Abstract

Objectives

Patients receiving therapy for haematological malignancies have a higher risk of invasive fungal infections (IFIs). Antifungal prophylaxis is an effective strategy against IFIs, but relative effectiveness estimates across agents are inconclusive. A mixed treatment comparison (MTC) was conducted to estimate the relative effectiveness of all agents for a number of outcomes of interest.

Methods

A systematic review was performed to collect evidence from randomized controlled trials (RCTs) on the risk of IFIs and on mortality after antifungal prophylaxis. The agents analysed were no prophylaxis/placebo, fluconazole, itraconazole, micafungin, caspofungin, liposomal amphotericin B and posaconazole. Meta-analyses and MTCs were used to synthesize the evidence. The primary outcome was the risk of proven or probable IFI. Secondary outcomes were risk of candidiasis/aspergillosis, risk of IFI mortality and risk of all-cause mortality.

Results

Antifungal prophylaxis was more effective than no prophylaxis/placebo in reducing IFI risk. The IFI risk after voriconazole or posaconazole was lower than after fluconazole [relative risk (RR) 0.38, 95% CI 0.14–0.83 and RR 0.34, 95% CI 0.14–0.83] or itraconazole tablets (RR 0.22 95% CI 0.06–0.72 and RR 0.20 95% CI 0.05–0.72). Posaconazole was also found to be more effective than no prophylaxis/placebo in reducing all-cause mortality (RR 0.56, 95% CI 0.30–0.98). Posaconazole had the highest probability of being the most effective agent in reducing IFI risk and all-cause mortality.

Conclusions

IFI prophylaxis has a positive effect on IFI risk reduction. However, its effect on all-cause mortality is not as pronounced. The analysis has additionally pinpointed posaconazole as potentially the most effective IFI prophylaxis in neutropenic patients.

Introduction

Patients with haematological malignancies undergoing chemotherapy or haematopoietic stem-cell transplantations (HSCTs) are at high risk of acquiring invasive fungal infections (IFIs). Because of the increasing population of this and other at-risk patient groups, the incidence of IFIs has increased over the last two decades.1–4 In acute leukaemia and HSCT patients, overall incidence rates of 5%–40% have been reported across different countries and institutions.5–7

Like all serious complications, IFIs are associated with high morbidity and mortality. The predominant causative agents of IFI are the Candida and Aspergillus species. In most epidemiological studies, Candida albicans is the most common cause of candidiasis, but cases of non-albicans Candida species are increasing.8,9 In a prospective, multicentre surveillance study, almost half of all nosocomial invasive candidaemia and 65% of breakthrough candidaemia cases were due to non-albicans Candida species.10 Epidemiological studies have also reported an increase in mould infections over the last decade.11 The incidence of invasive aspergillosis (IA) infections may be as high as 10%–20% in patients undergoing allogeneic HSCT.12 Moreover, mortality rates from aspergillosis have been reported to be at least 50% in patients with neutropenia alone and 86% in patients who have undergone HSCT.13

Early treatment of IFIs is a key factor in improving health outcomes. However, antifungal treatment is often delayed because of difficulties relating to non-specific symptoms of IFIs, diagnostic challenges and the high cost of IFI treatment. Despite advances in diagnostic tools, the accurate and timely diagnosis of IFIs remains difficult.14,15Although antifungal prophylaxis is proven to be an effective strategy in reducing the incidence of IFIs in high-risk settings in haemato-oncology, the choice of antifungal prophylaxis remains controversial.16,17 Furthermore, there are several concerns with regards to prophylaxis, including higher costs, toxicity, emergence of drug resistance and shifting fungal epidemiology.18

Evidence-based, clinical or policy-related decisions require the comparison of all relevant treatment options. This requirement, however, is almost never met within the randomized controlled trial (RCT) setting, which is prohibited by cost and regulatory approval-driven strategies. As a consequence, new treatments are not compared with each other, but are compared with standard therapies or placebo.19,20

Mixed treatment comparison (MTC) models (also called network meta-analyses) allow the comparison of all relevant treatment options, including those for which head-to-head comparisons do not exist.21,22 The synthesis of a network of evidence serves two functions. First, MTC models may strengthen the relative efficacy estimate between two treatments through the inclusion of both direct and indirect comparisons.21 Second, MTCs facilitate the simultaneous comparison of all treatments and allow the designation of the best treatment. The methods used in MTC are consistent with and represent a generalization of standard pair-wise meta-analysis methods.21

