Pre‐emptive antifungal therapy versus empirical antifungal therapy for febrile neutropenia in people with cancer

Abstract

Background

Intensive cytotoxic chemotherapy for people with cancer can cause severe and prolonged cytopenia, especially neutropenia, a critical condition that is potentially life‐threatening. When manifested by fever and neutropenia, it is called febrile neutropenia (FN). Invasive fungal disease (IFD) is one of the serious aetiologies of chemotherapy‐induced FN. In pre‐emptive therapy, physicians only initiate antifungal therapy when an invasive fungal infection is detected by a diagnostic test. Compared to empirical antifungal therapy, pre‐emptive therapy may reduce the use of antifungal agents and associated adverse effects, but may increase mortality. The benefits and harms associated with the two treatment strategies have yet to be determined. 

Objectives

To assess the relative efficacy, safety, and impact on antifungal agent use of pre‐emptive versus empirical antifungal therapy in people with cancer who have febrile neutropenia.

Search methods

We searched CENTRAL, MEDLINE Ovid, Embase Ovid, and ClinicalTrials.gov to October 2021.

Selection criteria

We included randomised controlled trials (RCTs) that compared pre‐emptive antifungal therapy with empirical antifungal therapy for people with cancer.

Data collection and analysis

We identified 2257 records from the databases and handsearching. After removing duplicates, screening titles and abstracts, and reviewing full‐text reports, we included seven studies in the review. We evaluated the effects on all‐cause mortality, mortality ascribed to fungal infection, proportion of antifungal agent use (other than prophylactic use), duration of antifungal use (days), invasive fungal infection detection, and adverse effects for the comparison of pre‐emptive versus empirical antifungal therapy. We presented the overall certainty of the evidence for each outcome according to the GRADE approach.

Main results

This review includes 1480 participants from seven randomised controlled trials. Included studies only enroled participants at high risk of FN (e.g. people with haematological malignancy); none of them included participants at low risk (e.g. people with solid tumours). 

Low‐certainty evidence suggests there may be little to no difference between pre‐emptive and empirical antifungal treatment for all‐cause mortality (risk ratio (RR) 0.97, 95% confidence interval (CI) 0.72 to 1.30; absolute effect, reduced by 3/1000); and for mortality ascribed to fungal infection (RR 0.92, 95% CI 0.45 to 1.89; absolute effect, reduced by 2/1000). Pre‐emptive therapy may decrease the proportion of antifungal agent used more than empirical therapy (other than prophylactic use; RR 0.71, 95% CI 0.47 to 1.05; absolute effect, reduced by 125/1000; very low‐certainty evidence). Pre‐emptive therapy may reduce the duration of antifungal use more than empirical treatment (mean difference (MD) ‐3.52 days, 95% CI ‐6.99 to ‐0.06, very low‐certainty evidence). Pre‐emptive therapy may increase invasive fungal infection detection compared to empirical treatment (RR 1.70, 95% CI 0.71 to 4.05; absolute effect, increased by 43/1000; very low‐certainty evidence). Although we were unable to pool adverse events in a meta‐analysis, there seemed to be no apparent difference in the frequency or severity of adverse events between groups.

Due to the nature of the intervention, none of the seven RCTs could blind participants and personnel related to performance bias. We identified considerable clinical and statistical heterogeneity, which reduced the certainty of the evidence for each outcome. However, the two mortality outcomes had less statistical heterogeneity than other outcomes.

Authors' conclusions

For people with cancer who are at high‐risk of febrile neutropenia, pre‐emptive antifungal therapy may reduce the duration and rate of use of antifungal agents compared to empirical therapy, without increasing over‐all and IFD‐related mortality; but the evidence regarding invasive fungal infection detection and adverse events was inconsistent and uncertain. 

Plain language summary

Pre‐emptive versus empirical antifungal therapy in people with cancer who have febrile neutropenia

Intensive chemotherapy for people with cancer can cause severe and prolonged cytopenia (lower‐than‐normal number of blood cells), especially neutropenia (lower‐than‐normal white blood cells, which help to fight infection), a critical condition that is potentially life‐threatening. When a person has both fever and neutropenia, it is called febrile neutropenia (FN). Invasive fungal disease (IFD; an infection caused by a fungus) is one of the serious causes of chemotherapy‐induced FN. 

There are two treatment strategies in this situation. In empirical antifungal therapy, an antifungal medicine is given when the doctor first suspects there is a fungal infection (e.g. person still has a fever after four to seven days of antibiotic therapy, or the doctor is still trying to pinpoint the cause of the fever). In pre‐emptive therapy, the doctor uses a series of laboratory screening tests to find what is causing the infection before starting the antifungal medicine. 

Compared to empirical therapy, pre‐emptive therapy may reduce the use of antifungal medicines and the adverse effects they can cause, but may increase the number of deaths. The benefits and harms associated with the two treatment strategies have yet to be determined.

Who will be interested in this review?
Healthcare professionals, including clinical oncologists; people with cancer and the people around them.

What question does this review aim to answer?
This systematic review aimed to find and evaluate the evidence for the relative efficacy (how well they work); safety (the number and severity of side effects); and the impact of pre‐emptive versus empirical antifungal therapy on the use of antifungal medicines in people with cancer who have FN.

Which studies were included in the review?
We searched electronic medical databases to find all relevant studies that included adults with cancer who had FN. To be included, the studies had to be randomised controlled trials (RCTs), which means the participants were divided at random (by chance alone), to receive either empirical or pre‐emptive antifungal medicine (last search October 2021). We included seven studies, involving 1480 people, which compared empirical and pre‐emptive antifungal treatment strategies. 

What does the evidence from the review tell us?
For people with cancer and febrile neutropenia, there may be little or no difference in the number of deaths between those receiving pre‐emptive and those receiving empirical antifungal therapy. Pre‐emptive therapy may increase the rate of identifying IFD, and reduce the duration and rate of use of antifungal medicines, but has not been shown to reduce adverse events. The certainty of the evidence ranged from very low to low; at best, our confidence in the effect estimate is limited.

 What should happen next?
Pre‐emptive therapy may be a promising treatment method for people with cancer who have FN. Since the trials reported different treatments, standardising treatment protocols will help to determine a more valid evaluation of the treatment effects.

Summary of findings

Summary of findings 1. Pre‐emptive antifungal therapy versus empirical antifungal therapy.

Pre‐emptive antifungal therapy versus empirical antifungal therapy for febrile neutropenia in people with cancer
Patient or population: people with cancer who have febrile neutropenia Settings: intensive cytotoxic chemotherapy at hospital Intervention: pre‐emptive antifungal therapy Comparison: empirical antifungal therapy
Outcome	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No. of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed risk with empirical antifungal therapy	Corresponding risk with pre‐emptive antifungal therapy
All‐cause mortality	108 per 1000	105 per 1000 (78 to 141)	RR 0.97 (0.72 to 1.30)	1459 (6 RCTs)	⊕⊕◯◯ lowa	‐
Mortality ascribed to fungal infection	29 per 1000	27 per 1000 (13 to 55)	RR 0.92 (0.45 to 1.89)	1407 (5 RCTs)	⊕⊕◯◯ lowa	‐
Proportion of antifungal agent use (other than prophylactic use)	431 per 1000	306 per 1000 (202 to 452)	RR 0.71 (0.47 to 1.05)	1480 (7 RCTs)	⊕◯◯◯ very lowb	‐
Duration of antifungal use  (days)	The mean duration of antifungal use ranged across control groups from 7 to 20 days	The mean duration of antifungal use ranged across intervention groups from 4.5 to 13.8 days	MD ‐3.52 (‐6.99 to ‐0.66)	764 (3 RCTs)	⊕◯◯◯ very lowb	‐
Invasive fungal infection detection	60 per 1000	103 per 1000 (43 to 244)	RR 1.70 (0.71 to 4.05)	1480 (7 RCTs)	⊕◯◯◯ very lowb	‐
Adverse effects	Hebart 2009 reported 39/152 (25.7%) cases of nephrotoxicity in the pre‐emptive therapy group, and 28/92 cases (30.4%) in the empirical therapy group.  Morrissey 2013 reported 12/118 (10%) cases of hepatotoxicity in the pre‐emptive therapy group, and 21/122 (17%; P = 0.11) cases in the empirical therapy group; and 60/118 (51%) cases of nephrotoxicity in the pre‐emptive therapy group, and 52/122 (43%; P = 0.21) cases in the empirical therapy group.  Cordonnier 2009 reported the mean decrease in creatinine clearance ± SD was larger in the empirical therapy group (‐8.7 ± 20.8) than in the pre‐emptive therapy group (‐5.8 ± 27.2). Severe adverse events occurred in similar proportions in the two groups.  Yuan 2016 reported that 1/26 courses (3.8%) of antifungal treatment resulted in adverse events leading to discontinued treatment in the pre‐emptive therapy group, and 2/43 courses (4.7%) resulted in adverse events leading to discontinued treatment in the empirical therapy group (P = 0.874).			1074  (4 RCTs)	⊕◯◯◯ very lowc	‐
*The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the intervention group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RR: risk ratio; MD, mean difference.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

^aWe downgraded the certainty of the evidence by two levels. First, since there were no blinded studies, this could lead to considerable risk of bias; second, the clinical heterogeneity of the interventions was a critical factor of indirectness. 
^bWe downgraded the certainty of the evidence by three levels. First, since there were no blinded studies, this could lead to considerable risk of bias; second, the clinical heterogeneity of the interventions was critical factor of indirectness; third, the results were inconsistent, marked with considerable heterogeneity (I² =75% to 88%). 
^cWe downgraded the certainty of the evidence by three levels. First, since there were no blinded studies, this could lead to considerable risk of bias; second, the clinical heterogeneity of the interventions was critical factor of indirectness; third, meta analysis was not performed. 

