Two-thirds of all children visit a physician for fever before they reach the age of 2 years. Fever in the pediatric population is usually grouped into 4 categories:
- •
Fever in the neonate
- •
Fever with localizing signs
- •
Fever without source (FWS)
- •
Fever of unknown origin (FUO)
Pathophysiology of Fever
Temperature is controlled by the thermoregulatory center, located in the preoptic area of the anterior hypothalamus. The thermoregulatory center receives input from peripheral receptors and the temperature of the blood bathing the hypothalamus and acts on autonomic, endocrine, and behavioral mechanisms to maintain the body temperature at a particular set point. The hypothalamic set point normally maintains body temperature around 37°C, but there can be significant variation among individuals. Normal temperatures range from 36-37.8°C, depending on the time of day, with the peak in the afternoon (5-7 p.m. ) and the trough in the early morning (2-6 a.m. ). Although the circadian rhythm is not well established in infancy, it becomes more reliable by the 2nd year of life.
The febrile response not only produces an elevation in body temperature but also causes physiologic changes that enhance the individual’s ability to eliminate infection. Production of acute-phase reactants and alterations in metabolism and endocrine function are examples of these changes. Acute-phase reactants —proteins that are produced in response to infection or injury—include ceruloplasmin, C-reactive protein, haptoglobin, amyloid A, complement, and fibrinogen. Hormones and cytokines, some of which are endogenous pyrogens, regulate the production of acute-phase proteins. Exogenous pyrogens, such as bacteria or endotoxins, generate the production of endogenous pyrogens, which play a vital role in prostaglandin-related set point elevation and regulation of acute-phase responses.
Fever results when the thermoregulatory set point is elevated above the normal set point; the hypothalamus then generates physiologic changes involving endocrine, metabolic, autonomic, and behavioral processes. Diversion of blood from peripheral vessels to central vessels causes coolness of the extremities but helps increase core temperature. Shivering increases metabolic activity and heat production. The affected patient may feel cold and seek a warmer environment or add clothing to feel warmer and prevent heat loss. Once these processes have resulted in increasing the core temperature to match the elevated set point, the thermoregulatory center works to maintain the temperature as it does during normothermia. The thermoregulatory point returns to normal once the infection is resolved. The hypothalamus then produces physiologic changes to decrease the core temperature; these include sweating, dilation of cutaneous blood vessels, and the sensation of feeling hot, which may lead to behaviors such as removing clothing or seeking a cooler environment.
Fever has both positive and negative effects. High body temperatures may impair the reproduction and survival of some invading microorganisms by decreasing required nutrients, such as free iron, or by increasing immunologic responses such as phagocytosis. However, at extremely high temperatures, immunologic responses may be impaired. Fever increases the basal metabolic rate by 10-12% for each degree Celsius elevation of temperature. This increases oxygen consumption, carbon dioxide production, and fluid and caloric needs. Fluid requirements increase 100 mL/m 2 /day for each 1°C rise in temperature above 37.8°C.
Heat illness must be distinguished from fever as a cause for elevated body temperature. In heat illness, there is an unregulated rise in body temperature, despite the fact that the hypothalamic set point is normal. It can result from excessive heat production or inadequate heat dissipation. Temperatures may reach extreme heights and can result in multiorgan dysfunction and death. Restoration of normal body temperature in heat illness is mandatory ( Table 39.1 ).
| Excessive Heat Production |
|
| Diminished Heat Dissipation |
|
| Hypothalamic Dysfunction * |
|
Fever Without Source
A child with fever of recent onset with no obvious historical or physical explanation for the fever is said to have fever without source (FWS). Bacterial pathogens account for a small but clinically significant number of cases. The risk of bacterial infection decreases with increasing age and is highest for infants less than 3 months of age, compared to infants and toddlers 3-36 months of age, and even lower for children over the age of 36 months. Most of the patients in all age groups have a self-limited viral illness. The challenge is to identify which children have fever caused by bacterial pathogens, or other pathogens requiring treatment, in order to avoid the morbidity and mortality associated with delayed treatment, balanced against the risks of testing or treatment when neither is needed. Bacterial infection must be considered in immunocompromised patients or those with central lines or shunts. Studies in adults suggest that patients with high fever (>105 o F) and rigors have a higher risk of bacterial infection; exceptions to this include influenza and adenoviral infections.
History
A detailed history may reveal a potential source for infection. A complete history addresses several important issues: (1) onset and duration of fever; (2) degree of temperature; (3) by what method and in which anatomic site the temperature was taken; (4) medications given, including antipyretics, antibiotics, or home remedies; (5) environmental exposures; (6) associated symptoms; (7) ill contacts; (8) recent immunizations, and (9) recent travel. Inquiry into the child’s medical history may reveal important information such as recurrent febrile illnesses, primary or acquired immunodeficiency, or medications such as chemotherapy that alter host defenses.
