Fever in children with chemotherapy-induced neutropenia

INTRODUCTION — Infection is a major cause of morbidity and mortality in patients with cancer. Fever may be the first manifestation of a life-threatening infection, particularly during periods of neutropenia.

Evaluation and management of fever in children with chemotherapy-induced neutropenia are reviewed here. The types of infections and management of fever in the child with other forms of neutropenia are discussed separately. (See "Evaluation of children with non-chemotherapy-induced neutropenia and fever" and "Management of children with non-chemotherapy-induced neutropenia and fever".)

DEFINITIONS

●Neutropenia – For the purposes of management of fever in children with cancer or hematopoietic cell transplant, neutropenia is defined as an absolute neutrophil count (ANC) <500 cells/microL or an ANC that is expected to decrease to <500 cells/microL during the next 48 hours [1,2].

The ANC is calculated using the following formula (calculator 1):

●Fever – In neutropenic patients, fever is defined as a single oral temperature ≥38.3°C (101°F), a temperature ≥38°C (100.4°F) for longer than one hour, or two elevations >38°C (100.4°F) during a 12-hour period [1,3,4].

Rectal temperature measurement should be avoided in neutropenic patients to avoid the possibility of introducing infection through mucosal trauma. Oral temperature measurement is preferred to other nonrectal sites.

An axillary temperature is acceptable if the patient is unable to use an oral thermometer. Many electronic thermometers automatically adjust for oral, axillary, or rectal location If the electronic thermometer automatically adjusts for location, the thresholds stated above for oral temperature measurements should be used. If the axillary thermometer does not automatically adjust, some centers consider fever as a single axillary temperature of ≥37.7°C (99.9°F) or axillary temperature ≥37.4°C (99.4°F) for longer than one hour [5].

EPIDEMIOLOGY — Fever occurs in approximately one-third of neutropenic episodes in children with cancer or hematopoietic cell transplantation (range 10 to 60 percent) [6]. The approximate rate of occurrence is 0.76 episodes per every 30 days of neutropenia.

In the United States between 2007 and 2014, the rate of hospitalizations for fever and neutropenia in pediatric cancer patients ranged from 13 to 18 per 100,000 population and increased over time [7]. The median length of stay was four to five days. The overall mortality rate was 0.75 percent. Sepsis, pneumonia, meningitis, and mycosis were the comorbidities most frequently associated with mortality.

FEVER AND INFECTION IN CHILDREN WITH CANCER

Clinical presentation and risk factors — Fever is often the only sign of occult infection in the neutropenic host [1]. However, serious infection may occur in the absence of fever or neutropenia. Children with infection who are receiving glucocorticoids may present with a lower and/or intermittent temperature elevation.

Serious infection must be considered and promptly treated in pediatric cancer patients or hematopoietic cell transplant (HCT) recipients who are [1,2,8]:

●Febrile and neutropenic

●Febrile and not neutropenic

●Afebrile and neutropenic with signs of infection or clinical deterioration (eg, hypothermia, hypotension, listlessness, confusion)

Risk factors for serious infection in children with cancer or HCT recipients include [1,9]:

●Chemotherapy-induced neutropenia

●Central venous catheters (whether or not the child is neutropenic); the risk of serious bacterial infection is lower with totally implantable catheters (eg, mediports, portacaths) than with tunneled external catheters or peripherally inserted catheters [8,10]

●"Functional neutropenia" (eg, hematologic malignancies that impair phagocytosis and killing of pathogens, even if the absolute neutrophil count is normal)

●Breakdown of skin and mucosal barriers (eg, mucositis)

●Altered humoral and cellular immunity, particularly if neutropenia is anticipated to last for >7 days [11]

Risk factors for invasive fungal infection are discussed below. (See 'Addition of antifungal therapy' below.)

Infectious causes of fever — In children with chemotherapy-induced neutropenia, the rate of documented infection ranges between 10 and 40 percent [6,12-17]. The most common documented infection is bacteremia, which accounts for approximately 20 to 50 percent; the proportion varies with the underlying malignancy and the institution [13,14,18-22]. Other sites of infection include the gastrointestinal tract (related to oral or intestinal mucositis), upper and lower respiratory tract, urinary tract, and skin and soft tissues [12,23,24].

Gram-positive and gram-negative organisms are frequently isolated from the blood in children with fever and chemotherapy-induced neutropenia [6,14,25]. The frequency of isolation of specific organisms varies from institution to institution. However, with increased use of prophylactic antimicrobials and indwelling venous catheters, there has been a general shift toward dominance of gram-positive organisms [26].

●Gram-positive bacteria – The most common gram-positive pathogens are coagulase-negative staphylococci, viridans streptococci, and Staphylococcus aureus (including methicillin-resistant S. aureus) [6,14,17,20,27].

●Gram-negative bacteria – Aerobic gram-negative bacilli account for approximately one-third to one-half of bacteremic episodes, with Escherichia coli, Klebsiella spp, Pseudomonas spp, Acinetobacter spp, and Enterobacter spp among the more common isolates [6,14,17,20,27].

●Fungi – Fungi, typically Candida spp, are more likely to be recovered after prolonged courses of broad-spectrum antibiotics but occasionally may be the primary pathogen [6]. Other potential fungal organisms include Aspergillus spp, Zygomycetes, and Cryptococcus spp [28,29]. Fusarium spp and dematiaceous fungal infections also have been increasingly recognized in these hosts [30,31].

The increasing use of antifungal prophylaxis has shifted the distribution of fungal isolates away from Candida and Aspergillus spp toward more unusual mold infections [32]. This is particularly true in HCT recipients.

●Viruses – Respiratory viruses, herpes simplex virus, and varicella-zoster virus are relatively common viral causes of infection in children with fever and chemotherapy-induced neutropenia [14,33-40].

