clinical practice guidelines Annals of Oncology 27 (Supplement 5): v111v118, 2016

doi:10.1093/annonc/mdw325

Management of febrile neutropaenia: ESMO Clinical

Practice Guidelines
J. Klastersky1, J. de Naurois2, K. Rolston3, B. Rapoport4, G. Maschmeyer5, M. Aapro6 &
J. Herrstedt7 on behalf of the ESMO Guidelines Committee*
1
Institut Jules BordetCentre des Tumeurs de lULB, Brussels, Belgium; 2St Luke's Cancer Centre, Royal Surrey County Hospital, Guildford, UK; 3M.D. Anderson Cancer
Center, Houston, TX, USA; 4Medical Oncology Centre of Rosebank, Johannesburg, South Africa; 5Department of Hematology, Oncology and Palliative Care, Ernst von
Bergmann Hospital, Potsdam, Germany; 6Multidisciplinary Institute of Oncology, Clinique de Genolier, Genolier, Switzerland; 7Department of Oncology, Odense University
Hospital (OUH), Odense, Denmark

Downloaded from http://annonc.oxfordjournals.org/ by guest on January 2, 2017

denition of febrile neutropaenia morbidity and mortality rates. Other factors having a similar
role are as follows:
Febrile neutropaenia (FN) is dened as an oral temperature of
>38.3C or two consecutive readings of >38.0C for 2 h and an advanced disease,
absolute neutrophil count (ANC) of <0.5 109/l, or expected to history of prior FN,
fall below 0.5 109/l. no antibiotic prophylaxis or granulocyte colony-stimulating
factor (G-CSF) use [III, IV],
mucositis,

clinical practice
incidence, morbidity, mortality and poor performance status and/or

guidelines
microorganisms cardiovascular disease [III, IV].
Despite major advances in prevention and treatment, FN The risk of FN and its complications increases when one or
remains one of the most frequent and serious complications of several co-morbidities are present in the patient. These consid-
cancer chemotherapy (ChT). It is a major cause of morbidity, erations will be instrumental in deciding whether a ChT-treated
healthcare resource use and compromised treatment efcacy patient should receive primary prophylaxis to decrease the po-
resulting from delays and dose reductions of ChT. Mortality tential risk of FN.
from FN has diminished steadily, but remains signicant. In the case of FN, prognosis is worst in patients with proven
Most standard-dose ChT regimens are associated with 68 bacteraemia, with mortality rates of 18% in Gram-negative and
days of neutropaenia, and FN is observed in 8 cases per 1000 5% in Gram-positive bacteraemia [for bacteraemias due to coagu-
patients receiving cancer ChT. FN is responsible for consider- lase-negative Staphylococcus (CNS) only, no attributable mortality
able morbidity as 20%30% of patients present complications has been reported] [1]. The presence of a focal site of presumed
that require in-hospital management, with an overall in-hospital infection (e.g. pneumonia, abscess, cellulitis) also makes the
mortality of 10%. The mean cost per hospitalisation in outcome worse. Mortality varies according to the Multinational
Western countries is 13 500E (15 000 US$). Association of Supportive Care in Cancer (MASCC) prognostic
There is a clear relationship between the severity of neutro- index (Table 1): lower than 5% if the MASCC score is 21, but
paenia (which directly inuences the incidence of FN) and the possibly as high as 40% if the MASCC score is <15 [2].
intensity of ChT. Currently, the different regimens are classied Positive microbiological detection rates by standard blood
as producing a high risk (>20%), an intermediate risk (10% cultures vary depending on whether or not patients have
20%) or a low risk (<10%) of FN. received prophylactic antibiotics. Overall, bacteraemia can be
It has been shown that several factors, other than ChT itself, detected in 20% of patients with FN; this obviously helps to
are responsible for increasing the risk of FN and its complica- further adjust antibiotic therapy.
tions. Among them, age plays a major role [II, III] with older It is crucial to understand that different centres experience dif-
patients having a higher risk of FN following ChT, with worse ferent patterns of frequency of causative pathogens. Consequently,
these guidelines are intended for use alongside appropriate local
*Correspondence to: ESMO Guidelines Committee, ESMO Head Ofce, Via L. Taddei 4, antimicrobial policies adapted to the epidemiology of the centre.
6962 Viganello-Lugano, Switzerland. Over the last few decades, a shift has occurred from FN asso-
E-mail: [email protected]
ciated mainly with Gram-negative bacteria to FN associated

Approved by the ESMO Guidelines Committee: October 2008, last update August with Gram-positive organisms. At the present time, most
2016. This publication supersedes the previously published versionAnn Oncol 2010; 21 centres report Gram-positive and Gram-negative bacteraemia in
(Suppl. 5): v252v256.