Fever: Temperature Measurement
Rectal temperature measurement is considered to be the gold standard for children 3 years of age or younger. The most widely accepted definition of fever is rectal temperature of 38°C (100.4°F) or higher. It is important to consider that infants, especially those younger than 2 months of age, may have a blunted febrile (or hypothermic) response to infection. Hence, lack of fever should not be used as a criterion for ruling out infection in infants. Although rectal temperature measurement is the gold standard, it should be avoided in neutropenic immunocompromised patients, in whom rectal manipulation may seed the blood with bacteria.
Oral thermometry can be considered for cooperative patients who are older than 4-5 years of age. Axillary temperatures are commonly done and tympanic membrane and temporal artery temperatures are newer modalities with some studies examining their reliability. Axillary temperatures are less precise than rectal temperatures. There is a correlation between axillary and rectal temperature measurements; the axillary temperature is usually 0.5-0.85°C lower . Tympanic membrane thermometers are often inaccurate in children. Temporal artery temperature measurement correlates well with rectal temperature in some studies, but has been shown to be inferior when patients are febrile. It can be considered in settings when children are not likely to be febrile and are over 3 months of age. When detection of fever is critical for diagnosis and management, rectal temperatures should be used in the child 3 years of age and younger.
Physical Examination
Many children will have a source for fever identified on their history and/or physical examination. If no focus of infection is found on the physical examination, the clinician must rely on history and observation to determine the appropriate next steps. The child may appear ill or well. Ill-appearing children are typically lethargic or irritable. They may show signs of shock, including weak peripheral pulses, tachycardia, poor perfusion, respiratory distress, mottling, cyanosis, or decreased mental status ( Table 39.2 ). After thorough clinical and laboratory evaluation, ill-appearing children should be admitted to the hospital, and will likely need empiric antibiotic treatment.
| Infection | Suspected or Proven Infection or a Clinical Syndrome Associated with High Probability of Infection |
|---|---|
| SIRS | Two of 4 criteria, 1 of which must be abnormal temperature or abnormal leukocyte count:
|
| Sepsis | SIRS Plus a Suspected or Proven Infection |
| Severe sepsis | Sepsis plus 1 of the following:
|
| Septic shock | Sepsis plus cardiovascular organ dysfunction as defined above |
| MODS | Presence of altered organ function such that homeostasis cannot be maintained without medical intervention |
Observational Scales
Infants and children with fever without source who do not appear ill create important decision processes in terms of evaluation and management. The physician’s ability to make a hypothesis about the child’s degree of illness, on the basis of observation, is critical in the evaluation. An objective scoring measure may be used in an effort to assess serious illness in young febrile children. The Acute Illness Observation Scale (AIOS) ( Table 39.3 ), also known as the Yale Observation Score, is a 6-item predictive model graded on a scale of 1-5. Use of the AIOS in conjunction with the history and physical examination has a higher sensitivity for identifying serious illness than history and physical examination alone. The AIOS is most useful in patients younger than 24-36 months; it has not been shown to provide sufficient data to identify serious illness in 4- to 8-week-old infants, and has not been evaluated in infants less than 4 weeks old.
| 1 | 3 | 5 | |
|---|---|---|---|
| Observation Item | Normal | Moderate Impairment | Severe Impairment |
| Quality of cry |
|
|
|
| Reaction to parent stimulation |
| Cries off and on |
|
| State variation |
|
|
|
| Color | Pink |
|
|
| Hydration |
|
|
|
| Response (talk, smile) to social overtures |
|
|
|
Differential Diagnosis
Most children who present with fever without source (FWS) are subsequently determined to have a self-limited benign viral infection. In 1 study in 2-36 month old children who presented with FWS, 76% had 1 or more known pathogenic viruses found; 57% had adenovirus, human herpesvirus 6 (HHV-6: roseola), enterovirus, or parechovirus detected. Other identifiable viruses include respiratory syncytial virus, parainfluenza viruses, influenza viruses, varicella (chickenpox), human metapneumovirus and parvovirus (fifth disease/erythema infectiosum). Measles, mumps, and rubella are uncommon in developed countries but have been reported in epidemics following imported cases or in underimmunized communities. Although rapid testing for viral pathogens is often readily available, a detailed investigation to identify a viral pathogen is not necessary unless the confirmation of a viral infection will change the acute diagnostic plan; treatment with antivirals is an option (HSV, influenza) if the fever is prolonged and evolves into FUO or if there is end-organ involvement, as in hepatitis, myocarditis, encephalitis, or meningitis.
