Current antimicrobial practice in febrile neutropenia across Europe and Asia: the EBMT Infectious Disease Working Party survey

The aim of this survey was to summarize the current antimicrobial practice in febrile neutropenia and the presence of key aspects of antimicrobial stewardship. A questionnaire was sent to 567 centers, and complete responses were obtained from 194 (34.2%). Fluoroquinolone and co-trimoxazole prophylaxis are used in 57.1% and 89.1%, respectively. In 66.4%, the first-line empirical therapy is piperacillin/tazobactam, whereas 10.9% use carbapenems. Empirical combination therapy is used in stable patients without history of resistant pathogens in 37.4%. De-escalation to monotherapy is performed within 3 days in 35.3% and after 10 days in 19.1%. Empirical addition of a glycopeptide is performed when fever persists more than 2–3 days in 60.8%. Empirical escalation to a broader spectrum agent is performed when fever persists more than 3–5 days in 71.4%. In case of positive blood cultures with a susceptible pathogen and uncomplicated presentation, 76.7% of centers de-escalate and 36.6% discontinue before neutrophil recovery. In fever of unknown origin with uncomplicated presentation, 54.1% of centers de-escalate and 49.5% discontinue before neutrophil recovery. Recommendations put forward in the ECIL guidelines are not widely implemented in clinical practice. Specific problems include overuse of carbapenems and combination therapy and unjustified addition of glycopeptides without further de-escalation or discontinuation.


Introduction
Febrile neutropenia remains a challenging complication in hematopoietic stem cell transplant (HSCT) recipients. Although in more than half of cases an underlying infection cannot be identified, prompt administration of empiric broad-spectrum antibiotic therapy is standard of care as it decreases morbidity and mortality [1][2][3].
The worldwide presence of rising antimicrobial resistance has been confirmed in hematology patients and HSCT recipients [4,5]. This led the European Conference on Infections in Leukemia (ECIL) to issue guidelines suggesting strategies of de-escalation and discontinuation of broad-spectrum antibiotic therapy under certain conditions [6]. The safety of de-escalation/discontinuation strategies prior to neutrophil recovery in high risk patients has been confirmed in several publications [7][8][9]. A recent critical appraisal on the use of fluoroquinolone prophylaxis (FP) concluded that the possible benefits on lowering the rate of bloodstream infections should be weighed against its impact in terms of toxicity and risk of increased colonization/infection with fluoroquinolone or multidrug resistant strains [10].
Implementation of the available guidelines is strongly affected by many factors, including local microbial epidemiology and resistance patterns, presence of rapid reporting of microbiological results, experience in management of infectious diseases, and historical antibiotic prescription habits [11]. The aim of this survey was to assess current general practice patterns with respect to antibiotic prophylaxis and empiric antibiotic treatment in febrile neutropenia, as well as the presence of key aspects of antimicrobial stewardship in EBMT centers.

The questionnaire
In August 2017, a total of 567 registered EBMT member centers in 57 countries were invited to complete the questionnaire and return it to the EBMT Data Center.
The questionnaire was designed to capture sufficient information to assess current practices in febrile neutropenia and gain an accurate impression on implementation of recent guidelines. It consisted of 54 questions (Supplementary Survey) and was divided into five sections, which included general information of the center, presence of key aspects of antimicrobial stewardship, policies on antibiotic prophylaxis, policies on empiric antibiotic therapy, and implementation of de-escalation/discontinuation strategies.

Statistical analysis
Descriptive analysis was performed for all variables with absolute and percentage frequencies being reported. If applicable, percentage of missing values was reported and rates were calculated for number of valid responses as denominator.
A statistical comparison was performed to detect differences in the proportion of responses according to center characteristics such as performing autologous or allogeneic transplants, treating adults or children, geographical region and involvement of infectious diseases or microbiology departments (IDM) in writing guidelines and/or decisionmaking. Categorical variables were compared using chisquare or Fisher's exact test if appropriate. Continuous variables were compared using Mann-Whitney test. A p value ≤ 0.05 was considered statistically significant.

Implementation of de-escalation/discontinuation strategies
In 35.3% (24/68) of centers using empirical combination therapy de-escalation to monotherapy is performed within 3 days, whereas in 19.1% (13/68) combination therapy is continued for more than 10 days ( When confronted with patients with probable/proven pulmonary aspergillosis, either with uncomplicated or severe presentation, who became afebrile on antifungal therapy, centers discontinue antibiotics prior to neutrophil recovery in 22.4% (41/183) and 18.0% (33/183), respectively.
Whereas most centers generally continue antibiotic therapy for 7-14 days, a third of centers continue treatment until the end of neutropenia in any given scenario. In centers where IDM are involved in decision-making, antibiotics are more frequently discontinued before neutrophil recovery There were no significant differences in treatment duration in relation to the center's policy on the use of FP, combination therapy, or carbapenems in first line, empirical association of a glycopeptide or empirical escalation to a broader spectrum agent.