The Author 2016. Published by Oxford University Press on behalf of the European Society for Medical Oncology.
All rights reserved. For permissions, please email: [email protected].
clinical practice guidelines Annals of Oncology

Table 1. MASCC febrile neutropaenia risk index Assess frequency of FN associated with the planned chemotherapy regimen

Characteristics Score

Burden of illness: no or mild symptoms 5

Burden of illness: moderate symptoms 3 FN risk 20% FN risk 10-20% FN risk 10%

Burden of illness: severe symptoms 0

No hypotension (systolic BP > 90 mmHg) 5
No chronic obstructive pulmonary disease 4 Assess factors that increase the frequency/risk of FN
Solid tumour/lymphoma with no previous fungal infection 4 Age > 65 years
Other comorbities
No dehydration 3 Reassess at
Outpatient status (at onset of fever) 3 each cycle

Age <60 years 2

Define the patients overall FN risk for planned
chemotherapy regimen
Patients with scores 21 are at low risk of complications. Points attributed
to the variable burden of illness are not cumulative. The maximum
theoretical score is therefore 26 [2]. Reprinted with permission. 2000
Overall FN risk 20% Overall FN risk < 20%
American Society of Clinical Oncology. All rights reserved.
BP, blood pressure.

Prophylactic G-CSF recommended G-CSF prophylaxis not indicated

Downloaded from http://annonc.oxfordjournals.org/ by guest on January 2, 2017

50% of patients with FN, although centres that do not use
uoroquinolone prophylaxis report a predominance of Gram- Figure 1. Algorithm to decide primary prophylactic granulocyte colony-
negative bacteria. An increase in antibiotic-resistant strains has stimulating factor usage, adapted from European Organisation for Research and
been noted, such as extended spectrum -lactamase (ESBL) Treatment of Cancer guidelines. FN, febrile neutropaenia; G-CSF, granulocyte
producing Gram-negative bacteria, vancomycin-resistant en- colony-stimulating factor. Reprinted from [8], with permission from Elsevier.
terococci (VRE) and methicillin-resistant Staphylococcus aureus
(MRSA). Increasing numbers of infections with uconazole-re- tumours without signicantly affecting tumour response or
sistant Candida strains (e.g. Candida krusei and Candida glab- overall survival [I] [57].
rata) have also been reported [3]. Most guidelines recommend that G-CSF be administered
prophylactically if the risk of FN is >20% for all planned cycles
of treatment [I, A]. Classications of the risk according to the
chemoprophylaxis type of ChT have been published and updated [8]. For patients
Antimicrobials (rst non-absorbable antibiotics and later, co-tri- with an intermediate risk (10%20%), it is important to con-
moxazole) have been used for a long time for the prevention of sider the patients age and particularly any coexisting morbid-
episodes of FN in ChT-treated patients. This approach has been ities, as already mentioned [810].
somewhat successful, but has also led to the emergence of resistant An algorithm for the decisions about primary prophylactic
strains, limiting its efcacy. Since the 1990s, uoroquinolones have G-CSF use is presented in Figure 1.
been used extensively for chemoprophylaxis. Most studies have Besides this approach, G-CSF can be considered in patients
shown that uoroquinolones reduce the incidence of infection with reduced bone marrow reserve due to extensive radiother-
and, in some studies, also the infection-related mortality, but at the apy [III] or patients who are neutropaenic in the context of HIV
expense of the emergence of quinolone-resistant strains. This infection [II].
should, in the end, render the prophylaxis useless; moreover, these Recent meta-analysis of randomised, controlled trials [11] and ex-
strains jeopardise the use of uoroquinolones as a therapy of FN in perience in real-world settings [12] conrm the outstanding (>50%
low-risk patients, as will be discussed elsewhere. For all of these success) of primary prophylaxis with lgrastim or peglgrastim.
reasons, the use of antimicrobials, including uoroquinolones, With most ChT used for the treatment of common tumours, the
should be discouraged. Guidelines from the EORTC (European risk of FN is maximal during the rst course; thus, it makes sense
Organisation for Research and Treatment of Cancer) and to recommend primary prophylaxis for the patients at risk rather
American Society of Clinical Oncology (ASCO) recommend that than to systematically resort to secondary prophylaxis. Secondary
clinicians limit the use of antibacterial prophylaxis to patients at prophylaxis (i.e. G-CSF given for a course of ChT following a
high risk for FN; others recommend the mere avoidance of such course with FN) is indicated if dose reduction below threshold or
practices for the prevention of FN. The most recent update of the delay of ChT is not desirable (e.g. treatment with a curative intent).
Cochrane meta-analysis still recommended the use of ciprooxacin There are few complications associated with G-CSF adminis-
or levooxacin in cancer patients undergoing intensive ChT [4]. tration; the most common adverse effect is minor or moderate
bone pain that can usually be handled with standard analgesics.
indications for primary prophylaxis of FN
with G-CSF dose schedule, route of application of
Several meta-analyses indicate that primary prophylaxis with G-
G-CSF and peglgrastim
CSF (i.e. G-CSF administered immediately after cycle 1 of ChT) Use 5 g/kg/day of G-CSF subcutaneously (s.c.) 2472 h after
reduces the risk of FN by at least 50% in patients with solid the last day of ChT until sufcient/stable post-nadir ANC