Noninfectious causes of fever — Noninfectious causes of fever in children with cancer include drug fever, cancer-related fever, deep vein thrombosis, pulmonary embolus, transfusion reaction, and dysautonomia (in children with central nervous system disease).

EVALUATION — Infection is a major cause of morbidity and mortality in patients with fever and chemotherapy-induced neutropenia. Rapid and thorough evaluation is critical so that empiric antimicrobial therapy can be administered promptly (as soon as possible and always within 60 minutes of triage). (See 'Prompt initiation of antimicrobial therapy' below.)

History and examination

●History – Important aspects of the history in a child with fever and neutropenia include [1]:

•New site-specific symptoms of localized infection

•Antimicrobial prophylaxis (may decrease the risk of infection and modify the choice of empiric antimicrobial therapy)

•Infection exposures

•Recent documented infections or colonization (may affect the choice of empiric therapy)

•Underlying comorbid conditions, such as diabetes or recent surgery (increase the risk of infection)

•Previous chemotherapy, agents used, and the stage of therapy (to anticipate the length of the neutropenic episode)

•Current medications, including glucocorticoids and immunosuppressive agents for children with hematopoietic cell transplantation (HCT) (may extend the risk of serious infection beyond the duration of neutropenia)

•Intravascular catheters or other devices (see 'Clinical presentation and risk factors' above)

•Concomitant noninfectious cause of fever (eg, receipt of blood products) (see 'Noninfectious causes of fever' above)

•Vaccination history (affects the infections to be considered in the differential diagnosis)

●Physical examination – The physical examination should be thorough and focused on the most common sites of infection (listed below) [1]. In children with neutropenia, clinical findings of inflammation may be subtle; mild erythema or tenderness should not be ignored. The physical examination should be repeated at least daily. Signs of inflammation may become evident as the neutrophil count recovers.

Important aspects of the physical examination in children with fever and chemotherapy-induced neutropenia include:

•Abnormal vital signs (eg, hypotension, tachycardia); wide pulse pressure may indicate septic shock; tachycardia out of proportion to the degree of fever or tachycardia that persists despite defervescence may be a sign of early septic shock (even without hypotension)

•Skin (looking for poor perfusion; slow capillary refill and signs of localized infection, especially in skin folds, areas surrounding nail beds, central venous catheter exit sites and subcutaneous tunnel [if present]; and sites of bone marrow aspiration and lumbar puncture)

•Sinuses

•Oropharynx, including the gums, which may be a source for bacteremia [41]

•Lungs (particularly for signs of pneumonia [eg tachypnea, crackles, increased work of breathing])

•Abdomen

•Perineum, particularly the perianal and labial regions; digital rectal examination should be avoided

•Mental status changes

Routine laboratory and imaging studies — The laboratory evaluation for the child with fever and neutropenia should include (at minimum) [1,42]:

●Complete blood count with differential and platelet count

●Electrolytes, creatinine, and blood urea nitrogen

●Liver transaminases and total bilirubin

●Blood cultures

●Viral panel, including severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and influenza as indicated by local epidemiology and/or compatible symptoms

Additional cultures, molecular diagnostic assays, and/or imaging should be obtained if localized infection is suspected. (See 'Additional studies as indicated' below.)

Treatment should be initiated as soon as possible after blood cultures are obtained and always within 60 minutes of triage. It should not be delayed while waiting for results or performing additional studies. (See 'Prompt initiation of antimicrobial therapy' below.)

Although we do not routinely obtain them, in addition to the tests listed above, some experts also include C-reactive protein (CRP), lactate, and urinalysis in children <5 years as part of the initial evaluation [2]. CRP has been used in risk categorization strategies in several pediatric studies [43,44].

●Blood cultures – Blood cultures should be obtained without delay. Blood cultures should be taken from each lumen of the central catheter when such access is available [42,45]. In observational studies, approximately 30 to 40 percent of positive cultures from multiple lumen catheters were positive from only one lumen [46,47].

We also obtain peripheral blood cultures. Culturing blood from both peripheral and central sites helps to distinguish true bacteremia from infection related to the catheter and to guide decisions about catheter removal [48]. Peripheral cultures increase the identification of true bacteremia. In a meta-analysis of observational studies, 13 percent of bloodstream infections in patients with cancer were detected only in the peripheral culture [49]. This potential increased yield is balanced against pain and potential isolation of contaminants [42]. Diagnosis of intravascular catheter-related infection is discussed separately. (See "Intravascular non-hemodialysis catheter-related infection: Clinical manifestations and diagnosis", section on 'Diagnosis'.)

Obtaining more than one blood culture is helpful in the interpretation of blood culture results. If coagulase-negative staphylococci are isolated from two or more blood cultures, true bacteremia is more likely than contamination of the specimen, which may be reflected by a single positive blood culture.

Repeat blood cultures are discussed below. (See 'Reevaluation during treatment' below.)

●Urine culture – We obtain urine cultures in children with urinary symptoms, history of urinary tract infection (UTI), and/or urinary tract abnormalities. Other experts routinely obtain urine cultures in females or young children who may not complain of urinary tract symptoms.

Urinalysis should not be used to determine the need for urine culture. In a small observational study of children with UTI, pyuria was found in only one of 23 neutropenic patients, compared with 21 of 31 without neutropenia [50].

Additional studies as indicated — Additional studies should be obtained as clinically indicated for signs or symptoms of localized infection. Examples include:

●Respiratory signs and symptoms – Respiratory virus molecular analysis (including for SARS-CoV-2 and seasonal respiratory viruses [eg, influenza, respiratory syncytial virus] as indicated by local epidemiology and/or compatible symptoms, if not already obtained [51]) or culture of nasopharyngeal wash [38] and chest radiograph [42,52,53]

In children with respiratory signs and symptoms, chest radiographs without infiltrates should be interpreted with caution. The appearance of infiltrates may be delayed until neutrophil counts are recovering.

