Neutropenic Fever

Neutropenic fever (NF), also called febrile neutropenia, is the simultaneous occurrence of fever and severe neutropenia in a patient receiving myelosuppressive cancer therapy. It is a common oncologic emergency requiring immediate recognition and treatment.

Defining the Syndrome

Fever is defined as a single oral temperature of ≥38.3°C (101°F) or a temperature of ≥38.0°C (100.4°F) sustained over a 1-hour period ^[3]D5. Neutropenia is defined as an absolute neutrophil count (ANC) <500 cells/μL or an ANC <1000 cells/μL with an expected decline to ≤500 cells/μL within 48 hours ^[3]D5. The combination of these two criteria, occurring after recent chemotherapy, targeted therapy, or hematopoietic stem cell transplantation (HSCT), constitutes neutropenic fever.

Grading of Neutropenia

The severity of neutropenia is graded according to the Common Terminology Criteria for Adverse Events (CTCAE), which informs the risk of developing NF.

Most clinical trials and guidelines, including those cited in this article, define NF using Grade 4 neutropenia (ANC <500). The duration of ANC <500, especially if prolonged (>7 days), further escalates infection risk ^[1]A1b.

Classification by Risk

Patients with NF are stratified into low-risk and high-risk categories to guide the intensity of management , including the need for hospitalization and intravenous antibiotics . The most widely used tool is the Multinational Association for Supportive Care in Cancer (MASCC) risk index:

• Low-risk: MASCC score ≥21, associated with a low probability of serious complications and mortality ^[6]B3b.
• High-risk: MASCC score <21, indicates a higher risk of complications, ICU admission, and death ^[6]B3b.

In a large retrospective cohort of solid tumor patients, the MASCC score had a sensitivity of 83.5% and specificity of 57.3% for predicting an uncomplicated course, emphasizing that clinical judgment remains essential ^[6]B3b. The Clinical Index of Stable Febrile Neutropenia (CISNE) score of 0 also identifies low-risk patients suitable for outpatient management ^[3]D5.

Clinical Significance

Neutropenic fever carries a crude mortality rate of 3% to 18% ^[3]D5. It is a leading cause of hospitalization and unplanned chemotherapy dose reductions or delays. Prompt recognition, risk stratification, and initiation of empiric broad-spectrum antibiotics are the cornerstones of management. The epidemiology and risk factors that predispose patients to this syndrome are discussed next.

Pearl: Risk stratification using the MASCC score (low-risk ≥21) guides outpatient vs inpatient management ^[6]B3b.

Defined by temperature and neutrophil thresholds, the clinical burden of neutropenic fever varies substantially across populations and treatment regimens. Among the most informative contemporary data, patients receiving neoadjuvant docetaxel , cisplatin , and 5‑fluorouracil (DCF) for locally advanced esophageal cancer experienced a first‑cycle febrile neutropenia (FN) incidence of 26% (41 of 156 patients) in a retrospective cohort, with the nadir occurring on median day 8 (range 5-12) ^[16]B3b. In the phase III JCOG1109 trial, the FN rate with DCF was 16.3%, compared with 1.0% for cisplatin‑5‑FU and 4.7% for chemoradiotherapy ^[16]B3b. These figures illustrate the regimen‑dependent nature of risk: any chemotherapy associated with a ≥20% FN incidence is considered high‑risk by major guidelines.

Risk Factors for Febrile Neutropenia

Risk stratification relies on both treatment‑related and patient‑related factors. The strongest modifiable factors are prophylaxis use itself. In multivariate analysis of the DCF cohort, prophylactic antibiotics reduced FN incidence from 40% to 17% (adjusted OR 0.34, 95% CI 0.16-0.73; p = 0.006), and prophylactic granulocyte colony‑stimulating factor (G‑CSF) reduced it from 35% to 6% (adjusted OR 0.14, 95% CI 0.04-0.47; p = 0.002) ^[16]B3b. The timing of G‑CSF also matters: in the same cohort, FN occurred in 35% of patients without G‑CSF, 13% with G‑CSF started on days 6‑8, and 0% with G‑CSF started on days 3‑4 (p < 0.001 for early vs. none) ^[16]B3b.

Patient‑related factors that impair antibiotic efficacy, and therefore increase the risk of FN complications, include older age and heavier body weight. In a pediatric FN trial (93 patients, 405 episodes), treatment success rates were significantly lower in patients ≥16 years (82.0% vs. 94.0%; p < 0.001) and ≥52 kg (82.7% vs. 94.0%; p = 0.002) ^[15]A1b. Bacteremia at presentation also predicted failure (success 52.0% vs. 76.3%; p = 0.0147) ^[15]A1b.

Nutritional status is a less appreciated but important risk factor. In 99 patients with limited‑stage small‑cell lung cancer undergoing chemoradiotherapy, ≥5% weight loss during treatment was independently associated with FN (OR 4.49, 95% CI 1.47-13.80; p = 0.009), and the absolute FN incidence was 80.0% in those with ≥5% weight loss versus 43.2% in those without (p = 0.002) ^[14]B3b.

Risk Factor Summary

Pearl: The two most actionable risk factors for febrile neutropenia are the absence of prophylactic G‑CSF and antibiotics; their combined effect can reduce FN incidence from ~40% to as low as 0% in high‑risk regimens like DCF, making prophylaxis the single most important modifiable driver of risk ^[16]B3b.

The risk factors outlined above converge on a single pathophysiologic defect: a profound deficiency of neutrophils, the primary cellular defense against bacterial and fungal pathogens. Neutropenia not only impairs phagocytosis and intracellular killing but also disrupts mucosal barriers, particularly in the gastrointestinal tract, allowing translocation of colonizing organisms into the bloodstream. Chemotherapy-induced mucositis, central venous catheters, and skin breakdown provide additional portals of entry.

Infectious versus Non-Infectious Causes

Febrile neutropenia is not always due to infection. Non-infectious causes include chemotherapy-induced mucositis, tumor fever, transfusion-related fever, drug fever, and graft-versus-host disease ^[18]A1a. In these cases, antibiotics may be unnecessary and contribute to resistance and toxicity. However, distinguishing these from true infection at presentation is often impossible, necessitating empiric broad-spectrum coverage.

Epidemiology of Documented Infections

Only 30% of febrile neutropenia episodes yield a documented bloodstream infection (BSI), and 40% have a clinically or microbiologically documented infection overall ^[21]D5. The majority of febrile episodes are classified as fever of unknown origin, though many of these likely represent transient or subclinical infections.