Background

Description of the condition

Intensive cytotoxic chemotherapy for people with cancer can cause severe and prolonged cytopenia, especially neutropenia, a critical condition that is potentially life‐threatening (Smith 2015). When it is manifested by fever and neutropenia, it is called febrile neutropenia (FN). The fever is defined as a single oral temperature reading that is above 38.3 °C (101 °F), or a temperature above 38.0 °C (100.4 °F) that is sustained for at least one hour. Neutropenia is defined as an absolute neutrophil count (ANC) of 500 cells/mm³ or less, or an ANC that is expected to decrease to less than 500 cells/mm³ during the next 48 hours (Freifeld 2011). Incidence rates of chemotherapy‐induced FN are 10% to 50% in people with solid tumours, and more than 80% with haematological malignancies (Al‐Tawfiq 2019; Klastersky 2004).

Invasive fungal disease (IFD) is one of the serious aetiologies of chemotherapy‐induced FN. It is challenging to diagnose IFD early in the course of FN, due to the low sensitivity and specificity of clinical symptoms (e.g. fever, cough, haemoptysis, maculopapular eruptions, nodules, erythema), and microbiological and radiological diagnostic tests (Maddy 2019). Treatment outcomes can potentially result in severe morbidity and mortality when the initiation of therapy is delayed. One autopsy study revealed that 69% of people with a history of prolonged FN had evidence of invasive fungal infections (Cho 1979). A more recent report suggested that the incidence rate of IFD, even in people at a high risk, was considered to be 10% or lower; there was considerable variation in the incidence rate among different cancer types (e.g. leukaemia, allogeneic stem cell transplant, etc), countries, and cancer hospitals (Maertens 2018).

To reduce the morbidity and mortality from IFD, researchers have investigated several antifungal drug strategies. An antifungal treatment strategy is often initiated by giving antifungal drugs prophylactically during chemotherapy. Numerous clinical trials have investigated a prophylactic approach and observed robust efficacy, reducing the risk of IFD and fungal infection‐related mortality (Bow 2002; Glasmacher 2003; Robenshtok 2007). Accordingly, the Infectious Diseases Society of America (IDSA) guidelines recommend antifungal prophylactic drugs (A‐1 recommendation) for people with high‐risk FN (Freifeld 2011).

Alternative antifungal treatment should be considered when fever persists or recurs despite the use of empirical antimicrobial therapy and antifungal prophylaxis. Empirical antifungal therapy refers to the administration of an antifungal agent at the first clinical suspicion of fungal infection (e.g. persistent fever after four to seven days of antibiotic therapy, or in the absence of a definitive diagnosis (Chen 2017)). In previous clinical trials, empirical therapy showed favourable survival outcomes (EORTC 1989; Pizzo 1982). This therapy is also strongly recommended (A‐1 recommendation) in the IDSA guidelines (Freifeld 2011). However, the practice of empirical administration of antifungal drugs leads to several concerns: possible overuse of antifungal agents, emergence of antifungal resistance, undesirable drug interaction, and adverse effects (i.e. nephrotoxicity and hepatotoxicity (Maertens 2012)).

To combat these concerns, researchers have developed another approach, called pre‐emptive or presumptive antifungal therapy. Pre‐emptive therapy is generally referred as the treatment that uses antifungal agents only in cases when fungal infection is detected by diagnostic tests, such as imaging tests (e.g. chest X‐ray, chest and sinus computed tomography (CT) scan), rapid sensitive microbiology assays (e.g. β‐D‐glucan, galactomannan antigen, serum Candida spp. or Aspergillus spp. polymerase chain reaction (PCR)), or both (Fung 2015). According to the IDSA guidelines, pre‐emptive antifungal management is an acceptable alternative to empirical antifungal therapy in a subset of people with high‐risk neutropenia (Freifeld 2011).

To date, there has been no examination of the difference between empirical and pre‐emptive administration of antifungal drugs for efficacy, safety, and impact on clinical resource use. The development of novel antifungal prophylactic agents, and the selection of antifungal prophylactic agents also confound efforts to determine the risks (harms) and benefits between the two treatment strategies.

Description of the intervention

Pre‐emptive therapy begins with serial screening, using rapid sensitive microbiology assays, to detect the presence of a putative pathogen or early subclinical infection. Physicians may also use an imaging test to support a diagnosis. Considering that a single diagnostic approach yields less sensitivity, a combination of microbiology assay and imaging studies may be promising for the early diagnosis of fungal infection (Asano‐Mori 2008; Escuissato 2005; Marr 2005). The combination of these tests leads to the existence of various pre‐emptive therapy protocols. When these tests are positive, antifungal therapy is initiated to avoid progression to an invasive disease.

How the intervention might work

In pre‐emptive therapy, physicians only initiate antifungal therapy when an invasive fungal infection is detected by a diagnostic test. Accordingly, pre‐emptive treatment is expected to decrease the use of antifungal drugs, the emergence of antifungal resistance, undesirable drug interactions, and adverse effects (Fung 2015). However, if the initiation of pre‐emptive treatment is delayed, it may lead to life‐threatening situations compared with empirical therapy.

Why it is important to do this review

Compared to empirical therapy, pre‐emptive therapy may reduce the use of antifungal agents and associated adverse effects, but may increase mortality. The benefits and harms associated with the two treatment strategies have yet to be determined. This systematic review aimed to provide a body of evidence regarding the relative efficacy, safety, and impact of antifungal agent use of pre‐emptive versus empirical antifungal therapy in people with cancer who have FN.

Objectives

To assess the relative efficacy, safety, and impact of antifungal agent use of pre‐emptive versus empirical antifungal therapy in people with cancer who have febrile neutropenia.

Methods

Criteria for considering studies for this review

Types of studies

We only included studies with a randomised controlled trial (RCT) design that compared pre‐emptive antifungal therapy with empirical antifungal therapy for people with cancer, in which participants were prospectively identified. Our eligibility criteria included cluster‐randomised and cross‐over trials. We did not include quasi‐randomised (i.e. using a method of allocation that is not truly random) or non‐randomised comparative studies. 

Types of participants

We included studies with adults, 18 years or older, with cancer, who had febrile neutropenia (FN) as a result of chemotherapy or bone marrow transplantation. We did not restrict participants for gender, ethnicity, underlying malignancies (e.g. solid tumours, haematological malignancies, or haematopoietic stem cell recipients), or the type of chemotherapy used. We also included studies with a paediatric population (younger than 18 years), as long as it was less than 33.3% of the total population of the trial.

We defined fever as a single oral temperature reading above 38.3 °C (101 °F), or a temperature above 38.0 °C (100.4 °F) that was sustained for at least one hour; and neutropenia as an absolute neutrophil count (ANC) of 500 cells/mm³ or less, or an ANC that was expected to decrease to less than 500 cells/mm³ during the next 48 hours (Freifeld 2011). When studies used alternative definitions, we included such studies, provided their definitions meant the same as ours. 

Types of interventions

Experimental intervention

Pre‐emptive antifungal therapy referred to treatment that started only when diagnostic tests explicitly suggested the presence of a fungal infection during FN. To investigate the fungal infection, studies were to have used a serological test (e.g. blood β‐D‐glucan, blood galactomannan antigen, serum Candida spp. or Aspergillus spp. detected by polymerase chain reaction (PCR)), an imaging study (e.g. chest X‐ray or pulmonary or sinus computed tomography (CT)), or both.

Comparison interventions

Empirical antifungal therapy referred to treatment with broad‐spectrum antibacterials that was initiated after several consecutive days of persistent FN.

If studies used alternative definitions, we included such studies, provided their definitions meant the same as ours.

Types of outcome measures

We evaluated one primary outcome and five secondary outcomes.

Primary outcomes

We considered all‐cause mortality as the primary outcome (survival period was defined as the time interval from random treatment assignment to either death from any cause or the last follow‐up).

We did not separate the primary outcome data into short‐ (2 to 6 weeks), medium‐ (7 to 16 weeks), and long‐term (17 to 48 weeks) mortality (Higgins 2022). When possible, we also estimated mortality that occurred during hospitalisation.

Secondary outcomes

1. Treatment failure

   1. Mortality ascribed to fungal infection: defined as death with clinical or autopsy diagnosis of refractory invasive fungal disease (IFD), in the absence of other causes

2. Other outcomes

   1. Proportion of antifungal agent use (other than prophylactic use), defined as the proportion of study participants who received antifungal agents other than prophylactic use during the study period.

   2. Duration of antifungal agent use (days)

   3. Invasive fungal infection detection, defined according to the criteria of the European Organization for Research and Treatment of Cancer/Mycoses Study Group, when available (Ascioglu 2002; Donnelly 2020; De Pauw 2008)

   4. Adverse events (e.g. treatment discontinuation, nephrotoxicity, hepatic dysfunction), according to the Common Terminology Criteria for Adverse Events (CTCAE) Version 5.0 2017, when available (National Cancer Institiute 2017)

We prepared a summary of findings table that reported the following outcomes, listed in order of priority.

1. All‐cause mortality

2. Mortality ascribed to fungal infection

3. Proportion of antifungal agent use (other than prophylactic use)

4. Duration of antifungal agent use

5. Invasive fungal infection detection

6. Adverse events

Search methods for identification of studies

Electronic searches

We searched the following databases:

• Cochrane Central Register of Controlled Trials (CENTRAL; 2021, Issue 10) in the Cochrane Library (searched 12 October 2021; Appendix 1);

• MEDLINE Ovid (1946 to September week 5 2021; Appendix 2);

• Embase Ovid (1980 to 2021 week 40; Appendix 3);

• PubMed (1966 to October 2021).

