Febrile neutropenia: grade, risk and treatment

Fever (IDSA 2010; ASCO/IDSA 2018): a single oral reading of 38.3 °C (101 °F) or more, or 38.0 °C (100.4 °F) or more sustained over 1 hour. Readings in °F are converted to °C and rounded to one decimal before they are classified, so 100.9 °F counts as 38.3 °C. IDSA's parenthesis rounds that cutoff to 101 °F; the tool uses the lower figure so no fever is missed.

Neutropenia (IDSA 2010): ANC under 500 cells/µL, or an ANC expected to fall under 500 during the next 48 h. ASCO/IDSA 2018 calls an ANC under 1,000 neutropenia, under 500 severe and under 100 profound. The expected-fall question appears for any ANC of 500 or more. The 500 cutoff is judged on the exact ANC (0.4995 ×10⁹/L, or 0.9 × 55.5%, counts as under 500). IDSA adds that its definitions "are not hard-and-fast rules", and that "functional neutropenia" from a hematologic malignancy also carries increased risk despite a normal count; an ANC of 500–999 not expected to fall is shown as below the treatment definition, never as "not neutropenic".

High and low risk (IDSA 2010 full text, high-risk criteria list): high risk is expected prolonged (over 7 days) and profound (100 or less) neutropenia, or comorbidity: hemodynamic instability; mucositis that interferes with swallowing or causes severe diarrhea; gastrointestinal symptoms (abdominal pain, nausea and vomiting, or diarrhea); new neurologic or mental-status change; intravascular catheter infection; new pulmonary infiltrate, hypoxemia or underlying chronic lung disease; aminotransferases over 5× normal or CrCl under 30. MASCC is Rec 4. MASCC under 21 is high risk; 21 or more is low risk (Rec 4, B-I). Where IDSA and ASCO draw a boundary differently (100 or less against under 100; 30 or less against under 30), the tool uses the more cautious one.

Not neutropenic or not febrile: the tool stops and says so. Afebrile patients with new signs of infection are treated as high risk (IDSA Rec 13, B-III).

CTCAE v6.0 shifted every neutrophil grade by one level, so the result names the ANC band first. NCI (FAQ, 9 Sep 2025): v6.0 is ready for immediate use; NCI CTEP and DCP studies reporting in v5.0 stay on v5.0 for their whole life; new CTEP/DCP studies use v6.0 once its Rave release ships, tentatively July 2026. Most current trials and label dose tables still use v5.0. Febrile neutropenia (same in v5 and v6) has no grade 1 or 2: grade 3 is an ANC under 1,000 with a single reading over 38.3 °C (101 °F), or 38 °C or more for over 1 hour; grade 4 means life-threatening consequences with urgent intervention indicated; grade 5 is death. A single reading of exactly 38.3 °C meets the IDSA/ASCO fever definition but not the CTCAE wording.

MASCC (Klastersky 2000), maximum 26: burden of illness with no or mild symptoms 5, or moderate 3 (not added together; severe or moribund 0); no hypotension (systolic over 90) 5; no COPD 4; solid tumour, or hematologic cancer without previous fungal infection, 4; no dehydration needing IV fluids 3; outpatient when fever began 3; age under 60, 2. 21 or more is low risk. In pooled data, serious complications still occurred in up to 11% of patients classified low risk (ASCO 2018). A score with any item missing shows its possible range; it is high risk once the range cannot reach 21.

CISNE (Carmona-Bayonas 2015), 0–8, for solid-tumour patients on mild-to-moderate chemotherapy who look clinically stable: ECOG 2 or worse 2; stress-induced hyperglycemia 2; COPD 1; chronic cardiovascular disease 1; mucositis grade 2 or more 1; monocytes under 200/µL 1. Classes: low 0, intermediate 1–2, high 3 or more. In validation, serious complications were 1.1%, 6.2% and 36% by class. ASCO 2018 Figure 1: CISNE 3 or more means inpatient care. The derivation paper defines stress hyperglycemia as glucose 121 mg/dL or more (250 or more with diabetes; it prints the unit as mg m⁻², read here as mg/dL).

Talcott's rules (also endorsed by ASCO 2.1b) are not scored here.