Organism Distribution and Resistance Patterns

Among BSIs, Gram-negative bacteria predominate, accounting for 56.7% of isolates in a recent cohort of patients with hematologic malignancies ^[24]B3b. The most prevalent Gram-negative organism is Escherichia coli (31%), followed by other Enterobacterales, Pseudomonas aeruginosa , and Klebsiella species. A striking 49.3% of Gram-negative isolates produce extended-spectrum beta-lactamases (ESBL), and 20.2% demonstrate carbapenem resistance ^[24]B3b. Gram-positive organisms, including coagulase-negative staphylococci, viridans group streptococci, and Staphylococcus aureus , are also important but now less common than Gram-negative bacilli in many centers.

Sources of Infection

The gastrointestinal tract is the primary reservoir for Gram-negative bacilli, and gut decontamination strategies (e.g., fluoroquinolone prophylaxis) have been shown to reduce febrile neutropenia and BSI ^[17]B2a. The skin and central venous catheters are common sources for Gram-positive cocci, and decolonization with mupirocin nasal drops can reduce coagulase-negative staphylococci and methicillin-resistant S. aureus carriage ^[20]A1b.

Clinical Implications of Resistance

Carbapenem resistance in Gram-negative BSIs is independently associated with both 30-day ICU admission and 30-day all-cause mortality ^[24]B3b. This finding underscores the importance of local surveillance data in guiding empiric antibiotic selection, as discussed in the Empiric Antibiotic Therapy section.

Pearl: In febrile neutropenia, Gram-negative bacilli now outnumber Gram-positive organisms in most centers, and carbapenem resistance, not ESBL production alone, is the key driver of adverse outcomes, making local resistance data essential for appropriate empiric coverage.

In the absence of a functioning neutrophil pool, the classic signs of inflammation, purulence, rubor, tumor, and calor, are muted or absent. Fever is the cardinal and often the only objective finding. Fever is defined as a single oral temperature ≥38.3°C or ≥38.0°C sustained for ≥1 hour. Chills, rigors, or diaphoresis may be present, but localizing symptoms are frequently absent. The absence of typical infectious complaints should never delay the initiation of empiric antibiotics .

History

A focused history must capture the chemotherapy regimen, the date of the last cycle (the nadir period typically occurs 7-14 days after myelosuppressive therapy), and the presence of a central venous catheter . Recent antibiotic use, prophylactic antimicrobials, prior hospitalizations, travel, and sick contacts are also critical. Obesity is a recognized risk factor: in the GAIN study, 15% of obese patients receiving full-dose chemotherapy experienced febrile neutropenia versus 6% of those with adjusted dosing (P = 0.003) ^[26]B3b. Patients should be asked about perianal pain, dysphagia, odynophagia, cough, dyspnea, dysuria, and skin lesions.

Physical Examination

A systematic but targeted examination is essential. Key areas include:

• Oropharynx: inspect for mucositis, ulcers, or thrush.
• Lungs: auscultate for crackles or wheezes; radiography may be normal initially.
• Heart: new murmur may suggest endocarditis.
• Abdomen: tenderness, especially right upper quadrant or perianal.
• Perianal region: gentle inspection for induration, fissures, or abscess (avoid digital rectal exam in severe neutropenia due to risk of bacteremia).
• Skin: examine all catheter exit sites, the entire integument for cellulitis or echymoses.
• Neurologic: altered mental status can indicate sepsis or central nervous system infection.

In elderly patients or those receiving corticosteroids, the fever response may be blunted, and hypothermia (temperature <36.0°C) can be an equivalent sign. The initial assessment also includes a review of the complete blood count with differential to confirm the absolute neutrophil count and to document the severity of neutropenia.

Serum biomarkers such as C-reactive protein and procalcitonin have been investigated for risk stratification at presentation, but their role in initial assessment remains limited due to marked heterogeneity in published data ^[25]B2a. The clinical picture, not the biomarker level, should drive the decision to start antibiotics.

Pearl: The single most important clinical rule: in a neutropenic patient, any fever is a medical emergency; do not wait for localizing signs before initiating empiric antibiotics. The history and physical examination described here lay the groundwork for the diagnostic workup, which is initiated concurrently with the first dose of broad-spectrum antimicrobial therapy.

Once the clinical suspicion of neutropenic fever is raised, the diagnostic workup proceeds in parallel with the initiation of empiric antibiotics . The goal is to identify the source of infection, guide targeted therapy, and stratify risk. The approach is systematic and time-sensitive, with cultures and imaging obtained before or shortly after antibiotic administration.

Blood Cultures

Blood cultures are the cornerstone of microbiologic diagnosis. Two sets should be obtained from a peripheral venipuncture and from each lumen of a central venous catheter if present. The presence of a recognized pathogen in at least one blood culture defines bacteremia ^[22]B2a (as referenced in ^[32]B2a). In a prospective pediatric trial, positive blood cultures were found in 25 of 405 episodes (6.2%), and treatment success was significantly lower in patients with positive cultures (52.0% vs 76.3%; P = 0.0147) ^[15]A1b. Organisms isolated included viridans group streptococci, Pseudomonas aeruginosa , and coagulase-negative staphylococci, among others ^[15]A1b. Blood cultures should be drawn before the first dose of antibiotics if this can be done without delaying therapy by more than 30 minutes. Delays beyond 60 minutes have been associated with worse outcomes in patients with severe disease at presentation, though the relationship is confounded by triage bias ^[32]B2a.

Other Cultures

• Urine culture: Obtain a clean-catch or catheterized specimen before antibiotics. Asymptomatic bacteriuria is common in neutropenic patients and may be a source of fever.
• Stool culture: Indicated when diarrhea is present, especially if Clostridioides difficile infection is suspected (although fluoroquinolone prophylaxis does not appear to increase CDI risk ^[17]B2a).
• Sputum culture and bronchoalveolar lavage: Reserved for patients with respiratory symptoms or abnormal chest imaging.
• Cultures from indwelling devices: If a catheter exit site shows signs of infection, swab cultures should be obtained.

Chest Imaging

A chest radiograph (posteroanterior and lateral) is recommended for all patients with neutropenic fever, even in the absence of respiratory symptoms. Pulmonary infiltrates, including those from opportunistic infections, may be subtle. Repeat imaging is indicated if the fever persists or respiratory symptoms develop.

Computed Tomography (CT) Scans

CT of the chest, abdomen, and pelvis is indicated when:

• Fever persists beyond 72 hours despite broad-spectrum antibiotics.
• There is localizing pain or organomegaly.
• There is concern for a fungal infection (e.g., hepatosplenic candidiasis, pulmonary aspergillosis).
• Blood cultures are negative and the source remains unclear.

CT has higher sensitivity than chest radiography for detecting invasive pulmonary aspergillosis, especially the halo sign. In pediatric patients, CT may be deferred if the patient is stable and the chest radiograph is normal, but it should be performed early in high-risk patients with prolonged neutropenia.