We searched trial databases, specifically current controlled trials in the National Institutes of Health database (clinicaltrials.gov/), for ongoing and unpublished trials.

Searching other resources

We reviewed the references of all included studies for additional trials. We handsearched the following conference proceedings:

• European Congress of Clinical Microbiology and Infectious Diseases (2001 to Oct 2021);

• Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC) and American Society of Mircrobiology (ASM) Microbe (2001 to Oct 2021);

• Annual Meeting of the Infectious Diseases Society of America (IDSA) and ID Week (2001 to Oct 2021);

• European Society for Medical Oncology Congress (2001 to Oct 2021);

• American Society of Clinical Oncology Annual Meeting (2001 to Oct 2021); and

• American Society of Haematology (ASH) Annual Meeting (2001 to Oct 2021).

Data collection and analysis

Selection of studies

We downloaded all titles and abstracts retrieved by electronic searching into Covidence and removed duplicates. Two review authors (YU and HI) independently screened the titles and abstracts, and excluded those that clearly did not meet the following inclusion criteria:

1. Is the study described as an RCT?

2. Does the study compare pre‐emptive therapy with empirical therapy?

3. Were the participants identified as people with cancer?

We obtained full‐text copies of any potentially relevant references and assessed the eligibility of the retrieved articles. We resolved any disagreement through discussion, or, if required, we consulted a third person (YM). We identified and excluded duplicate reports and collated multiple reports of the same study, so that each study rather than each report was the unit of interest in the review. We recorded the selection process in sufficient detail to complete a PRISMA flow diagram and the characteristics of included and excluded studies tables (Moher 2009).

Data extraction and management

The same review authors (YU and HI) designed a data extraction form (Appendix 4). We extracted the following information and data.

• General information (i.e. title, authors, contact information, country, language and year of publication, sponsors and conflict of interests)

• Trial characteristics (e.g. study design, inclusion and exclusion criteria, publication status, study years, number of centres, randomisation procedure, allocation concealment, blinding procedure, early termination, completeness of data, presence of selective reporting, assessment of compliance, handling of withdrawals, losses to follow‐up, type of analysis, and presence of other sources of bias)

• Participant characteristics (i.e. type of population, inclusion and exclusion criteria, and baseline characteristics)

• Information regarding the intervention (i.e. what test was used before the antifungal drug administration started)

• Information regarding outcomes (e.g. all‐cause mortality, mortality ascribed to fungal infection, IFD detection, proportion of antifungal agent use (other than prophylactic use), adverse events, cost, and other reported outcomes)

• All individuals randomised in the outcome assessment, preferably by intention‐to‐treat; when a single study provided data for multiple measures of the same outcome, we chose the measure that best fit our outcome definitions

• For dichotomous outcomes, we recorded the number of participants manifesting the outcome in each group, as well as the number of evaluated participants per group

• For continuous outcomes, we documented values as well as the measure chosen to report data (e.g. mean with standard deviation (SD) and median with interquartile range)

The two review authors (YU and HI) entered the collected data into the Review Manager 5 file (RevMan 5 (Review Manager 2020)). We double‐checked if the data were entered correctly by comparing the data in RevMan 5 with the study reports. A third review author (YM) independently reviewed the extracted data for accuracy, and resolved any disagreements.

Assessment of risk of bias in included studies

The two authors (YU and HI) independently assessed risk of bias in each study by using the following items:

• selection bias (random sequence generation and allocation concealment),

• performance bias (blinding of participants and personnel),

• detection bias (blinding of outcome assessment),

• attrition bias (incomplete outcome data; defined as high risk when the number of randomised participants was not provided, or the outcome reported was less than 80% of the randomised population without justification),

• reporting bias (selective reporting), and other biases.

We expressed the classification of the specific risks of bias as low, high, or unclear risk of bias, using the criteria suggested in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017). We resolved any disagreement by discussion. We detailed the risk of bias assessment in the Characteristics of included studies table, and graphically presented the results in risk of bias summary tables.

Measures of treatment effect

For continuous outcomes, we extracted means with SD from the studies. If reported differently in a trial, we contacted the trial authors to request the means and SD. We computed the mean difference (MD) and the 95% confidence intervals (CI) for each study. For all dichotomous outcomes, we calculated the risk ratios (RR) for individual studies, with 95% CI. We did not plan to analyse time‐to‐event data (e.g. hazard ratios).

Unit of analysis issues

We presented the included studies in a parallel‐group design; the unit of analysis was expected to be the individual participant. If we included cluster‐randomised trials, we planned to use the published effect estimates, with clustering taken into account. When the original report correctly analysed the cluster‐randomised study, we planned to enter the effect estimate and standard error, using the generic inverse‐variance method in RevMan 5.

If the original report failed to adjust for cluster effects, we planned to include the studies only when we could extract the following information:

• number of clusters randomised to each intervention, or the average size of each cluster,

• outcome data, ignoring the cluster design, for the total number of participants, and

• estimate of the intracluster correlation coefficient (ICC).

We planned to use the ICC from similarly designed studies when such were available. Then, we planned to conduct the approximately correct analysis following the procedures described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2022). If multiple‐arm trials were included, we planned to consider within‐study correlation, by calculating the average of the relevant pairwise comparisons from the study, and calculating a variance for the study, accounting for the correlation between the comparisons. If the control group was a common comparator in a comparison, we planned to divide the participants by the number of relevant intervention groups, to avoid double counting the same data. For trials that had a cross‐over design, we planned to only consider results from the first randomisation period, to avoid period effects.

Dealing with missing data

We requested missing information from the study authors, but were not successful in all cases. The primary analysis was an intention‐to‐treat analysis (ITT), wherein all participants randomised to treatment were included in the denominator. This analysis assumed that all losses to follow‐up had favourable outcomes. To explore the impact of the missing data on the summary effect estimate for death, we planned to conduct a sensitivity analysis. 

If the data were not available for an ITT analyses, we analysed the data with a per‐protocol approach. We described and critiqued any methods, such as imputation, reported in the publication to address incomplete data. In the discussion section of the review, we considered the impact of missing data on the review findings. If possible, we planned to impute missing SDs from other reported studies, by using the maximum reported values in those studies. In all cases, we considered the most conservative assumptions.

Assessment of heterogeneity

We tested for statistical heterogeneity between trials by using the Chi² test (P < 0.1 was considered statistically significant), and the I² statistic, along with a visual inspection of the forest plots. We interpreted I² values in accordance with the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2022). An approximate guide to interpretation was:

• 0% to 40% might not be important;

• 30% to 60% may represent moderate heterogeneity;

• 50% to 90% may represent substantial heterogeneity;

• 75% to 100% indicates considerable heterogeneity.

We also kept in mind that the importance of the observed value in the I² statistic depends on the magnitude and direction of effects and the strength of evidence for heterogeneity.  We conducted subgroup and sensitivity analyses if the I² was above 50%.

Assessment of reporting biases

Had we pooled more than 10 studies, we planned to proceed with the relevant statistical test for funnel plot asymmetry to explore possible small study biases. Since we pooled fewer than 10 trials, we included a narrative description of the risk of reporting bias.

Data synthesis

We analysed the data using RevMan 5 (Review Manager 2020). We individually combined included studies in a meta‐analysis. Given the nature of the intervention, we used a random‐effects model. When we detected substantial and considerable heterogeneity (I² > 75%), we investigated potential source of the heterogeneity and downgraded the evidence.

For dichotomous data, we pooled RRs by using the Mantel–Haenszel method. For continuous data, we calculated the mean difference (MD) when data were reported using the same scale, or standardised mean difference (SMD) for data reported in different scales. We reported pooled estimates with 95% CI.

Subgroup analysis and investigation of heterogeneity

We pre‐planned and addressed the following potential effect modifiers of the primary outcome (all‐cause mortality). 

• documented infection (clinically or microbiologically) versus fever of unknown aetiology

• high‐risk participants versus low‐risk participants, preferably defined by a Multinational Association for Supportive Care (MASCC) score, but we accepted and documented the study definitions

• people with solid tumour versus people with haematological malignancy

• severity of neutropenia (ANC < 100 cells/mm³ versus ANC between 100 cells/mm³ and 500 cells/mm³), referring to the lowest ANC documented (nadir of neutropenia)

• anti‐mould prophylaxis versus non‐anti‐mould prophylaxis

• selection of antifungal agents

• people undergoing stem‐cell transplantation versus people undergoing chemotherapy (we expected stem‐cell transplantation)

Sensitivity analysis

We planned sensitivity analyses for the primary outcome, based on risk bias (only including data from trials at low risk of bias); missing data (only including data from trials without missing data), and the presence of heterogeneity (only including data from trials with homogenous interventions). We added sensitivity analyses for all‐cause mortality reported in the short‐term (2 to 6 weeks), and medium‐term (7 to 16 weeks). Since Hebart 2009 reported all‐cause mortality at both short‐ and medium‐term time points, we added another sensitivity analysis in which we replaced medium‐term data for short‐term data.

Summary of findings and assessment of the certainty of the evidence

We presented the overall certainty of the evidence for each outcome according to the GRADE approach, which took into account issues related to internal validity (risk of bias, inconsistency, imprecision, publication bias), and external validity, such as directness of results (Langendam 2013; Schünemann 2011). We created a summary of findings table, following the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2022), using GRADEPro GDT software. We downgraded the evidence from high certainty by one level for serious (or by two for very serious) concerns for each limitation. We used the GRADE approach and GRADE Working Group certainty of evidence definitions, listed below (Meader 2014). 

