Febrile Neutropenia - Treatment

Treatment

               The management of the febrile neutropenia syndrome has evolved progressively over the years. In the 1960s, the initiation of broad-spectrum antibiotics at the very onset of fever in neutropenic patients, decreased dramatically the mortality. This concept of empiric therapy is still one of the major elements of standard care of the febrile neutropenia nowadays. Until the 1980s, a combined therapy of a ß-lactam and an aminoglycoside has prevailed as the regimen of choice for empiric therapy, justified by a predominance of Gram-negative bacteria causing severe infection in neutropenic patients. Moreover, the available ß-lactam antibiotics at this period had a modest intrinsic activity against P. aeruginosa. However, since the mid 1980s, and in parallel to the shift from a predominance of Gram-negative to that of Gram-positive bacteria causing bacteremia in neutropenic patients, several comparative studies were able to show equivalent efficacy between a ß-lactam alone and its combination with an aminoglycoside. In fact, the introduction of new ß-lactam drugs with better intrinsic activity against Gram-negative bacteria in general, and against P. aeruginosa in particular, contributed to decrease the role of combination with aminoglycosides and its potential for nephrotoxicity and ototoxicity. More recently, since the mid 1990s, the attitude consisting in extra safety measures with hospitalization and intravenous antibiotic therapy for all neutropenic patients began to wane in favor of a risk-adapted therapy based on prediction rules of complication rates during the course of a febrile neutropenic episode. Refinement of selection of patients at low-risk of complications was achieved and oral antibiotic therapy with early discharge was applied to this low-risk population. Thus, the risk categorization of the febrile neutropenic episode allows to determine if the patient should be kept in the hospital and treated intravenously, or if he can receive oral antibiotics and be discharged early. Additional key elements are to be addressed for those who receive intravenous empiric antibiotics in order to know which ß-lactam to choose, in monotherapy or in combination with aminoglycosides or other antibiotics (Figure 2).

Figure 2: Factors that Influence the Choice of Empiric Therapy

 GN-ESBL : Gram-negative bacilli producing extended-spectrum ß-lactamases

 MRSA : Methicillin-resistant Staphylococcus aureus

 Pen-R : Penicillin-resistant

 AB(s) : Antibiotic(s)

 VRE : Vancomycin resistant enterococci

 FUO : Fever of unknown origin

    Treatment of Patients at Low-Risk

Once a febrile neutropenic episode has occurred a risk-assessment of complications is undertaken. For this purpose, the MASCC index scoring system is the most widely used and allows selecting low-risk patients with good sensitivity and specificity. Once the patient is categorized as being at low-risk, the next question then is whether this patient is eligible for oral therapy. In fact, several conditions should be present including the absence of severe oral mucositis, diarrhea and vomiting. Two large, prospective randomized trials compared oral and intravenous antibiotics in patients with febrile neutropenia at low risk of complications, both using amoxicillin-clavulanate plus ciprofloxacin for oral therapy and both showed equivalent efficacy and safety between the oral and the intravenous arms. Based on these results, the association of amoxicillin-clavulanate plus ciprofloxacin has been recommended for oral therapy providing that patients have not received fluoroquinolone prophylaxis. Patients who are allergic to penicillin may benefit from an alternative regimen associating ciprofloxacin and clindamycin although creatinine elevation has been reported with this regimen in neutropenic patients. Two small studies evaluated new-generation fluoroquinolone monotherapy for oral empiric therapy of low risk febrile neutropenic patients. The first one included 54 patients who received oral moxifloxacin 400 mg once daily with a response rate of 91 %. The second one enrolled 43 patients treated with oral gatifloxacin 400 mg once daily and showed a response rate of 95 %. Although promising, these results must be confirmed by larger randomized trials. For patients at low-risk who received intravenous antibiotics for temporary reasons, a step-down strategy with a rapid shift to oral antibiotics may be applied. Outpatient management for patients with febrile neutropenia at low-risk, with oral antibiotic therapy has been evaluated recently in a randomized trial which compared oral antibiotics, followed by early discharge after a 24 h observation period, with in-patient intravenous antibiotics until resolution of the febrile neutropenic episode. The efficacy with both strategies was equivalent and the readmission rate in the outpatient arm was low. Whether some patients at low-risk would benefit from immediate discharge, without a period of observation is not yet established and the safety of such a procedure needs to be carefully assessed before it could be generalized.