Biomarkers

• C-reactive protein (CRP): Routinely measured in many centers. A rising CRP suggests bacterial infection, but it is not specific. In the BEATLE trial, CRP levels were used to assess response, but the trial did not find a significant difference in treatment success between extended and intermittent infusion of β-lactams ^[33]A1b.
• Procalcitonin: More specific for bacterial infection than CRP. Levels >0.5 ng/mL are suggestive of a bacterial etiology. However, procalcitonin has not been validated as a sole guide for antibiotic initiation or discontinuation in neutropenic fever.
• Presepsin: A novel biomarker that may be superior to procalcitonin in some settings, but data in neutropenic fever are limited ^[15]A1b.

Diagnostic Algorithm

Step 1: Upon presentation, obtain blood cultures (two sets), urine culture, and a chest radiograph. Step 2: Assess the patient for severity, hypotension, altered mental status, tachypnea, or signs of organ dysfunction indicate severe disease. Step 3: Start empiric broad-spectrum antibiotics immediately in severe disease; otherwise, start within 60 minutes ^[32]B2a. Step 4: Re-evaluate at 48-72 hours. If the patient is afebrile and cultures are negative, de-escalate to a narrower regimen. If fever persists, repeat cultures, obtain CT imaging, and consider biomarkers to guide therapy modification.

Pearl: The diagnostic workup is most valuable when cultures are obtained before antibiotics, but the presence of positive blood cultures identifies a subgroup at highest risk of treatment failure (success rate 52% vs 76%), mandating close follow-up and rapid escalation if clinical response is inadequate ^[15]A1b.

Following the diagnostic workup, the next critical step is to stratify the patient's risk of complications to guide the intensity of monitoring, the need for hospitalization, and the choice of empiric antibiotics . The Multinational Association for Supportive Care in Cancer (MASCC) score is the most extensively validated tool for this purpose, though its performance remains imperfect, and clinical judgment is indispensable ^{[6]B3b[39]B3b[41]B3b}.

MASCC Score

The MASCC score classifies patients as low-risk (≥21) or high-risk (<21) based on clinical variables at presentation. In a cohort of 329 patients with solid tumors, the score demonstrated a sensitivity of 83.5% (95% CI 77.8-88.2%) and a specificity of 57.3% (95% CI 47.8-66.4%) for predicting an uncomplicated clinical course ^[6]B3b. Low-risk patients had significantly lower rates of ICU admission (0.4% vs. 32.7%, p<0.001) and mortality (0.9% vs. 16.8%, p<0.001), with a median hospitalization of 4 days [IQR 3-6] compared to 6 days [IQR 4-10] in high-risk patients ^[6]B3b. Despite the moderate specificity, using the MASCC score to guide outpatient eligibility could have prevented 486 hospitalization days across 161 patients in that study, with 80.7% experiencing no complications ^[6]B3b.

In a separate inpatient cohort, only 2% of patients had a documented MASCC score, yet when calculated retroactively, a high-risk score (<21) was associated with a 33% higher likelihood of requiring intermediate or ICU level of care (95% CI 23-44%) and a 19% higher risk of in-hospital death (95% CI 10-27%) compared to low-risk patients ^[39]B3b. These data underscore the score's potential utility even in hospitalized patients, where it is underutilized.

Cut-Off for Mortality Prediction

A more recent analysis of 354 hospitalized FN patients proposed a modified MASCC threshold for predicting all-cause mortality. A score below 17 yielded a sensitivity of 83.6% and a specificity of 94.1% for mortality, with 30-day, 60-day, and 90-day mortality rates of 25.1%, 30.2%, and 32.7%, respectively ^[41]B3b. The MASCC score was the only independent risk factor for mortality in multivariate Cox regression ^[41]B3b. This lower cut-off may better identify patients who require intensive monitoring and early goals-of-care discussions.

Biomarker-Based Stratification

Serum biomarkers may refine risk stratification, particularly in pediatric populations. In a cohort of 79 FN episodes in children with cancer, the combination of procalcitonin (PCT) ≥0.425 ng/ml and interleukin-10 (IL-10) ≥4.37 pg/ml at presentation achieved a sensitivity of 100% (95% CI 68.8-100%) and specificity of 89% (95% CI 80.0-95.0%) for predicting bacteraemia, correctly identifying all 8 episodes ^[37]B2b. CRP and IL-6 were not significantly associated with bacteraemia in this study ^[37]B2b. PCT and IL-10 are not yet incorporated into standard risk tools, but they hold promise for identifying truly low-risk patients who may be candidates for oral antibiotic therapy and early discharge.

Clinical Judgment and Criteria for Low-Risk Outpatient Management

No single risk tool is perfectly accurate. The moderate specificity of the MASCC score (57.3%) means that approximately 4 in 10 patients classified as low-risk will still develop complications ^[6]B3b. Therefore, clinical judgment remains essential. Patients considered for outpatient management should meet all of the following criteria:

• MASCC score ≥21 (or, in selected settings, a clinically low-risk profile)
• No hemodynamic instability or signs of sepsis
• No focal infection requiring parenteral therapy (e.g., pneumonia, intra-abdominal infection)
• Adequate oral intake and ability to adhere to follow-up
• Reliable social support and access to 24-hour medical care

In a randomized trial of low-risk FN in children with cancer, oral levofloxacin was safe and cost-effective compared to amoxicillin-clavulanate / ciprofloxacin for home-based management, reinforcing that outpatient therapy is feasible when strict eligibility criteria are met ^[38]A1b. The decision to pursue outpatient management should be shared with the patient and family, with clear instructions to return if fever recurs or clinical status deteriorates.

After risk stratification is complete, the next step is to select the appropriate empiric antibiotic regimen, which is tailored to the patient's risk category and local epidemiology .

Pearl: Combined procalcitonin (≥0.425 ng/ml) and IL-10 (≥4.37 pg/ml) at presentation can predict bacteraemia with 100% sensitivity and 89% specificity in children with febrile neutropenia, complementing clinical risk assessment.

Once risk stratification is complete, antibiotic selection follows a structured approach based on the patient’s risk category, local resistance patterns, and the presence of specific clinical syndromes. The goal is to initiate broad-spectrum antipseudomonal coverage within 60 minutes of presentation in high-risk patients, although recent evidence challenges the use of strict time-to-antibiotic thresholds as quality measures ^[32]B2a.