• High certainty: we are very confident that the true effect lies close to that of the estimate of the effect.

• Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.

• Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect.

• Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

We evaluated the evidence for all‐cause mortality, mortality ascribed to fungal infection, proportion of antifungal agent use (other than prophylactic use), duration of antifungal use (days), invasive fungal infection detection, and adverse effects, for the comparison between empirical antifungal therapy and pre‐emptive antifungal therapy.

Since we were unable to undertake meta‐analysis for all outcomes, we presented our results in a hybrid summary of findings table format (both quantitative and narrative), such as that used by Chan 2011.

Results

Description of studies

Results of the search

We retrieved 2254 reports from the systematic literature search, one abstract from an academic conference, and two studies by handsearching. After Covidence removed 174 duplicate records, we screened 2083 records. We unanimously included seven records after title and abstract screening, and two after discussing 19 conflict points. We assessed the full text of nine reports for eligibility (Figure 1).

1.

PRISMA flow diagram, illustrating study selection

Included studies

We included seven trials included in the review. See Characteristics of included studies for details. 

The included studies only enroled 1480 people with high‐risk febrile neutropenia (FN; e.g. people with haematological malignancy); none with low‐risk FN (e.g. people with solid tumour). We found variability in the definition of pre‐emptive treatment (e.g. laboratory tests with or without quantitative real‐time PCR assays, ELISA, galactomannan test; imaging with or without high‐resolution computed tomography; prophylaxis; antifungal agents). Three were single‐centred, and four were multicentre trials.

Excluded studies

We excluded two reports after full‐text review, and described the reasons in the Characteristics of excluded studies; one had the wrong intervention, and one was not a randomised trial. 

Risk of bias in included studies

For details of the risk of bias judgements for each study, see Characteristics of included studies. See Figure 2 for the graphical representation of the overall risk of bias across included studies.

2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies

Allocation

Sequence generation
Four studies described the method of randomisation in detail (low risk of bias); three did not, and the process was unclear (unclear risk of bias).

Allocation concealment
Appropriate allocation concealment was performed in five studies (low risk of bias), and the process was unclear in two studies that did not provide a detailed description (unclear risk of bias).

Blinding

Blinding of participants and personnel

Blinding of participants, healthcare professionals, and investigators was not feasible due to the nature of the intervention. Non‐blinded or open label was clearly stated in four studies, and not in three studies (high risk of bias).

Blinding of outcome assessment

Five of the seven studies reported a blinded outcome assessor, so we assessed them at low risk of bias; we assessed the other two as unclear risk of bias, because of lack of information.

Incomplete outcome data

All studies clearly stated the process from participant enrolment to allocation and outcome assessment (low risk of bias), but there was a serious risk of bias in Blennow 2010; only 21 out of 41 enroled people were randomised (high risk of bias).

Selective reporting

Four studies had clinical trial registration. We did not identify any apparent selective reporting, and judged them at low risk of bias. Since there were no protocols or clinical trial registrations for the other three studies, we rated them at unclear risk of bias.

Other potential sources of bias

Five studies transparently described the study process, and we did not identify other potential biases (low risk of bias). The other two studies lacked sufficient information to assess other biases (unclear risk of bias).

Effects of interventions

See: Table 1

See: Table 1

Six included studies reported the effect on all‐cause mortality; five reported on mortality ascribed to fungal infection; seven reported on the proportion of antifungal agent use (other than prophylactic use); three reported on the duration of antifungal use in days; seven reported on invasive fungal infection detection; and four reported adverse events. 

Primary outcome

All‐cause mortality

There may be little to no difference between groups for all‐cause mortality (risk ratio (RR) 0.97, 95% confidence interval (CI) 0.72 to 1.30; absolute effect – reduction of 3 out of 1000; 6 studies, 1459 participants; low‐certainty evidence; Analysis 1.1). No evidence of heterogeneity was identified for this outcome (Chi² = 3.31, I² = 0%). 

1.1. Analysis.

Comparison 1: Pre‐emptive versus empirical antifungal therapy, Outcome 1: All‐cause mortality

Secondary outcomes

Mortality ascribed to fungal infection

There may be little to no difference between groups for mortality ascribed to fungal infection (RR 0.92, 95% CI 0.45 to 1.89; absolute effect – reduction of 2 out of 1000; 5 studies, 1407 participants; low‐certainty evidence; Analysis 1.2). No evidence of heterogeneity was identified for this outcome (Chi² = 4.48, I² = 11%). 

1.2. Analysis.

Comparison 1: Pre‐emptive versus empirical antifungal therapy, Outcome 2: Mortality ascribed to fungal infection

Proportion of antifungal agent use (other than prophylactic use)

Pre‐emptive therapy may reduce the proportion of antifungal agent used (other than prophylactic use; RR 0.71, 95% CI 0.47 to 1.05; absolute effect – reduction of 125 out of 1000; 7 studies, 1480 participants; very low‐certainty evidence: Analysis 1.3). There was considerable heterogeneity for this outcome (Chi² = 49.0, I² = 88%); intervention protocols for antifungal agents used were heterogeneous across included studies. 

1.3. Analysis.

Comparison 1: Pre‐emptive versus empirical antifungal therapy, Outcome 3: Proportion of antifungal agent use (other than prophylactic use)

Duration of antifungal use (days)

Pre‐emptive therapy may reduce the duration of antifungal use more than empirical therapy (mean difference (MD) ‐3.52 days, 95% CI ‐6.99 to ‐0.66; 3 studies, 764 participants; very low‐certainty evidence; Analysis 1.4). There was considerable heterogeneity for this outcome (Chi² = 14.64, I² = 86%). 

1.4. Analysis.

Comparison 1: Pre‐emptive versus empirical antifungal therapy, Outcome 4: Duration of antifungal use (days)

Invasive fungal infection detection

 Pre‐emptive therapy may increase invasive fungal infection detection more than empirical treatment (RR 1.70, 95% CI 0.71 to 4.05; absolute effect – increased by 43 out of 1000; 7 studies, 1480 participants; very low‐certainty evidence; Analysis 1.5). There was considerable heterogeneity for this outcome (Chi² = 23.54, I² = 75%). 

1.5. Analysis.

Comparison 1: Pre‐emptive versus empirical antifungal therapy, Outcome 5: Invasive fungal infection detection

Adverse events

Four of the seven trials reported adverse events. Since the adverse events and evaluating methods were different among the trials, we did not undertake a meta‐analysis. Narrative summary of the adverse events follows.

Hebart 2009 reported 39/152 (25.7%) cases of nephrotoxicity in the pre‐emptive therapy group, and 28/92 (30.4%) cases in the empirical therapy group. 

Morrissey 2013 reported 12/118 (10%) cases of hepatotoxicity in the pre‐emptive therapy group, and 21/122 (17%; P = 0.11) in the empirical therapy group. They also reported 60/118 (51%) cases of nephrotoxicity in the pre‐emptive therapy group, and 52/122 (43%; P = 0.21) in the empirical therapy group. 

In Cordonnier 2009, the mean decrease of creatinine clearance ± standard deviation (SD) was larger in the empirical therapy group (‐8.7 ± 20.8) than in the pre‐emptive therapy group (‐5.8 ± 27.2). Severe adverse events occurred in similar proportions among the two groups. 

In Yuan 2016, 1/26 courses (3.8%) of antifungal treatment administered during pre‐emptive therapy, and 2/43 courses (4.7%) of empirical therapy led to adverse events, for which antifungal treatment was discontinued (P = 0.874).

Subgroup analysis

We had planned a number of subgroup analyses for the primary outcome (prophylaxis, cancer type, risk of FN, selection of antifungal agents, severity of neutropenia, type of chemotherapy), but the reports all lacked sufficient information to enable us to do so. Our enquiries to corresponding authors of each of the studies did not garner enough data. Therefore, we did not perform any subgroup analyses.

Sensitivity Analysis

Results were consistent for all‐cause mortality when we excluded data from trials with higher risk of bias (Analysis 2.1); replaced medium‐term with short‐term outcome data for Hebart 2009 (Analysis 2.2), or separately analysed short‐term (2 to 6 weeks; Analysis 2.3), and medium‐term (7 to 16 weeks) outcome data (Analysis 2.4). 

2.1. Analysis.

Comparison 2: Sensitivity analyses for all‐cause mortality, Outcome 1: RCTs at lower risk of bias

2.2. Analysis.

Comparison 2: Sensitivity analyses for all‐cause mortality, Outcome 2: Replacing outcomes for Hebart 2009

2.3. Analysis.

Comparison 2: Sensitivity analyses for all‐cause mortality, Outcome 3: Short‐term time point

2.4. Analysis.

Comparison 2: Sensitivity analyses for all‐cause mortality, Outcome 4: Medium‐term time point

We were unable to undertake sensitivity analyses for trials with no missing data, or those with homogenous interventions, as planned, due to lack of data.

Reporting bias

We did use a funnel plot to assess reporting bias because we only included seven trials. It was difficult to fully evaluate reporting bias, because not all trials had clinical trial registrations, and published protocols.

Discussion

Summary of main results

See Table 1  for the comparison, pre‐emptive versus empirical antifungal therapy for febrile neutropenia in people with cancer.

We included seven randomised controlled trials, with 1480 participants in this review. 

There may be little to no difference between groups for all‐cause mortality (risk ratio (RR) 0.97, 95% confidence interval (CI) 0.72 to 1.30); and mortality ascribed to fungal infection (RR 0.92, 95% CI 0.45 to 1.89). 