Most viral infections do not have simultaneous co-infection with a bacterial pathogen. Exceptions include croup due to parainfluenza virus, which may predispose to bacterial tracheitis and influenza, which may predispose to bacterial pneumonia. The sequence may be biphasic with viral symptoms followed by improvement, followed by worsening symptoms of the bacterial superinfection, or both phases may not be apparent as the child demonstrates no improvement or deterioration. Respiratory syncytial virus (RSV) may predispose patients to otitis media.
Noninfectious conditions manifesting with FWS are extremely rare. Historical clues (recurrences, chronicity) or systemic signs usually indicate malignancy or rheumatic disorders. If the history and physical examination are not suggestive, these diagnoses need not be pursued. Heat-related illness or drug ingestion may be considered if supported by the history. Fever caused by immunizations may not be accompanied by other signs or symptoms, but the history should suggest immunization as the cause.
Urinary Tract Infections (UTIs)
UTIs are the most common serious bacterial infection in children less than 36 months of age who present with FWS. UTIs are almost always occult in children younger than 24 months because the symptoms, except for fever, are nonspecific or nonexistent. UTI occurs in 7% of febrile children younger than 2 years. The prevalence of UTI varies by height of the fever, duration of the fever, age, gender, race, and circumcision status. Children with fever greater than 39°C are at a higher risk of UTIs. Boys with fever for more than 2 days and girls with fever for more than 1 day are more likely to have a UTI. Higher rates of UTIs are found in girls, especially those younger than 12 months of age. For febrile boys younger than 3 months of age, 20.1% of those who are uncircumcised have a UTI; for circumcised boys the rate is 2.4%. UTI rates are higher among white infants than among black infants and among children with abnormal genitourinary tract anatomy or neurogenic bladder.
Urine specimens should be obtained from the following children with FWS: those with a history of UTI, those with a history of urinary tract anomalies or vesicoureteral reflux, all infants younger than 2 months, girls younger than 12-24 months, uncircumcised boys younger than 12 months, and circumcised boys younger than 6 months. There is an age-associated risk of bacteremia with UTIs, particularly in infants. The incidence of bacteremia in patients younger than 2 months with UTI is 10%. The incidence of bacteremia in patients younger than 2 months with UTI ranges from 4-15% depending on the setting. Opinions regarding when to obtain blood cultures in infants with UTI differ, but a reasonable approach would be to obtain blood cultures in children younger than 2-6 months with suspected UTI, and in older infants with UTI if they are ill-appearing (urosepsis).
Bacteremia
Occult bacteremia is defined by the presence of a positive blood culture for pathogenic bacteria in a febrile patient who does not appear extremely ill and who has no focus of infection, excluding otitis media. Following the introduction of the 7-valent pneumococcal vaccine in 2007, invasive pneumococcal disease decreased dramatically. Pneumococcal bacteremia decreased from 80% of the cases of bacteremia to 30%. Most cases of bacteremia in children were not occult. Bacteremic children were either ill or had a focus of infection, such as a UTI. In 1 study, the rate of occult bacteremia after 2007 was 0.25%. After the 13-valent pneumococcal vaccine was introduced in 2010, the incidence of invasive pneumococcal disease in children less than 5 years old decreased again with 1 state-based population study showing incidence rates dropping from 46/100,000 to 23/100,000 with the age group most involved being children 2-23 months of age. Escherichia coli is the most common cause of bacteremia in children aged less than 12 months, all due to UTIs. Other less common causes of bacteremia in young children are N. meningitidis , nontyphoidal Salmonella , Staphylococcus aureus , and group A streptococcus. Neisseria meningitidis bacteremia is frequently associated with serious sequelae. Children with N. meningitidis bacteremia are much more likely to progress to meningitis than are those with S. pneumoniae bacteremia. Nontyphoidal Salmonella bacteremia is often accompanied or preceded by enteritis. In some instances, particularly in young infants, the diarrhea is mild or even absent. The prevalence of Salmonella bacteremia among patients with Salmonella enteritis has been reported to be between 2% and 45%; fever is not always present. Salmonella infection seldom causes serious complications in patients with normal host defenses and resolves spontaneously. Infants younger than 3 months, malnourished, and immunocompromised individuals are exceptions.
Role of Diagnostic Testing in Patients with Fever Without Source
Evaluation is usually divided into 4 different age ranges: younger than 1 month, 1-3 months, 3-36 months, and older than 36 months. Testing for each individual age group is based on risks for diseases and prevalence of pathogens.