Monotherapy with an antipseudomonal beta-lactam: cefepime, piperacillin-tazobactam, meropenem or imipenem-cilastatin (IDSA Rec 9, A-I; ASCO 1.1h). Only cefepime has a US febrile neutropenia indication; the others are guideline-endorsed and off-label for this use. The cefepime label notes that monotherapy may not be appropriate in its high-risk groups (recent marrow transplant, hypotension at presentation, hematologic cancer, severe or prolonged neutropenia), and the guidelines add agents for complications.

Gram-positive cover is not routine (IDSA Rec 10, A-I); agents "should be considered" for specific indications, including penicillin-resistant pneumococcal colonization. IDSA Table 4 indications: hemodynamic instability or severe sepsis; pneumonia on imaging; gram-positive bacteria in blood culture before identification; clinically suspected serious catheter infection; skin or soft-tissue infection; colonization with MRSA, VRE or penicillin-resistant pneumococcus; severe mucositis only if fluoroquinolone prophylaxis was given and ceftazidime is the empirical agent. Stop it after 2 days if no gram-positive infection is found (Rec 18, A-II).

Resistant organisms (IDSA Rec 11, B-III; ASCO 1.1h): MRSA, vancomycin, linezolid or daptomycin (not daptomycin for pneumonia); VRE, linezolid or daptomycin; ESBL, a carbapenem; carbapenemase producers, a newer beta-lactam active against resistant gram-negatives (ASCO 2018), chosen with infectious diseases. IDSA 2010's colistin or tigecycline advice predates those agents.

Allergy (IDSA Rec 12, A-II): most patients with penicillin allergy tolerate cephalosporins. With immediate-type hypersensitivity IDSA avoids beta-lactams and carbapenems: ciprofloxacin plus clindamycin, or aztreonam plus vancomycin. The four allergy answers and their rules match the MDRO tool: unverified, non-anaphylactic penicillin allergy removes penicillins pending evaluation and keeps cephalosporins, carbapenems and aztreonam (the 2022 drug-allergy practice parameter suggests a cephalosporin may be given without testing here; conditional, moderate certainty; secondary summary). Immediate or anaphylactic penicillin or non-ceftazidime cephalosporin allergy removes penicillins and cephalosporins by default but keeps carbapenems and aztreonam, which the parameter suggests may be given without testing (conditional); after penicillin anaphylaxis it also allows a cephalosporin with a dissimilar R1 side chain, so an allergist-cleared cephalosporin is an option. Ceftazidime allergy removes aztreonam (shared side chain) and, conservatively, other penicillins and cephalosporins; carbapenems stay. Severe delayed reactions (SJS/TEN, DRESS) remove every beta-lactam and call for allergy and ID input.

Fluoroquinolone prophylaxis "strictly precludes the subsequent use of fluoroquinolones for initial empirical therapy" (IDSA 2010; ASCO 2018 agrees): ciprofloxacin and levofloxacin are removed from every line, IV and oral.

Outpatient oral: ciprofloxacin plus amoxicillin-clavulanate (IDSA 14i, A-I; ASCO 5.1, moderate); ASCO also allows levofloxacin. With penicillin allergy, a fluoroquinolone plus clindamycin (ASCO 5.1; IDSA lists ciprofloxacin plus clindamycin among less-studied regimens, B-III). Not a fluoroquinolone alone. Neither guideline gives oral doses, so none are shown.

Abdominal pain or diarrhea (IDSA 2010): CT for neutropenic enterocolitis; monotherapy with piperacillin-tazobactam or a carbapenem, or an antipseudomonal cephalosporin plus metronidazole, so cefepime is listed only with metronidazole and never first. Diarrhea: C. difficile toxin assay; IDSA 2010 allows empirical treatment while awaiting results; treat per IDSA/SHEA 2021 (fidaxomicin preferred, conditional, moderate certainty; oral vancomycin "remains an acceptable alternative"; metronidazole no longer first-line). The VRE answer removes IV vancomycin as the gram-positive agent only. With prior ESBL, aztreonam is removed too (ESBLs hydrolyse it; IDSA Rec 11 advises a carbapenem; the MDRO tool's ESBL-E rule offers no aztreonam). In enterocolitis, the allergy alternatives are not IDSA regimens: aztreonam needs added metronidazole, and either is a decision for ID. Oral ulcers or esophagitis in high-risk patients: empirical acyclovir and/or fluconazole are appropriate; HSV/VZV treatment only with evidence of active disease (Rec 36, C-III). Unstable with persistent fever: a carbapenem, plus an aminoglycoside, ciprofloxacin or aztreonam with vancomycin, and anti-Candida cover (IDSA 2010), each subject to the same exclusions.

