The term FN entails the cooccurrence of fever and neutropenia.3–5 The definition of FN allows the establishment of validated recommendations for its management. The recommendations of this guideline may not be the most appropriate in other situations in which there is risk or suspicion of infection (e.g., fever without severe neutropenia), for which specific protocols should be established.

The relative frequency of isolation of gram-positive cocci (GPC) and gram-negative bacilli (GNB) depends on various factors, such as the use of central venous catheters and their care (coagulase-negative staphylococci), the emergence of Viridians streptococci in association with haematopoietic stem cell transplantation (HSCT), more aggressive induction chemotherapy regimens and the presence of mucositis,1,2,6,7 with a greater frequency of GPC in blood compared to solid tumours and in patients who undergo HSCT and aggressive induction therapy.

We are witnessing an increase in the prevalence of MDR pathogens, chief among them extended spectrum beta-lactamase (ESBL)-producing bacteria6,8–11 and carbapenem-resistant bacteria (especially Pseudomonas aeruginosa and Klebsiella pneumoniae),6–8,12 which is associated with an increased severity and mortality, and the prevalence of methicillin-resistance Staphylococcus aureus in Spain has remained unchanged. In consequence, empiric antibiotherapy must be selected taking into account individual risk as well as local epidemiological trends.

Viral detection is associated with less severe disease and a decreased length of stay,13,14 but its impact on the discontinuation of antibiotherapy in these patients remains small,14 probably due to the impossibility of ruling out bacterial coinfection with absolute certainty and the prolonged shedding of respiratory viruses in these patients, which complicates the interpretation of viral detection. On the other hand, the detection of a viral pathogen is useful for the purpose of implementing infection control measures and to initiate effective antiviral therapy in the case of viruses for which it is available. In addition, in patients at low risk who are stable and improving, viral detection could guide decision-making regarding antimicrobial de-escalation or discontinuation.5

There are several risk stratification models validated in the paediatric population (Table 2) that allow the differentiation of groups of patients with FN with a probability of bacterial infection of less than 10% and a low risk of complications. Their application allows the implementation of outpatient care protocols and/or early discharge in patients with low-risk FN. Although many studies have excluded patients aged less than 1 year or who develop fever while hospitalised, the study by Haeusler et al. included these subsets of patients.15

These models have exhibited a good sensitivity and negative predictive value and therefore can be used to identify patients eligible for initiation of treatment at the outpatient level or for antimicrobial de-escalation and early discharge (after 24–48 h of clinical stability),16,17 with an admission/readmission rate of less than 10% and a very low probability of significant adverse events.

In Spain, the multicentre clinical trial e-STOP (EudraTC 2018-000775-32) is currently underway to assess the safety of early discontinuation of antibiotherapy in patients with FN and validate a clinical scoring system in the Spanish population.

More recently, an algorithm for risk stratification has been developed and published (DISCERN-FN) that has exhibited a high sensitivity (95%–98%) and an acceptable specificity (38%–58%) in the detection of severe infection, which has yet to be validated in Spain.18

The initial assessment of paediatric oncological/haematological patients with FN must include an exhaustive history-taking, both general and focused on identifying potential foci of, an assessment of immunocompetence and associated risks, a careful and detailed physical examination—including vital signs and assessment of signs of sepsis—and blood tests including a complete blood count, blood chemistry panel and measurement of inflammatory markers, such as C-reactive protein (CPR)5 and procalcitonin (PCT),18 with the option of enhancing their sensitivity by expanding the assessment with additional markers (e.g., interleukin-6, interleukin-8, etc.).7Table 3 summarises the recommendations regarding the history-taking, physical examination and diagnostic tests.