Complete Blood Count and Other Markers of Inflammation
The white blood cell (WBC) count is the most commonly used test in young children with FWS. Complete blood count (CBC) is less useful as a marker for invasive disease caused by E. coli than by S. pneumoniae , thus its utility has declined with the reduction of the incidence of invasive pneumococcal disease. Similarly, band counts are less commonly used, except in the 29-60 day old infant as part of identifying a low-risk cohort. A WBC count of 5,000-15,000 is generally considered normal for children over 1 month of age. A WBC count less than 15,000/mm 3 or even leukopenia may be found in children with N. meningitidis bacteremia. A minority of children with occult nontyphoidal Salmonella bacteremia have been found to have a WBC count exceeding 15,000/mm 3 . C-reactive protein (CRP) and procalcitonin combined with a urine dipstick (the lab score) can be used to screen for bacterial infection. This combination of tests has been validated for children 7 days to 36 months of age.
Polymerase Chain Reaction (PCR)
PCR is useful in identifying the cause of fever for common viruses such as respiratory syncytial virus, influenza viruses, parainfluenza viruses, enteroviruses, parechovirus, adenoviruses, or herpes simplex virus.
Additional methods available or in development that may be helpful to identify serious bacterial infections and distinguish bacterial from viral infections utilize molecular microbiology methods. Gene expression profiles of the patient’s peripheral blood leukocytes demonstrate different biosignatures of RNA production that may differentiate bacterial from viral infections. This method does not identify the specific pathogen. Rapid multiplex PCR combined with standard blood culture methods may identify a specific pathogen much sooner (~20 hours) than standard blood culture techniques. Specific bacteria may be identified using 16S ribosomal RNA bacterial gene detection. This method does not require bacterial growth. 16S rRNA detection may be helpful when antibiotics were administered before the sample was obtained, and in patients with ventilator-associated pneumonia or bacteria that grow poorly or are present in effusions or tissues (heart valves).
Blood Cultures
Blood cultures are the gold standard for determination of bacteremia. Although blood cultures do not provide immediate results, methods allow for continuous and more rapid detection of bacterial growth. Blood cultures are easy to perform and provide essential information in the diagnosis and management of patients with possible bacteremia. Preliminary blood culture results are typically available within 24 hours, with positive identification of most organisms within 48 hours.
False-negative blood culture results may be due to prior treatment with antibiotics, missing an episode of bacteremia if it is intermittent, and inoculation of too little blood into the culture media. Alternatively, too much blood inoculated into the blood culture bottle may yield a false-negative result because of ongoing killing of bacteria by neutrophils. Three to 5 mL of blood should be added to each blood culture bottle. False-positive results may be due to inadequate skin preparation, leading to contamination with skin flora.
Urinalysis and Urine Culture
A positive urine culture was once considered the gold standard; current recommendations include a urinalysis that has pyuria (defined as >5 WBCs/high-power field [hpf] on the microscopic examination or a positive leukocyte esterase on dipstick) and a positive urine culture for a uropathogen in an appropriately collected specimen. Fifty to 100,000 colonies of a single organism is considered positive (see Chapter 18 ). Children should have a catheterized urine specimen obtained, unless they are toilet-trained and can supply a clean voided specimen. Suprapubic aspiration is acceptable but requires technical expertise, and parents often perceive it as unsuitably invasive; it may be the only alternative for boys with severe phimosis. The use of plastic receptacles attached to the perineum should be discouraged because contamination from skin and fecal flora commonly occurs.
Lumbar Puncture
Lumbar puncture is indicated if the patient is younger than 28 days or if a diagnosis of sepsis, meningitis, or encephalitis is considered, regardless of the child’s age. Normal cerebrospinal fluid (CSF) findings, including chemistry, cell count with differential, Gram stain, PCR, and culture, help exclude the diagnosis of meningitis. Less than 1% of children with normal preliminary CSF results have a positive culture; in most of these, the pathogen is N. meningitidis . Thus, even in the presence of normal preliminary CSF results, close follow-up is essential.
Chest Radiographs
Chest radiographs are usually normal in children who have FWS. Respiratory signs or symptoms, such as tachypnea, retractions, crackles, wheezing, rhonchi, nasal flaring, grunting, cough, or hypoxia, may predict chest radiograph findings consistent with pneumonia. In practice, pneumonia can often be diagnosed solely on the basis of the clinical findings of fever, tachypnea, and crackles; chest radiographs are not always necessary. However, chest radiographs may be useful in evaluating for the presence of pleural effusion or other complications of pneumonia.