    Treatment of Patients at High-Risk

The standard care of febrile neutropenia at high-risk of complications includes in-patient management with intravenous broad-spectrum antibiotics. A ß-lactam agent active against Gram-negative bacteria including P. aeruginosa is the corner stone around which the whole management strategy is build up. Although detailed global and updated epidemiological data are lacking in the neutropenic population, it is more likely that local institution epidemiology would be more appropriate for the selection of initial ß-lactam agents for empiric therapy. Among the different ß-lactams, the highest percentage of susceptibility of P. aeruginosa is observed for piperacillin-tazobactam and meropenem although no one single ß-lactam delineates 100 % activity against P. aeruginosa anymore. Another epidemiologic feature to be taken into account, is the rising incidence of extended spectrum ß-lactamase (ESBL)-producing Gram-negative bacilli, especially E. coli and Klebsiella spp., in the neutropenic patients. Among the few ß-lactam agents that are suitable for empiric therapy of febrile neutropenia, it is important to make a distinction between molecules such as ceftazidime, cefepime or piperacillin-tazobactam among which the probability of cross-resistance is high, and the carbapenems such as imipenem or meropenem which retain activity in case of emergence of resistance to other ß-lactams. Based on these considerations, it is recommended to use a strategy that differentiates between first-line therapy based on molecules such as cefepime or piperacillin-tazobactam, and second-line therapy constituted of the carbapenems. This allows keeping a valuable alternative in case of emergence of resistance. However, such a strategy can be implemented only if the baseline incidence of ESBL-producing Gram-negative bacilli is low. In fact, carbapenems are the most active drugs against these bacteria, and reduced the mortality in ESBL-producing K. pneumoniae bacteremia. Among the first-line ß-lactam agents, there are some theoretical advantages of piperacillin-tazobactam and cefepime over ceftazidime. These include a better activity against streptococci and methicillin-susceptible S. aureus with less need for the addition of a glycopeptide. Furthermore, less induction and decreased emergence of ESBLs, are reported with cefepime and piperacillin-tazobactam.

Several randomized comparative trials have assessed the potential of each the ß-lactams, namely ceftazidime, cefepime, piperacillin-tazobactam, imipenem, meropenem and aztreonam, for empiric therapy of primary episodes of febrile neutropenia at high-risk. No significant differences in response rates or mortality were observed in these individual trials. So, the clinical evidence-based medicine assessment of these different ß-lactams will result in a strong recommendation with good evidence to support their use. But, the continuously changing microbial distribution and the emergence of new mechanisms of resistance that occurred after the completion of many of these studies, contribute to obscure the interpretation of results and their relevance to the actual epidemiology. Nevertheless, a recent meta-analysis focusing on response rate in the different comparative trials of empiric therapy of febrile neutropenia, showed superiority of the carbapenems and piperacillin-tazobactam over ceftazidime. In some situations, specific anti-anaerobic coverage is indicated. These include severe mucositis or gingivitis, typhlitis, peri-anal and allogeneic HSCT where up to 17 % of bloodstream infections were anaerobes. Piperacillin-tazobactam and carbapenems cover the majority of anaerobes, otherwise, for cephalosporins or aztreonam, metronidazole should be added. Although this is not a current first choice, in case of penicillin-allergy mediated by IgE where the risk of anaphylaxis is important, aztreonam combined with a glycopeptide is an acceptable alternative.