Pre‐emptive therapy may decrease antifungal agent use (other than prophylactic use; RR 0.71, 95% CI 0.47 to 1.05); and the duration of antifungal use (mean difference (MD) ‐3.52, 95% CI ‐6.99 to ‐0.06) over pre‐emptive therapy. 

Although we could not undertake a meta‐analysis of adverse events, narratively, there seemed to be no apparent difference in the frequency or severity of adverse events between groups.

Overall completeness and applicability of evidence

Our conclusions were based on data from seven studies that met our inclusion criteria. We completed sensitivity analyses for the primary outcome, all‐cause mortality, for risk of bias and measurement time point. We found that these results were consistent with the primary analysis, which supported the robustness of the effect. We were unable to complete our pre‐planned subgroup analysis, due to lack of data. 

Each study used their own specific antifungal protocols for both pre‐emptive and empirical antifungal therapies. The results from this review were based on clinically heterogeneous evidence, and did not provide a clear answer as to which protocol should be implemented in clinical practice. 

The two outcomes associated with mortality exhibited less statistically heterogeneity, suggesting that whatever protocol chosen may have little impact on mortality. The two outcomes of clinical resource utilisation, and the identification rate of invasive fungal disease (IFD) showed considerable statistical heterogeneity (I² > 75%), which might reflect the effects of different intervention and outcome evaluation protocols.

Quality of the evidence

We judged all seven of the included studies to have some risk of bias, in particular, none of the participants or personnel were blinded. We downgraded the certainty of the evidence for all outcomes by one level due to lack of blinding. We identified non‐negligible clinical and statistical heterogeneity for each outcome, which was a critical factor of indirectness, for which we downgraded the certainty by another level, or two (for serious indirectness). Since we were unable to pool the data for adverse events, we downgraded by another level for sparse data.

Potential biases in the review process

First, our literature search strategy seemed have limited sensitivity. Although it was not specified in the title and abstract as pre‐emptive treatment, there were studies that conceptually, could be regarded as pre‐emptive treatment (Blennow 2010; Morrissey 2013). These studies were included in a previous review by Fung 2015; we identified them by handsearching.

Second, we were unable to conduct subgroup analyses. Efficacy of pre‐emptive treatment can be modified by various factors, including our pre‐planned factors. Since subgroup analysis contributes to the identification of suitable treatment populations, future studies would be encouraged to report the results of various subgroup analysis in detail.

Third, we meta‐analysed various evaluation periods for the primary endpoints. Initially, we planned analyses for short‐term, medium‐, and long‐term outcomes, but we only had seven trials. The main analysis had low statistical heterogeneity, but the short‐term and medium‐term evaluations, as sensitivity analysis, had higher statistical heterogeneity. It would be reasonable to develop a consensus on the clinically optimal evaluation period.

Finally, the novel concept of antifungal treatment strategy for FN has been developed in recent years. Kanda 2020 developed D‐Index‐Guided Early Antifungal Therapy and demonstrated promising efficacy. When we update the current review, it will be necessary to update the review questions, including not only a comparison between pre‐emptive and experiential therapy, but also the novel treatment strategies. 

Agreements and disagreements with other studies or reviews

Fung 2015 conducted a systematic review and meta‐analysis of the same research question as this review in 2015, which included nine studies: five RCTs, and four non‐RCT study designs. We included these five RCTs, and we also added two further studies.

They concluded that "Compared to empirical antifungal therapy, pre‐emptive strategies were associated with significantly lower antifungal exposure (RR 0.48, 95% CI 0.27 to 0.85) and duration, without an increase in IFD‐related mortality (RR 0.82, 95% CI 0.36 to 1.87) or overall mortality (RR 0.95, 95% CI 0.46 to 1.99)". Our results were similar. The statistical heterogeneity associated with mortality improved in our review, which may be due to the focus on randomised controlled trials and the addition of new studies. Both reviews showed considerable statistical heterogeneity in IFD identification and antifungal agent use. The duration of antifungal use was only meta‐analysed in our review.

Authors' conclusions

Implications for practice.

For people with cancer and high‐risk febrile neutropenia, pre‐emptive antifungal therapy may reduce the duration and rate of use of antifungal agents without deteriorating all‐cause and invasive fungal disease‐related mortality, compared to empirical therapy. But the evidence regarding invasive fungal infection detection and adverse events was inconsistent and uncertain. 

Implications for research.

Future research should focus on:

• Developing a standardised protocol for pre‐emptive antifungal therapy;

• Defining the target population for pre‐emptive antifungal therapy, reporting the detailed and various subgroup analysis;

• Developing a consensus on the clinically optimal evaluation period for mortality and other relevant outcomes.

History

Protocol first published: Issue 5, 2020

Appendices

Appendix 1. CENTRAL search strategy

#1 MeSH descriptor: [Neoplasms] explode all trees         
#2 cancer* or tumor* or tumour* or neoplas* or malignan* or carcinoma* or adenocarcinoma* or choriocarcinoma* or leukemia* or leukaemia* or metastat* or sarcoma* or teratoma*     
#3 #1 or #2        
#4 MeSH descriptor: [Invasive Fungal Infections] explode all trees            
#5 MeSH descriptor: [Antifungal Agents] explode all trees           
#6 (antifung* or anti fung* or anti‐fung*) near/5 (therap* or treat* or agent* or manag* or method* or car* or service* or plan* or intervention*)
#7 fung* near/5 infect*              
#8 (pre‐emptive* or pre emptive* or preemptive* or presumptive*) near/5 therap* 
#9 (empirical or empiric) near/5 therap*
#10 #4 or #5 or #6 or #7 or #8 or #9       
#11 MeSH descriptor: [Febrile Neutropenia] explode all trees     
#12 febrile neutrop* or febrile‐neutrop* or FN* or neutropenia*            
#13 (fever* or pyrexia* or temperature* or hyperthermia) near/8 neutrop*              
#14 #11 or #12 or #13   
#15 #3 and #10 and #14              

Appendix 2. MEDLINE search strategy

1. exp Neoplasms/
2. (cancer* or tumor* or tumour* or neoplas* or malignan* or carcinoma* or adenocarcinoma* or choriocarcinoma* or leukemia* or leukaemia* or metastat* or sarcoma* or teratoma*).mp.
3. 1 or 2
4. exp Invasive Fungal Infections/
5. exp Antifungal Agents/
6. ((antifung* or anti fung* or anti‐fung*) adj5 (therap* or treat* or agent* or manag* or method* or car* or service* or plan* or intervention*)).mp.
7. (fung* adj5 infect*).mp.
8. ((pre‐emptive* or pre emptive* or preemptive* or presumptive*) adj5 therap*).mp.
9. ((empirical or empiric) adj5 therap*).mp.
10. 4 or 5 or 6 or 7 or 8 or 9
11. exp Febrile Neutropenia/
12. (febrile neutrop* or febrile‐neutrop* or FN* or neutropenia*).mp.
13. ((fever* or pyrexia* or temperature* or hyperthermia) adj8 neutrop*).mp.
14. 11 or 12 or 13
15. 3 and 10 and 14
16. randomized controlled trial.pt.
17. controlled clinical trial.pt.
18. randomized.ab.
19. placebo.ab.
20. drug therapy.fs.
21. randomly.ab.
22. trial.ti.
23. groups.ab.
24. 16 or 17 or 18 or 19 or 20 or 21 or 22 or 23
25. (animals not (humans and animals)).sh.
26. 24 not 25
27. 15 and 26

Appendix 3. Embase search strategy

1. exp neoplasm/
2. (cancer* or tumor* or tumour* or neoplas* or malignan* or carcinoma* or adenocarcinoma* or choriocarcinoma* or leukemia* or leukaemia* or metastat* or sarcoma* or teratoma*).mp.
3. 1 or 2              
4. exp systemic mycosis/             
5. exp antifungal agent/ 
6. ((antifung* or anti fung* or anti‐fung*) adj5 (therap* or treat* or agent* or manag* or method* or car* or service* or plan* or intervention*)).mp. 
7. (fung* adj5 infect*).mp.          
8. ((pre‐emptive* or pre emptive* or preemptive* or presumptive*) adj5 therap*).mp.     
9. ((empirical or empiric) adj5 therap*).mp.        
10. 4 or 5 or 6 or 7 or 8 or 9        
11. exp febrile neutropenia/       
12. (febrile neutrop* or febrile‐neutrop* or FN* or neutropenia*).mp.    
13. ((fever* or pyrexia* or temperature* or hyperthermia) adj8 neutrop*).mp.              
14. 11 or 12 or 13            
15. 3 and 10 and 14        
16. crossover procedure/            
17. double‐blind procedure/      
18. randomized controlled trial/ 
19. single‐blind procedure/         
20. random*.mp.            
21. factorial*.mp.           
22. (crossover* or cross over* or cross‐over*).mp.           
23. placebo*.mp.            
24. (double* adj blind*).mp.      
25. (singl* adj blind*).mp.           
26. assign*.mp. 
27. allocat*.mp. 
28. volunteer*.mp.        
29. 16 or 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26 or 27 or 28              
30. 15 and 29