Stool Cultures
Most acute diarrhea and fever is caused by viral pathogens in developed countries. Obtaining a stool culture is indicated if bacterial enteritis is indicated by the presence of risk factors in the history, such as blood in the stool or certain exposures (petting zoos) (see Chapter 11 ).
Evaluation and Management
Children Younger Than 3 Months
Febrile infants younger than 3 months have a higher incidence of serious bacterial infections than older infants. The relatively high incidence of bacterial disease probably results from a combination of factors unique to this age group: decreased opsonin activity; decreased macrophage function; decreased neutrophil function; poor immunoglobulin G antibody response to encapsulated bacteria; and susceptibility to bacterial pathogens such as group B streptococci (GBS), gram-negative enteric organisms, and Listeria monocytogenes . The incidence of early-onset group B streptococcal infections has decreased with routine screening and the intrapartum treatment of GBS-positive pregnant women; the incidence of late-onset GBS (>1 week) has not decreased. E. coli is the most common organism causing bacterial infections in neonates and young infants.
In very young infants, clinical evaluation alone is inadequate for excluding serious bacterial infections. Management of febrile infants younger than 28 days includes a sepsis evaluation and hospitalization for parenteral antimicrobial therapy pending culture results. The reasoning for this conservative approach lies in the difficulty in evaluating the behavioral state of neonates, the rapid clinical deterioration of infants with bacterial infections, the immature neonatal immune system, and the possibility of life-threatening viral infections caused by herpes simplex viruses (HSV) or enteroviruses. Sepsis evaluation should include culture of the CSF, blood, and urine; a complete blood cell count with differential; examination of the CSF for cells, protein, and glucose; and urinalysis. A chest radiograph should be considered if the patient has signs or symptoms of a respiratory infection. Testing (blood and CSF PCR for HSV) and treatment for possible HSV infection should be considered in ill-appearing infants, those with a seizure prior to presentation, and those with a vesicular rash consistent with HSV.
A combination of clinical evaluation and laboratory studies can be used to define a specific population of infants aged 29-60 days who do not appear ill and are at low risk for bacterial infections. Infants at low risk for bacterial infections are those who are previously healthy with no focus of bacterial infection on physical examination and who have negative laboratory screening results. A number of prospective studies have contributed to the development of specific low-risk screening criteria ( Table 39.4 ). The age groups included vary by study, ranging from 0-90 days to 29-56 days. Because there are differences in study criteria used to define infants at low risk for bacterial infections the most conservative values have been used in the guidelines presented in this chapter. Negative laboratory screening results consist of a WBC count of 5000-15,000/mm 3 ; fewer than 1500 bands/mm 3 or a band-to-neutrophil ratio of less than 0.2; fewer than 10 WBCs/hpf and no organisms on urinalysis; and fewer than 8 WBCs/hpf and no organisms on CSF Gram stain. Some experts also include a negative chest radiograph and, when diarrhea is present, a stool examination with fewer than 5 WBCs/hpf.
Most experts suggest that febrile infants 29-60 days old who meet the low-risk criteria and have access to close follow-up can be managed as outpatients. Blood, urine, and CSF cultures should be obtained before empirical antibiotic treatment so that viral and bacterial causes may be distinguished. An alternative strategy is to manage such infants as outpatients, without empirical antibiotic therapy, after blood, CSF, and urine cultures are obtained. Although most of the original studies on outpatient management of febrile infants included infants aged 2-3 months, many experts agree that infants aged 2-3 months can be managed safely according to the guidelines for infants and children aged 3-36 months ( Table 39.4 ).
| Boston Criteria |
| Infants are at low risk if they appear well, have a normal physical examination, and have a caretaker reachable by telephone, and if laboratory tests are as follows: |
|
| Philadelphia Protocol |
| Infants are at low risk if they appear well and have a normal physical examination, and if laboratory tests are as follows: |
|
| Pittsburgh Guidelines |
| Infants are at low risk if they appear well and have a normal physical examination, and if laboratory tests are as follows: |
|
| Rochester Criteria |
| Infants are at low risk if they appear well and have a normal physical examination, and if laboratory findings are as follows: |
|
Regardless of whether the clinician chooses to treat the patient with empiric antibiotics, all low-risk infants should be re-evaluated within 24 hours. Those who appear ill or who have positive culture results should be admitted for parenteral antibiotics. If a child appears well and all culture results are negative, close follow-up should be continued and a 2nd return visit made in 24 hours.
