Chapter 3: The Febrile- or Septic-Appearing Infant or Child

• Children typically have four to six febrile illnesses in the first two years of life with most of these being nonspecific viral illnesses.

• Widespread vaccination for Haemophilus influenzae type b and Streptococcus pneumoniae has had a profound impact on the incidence of occult bacteremia and serious bacterial illness (SBI).

• Meticulous history and physical examination, most notably the initial visual inspection of the febrile infant or child is the foundation of accurate diagnosis.

• In the 28- to 56-day-old febrile infant, risk stratification criteria can be used to identify those at risk for SBI.

• Focusing parental attention to “controlling the fever” is generally impractical and misguided and contributes to fever phobia.

Fever is one of the most common complaints of children presenting to the emergency department (ED). In fact, children typically have four to six febrile illnesses in the first two years of life. In most cases, the etiology of the fever is either a benign, nonspecific viral infection or a self-limited focal bacterial infection such as otitis media or streptococcal pharyngitis. However, parents and practitioners are always concerned with the possibility of an underlying serious condition, such as meningitis or sepsis, and therefore often approach the febrile child with some degree of trepidation. In addition, the evaluation of the febrile child is complicated by age-related variations such as the patient's ability to communicate, immune system development, and immunization status. Therefore, it is important for the practitioner to understand the underlying pathophysiology of fever, the risk factors associated with serious bacterial illness (SBI), and the relative frequency of certain sepsis and sepsis-like conditions to develop a rational approach to management, including the appropriate use of laboratory testing and antibiotics.

The evaluation of the febrile child is often determined on presentation by either the age of the child, the presence of underlying chronic illness or immunocompromise. Due to the relative lack of clinical signs and an immature immune system, neonates have specific age-appropriate management strategies (see Chapter 2). However, there is practice variation for 28- to 56-day-old infants. They are either included with neonates or they are considered as a separate group and managed uniquely. Fever may be the only presenting symptom of serious bacterial illness in children with underlying illnesses such as sickle cell disease, immunodeficiency, or cancer as well as those children with indwelling devices such as a ventriculoperitoneal shunt or a central line. Therefore, the presence of fever in these children often requires laboratory testing and empiric antibiotic treatment until bacterial infection can be ruled out. However, the majority of well-appearing, otherwise healthy children, require only a thorough history and physical examination, and in the absence of a bacterial source of infection, can be managed by antipyretic therapy and observation until the course of the illness resolves. This chapter deals specifically with the otherwise healthy febrile child between the ages of 1 month and 3 years.

Normal body temperature is generally considered to be 98.6°F (37°C). However, body temperature varies with age, time of day, gender, and other host-specific factors. In addition, nondisease causes of variation include bundling, elevated ambient air temperature, and exercise. Thus, there is no true single value for normal body temperature. However, as an effort to standardize protocols and guidelines, fever in young infants and children is often defined as a body temperature greater than 100.4°F (38°C) rectally.¹

The anatomic site at which the temperature is taken also affects measurement. Commonly used sites include rectal, oral, tympanic, and axillary. Significant variation may occur between temperatures measured externally (axillary) and those measured closer to core areas (rectal), which more accurately and consistently reflect overall body temperature status. Sites other than rectal, such as axillary, oral, tympanic, or skin, as well as use of other devices, such as infrared or liquid crystal, are more convenient and less invasive. They may serve as effective screening tools but are less accurate in the young infant and child.^2,3

Parents often rely on palpation to detect the presence of fever, even if they own a thermometer.^4,5 Although this subjective measure may be accurate for determining the absence of fever, it overestimates the actual presence of fever. Therefore, “fever to touch” is generally unreliable as a sole measure to determine the presence of fever but may be useful to exclude its presence.

Body temperature is tightly regulated. Increases above normal are termed either hyperthermia, if environmentally produced, or fever. Maintenance of body temperature involves a complex interplay of autonomic, endocrinologic, and behavioral mechanisms. The thermoregulatory center is located in the preoptic region of the hypothalamus and maintains the body's “set point.” Damage to this area may result in extreme body temperature variations.

