Infectious Disease Compendium

Neutropenia and Fever


IDSA Guidelines

Fevers (T > 38) that occur with cytotoxic chemotherapy.

Absolute neutrophil count < 1000 makes one neutropenic; < 100 makes one sentient culture media.

Oral temperatures in patients with mucositis may be a degree above core temperature and are not reliable (PubMed).

Epidemiologic Risks

Chemotherapy. Most of what follows applies to the hematologic malignancies and neutropenia expected to last greater than 5 days. Between advances in chemotherapy and colony stimulating factors, these are the only ones who get admitted. Bummer for ID docs.

An ounce of prevention is worth a pound of cure so follow the guidelines and wash your damn hands (PubMed).

microbiology and empiric Therapy: different stages of the neutropenic febrile state have different organisms and require different empiric antibiotics.

To wit:

Pre-Neutropenia: as the counts are falling it is reasonable to have the patient on prophylactic oral agents to prevent bacterial and fungal infections. As a rule that is ciprofloxacin 750 po bid and an antifungal. Based on the 2005 NEJM studies I do not think the bang is worth the buck for quinolones when used as prophylactic in neutropenic patients; you breed too much resistance without decreasing death (PubMed). On the other hand, an Annals meta analysis (and I never meta analysis I really liked) of quinolone prophylaxis showed decrease in all cause death 33% and infection related death 44% (PubMed).

Quinolone use leads to way resistant viridans streptococcal bacteremia/sepsis especially when the patient has mucositis.

What antifungal? Fluconazole < itraconazole < voriconazole < posaconazole (PubMed), (PubMed), (PubMed), (PubMed), (PubMed). Until there is a head to head study, either posaconazole or voriconazole is probably the antifungal of choice; although a review of the literature gives the nod to posaconazole (PubMed). Given the prices pharmacists are weeping across the nation.

To prevent pulmonary aspergillosis, inhaled lipid amphotericin B at a dose of 2.5 mL of a 5-mg/mL solution of liposomal amphotericin B is effective (PubMed). I have never seen this done.

In neutropenics, there may be greater breakthrough when treated with caspofungin (PubMed).

If lymphoma or other hematologic malignancy, tmp/sulfa 1 ds po qd for PJP prophylaxis is required (and often forgotten) especially if on steroids. My bias is qod tmp/sulfa is the optimal dose to breed resistance of everything, including PJP.

First Fever: E. coli, K. pneumonia, (especially first couple of hospitalizations), P. aeruginosa (more commonly in later hospitalizations) are the most common gram-negative rods. If you kill the Pseudomonas you will kill the others, but not vice versa. Stop the po prophylactic antibiotics and change to an anti-pseudomonal beta lactam in the following order carbapenem > cefepime = ceftazidime > piperacillin = piperacillin /tazobactam >> aztreonam. Keep reading for what to do about vancomycin.

Giving the beta-lactams as an extended infusion instead of a bolus is associated with better outcomes (PubMed).

Prolonged and profound neutropenia also worry re: viridans Streptococci (PubMed) Streptococcus mitis and S. oralis (PubMed).

There is one meta-analysis to suggest cefepime is INFERIOR to carbapenems (PubMed).

Prior antibiotics increase resistance, so, for example, if they were on piperacillin /tazobactam last admit, don't use it this admit (PubMed).

Vancomycin is problematic. MRSA rates are 73% where I am so I can no longer rely on the beta lactam to hold the fort while waiting on cultures. So if the patient has a central venous catheter and looks sick and thou art worried about S. aureus and coagulase negative staphylococci, INCLUDE vancomycin. The official indications for vancomycin are

  • Hemodynamic instability or other evidence of severe sepsis
  • Clinically suspected, serious, catheter-related infection
  • Positive blood culture for Gram-positive bacteria before final identification and susceptibility testing if available
  • Known colonization with MRSA or penicillin-resistant Streptococcus pneumoniae
  • Pneumonia documented radiographically

Prolonged fevers: fevers greater than 5 to 7 days, infections due to Candida (especially non-albicans) and moulds (especially Aspergillus and Mucormycosis) increase. Cultures are negative and signs and symptoms can be subtle. Empiric antifungals are given at this stage: amphotericin B = voriconazole or other anti-Aspergillus azole and amphotericin B = caspofungin.

The big problem with voriconazole and caspofungin is no Mucormycosis coverage. So guess what these patients get when placed on caspofungin or voriconazole? Mucor. No good deed ever goes unpunished. But posaconazole will cover mucor. Currently I am leaning towards posaconazole > voriconazole unless the patient has extensive prior exposure to azole antibiotics. But there is an nice review of the NEJM studies that cast doubt on the validity of the voriconazole use instead of amphotericin B (Online text).

As azoles get used for prophylaxis, patients can still get fungal infections, but rare and resistant fungi (PubMed); ID is nothing if not applied evolution.

"Only 2 strategies have been shown in prospective studies to improve survival from mold infection in patients with acute myelogenous leukemia or myelodysplastic syndrome: (1) preemptive initiation of antifungal therapy at first sign of invasive aspergillosis on computed tomography (CT) scan and (2) antifungal prophylaxis with posaconazole (PubMed)."

For prophylaxis, echinocandin ­based prophylaxis for AML may be associated with higher risk of breakthrough invasive fungal infections (PubMed).

Post neutropenic fevers: hepatosplenic Candida.

It used to be that antibiotics are continued until ANC is > 500 (or as needed for specific infections), but data suggests that if the patient becomes afebrile, then 72 hours of antibiotics is all that is needed: "In high-risk patients with haematological malignancies and febrile neutropenia, EAT can be discontinued after 72 h of apyrexia and clinical recovery irrespective of their neutrophil count . Even in high-risk patients with hematological malignancies (PubMed)."

Whilst we always worry about bacteria and fungi, viral causes of fever may be more common that suspected. A viral pathogen can be detected in 35% of patients, with 18% of pharyngeal aspirates and 22% of blood samples positive of a wide variety of pathogens (PubMed).


To state the obvious, keep looking for a source of infections. Signs and symptoms are blunted as there is are no WBC's. ANY new pain anywhere needs a careful evaluation, as pain is often the only sign of infection.

Never narrow the therapy once started. Never. To my way of thinking, broad spectrum antibiotics are in part replacing the missing WBC.

Pulmonary molds may best be found early with a chest CT (PubMed) that demonstrates a halo (Aspergillus) or reverse halo sign (Mucormycosis). Aspergillus can be diagnosed with the serum galactomannnan assay. With return of the WBC patients with Aspergillus pneumonia can get acutely worse and that, it turns out, is a good sign (PubMed). CT guided aspiration of pulmonary infiltrates has a reasonable yield in one study (PubMed).

CT's done early can be false positive (PubMed); small nodules are common.

Any red bump should be biopsied as it can be the ONLY manifestation of a disseminated fungus.

RLQ pain may be a sign of typhlitis.

Dosing antibiotics in cancer patients is different from 'normal' people and most should get the maximum dose (PubMed).


Fevers often do not resolve until the ANC returns. I always tell patients that the goal of antibiotics is not to get rid of the fever, but to prevent the patient from dying.

Curious Cases

Relevant links to my Medscape blog

The Mitis Touch

Pulling the Plug