Children Aged 3 to 36 Months
The risk of bacteremia for children with FWS in this age group has decreased with the routine use of pneumococcal vaccines. The most common occult bacterial infection in this age group is UTI. For children in this age group who appear ill, a full sepsis evaluation should be undertaken ( Table 39.5 ).
| Group | Management |
|---|---|
| Any toxic-appearing child 0–36 mo and temperature ≥38°C (100.4°F) | Hospitalize, broad cultures plus other tests, * parenteral antibiotics |
| Child <1 mo and temperature ≥38°C (100.4°F) | Hospitalize, broad cultures plus other tests, * parenteral antibiotics |
| Child 1–3 mo and temperature ≥38°C (100.4°F) | Two-Step Process
|
| Child 3–36 mo and temperature 38–39°C (100.4–102.2°F) | Reassurance that diagnosis is likely self-limited viral infection, but advise return with persistence of fever, temperatures >39°C (102.2°F), and/or new signs and symptoms |
| Child 3–36 mo and temperature >39°C (102.2°F) | Two-Step Process
|
* Other tests may include chest radiograph, stool studies, herpes simplex virus polymerase chain reaction.
Screening urinalysis (UA) for UTI should be considered in children with a history of UTI, children with a history of urinary tract anomalies or vesicoureteral reflux, girls younger than 12-24 months, especially when the temperature is greater than 39.0°C, uncircumcised boys younger than 12 months, and circumcised boys younger than 6 months. Blood cultures are recommended for children with probable UTIs who are less than 6 months of age. A febrile child with moderate leukocyte esterase on urine dipstick testing or pyuria on an appropriately collected specimen should be treated presumptively for a UTI. Urine cultures should be obtained for any patient with a suspected UTI. The choice of antibiotics should be guided by knowledge of the common pathogens that cause UTIs and by patterns of antibiotic sensitivity in the community. Hospitalization should be considered for the child who is vomiting, is dehydrated, or appears ill; for those in whom compliance is likely to be poor; and for any patient with underlying renal or urologic anomalies.
Examination and culture of the CSF are the only tests to exclude the diagnosis of meningitis and encephalitis. They should be considered in any child in whom the diagnosis of sepsis, meningitis or encephalitis is suspected on the basis of the history, observation, and physical examination findings. Outpatient management of children with FWS is acceptable for those with a low probability of meningitis, good follow-up, and reliable caregivers. Blood cultures should be obtained for all children in whom sepsis or meningitis is suspected. Empiric treatment with antibiotics should be considered in those suspected of sepsis or meningitis after appropriate cultures are obtained.
In summary, management of children aged 3-36 months with fever is based on clinical experience and numerous study results:
- •
Child who appears ill on initial evaluation or on follow-up: Admit to the hospital for parenteral antibiotics after appropriate laboratory evaluation.
- •
Well-appearing children with FWS should be screened for UTIs, based on their number of risk factors. Risk factors for girls are: age <12 months, white race, temperature greater than 39°C, and fever for 2 or more days. Girls 2-24 months of age with 1 or more of these risk factors have a greater than 1% probability of having a UTI, and should be screened for a UTI.
For boys, the risk factors are uncircumcised status, nonblack race, temperature greater than 39°C, and fever for over 24 hours. All uncircumcised boys less than 12 months old, even if they don’t have other risk factors, should be screened for a UTI. For boys who are circumcised, 2 or more of the other risk factors increases the risk to over 1% and they should be screened.
- •
Child with positive blood culture: Reevaluation should occur in any child whose blood culture is presumptively positive. If the blood is found to contain N. meningitidis or Haemophilus influenzae (which has been rare since the advent of H. influenzae b immunization), a CSF sample and a repeat blood culture should be obtained, and the child should be admitted to the hospital for parenteral antibiotics, pending the results of the cultures. The child with occult pneumococcal bacteremia who appears well and is afebrile when returning for a follow-up may be managed as an outpatient with parenteral ceftriaxone followed by oral antibiotics according to the sensitivity of the organism. Because of the concern of pneumococcal resistance to penicillin, a 2nd dose of intramuscular ceftriaxone may be given until sensitivity results are available. If the culture is positive for nontyphoidal Salmonella organisms and the child is younger than 3 months, full sepsis evaluation and intravenous antibiotics are recommended. Oral antibiotics and close follow-up are recommended for older children with Salmonella bacteremia.
- •
Child with positive urine culture: If the child is afebrile and appears well, treatment with oral antibiotics is recommended, according to the sensitivity of the organism.
Children Older Than 36 Months
Evaluation and management of ill-appearing children older than 36 months with fever without source are similar to those of younger children. For children in this age group who do not appear ill, no screening diagnostic tests are indicated. Close attention should be paid to environmental exposures and ill contacts because of the high likelihood of increased contacts in this school-aged cohort.