    Is There Still a Role for Combination Therapy of a ß-lactam Plus an Aminoglycoside for Empiric Therapy?

This question has been addressed in two recent meta-analyses which reviewed the studies that compared monotherapy with combination therapy. The conclusion was that of no advantage of the combination with an excess of toxicity. However, it should be mentioned that mortality is highest in the subgroup of febrile neutropenic patients with sepsis, varying between 18 and 40 % as compared with 2.8 % for non-septic febrile neutropenic patients. If any benefit is to be expected from combination therapy, it is within this subgroup that would occur. But, no single study targeted specifically such patients. Thus, if for the majority of patients with febrile neutropenia, a combination therapy is not justified by the existing data, in the subgroup of patients with sepsis and high mortality, the benefits of combining a ß-lactam with an aminoglycoside may outweigh the risk of toxicity. Another important question has been the addition of a glycopeptide to the initial regimen. In fact, among the Gram-positive pathogens that cause infection or bacteremia during neutropenia, very few cause a fulminant infection course with significant morbidity and mortality. These include viridans streptococci, S. pneumoniae and S. aureus, which need optimal coverage upfront. Again, the local epidemiology and the penicillin or methicillin-resistance of these organisms, which is highly variable between institutions, plays an important role. A recent Cochrane review of 13 randomized trials comparing the addition of an anti-Gram-positive antibiotic to the initial empiric regimen, did not show any benefit in reducing treatment failure, all cause mortality or Gram-positive superinfections. However, there are circumstances in which a glycopeptide or another antibiotic active against resistant Gram-positive bacteria, should be added up-front. These include patients who are already known to be colonized by MRSA, if MRSA is endemic in the institution and in several skin infections that could be caused by resistant staphylococci as folliculitis, furonculosis and periporth cellulitis and also if penicillin-resistant viridans streptococci are prevalent in the institution.

    When and How to Modify Initial Empiric Therapy in Persistently Febrile Neutropenic Patients?

Few studies have specifically addressed this question. According to the study by Cometta et al  conducted by the anti-infection group of the EORTC, the mean time for defervesence on piperacillin-tazobactam monotherapy was 5 days. Therefore, for patients without hypotension or sepsis and in whom no resistant pathogen is isolated, it seems reasonable to wait until day 5 before modification. On the contrary, for those who develop signs of sepsis, hypotension or other early findings of deterioration, at any time, a shift to non-cross resistant antibiotic is recommended. After day 5, those who remain febrile and had no non-infectious cause of fever, may benefit from a shift in ß-lactam coverage i.e. from a first-line treatment with piperacillin-tazobactam or cefepime to a carbapenem. Simultaneously, a thorough investigation including chest and sinus CT scan, galactomannan test and other diagnostic tests for viral or parasitic infections, should be undertaken. Algorithms 1, 2, 3 and 4 provide general lines of management according to specific situations of febrile neutropenia.

Algorithm 1:  Approach to the Presence of Clinical Signs of Febrile Neutropenia

Febrile neutropenia Presence of specific clinical sign

                      ↓                                                                         ↓                                                    

          Yes                                                             No (see algorithms 2-4)

          ↓

Adapted empirical treatment               

Algorithm 2:  Approach of Empiric Antibiotic Therapy for Febrile Neutropenia by MASCC score