Monotherapy versus Dual Therapy

For high-risk febrile neutropenia, guidelines recommend empiric monotherapy with an antipseudomonal β-lactam: cefepime , piperacillin/tazobactam , or a carbapenem such as meropenem ^[44]A1c. A systematic review and meta-analysis of five pediatric RCTs (470 episodes) found no significant difference in treatment success between piperacillin/tazobactam and cefepime (RR 1.02; 95% CI 0.89-1.18; P = 0.76) ^[31]A1a. In a prospective pediatric trial (405 episodes), 2-hour infusions of meropenem 120 mg/kg/day (max 3 g/day) and piperacillin/tazobactam 360 mg/kg/day (max 18 g/day) achieved similar success rates (74.9% vs 74.8%) ^[15]A1b. Dual therapy, adding an aminoglycoside (e.g., gentamicin or amikacin ) to an antipseudomonal β-lactam, is reserved for patients with septic shock, suspected resistant gram-negative infection, or in institutions with high rates of multidrug-resistant organisms ^[44]A1c. No randomized trial has demonstrated superiority of dual therapy over monotherapy in unselected high-risk populations.

Indications for Vancomycin

Empiric addition of vancomycin is indicated when there is clinical suspicion of a catheter-related bloodstream infection, skin or soft-tissue infection, severe pneumonia, or in centers with high prevalence of methicillin-resistant Staphylococcus aureus (MRSA) ^[44]A1c. Vancomycin should be discontinued after 48-72 hours if cultures are negative and no resistant gram-positive infection is identified.

Oral Options for Low-Risk Patients

In carefully selected low-risk patients (e.g., those who are hemodynamically stable, have no focus of infection, and adequate social support), oral antibiotic therapy is safe and effective. A randomized trial in children with low-risk febrile neutropenia found that oral levofloxacin was a cost-saving strategy compared with amoxicillin-clavulanate plus ciprofloxacin , with similar outcomes over a 7-day horizon ^[38]A1b. Guidelines recommend oral ciprofloxacin plus amoxicillin -clavulanate as an alternative for low-risk adults ^[44]A1c.

Duration and De-escalation

Historically, broad-spectrum antibiotics were continued until neutrophil recovery (ANC ≥500 cells/μL). However, a meta-analysis of 10 studies (1884 patients) demonstrated that early de-escalation (within 2-5 days) significantly reduced mortality (OR 0.20; 95% CI 0.06-0.69) without increasing infection-related ICU admissions, bacteremia, recurrent fever, or Clostridioides difficile infection ^[18]A1a. Subgroup analysis showed benefit particularly in patients >55 years (OR 0.42; 95% CI 0.18-0.98) ^[18]A1a. Guidelines from the European Conference on Infections in Leukemia (ECIL) recommend stopping empiric antibiotics after 72 hours in hemodynamically stable patients who are afebrile for ≥48 hours, regardless of ANC ^[18]A1a.

Extended Infusion Considerations

Extended infusion (3-4 hours) of β-lactams improves pharmacokinetic/pharmacodynamic target attainment (100% ƒT>MIC) compared with standard 30-minute infusions, as shown by Monte Carlo simulations ^[34]D5. However, the BEATLE randomized trial (150 patients) found no improvement in day-5 treatment success with extended versus intermittent infusion (50.6% vs 63.0%; risk difference -12.4%; 95% CI -29.4 to 4.7; P = 0.17) ^[33]A1b. Extended infusion may still be considered for patients with infections caused by less susceptible organisms or in septic shock ^[33]A1b.

Special Considerations: Anaerobic Coverage in Allogeneic HSCT

In patients undergoing allogeneic hematopoietic stem cell transplantation (HSCT), antibiotics with potent anaerobic coverage (e.g., piperacillin/tazobactam, meropenem) may increase the risk of acute graft-versus-host disease (GVHD) compared with agents having limited anaerobic activity (RR 1.33; 95% CI 1.17-1.51) ^[43]B2a. In this population, choosing a narrow-spectrum β-lactam with limited anaerobic coverage may be prudent when clinical suspicion for anaerobic infection is low.

Controversies and Guideline Disagreement

Dosing Table

Pearl: For high-risk febrile neutropenia, initiate empiric monotherapy with an antipseudomonal β-lactam (cefepime, piperacillin/tazobactam, or meropenem) within 60 minutes; in hemodynamically stable patients who are afebrile for ≥48 hours, consider early de-escalation or discontinuation regardless of neutrophil count, as this strategy is associated with reduced mortality without increased adverse events.

Once empiric broad-spectrum antibiotics are underway, the next critical step is tailoring therapy based on clinical response, culture results, and risk stratification. The goal is to de-escalate or narrow coverage as soon as it is safe, reducing unnecessary antimicrobial exposure, toxicity, and selection pressure.

Step 1: De-escalation When Cultures Are Negative and Patient Is Stable

If blood cultures remain sterile after 48-72 hours and the patient is hemodynamically stable with resolved fever, de-escalation should begin ^[18]A1a. A 2025 meta-analysis of ten studies (867 patients in early de-escalation groups) found that early de-escalation significantly reduced mortality risk (OR 0.20, 95% CI 0.06-0.69) ^[18]A1a. The benefit was greatest when de-escalation occurred within 3 days of starting therapy (OR 0.14, 95% CI 0.03-0.66) ^[18]A1a. There was no significant increase in infection-related ICU admissions, bacteremia, recurrent fever, or Clostridium difficile infection ^[18]A1a.

De-escalation options include:

• Switching from an antipseudomonal β-lactam (e.g., cefepime , piperacillin-tazobactam , meropenem ) to a narrower-spectrum β-lactam (e.g., ceftriaxone ) or to prophylactic fluoroquinolones (e.g., levofloxacin ).
• Discontinuing combination therapy (e.g., stopping an aminoglycoside or glycopeptide) if monotherapy was initiated.
• Stopping all antibiotics if no infection is documented and the patient is clinically well.

Figure 1: De-escalation algorithm for high-risk febrile neutropenia. Adapted from ^[18]A1a and ^[45]A1b.

Step 2: Duration of Therapy - Beyond Neutrophil Recovery

The traditional practice of continuing antibiotics until the absolute neutrophil count (ANC) exceeds 500 cells/μL is no longer mandatory. The How Long study (2017) randomized 157 high-risk patients to discontinue empirical antimicrobial therapy (EAT) after 72 hours of apyrexia and clinical recovery regardless of ANC, versus continuing until ANC ≥ 500 cells/μL ^[45]A1b. The clinical approach resulted in significantly more EAT-free days (mean 16.1 vs 13.6, p=0.026) without excess mortality (1 vs 3 deaths) ^[45]A1b. The authors concluded that this strategy is safe and reduces unnecessary antimicrobial exposure ^[45]A1b.

For patients with a documented infection, the duration should be guided by the infection source (e.g., 7-14 days for bloodstream infection, 10-14 days for pneumonia).