Central Nervous System Infections
Bacterial Meningitis
Bacterial meningitis is usually a disease of infants and young children. The attack rate is highest between the ages of 3 and 8 months; 66% of cases occur in children younger than 5 years of age. Bacterial meningitis is seen during all seasons; however, there may be a seasonal correlation between the presence of preceding respiratory pathogens in the upper respiratory tract and the subsequent development of bacterial meningitis. Bacterial meningitis usually occurs sporadically. Clusters of cases have been noted in day care centers, colleges, and other closed communities. Bacterial meningitis occurs more frequently in children with traumatic fractures of the cribriform plate or paranasal sinuses or with a cochlear implant (pneumococci); in children who have undergone neurosurgical procedures such as ventricular shunts ( S. aureus , S. epidermidis , Corynebacterium species); in children with congenital or acquired immunodeficiencies (pneumococci, L. monocytogenes , meningococci); in children with anatomic or functional asplenia (pneumococci, meningococci); and in children with sickle hemoglobinopathies (pneumococci). There may be a genetic predisposition in some groups to the development of meningitis, inasmuch as there is an increased incidence of H. influenzae type b meningitis in Navahos and Eskimos.
Bacterial meningitis manifests in 2 patterns. In the 1st, the symptoms develop slowly over several days, the initial symptoms being those of a nonspecific illness. The signs and symptoms of meningitis develop subsequently. In the 2nd pattern, the disease develops suddenly and quickly, the 1st indications of illness being the signs and symptoms of meningitis and/or sepsis.
The manifestations of meningitis depend on the child’s age. In infants, the findings are usually nonspecific and may be subtle; they include vomiting, diarrhea, irritability, lethargy, poor appetite, respiratory distress, seizures, hypothermia, and jaundice. Only 50% of affected infants have fever; some present only with fever. It is uncommon for affected young infants to have a stiff neck; only 30% have a bulging fontanel.
Older children present with more specific meningeal signs. They complain of a headache that is described as being severe, generalized, deep-seated, and constant. They complain about neck stiffness, caused by inflammation of the cervical dura and reflex spasm of the extensor muscles of the neck. There is pain and limitation of motion on flexion of the neck, but lateral movement of the neck may be normal and pain-free. They also complain of nausea, vomiting, anorexia, and photophobia.
On examination, they demonstrate irritability, mental confusion or altered consciousness, nuchal rigidity, and, occasionally, hyperesthesia and ataxia. The clinician demonstrates nuchal rigidity by feeling resistance and observing a painful response while flexing the patient’s neck. The stiffness may not be recognized until the end of flexion. The neck usually can be rotated without symptoms. In the child who is crying and tensing the muscles, nuchal rigidity may be demonstrated if the examiner places 1 hand under the occiput of the supine patient and lifts the child. If the neck does not flex, it is stiff. Alternatively, a sitting child may be observed following an object as it falls to the floor. The child who flexes the neck to look at the object does not have nuchal rigidity. In the presence of meningitis, flexion of the neck causes spontaneous flexion of the legs at the hips and knees, the Brudzinski sign ( Fig. 39.1 ). The Kernig sign is elicited when the patient lies supine and, with the knee flexed, the leg is flexed at the hip. The knee is then extended. A positive sign is present if this movement is limited by contraction of the hamstrings and causes pain. Absence of nuchal rigidity is found in 1.5% of older children with meningitis; it may be absent in children who have overwhelming infections, are deeply comatose, or who have focal or global neurologic impairment.
As many as 15% of children with bacterial meningitis initially present in a comatose or semicomatose state (see Chapter 31 ). Because of the short duration and inconsistent development of increased intracranial pressure, papilledema is usually not seen at presentation. When it is present, venous sinus thrombosis, subdural effusion, or an intracranial abscess must be considered. Seizures occur before hospital admission in up to 20% of affected patients.
Children with meningitis may also present with cutaneous findings. Although commonly associated with meningococcal disease, purpura, petechiae, or a diffuse nonspecific maculopapular rash may be present in meningitis caused by any of the common bacterial pathogens (see Chapter 40 ).
Septic arthritis may be seen simultaneously with bacterial meningitis. This has been assumed to be caused by simultaneous localizing infection after a primary bacteremia. Reactive arthritis caused by immune complex deposition is also seen with bacterial meningitis. This arthritis affects 1 large joint and appears 5-7 days after treatment for meningitis has started. In general, arthritis occurring acutely with meningitis should be assumed to be infectious (see Chapter 33 ).
Various eye disorders have also been described with acute bacterial meningitis, including transient cataracts, paralysis of the extraocular muscles, pupillary dysfunction, dendritic ulcers, endophthalmitis, cortical blindness, and conjunctivitis.