Appendix 4. Data extraction form

Assessor			YU	YU
Date of Assessment			29/08/2019	29/08/2019
		Study ID	ABC 2019	ABC 2019
		Intervention or control	Intervention	Control
		Labels for intervention and control groups	Pre‐emptive	Empirical
General Information		Title
		Authors
		Contact person/affiliation
		Contact address
		Country
		Language
		Year of publication
		Sponsors
		Conflict of interests
Trial Characteristics		RCT type
		inclusion criteria
		exclusion criteria
		Publication status
		Study years
		Number of centres
Participants' characteristics		Population
		No. randomised
		Age: mean (SD)
		female (%)
		Haematological malignancy (%)
		Patients undergoing stem‐cell transplantation (%)
		High‐risk participants defined by MASCC score (%)
		Severe neutropenia (ANC < 100; %)
		Antifungal prophylaxis (%)
		Anti‐mould prophylaxis (%)
Information regarding the intervention		Pre‐emptive protocol
		Empirical protocol
		Diagnostic testing
		Escalation antifungal drug
Outcomes 1: All‐cause mortality	post‐intervention	timing
		definition of events
		number of events (if only continuous data were provided, write down mean (SD)
		number in denominator
		indirectness of the outcome
Outcome 2: Mortality ascribed to fungal infection	post‐intervention	timing
		definition of events
		number of events (if only continuous data were provided, write down mean (SD)
		number in denominator
		indirectness of the outcome
Outcome 3: Proportion of antifungal agent use	post‐intervention	timing
		definition of events
		number of events (if only continuous data were provided, write down mean (SD)
		number in denominator
		indirectness of the outcome
Outcome 4: Duration of antifungal agent use	post‐intervention	timing
		definition of events
		number of events (if only continuous data were provided, write down mean (SD)
		number in denominator
		indirectness of the outcome
Outcome 5: Invasive fungal infection detection	post‐intervention	timing
		definition of events
		number of events (if only continuous data were provided, write down mean (SD)
		number in denominator
		indirectness of the outcome
Outcome 6: Adverse events	post‐intervention	timing
		definition of events
		number of events (if only continuous data were provided, write down mean (SD)
		number in denominator
		indirectness of the outcome
Outcome X: additional probably important outcome 1	post‐intervention	timing
		definition of events
		number of events (if only continuous data were provided, write down mean (SD)
		number in denominator
		indirectness of the outcome
Outcome Y: additional probably important outcome 2	post‐intervention	timing
		definition of events
		number of events (if only continuous data were provided, write down mean (SD)
		number in denominator
		indirectness of the outcome
Risk of bias		Random sequence generation
		Allocation concealment
		Blinding
		Incomplete outcome data
		Selective reporting
		Other bias
		Overall RoB
Indirectness		participants
		intervention
		control condition
		overall indirectness

Data and analyses

Comparison 1. Pre‐emptive versus empirical antifungal therapy.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 All‐cause mortality	6	1459	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.72, 1.30]
1.2 Mortality ascribed to fungal infection	5	1407	Risk Ratio (M‐H, Random, 95% CI)	0.92 [0.45, 1.89]
1.3 Proportion of antifungal agent use (other than prophylactic use)	7	1480	Risk Ratio (M‐H, Random, 95% CI)	0.71 [0.47, 1.05]
1.4 Duration of antifungal use (days)	3	764	Mean Difference (IV, Random, 95% CI)	‐3.52 [‐6.99, ‐0.06]
1.5 Invasive fungal infection detection	7	1480	Risk Ratio (M‐H, Random, 95% CI)	1.70 [0.71, 4.05]

Comparison 2. Sensitivity analyses for all‐cause mortality.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 RCTs at lower risk of bias	5	1191	Risk Ratio (M‐H, Random, 95% CI)	0.93 [0.68, 1.26]
2.2 Replacing outcomes for Hebart 2009	6	1459	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.52, 1.38]
2.3 Short‐term time point	3	964	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.25, 3.55]
2.4 Medium‐term time point	2	455	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.65, 1.50]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Aguado 2015.

Study characteristics
Methods	Study design: open‐label, parallel‐group, randomised trial Country: Spain Number of centres: 13
Participants	Adults (≥ 18 years) with AML or high‐risk MDS undergoing remission induction therapy for newly diagnosed or relapsed or refractory disease and people undergoing allo‐HSCT. People with the diagnosis of IA within 6 months prior to or at enrolment, the receipt of antifungal therapy with anti‐Aspergillus activity within 30 days prior to or at enrolment, and a history of hypersensitivity to azoles, were excluded.
Interventions	Pre‐emptive therapeutic protocol:  Serum GM and Aspergillus species quantitative rtPCR assays were performed twice a week until neutrophil recovery reached ≥ 0.5 × 10^3 cells/mm^3 for people with AML/MDS, or until days 180 or 100 for allo‐HSCT recipients with or without GVHD, respectively. The results of both GM and rtPCR assays were provided to the attending physicians of participants allocated into the GM‐PCR group; in the GM group, only the results of the GM assay were provided. In both groups, these results were available within 24 to 48 hours from sampling. When the result of at least 1 of the tests were positive, ordering a thoracic HRCT scan was mandatory and an antifungal agent with activity against Aspergillus species had to be initiated, even if the HRCT scan revealed no radiological signs suggestive of IA, according to the EORTC/MSG criteria. Of note, the rtPCR results generated from the GM group had no influence on the clinical management of these participants, or on the decision of whether to order an imaging study. Febrile neutropenia episodes were managed with broad‐spectrum antibiotics according to the local practice at each centre. Attending physicians were allowed to order additional HRCT scan examinations whenever they deemed them clinically necessary.  Empirical therapeutic protocol:  Empirical antifungal therapy had to be prescribed after 72 hours of the onset of refractory febrile neutropenia, even if no evidence of IA was present, and subsequently, the participant's withdrawal from the study was mandatory.
Outcomes	Time points for assessment: the follow‐up period extended from the date of first monitoring until 30 days after the resolution of neutropenia for participants with AML/MDS, or until days 210 or 130 days for allo‐HSCT recipients with or without GVHD, respectively. Primary outcome:   1. Cumulative incidence of proven or probable IA according to the EORTC/MSG criteria. Secondary outcome: 1. All‐cause and proven or probable IA‐attributable mortality 2 Proven or probable IA‐free survival 3. Cumulative incidence of any EORTC/MSG category of IA (possible, probable, or proven)  4. Antifungal consumption
Notes	Study period: February 2011 to September 2012 Funding source: this study was supported by an unrestricted grant from Pfizer. Declarations of interest among the researchers: J. M. A. received grant support from Astellas Pharma, Gilead Sciences, Merck Sharp & Dohme (MSD), Pfizer, Instituto de Salud Carlos III (Spanish Ministry of Economy and Competitiveness), and the Mutua Madrileña Foundation; was an advisor/consultant to Astellas Pharma, Gilead Sciences, MSD, and Pfizer; and received honoraria for talks on behalf of Gilead Sciences, MSD, Pfizer, and Astellas Pharma. L. V. received grant support from Astellas Pharma, Gilead Sciences, MSD, Pfizer, Instituto de Salud Carlos III (Spanish Ministry of Economy and Competitiveness), and the Mutua Madrileña Foundation; was an advisor/consultant to Astellas Pharma, Gilead Sciences, MSD, and Pfizer; and received honoraria for talks on behalf of Gilead Sciences, MSD, Pfizer, and Astellas Pharma. M. F.‐R. received honoraria for talks on behalf of Pfizer. I. R.‐C. received honoraria for talks on behalf of Pfizer, MSD, Gilead, Astellas, and Novartis. P. B. received honoraria for talks on behalf of Gilead Sciences, Pfizer, and Astellas Pharma. M. B. received grant support from Pfizer. C. S. received grant support from Astellas Pharma, Gilead Sciences, Pfizer, and Instituto de Salud Carlos III (Spanish Ministry of Economy and Competitiveness); was an advisor/consultant to Astellas Pharma, Gilead Sciences, MSD, and Pfizer; and received honoraria for talks on behalf of Gilead Sciences, MSD, Pfizer, and Astellas Pharma. D. G. received grant support from Instituto de Salud Carlos III (Spanish Ministry of Economy and Competitiveness) and the Mutua Madrileña Foundation. R. V. was an advisor/consultant to Gilead Sciences. C. V. received grant support from Pfizer; was an advisor/consultant to Astellas Pharma, Gilead Sciences, MSD, and Pfizer; and received honoraria for talks on behalf of Gilead Sciences, MSD, and Pfizer. M. C.‐E. received grant support from Astellas Pharma, bioMérieux, Gilead Sciences, MSD, Pfizer, Schering‐Plough, Soria Melguizo SA, Ferrer International, the European Union, the ALBAN program, the Spanish Agency for International Cooperation, the Spanish Ministry of Culture and Education, the Spanish Health Research Fund, Instituto de Salud Carlos III (Spanish Ministry of Economy and Competitiveness), the Ramon Areces Foundation, and the Mutua Madrileña Foundation; was an advisor/consultant to the Pan American Health Organization, Astellas Pharma, Gilead Sciences, MSD, Pfizer, and Schering‐Plough; and was paid for talks on behalf of Gilead Sciences, MSD, Pfizer, Astellas Pharma, and Schering‐Plough. All other authors reported no potential conflicts.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	An independent statistician performed the procedure by using a computer‐generated schedule of randomly permuted blocks.
Allocation concealment (selection bias)	Low risk	Phone confirmation of the allocated group was provided by independent statistician.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Open‐label, controlled, parallel‐group randomised trial
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Laboratory technicians who performed the monitoring assays were blinded to the intervention, and all the study outcomes were reviewed by an independent adjudication committee.
Incomplete outcome data (attrition bias) All outcomes	Low risk	This study clearly stated the process from participant enrolment to allocation and outcome assessment; attrition rate was 7.3% (16/219).
Selective reporting (reporting bias)	Unclear risk	There was a clinical trial registration and no apparent selective reporting was identified.
Other bias	Low risk	This study transparently described the study process, and other apparent biases were not identified.

Blennow 2010.