Step 3: Switching to Oral Therapy in Low-Risk Patients

Low-risk patients (e.g., MASCC score ≥ 21, clinically stable, no focus of infection) can be switched to oral antibiotics early. A systematic review of pediatric low-risk febrile neutropenia (37 studies, 3205 episodes) found no significant difference in treatment failure between oral and intravenous therapy (OR 1.05, 95% CI 0.74-1.48) ^[47]B2a. The estimated failure rate with oral therapy was 10.5% (95% CI 8.9-12.3%) ^[47]B2a. Safety events (death or ICU admission) occurred in only 0.1% of episodes ^[47]B2a.

Common oral regimens for low-risk patients include ciprofloxacin plus amoxicillin-clavulanate or levofloxacin monotherapy (in penicillin-allergic patients).

Step 4: When to Broaden Therapy

Broadening empiric coverage is indicated when:

• Persistent or recurrent fever after 72-96 hours despite adequate empiric therapy (see “Persistent Fever and Antifungal Therapy” for the next step).
• New clinical deterioration (hypotension, organ failure, new focal signs).
• Culture results reveal a resistant organism requiring escalation.

For patients with persistent fever without a source, 18F FDG-PET-CT has been shown to guide antimicrobial rationalization more effectively than conventional CT. The PIPPIN trial (2022) randomized 147 patients with persistent/recurrent neutropenic fever to PET-CT or CT; antimicrobial rationalization occurred in 82% vs 65% (OR 2.36, 95% CI 1.06-5.24) ^[46]A1b. PET-CT supports decisions to stop or narrow therapy when no infection is identified ^[46]A1b.

Step 5: Pediatric Considerations

In children, the evidence is consistent with adults. A systematic review of 13 pediatric studies found no safety difference between early cessation of antibiotics versus continuation when no infection is proven, and no difference between oral and intravenous therapy in low-risk patients ^[22]B2a. However, no pediatric studies specifically examined de-escalation of empiric broad-spectrum antibiotics ^[22]B2a. Most protocols use a risk-stratified approach with a 48-hour observation period before de-escalation.

Pearl: De-escalate or stop empiric antibiotics after 72 hours of apyrexia and clinical stability, regardless of neutrophil count; this improves outcomes and reduces antimicrobial resistance without increasing mortality (How Long study, meta-analysis ^{[18]A1a[45]A1b}).

When fever persists or recurs after 72-96 hours of broad-spectrum antibiotics , the likelihood of invasive fungal infection (IFI) increases, and antifungal therapy should be initiated. This section addresses the definition of persistent fever, risk factors for IFI, diagnostic evaluation, and empiric antifungal choices.

Definition of Persistent Fever

Persistent fever is defined as fever that continues beyond 72-96 hours of empiric antibacterial therapy despite adequate antibiotic coverage, without evidence of non-infectious causes. Recurrent fever after an initial defervescence also warrants evaluation for IFI.

Risk Factors for Invasive Fungal Infection

• Prolonged neutropenia (>7 days) - the strongest predictor.
• High-risk hematologic malignancy (acute myeloid leukemia, myelodysplastic syndrome).
• Allogeneic hematopoietic cell transplantation (HCT) with graft-versus-host disease.
• Prior fungal colonization (e.g., Aspergillus, Candida species ).
• Use of corticosteroids or other immunosuppressive agents.
• Central venous catheters and prolonged parenteral nutrition.

Diagnostic Workup

Serum biomarkers and imaging guide the decision to start antifungal therapy. The 2023 International Pediatric Fever and Neutropenia Guideline Panel made a conditional recommendation for pre-emptive antifungal therapy in high-risk patients not receiving antimold prophylaxis, using serial screening with serum galactomannan and/or β-D-glucan ^[50]A1c.

Caption: Diagnostic approach to persistent fever in neutropenic patients (adapted from ^{[50]A1c[54]A1b}).

• Serum galactomannan - sensitivity ~60-70% for invasive aspergillosis ; positive threshold ≥0.5.
• β-D-glucan - detects a broad range of fungi including Candida, Aspergillus, Pneumocystis; false positives with hemodialysis or certain antibiotics.
• CT chest - typical findings: nodules, halo sign, air crescent sign (suggestive of angioinvasive aspergillosis).
• CT sinuses - if sinus tenderness or nasal ulceration.

Empiric Antifungal Therapy

Once the decision to treat is made, echinocandins are the preferred first-line agents. A meta-analysis of six RCTs (four evaluating caspofungin , two evaluating micafungin) found that echinocandins reduced all-cause mortality compared with non-echinocandins (risk ratio [RR] 0.70, 95% CI 0.49-0.99) and decreased adverse events (RR 0.48, 95% CI 0.33-0.71) ^[57]A1a. When compared specifically with liposomal amphotericin B, mortality was also lower (RR 0.68, 95% CI 0.46-0.99) ^[57]A1a.

Pre-emptive vs Empirical Approach

Two strategies exist: empiric (start antifungal in all patients with persistent fever) vs pre-emptive (start only when biomarkers or imaging suggest IFI). The CEDMIC trial (N=413) showed that D-index-guided early therapy (DET) using micafungin 150 mg/day was noninferior to empiric therapy for preventing proven/probable IFI (0.5% vs 2.5%; difference -2.0%, 90% CI -4.0% to 0.1%) and significantly reduced antifungal use (60.2% vs 32.5%, P < 0.001) ^[54]A1b. In pediatric patients, a single RCT (n=149) found no difference in overall mortality (8% empirical vs 5% pre-emptive, P=0.47) or IFI-related mortality (3% each), while antifungal duration was shorter in the pre-emptive arm (6 vs 11 days, P < 0.001) ^[55]A1b. A 2024 systematic review confirmed this paucity of data but noted moderate-certainty evidence that both approaches are comparable ^[56]A1a.

Controversies and Guideline Disagreement

Pearl: In patients with persistent fever beyond 72-96 hours, initiate an echinocandin (caspofungin 70 mg/50 mg or micafungin 150 mg daily) while awaiting serum galactomannan and CT chest; a pre-emptive strategy using biomarker screening can reduce antifungal overuse without increasing mortality, particularly in high-risk patients receiving antimold prophylaxis ^{[50]A1c[54]A1b[57]A1a}.

While antifungal therapy addresses persistent fever in high-risk patients, supportive care measures, including granulocyte colony-stimulating factor (G-CSF), antipyretics, fluid resuscitation, and infection control, form the foundation of management for all patients with neutropenic fever.