Fever is a host response to combat infection. Exogenous pyrogens are released during viral or bacterial invasion. These pyrogens elicit a complex cascade of processes that result in the elevation of the set point in the thermoregulatory center. As a response, the body increases heat generation and heat conservation to increase the body temperature to the new set point. Since this process takes many minutes to raise the core temperature, it is not unusual to see rigors (shivering, cool extremities due to vasoconstriction) before the detection of fever at a peripheral site. A summary of the febrile response is shown in Figure 3-1.

Much focus has been placed on the height and duration of fever as a predictor of serious bacterial illness. Although the height of the fever is often considered a worrisome sign by parents and clinicians alike, in an otherwise healthy child even temperatures of >39°C have relatively low sensitivity (10%–14%) and predictive value (4%–40%) for serious illness in infants over 6 months and limited usefulness in infants 3 to 6 months.⁶ In fact, other clinical signs such as age, appearance, and peripheral perfusion are better predictors. Similarly, although there may be a weak converse association between duration of fever and bacteremia, overall, the duration of fever is a poor predictor of serious bacterial illness.⁶ Finally, the change in temperature 1 to 2 hours after antipyretic administration does not help identify children with serious bacterial illness.^6–9 However, antipyretic response may allow the clinician to better assess the child's clinical state and therefore aid in determining which child is at risk for serious bacterial illness.

Fever is often a source of anxiety for both parents and clinicians. Parents are often overly concerned with the presence of fever, fearing seizures or brain damage, and do not understand the beneficial aspects. Likewise, clinicians may fear missing the rare occurrence of serious illness or succumb to parental pressure to perform laboratory tests and provide empiric antibiotics. In fact, parents who are “very worried” about fever are more likely to bring their child for medical evaluation, more likely to have blood tests performed on their child, and more likely to perceive their doctors to be very worried about fever.¹⁰ Therefore, parental education should attempt to provide parents with an understanding of the use of antipyretics to treat symptoms and not to cure the disease. When possible, parents should also be brought to understand the recurrent nature of fever and the lack of relationship of height of temperature to the presence of serious illness. In fact, many health care providers believe that the risk of heat-related adverse outcomes during a simple febrile illness is increased with temperatures above 40°C, although this belief is not justified.⁹ The likelihood of a benign viral illness as the cause for most febrile illness in infants and children should be stressed.^11,12 Significant parental coaching is often warranted to avoid over- or underevaluation and treatment. In addition, during the course of the illness, parents should be instructed to monitor the child's clinical state and the appearance of new or worsening symptoms as indicators of when to follow-up with up with their medical provider.

Pediatric sepsis is a common cause of death in infants and children worldwide and may be caused by bacterial, viral, parasitic, or fungal organisms.¹³ Although most infections are contained by a local inflammatory response, occasionally, the host response is of a more widespread nature characterized by the dissemination of inflammatory cytokines and resultant activation of the immune system systemically. The clinical symptoms of this systemic inflammatory response syndrome (SIRS) include temperature instability, age-inappropriate tachycardia or tachypnea, as well as an abnormal leukocyte count.¹⁴ Sepsis is defined as SIRS in response to an infectious process. Severe sepsis includes sepsis-induced organ dysfunction or tissue hypoperfusion resulting in hypotension, liver dysfunction, changes in mental status, and elevated lactate. Septic shock is severe sepsis with hypotension that does not respond to intravenous fluid administration alone. Thus, there is a continuum from infection to bacteremia, SIRS, severe sepsis, septic shock, and finally multiorgan dysfunction syndrome (MODS). When assessing febrile young infants and children, the goal is to determine which of the above conditions apply or whether it is a self-limited viral illness.¹⁵ Because decompensation can be rapid, there is increased emphasis on the importance of “stop sepsis,” which requires early identification of those children who manifest early signs of SIRS and severe sepsis in an effort to begin goal-directed treatment promptly with intravenous fluids, antibiotics, and pressors if needed.^16,17