US label, empiric therapy for febrile neutropenia: 2 g IV every 8 hours for 7 days or until neutropenia resolves, with CrCl over 60 mL/min. The first dose is not reduced for kidney function: label §2.3 says the initial dose "should be the same as in patients with CrCL greater than 60 mL/min except in patients undergoing hemodialysis". Table 2 maintenance after a 2 g first dose: CrCl 30–60, 2 g every 12 h; 11–29, 2 g every 24 h; under 11, 1 g every 24 h; CAPD, 2 g every 48 h. Hemodialysis: 1 g every 24 h, given after dialysis on dialysis days. CRRT is not in the label. CrCl is rounded to a whole number before the row is chosen. Encephalopathy, myoclonus, seizures and nonconvulsive status epilepticus occur mostly when the dose is not adjusted for kidney function (label §5.2).

Secondary evidence: a 2026 Bayesian meta-analysis of 110 trials found a 94.4% probability of higher mortality with cefepime than other beta-lactams (OR 1.10, 95% credible interval 0.98–1.24). No guideline has changed on the strength of it.

Duration (IDSA): documented infection, by site and organism and at least until ANC is 500 or more (Rec 22, B-III); unexplained fever, until clear marrow recovery (Rec 23, B-II); or, after a full course with symptoms resolved, resume fluoroquinolone prophylaxis until recovery (Rec 24, C-III). ECIL-4 (2013) instead suggests stopping after 72 h or more in fever of unknown origin if the patient has been hemodynamically stable since presentation and afebrile for 48 h or more, whatever the ANC; the How Long trial (2017) supported stopping after 72 h afebrile with clinical recovery.

Persistent fever: after 4–7 days of antibiotics with neutropenia expected to last over 7 days, consider empirical antifungal therapy and a fungal workup (IDSA Rec 28, A-I), or a pre-emptive approach for stable patients with negative CT and serology (Rec 29, B-II). Chest CT whenever invasive aspergillosis is suspected, and serum galactomannan in hematologic cancer or HSCT, but not as routine screening on mold-active prophylaxis (IDSA 2016 aspergillosis Recs 9, 10, 13). Empirical options: a lipid amphotericin B, caspofungin or micafungin, or voriconazole (Rec 74). Not routine at low risk (IDSA Rec 30).

G-CSF (ASCO 2026): not routinely added to antibiotics (3.2, strong) and not for afebrile neutropenia (3.1, strong); may be offered at high risk of poor outcome (3.3, conditional, low-quality evidence). Box 2 examples: sepsis syndrome, age 65 or more or frailty, ANC under 100, neutropenia expected over 7 days, a previous febrile neutropenia, pneumonia, invasive fungal or other documented infection, and being in hospital when fever began. Filgrastim, pegfilgrastim, eflapegrastim or biosimilars (6.1); avoid with concurrent chemoradiation, especially to the mediastinum (5.1, conditional).

Children: a separate guideline applies (Lehrnbecher 2023); this tool is for adults. HSCT: engraftment-period and graft-versus-host specifics, viral reactivation and transplant prophylaxis are outside it; HSCT and acute leukemia induction patients are unlikely outpatient candidates (ASCO 2.1). Pregnancy: pregnant or nursing patients are excluded from outpatient care (ASCO Table 1); drug choice in pregnancy is not covered. Liver impairment: only the Table 1 criteria are used; dose with pharmacy. Antibacterial, antifungal and antiviral prophylaxis; outpatient IV regimens; neutropenic enterocolitis; fever after CAR-T or bispecific therapy (cytokine release syndrome); doses other than cefepime; agent choice for carbapenem-resistant organisms (use the MDRO tool); local antibiograms.

Guideline check, October 2026: IDSA 2010 and ASCO/IDSA 2018 remain the current US febrile neutropenia guidelines; ASCO updated its growth factor guideline in 2026; NCI published CTCAE v6.0 in July 2025. NCCN guidance was not used because its current text could not be opened.