count of less than 15,000 cells/μL

• Band-neutrophil ratio of less than 0.2

• Urinalysis reveals less than 10 WBC/HPF on microscopic examination and a negative urine Gram stain result

• CSF has less than 8 WBCs/μL and a negative CSF Gram stain finding

• Chest radiograph does not have an infiltrate (if a radiograph was obtained)

• Stool has no blood and few or no WBCs on the smear

The laboratory evaluation of children aged 3 months to 3 years consists of 2 parts.

The first part is dictated by epidemiologic and focal findings uncovered by the history and physical examination. Children with severe cough and respiratory distress should undergo chest radiography as part of their evaluation; children with limp or evidence of focal infection should undergo appropriate imaging studies.

The second part of the evaluation is designed to identify patients at low risk for serious bacterial infections. Children who appear clinically ill, have a history of vomiting, have chest wall retractions with tachypnea, or have delayed capillary refill times are at increased risk for bacterial infection. Increasing fever also increases the risk of bacterial infection. Children aged 3-24 months with fever exceeding 39^o C and children aged 24-36 months with fever exceeding 39.5^o C have a higher risk of bacteremia than children with lower temperatures. ^[18]

Urine studies

Occult urinary tract infection (UTI) is the most common cause of unexplained fever and bacteremia in females and uncircumcised males in this age group; thus, all of such children should undergo a urine culture. This is the only routine laboratory study recommended by most experts. Circumcised males should have a urinalysis and if abnormal, culture of the urine.

Catheterized, unspun urine samples with more than 10 WBCs per high-power field (HPF) or bacteria found in any of 10 oil immersion fields of a Gram-stained sample is highly suggestive of infection. The presence of leukocyte esterase or nitrites by dipstick of unspun samples also suggests infection. Febrile children with suspicious urinalysis findings require cultures of both urine and blood. Urine cultures that yield more than 50,000 colony-forming units/mL are diagnostic of infection. ^[19]

Laboratory low-risk criteria for occult bacteremia

White blood cell (WBC) counts, absolute neutrophil counts (ANC) (see the Absolute Neutrophil Count calculator), serum C-reactive protein (CRP) concentration, and serum procalcitonin concentration are additional metrics that help to distinguish further children at low risk for serious bacterial infection.

Total WBC counts below 15,000 cells/μL, ANCs below 10,000 neutrophils/μL, CRP concentrations less than 40 mg/L, and serum procalcitonin levels less than 0.5 ng/mL identify children at low risk for serious bacterial infection. ^{[6, 20, 21, 22]}

Laboratory findings suggestive of serious bacterial infections include the following:

• Urinalysis (unspun): More than 10 WBCs/HPF, bacteria in any of 10 HPFs, or positive leukocyte esterase and nitrite findings

• WBC count: More than 15,000 cells/μL

• ANC: More than 10,000 neutrophils/μL

• CRP level: More than 40 mg/L

• Procalcitonin level: More than 0.5 ng/mL