A 31-year-old woman is hospitalized with a new diagnosis of acute myelogenous leukemia. She recently completed standard induction chemotherapy with cytarabine and daunorubicin. She has no other relevant medical history. Her absolute neutrophil count (ANC) has been below 500 cells/mm3 for the past 8 days, and she has been receiving ciprofloxacin, acyclovir, and antifungal prophylaxis. She develops a temperature to 38.5 °C (101.3 °F) with associated tachycardia (115 beats/min) but otherwise is asymptomatic. Her physical exam is normal, including intact oropharyngeal mucosa, a port site without erythema or tenderness, normal external perirectal exam, and no rash.
On the day of her first febrile episode, her ANC is 100 cells/mm3, her platelet count is 45,000 cells/mm3, and her chemistry panel and liver function panel are within normal limits. Chest X-ray is obtained; blood and urine cultures are drawn; and she is started on cefepime, 2 g IV every 8 hours. Her cultures are negative, but she remains febrile over the next 3 days. Repeat cultures are drawn. Four days later (7 days into her persistent fever), she develops a dry cough, her oxygen saturation on room air decreases to 90%, and a chest X-ray shows increased haziness at the right base. Vancomycin is added to her regimen. She undergoes a CT of the chest and bronchoscopy. In addition to bacterial and fungal cultures, bronchoalveolar lavage fluid is sent for an aspergillus galactomannan assay. A serum aspergillus galactomannan is also drawn. She is empirically started on voriconazole.
Patients with hematologic malignancies undergoing chemotherapy are at substantially increased risk of infection, particularly in the setting of neutropenia. Fever is a relatively sensitive although nonspecific sign of infection. While there is individual variation in clinical practice, fever is generally defined as an oral temperature greater than 38.3 °C (101 °F) or a temperature greater than 38.0 °C (100.4 °F) sustained over an hour. Neutropenia is defined as an ANC less than 500 cells/mm3 or an ANC that is expected to drop below 500 cells/mm3 in the next 48 hours. Profound neutropenia is defined as an ANC less than 100 cells/mm3, while prolonged neutropenia is defined as an ANC less than 500 cells/mm3 for 7 or more days (1). Patients with profound and/or prolonged neutropenia, as well as those with a rapid rate of decline in neutrophil count, are at higher risk for more severe infection.
Identifying high- versus low-risk patients with neutropenic fever is critical to determining the appropriate level of monitoring and treatment. Patients are considered to be at high risk for complications and death if they have:
- profound or prolonged neutropenia,
- certain high-risk malignancies and history of corresponding therapies (acute leukemia, allogeneic hematopoietic stem-cell transplantation, alemtuzumab therapy), and/or
- the presence of comorbidities (including graft-vs.-host disease, pneumonia, hypotension, new-onset abdominal pain, oral or gastrointestinal mucositis that impairs absorption, significant diarrhea, aminotransferase levels >5 times the upper limit of normal, renal insufficiency with a glomerular filtration rate <30 mL/min, or neurologic symptoms such as altered mental status) (1, 2).
Low-risk patients are those who are expected to have fewer than 7 days of neutropenia, are clinically stable, and do not have any of the medical comorbidities described above. Such patients are often those with solid tumors who are receiving outpatient chemotherapy.
In addition to clinical assessment, the Multinational Association for Supportive Care in Cancer (MASCC) Risk Index Score is a validated tool to stratify patients with neutropenic fever into high- and low-risk categories (Table). The scoring system takes into account severity of symptoms, hypotension, chronic obstructive pulmonary disease, prior fungal infection, dehydration, outpatient status, and age. (A calculator is online.) Low-risk patients are those with a score of 21 or greater, whereas high-risk patients have a score below 21.
Patients who are classified as high risk based on clinical assessment or by the MASCC Risk Index Score should be admitted to the hospital for treatment and monitoring. Low-risk patients can potentially be treated with broad-spectrum oral antibiotics as outpatients if they have reliable follow-up.
Because infectious symptoms are attenuated in neutropenic patients, a thorough and rigorous history and physical exam are essential to localize a possible infection. Information should be obtained regarding:
- the type of chemotherapy the patient has received,
- the time since last chemotherapy administration,
- receipt of antimicrobial prophylaxis,
- recent administration of colony-stimulating factors and blood products,
- history of prior infections, and
- colonization with drug-resistant pathogens.