●Any site with purulent drainage – Gram stain and culture

●Abdominal pain and diarrhea (suspected neutropenic enterocolitis [typhlitis]) – Abdominal ultrasonography and anaerobic blood culture

●Diarrhea

•Children with HCT – Testing for Clostridioides difficile and viral pathogens (see "Acute viral gastroenteritis in children in resource-abundant countries: Clinical features and diagnosis", section on 'Etiologic diagnosis' and "Clostridioides difficile infection in children: Clinical features and diagnosis", section on 'Testing for C. difficile')

•Children with ≥3 loose bowel movements per day without a known source – Testing for C. difficile (see "Clostridioides difficile infection in children: Clinical features and diagnosis", section on 'Testing for C. difficile')

●Altered mental status or meningeal signs – Lumbar puncture; platelet transfusion may be necessary before lumbar puncture in patients with neutropenia

●Skin and soft tissue lesions – Aspiration and/or biopsy for microbial staining, cultures and other microbiologic studies for suspected pathogens, histology, and cytology [24]

REEVALUATION DURING TREATMENT — Children with fever and chemotherapy-induced neutropenia should be reevaluated at least once daily during antimicrobial therapy. The evaluation typically includes:

●Review of systems for infection

●Physical examination with particular attention to the sites most frequently infected (see 'History and examination' above)

●Complete blood count with differential and platelet count daily during hospital admission

●For children who remain febrile after initiation of empiric antimicrobial therapy, repeat blood cultures once daily for the next two days and then if there is deterioration of clinical condition [1,16,54]; some experts recommend daily blood cultures regardless of clinical status

●Repeat blood culture for recurrent fever (after being afebrile for ≥24 hours)

●Cultures and/or imaging for any suspected sites of infection

Additional evaluation for children with clinical worsening or instability is discussed below. (See 'Clinical worsening or instability' below.)

RISK STRATIFICATION — Patients with fever and neutropenia can be categorized as high- and low-risk for severe infection or complications based upon presenting signs and symptoms, absolute neutrophil count (ANC), underlying cancer, anticipated duration of neutropenia (varies with type of chemotherapy), and medical comorbidities [1,42].

Several risk stratification strategies that incorporate one or more of these factors have been validated in children and are used at various institutions [13,43,44,55-57]. The choice of strategy varies by institution [42,58]. We provide some general guidelines below (table 1).

●High-risk – High-risk patients have an increased risk of severe infection or complications and should be admitted to the hospital for empiric antimicrobial therapy. (See 'Prompt initiation of antimicrobial therapy' below.)

We consider patients with any of the following to be high risk (table 1) [1,42]:

•Neutropenia (ANC <500 cells/microL) anticipated to last >7 days

The relative risk of infection is related to both the degree and duration of neutropenia. An increased risk becomes apparent at an ANC <1000 cells/microL, is greater at an ANC <500 cells/microL, and greatest at an ANC ≤100 cells/microL (profound neutropenia) [6,59]. Patients with neutropenia projected to last for more than seven days also are at a higher risk of infection than are those with neutropenia of shorter duration [11,60].

Formal studies to clearly differentiate between patients with an ANC <500 cells/microL and ANC ≤100 cells/microL are lacking. Our ANC threshold for high-risk differs from that in the 2010 Infectious Diseases Society of America guidelines (≤100 cells/microL anticipated to last >7 days) [1].

•Evidence of hepatic insufficiency (aminotransferase levels >5 times normal values)

•Evidence of renal insufficiency (creatinine clearance <30 mL/min)

•Comorbid medical problems including, but not limited to:

-Hemodynamic instability

-Oral or gastrointestinal mucositis that interferes with swallowing or causes diarrhea

-Gastrointestinal symptoms, including abdominal pain, nausea, vomiting, or diarrhea

-New-onset neurologic or mental status changes

-Signs of intravascular catheter infection, particularly catheter tunnel infection (eg, erythema, swelling, and/or tenderness at the insertion site, rigors or chills associated with catheter flushing or infusion)

-New pulmonary infiltrate or hypoxemia or underlying chronic lung disease

-New findings or symptoms associated with localized infection

•Infants with acute lymphoblastic leukemia

•Patients with acute myeloid leukemia [61]

•Patients within 30 days of hematopoietic cell transplant (HCT); some centers extend this period to within 100 days of HCT

●Low-risk — Low-risk patients are those with all of the following (table 1) [1]:

•Neutropenia expected to resolve within seven days (ie, neutrophil count increasing)

•Stable and adequate hepatic and renal function

•No active comorbidities (eg, hemodynamic instability, mucositis, gastrointestinal symptoms)

Children initially assessed to be at low risk can move into the high-risk category after initial presentation if they develop a high-risk feature.

Carefully selected low-risk patients may be candidates for oral empiric therapy or outpatient treatment. (See 'Empiric therapy for low-risk patients' below.)

PROMPT INITIATION OF ANTIMICROBIAL THERAPY — Prompt initiation (ie, as soon as possible and always within 60 minutes of triage) of empiric-broad spectrum antibiotics is the cornerstone of therapy for patients with fever and chemotherapy-induced neutropenia. In an observational study in pediatric cancer patients with fever and neutropenia, receipt of antibiotics within 60 minutes of presentation was associated with decreased rates of consultation for or admission to the intensive care unit [62]. Conversely, administration of antibiotic therapy more than 60 minutes after presentation has been associated with increased adverse outcomes and length of stay [63-65].

Our approach to the selection of empiric antibiotic therapy in children with fever and chemotherapy-induced neutropenia, detailed in the sections below, is generally consistent with the recommendations of expert guidelines (table 2A-B) [1,2,42]. (See 'Society guideline links' below.)

