• –

  Meningitis: In patients with suspected meningitis, BC has a high diagnostic yield and should always be performed.9

• –

  Pneumonia: The prevalence of bacteraemia in children with community-acquired pneumonia ranges between 2.1% and 7%, and exceeds 10% in cases of complicated pneumonia and 20% in cases of empyema.10 Drawing a BC is recommended in all patients requiring admission due to their clinical condition or in whom complicated pneumonia is suspected.

• –

  Skin and soft tissue infection (SSTI): In a recent study on uncomplicated superficial SSTIs, none of the patients had bacteraemia, compared to 12.5% of patients with complicated infections.4 Thus, a BC should only be drawn under specific circumstances (Table 2).

• –

  Osteoarticular infections: Drawing of blood for culture is indicated prior to initiating antibiotherapy due to its high diagnostic yield (>50%).11

• –

  Urinary tract infections (UTIs) with fever: The rate of bacteraemia depends on age, reaching up to 20% in newborns and less than 5% in children older than 3 months. Thus, a BC should be drawn in infants aged less than 3 months with suspected UTI and in older children requiring intravenous therapy.2,12

• –

  Endocarditis (IE): Extraction of sample for BC is indicated in any child with suspected IE.

• –

  Intraabdominal infections: In case of suspected acute peritonitis, extraction of a sample for BC is recommended, although there is not sufficient data on its diagnostic yield and a sample of peritoneal fluid/exudate is the gold standard for culture.

• –

  Complicated infection of the ear-nose-throat (ENT) region: The option of drawing a BC should be considered in complicated infections of the ENT region requiring admission.

There is enough evidence to support the recommendation of drawing BCs in immunocompromised patients with a temperature of 38°C or higher, or even with a sustained low-grade fever.25,26 The real challenge arises if the isolated microorganism is part of the normal flora of the patient, for example, coagulase-negative staphylococci, viridians group streptococci, Corynebacterium species, Propionibacterium acnes, Bacillus species and some Clostridium species.26,27 In these cases, it is hard to differentiate contamination from true bacteraemia. One factor to consider is the number of BCs in which the same organism is isolated. Assuming that not all samples have been collected from the same contaminated source, the isolation of the same microorganism in more than one culture increases the probability of a true bacteraemia. Conversely, the presence of only one positive BC in two or more serial draws performed in a short period of time would be more suggestive of contamination.26,27 In any event, there is no definitive method to differentiate contamination and true bacteraemia, and in each case the interpretation of the results of BC will ultimately rest with the clinician.

In children with central vascular catheters or port-a-cath systems, the simultaneous drawing of two or more BC samples is recommended, with at least one central venous catheter sample and one venipuncture sample. The isolation of the same bacterium in both BCs increases the probability of a true bacteraemia, as well as the therapeutic recommendation of removing the colonised catheter.15,27 Faster rates of bacterial growth in the BC correlate to higher probabilities of true bacteraemia.28

We recommend that anaerobic cultures be ordered only in situations with a high degree of clinical suspicion17,29,30 (Table 3).

The epidemiology of bacteraemia is subject to constant changes that are related to antibiotic use, vaccination, geographical differences and emerging microorganisms.31–33 Currently, the most frequently involved pathogens are: (a) during the neonatal period: Streptococcus agalactiae, Escherichia coli, Klebsiella species and Listeria monocytogenes, and (b) starting from age 30 days: Streptococcus pneumoniae, enterobacteria, Staphylococcus aureus and Neisseria meningitidis (Table 4). In immunocompromised children, Pseudomonas aeruginosa and Candida species should also be considered. In recent years, advances in molecular biology methods have contributed to the identification of other aetiological agents, such as Kingella kingae or Ureaplasma species.31

The bacteria in the commensal skin flora are the most common contaminants leading to false-positive results in healthy children (table 4). However, we must take into account that they can be the cause of nosocomial or health care-related infections, especially in catheterised preterm infants and immunocompromised children with intravascular catheters.

With the advent of automated systems, there has been significant improvement in the time to detection of growth of pathogens that cause bacteraemia. The time to positivity (TTP) for 95% of these pathogens is less than 24h. While clinical criteria still prevail, the rapid detection of positive cultures is a factor to consider in the decision-making process. When the growth corresponds to contaminating organisms, the TTP is markedly longer, exceeding 24h in most cases.34

In children with central venous catheters, the difference in the TTP between BCs from catheter samples and from venipuncture samples allows the diagnosis of the focus of the bacteraemia (isolation ≥2h earlier in catheter samples indicates that the catheter is the focus of the bacteraemia). In order to assess the differential TTP, samples must be collected simultaneously and the same volume of blood inoculated in each of the BC bottles.35 The TTP has also proven useful for the differentiation between contamination and bacteraemia in BCs positive for coagulase-negative staphylococci; a TTP of 15h or less usually indicates a clinically significant result, while a TTP of 22h or more is suggestive of contamination.36

• –

  Persistent fever: If the source is not identified, a new BC must be drawn, performance of a lumbar puncture strongly considered in children aged less than 18 months, and parenteral antibiotherapy administered. In patients with good general health, after antibiotic treatment initiation, therapy can be maintained in the home for 7–10 days.

• –

  Afebrile patients: These patients may be treated with oral antibiotics at home under close monitoring. Performance of another BC prior to initiating antibiotherapy should be considered, as the risk of persistent bacteraemia is 9%.39 The initial antibiotherapy regimen should be:

  • -

    Amoxicillin: 80–90mg/kg/day in 3 doses (maximum, 3g/day).

  • -

    Patients allergic to penicillin will be prescribed treatment with a macrolide or cefuroxime axetil (30mg/kg/day) if there is no history of severe anaphylactic reaction.

In cases in which appropriate antibiotic treatment has already been initiated, the treatment will continue, and the decision to draw a follow-up BC made on a case-to-case basis. All cases will be closely monitored for 7–10 days by a primary care paediatrician.

The published studies on bacteraemia caused by organisms other than S. pneumoniae suggest that administration of oral antibiotherapy does not prevent severe bacterial infection even in afebrile children in good general health. Furthermore, the risk of meningitis is greater in patients with bacteraemia caused by N. meningitidis. Consequently, hospital admission is required in patients with a BC positive for N. meningitidis, Hib, S. aureus, Gram-negative bacteria or other pathogens. A lumbar puncture is recommended not only in patients with clinical manifestations of meningitis, but also in patients with a BC positive for N. meningitidis and infants aged 3–6 months with bacteraemia caused by group B Streptococcus. Conversely, a lumbar puncture is not routinely required in children in good general health aged more than 3 months with cultures positive for E. coli or S. aureus.

• –

  Duration of antibiotic treatment: Until the results of the antibiogram become available, antibiotic treatment will be initiated based on the suspected diagnosis and the sensitivity of the microorganism in each medium. The recommended duration is 7–10 days for uncomplicated bacteraemias, and the customary duration if the focus of infection is found.

• –

  Indication for extraction of a follow-up BC sample: The benefits of follow-up BCs for establishing the microbiological cure of bacteraemia have not been proven.40 It is recommended that follow-up BCs are performed 48–96h after initiation of antibiotic treatment in the following cases:

  • -

    Bacteraemia by S. aureus (especially MRSA) or multidrug-resistant Gram-negative bacilli.

  • -

    Candidaemia.

  • -

    Persisting fever or absence of clinical improvement after 48–96h of appropriate antibiotic treatment.

  • -

    Recurrence of fever.

  • -

    Suspected IE.