Granulocyte Colony-Stimulating Factor (G-CSF)

Prophylactic use is recommended for patients receiving chemotherapy with a febrile neutropenia (FN) risk ≥20% ^[58]A1c. Primary prophylaxis with G-CSF significantly reduces FN incidence (risk difference 0.22, 95% CI 0.01-0.43; p=0.04) ^[10]A1a. A single dose of pegylated G-CSF (PEG G-CSF) is strongly preferred over multiple doses of non-PEG G-CSF because it lowers FN incidence more effectively ^[11]A1a. In dose-dense chemotherapy for early-stage breast cancer, prophylactic pegfilgrastim maintains survival trends without increasing infection risk, though bone pain is more common (OR 2.57, 95% CI 1.00-6.62; p=0.05) ^[9]A1a. For urothelial cancer receiving dose-dense MVAC, primary prophylaxis is weakly recommended ^[13]D5. In contrast, primary G-CSF prophylaxis is inappropriate for colorectal cancer chemotherapy ^[12]B2a.

Therapeutic use of G-CSF during established FN is not routinely recommended. A meta-analysis found no significant reduction in infection-related mortality (RR 0.83, 95% CI 0.27-2.58; p=0.54) ^[59]B2a. G-CSF may be considered only for high-risk patients (e.g., prolonged neutropenia, pneumonia, hypotension, multiorgan dysfunction) ^[59]B2a.

Antipyretic Therapy

Acetaminophen (650 mg orally every 4-6 hours as needed) is the antipyretic of choice. Nonsteroidal anti-inflammatory drugs (NSAIDs) are avoided because they increase the risk of renal impairment, gastrointestinal bleeding, and may mask fever trends.

Fluid Resuscitation

For patients with hypotension or sepsis, initial fluid resuscitation with 30 mL/kg of crystalloid (e.g., lactated Ringer's or normal saline) is given, targeting a mean arterial pressure ≥65 mm Hg. Careful monitoring for fluid overload is essential, especially in patients with cardiac or renal comorbidities.

Infection Control Measures

Standard infection control practices reduce nosocomial transmission. Hand hygiene before and after patient contact is mandatory. Protective isolation (single room with HEPA filtration) is reserved for high-risk patients with prolonged severe neutropenia (absolute neutrophil count <100/μL). Low-risk patients may be managed in an outpatient setting with oral antibiotics ; levofloxacin is a cost-effective option compared with amoxicillin-clavulanate / ciprofloxacin ^[38]A1b.

Pearl: For therapeutic G-CSF, the evidence does not support routine use; reserve it for high-risk patients with complicated neutropenic fever (e.g., pneumonia, hypotension, prolonged neutropenia) where the potential benefit may outweigh the lack of proven mortality reduction ^[59]B2a.

The preceding section emphasized supportive care; however, specific populations demand tailored strategies that modify every step from risk stratification to antibiotic de-escalation. Age, underlying disease, and immune status shift the epidemiology , pharmacokinetics, and tolerance of therapy, requiring clinicians to adjust thresholds and choices.

Pediatrics

Children with febrile neutropenia differ from adults in several key aspects. The most common underlying malignancies are acute lymphoblastic leukemia (ALL) and acute myelogenous leukemia (AML), and the incidence of FN is higher during induction chemotherapy. Fluoroquinolone prophylaxis during induction for pediatric ALL reduces the risk of FN from 64.9% to 46.1% (OR 0.44; 95% CI 0.33-0.59) and of bloodstream infection (BSI) by half (OR 0.50; 95% CI 0.32-0.81) without increasing Clostridioides difficile infection ^[17]B2a.

For empiric monotherapy, both piperacillin-tazobactam (PIPC/TAZ) and meropenem (MEPM) are effective. A prospective randomized trial of 93 children (405 episodes) using MEPM 120 mg/kg/day as a 2-hour drip infusion three times daily or PIPC/TAZ 360 mg/kg/day as a 2-hour drip infusion four times daily reported first-line success rates of 74.9% and 74.8%, respectively ^[15]A1b. Grade 3 diarrhea was more common with PIPC/TAZ (15 episodes vs. 4 with MEPM, P < 0.001) ^[15]A1b. A meta-analysis of five RCTs (470 episodes) comparing PIPC/TAZ with cefepime found no difference in treatment success (RR 1.02; 95% CI 0.89-1.18) or mortality, but treatment duration was 0.9 day shorter with cefepime ^[31]A1a.

Adolescents and young adults (≥16 years, weight ≥52 kg) have lower antibiotic efficacy, a combined first- and second-line success rate of 82.0% vs. 94.0% in younger children (P < 0.001) ^[15]A1b. The reasons are unclear but may involve lower relative dosing or higher fungal infection risk. Rapid molecular diagnostics (e.g., Sepsis qPCR MX-30® panel) can identify pathogens and resistance genes earlier but have limited sensitivity (22.2%) and cannot replace blood cultures ^[61]C4.

Elderly

Age-related decline in organ function and polypharmacy increase the risk of drug-drug interactions and adverse effects. Early antibiotic de-escalation (within 2-3 days) is particularly beneficial in patients >55 years, reducing mortality risk (OR 0.42; 95% CI 0.18-0.98) ^[18]A1a. The benefit is thought to arise from reduced drug exposure and lower risk of resistant organisms. Empiric therapy should follow the same antipseudomonal β-lactam monotherapy as younger adults, but renal function must guide dosing (e.g., piperacillin-tazobactam dose adjustment for creatinine clearance <40 mL/min). The BEATLE trial (150 patients, median age not reported but included acute leukemia and HSCT) showed no benefit of extended infusion over intermittent infusion for day 5 success (50.6% vs. 63.0%; risk difference -12.4%, 95% CI -29.4 to 4.7), though extended infusion achieved higher pharmacokinetic/pharmacodynamic targets ^[33]A1b. Because elderly patients often have lower physiological reserve, maintaining a low threshold for escalation to carbapenems when local gram-negative resistance rates are high is prudent.

Hematologic Malignancy and Stem Cell Transplant

Patients with hematologic malignancies (especially AML, ALL, and multiple myeloma ) and those undergoing hematopoietic stem cell transplantation (HSCT) are at highest risk for prolonged, profound neutropenia and multidrug-resistant infections. A retrospective cohort of 164 FN episodes with BSI in hematologic patients found that 56.7% of BSIs were gram-negative, with E. coli the most common pathogen; ESBL positivity was 49.3% and carbapenem resistance 20.2% ^[24]B3b. Carbapenem resistance, but not ESBL, was independently associated with 30-day ICU admission and 30-day mortality ^[24]B3b.

Empiric antibiotic choice should be guided by local antibiograms. In centers with high carbapenem resistance, initial therapy with a carbapenem (e.g., meropenem) may be warranted. Antifungal prophylaxis is also essential: in children undergoing HSCT, intravenous micafungin 1 mg/kg/day (max 50 mg/day) from conditioning until neutrophil recovery, then oral fluconazole 10 mg/kg/day (max 400 mg/day) is recommended ^[15]A1b.