Study characteristics
Methods	Study design: prospective, randomised, non‐blinded study Country: Sweden Number of centres: 1
Participants	People undergoing HSCT after RIC at the Karolinska University Hospital, Huddinge, were included. People with hypersensitivity to liposomal amphotericin B were excluded.
Interventions	Participants were followed once a week with fungal PCR for the first 100 days after transplantation. Pre‐emptive therapeutic protocol:  Participants with a positive PCR result were treated with liposomal amphotericin. Liposomal amphotericin B was given at a dose of 3 mg/kg/day for participants with positive PCR for Aspergillus. After 3 days, the dose was reduced to 1 mg/kg/day if the participant was in a stable clinical condition. For Candida, the treatment dose was 1 mg/kg/day. The treatment was continued until one PCR test was found to be negative in participants without fever and neutropenia, or until two consecutive PCRs were negative in participants with fever and/or neutropenia. Depending on the participant’s clinical status, the treatment could be prolonged according to the judgement of the treating physician. Empirical therapeutic protocol:   Participants with fever above 38.5 °C for 3 days despite treatment with antibiotics, and with unknown aetiology, were considered eligible for antifungal treatment regardless of the PCR results.
Outcomes	Primary outcome:   1. Success of PCR‐guided antifungal treatment defined as a reduction of proven and probable IFIs in participants allocated to liposomal amphotericin B treatment within 100 days of HSCT Secondary outcome: 1. Survival at 100 days  2. Survival at 1 year 3. 1‐year incidence of IFI 4. Risk factors for IFI
Notes	Study period: April 2002 and November 2006 Funding source: Sweden Orphan Declarations of interest among the researchers: none
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	High risk	No clear statement of the blinding; but due to the nature of the intervention, it would be difficult to blind participants, healthcare professionals and investigators.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Laboratory test seemed to be performed at an independent department, however, outcome assessment was unclear.
Incomplete outcome data (attrition bias) All outcomes	High risk	Only 21.2% (21/99) of participants was randomised.
Selective reporting (reporting bias)	Unclear risk	There was no protocol or clinical trial registration for this study.
Other bias	Unclear risk	This study lacked sufficient information to assess the other biases.

Cordonnier 2009.

Study characteristics
Methods	Study design: a prospective, randomised, open‐label, non‐inferiority trial Country: France Number of centres: 13
Participants	People aged ⩾18 years, with haematological malignancies, and were scheduled for chemotherapy or autologous stem cell transplantation that was expected to cause neutropenia (neutrophil count, < 500 cells/mm3) for at least 10 days.  Exclusion criteria: a planned allogeneic transplantation, a history of, or symptoms consistent with IFI, previous severe toxicity from intravenous polyenes, a Karnofsky score < 30%, and HIV seropositivity.
Interventions	Pre‐emptive therapeutic protocol:  Antifungal treatment was guided by any of the following occurrences, at any time after 4 days of fever and antibacterial treatment: clinically and imaging‐documented pneumonia or acute sinusitis, mucositis of grade 3, septic shock, skin lesion suggesting IFI, unexplained CNS symptoms, periorbital inflammation, splenic or hepatic abscess, severe diarrhoea, Aspergillus colonisation, or ELISA results positive for galactomannan antigenemia.  Empirical therapeutic protocol:   Persistent or recurrent fever led to initiation of antifungal therapy.
Outcomes	Primary outcome:   1. Proportion of participants alive at 14 days after recovery from neutropenia, or for participants with persistent neutropenia at 60 days after inclusion in the study, or an SAE, at the time that these participants were censored Secondary outcome: 1. Fever duration and the proportion of participants with proven or probable IFI 2.  Survival at 4 months after inclusion
Notes	Study period: April 2003 to February 2006 Funding source: French Ministry of Health Declarations of interest among the researchers: C.C. received travel grant and research support from Pfizer Gilead, Merck Sharp & Dohme‐Chibret; and was a consultant for Pfizer Schering‐Ploug, Gilead, Merck Sharp & Dohme‐Chibret, and Zeneuu Pharma
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were assigned in a 1:1 ratio to the empirical or pre‐emptive treatment arms with use of a computer‐generated randomisation scheme with blocks of 4.
Allocation concealment (selection bias)	Low risk	Participants were assigned in a 1:1 ratio to the empirical or pre‐emptive treatment arms with use of a computer‐generated randomisation scheme with blocks of 4.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Open‐label study
Blinding of outcome assessment (detection bias) All outcomes	Low risk	An independent, blinded adjudication committee reviewed the reasons for starting antifungal therapy, diagnoses of IFI, and causes of death.
Incomplete outcome data (attrition bias) All outcomes	Low risk	This study clearly stated the process from participant enrolment to allocation and outcome assessment. Protocol violation occurred in 4.1% (12/293) of the participants; no neutropenia occurred in 4.8% (14/293) of the participants.
Selective reporting (reporting bias)	Low risk	There was a clinical trial registration and no apparent selective reporting was identified.
Other bias	Low risk	This study transparently described the study process, and other apparent biases were not identified.

Hebart 2009.

Study characteristics
Methods	Study design: prospective randomised controlled multicenter trial Country: Germany Number of centres: 5
Participants	People who were required to receive Allo‐SCT (BMT or peripheral blood progenitor cell transplantation from related and unrelated donors). People with documented allergy against liposomal amphotericin B, history of IFI at the time of randomisation or before the study entry were excluded.
Interventions	Pre‐emptive therapeutic protocol:  Participants were treated with liposomal amphotericin B after one positive PCR result. If Participants had febrile neutropenia for more than 120  h that was not responding to broad‐spectrum antibacterial therapy, they also received liposomal amphotericin B, even if PCR remained negative. Antifungal treatment lasted for a minimum of 7 days. PCR‐based liposomal amphotericin B therapy was stopped as soon as the granulocyte count had recovered (4500/mL for 3 consecutive days), or if the participant had been afebrile for a minimum of 3 days, presented no signs and symptoms of IFI, and two consecutive PCR results were negative. Empirical therapeutic protocol:   Participants received empirical antifungal therapy after 120 h of febrile neutropenia not responding to broad‐spectrum antibacterial therapy. Antifungal treatment lasted for a minimum of 7 days. Empirical liposomal amphotericin B was stopped if the granulocytes had recovered (4500/mL for 3 consecutive days), or if the participant had been afebrile for a minimum of 3 days and presented no radiological signs of invasive aspergillosis.
Outcomes	Primary outcome:   1. Incidence of IFIs during the observation period of 100 days after transplantation. Secondary outcome: 1. Survival at day 100 (overall mortality) 2. Survival at day 30 (overall mortality) 3. IFI‐related mortality,  4. Incidence and spectrum of adverse drug reactions
Notes	Study Period: 1998 to 2001 Funding source: Gilead Declarations of interest among the researchers: none
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Centrally randomised, but detailed randomisation process was unclear
Allocation concealment (selection bias)	Low risk	Central allocation seemed to be performed
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not described, but it seemed to be an open‐label study
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All radiological examinations were reviewed by radiologists blinded to the treatment group of the participant
Incomplete outcome data (attrition bias) All outcomes	Low risk	This study clearly stated the process from participant enrolment to allocation and outcome assessment; low attrition rate
Selective reporting (reporting bias)	Unclear risk	There was no protocol or clinical trial registration for this study.
Other bias	Low risk	This study transparently described the study process, and other apparent biases were not identified.

Morrissey 2013.