EMPIRIC THERAPY FOR HIGH-RISK PATIENTS

Broad-spectrum monotherapy — Children with fever and chemotherapy-induced neutropenia who are at high risk for severe infection or complications (table 1) should be hospitalized for empiric intravenous (IV) antimicrobial therapy. The initial empiric regimen is individualized (eg, for drug allergies, organ dysfunction, antimicrobial prophylaxis) [1,2,42,66].

●For high-risk children with fever and chemotherapy-induced neutropenia who are clinically stable and in whom there is no suspicion of resistant infection (eg, methicillin-resistant S. aureus [MRSA]), we recommend initial broad-spectrum monotherapy with an antipseudomonal beta-lactam agent (eg, cefepime, meropenem, piperacillin-tazobactam [doses are provided below]) to provide activity against the pathogens that are associated with serious or life-threatening infections in neutropenic patients.

●For high-risk children with fever and neutropenia following hematopoietic cell transplantation (HCT), cefepime is preferred to meropenem and piperacillin-tazobactam [67]. In observational studies, the use of antibiotics with anaerobic activity (eg, meropenem, piperacillin-tazobactam) eradicates beneficial gut microbiota and has been associated with increased rates of graft-versus-host disease and overall mortality [68-71].

Ceftazidime is no longer recommended for empiric monotherapy in high-risk patients because of increased resistance among many gram-negative pathogens and weak activity against viridans streptococci.

Combination therapy is reserved for children with clear indications for additional activity against gram-positive, gram-negative, or anaerobic bacteria (eg, hemodynamic instability, localized infection, concern for resistant pathogens) (table 2B) [1,2,42]. (See 'Limited indications for combination therapy' below.)

The doses for antimicrobials used for initial empiric therapy in patients with normal renal and hepatic function are as follows:

●Cefepime – 50 mg/kg IV every 8 hours up to a maximum of 2 g per dose, or

●Meropenem (for children ≥3 months of age), or

•For noncentral nervous system infections – 20 mg/kg IV every 8 hours up to a maximum of 1 g per dose

•For central nervous system infection – 40 mg/kg IV every 8 hours up to a maximum of 2 g/dose

●Piperacillin-tazobactam (dosed according to the piperacillin component)

•<2 months of age – 80 mg/kg IV every 6 hours

•2 to 9 months of age – 80 mg/kg IV every 8 hours

•>9 months

-Weight <40 kg – 100 mg/kg IV every 6 to 8 hours (maximum of 16 g/day)

-Weight ≥40 kg – 3 g IV every 6 hours; this regimen aims to optimize drug levels for organisms with higher minimum inhibitory concentrations [72]

Most patients with penicillin allergy can tolerate cephalosporins. However, those with a history of immediate-type penicillin allergy (table 3) or serious delayed reactions (table 4) should not receive cephalosporins or carbapenems. Although carbapenems often are avoided in patients with penicillin allergy because of potential cross reactivity, few studies describing the cross-reactivity between penicillin and carbapenems have been performed [73]. Appropriate alternatives for children with immediate-type hypersensitivity reaction to penicillin include either ciprofloxacin plus clindamycin or aztreonam plus vancomycin [1]. (See "Penicillin allergy: Immediate reactions".)

Monotherapy with cefepime, meropenem, or piperacillin-tazobactam provides activity against a broad spectrum of potential pathogens, particularly those that are associated with serious or life-threatening infections in neutropenic patients. (See 'Infectious causes of fever' above.)

In randomized trials and systematic reviews, empiric monotherapy with these agents is as efficacious as combination therapy but with fewer adverse events [74-76]. In the systematic review that was limited to children, treatment failure rates, overall mortality, and infection-related mortality were similar with monotherapy and aminoglycoside-containing combination therapy [74]. Similar rates of treatment failure, rates of infection-related mortality, and duration of fever were observed in direct comparisons of monotherapy with fourth generation cephalosporins (eg, cefepime) and monotherapy with piperacillin-tazobactam. In a subsequent randomized trial, the efficacy and safety of piperacillin-tazobactam and meropenem were similar [77].

Aminoglycoside monotherapy is not recommended as initial empiric therapy because of the rapid development of resistance. In addition, aminoglycosides are associated with oto-and nephrotoxicity and their use requires monitoring of drug levels and kidney function [78]. They may be added to initial therapy in certain patients with complications. (See 'Limited indications for combination therapy' below.)

The use of vancomycin monotherapy for initial empiric therapy of febrile neutropenia is limited to reduce the risk of colonization or infection with vancomycin-resistant Enterococcus and because it has not been associated with clinical benefit [1,2,79,80]. In randomized trials in cancer patients with febrile neutropenia, morbidity and mortality were similar whether or not vancomycin was included in the initial antibiotic regimen [76,81-83]. However, vancomycin may be added to initial therapy in certain patients with complications. (See 'Limited indications for combination therapy' below.)

Limited indications for combination therapy — Additions to empiric therapy may be necessary for certain high-risk patients (eg, clinical instability, specific complications) (table 1 and table 2B). When adding agents to empiric therapy, the combination of vancomycin and piperacillin-tazobactam should be avoided if possible because it has been associated with increased risk of acute kidney injury [84,85].

Examples of clinical presentations that may warrant broad-spectrum combination empiric therapy are listed below [1,14,42]. In these circumstances, suggested antimicrobial agents should be added to initial monotherapy.

Signs of sepsis — Signs of sepsis include fever, hypotension, unexplained tachycardia and/or widened pulse pressure, mental status changes, and respiratory dysfunction. After consultation with an expert in pediatric critical care for supportive management of septic shock and with an expert in pediatric infectious diseases, the empiric regimen for children with signs of sepsis should be modified according to institutional clinical practice guidelines and clinical features, including previous culture results, concern for multidrug-resistant organisms, and local resistance patterns [86]. Addition of antimicrobial agents with activity against gram-negative bacteria (eg, aminoglycoside or fluoroquinolone) and MRSA (eg, vancomycin (table 5), linezolid if known to be colonized with or previously infected with vancomycin-resistant organisms) may be warranted. As examples:

●For children with a history of multidrug-resistant organisms, we provide broad-spectrum antipseudomonal therapy with meropenem or a novel beta-lactamase inhibitor combination (eg, ceftazidime-avibactam, ceftolozane-tazobactam [87,88]).