v | Klastersky et al. Volume 27 | Supplement 5 | September 2016

Annals of Oncology clinical practice guidelines
recovery (achieving a target ANC of >10109/l is not necessary). Table 2. Initial assessment and investigations
Peglgrastim, injected s.c. as a single dose of either 100 g/kg
1 Note the presence of indwelling i.v. catheters
(individualised) or of a total dose of 6 mg (general approach), is
2 Symptoms or signs suggesting an infection focus:
considered equally effective [I, A]. The equivalent dose of lgras-
Respiratory system
tim is 5 g/kg/day for 10 days. There are no adequate data for
Gastrointestinal tract
reduced numbers or days or alternate days of G-CSF instead of
Skin
standard, neither for use on day 1 instead of on day 2. EMA/
Perineal region/genitourinary discharges
FDA approved biosimilars can be considered.
Oropharynx
Central nervous system
use of G-CSF in high-risk situations 3 Knowledge of previous positive microbiology results by checking
clinical records
The therapy of acute leukaemias, autologous and allogeneic 4 Routine investigations:
stem cell transplantations (TPLs) leads to higher risks of FN and Urgent blood testing to assess bone marrow, renal and liver
potentially lethal complications [13]. function
The incidence of FN in high-risk situations is as follows: Coagulation screen
common during autologous and allogeneic peripheral blood C-reactive protein
stem-cell (PBSC) TPLs and bone marrow TPL, during graft Blood cultures (minimum of two sets) including cultures from
failure, indwelling i.v. catheter
in 35%48% of acute myeloid leukaemia (AML) cases at diag- Urinalysis and culturea

Downloaded from http://annonc.oxfordjournals.org/ by guest on January 2, 2017

nosis and Sputum microscopy and culturea
Stool microscopy and culturea
in 13%30% during acute lymphoblastic leukaemia (ALL) in-
Skin lesion (aspirate/biopsy/swab)
duction ChT.
Chest radiograph
FN-related mortality is described as follows: 5 Further investigations (profound/prolonger neutropaenia/following
allografts)
0%10% in autologous TPL,
High-resolution chest CT (if pyrexial despite 72 h of appropriate
highly variable in allogenic TPL,
antibiotics)
80% during graft failure, Bronchoalveolar lavage
20%26% during the rst 2 months in AML and
2%10% during induction ChT of ALL. i.v., intravenous; CT, computed tomography.
a
Urinalysis, sputum and stool cultures only in case of suspected focus of
infection at these sites.
management of FN: patient education
and local policies
An initial assessment (Table 2) of circulatory and respiratory
Success in FN management requires prompt recognition of, and function, with vigorous resuscitation where necessary, should be
reaction to, potential infection. It is vital to educate outpatients to followed by careful examination for potential foci of infection.
monitor their symptoms, including body temperature, and to Signs and symptoms of infection in neutropaenic patients can
provide clear written instructions on when and how to contact the be minimal, particularly in those receiving corticosteroids, or in
appropriate service in the event of concerns. In addition, effective elderly patients who often may present with a confusional state.
written local policies are essential to ensure a rapid response Vigilance is required in patients at risk for FN who present
whenever FN is suspected. Some patients may present with FN at unwell, are hypotensive (compared with the known previous
the Emergency Department, and in this situation, clear protocols blood pressure readings), with a low-grade temperature or afeb-
must again be in place to manage these patients appropriately. rile, as they may be developing Gram-negative septicaemia, re-
The rst administration of therapy should be given in the hospital quiring prompt treatment.
within 1 h from the admission of a patient with FN. Delay in anti- Urgent full blood counts, to ascertain the ANC along with
biotic administration has been associated with signicant pro- other investigations listed in Table 2, are crucial in guiding early
longation of the hospital stay and increased mortality. management.
As already mentioned, the spectrum of infection in cancer Two sets of blood cultures from a peripheral vein and any
patients is different from place to place and changes over time; indwelling venous catheters should be taken as well as speci-
therefore, paying attention to local epidemiology is crucial [14]. mens for microbiological testing from any suspected sites of in-
fection, before the prompt institution of empirical broad-
spectrum antimicrobial therapy. Urinary tract infections have to
initial assessment and investigations
be suspected even in asymptomatic patients with a past history
A detailed history should be taken including the nature of the of such infections.
ChT given, prior prophylactic antibiotics, concomitant steroid
use, recent surgical procedures and the presence of allergies. To
guide therapy, it is important to check the clinical record for
outcome risk assessment
past positive microbiology, in particular previous presence of The vast majority of FN cases, as managed according to the al-
antibiotic-resistant organisms or bacteraemia. gorithm set out in Figure 2, respond promptly to empirical