Study characteristics
Methods	Study design: open‐label, parallel‐group, randomised controlled trial Country: Australia Number of centres: 6
Participants	Participants were aged 18 years or older, were undergoing allogeneic stem‐cell transplantation or intensive combination induction‐consolidation chemotherapy for acute myeloid or lymphoblastic leukaemia, and had given written informed consent. Participants with any other immunocompromised state, a previous diagnosis of proven or probable invasive aspergillosis, or active invasive fungal disease, or if they had previously been enroled in the trial, or were currently enroled in another trial of antifungal treatment, were excluded
Interventions	Pre‐emptive therapeutic protocol:  Either a single positive galactomannan, or PCR result, or serially negative results for both tests (in participants with persistently neutropenic fevers) prompted a high‐resolution chest CT. Authors assigned the same weight to PCR that was assigned to galactomannan by the EORTC/MSG criteria; adopted stricter criteria for possible invasive aspergillosis, which required mycological evidence in addition to clinical factors; interpreted intermittently positive results, as recommended by Halliday and colleagues; and classified a single positive result with no clinical features as not indicative of invasive aspergillosis. Antifungal treatment was recommended when the criteria for probable or possible invasive aspergillosis, based on these results, were met. Because authors recognised that situations might occur in which the treating clinician wished to use antifungal treatment for a participant assigned to the biomarker diagnosis group, the protocol specified that treatment could be started after discussion with the principal investigators, while results of the investigations were awaiting. Empirical therapeutic protocol:   Consistent with the 2002 guidelines for antimicrobial use in people with neutropenia with cancer. When an invasive fungal disease was suspected (e.g. because of persistent fever, defined as at least one temperature per day of 38·3 °C or higher, or a temperature of 38·0 °C or higher sustained for at least 1 h on 3 to 5 consecutive days), cultures of blood, urine, sputum, and faeces, and high‐resolution CT scans of the chest were done. Bronchoscopy and biopsy sampling were done in accordance with institutional protocols. Empirical treatment with antifungal drugs was recommended while participants were undergoing these investigations, and was continued, de‐escalated to prophylaxis, or the antifungal drug was changed to treat definite, probable, or possible invasive aspergillosis or other invasive fungal disease according to test results.
Outcomes	Time point for assessment was 26 weeks after randomisation Primary outcome:   1. the proportion of participants who received at least one course of empirical antifungal treatment  Secondary outcome: 1. All‐cause mortality 2. Mortality related to invasive aspergillosis 3. Mortality related to other invasive fungal disease 4. Incidence of proven, probable, and possible invasive aspergillosis or other invasive fungal disease 5. Frequency of hepatotoxic effects (defined as a three‐times increase in the results of baseline liver function tests) and nephrotoxic effects (defined as a two‐times increase in baseline serum creatinine concentration) 6. Number of courses of empirical antifungal treatment given
Notes	Study Period: 30 September 2005 to 19 November 2009 Funding source: Australian National Health and Medical Research Council, Cancer Council New South Wales, Pfizer, Merck, Gilead Sciences. The sponsors of the study had no role in trial design, data collection, data analysis, data interpretation, or writing of the report. Declarations of interest among the researchers: COM was a member of advisory boards for, received investigator‐initiated grants from, and given lectures for Gilead Sciences, Pfizer, Merck Sharp & Dohme, and Orphan Australia. SC‐AC received investigator‐initiated grants from, and was a member of advisory boards for Pfizer, Gilead Sciences, and Merck Sharp & Dohme. TCS was a member of advisory boards for, and has received investigator‐initiated grants from Pfizer and Merck Sharp & Dohme. KFB received consultancy payments from Pfizer, Merck Sharp & Dohme, and Gilead Sciences. NMG was a member of advisory boards for Schering Plough (2006) and Pfizer (2012), and gave talks at events sponsored by Pfizer, Merck Sharp & Dohme, and Gilead Sciences. AB received a travel grant to a conference from Merck Sharp & Dohme. ART received a travel grant to a conference from Pfizer. MAS was a member of advisory boards for, and received research funding from Pfizer, Merck Sharp & Dohme, Schering Plough, and Gilead Sciences. SM, PGB, JS, CLH, JM, APS, SG, and TS declare that they have no conflicts of interest.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The computer‐generated, sequentially numbered, randomisation schedule was prepared by the Australasian Leukaemia Lymphoma Group trial centre to ensure that roughly equal numbers of participants were assigned to each group within each centre.
Allocation concealment (selection bias)	Low risk	Centrally allocated
Blinding of participants and personnel (performance bias) All outcomes	High risk	Open‐label study
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The assigned diagnostic group was concealed from the independent data review committee
Incomplete outcome data (attrition bias) All outcomes	Low risk	This study clearly stated the process from participant enrolment to allocation and outcome assessment; attrition rate was 12.9% (31/240)
Selective reporting (reporting bias)	Low risk	There was a clinical trial registration and no apparent selective reporting was identified.
Other bias	Low risk	This study transparently described the study process, and other apparent biases were not identified.

Tan 2011.

Study characteristics
Methods	Study design: prospective, randomised, non‐blinded study Country: Singapore Number of centres: 1
Participants	People at least 12 years of age or older, with a Karnofsky score of 70% and any of the following conditions. Newly diagnosed or relapsed acute leukaemia, or high‐risk MDS receiving induction or salvage chemotherapy with expected duration of neutropenia (ANC < 0.5 × 109/L) of at least 10 days  SAA receiving chemotherapy or immunosuppressive therapy using antithymocyte globulin Receiving allogeneic/autologous HSCT using myeloablative conditioning regimens Consolidation regimens HyperCVAD type B or HIDAC with expected duration of neutropenia (ANC < 0.5 × 109/L) of at least 10 days
Interventions	Pre‐emptive therapeutic protocol:  Serial GM monitoring twice a week. The study co‐ordinator informed the primary haematology team of the participant’s inclusion in the study and the arm into which the participant was randomised, and documented this in the case records. A positive GM index, available the next day, triggered an immediate repeat, as well as an urgent CT scan of the thorax. A second consecutive positive GMI, or a single positive GMI plus a highly suggestive CT thorax15 would lead to the commencement of CAS, AmB, LAmB, or VCZ. Otherwise, GMmonitoring was continued Empirical therapeutic protocol:   Standard of care in accordance with international and departmental guidelines.
Outcomes	Primary outcome:   1. The use of a broad spectrum antifungal agent, either CAS, AmB, LAmB or VCZ. Secondary outcome: 1. Twelve‐week survival
Notes	Study Period: 14 June 2006 to 15 October 2007 Funding source: the National Medical Research Council (Singapore) via grant NMRC/0984/2005 and a SingHealth Foundation grant. Declarations of interest among the researchers: Dr Tan received travel grants from Pfizer. Dr Goh lectured at symposia organized by MSD and Pfizer. Dr Lin owns stock in Pfizer. Dr Goh received research funding from Novartis Oncology and Janssen Cilag and honoraria from Novartis Oncology, Janssen Cilag, Celgene, and Bristol Myers Squibb. Dr Kurup received travel grants from Pfizer.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomisation list was generated by the statistician.
Allocation concealment (selection bias)	Low risk	The study coordinator enroled the participants according the randomisation list.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Open‐label study
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Laboratory test was handled at an independent facility
Incomplete outcome data (attrition bias) All outcomes	Low risk	This study clearly stated the process from patient enrolment to allocation and outcome assessment; low attrition
Selective reporting (reporting bias)	Low risk	There was a clinical trial registration and no apparent selective report was identified.
Other bias	Low risk	This study transparently described the study process, and other apparent biases were not identified.

Yuan 2016.

Study characteristics
Methods	Study design: open‐label, parallel‐group, randomised trial Country: China Number of centres: one
Participants	Older than 18 years, had haematological malignancy, had a plan of chemotherapy treatment, and had severe neutropenia (less than 0.5 × 109 neutrophils/L for a minimum duration of 10 days).  Exclusion criteria: had received haematopoietic stem cell transplantation, had experience or symptoms of IFD, demonstrated low tolerance to triazoles, HIV‐positive, and Karnofsky Performance Status score less than 30, according to the accepted scale system, had no neutropenia or fever, were treated with antifungal therapy within 4 days of developing a fever, started antifungal therapy without proof of IFD at 4 to 14 days, or had no antifungal treatment but had proof of IFD at 4 to 14 days.
Interventions	Pre‐emptive therapeutic protocol:  Antifungal therapy was initiated immediately if any of the flowing situations occurred: clinical or imaging examination suggested pneumonia, acute sinusitis, stage III mucositis, or infectious shock, IFD‐related skin damage, central nerve system symptoms due to unknown reason, periorbital inflammation, abscess of liver or spleen, severely diarrhoea, colonisation by Aspergilloma, or (1, 3)‐b‐D‐glucan test (G test)‐positive and galactomannan test (GM test)‐positive Empirical therapeutic protocol:   Antifungal therapy was started within 4 days of start of persistent fever and antibacterial treatment. If there was recurrent fever between day 4 and day 14, antifungal therapy was initiated on the day of the recurrence.
Outcomes	Time points for assessment:  Primary outcome:   1. The survival rate of participants who recovered from neutropenia by Day 14 Secondary outcome: 1.  The ratio of participants diagnosed with IFD and duration of persistent fever 2. Rate of participants quitting treatment due to ineffectiveness or adverse events during the treatment
Notes	Study period: October 2013 to December 2014 Funding source: departmental sources Declarations of interest among the researchers: none
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	High risk	Open‐label study
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low attrition
Selective reporting (reporting bias)	Unclear risk	There were no protocols or clinical trial registrations for this study.
Other bias	Unclear risk	This study lacked sufficient information to assess the other biases.

Abbreviations: acute myeloid leukaemia, AML; myelodysplastic syndrome, MDS; haematopoietic stem cell transplantation, HSCT; galactomannan, GM; graft‐vs‐host disease, GVHD; real‐time Polymerase Chain Reaction, rtPCR; high‐resolution computed tomography, HRCT; European Organization for Research and Treatment of Cancer/Mycoses Study Group, EORTC/MSG; invasive fungal infection, IFI; invasive fungal disease, IFD;central nervous system, CNS; Enzyme‐Linked Immuno Sorbent Assay, ELISA; severe adverse event, SAE; Bone Marrow Transplantation, BMT; absolute neutrophil count、ANC, caspofungin, CAS; amphotericin B, AmB; liposomal amphotericin, LAmB; voriconazole, VCZ; SAA, severe aplastic anaemia

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Kanda 2020	Authors' discussions and enquiries to Kanda concluded that D‐Index‐Guided Early Antifungal Therapy is different in concept from pre‐emptive therapy.
Pauw 2006	Not a RCT report

Differences between protocol and review

Due to lack of data, we were unable to perform any pre‐planned subgroup analyses. Since the number of studies identified was small, we only completed a meta‐analysis of the primary end point, without dividing the evaluation period. We could only complete sensitivity analyses for the risk of bias and evaluation period. 

Contributions of authors

YU was the contact person with the editorial base.
YU co‐ordinated the contributions from the co‐authors and wrote the final draft of the protocol.
YU, HI, YM, KT, NW worked on the methods sections.
YU, HI, YM, KT, NW drafted the clinical sections of the background and responded to the clinical comments of the referees.
YU, HI, YM, KT, NW responded to the methodological and statistical comments from the referees.
YU, HI, YM, KT, NW contributed to writing the protocol.
HI, YM, and KT were the consumer co‐authors and checked the protocol for readability and clarity. They also ensured that the outcomes were relevant to consumers.
YU is the guarantor of the final review.

Sources of support

Internal sources

• none to declare, Other

  none

External sources

• National Institute of Health Research , UK

  GNOC Infrastrucutre Funding

Declarations of interest

Yu Uneno: none known
Haruki Imura: none known
Yosuke Makuuchi: has received lecture fees from Merck, Sharp and Dohme and Bristol‐Myers Squibb on topics outside the scope of this review.
Kentaro Tochitani: none known
Norio Watanabe: none known

New