●For children with a history of or concern for carbapenem-resistant Enterobacterales, we provide broad-spectrum antipseudomonal therapy with a novel beta-lactamase inhibitor combination (eg, ceftazidime-avibactam).

Gastrointestinal symptoms — Abdominal pain, rectal pain, perineal inflammation, or blood per rectum – Add anaerobic coverage (eg, metronidazole) if not already receiving antibiotics active against anaerobes (eg, meropenem, piperacillin-tazobactam) [89].

If infection with C. difficile is suspected, modifications to the empiric regimen vary with the clinical scenario [90,91]:

●Nonsevere initial episode – Add metronidazole or oral vancomycin

●Severe initial episode – Add oral vancomycin

●Recurrent episode – Add oral vancomycin or oral fidaxomicin

The treatment of C. difficile in children is discussed separately. (See "Clostridioides difficile infection in children: Treatment and outcome".)

Positive blood culture pending identification

●Gram-negative bacteria – Change the baseline agent to a carbapenem (eg, meropenem) if a carbapenem was not chosen initially and add an aminoglycoside

●Gram-positive bacteria – Add vancomycin (table 5) (or linezolid if the patient is known to be colonized with or previously infected with vancomycin-resistant gram-positive organisms)

Other indications for combination therapy — Other indications for addition of an agent with additional activity against gram-positive organisms (eg, vancomycin, clindamycin, linezolid) are listed below [1,14]. Clindamycin should be chosen only if the rates of clindamycin resistance in the cancer center and community center are acceptably low. Linezolid should be chosen only if the child is known to be colonized with or has a history of infection with vancomycin-resistant gram-positive organisms; linezolid should be used with caution because it has been associated with neutropenia.

●Evidence of meningitis or new onset of neurologic signs and symptoms suggestive of meningitis – Add vancomycin (table 5) to broad-spectrum antipseudomonal therapy with cefepime or meropenem; if meropenem is used as the broad-spectrum antipseudomonal agent, use the dose for central nervous system infections: 40 mg/kg IV every eight hours (maximum of 2 g per dose).

●Radiographically documented pneumonia – Add vancomycin if MRSA is suspected (table 5).

●Clinically suspected central venous catheter-related infection (eg, chills or rigors with infusion through the catheter and cellulitis around the catheter entry or exit site) – Add vancomycin (table 5) to broad-spectrum antipseudomonal therapy [92]. Alternatives to vancomycin include ceftaroline or daptomycin, although these agents are not used routinely.

●Skin or soft tissue infection at any site – Add vancomycin, clindamycin, or linezolid [24].

●Known colonization with MRSA or penicillin- and cephalosporin-resistant Streptococcus pneumoniae or previous history of infection with penicillin-resistant pneumococci – Add vancomycin, clindamycin, or linezolid.

●Recent intensive chemotherapy associated with a high risk for infection with penicillin-resistant streptococci (eg, viridans streptococci following high-dose cytarabine) – Add vancomycin, clindamycin, or linezolid [14,93-95].

●Prophylaxis with quinolones during afebrile neutropenia – Add vancomycin, clindamycin, or linezolid [1,95-98].

EMPIRIC THERAPY FOR LOW-RISK PATIENTS

Inpatient therapy — Children with fever and chemotherapy-induced neutropenia who are low risk for serious infection or complications and are not candidates for outpatient therapy are hospitalized for IV broad-spectrum monotherapy with cefepime, meropenem, or piperacillin-tazobactam (table 1 and table 2A) [1,42]. (See 'Broad-spectrum monotherapy' above.)

Outpatient therapy — Outpatient management of fever and neutropenia with IV or oral antibiotics may be an option for carefully selected low-risk patients if daily follow-up is ensured [1,42]. Several studies using strict eligibility criteria suggest that outpatient treatment may be safe and appropriate for children at low risk of serious infection [99-105].

When children with fever and chemotherapy-induced neutropenia are treated as outpatients, they should receive the first dose of antimicrobial therapy in a clinical or hospital setting and be observed for ≥4 hours before discharge [1]. They should be readmitted for hemodynamic instability, signs and symptoms of new or worsening infection, or, in the absence these, persistent fever (ie, >48 to 96 hours).

A systematic review of four randomized trials comparing inpatient and outpatient treatment for children with cancer and low-risk febrile neutropenia [18,106-108] found no clear evidence of different rates of treatment failure, mortality, or adverse drug reactions. However, outpatient treatment in these trials consisted of early discharge after surveillance for 24 to 72 hours rather than exclusive outpatient therapy [103].

●Oral therapy for select low-risk patients – Decisions regarding outpatient oral therapy for low-risk children with fever and chemotherapy-induced neutropenia are made on a case by case basis [1,42].

•Criteria – Outpatient oral therapy is an option for children who meet all of the following criteria [1]:

-Able to take and absorb oral antibiotics

-Has a caregiver and a telephone

-Live relatively close to their local medical facility in the event of clinical worsening (eg, within one hour)

-Able to adhere to daily outpatient follow-up

-Patient or caregiver(s) and clinician agree to oral outpatient therapy

-Has not been receiving fluoroquinolone prophylaxis

•Oral regimen – For low-risk children who receive oral outpatient therapy and have not been receiving fluoroquinolone prophylaxis, we suggest oral empiric therapy with ciprofloxacin plus amoxicillin-clavulanate rather than other oral regimens [1,109-111].