Volume 27 | Supplement 5 | September 2016 doi:10.1093/annonc/mdw325 | v

clinical practice guidelines Annals of Oncology

Temperature > 38.3C and ANC < 0.5109/l Table 3. Key recommendations for the management of FN
Prompt assessment and vigorous
resuscitation if needed FN is observed in 1% of patients receiving ChT; it is associated with
considerable morbidity (20%30%) and mortality (10%)
FN can be effectively prevented by the use of G-CSFs; it is
Calculate MASCC score
recommended to use these agents in patients receiving
chemotherapies with a >20% risk of developing FN and in those
having serious co-morbidities and/or aged >60 years [I, A]
Patients with FN should be assessed for the risk of complications
using a validated predictive tool, such as the MASCC score [I, A]
High risk Low risk
Patients with FN at a low risk of complications can often be treated
with oral antibiotics and possibly as outpatients, if adequate follow-up
is available [I, A]
Inpatient broad spectrum Inpatient oral antibacterial Patients with FN at a high risk of complications should be hospitalised
intravenous antibacterial therapy therapy for some cases
and treated without delay with broad spectrum antibiotics; these
patients should be closely monitored for instability (pre-shock) [I, A]
Figure 2. Initial management of febrile neutropaenia. ANC, absolute neu-
trophil count; MASCC, Multinational Association of Supportive Care in
Cancer. FN, febrile neutropaenia; ChT, chemotherapy; G-CSF, granulocyte
colony-stimulating factor; MASCC, Multinational Association of

Downloaded from http://annonc.oxfordjournals.org/ by guest on January 2, 2017

Supportive Care in Cancer.
therapy, suffering no major complications. A number of instru-
ments have been developed in attempts to predict these low-risk
cases where complications are not likely. The most widely used reduction in the incidence of nosocomial infections. There is
instrument, the MASCC index, allows the clinician to rapidly also evidence to support an early discharge policy in these low-
assess, on just a clinical basis, the risk of a patient with FN. The risk cases once they have become clinically stable, symptomatic-
MASCC score has been prospectively validated in several ally better and there is evidence of fever lysis after a minimum of
studies. The criteria and weighting scores are listed in Table 1. 24 h in hospital [II, B], and provided that there is an adequate
Low-risk cases are those scoring 21. The serious medical com- understanding of the risks and that patient surveillance is avail-
plication rate in low-risk cases is estimated to be 6% and the able [1619].
mortality rate to be below 1%. If an obvious focus of infection is
apparent, antibacterials should be tailored accordingly. high-risk patients
Patients with FN who are at high risk as assessed by the MASCC
low-risk patients criteria (<21), or have high-risk features as judged by the admit-
ting doctor, should be admitted and commenced on broad-
oral therapy spectrum i.v. antibiotics, since the risk of bacterial sepsis is very
A recent review has concluded that inpatient oral antibacterial high [20].
therapy can be safely substituted for conventional intravenous
(i.v.) treatment in some low-risk FN patients, namely those who choice of i.v. antibacterial
are haemodynamically stable, Local epidemiological bacterial isolate and resistance patterns
do not have acute leukaemia or evidence of organ failure and are crucially important in determining the rst-choice empirical
do not have pneumonia, an indwelling venous catheter or therapy, since coverage for MRSA or resistant Gram-negative
severe soft tissue infection [I, A]. bacteria may be required [21]. A meta-analysis comparing
monotherapy (e.g. an anti-pseudomonal cephalosporin like cef-
Precise criteria were not dened as they varied between the trials tazidime or cefepime, imipenem, meropenem or piperacillin
reviewed. Single-agent quinolones (moxioxacin) were not in- tazobactam) with combination therapy found equivalent efcacy
ferior to combinations (quinolone with amoxicillin plus clavula- [I, A] [22, 23]. This is less clear in the subsets at high risk of pro-
nic acid), but the latter are preferred given the rise in Gram- longed neutropaenia and those with bacteraemia, where the bac-
positive FN episodes. Oral quinolone therapy should not be tericidal activity and synergistic effect of a -lactam antibiotic in
used in patients who have taken a quinolone antibacterial as combination with an aminoglycoside might be preferable;
prophylaxis. The safety of early change to oral combinations in namely, in case of Pseudomonas aeruginosa sepsis or in centres
afebrile patients after 48 h on i.v. therapy is supported in the with known intermediate susceptibility of Gram-negative bacilli
review and preferred by many physicians. Some low-risk to -lactams [3].
patients may be treated with outpatient parenteral regimens. Key recommendations about the management of febrile neu-
tropenia are summarised in Table 3.
outpatient and early discharge policies
The possibility of exclusive oral outpatient management for specic indications for alternative therapy
low-risk FN cases has become increasingly appealing on the Apart from the standard treatment with broad-spectrum anti-
grounds of the patients convenience, economy [15] and bacterial agents, there are a number of situations, in clinical