Early studies on fever noted that approximately 3% of well-appearing children (3–36 months old), who were initially managed as outpatients, had bacteremia.^18,19 Up to 80% of bacteria isolated in these studies were represented primarily by Streptococcus pneumoniae, and, to a lesser degree, Haemophilus influenzae type b and Neisseria meningitides. Although the vast majority of these cases of “occult” or “unsuspected” bacteremia resolved spontaneously, without antibiotic therapy, the concern that a small percentage would develop SBI (e.g., meningitis, osteomyelitis, and pneumonia) led to various policies advocating routine laboratory testing and selective use of empiric antibiotics for febrile, well-appearing young children without an identifiable source of infection.¹¹ However, presumptive antibiotic therapy has not been shown to prevent this occult bacteremia.^18–20

Widespread vaccination for H. influenzae type b and S. pneumoniae has had a profound impact on the incidence of occult bacteremia. Pneumococcal bacteremia has been reduced to less than 0.2% in vaccinated healthy children but remains at about 1.5% in unimmunized children.^21–23 Therefore, based on the existing data, it appears that for the well-appearing, otherwise healthy, febrile young child with no source of infection, the risks and costs of routine testing and empiric antibiotic therapy likely exceed the potential benefit.¹⁹ In fact, over the past decade, there has been a national decline in obtaining a complete blood count (CBC) as well as a trend toward less blood cultures in the evaluation of children 2 to 24 months presenting to the ED with fever and no source of infection.²⁴ Given the extremely low risk of bacteremia and its sequelae, the best expectant therapy in the well-appearing child is close observation by a reliable caretaker.

Other Occult Infections

The most common bacterial cause of fever in young children without a source of infection on physical examination is a urinary tract infection (UTI), which has an overall prevalence of about 5% in febrile children 2 to 24 months old.²⁵ The prevalence of UTI varies by several factors including age, gender, and certain clinical symptoms. The single most important risk factor for boys is circumcision. Boys who are older than 1 year and are circumcised are at very low risk for UTI and generally do not need to be tested in the absence of specific symptoms or historical factors. Uncircumcised infant boys have a tenfold greater risk for UTI ²⁶, whereas those over 1 year of age have a three- to fourfold increased risk. Other risk factors for boys include prior history of UTI, fever >39.0°C, ill appearance, fever for more than 24 hours, and nonblack race.²⁵ For febrile young girls, risk factors include age less than 1 year, white race, fever >39.0°C, and fever for at least 2 days.²⁵ Other important historical risk factors include, previous UTI, renal or bladder abnormality or dysfunction, and foul smelling urine. Urine specimens should be obtained by catheterization or suprapubic aspiration. Bag urine collection has a very high contamination rate and is unacceptable. Urine should be tested with both dipstick and microscopic examination. An enhanced urinalysis, which includes microscopic examination for white blood cells and a gram stain for bacteria, is more sensitive for detecting a UTI in young children than the presence of leukocyte esterase and nitrates on dipstick examination.²⁵ Urine culture, the gold standard for diagnosing a UTI, should be sent for confirmation regardless of the initial urinalysis result. The sensitivity testing of positive specimens can guide antibiotic choice for nonresponding infants.

Pneumonia is often clinically apparent by recognition of clinical signs of lower respiratory tract disease (rales, grunting, decreased breath sounds). Yet, approximately 5% to 6% of children with fever and no lower tract findings, tachypnea, or respiratory distress have “occult” pneumonia identified on chest radiograph.^27,28 Historically, children with fever >39.0°C and a WBC count of over 20,000/mm³ were at risk of occult pneumonia, although this rate may have declined since universal vaccination for pneumococcus. The likelihood of occult pneumonia increases with increasing duration of fever and prolonged or worsening cough but predictive values of these findings remain low. However, studies show that there is limited utility in obtaining a chest radiograph for occult pneumonia in febrile children without a cough.^27,28

Biomarkers For Bacterial Infection

Historically and currently many biomarkers have been studied as potential touchstones of bacterial infection. These include the white blood cell count and differential, the erythrocyte sedimentation rate, C-reactive protein (CRP), and procalcitonin.^29–32 All lack the specificity necessary to safely rule out SBI.