Early de-escalation of broad-spectrum antibiotics (BSA) before hematopoietic recovery is feasible and reduces mortality (OR 0.20; 95% CI 0.06-0.69) in this population, based on a meta-analysis of 10 studies (mainly retrospective) ^[18]A1a. The effect was strongest when de-escalation occurred within 3 days (OR 0.14; 95% CI 0.03-0.66) ^[18]A1a. De-escalation strategies include switching to prophylactic fluoroquinolones, narrowing to monotherapy, or stopping all antibiotics, provided the patient is hemodynamically stable and afebrile for ≥48 hours ^[18]A1a.

Catheter-Related Infections

Central venous catheters (CVCs) are common in oncology patients and may be the source of FN. Paired blood cultures from the CVC and a peripheral vein should be obtained; a differential time to positivity >2 hours suggests catheter-related BSI. Empiric therapy should include an antipseudomonal β-lactam plus an agent active against gram-positive organisms (e.g., vancomycin ) if the patient has a CVC and is hemodynamically unstable, or if there is skin or soft tissue infection at the exit site. If cultures confirm a catheter-related BSI, catheter removal is indicated for infections with S. aureus, Pseudomonas aeruginosa , or Candida species , or if bacteremia persists >72 hours despite appropriate antibiotics.

Pearl: In children with febrile neutropenia, adolescents ≥16 years have lower antibiotic efficacy (82% vs 94%) and may benefit from earlier escalation to second-line therapy; in elderly patients, early de-escalation (within 2-3 days) reduces mortality (OR 0.42).

For these vulnerable groups, outcomes depend on the interplay of host factors, microbial virulence, and timeliness of care. The prognosis of febrile neutropenia (FN) has improved dramatically over the past decades, but mortality and morbidity remain substantial in specific subsets.

Overall Mortality

Contemporary mortality in pediatric FN, treated in high-resource centers, is low, approximately 2-3% overall ^[32]B2a. In the largest international individual-patient data meta-analysis (5433 episodes from 15 centers), safety-relevant events (SREs), defined as bacteremia, ICU admission, or death, occurred in 17% of all episodes and 34% of episodes with severe disease at presentation ^[32]B2a. Among those with severe disease, the case-fatality rate was 2.6% (9 deaths out of 345 episodes) ^[32]B2a. In hematologic malignancy patients, early de-escalation of broad-spectrum antibiotics (before hematopoietic recovery) was associated with a significant reduction in mortality (OR 0.20, 95% CI 0.06‑0.69; NNT not calculable from reported data) ^[18]A1a. This benefit was most pronounced in patients older than 55 years (OR 0.42, 95% CI 0.18‑0.98) ^[18]A1a. Notably, fluoroquinolone prophylaxis during induction chemotherapy for pediatric acute lymphoblastic leukemia significantly reduces FN and bloodstream infection but does not significantly alter all-cause mortality (OR 1.04, 95% CI 0.20‑5.38) ^[17]B2a.

Risk Factors for Poor Outcome

Several factors independently predict adverse outcomes:

• Severe disease at FN presentation, severe sepsis, reduced clinical condition, or high triage category. In the IPD meta-analysis, SREs occurred in 34% of such episodes vs. 17% overall ^[32]B2a.
• Weight loss ≥5% during chemoradiotherapy, associated with a 4.5-fold increased odds of FN (OR 4.49, 95% CI 1.47‑13.80) ^[14]B3b.
• Prolonged neutropenia, duration >14 days is a classic risk factor, though subgroup analyses have not always shown a significant interaction with de-escalation outcomes ^[18]A1a.
• Older age, patients >55 years derive mortality benefit from early de-escalation, suggesting higher baseline risk ^[18]A1a.
• Carbapenem-resistant Enterobacteriaceae bacteremia, associated with high mortality ^[18]A1a.
• Radiation esophagitis (grade ≥2), FN incidence 72.7% vs. 45.5% among those with grade 0-1 esophagitis ^[14]B3b.

Complications

Complications of FN include septic shock, ICU admission, organ failure, and invasive fungal infections. In the IPD meta-analysis, 50 ICU admissions occurred among 345 severe episodes (14.5%) ^[32]B2a. Bacteremia was found in 95 of 345 severe episodes (27.5%) ^[32]B2a. Recurrent fever after initial defervescence occurs in approximately 20-30% of patients, though early de-escalation does not increase this risk (OR 0.88, 95% CI 0.56‑1.38) ^[18]A1a.

Impact on Chemotherapy

FN episodes frequently lead to chemotherapy dose delays, dose reductions, or treatment discontinuation, especially in patients receiving curative-intent regimens. Weight loss ≥5% during chemoradiotherapy is both a risk factor for FN and a consequence of FN-driven toxicity, creating a cycle of treatment intolerance ^[14]B3b. In low- and middle-income countries, where supportive care resources are limited, FN is a major contributor to the survival gap ^[60]B2b.

Prognostic Factors Table

Pearl: In pediatric FN, the strongest predictor of a safety-relevant event is severe disease at presentation, mortality is <3% overall but primarily clusters in this group; early de-escalation of antibiotics before hematopoietic recovery, particularly in older patients, may reduce mortality without increasing complications.

Given the substantial morbidity and mortality from neutropenic fever, prevention strategies are a critical component of care. Two pharmacologic approaches, prophylactic antibiotics and granulocyte colony‑stimulating factors (G‑CSF), are supported by high‑quality evidence, and their use should be guided by patient‑ and regimen‑specific risk assessment.

Antibiotic Prophylaxis

In patients with newly diagnosed multiple myeloma starting active treatment, the TEAMM trial demonstrated that oral levofloxacin 500 mg once daily for 12 weeks significantly reduced the incidence of first febrile episode or death from 27% to 19% (HR 0.66, 95% CI 0.51-0.86; p=0.0018; NNT=13 over 12 weeks) ^[63]A1b. Serious adverse events were similar between groups, with five episodes (1%) of mostly reversible tendonitis in the levofloxacin arm ^[63]A1b.

In pediatric acute lymphoblastic leukemia (ALL) during induction, a systematic review and meta‑analysis of seven studies (991 patients) found that fluoroquinolone prophylaxis (levofloxacin or ciprofloxacin ) reduced febrile neutropenia (FN) from 64.9% to 46.1% (OR 0.44; 95% CI 0.33-0.59; NNT=6) and bloodstream infections (BSI) from a pooled rate (OR 0.50; 95% CI 0.32-0.81; NNT not calculable from reported data) without a significant increase in Clostridioides difficile infection or all‑cause mortality ^[17]B2a. An interim analysis of a randomized trial in Brazilian children with ALL confirmed the safety of levofloxacin during induction, showing no increase in carbapenemase‑producing Enterobacteriaceae colonization or C. difficile diarrhea ^[66]A1b.