v | Klastersky et al. Volume 27 | Supplement 5 | September 2016

Annals of Oncology clinical practice guidelines
practice, that require a specic regimen. The duration of treat- response to antibacterial therapy is essential, and, in the absence
ment may vary and local antibacterial guidelines should be fol- of prompt improvement, further investigations are warranted. If
lowed in these circumstances. invasive aspergillus is suspected, a high-resolution chest
computed tomography (CT) scan should be carried out on the
central i.v. catheters. If a patient has an i.v. catheter, catheter- same day, looking for typical features such as nodules with halos
related infection (CRI) should be suspected, and blood must be or ground-glass change, and galactomannan should be
cultured from the catheter and peripherally to measure the measured in serum. If any inltrate is found, bronchoalveolar
differential time to positivity (DTTP), which is the difference in lavage should be undertaken if possible.
time between positivity of results between catheter culture and Advice from an infectious diseases (ID) specialist or clinical
peripheral blood culture. A DTTP of 2 h is a highly sensitive microbiologist is recommended, and an appropriate therapy
and specic indicator of catheter-related bacteraemia [I, A] [24]. against infection with fungi or Pneumocystis species should be
All cases of CRI in the setting of FN require decision-making instituted. The choice of antifungal agents will depend on
on the choice and duration of i.v. antibiotics, and the need for centres, individual patients and use of prior prophylactic
catheter removal. When CRI is suspected, and the patient is therapy [28].
stable, the catheter should not be removed without microbio- Therapy for presumed aspergillosis (for cases with typical
logical evidence of infection [25]. inltrates on CT) could consist of either voriconazole or liposo-
A glycopeptide such as vancomycin should be administered mal amphotericin B [I, A] [29, 30]. These antifungals can be
through the line when possible to cover Gram-positive organ- combined with an echinocandin in unresponsive disease [IV,
isms [III, A]. Teicoplanin is a useful alternative as it can be B]. A precise microbiological diagnosis is highly desirable in

Downloaded from http://annonc.oxfordjournals.org/ by guest on January 2, 2017

administered once daily as a line lock. Success in treating CRI patients suspected of invasive fungal infection, as the sensitivity
without removing the catheter depends on the pathogen isolated to various antifungal agents is variable among different species.
in the blood cultures. High-dose co-trimoxazole is the treatment of choice for sus-
In CRI due to CNS, an attempt at preserving the catheter can pected Pneumocystis infection [I, A].
be made if the patient is stable [III, B]. Catheter retention does
not have an impact on the resolution of CNS bacteraemia but is vesicular lesions/suspected viral infection. After appropriate
a signicant risk factor for recurrence in those patients in whom samples are taken, therapy with aciclovir should be initiated [I,
the catheter was retained. A]. Ganciclovir (or foscarnet) should be substituted only when
Removal of the line is indicated in the context of tunnel infec- there is a high suspicion of invasive cytomegalovirus infection
tions, pocket infections (implanted port system) [III, B], persist- [I, A] [31, 32].
ent bacteraemia despite adequate treatment, atypical
mycobacterial infection and candidaemia. With regard to line suspected meningitis or encephalitis. Lumbar puncture (if in
infections caused by S. aureus, the literature is divided. The any way possible before the institution of antibiotics) is
desire to preserve the line must be balanced against the risk of mandatory in these rare cases. Bacterial meningitis should be
metastatic spread by bloodstream seeding. The recommendation treated with ceftazidime plus ampicillin (to cover for Listeria
should be to remove the line if at all possible, while recognising monocytogenes) or meropenem [II, A]. Viral encephalitis is
that, with careful management, it might be possible to maintain treated with a high dose of aciclovir.
it for a short period. Persistent fever and bacteraemia despite ap-
propriate antibiotics are indications for line removal. cellulitis. The addition of vancomycin broadens the cover
against skin pathogens [V, D]. Linezolid and daptomycin are
pneumonia. If pneumonia in an outpatient is diagnosed either emerging alternatives to glycopeptides; however, more clinical
on clinical grounds and/or on the basis of radiological imaging, experience is needed, especially in neutropaenic patients.
antibiotic cover may be extended to treat atypical organisms
such as Legionella and Mycoplasma by adding a macrolide or a intra-abdominal or pelvic sepsis. If clinical or microbiological
uoroquinolone antibiotic to a -lactam antibiotic [V, D]. evidence of intra-abdominal or pelvic sepsis exists,
Consideration for infection with Pneumocystis jirovecii should metronidazole should be commenced [V, D], unless the patient
be given in patients who present with high respiratory rates and/ is on a carbapenem or piperacillintazobactam, which have
or desaturate readily off oxygen or on minimal exertion. adequate anaerobic coverage.
Predisposing factors include prior corticosteroid therapy, use of
immune suppressants after organ TPL and exposure to purine diarrhoea. Assessment for Clostridium difcile is needed and, if
analogues, as well as lack of reliable chemoprophylaxis with co- suspected, oral vancomycin or metronidazole treatment should
trimoxazole [26]. In high-risk patients with profound prolonged be administered [V, D].
neutropaenia and lung inltrates, early treatment with a mould-
active antifungal agent is recommended. candidiasis. Patients at risk of disseminated candidiasis are
those with prolonged neutropaenia and especially those with
lung inltrates. Patients with AML during remission induction haematological malignancies undergoing myeloablative therapy
ChT and those undergoing allogeneic haematopoietic stem cell [33]. Candidaemia can be diagnosed on blood culture; however,
TPL with prior conditioning ChT are at risk of invasive fungal cultures may take several days to become positive. Empirical
infections (namely aspergillosis) due to prolonged and initiation of antifungal therapy is recommended in patients
profound neutropaenia [27]. Frequent assessment of initial whose fever fails to respond to broad-spectrum antibiotics after

Volume 27 | Supplement 5 | September 2016 doi:10.1093/annonc/mdw325 | v

clinical practice guidelines Annals of Oncology

Review of therapy at 48 hours

Patient afebrile and Pyrexia continues

ANC 0.5109/l

Low risk High risk

On oral antibacterials: On.iv. antibacterials: Pathogen not identified: Pathogen identified: Patient deteriorating:
continue therapy consider continuing discontinue consider specific Patient stable: seek expert advice
and consider therapy with aminoglycoside, antibacterial therapy continue same therapy from ID physician or
early discharge oral antibacterials continue i.v. therapy clinical microbiologist

Figure 3. Assessment of response and subsequent management. ANC, absolute neutrophil count; i.v., intravenous; ID, infectious disease.