Meticulous history is the foundation of accurate diagnosis in infants and young children. The caretaker should be asked about methods of fever measurement as well as fever onset, height, pattern, and duration. The method of fever measurement (by thermometer or by “touch”), and the site of measurement (rectal, oral, axillary) are important. Of note, forehead chemical thermometers are generally unreliable.⁶ Parental perception of fever based on “touch” has similar sensitivity and specificity as those reported for axillary and tympanic thermometers and therefore should be considered as fever.⁶ Because of the variety of antipyretics and differing concentrations and dosing intervals, it is important to obtain an accurate medication history, not only to determine whether the pattern of the fever may be altered by antipyretic use, but also to ascertain if there was under- or overdosing. The caretaker's perception of illness is vitally important as a background description of normal and abnormal behavior in the patient. Specific observations include activity level, feeding behavior, and interactive behavior. It is important to use common terms familiar to parents, such as fussiness and drowsiness, rather than medical terms such as irritability or lethargy, which may be misinterpreted or misunderstood. Caretaker report of “chills” and “sweats” are common and likely represent the physiologic effects of the various febrile states (development of fever–chill phase, resolution of fever–flush phase). Fever due to a self-limited viral infection typically peaks in 1 to 2 days and resolves within 3 to 5 days, whereas fever due to bacterial illness generally worsens if untreated. Fevers lasting for longer than 1 week suggest a more chronic infection, immunologic, rheumatologic, or neoplastic etiology.

A detailed review of systems to elicit clues for specific infections follows. Rashes are extremely common and typically herald benign illnesses. However, the most ominous skin symptom or sign is the presence of petechiae or purpura, which may indicate disseminated intravascular coagulation (DIC) and potential life-threatening sepsis. Signs and symptoms suggesting early meningococcal sepsis include leg pain, cold hands and feet, and abnormal skin color (e.g., pallor or mottling).³³

Many infectious diseases are highly contagious, thus knowledge of the local epidemiology of specific infectious illnesses usually aid in diagnosis. Exposures from family members, friends, or daycare exposures may provide important clues to diagnosis. Immunization history should be reviewed, in particular immunization against H. influenzae type b and S. pneumoniae. In addition, a history of recent immunization adds the possibility of febrile reaction to the differential and may prevent unnecessary evaluation and treatment.

Finally, medication history is vital in properly evaluating fever. Antipyretics may improve a child's symptoms and provide a false sense of security on presentation. Antibiotic treatment may cause false-negative culture evaluation of a patient and result in either undertreatment of a serious illness such as meningitis or result in overtreatment of an illness due to the inability to rule out such an illness.

Skill at determining risk for SBI on physical examination results from a combination of careful observation, training, experience, and clinical judgment. Despite variability and subjectivity, some important reproducible principles do exist.

Initial assessment consists of vital signs and general appearance. Often, the first clues to significant illness are abnormal vital signs. Vital signs should be corrected for the height of fever and then compared with published age-appropriate norms for heart and respiratory rate (see Tables 1-2, 1-3, and 1-4, Chapter 1). Heart rate is known to be affected by many factors, including fever, which may elevate the pulse rate up to 10 beats per minute per degree.³⁴ However, tachycardia alone, even in the presence of normal blood pressure, may indicate shock as a physiologic response to maintain cardiac output. Bradycardia is an ominous sign, potentially indicating impending cardiorespiratory arrest. Tachypnea may be the result of fever, hypoxia, or acidosis associated with serious illness. Bradypnea is also a cardinal sign of impending respiratory arrest and mandates immediate airway stabilization. Tachycardia and tachypnea may be the only indications of a serious illness in an otherwise well-appearing febrile child.

