The causes, evaluation, and management of possible occult bacteremia vary by children's age and immunization status. See also Fever in Infants and Children.
In the era before conjugate vaccines, about 3 to 5% of children aged 3 to 36 months with a febrile illness (temperature ≥ 39° C) and no localizing abnormalities (ie, fever without a source) had occult bacteremia. In contrast, children > 36 months with bacteremia almost always looked ill and had an identifiable (ie, non-occult) focus of infection. The majority (80%) of occult bacteremia prior to routine conjugate immunization was caused by Streptococcus pneumoniae. A smaller percentage (10%) was caused by Haemophilus influenzae type b, and an even smaller percentage (5%) by Neisseria meningitidis.
Occult bacteremia is a concern because about 5 to 10% of the children develop serious bacterial infections (SBIs)—typically defined as sepsis, meningitis, and urinary tract infection (UTI), but also including septic arthritis and osteomyelitis. Such infections could be minimized by early identification and treatment of the bacteremia. The likelihood of progression to serious focal illness depended on the cause: 7 to 25% for bacteremia caused by H. influenzae type b but 4 to 6% for bacteremia caused by S. pneumoniae.
Currently in the US and Europe, routine vaccination of infants with polysaccharide conjugate vaccines against S. pneumoniae and H. influenzae type b has eliminated (> 99%) H. influenzae type b infections and substantially reduced (≥ 70% overall and ≥ 90% vaccine-type) invasive S. pneumoniae infections. Thus, in this age group, occult bacteremia has become rare except in underimmunized or nonimmunized children, and in children with immunodeficiency.
In contrast, febrile infants < 3 months of age continue to have a greater risk of serious bacterial infection (SBI) than older infants, about 8 to 10%. In the past, SBIs in young infants < 3 months of age were more commonly caused by group B beta-hemolytic Streptococcus, S. pneumoniae, and H. influenzae type b. However, chemoprophylaxis during labor in pregnant women colonized with group B beta-hemolytic Streptococcus has reduced early-onset (infection occurring at < 7 days of age) group B streptococcal disease by > 80%. In addition, routine conjugate immunization has decreased colonization among older siblings immunized against S. pneumoniae and H. influenzae type b such that the rate of SBI caused by those organisms has decreased as well (herd immunity).
Notably, late-onset (infection occurring at > 7 days of age) group B streptococcal infection is not affected by chemoprophylaxis during labor, and other serious bacterial illnesses such as UTI (most commonly caused by Escherichia coli) and occasional cases of Salmonella bacteremia continue to be important causes of fever without apparent source on physical examination in infants < 3 months.
The major symptom of occult bacteremia is fever—temperature ≥ 39° C (≥ 38° C for infants < 3 months). By definition, children with apparent focal disease (eg, cough, dyspnea, and pulmonary crackles suggesting pneumonia; skin erythema suggesting cellulitis or septic arthritis) are excluded (ie, because their disease is not occult). A toxic appearance (eg, limpness and listlessness, lethargy, signs of poor perfusion, cyanosis, marked hypoventilation or hyperventilation) suggests sepsis or septic shock; bacteremia in such children is also not classified as occult or fever without a source. However, early sepsis can be difficult to distinguish from occult bacteremia.
Diagnosis of bacteremia requires blood cultures; ideally, two samples are taken from separate sites, which helps minimize the problem of false positives due to skin contaminants, and results should be made available within 24 hours.
Recommendations for testing and choice of tests vary with age, temperature, and clinical appearance; the goal is to minimize testing without missing an SBI. Children who have indications of focal infection on history or physical examination are evaluated based on those findings.
When available, rapid diagnostic tests for enteroviruses, respiratory syncytial virus, and influenza virus are useful in the evaluation of infants with fever without apparent source, because infants whose test results are positive for these viruses likely have fever resulting from that virus and require few or no further tests for SBI. There also are rapid tests for other viruses but these have not been studied sufficiently to justify using their results to alter testing for SBI.
In infants with SBI, the CBC usually shows an elevated white blood cell (WBC) count; however, only about 10% of children with WBC counts of > 15,000/mcL (> 15 × 109/L) are bacteremic, so specificity is low. Acute-phase reactants (eg, erythrocyte sedimentation rate, C-reactive protein with or without procalcitonin) are used by some clinicians but add little information; some clinicians believe that elevated procalcitonin levels may be more specific for serious illness. In children < 3 months, band counts > 1500/mcL (> 1.5 × 109/L) and either low (< 5000/mcL [5 × 109/L]) or high (> 15,000/mcL [> 15 × 109/L]) WBC counts may indicate bacteremia.
It is important to note that any febrile infant, regardless of immunization history, who appears seriously ill or toxic requires complete clinical and laboratory evaluation (CBC with differential, blood cultures, urine cultures, lumbar puncture, and, in most cases, admission to the hospital with empiric antimicrobial therapy). Unimmunized, underimmunized, and immunocompromised febrile infants in this age range are more susceptible to SBI than their peers and also typically require the same full clinical and laboratory evaluation for SBI and empiric antibiotics. Children with dyspnea or low oxygen saturation should also have chest x-ray.