Downloaded from http://annonc.oxfordjournals.org/ by guest on January 2, 2017

37 days of appropriate treatment [I, A]. A CT scan of the liver Table 4. Levels of evidence and grades of recommendation
and spleen should be carried out before commencing anti- (adapted from the Infectious Diseases Society of AmericaUnited
Candida treatment, looking for typical changes. States Public Health Service Grading Systema)
First-line empirical treatment depends on what is known Levels of evidence
about the patient. Liposomal amphotericin B and an echinocan-
din antifungal such as caspofungin are appropriate rst-line I Evidence from at least one large randomised, controlled trial of
treatments if the patient has already been exposed to an azole or good methodological quality (low potential for bias) or meta-
if the patient is known to be colonised with non-albicans analyses of well-conducted randomised trials without
Candida [I, A]. Fluconazole can be given rst line provided the heterogeneity
patient is at low risk of invasive aspergillosis; local epidemio- II Small randomised trials or large randomised trials with a suspicion
of bias (lower methodological quality) or meta-analyses of such
logical data suggest low rates of azole-resistant isolates of
trials or of trials with demonstrated heterogeneity
Candida and the patient has not received an azole antifungal as
III Prospective cohort studies
prophylaxis. Once begun, antifungal treatment should be con-
IV Retrospective cohort studies or casecontrol studies
tinued until neutropaenia has resolved, or for at least 14 days in
V Studies without control group, case reports, experts opinions
patients with a demonstrated invasive Candida infection.
Specic needs for preventing other opportunistic infections are Grades of recommendation
required in patients with haematological malignancies, namely A Strong evidence for efficacy with a substantial clinical benefit,
those undergoing haematopoietic stem cell transplants [34]. strongly recommended
B Strong or moderate evidence for efficacy but with a limited clinical
benefit, generally recommended
daily follow-up and assessment of C Insufficient evidence for efficacy or benefit does not outweigh the
risk or the disadvantages (adverse events, costs, ...), optional
response D Moderate evidence against efficacy or for adverse outcome,
The frequency of clinical assessment is determined by severity generally not recommended
but may be required every 24 h in cases needing resuscitation. E Strong evidence against efficacy or for adverse outcome, never
Daily assessment of fever trends, bone marrow and renal func- recommended
tion is indicated until the patient is afebrile and has an ANC of a
0.5 109/l (Figure 3) for 24 h. Repeated imaging may be By permission of the Infectious Diseases Society of America [34].
required in patients with persistent pyrexia.
If the patient is afebrile and has an ANC of 0.5 109/l at
48 h, has low risk and no cause of infection has been found, con- units will add a glycopeptide to the regimen, while other centres
sider changing to oral antibiotics [II, A]. If the patient is at high will change the regimen to imipenem or meropenem and a gly-
risk with no cause found and is on dual therapy, aminoglycoside copeptide. This group of patients with persistent fever is at a
may be discontinued [V, D]. When the cause is found, continue high risk of serious complications, and prompt advice from an
on appropriate specic therapy [II, A]. ID physician or clinical microbiologist should be sought.
If the patient is still febrile at 48 h, but clinically stable, initial Unusual infections should be considered, particularly in the
antibacterial therapy should be continued. If the patient is clin- context of a rising C-reactive protein, with a view to proceeding
ically unstable, antibacterial therapy should be rotated or broa- to imaging of the chest and upper abdomen, to exclude probable
dened if clinical developments justify this. Some haematology fungal or yeast infection, or abscesses. When the pyrexia lasts