Although this regimen has not been well-studied in children, it appears to be as effective as IV therapy in low-risk patients [112]. In two randomized trials, the rate of treatment failure and duration of fever, neutropenia, and antimicrobial therapy were similar in patients treated with oral ciprofloxacin/amoxicillin-clavulanate versus IV ceftriaxone or cefepime [109,111].

There are few data to guide the choice of outpatient regimen for low-risk children with contraindications to ciprofloxacin combined with amoxicillin-clavulanate or low-risk children who have been receiving fluoroquinolone prophylaxis. Consultation with an expert in infectious diseases is suggested to optimize the regimen for individual patients. Hospitalization for IV therapy may be necessary.

ONGOING ANTIMICROBIAL THERAPY — Modification of the empiric regimen may be warranted based upon a variety of clinical scenarios, as discussed in the sections below (algorithm 1) [1,2].

Clinical worsening or instability — Children with cancer and febrile neutropenia who worsen or become clinically unstable at any point during antimicrobial therapy should undergo reevaluation for new or worsening sites of infection [1]. If they have been treated in the outpatient setting, they should be admitted for IV therapy.

Reevaluation includes thorough examination and imaging or repeat imaging for new or worsening sites of infection, as well as culture, biopsy, or drain of sites of worsening infection (assessing for bacterial, viral, and fungal pathogens). Antimicrobial therapy should be reviewed for adequacy of dosing and spectrum. (See 'History and examination' above.)

For children with hemodynamic instability and no identified source of fever, antimicrobial therapy should be broadened (eg, to increase activity against gram-positive, gram-negative, resistant, and anaerobic bacteria). Appropriately broadened antibacterial therapy may be achieved by changing cefepime to meropenem, adding vancomycin (table 5), and increasing gram-negative activity by adding an aminoglycoside, ciprofloxacin, or aztreonam [2].

The addition of antifungal therapy may be warranted for children who remain febrile for ≥96 hours. (See 'Addition of antifungal therapy' below.)

Documented infection — For children with fever, chemotherapy-induced neutropenia, and documented infection at any point during treatment, antimicrobial therapy is modified according to culture and susceptibility results and/or site of infection [1]. Infection may be documented microbiologically (eg, positive culture) or clinically (eg, cellulitis or pneumonia without isolated pathogen).

●Clinical worsening or instability – Additional evaluation and management of children with documented infection who worsen clinically or whose clinical status becomes unstable are discussed above. (See 'Clinical worsening or instability' above.)

●Clinically stable and improved – For those who remain stable and are clinically improved, we tailor the antimicrobial regimen to the infection and continue therapy for 7 to 14 days as appropriate for the infection and until the absolute neutrophil count (ANC) is >500 cells/microL and increasing [1]. Criteria for switching to oral antimicrobial therapy are discussed below. (See 'Switching to oral therapy' below.)

No documented infection — For children with fever, chemotherapy-induced neutropenia, and no documented infection (ie, unexplained fever) after 48 hours, changes to antimicrobial therapy depend upon the child's clinical status (algorithm 1):

●Clinical worsening or instability – Evaluation and management of clinical worsening is discussed above. (See 'Clinical worsening or instability' above.)

●Clinically stable and afebrile – For children with chemotherapy-induced neutropenia and no documented infection who are clinically stable and have been afebrile for ≥24 hours, we continue initial empiric therapy until the ANC is >500 cells/microL. However, we discontinue agents with additional gram-positive activity (eg, vancomycin, clindamycin, linezolid) and/or additional gram-negative activity (eg, aminoglycosides) after 48 hours if they were added to the initial regimen, given the potential harms (eg, nephrotoxicity, colonization or infection with resistant pathogens).

Switching to oral therapy is discussed below. (See 'Switching to oral therapy' below.)

●Clinically stable, persistent fever <96 hours (high- or low-risk) – For children with chemotherapy-induced neutropenia and no documented infection who are clinically stable but remain febrile, we continue initial empiric therapy and daily reevaluation. However, we discontinue agents with additional gram-positive activity (eg, vancomycin, clindamycin, linezolid) and/or additional gram-negative activity (eg, aminoglycosides) after 48 hours if they were added to the initial regimen, given the potential harms (eg, nephrotoxicity, colonization or infection with resistant pathogens). We adjust the antimicrobial regimen if clinical status changes or an infection is identified. (See 'Reevaluation during treatment' above and 'Clinical worsening or instability' above and 'Documented infection' above.)

●Clinically stable, persistent fever for ≥96 hours or recurrent fever (high- or low risk) – For children with chemotherapy-induced neutropenia and no documented infection who are clinically stable with persistent fever for ≥96 hours despite broad-spectrum empiric antibacterial therapy or who have recurrent fever, we continue empiric antimicrobial therapy and daily reevaluation [1]. Additional management depends upon anticipated bone marrow recovery.

•Bone marrow recovery imminent – For clinically stable children with no documented infection, persistent fever for ≥96 hours or recurrent fever, and imminent bone marrow recovery, we modify the antimicrobial regimen only if clinically indicated (eg, examination, laboratory, or imaging studies suggest a new infection).

The duration of antimicrobial therapy is discussed below. (See 'Duration of antimicrobial therapy' below.)

•Bone marrow recovery not imminent – For clinically stable children with no documented infection, persistent fever for ≥96 hours or recurrent fever, and whose bone marrow recovery is not imminent, we generally add antifungal therapy. In addition, we consider noninfectious causes of fever. (See 'Addition of antifungal therapy' below and 'Noninfectious causes of fever' above.)

Addition of antifungal therapy — Clinically occult fungal infection must be considered in children with recurrent fever or persistent fever (ie, ≥96 hours) and neutropenia despite empiric antibacterial therapy [1,42,113]. Invasive fungal infection is associated with decreased survival [114].