Laboratory analysis should include a complete blood count with differential, a complete chemistry panel, liver function tests, at least 2 sets of blood cultures (including cultures from all lumens of indwelling vascular devices and at least one peripheral blood culture), urinalysis, and a urine culture (even if the urinalysis is normal, since neutropenic patients are often unable to produce pyuria). A chest X-ray should be considered for all patients and should always be obtained for patients with respiratory symptoms. Patients with abnormalities on chest X-ray should undergo a chest CT to further characterize the pathology. Similarly, there should be a low threshold to obtain a CT of the abdomen in patients with abdominal symptoms in order to assess for neutropenic enterocolitis (typhlitis).
Most patients with neutropenic fever are likely to require hospitalization. In the absence of localizing signs and symptoms, empiric antibiotic coverage should be broad-spectrum with antipseudomonal activity. First-line regimens (1) include:
- piperacillin/tazobactam (if a patient is at risk for invasive aspergillosis, an alternative agent may be considered, as this antibiotic can lead to a false-positive galactomannan test), and
Vancomycin is not recommended as part of the initial empiric antibiotic regimen unless there is evidence of a skin or soft-tissue infection, a catheter-associated infection, pneumonia, gram-positive bacteremia, hemodynamic instability, or the concurrence of severe mucositis and empiric gram-negative therapy with either ceftazidime or aztreonam (which have little and no gram-positive activity, respectively). Patients who have a history of colonization with methicillin-resistant Staphylococcus aureus, vancomycin-resistant enterococci, or extended-spectrum beta-lactamase-producing organisms should initially be treated with antibiotics that have activity against these pathogens.
For patients who respond clinically to empiric therapy, antibiotics should be continued until ANC is above 500 cells/mm3 for 2 consecutive days, with some guidelines recommending a cutoff of above 1,000 cells/mm3. Although antibiotics should be continued until neutrophils recover, coverage can be narrowed before then based on culture results (1). For example, if no resistant, gram-positive organisms have grown from cultures after 48 hours and the patient is clinically stable, vancomycin can be discontinued. For some low-risk patients with reliable gastrointestinal absorption, IV antibiotics can be transitioned to oral ciprofloxacin plus amoxicillin-clavulanic acid (or plus clindamycin, if penicillin-allergic) (1). In patients with an identified source of infection (e.g., skin or soft-tissue infection, pneumonia, bacteremia) antibiotics should be continued to complete the standard duration of therapy for the corresponding infectious syndrome, even if neutrophils recover sooner.
For high-risk patients who do not respond to initial therapy and remain febrile after 72 to 120 hours, an additional workup for other sources of infection should be undertaken (repeat bacterial cultures and order fungal cultures and CT scans of the chest and sinuses if not already completed) and expansion of antimicrobial coverage should be considered. Antibiotics may be broadened to include coverage for resistant gram-negative, resistant gram-positive, and anaerobic organisms. Coverage for Candida and/or for molds should be also considered.
The use of hematopoietic growth factors in patients with neutropenic fever as an adjunct to antibiotics is controversial. A 2014 Cochrane review found that use of these factors during neutropenic fever may decrease the duration of neutropenia, duration of fever, and length of hospital stay (4), but the magnitude of these effects is small, and there has been no documented improvement in survival (1, 2). These findings, coupled with the costs and side effects of these therapies, have led most guidelines (5, 6) to recommend against their routine use in the setting of neutropenic fever.
Deciding whether to prescribe antimicrobial prophylaxis during the course of neutropenia requires weighing the benefits of preventing invasive infection against the risks of medication-associated toxicity and the development of drug-resistant pathogens. Prophylaxis may be considered in the following situations (1, 2):
- Patients who are expected to have prolonged (>7 days) or profound (<100 cells/mm3) neutropenia should receive antibacterial prophylaxis with a fluoroquinolone.
- Patients undergoing allogeneic stem-cell transplantation require prophylaxis against Candida infections with a triazole or echinocandin.
- Patients receiving intensive induction or salvage chemotherapy for acute leukemia require prophylaxis against both Candida and Aspergillus infections with posaconazole.
- Antiviral prophylaxis against herpes simplex virus is given routinely in patients undergoing allogeneic or autologous stem-cell transplantation and in those undergoing induction therapy for acute leukemia.
- Antiviral prophylaxis against cytomegalovirus is generally not recommended but can be considered for patients undergoing allogeneic stem-cell transplantation and for those receiving alemtuzumab therapy.
Back to the patient
Over the next week, the patient's fever curve slowly improves. Her chest CT shows right greater than left diffuse pulmonary nodules, some of which have surrounding ground-glass opacities. Her galactomannan is positive based on her bronchoalveolar lavage specimen but negative from the serum. All cultures are negative. She is diagnosed with invasive pulmonary aspergillosis and continued on treatment-dosed voriconazole with therapeutic levels confirmed.