This study aimed to identify the optimal antifungal prophylactic strategy in preventing IFIs among patients undergoing chemotherapy or HSCT procedures for haematological malignancies. MTC methodologies are used to capture the whole body of available RCT evidence within a single evidence-synthesis framework. Bayesian approaches to fixed- and random-effect MTC regression models are used to estimate the relative effectiveness of fluconazole, itraconazole, voriconazole, posaconazole, micafungin, caspofungin, liposomal amphotericin B and placebo against a number of outcomes, including overall IFIs, invasive candidiasis (IC), IA and IFI- and non-IFI-related mortality. In addition, the strategies are ranked in accordance with their probability of being the best treatment option.

Methods

Search strategy and study selection

A systematic literature review was performed to collect evidence to create a network of relative effectiveness of antifungal agents. In particular, searches were conducted in MEDLINE, Embase, the US NIH clinical trials registry (http://www.clinicaltrials.gov) and Google Scholar during April 2013 using the following search terms: (‘invasive fungal infections’ or ‘ifi’ or ‘fungus’ or ‘fungal’ or ‘fungemia’ or ‘mycosis’ or ‘candidiasis’ or ‘Candida’ or ‘Aspergillus’ or ‘invasive mould infections’ or ‘imi’ or ‘aspergillosis’) and (‘prophylaxis’ or ‘prophylactic’ or ‘prevention’) and (‘antifungal’ or ‘amphotericin’ or ‘azoles’ or ‘triazoles’ or ‘fluconazole’ or ‘itraconazole’ or ‘isavuconazole’ or ‘voriconazole’ or ‘posaconazole’ or ‘ravuconazole’ or ‘echinocandin’ or ‘micafungin’ or ‘caspofungin’ or ‘anidulafungin’). The search was limited to randomized trials using the Cochrane Highly Sensitive Search Strategy for identifying randomized trials.23 In addition, bibliographies of selected studies and systematic reviews on the topic were perused for relevant studies.

All RCT studies on antifungal agents that were newly introduced or are currently being used as IFI prophylaxis among adult patients with haematological malignancies undergoing chemotherapy or HSCT were eligible for inclusion.24 However, studies were limited to those written in English and published in international peer-reviewed journals. Because the focus was on IFI, studies that reported only non-invasive, single-site fungal infections were excluded. Additionally, studies that investigated antifungal agents that are not suggested for prophylactic use by current guidelines or agents of outdated formulations such as (non-liposomal) amphotericin B, ketoconazole, miconazole and nystatin were excluded. However, placebo-designed studies were included to serve as a baseline measure for estimating the relative effectiveness of each antifungal agent. In the analysis, no relative effectiveness difference between placebo and no prophylaxis was assumed. Studies that analysed only graft-versus-host disease patients were excluded as this patient population is at a significantly higher risk of IFI and inclusion of such studies was expected to increase the heterogeneity within the evidence network.25,26 Finally, because of well-documented pharmacokinetic differences between the available administration forms of itraconazole,27,28 we distinguished between itraconazole tablets and itraconazole intravenous/oral solution.

Two investigators (P. P. and H. H. L.) independently scanned all titles and abstracts for RCT studies investigating the prophylactic use of antifungal agents among high-risk neutropenic patients. Following identification of relevant abstracts, the full text of each publication was obtained and thoroughly evaluated for inclusion. Risk of bias was assessed on the basis of sequence allocation, allocation concealment, blinding, incomplete outcomes and intention-to-treat analysis in accordance with the Cochrane tool for bias assessment.23 Studies that reported no cases of an outcome in both treatment arms were excluded from the analysis on that outcome.

Data collection

The same investigators also independently conducted the data extraction. Using pre-formulated data extraction sheets, the primary and secondary outcome measures were recorded. The primary outcome was the total number of proven or probable IFIs, while secondary outcomes were the number of proven or probable IFIs caused by Candida or Aspergillus as well as IFI-related and all-cause mortality. Whenever possible the definitions from the European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group and the National Institute of Allergy and Infectious Diseases Mycoses Study Group (EORTC/MSG) Consensus Group were used for a proven or probable IFI. Covariates such as age, gender, dosing regimens, maximum days of follow-up, average duration of prophylaxis and neutropenia, proportion of chemotherapy and HSCT patients, and proportion of patients with acute myeloid leukaemia (AML), myelodysplastic syndromes, acute lymphoblastic leukaemia, chronic myeloid leukaemia, chronic lymphocytic leukaemia, non-Hodgkin lymphoma and multiple myeloma as the underlying disease were recorded. Medians were used when means were not reported for age and duration of prophylaxis and neutropenia. Inconsistencies in data extraction were resolved through discussion and review of the material.