●Indications – We initiate empiric antifungal therapy in children who have recurrent or persistent fever for ≥96 hours after initiation of empiric broad-spectrum antimicrobial therapy, no identified source of fever, and ANC that is not increasing [1]. In a small, randomized trial, the addition of amphotericin B to antibacterial therapy in children with fever and neutropenia ≥7 days was effective in controlling clinically occult fungal infections and preventing fungal superinfections [11].

Other experts suggest that empiric antifungal therapy may be withheld from such children who are not at increased risk for invasive fungal disease [42] or who do not meet certain prespecified criteria (eg, pulmonary infiltrates, skin lesions suggestive of invasive fungal disease, positive mycology studies) [115,116].

The risk of invasive fungal infection is increased in children with one or more of the following [32,42,113,114,117-121]:

•Acute myeloid leukemia

•Relapsed or high-risk acute lymphoblastic leukemia (see "Prognostic factors and risk group stratification for acute lymphoblastic leukemia/lymphoblastic lymphoma in children and adolescents")

•Prolonged neutropenia (>10 days)

•Highly myelosuppressive chemotherapy

•High-dose glucocorticoid therapy (usually defined as prednisone ≥20 mg per day, or ≥2 mg/kg per day for patients weighing <10 kg) for ≥14 days (or equivalent [122])

•Allogeneic hematopoietic cell transplant recipient

•Graft-versus-host disease

Support for withholding antifungal therapy for certain children is provided by a trial in which children with persistent high-risk febrile neutropenia were randomly assigned to empiric antifungal therapy or antifungal therapy if, at any time during follow-up, they met prespecified clinical, radiographic, or laboratory criteria (pre-emptive antifungal therapy) [116]. Although only 42 percent of the pre-emptive group received antifungal therapy and the duration of antifungal therapy was shorter (6 versus 11 days), no differences were detected in the rates of invasive fungal disease (12 percent) or mortality related to invasive fungal disease (3 percent).

●Evaluation for fungal infection – The suggested evaluation for fungal infection, performed concomitantly with initiation of antifungal therapy, includes [1,42]:

•Ultrasonography and/or computed tomography (CT) of the liver and spleen

•CT of the chest

•CT of the sinuses (in patients with localizing signs or symptoms)

•Biopsy of any suspicious lesions

Laboratory-based diagnostic tests for fungal infection include galactomannan, beta-D glucan, and polymerase chain reaction (PCR) assays. These tests have variable sensitivity, specificity, and predictive values [42,123]. None is a reliable predictor of fungal infection in children. Galactomannan is specific for Aspergillus molds; results may be negative in children with non-Aspergillus fungal infection. Clinical experience with beta-D glucan as a fungal biomarker in children is limited. PCR testing for fungal infection has poor standardization and has poor positive and negative predictive value.

●Antifungal regimen – Appropriate antifungal agents include (one of the following) [1,42]:

•Amphotericin B

We prefer lipid formulations of amphotericin B (liposomal amphotericin or amphotericin B complex) to amphotericin B deoxycholate. A meta-analysis of four randomized trials [124-127] found lipid formulations of amphotericin B to have similar rates of breakthrough fungal infections but less nephrotoxicity and infusion-related toxicity (liposomal formulations) than conventional amphotericin B [128]. If lipid formulations of amphotericin B are not available, amphotericin B deoxycholate is acceptable [1,42].

•Echinocandins (eg, caspofungin, micafungin, anidulafungin)

•Triazole derivatives (eg, voriconazole, posaconazole)

The duration of antifungal therapy is discussed below. (See 'Total duration' below.)

There is little evidence to guide the choice of antifungal agents for children with febrile neutropenia. Factors that influence the choice of empiric antifungal therapy include antifungal prophylaxis and renal or hepatic dysfunction. For empiric antifungal therapy in children who have been receiving antifungal prophylaxis, we choose an antifungal agent from a different class than the agent used for prophylactic therapy (eg, if they have been receiving an echinocandin, we choose amphotericin B or a triazole). We generally prefer lipid formulations of amphotericin for children with renal dysfunction and anidulafungin for children with hepatic and/or renal dysfunction.

Data regarding the efficacy and safety of echinocandins and triazole derivatives are limited. In a multicenter randomized trial comparing amphotericin B and caspofungin in 82 children with persistent fever and neutropenia, rates of response and adverse effects (including infusion-related toxicity and nephrotoxicity) were similar [129]. However, the trial was not powered to detect statistically significant differences. A systematic review comparing amphotericin B with voriconazole in cancer patients with neutropenia concluded that amphotericin B is more effective for empiric therapy [130].

DURATION OF ANTIMICROBIAL THERAPY

Total duration — The total duration of empiric antimicrobial therapy depends upon the clinical response and myeloid recovery.

●Antibacterial therapy – We generally continue antimicrobial therapy until blood cultures have been negative for ≥48 hours, the patient has been afebrile for ≥24 hours, and the ANC is >500 cells/microL and increasing [1,42,131].

However, shorter durations of therapy may be appropriate for select low-risk patients [14,56,132]. The 2017 International Pediatric Fever and Neutropenia Guidelines permit low-risk patients who have received IV antibiotics for ≥72 hours, have been afebrile for ≥24 hours, and have no identified source of infection to be discharged from the hospital without antibiotics regardless of marrow recovery if follow-up is ensured [42], as supported by a small retrospective study [133].

●Antifungal therapy – We usually continue empiric antifungal therapy until resolution of neutropenia in the absence of evidence of invasive fungal infection [1,42]. However, there is little evidence to guide this decision.

Switching to oral therapy — For children with adequate gastrointestinal absorption, decisions regarding the switch from IV to oral therapy (and hospital discharge) are individualized according to clinical status. Children who have recurrent fever within 48 hours of initiation of oral therapy should be readmitted to the hospital and managed as high-risk patients [2].