Discussion
The management of febrile neutropenia in HSCT recipients remains challenging as there are many possible infectious and noninfectious causes for fever in these patients. In view of rising antimicrobial resistance in hematology patients, ECIL issued guidelines in 2011 introducing the concept of escalation/de-escalation of empirical therapy and suggesting discontinuation of broad-spectrum antibiotic therapy under certain conditions, which has been confirmed safe in several This table summarizes all valid responses from centers on questions concerning Implementation of de-escalation/discontinuation strategies. recent publications [4][5][6][7][8][9]. Compliance with these recommendations has never been assessed and this survey was performed to quantify their current implementation rate in EBMT centers. In the survey there were no specifications made about circumstances or timing in the transplant journey, but typically questions on febrile neutropenia involve mainly pre-engraftment period. The rate of response to this questionnaire was similar to prior surveys [12]. It demonstrated important discrepancies between guidelines and practices that should be addressed in order to optimize antimicrobial therapy and diminish development of resistance. Local definitions for fever, neutrophil cutoff and interpretation of clinical status of the patient (stable versus severe presentation) were used. This might cause some limitations to the interpretation of the data, but on the other hand represents the real world situation where definitions and practices might differ between centers and even between single clinicians within a center. ECIL guidelines recommend an escalation approach, meaning monotherapy with noncarbapenem beta-lactams in stable patients without history of colonization/infection with resistant bacteria. Still, a third of centers use empirical combination therapy in first line and 10% use carbapenems. This finding does not correlate with geographical location and cannot be explained solely by epidemiological presence of multidrug resistant bacteria. Centers using empirical combination therapy and/or carbapenems often did not define themselves as using a de-escalation approach, indicating the need for education on escalation/de-escalation principles.
Streamlining of initial combination therapy is recommended after 72 h, including discontinuation of combination therapy if resistant bacteria were not cultured. However, only a third of centers using combination therapy, de-escalate within 3 days and SE centers are more likely to extend combination therapy for more than 10 days. Although timing of de-escalation depends on speed of antimicrobial susceptibility testing, the large majority of centers (79.7% of total and 84.1% of SE centers) noted swift reporting of resistance figures. In case of positive blood cultures with a susceptible pathogen, many centers do not de-escalate, indicating clinicians may not feel confident enough in results of susceptibility testing and/or fear a decline after de-escalation.
ECIL guidelines emphasize that there is no need to escalate treatment in stable patients with persisting fever. This recommendation is supported by previous clinical trials which noted that the median time to defervescence in high risk neutropenic patients is 5 days [13]. Nevertheless, more than half of centers add a glycopeptide (60.8%) and/or escalate to a broader spectrum agent (71.4%) empirically and the minority of these centers de-escalate later on. These empirical additions/escalations are performed significantly more frequently in centers from SE Europe and Asia, centers using FP and centers without IDM involvement.
The duration of empirical treatment is a debatable issue. IDSA and ESMO guidelines recommend continuing antibiotic therapy until neutrophil recovery [14,15]. ECIL guidelines recommend considering treatment discontinuation after 72 h or later in hemodynamically stable patients with FUO who are afebrile for at least 48 h, irrespective of neutrophil count or expected duration of neutropenia [6]. However, results of this survey show that in case of FUO, discontinuation of antibiotic therapy within 7 days is only performed in 17.6% and one-third of centers continue treatment until neutrophil recovery. Centers from SE Europe and Asia are less likely to discontinue antibiotic therapy before neutrophil recovery.
Although FP is currently recommended by several guidelines, the benefits should be weighed against the risks [10]. Our survey demonstrated that half of centers provide FP, significantly more frequently in centers treating adults and in SE Europe and Asia. Reduced value of FP is expected in the presence of the higher resistance rates reported in the southern regions and one would expect changes in practices accordingly. However, the results of our survey show the opposite, which can be due to historically higher rates of FP use in these regions. Centers using FP are more likely to empirically add a glycopeptide and/or escalate to a broader spectrum agent in stable patients with persistent fever. Nonetheless these centers are not more likely to de-escalate or discontinue the empirical antibiotic treatment.
Centers from SE Europe more frequently use FP, empirically add a glycopeptide and/or escalate to a broader spectrum agent in stable patients with persistent fever and continue combination therapy for a longer period of time. On the other hand they are less likely to de-escalate or stop antibiotic therapy prior to neutrophil recovery. These practices may result from higher resistance rates [4,5], but can also be the cause of these resistance rates. Nonetheless, they create a vicious circle of higher antibiotic consumption and higher resistance rates, which can only be broken by implementing antimicrobial stewardship.
Most key aspects of antimicrobial stewardship are available in a great majority of centers, although only around half are advised by IDM when writing guidelines or taking decisions on antimicrobial treatment. This is an important finding, as empirical escalation to a broader spectrum agent in stable patients with persistent fever was performed less frequently and de-escalation/discontinuation rates were significantly higher in centers with IDM involvement. This emphasizes the importance of collaboration between the hematology departments and their IDM counterparts to improve antimicrobial stewardship as a first step towards lowering resistance rates.

Conclusion
Recommendations put forward in the ECIL guidelines are not widely implemented in clinical practice throughout EBMT centers, less frequently in SE Europe and Asia, which may correlate with higher resistance rates. Specific problems include overuse of combination therapy and carbapenems and unnecessary addition of glycopeptides without further de-escalation. Despite growing evidence on the safety of treatment discontinuation, a large proportion of centers continue antibiotic therapy until neutrophil recovery. Education on escalation and de-escalation practices, dissemination of the guidelines and available evidence can lead to a change in practice, which is important to diminish growing resistance rates. This should be accompanied by continuous monitoring of resistance patterns and clinical patient outcome data.