The performance of other diagnostic tests, which should not delay initiation of treatment, may include collection of samples for blood culture from all central line lumens. In addition, consider collection of peripheral blood samples for culture, especially in the case of suspected catheter-related bloodstream infection (for differential quantitative or qualitative culture, in this case, ensuring collection of the same volume and culture incubation time), non-invasive urine sample and sample for detection of respiratory viruses during epidemic seasons and/or in patients with respiratory symptoms. The performance of additional diagnostic tests will be based on the signs and symptoms.3–5

Furthermore, monitoring of clinical manifestations, haemodynamic status and laboratory values helps identify previously undetected foci of infection and complications and assess the response to treatment, which can then be adjusted as needed.7

Early initiation of broad-spectrum antibiotherapy in these patients is a crucial aspect that has an impact on patient outcomes.3,4 Given the high risk of infection by P. aeruginosa and the high associated morbidity and mortality, we recommend selecting agents active against this pathogen for empiric antibiotherapy. Fig. 1 presents the algorithm to select treatment based on individual risk, and Table 4 the dosage of the most frequently used antimicrobial agents.

Figure 1.

Algorithm for the management of episodes of febrile neutropenia in paediatric oncological/haematological patients.

GNB, gram-negative bacilli; GPC, gram-positive cocci; MDR, multidrug-resistant.

Source: adapted from Lehrnbecher et al., 2014.4

(0.44MB).

The mean time to positivity for blood cultures in patients with FN is of approximately 12 h, with a majority turning positive in the first 24 h.25 A time to positivity greater than 24–48 h is usually indicative of infection by slow-growing microorganisms (Candida spp., anaerobes or CoNS) or contamination.

In adults with blood tumours and high-risk FN of unknown aetiology, the discontinuation of empiric antibiotherapy 72 h after the patient becomes afebrile and all signs of symptoms of infection have resolved, as opposed to the standard approach of maintaining antibiotic treatment until the neutrophil count recovers, has been found to be safe, and it allows to reduce the duration of antimicrobial exposure with no impact on mortality. The updated guidelines for adult patients of the Fourth European Conference on Infections in Leukemia (ECIL-4) established that in patients with FN and no clinical or microbiological evidence of infection, empiric antibiotherapy can be discontinued after a minimum of 72 h of intravenous treatment in patients who have been haemodynamically stable from onset and afebrile for at least 48 h, independently of the neutrophil count or the expected duration of neutropenia.

When it comes to the paediatric population, Kobayashi et al.26 analysed 170 episodes of FN in which antibiotherapy had been discontinued before neutrophil recovery to assess the risk of recurrence. In the multivariate analysis, a neutrophil count of less than 100 cells/mm3 at the time of discontinuation was an independent risk factor for recurrent fever. Similarly, an observational study in children with FN concluded that discontinuation of empiric antibiotherapy in those who did not have bacterial infection and with neutrophil counts greater than 100 mm3 after 24 h of being afebrile was safe.27 An international paediatric guideline considers discontinuation of empiric antibiotherapy in low-risk patients with no documented pathogen safe if they have been afebrile for 24 h and received antibiotherapy for 72 h, irrespective of the neutrophil count, and gave no recommendation for the specific scenario of persistent severe neutropaenia.3

Based on the available evidence, we recommend reassessing the need for empiric antibiotherapy 48–72 h after its initiation, considering de-escalation and discontinuation of combination therapy. Once empiric antibiotherapy is initiated, subsequent adjustments should be based, preferably, on culture results or the identified focus of infection to establish targeted therapy against the isolated microorganism and for infection control at the focus, if applicable (Fig. 1).

When it comes to the use of inflammatory markers in the follow-up, the current evidence is not sufficient to support their use in guiding decisions regarding the duration of antibiotherapy.28

The incidence of invasive fungal infection (IFI) in paediatric patients with cancer or who have undergone HSCT ranges from 2% to 20%.29,30 Individual risk depends on the underlying disease of the patient and associated factors (Table 6), of which prolonged severe neutropenia is the most relevant.