The following criteria for switching from IV to oral therapy have been safely used in observational studies [12,18,134-136]:

●Afebrile ≥24 hours

●Clinically stable and well-appearing

●Negative blood cultures

●Local infection (if present) under control

●ANC >100 cell/microL and evidence of bone marrow recovery (ie, sustained increase in platelet count and ANC or absolute phagocyte count [ANC plus absolute monocyte count])

●≥48 hours of IV therapy

For children with documented infection, the oral regimen is determined by the culture and susceptibility results and/or site of infection. For children with no documented infection who are switched to oral therapy, we use one of the following:

●Cefixime

●Cefpodoxime

●Ciprofloxacin (with or without amoxicillin-clavulanate)

●Levofloxacin [136]

COLONY STIMULATING FACTORS — Whether high-risk patients and/or those with predicted prolonged courses of neutropenia (ie, more than seven days) may benefit from the use of granulocyte colony stimulating factors (G-CSF) remains unanswered. Nonetheless, interventional G-CSF may be warranted for certain children with complicated episodes of fever and neutropenia [137-139]. One of the authors of this topic review (PF) routinely provides CSF for children with solid tumors who have fever and neutropenia; the other author of this topic review (NA) provides CSF predominantly in children with chemotherapy-induced neutropenia and decreasing absolute neutrophil count or anticipated prolonged neutropenia.

A substantial benefit of initiating CSFs during febrile neutropenia has not been firmly established. In a meta-analysis of three randomized trials, therapeutic CSF (G-CSF or granulocyte monocyte CSF [GM-CSF]) reduced duration of hospitalization by 1.42 days (95% CI 0.62-2.22) compared with placebo [74]. In individual randomized trials, G-CSF and GM-CSF also modestly reduced days of intravenous antibiotics or median time to resolution of neutropenia [140-142].

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Neutropenic fever in children with cancer".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or email these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient education" and the keyword[s] of interest.)

●Basics topics (see "Patient education: Neutropenia and fever in people being treated for cancer (The Basics)")

SUMMARY AND RECOMMENDATIONS

●Definitions – We use the following definitions in children with cancer or hematopoietic cell transplant (see 'Definitions' above):

•Neutropenia – Absolute neutrophil count (ANC (calculator 1)) <500 cells/microL or an ANC that is expected to decrease to <500 cells/microL during the next 48 hours

•Fever – Any of the following:

-A single oral temperature ≥38.3°C (101°F)

-A temperature ≥38°C (100.4°F) for longer than one hour

-Two elevations >38°C (100.4°F) during a 12-hour period

Rectal temperature measurements should be avoided in neutropenic patients.

●Cause of fever – Bloodstream infections are the most common cause of fever in children with chemotherapy-induced neutropenia and may be caused by gram-positive organisms, gram-negative organisms, or fungi. (See 'Infectious causes of fever' above.)

●Evaluation – Rapid and thorough evaluation is crucial for prompt administration of appropriate empiric antimicrobial therapy.

•The history is focused on infection (eg, symptoms, exposures, recent infection or colonization, antimicrobial prophylaxis, central venous catheters and other factors that predispose to infection). (See 'History and examination' above.)

•The physical examination should focus on the sites most commonly infected:

-Skin and mucous membranes – Skin folds, nail beds, central catheter sites, perineum (including rectal area), sites of recent procedures

-Sinuses

-Oropharynx, including the gums

-Lungs

-Abdomen

The review of systems and physical examination should be repeated at least daily. (See 'History and examination' above and 'Reevaluation during treatment' above.)

•Routine laboratory evaluation includes (see 'Routine laboratory and imaging studies' above):

-Complete blood count with differential and platelet count

-Electrolytes, creatinine, and blood urea nitrogen

-Liver transaminases and total bilirubin

-Blood cultures

Additional studies (eg, other cultures, molecular diagnostic assays, imaging) are obtained as clinically indicated (see 'Additional studies as indicated' above)

●Antimicrobial therapy

•Broad-spectrum empiric therapy should be initiated as soon as possible after blood cultures are obtained and always within 60 minutes of triage (table 2A-B). (See 'Prompt initiation of antimicrobial therapy' above.)

•For high-risk children (table 1) with fever and chemotherapy-induced neutropenia in whom there is no suspicion of resistant infection, we recommend initial broad-spectrum monotherapy with an antipseudomonal beta-lactam agent (eg, cefepime, meropenem, piperacillin-tazobactam) rather than combination therapy (Grade 1A). Combination therapy is reserved for children with clear indications for additional activity against gram-positive, gram-negative, or anaerobic bacteria (table 2B). (See 'Empiric therapy for high-risk patients' above and 'Limited indications for combination therapy' above.)

•For low-risk children (table 1) with fever and chemotherapy-induced neutropenia, the empiric antimicrobial regimen is influenced by the treatment setting (table 2A). Those who are treated as outpatients should receive the first dose of antimicrobial therapy in the hospital or clinic setting and be observed for ≥4 hours before discharge. (See 'Empiric therapy for low-risk patients' above.)

•Ongoing antimicrobial therapy is modified if there is clinical worsening or instability or documented clinical or microbiologic infection. For patients who remain stable and do not have a documented infection, continued antimicrobial therapy is modified according to resolution or persistence of fever and resolution or persistence of neutropenia (algorithm 1). (See 'Ongoing antimicrobial therapy' above.)

•We initiate empiric antifungal therapy in children who have recurrent or persistent fever for ≥96 hours after initiation of empiric broad-spectrum antimicrobial therapy, no identified source of fever, and ANC that is not increasing. Options for antifungal therapy in children include amphotericin B, echinocandins (eg, caspofungin, micafungin, anidulafungin), or a triazole derivative (eg, voriconazole, posaconazole). (See 'Addition of antifungal therapy' above.)

•We generally continue antimicrobial therapy until blood cultures have been negative for ≥48 hours, the patient has been afebrile for ≥24 hours, and the ANC is >500 cells/microL and increasing. (See 'Duration of antimicrobial therapy' above.)