Algorithm 3:  Approach to Febrile Neutropenia with Stable Hemodynamic Status

Algorithm 4:  Approach to Febrile Neutropenia with Unstable Hemodynamic Status

Antifungal Empiric Therapy

               The rationale for empiric antifungal therapy in persistently febrile neutropenic patients, is based on the fact that early diagnosis of invasive fungal infections is difficult to establish and the mortality is increased by delay in adequate therapy. The concept of empiric antifungal therapy in neutropenic patients with persistent fever despite broad-spectrum antibiotics, has been introduced following two studies with a limited number of patients, comparing amphotericin B with placebo and showing a decrease in fungal infection-related mortality in patients receiving amphotericin-B deoxycholate. These studies were accomplished at a period where antifungal prophylaxis was ineffective and CT-scans and ELISA galactomannan tests on PCR for Aspergillus detection were inexistent. Is this concept still true today? We just don’t know, because all the large studies that have been done thereafter addressing the question of empiric antifungal therapy were not placebo controlled. With this in mind, we see in the latest trials that the true rate of failure of empiric antifungal therapy which is the development of a breakthrough fungal infection and the absence of cure of base line fungal infection is 6.9 % for liposomal amphotericin-B, 3.6 % for voriconazole and 7.7 % for caspofungin. The need for empirical antifungal therapy during neutropenia has decreased secondarily to the implementation of more effective antifungal prophylaxis. In one of the most recent trials using posaconazole prophylaxis in neutropenic patients, the need for empiric therapy was only of 27 %. The next question is whether our new imaging techniques and new laboratory methods are able to detect invasive fungal infections early enough and whether they can serve to build-up a pre-emptive or a diagnostic-driven strategy in order to save unnecessary costs and drug exposure, without increase in mortality. One such a pre-emptive study by Maertens et al, based on positive ELISA galactomannan test or suggestive infiltrate on chest CT-scan, showed a reduction of 78 % in the use of empiric antifungal drugs. Another study by Cordonnier et al comparing prospectively the empiric with the pre-emptive strategy in high-risk febrile neutropenic patients showed again a decrease of 20 % in the use of antifungal drugs with the pre-emptive strategy, a similar mortality rate but surprisingly similar median medication costs. Thus, the pre-emptive or diagnostic-driven strategy is able to reduce the rate of antifungal overtreatment but the safety and cost effectiveness should be confirmed by large comparative prospective trials.

Adjunctive Therapy

               The recovery of neutrophils in patients with severe infection and hematological malignancy plays a major role in resolution and survival. In order to palliate a temporary deficit in neutrophils, a logical approach was the development of transfusion of donor neutrophils as adjunctive therapy to antibiotics and antifungals, in neutropenic patients with refractory infections. Although initial clinical successes were reported between 1970s and beginning 1980s, ganulocyte transfusion adjunctive therapy declined progressively due to several reasons. In fact, adverse effects such as allo-immunization, CMV transmission, GVHD and pulmonary reactions, have been reported. Besides, the availability of more advanced antimicrobials, the time and cost consuming procedure and a marginal effect demonstrated in randomized trials, all together contributed to this decline. Moreover, the lack of standard optimal granulocyte dose required daily (>1010 cells) and storage conditions and duration (8-24h) add to the uncertainty of study results. More recently, a revival in granulocyte transfusion use has occurred as donors were stimulated more efficiently by G-CSF at 5 to 10 µg/kg. Higher yields and improved phagocytosis and killing of neutrophils with prolonged intravascular survival are obtained after stimulation. However, the clinical efficacy of this new generation of granulocyte transfusion is still limited to case reports and small non randomized series. A recent meta-analysis of 8 randomized controlled trials involving granulocyte transfusions, given therapeutically to neutropenic patients showed inconclusive evidence to support or refute the generalized use granulocyte transfusion therapy. Thus, in the absence of definitive data, it is reasonable to provide granulocyte transfusions for neutropenic patients with hematological malignancy and documented bacterial or invasive fungal infection, not controlled by adequate antimicrobial therapy.

               The role of G-CSF and GM-CSF as adjunct therapy to antimicrobials in febrile neutropenic patients is controversial. Individual small studies have generated conflicting results. A recent meta-analysis on 13 controlled trials comparing CSFs plus antibiotics versus antibiotics alone for the treatment of established febrile neutropenia, showed a benefit in terms of length of hospitalization and time to neutrophil recovery, in favor of CSFs, but no advantage on overall mortality and a trend toward a decrease in infection-related mortality. This latter effect needs further investigation.