In the unstable child, signs of clinical toxicity include altered or decreased mental status, significantly abnormal vital signs, dyspnea, color changes such as cyanosis and pallor, and hypoxia as measured by pulse oximetry. These children require immediate stabilization including airway management, oxygen, intravenous access and administration of normal saline fluid bolus, temperature management, rapid examination, laboratory evaluation, and immediate empiric antibiotic therapy pending further diagnosis. Examination of the clinically stable child should include an assessment of mental status, including activity level and interaction with parents, as well as interaction with the health care provider. In the young infant, poor feeding behavior and inability to console may be important clues to serious illness. In the older child, anxiety may be normal but should be lessened when the child is consoled in the parent's arms. Paradoxical irritability, irritability that worsens when the child is carried or rocked, is highly suggestive of meningitis or elevated intracranial pressure. The presence of a social smile is an important indicator of wellness. Decreased activity or lethargy is always a sign of serious illness. Finally, character of the cry may be informative, with a lusty loud cry suggesting a healthy child and a weak, high-pitched cry or absent cry suggesting a child who is seriously ill.

Assessment of perfusion is invaluable in determining the risk of serious illness. Normal peripheral perfusion is indicated by capillary refill of less than 3 seconds in a warm extremity (in normal ambient air temperature). In the presence of dehydration or septic shock, perfusion will be delayed. Another hallmark of poor perfusion is decreased mental status. Reassessment of perfusion and mental status is mandatory as clinical condition may wax and wane in young infants and children due to cardiovascular reserve, but when the reserve is exhausted, sudden vascular collapse can occur. A hallmark of impending collapse is an ashen or mottled appearance of the skin. Any child with even a momentary decrease in mental status by history or examination should be considered seriously ill until proven otherwise.

The general assessment is followed by detailed physical examination of all anatomic regions using an age-appropriate approach (orifices last) searching for an identifiable source of infection.

Historically, most practitioners managed febrile 28- to 56-day-old infants identically as for neonates: a “sepsis” evaluation, hospitalization, and administration of empiric antibiotics. In the early 1990s, several screening tools were developed to aid in the identification of young infants who were at low risk for SBI and therefore could be spared both routine hospitalization and possible antibiotic therapy. These eponymous screening tools—in particular, the Boston,³⁵ Philadelphia,³⁶ and Rochester³⁷ criteria—all employed slightly different inclusion criteria for risk stratification. Although variations of these management strategies have continued to gain acceptance, care must be taken to obtain all necessary clinical and laboratory parameters, as practitioners often deviate from the published algorithm that they believe they are following for management of the febrile neonate.³⁸

An ill-appearing 28- to 56-day-old infant requires a “sepsis” evaluation, hospitalization, and administration of empiric antibiotics. However, if well appearing, the risk can be stratified into high-risk or low-risk categories based upon a combination of clinical and laboratory findings. A useful set of low-risk clinical criteria is shown in Table 3-1. The decision to perform a lumbar puncture is contentious in those who meet all other low-risk criteria. Although there are no definitive data, some practitioners will defer or omit CSF examination given the very low risk of bacterial meningitis in this selected low-risk group. To that end, practitioners must judge each infant individually, taking into account the history, clinical appearance indicators such as social smile and reaction to parent stimulation (which are variable in this age group), as well as the reliability of the caretakers to observe the neonate and follow-up appropriately.

The management of febrile children between 3 and 36 months of age is largely dependent on clinical appearance, risk factors for bacterial illness, and the presence or absence of an identifiable source of infection. For children who are clinically unstable, stabilization, including airway management, intravenous fluids, and oxygen administration, takes precedence. If a source is identified, such as UTI, pneumonia, skin, soft-tissue or bone infection, then directed antibiotic therapy is initiated. For those children who are septic or at high risk for SBI, empiric antibiotic therapy should begin as soon as possible, preferably within 1 hour of presentation. If meningitis is considered in an unstable child, the lumbar puncture should be deferred with the immediate initiation of antibiotic therapy for meningitis (see Chapter 58).