The fungi isolated most frequently in these patients are Candida spp. and Aspergillus spp., which are associated with a mortality of 10%–25% and of 20%–50%, respectively.29

Invasive fungal infection should be considered and suspected in patients with FN that persists for 96 h or longer despite appropriate antibiotherapy, especially those classified as being at high risk of IFI. Collection of samples, including blood and from the site of infection, is essential.31,32

We recommend performance of a serum galactomannan antigen test in patients with prolonged FN and risk of aspergillosis (cut-off point, 0.5). It is also possible to assess the presence of galactomannan in bronchoalveolar lavage and cerebrospinal fluid samples (cut-off point, 1).31–33 At present, routine testing for other biomarkers is not recommended for diagnosis of IFI in paediatric patients.31 Molecular tests (specific or panfungal [18S] PCR assays, T2MR) offer some advantages compared to traditional methods and need to be interpreted by an experienced microbiologist.31,34

A chest CT scan is indicated in paediatric patients at high risk of IFI with FN of more than 96 h’ duration or with respiratory manifestations and an inconclusive chest radiograph. In the case of abnormal findings in the chest CT scan, it is crucial to obtain bronchoalveolar lavage or lung biopsy samples for microbiological testing.31,35 Abdominal imaging should be considered to rule out invasive candidiasis if a focus is not identified, in addition to other imaging tests depending on the clinical presentation and level of risk. Based on the findings, IFI is classified as possible, probable or proven (criteria of the European Organization for Research and Treatment of Cancer [EORT]/ Mycoses Study Group [MSG]).36

Two strategies have been established for the management of IFI in paediatric oncological/haematological patients with FN:

Empiric antifungal therapy consists in initiating antifungal treatment as soon as fungal infection is suspected, without waiting for the results of the applicable diagnostic tests. Consider caspofungin or liposomal amphotericin B (standard of care if there is a high risk of filamentous fungal infection), and possibly switching to a different class if the patient had received antifungal prophylaxis in the past (especially in the case of azoles).

Pre-emptive therapy consists in carrying out the tests first (imaging and fungal biomarkers) and initiating antifungal treatment once one of the results has turned out positive based on the findings.31,37

Both strategies have proven safe in paediatric care,38 so the choice will depend on the characteristics of the patient, the suspected microorganism and the resources and epidemiological trends of the centre. In clinically unstable patients with additional risk factors for candidaemia, administration of empiric therapy against yeasts should be prioritised, while in clinically stable patients with prolonged FN or clinical suspicion of filamentous fungal infection, we recommend waiting for tests results if the turnaround time is short.

If fungal infection is confirmed, it is important to control the focus of infection, either by removing indwelling devices or surgical resection, if necessary, as well as predisposing factors.

Once treatment has been initiated and all necessary diagnostic tests have been done to identify the aetiological agent and assess disease extension, we recommend monitoring the response to treatment (clinical, laboratory and imaging findings). In the case of candidaemia, a follow-up blood culture should be performed every 48 h until resolution of fever and one or more cultures turn out negative.

When it comes to imaging for monitoring, the EORTC/MSG recommendation is carrying out imaging tests at 4–6 and 12 weeks of treatment. However, in some patients it may be necessary from the second week of treatment.31,32,34

Monitoring of plasma drug concentrations is essential in the case of voriconazole, posaconazole (oral suspension) and itraconazole, and should also be considered for isavuconazole until more data are available for the paediatric population.34,37

In the case of infection by Candida spp., treatment must be maintained for a minimum of 14 days after the first negative culture in patients with candidaemia, and longer in the case of infection of a solid organ.39 In the case of a catheter-related infection, the catheter must be removed. In the case of infection by filamentous fungi, we recommend a minimum duration of treatment of 6 weeks, to be prolonged as needed until the full clinical/laboratory/radiological resolution of the infection.31,32,34

Supportive care (Table 7) encompasses general measures, such as fluid replacement therapy, analgesic and antipyretic drugs, any other ongoing treatment the patient is receiving, consideration of the need of stress doses of hydrocortisone (patients with current or recent steroid therapy) and measures aimed at reducing the incidence of infection in neutropenic patients.40

Treatment with granulocyte colony-stimulating factor (G-CSF) shortens the duration of neutropenia, in addition to the duration of antibiotherapy and the length of stay, but there is no evidence that it reduces the frequency or severity of infection. It is generally recommended in patients with solid tumours or lymphoma, but only occasionally in patients with leukaemia.3