v | Klastersky et al. Volume 27 | Supplement 5 | September 2016

Annals of Oncology clinical practice guidelines
for >46 days, empirical initiation of antifungal therapy may be causing bacteremia in febrile neutropenia adult cancer patients in Lebanon, broad
needed [I, A]. spectrum antibiotics use as a major risk factor, and correlation with poor
prognosis. Front Cell Infect Microbiol 2015; 5: 11.
4. Gafter-Gvili A, Fraser A, Paul M, Leibovici L. Meta-analysis: antibiotic prophylaxis
duration of therapy reduces mortality in neutropenic patients. Ann Intern Med 2005; 142: 979995.
[Erratum, Ann Intern Med 2006; 144:704].
If the ANC is 0.5 109/l, the patient is asymptomatic and has 5. Clark OA, Lyman GH, Castro AA et al. Colony-stimulating factors for
been afebrile for 48 h and blood cultures are negative, antibac- chemotherapy-induced febrile neutropenia: a meta-analysis of randomized
terials can be discontinued [II, A]. controlled trials. J Clin Oncol 2005; 23: 41984214.
If the ANC is 0.5 109/l, the patient has suffered no complica- 6. Cooper KL, Madan J, Whyte S et al. Granulocyte colony-stimulating factors for
tions and has been afebrile for 57 days, antibacterials can be dis- febrile neutropenia prophylaxis following chemotherapy: systematic review and
continued except in certain high-risk cases with acute leukaemia meta-analysis. BMC Cancer 2011; 11: 404.
and following high-dose ChT when antibacterials are often contin- 7. Kuderer NM, Dale DC, Crawford J, Lyman GH. Impact of primary prophylaxis with
granulocyte colony-stimulating factor in febrile neutropenia and mortality in adult
ued for up to 10 days, or until the ANC is 0.5 109/l [II, A].
cancer patients receiving chemotherapy: a systematic review. J Clin Oncol 2007;
Patients with persistent fever despite neutrophil recovery 25: 31583167.
should be assessed by an ID physician or clinical microbiologist 8. Aapro MS, Bohlius J, Cameron DA et al. 2010 Update of EORTC guidelines for the
and antifungal therapy considered [II, A]. use of granulocyte-colony stimulating factor to reduce the incidence of
An overall algorithm for the assessment of response and sub- chemotherapy-induced febrile neutropenia in adult patients with
sequent management is proposed in Figure 3. lymphoproliferative disorders and solid tumours. Eur J Cancer 2011; 47: 832.
9. Smith TJ, Khatcheressian J, Lyman GH et al. 2006 update of recommendations

Downloaded from http://annonc.oxfordjournals.org/ by guest on January 2, 2017

for the use of white blood cell growth factors: an evidence-based clinical practice
methodology guideline. J Clin Oncol 2006; 24: 31873205.
10. Sung L, Nathan PC, Alibhai SM et al. Meta-analysis: effect of prophylactic
These clinical practice guidelines were developed in accordance
hematopoietic colony-stimulating factors on mortality and outcomes of infections.
with the ESMO standard operating procedures for clinical prac- Ann Intern Med 2007; 147: 400411.
tice guidelines development, http://www.esmo.org/Guidelines/ 11. Wang L, Baser O, Kutikova L et al. The impact of primary prophylaxis with
ESMO-Guidelines-Methodology. The relevant literature has granulocyte colony-stimulating factors on febrile neutropenia during
been selected by the expert authors. A summary of recommen- chemotherapy: a systematic review and meta-analysis of randomized controlled
dations is shown in Table 3. Levels of evidence and grades of trials. Support Care Cancer 2015; 23: 31313140.
recommendation have been applied using the system shown in 12. Mitchell S, Li X, Woods M et al. Comparative effectiveness of granulocyte colony-
Table 4. Statements without grading were considered justied as stimulating factors to prevent febrile neutropenia and related complications in
cancer patients in clinical practice: a systematic review. J Oncol Pharm Pract
standard clinical practice by the experts and the ESMO faculty.
2016; 22: 702716.
This manuscript has been subjected to an anonymous peer 13. Hmlinen S, Kuittinen T, Matinlauri I et al. Neutropenic fever and severe sepsis
review process. in adult acute myeloid leukemia (AML) patients receiving intensive chemotherapy:
causes and consequences. Leuk Lymphoma 2008; 49: 495501.
14. Nesher L, Rolston KV. The current spectrum of infection in cancer patients with
conict of interest chemotherapy related neutropenia. Infection 2014; 42: 513.
JK has declared speakers fees and consulting fees from TEVA. 15. Elting LS, Lu C, Escalante CP et al. Outcomes and cost of outpatient or inpatient
JdN has declared no potential conicts of interest. KR has management of 712 patients with febrile neutropenia. J Clin Oncol 2008; 26:
606611.
reported research support from Merck, Allergan, and JMI
16. Freifeld A, Marchigiani D, Walsh T et al. A double-blind comparison of empirical
Laboratories and participation in an advisory board for
oral and intravenous antibiotic therapy for low-risk febrile patients with neutropenia
Allergan. BR has reported advisory boards for Sandoz/Hexal, during cancer chemotherapy. N Engl J Med 1999; 341: 305311.
Amgen and Roche, research support from Sandoz/Hexal and 17. Innes H, Lim SL, Hall A et al. Management of febrile neutropenia in solid tumours
speaker's bureau for Teva, Amgen and Roche. GM has reported and lymphomas using the Multinational Association for Supportive Care in Cancer
personal fees (outside the submitted work) from Merck/MSD, (MASCC) risk index: feasibility and safety in routine clinical practice. Support Care
Astellas, Gilead, Pzer, F2G, Roche and Basilea. MA has Cancer 2008; 16: 485491.
reported consultancy for Amgen, Hospira, Pzer, Pierre Fabre, 18. Kern WV, Marchetti O, Drgona L et al. Oral antibiotics for fever in low-risk
Roche, Sandoz, Teva and honoraria for lectures at symposia for neutropenic patients with cancer: a double-blind, multicenter trial comparing
single daily moxooxacin with twice daily ciprooxacin plus amoxicillin-clavulanic
Amgen, Chugai, Hospira, Kyowa Hakko Kirin, Pierre Fabre,
acid combination therapy EORTC infectious diseases group trial XV. J Clin Oncol
Roche, Sandoz, Sano, Taiho and Teva. JH has declared that he 2013; 31: 11491156.
is a member of the rolapitant advisory board for Tesaro. 19. Pherwani N, Ghayad JM, Holle LM, Karpiuk EL. Outpatient management of febrile
neutropenia associated with cancer chemotherapy: risk stratication and treatment
review. Am J Health Syst Pharm 2015; 72: 619631.
references 20. Feld R. Bloodstream infections in cancer patients with febrile neutropenia. Int J
1. Klastersky J, Ameye L, Maertens J et al. Bacteraemia in febrile neutropenic Antimicrob Agents 2008; 32 (Suppl. 1): S30S33.
patients. Int J Antimicrob Agents 2007; 30(Suppl. 1): S51S59. 21. Montassier E, Batard E, Gastinne T et al. Recent changes in bacteremia in patients
2. Klastersky J, Paesmans M, Rubenstein EB et al. The Multinational Association for with cancer: a systematic review of epidemiology and antibiotic resistance. Eur J
Supportive Care in Cancer risk index: a multinational scoring system for identifying Clin Microbiol Infect Dis 2013; 32: 841850.
low-risk febrile neutropenic cancer patients. J Clin Oncol 2000; 18: 30383051. 22. Furno P, Bucaneve G, Del Favero A. Monotherapy or aminoglycoside-containing
3. Moghnieh R, Estaitieh N, Mugharbil A et al. Third generation cephalosporin combinations for empirical antibiotic treatment of febrile neutropenic patients: a
resistant Enterobacteriaceae and multidrug resistant gram-negative bacteria meta-analysis. Lancet Infect Dis 2002; 2: 231242.