In previously immunized febrile infants aged 3 to 36 months who appear well (nontoxic), the risk of bacteremia is now as low or even lower than the rate of false-positive blood cultures due to skin contaminants, leading many experts to forego blood cultures in these children. However, a urinalysis with microscopic examination and urine culture is typically recommended but additional laboratory examination (eg, CBC, chest-x-ray) is not. Although the vast majority of these children have a viral infection, a very small number of well-appearing children will have an early SBI, so caretakers should be advised to monitor the child's symptoms, give antipyretics, and follow up with the clinician (by visit or telephone depending on the circumstances and the caretakers' reliability) in 24 to 48 hours. Children who worsen or remain febrile should have testing done (eg, CBC with differential, blood cultures, possibly chest x-ray or lumbar puncture).
Toxic-appearing or seriously ill-appearing infants require immediate clinical evaluation and collection of blood, urine, and spinal fluid cultures and hospitalization for empiric antibiotic therapy. Unlike in older infants, in those < 3 months of age, a nontoxic clinical appearance does not routinely allow deferral of testing.
Algorithms have been developed to help guide evaluation of infants in this age group (for one example, see Figure: Evaluation and management of the febrile infant aged < 3 months). In using the algorithm, many experts consider age < 30 days by itself to be a high-risk criterion (and thus routinely admit them and do additional testing), whereas others do not and manage all infants < 90 days of age using the same criteria. This algorithm is sensitive for SBI but relatively nonspecific. Thus, given the relatively low incidence of SBI even among the population of febrile young infants, the algorithm has a high negative predictive value but a low positive predictive value (see Understanding Medical Tests and Test Results : Test Characteristics), making it much more effective in identifying children at low risk of infection who can be treated expectantly (ie, SBI or bacteremia ruled out) rather than in identifying children with true SBI or bacteremia.
Children who receive antibiotics before bacteremia is confirmed by blood culture seem less likely to develop focal infections, but data are inconsistent. However, because of the low overall incidence of bacteremia, many children would receive unnecessary treatment if all who were tested were empirically treated. As above, management varies by age and other clinical factors.
Regardless of age, all children are reexamined in 24 to 48 hours. Those with persistent fever or positive blood or urine cultures who have not been treated already have more cultures done and are hospitalized for evaluation of possible sepsis and parenteral antibiotic therapy. If new signs of focal infection are found on reexamination, evaluation and therapy are directed by the findings.
Antipyretics in weight-based dosage are given. Antibiotics are not given unless cultures are positive. For urinary tract infection, well-appearing children may receive oral antibiotics for pediatric UTI as outpatients; others (eg, those who appear more ill) may require admission for parenteral antibiotics.
One common system for management before culture results (see Figure: Evaluation and management of the febrile infant aged < 3 months) minimizes antibiotic use in most febrile infants who are not likely to have serious bacterial infection and provides antibiotics promptly to the few who need them. If urinalysis and urine cultures suggest urinary tract infection, well-appearing children may receive oral antibiotics for pediatric UTI as outpatients; others (eg, those who appear more ill) may require admission for parenteral antibiotics.
Of note, some authorities prefer to hospitalize all febrile infants < 1 month of age, do a full evaluation with cultures of blood, urine, and CSF, and give parenteral antibiotics (eg, ceftriaxone) pending culture results because febrile infants < 1 month of age are the age group with the greatest incidence of SBI.
Febrile infants and young children < 36 months of age, who have been appropriately immunized with H. influenzae type b and pneumococcal conjugate vaccines, and who look well and have no apparent foci of infection, are unlikely to have occult bacteremia or serious bacterial infection (SBI; eg, sepsis, meningitis).
Use blood cultures (2 samples from 2 separate sites) to diagnose occult bacteremia in selected febrile children.
All febrile infants < 36 months of age should be evaluated for urinary tract infection (UTI) with urinalysis and urine culture because UTI is now the most common cause of SBI with fever.
Toxic-appearing children (and perhaps all febrile infants < 1 month of age of age) also require cultures of blood and spinal fluid and hospitalization for empiric antibiotic therapy.
In children 3 to 36 months of age with a temperature ≥ 39° C and who have been appropriately immunized, testing other than urine culture is not indicated for those who appear well; others should have testing based on clinical findings and other circumstances (eg, rapid viral testing for influenza virus, respiratory syncytial virus, and enterovirus during the appropriate seasons).
In infants < 3 months of age with a temperature ≥ 38° C, good clinical appearance does not completely exclude an SBI, so testing, including urinalysis, complete blood count with differential, blood and urine cultures, and, if available (depending on the local epidemiology and season), perhaps rapid viral testing for influenza virus, respiratory syncytial virus, and enterovirus, is indicated for all in this age range.
Nontoxic, low-risk infants require close follow-up if they are not treated with antibiotics.