Recurrent episodes of bacterial meningitis rarely occur. Potential etiologies include congenital CSF fistulas (inner ear, dermal sinus, neuroenteric cysts, lumbosacral sinus tracts), traumatic or surgical CSF fistula (skull fracture, postoperative nasal surgery, cochlear implant), immunodeficiency states and parameningeal infections (mastoiditis, sinusitis, craniofacial osteomyelitis).
Diagnostic Studies
Lumbar Puncture and Cerebrospinal Fluid Analysis
The definitive diagnosis of meningitis is based on examination of the cerebrospinal fluid (CSF). The CSF is usually obtained via a lumbar puncture (spinal tap). The lumbar puncture is performed by introducing a small-bore, short-beveled, spinal needle with a stylet into the subarachnoid space at the L3-L4 or L4-L5 level ( Figs. 39.2 to 39.4 ). A needle with a stylet is used to minimize the risk of introducing a nest of epidermal cells into the subarachnoid space that may later grow into a cord-compressing epidermoid tumor. Approximately 3 mL of fluid is removed for analysis.
There are a few contraindications for the performance of a lumbar puncture. The 1st is cardiorespiratory compromise. Performance of the lumbar puncture requires that the child be held in flexion to open the intervertebral spaces. In seriously ill children or children with significant underlying cardiac or pulmonary disease, this positioning may be enough to cause respiratory compromise. The lumbar puncture may need to be postponed, be performed cautiously with continuous oxygen saturation monitoring, or performed with the patient in the sitting position.
Second, children with increased intracranial pressure from a focal central nervous system (CNS) lesion, such as brain abscess or tumor, or from illnesses associated with cerebral edema have a high risk of cerebral herniation after a lumbar puncture. If signs or symptoms of increased intracranial pressure are present, the lumbar puncture should be postponed until the increased pressure is lowered with appropriate treatment. If a lumbar puncture is delayed, appropriate antibiotic therapy should be initiated without further delay. Third, a lumbar puncture should not be done if the spinal needle must pass through an area of infection on its way to the subarachnoid space. To do so might introduce pathogens into the CNS that could cause meningitis.
Epidural hematomas causing lower limb paralysis may be a complication of lumbar punctures in children with bleeding disorders. Therefore, in children with hemophilia, disseminated intravascular coagulopathy, or thrombocytopenia, lumbar puncture should be postponed until the bleeding disorder is corrected, and extra care should be taken to avoid a traumatic lumbar puncture. Such children should be monitored after the procedure for the development of neurologic deficits. Empirical therapy may be started while the coagulopathy is corrected.
Other, rarer complications of lumbar puncture include cortical blindness from compression of the posterior cerebral artery against the tentorium cerebelli, causing ischemic infarction of the occipital lobes. Cervical spinal cord infarction, with respiratory arrest and flaccid tetraplegia, may occur if intracranial hypertension causes herniation of the cerebellar tonsils through the foramen magnum with resulting compression of the anterior spinal artery or its penetrating branches. Post–lumbar puncture headache may occur in up to 10% of older children and adults; it is presumably caused by persistent CSF leakage at the lumbar puncture site.
The CSF is examined for red blood cells (RBCs), white blood cells (WBCs) and differential, glucose, protein, and the presence (by culture, by Gram stain or other stain, or by antigen or DNA-PCR testing for specific agents) of pathogenic organisms. Opening pressure measurements are obtained with the head of the bed flat and with the child relaxed and in the lateral decubitus position with the back no longer tightly flexed. The upper limit of normal value in children 1-18 years of age is less than 25 cm of water. Opening pressure is less than 5 cm H 2 O in premature infants and less than 10 cm H 2 O in normal newborns. Opening pressure measurements are elevated if the lumbar puncture is performed with the patient in the sitting position and if the patient is combative or performing the Valsalva maneuver. Obstructive hydrocephalus, hyperventilation, or removal of fluid can all lead to lowering of the measurement. Children with bacterial meningitis usually have a mean opening pressure of 18 ± 7 cm H 2 O.
Normal CSF is clear and colorless ( Table 39.6 ). Blood in the CSF indicates a traumatic lumbar puncture or a CNS hemorrhage. Obtaining a RBC count on tubes 1 and 3 may differentiate the 2 conditions because the count is unchanged in CNS hemorrhage but may decline in traumatic taps. Centrifugation of the CSF sample may also help differentiate between a traumatic tap and a CNS hemorrhage. When blood has been present in the CSF for several hours, the CSF is xanthochromic after centrifugation. However, if the blood was recently mixed with CSF, as in the case of a traumatic tap, the supernatant is clear. Xanthochromic CSF can also be caused by icterus or an elevated CSF protein concentration.