In the well-appearing child who appears to have a specific clinical syndrome, rapid antigen testing may obviate additional laboratory investigation. These tests may include screening for influenza, RSV, and streptococcal antigen. UTI should still be ruled out in established risk groups.

In the well-appearing child with no obvious clinical source of infection, the clinician should consider a variety of decision modifiers as well as risk factors for SBI. Risk factors include unimmunized state; chronic disease such as sickle cell disease, congenital heart disease, inborn errors of metabolism, or conditions requiring CSF shunts and immunocompromise. When risk factors are present, an evaluation should include a CBC with leukocyte and immature neutrophil count and markers of inflammation including CRP, blood culture as well as urinalysis and urine culture when indicated. Lumbar puncture should be performed in those children who appear to have or are at risk for meningitis. Imaging for pneumonia may also be indicated based on the presence of respiratory symptoms. In the well-appearing child with no risk factors and no illness warranting rapid screening, no investigations are needed. The diagnosis of presumptive viral syndrome may be made, followed by discharge with close follow-up.

Most uncomplicated febrile illnesses last for 4 to 5 days.³⁹ Children with fever lasting for more than 5 to 7 days, in whom a source of infection is not evident, should be assessed with a systematic focused approach. In these children, Kawasaki disease must be considered although infectious illness remains most likely. As the duration of the fever increases, the likelihood of rheumatologic and neoplastic etiologies increase.⁴⁰

Clinically unstable patients with the presence or risk of SBI should be admitted for stabilization and empiric or directed antibiotic therapy. In well-appearing infants and young children with an identified source, outpatient symptomatic and directed care may be initiated with close follow-up. In those with no identified source, the chart should clearly reflect cardiorespiratory stability, adequate feeding behavior, urine output, and alertness at the time of discharge. Instructions for seeking follow-up if there is no improvement must be provided as well as advice to return to the ED if there is deterioration.

For the child with a chronic illness (e.g., cancer, HIV) or a child with an inborn error of metabolism such that the increase in the metabolic rate associated with fever creates a catabolic stress that may be poorly tolerated, antipyretic therapy is an important part of management to limit metabolic demand. However, for the otherwise healthy child, antipyretic therapy should be used primarily for the management of the child's discomfort during the febrile episode rather than for the sole aim of reducing body temperature.^6,9 Focusing parental attention to “controlling the fever” is generally impractical and misguided. There is no evidence that, in healthy children, reducing fever during a febrile illness reduces morbidity, mortality, or even the occurrence of febrile seizures.^6,9 Therefore, emphasis should be placed on monitoring the child's comfort, hydration status, and clinical state.

If antipyretics are used, either acetaminophen (15 mg/kg every 4 hours) or ibuprofen (10 mg/kg every 6 hours) is effective to treat documented fever causing discomfort.⁹ Ibuprofen may be used in febrile children older than 6 months. In addition, there is some concern about the nephrotoxicity of ibuprofen in children with dehydration, preexisting renal disease, cardiovascular disease, or simultaneous use of other nephrotoxic agents.⁴¹ Alternating antipyretics is a common practice by many health care providers. Studies have suggested that this routine may produce more temperature reduction at 4 to 6 hours but that reduction may not be clinically significant.^6,9 Furthermore, there is no evidence in overall improvement in other clinical outcomes and parents generally did not perceive efficacy of this approach.^6,9 Since the dosing amounts and the administration intervals vary between acetaminophen and ibuprofen, there is the potential for incorrect administration or overdosing. Finally, emphasis on antipyretic use likely focuses the caretaker on the presence and height of the fever rather than on the clinical state and comfort of the child, thereby increasing the possibility of “fever phobia.” A summary of some concepts about fever in healthy children is listed in Table 3-2.^6,9,18