Volume 27 | Supplement 5 | September 2016 doi:10.1093/annonc/mdw325 | v

clinical practice guidelines Annals of Oncology

23. Rolston KV, Bodey GP. Comment on: empirical antibiotic monotherapy for febrile 29. Cornely OA, Maertens J, Bresnik M et al. Liposomal amphotericin B as initial
neutropenia: systematic review and meta-analysis of randomized controlled trials. therapy for invasive mold infection: a randomized trial comparing a high-loading
J Antimicrob Chemother 2006; 58: 478; author reply: 479480. dose regimen with standard dosing (AmBiLoad Trial). Clin Infect Dis 2007; 44:
24. Seifert H, Cornely O, Seggewiss K et al. Bloodstream infection in neutropenic cancer 12891297.
patients related to short-term nontunnelled catheters determined by quantitative blood 30. Herbrecht R, Denning DW, Patterson TF et al. Voriconazole versus amphotericin
cultures, differential time to positivity, and molecular epidemiological typing with B for primary therapy of invasive aspergillosis. N Engl J Med 2002; 347:
pulsed-eld gel electrophoresis. J Clin Microbiol 2003; 41: 118123. 408415.
25. Raad I, Kassar R, Ghannam D et al. Management of the catheter in documented 31. Glenny AM, Fernandez Maulefnch LM, Pavitt S, Walsh T. Interventions for the
catheter-related coagulase-negative staphylococcal bacteremia: remove or retain? prevention and treatment of herpes simplex virus in patients being treated for
Clin Infect Dis 2009; 49: 11871194. cancer. Cochrane Database Syst Rev 2009; (1): CD006706.
26. Kovacs JA, Masur H. Evolving health effects of Pneumocystis: one hundred years 32. Torrez-Madriz G, Boucher HW. Immunocompromised hosts: perspectives in the
of progress of diagnosis and treatment. JAMA 2009; 301: 25782585. treatment and prophylaxis of cytomegalovirus disease in solid-organ transplant
27. Maschmeyer G, Carratal J, Buchheidt D et al. Diagnosis and antimicrobial therapy recipients. Clin Infect Dis 2008; 47: 702711.
of lung inltrates in febrile neutropenic patients (allogeneic SCT excluded): updated 33. van der Velden WJ, Blijlevens NM, Feuth T, Donnelly JP. Febrile mucositis in
guidelines of the Infectious Diseases Working Party (AGIHO) of the German Society haematopoietic SCT recipients. Bone Marrow Transplant 2009; 43: 5560.
of Hematology and Medical Oncology (DGHO). Ann Oncol 2015; 26: 2133. 34. Dykewicz CA. Summary of the guidelines for preventing opportunistic infections
28. Marr KA, Schlamm HT, Herbrecht R et al. Combination antifungal therapy for among haematopoietic stem cell transplant recipients. Clin Infect Dis 2001; 33:
invasive aspergillosis: a randomized trial. Ann Intern Med 2015; 162: 8189. 139144.

Downloaded from http://annonc.oxfordjournals.org/ by guest on January 2, 2017

v | Klastersky et al. Volume 27 | Supplement 5 | September 2016