Indications for antibiotic prophylaxis (based on the available evidence):

• Newly diagnosed multiple myeloma undergoing active treatment (levofloxacin 500 mg daily for 12 weeks) ^[63]A1b
• Pediatric ALL during induction chemotherapy (fluoroquinolone, typically levofloxacin or ciprofloxacin) ^{[17]B2a[66]A1b}
• Other high‑risk populations where the expected FN rate exceeds 20% and no competing contraindications exist (e.g. fluoroquinolone allergy, prior colonization with resistant organisms)

Limitations and considerations:

• Antibiotic prophylaxis selects for resistant organisms; surveillance for colonization remains prudent ^[66]A1b.
• There is no evidence from the retrieved studies to support routine prophylaxis in all patients with solid tumors unless the chemotherapy regimen carries a high FN risk (see below).

G‑CSF Prophylaxis

Primary prophylaxis with G‑CSF is recommended for chemotherapy regimens expected to have an FN risk >20% ^[67]B2a. The strongest evidence comes from a meta‑analysis of 68 studies showing that primary G‑CSF support reduced all‑cause mortality (RR 0.92; 95% CI 0.90-0.95; absolute risk difference -; NNT=30) and was particularly effective with dose‑dense regimens (mortality RR 0.86; 95% CI 0.80-0.92; p<0.0001) ^[68]A1a. However, G‑CSF was also associated with an increased risk of secondary malignancies, including acute myeloid leukemia and myelodysplastic syndrome (RR 1.85; 95% CI 1.19-2.88; absolute risk difference 0.47%; NNH=213) ^[68]A1a.

In breast cancer, a systematic review of eight RCTs found that primary G‑CSF prophylaxis significantly reduced FN incidence (risk difference 0.22; 95% CI 0.01-0.43; p=0.04) ^[10]A1a. For the TCH regimen ( docetaxel , carboplatin , trastuzumab ± pertuzumab) in HER2‑positive breast cancer, the pooled FN incidence without G‑CSF was 27.6% (95% CI 18.6-37.1) compared with 5.0% (95% CI 2.6-8.0) with G‑CSF, giving an absolute risk reduction of (NNT=5) ^[67]B2a.

A pragmatic trial comparing ciprofloxacin versus G‑CSF as primary prophylaxis in patients receiving docetaxel‑ cyclophosphamide for breast cancer met feasibility endpoints but was not powered to detect differences in FN rates; overall FN incidence was 21.4% ^[64]A1b. This underscores the need for individualized prophylaxis selection based on patient preference, cost, and institutional resistance patterns.

Indications for primary G‑CSF prophylaxis:

• Chemotherapy regimens with an expected FN risk >20% (e.g., TCH(P), dose‑dense regimens) ^[67]B2a
• Patients with prior FN episodes or high‑risk features (e.g., age >65, poor performance status, advanced disease, comorbidities)
• Secondary prophylaxis is warranted after a documented FN episode if the same chemotherapy regimen is continued

Risks:

• Secondary malignancies (particularly AML/MDS), absolute increase about 0.5% over long‑term follow‑up ^[68]A1a
• Bone pain, mild injection site reactions

Patient Education

• Instruct patients to report fever (single oral temperature ≥38.3°C or ≥38.0°C sustained for 1 hour) immediately, even before planned prophylaxis is initiated.
• Emphasize hand hygiene, avoidance of sick contacts, and prompt attention to any signs of infection.
• Discuss the rationale for antibiotic or G‑CSF prophylaxis, including expected benefits (reduced FN, hospitalizations) and potential harms (antibiotic resistance, tendonitis, secondary malignancy).

Controversies and Guideline Disagreement

Pearl: Primary prophylaxis with G‑CSF is indicated for chemotherapy regimens with FN risk >20%, reducing FN incidence from 27.6% to 5.0% (NNT=5) ^[67]B2a. Antibiotic prophylaxis with levofloxacin reduces FN in myeloma (NNT=13) and pediatric ALL (NNT=6) ^{[63]A1b[17]B2a}. Balance benefits against the small increased risk of secondary malignancies with G‑CSF (NNH=213) ^[68]A1a.

Building on these prevention strategies, multiple professional societies have published clinical practice guidelines that consolidate the evidence for the management of established neutropenic fever. The following table summarizes key recommendations from the most recent guidelines included in the provided literature, focusing on risk stratification, empiric therapy, G‑CSF use, and special considerations.

Practice‑Changing Updates

Several updates carry direct clinical implications. The 2023 pediatric guideline ^[50]A1c now permits earlier discontinuation of antibiotics in low‑risk patients (48 h negative cultures), a shift from prior practice of waiting until neutrophil recovery. The ASCO 2015 guideline ^[69]A1c explicitly discourages routine dose‑dense chemotherapy outside of trials unless efficacy data support it. ECIL‑9 ^[70]A1c highlights that the infectious risk of novel targeted therapies is low as monotherapy, but combination regimens (e.g., venetoclax‑HMA) require heightened vigilance. The AGIHO 2026 guideline ^[58]A1c extends G‑CSF recommendations to immunotherapy settings, where the risk of febrile neutropenia may differ from conventional chemotherapy.

Controversies and Guideline Disagreement

The AGIHO panel ^[58]A1c explicitly notes that international guidelines conflict on several aspects of G‑CSF prophylaxis, particularly regarding risk thresholds and patient subgroups. ASCO ^[69]A1c recommends a ≥20% risk threshold for primary prophylaxis, while NCCN ^[71]A1c and other European guidelines sometimes use lower thresholds (e.g., ≥10‑20% with additional risk factors). The lack of uniformity underscores the need for local adaptation based on patient population and chemotherapy regimen.

Resources and Clinical Tools

The MASCC score (detailed in the Risk Stratification section) is endorsed by MASCC and the pediatric guideline as a validated tool for identifying low‑risk FN patients suitable for outpatient management or early discharge ^{[50]A1c[72]D5}. The 2023 pediatric guideline ^[50]A1c also provides a risk‑stratification algorithm for antifungal pre‑emptive therapy. Telemedicine and hospital‑at‑home models, recommended by MASCC during the COVID‑19 pandemic ^[72]D5, remain valuable for reducing nosocomial exposure in eligible patients. Patient information resources are available through the MASCC website and the NCCN guidelines for patients.

Pearl: When selecting a guideline to follow, note that the pediatric 2023 update ^[50]A1c supports stopping antibiotics at 48 h in low‑risk, culture‑negative, afebrile patients even without marrow recovery, a significant departure from adult practice, and that the AGIHO 2026 guideline ^[58]A1c now provides specific G‑CSF recommendations for immunotherapy, a rapidly expanding area not covered by older guidelines.