Back to the febrile neutropenia management series.
Two questions frequently arise during the course of management, particularly with regards to febrile neutropenia (FN) in patients with haematological malignancies where the duration of neutropenia may be very long:
- In the presence of neutropenia, can there be “de-escalation” of antibiotic therapy if a sensitive bacterium is cultured (i.e. a patient with FN on IV piperacillin/tazobactam that was found to have pan-sensitive Escherichia coli bacteraemia – can the piperacillin/tazobactam be switched to IV ceftriaxone or even IV ampicillin)?
- If initial cultures are negative and the fever has resolved, how long should antibiotics be continued and can they be stopped prior to resolution of neutropenia?
The first question is intuitively easier to address although the supporting studies (specifically for FN) are scarce. The answer at this point in time is yes. This question was specifically posed in the European guidelines for empirical antibacterial therapy for febrile neutropenic patients published in 2013 (the European Conference on Infections in Leukemia took place in 2011) and the recommendation to de-escalate antibiotic therapy if a plausible pathogen was cultured was given an A1 rating (with the evidence inferred from intensive care unit and healthcare-associated pneumonia studies).
In clinical practice, however, this can be psychologically difficult. We attempted to implement this at the university hospital in Singapore and the data are captured in the febrile neutropenia database, which regrettably we have not published. The psychological barrier – for both infectious diseases physicians and haematologists – of switching from a broad-spectrum antibiotic to ceftriaxone or ampicillin while the patient remains profoundly neutropenic (and in some cases still febrile) is not to be underestimated. Nevertheless, we were successful with de-escalation for the most part and no patients deteriorated as a direct consequence.
The second question is actually harder to address in clinical practice. Much of the concern for stopping antibiotics in adult patients with neutropenia stem from an old 1979 observational study (with small patient numbers) that found significantly higher infections and mortality in those patients where antibiotics were stopped during the neutropenic period (after resolution of initial fever). Nearly 2 decades later, neutropenic children considered at low risk of bacterial infections either had antibiotics stopped at Day 3 or continued in an open randomised study, and stopping antibiotics was shown to be as safe as continuing them. Subsequent studies in both children and adults appeared to confirm the later findings.
The European recommendation (for adult patients with haematological malignancies) therefore was to discontinue antibiotics in hemodynamically stable patients after 72 hours or more of IV antibiotic therapy and 48 hours of fever resolution, regardless of neutrophil count. The patient should be kept hospitalised and under observation for a further 24-48 hours if neutropenic, however, with antibiotics restarted in the event of fever recurrence. The US (IDSA) guidelines are far more circumspect, suggesting antibiotic discontinuation only in patients with low-risk FN and evidence of marrow recovery (again after at least 3 days of empirical antibiotics).
Stopping antibiotics while patients with haematological malignancies are neutropenic is psychologically even more difficult than de-escalating antibiotics. The European recommendation was recently challenged by a French group reporting their aborted attempt at implementing antibiotic stoppage: of 7 patients, 4 recovered (with rapid recovery from neutropenia) whereas 3 had rapid recurrence of fever for which one required ICU care for severe bacterial sepsis. The European experts defended their recommendation, but one can imagine that a significant number of physicians will remain unconvinced until stronger evidence is available (and perhaps even then).
Our own experience with stopping antibiotics in patients with prolonged neutropenia was not particularly rosy (again unpublished). Fever recurred at a much higher frequency compared to those that were continued on some form of antibiotic coverage, especially broad-spectrum IV antibiotics, although none died or required ICU care as a direct consequence. Some kind of cost comparison study should be performed to determine if the cost of subsequent fever work-up and initiation of broad-spectrum antibiotic therapy (along with the potential risk of deterioration as in the French experience) will outweigh the cost of continuing broad-spectrum antibiotics in a larger group of patients. We subsequently took the more expedient (and psychologically easier) path of continuing antibiotics in these patients with prolonged neutropenia, albeit with de-escalation to narrower-spectrum antibiotics (including oral antibiotics such as amoxicillin/clavulanate and ciprofloxacin) as far as possible.
- De-escalation of antibiotics if plausible drug-susceptible organisms are cultured is safe and strongly recommended, even if the antibiotics in question have no activity against Pseudomonas aeruginosa and the patients remain neutropenic.
- Stopping antibiotics entirely while the patient remains neutropenic is controversial in the local setting. Our experience has not been entirely positive, although the evidence base (observational single centre audit) is not strong. De-escalating to narrower-spectrum antibiotics is a possible option. Further studies should be performed here.