Statistical analysis

Meta-regression and network meta-regression approaches were used to synthesize the evidence network.29 The regression approach was used in order to explain any observed between-study heterogeneity and to reduce evidence inconsistency.29 Initially, pairwise mixed-effects Poisson meta-regressions were performed for each outcome separately, using direct comparative information from the studies in the network.30 Subsequently, mixed-effects, Poisson network meta-regression models were developed within a Bayesian framework for each outcome separately. These models were used to estimate the relative effectiveness of each strategy on the outcomes of interest. A more detailed description of the structure of the MTC regression models is given below.

For a set of i RCT studies and a network of k treatments, the number of patients experiencing an event of interest (e.g. IFI) Yik was assumed to follow a Poisson distribution: 

formula
where Eik is the offset parameter controlling for differences in sample sizes and follow-up time in person (Nik) days (Tik) (hence Eik = Tik Nik) and λik is the corresponding mean parameter, which is linked to the prophylactic effect and the covariates Xik through the log link: 
formula

Above, the parameter μi is the study-specific baseline estimate and δik is the study- and treatment-specific, relative effectiveness, random effect estimates. The indicator function I indicates that δik exists only when k ≠ b, where b is the baseline, study-specific treatment. For δik we made the assumption that it is normally distributed with mean dk − dib and variance 1/τ2, where dk − dib is the average relative effectiveness of treatment k against the baseline treatment of study i and τ2 is the between-study precision parameter. Finally, βi represents the trial-specific, random coefficients of the covariates Xik. All continuous covariates were standardized to the mean. Additionally, an alternative model was applied to the above model, where δik was assumed to be fixed. The deviance information criterion (DIC) was used to facilitate the decision on the final form of the model (e.g. fixed or random effects and number of covariates included).31

Since a Bayesian approach to the estimation of the above model was used, prior probabilities were assumed for each parameter. However, because no prior information to direct the posterior effectiveness estimates was available, a vague normal prior probability [N(0,1000)] was used. Additionally, a gamma probability [Γ(0.001,0.001)] was used for the prior distribution describing the precisions of the random effect parameters. All models were simulated for 100 000 iterations. Treatment options were ranked at every iteration according to their effectiveness from best to worst. Subsequently, the proportion of iterations in which each treatment option was ranked best, second best and so on was calculated. This proportion served as a proxy for ranking the treatment options.

A major assumption of MTC models is that of consistency among direct and indirect comparisons in a network.31 In other words, MTC models make the assumption that direct and indirect comparisons do not disagree beyond chance. In order to investigate the potential presence of inconsistency within loops, an illustrative informal inconsistency check method was used.32 Convergence of the model was assessed through (i) visual inspection of the trace plots, (ii) calculation of potential scale reduction factors33 and (iii) inspection of convergence in multiple chains with different initial values.

All estimations were performed using the statistical software R, ver. 2.14.1 (R Development Core Team, Vienna, Austria) and WinBUGS, ver. 1.434 (MRC Biostatistics Unit, Cambridge, UK). Missing data in any of the covariates were handled using a cross-sectional imputation method.35

Results

Through the systematic review, 25 RCTs were identified that satisfied the inclusion and exclusion criteria.16,20,24,3657 One RCT comparing fluconazole and itraconazole solution in patients with allogeneic HSCT was excluded because of the language restriction.58 A flow diagram illustrating the systematic review process is presented in Figure 1. From the included studies, a network of evidence for nine alternative antifungal prophylaxis agents was constructed (Figure 2). The nodes represent antifungal agents and the links between the nodes represent direct comparisons in RCTs. The diameter of the nodes and the width of the links vary according to sample size and number of direct comparisons, respectively. The itraconazole formulations are represented in one node to simplify the figure. With the exception of the study by Cornely et al.,19 there was no multi-arm RCT identified. In that study, application of prophylactic agents was split within the study into different centres. Hence, treatment randomization was not performed across centres. To address this issue, the two comparisons (i.e. posaconazole versus itraconazole solution and posaconazole versus fluconazole) were treated as separate studies.

Figure 1.

Flow chart of the study selection process.

Figure 1.

Flow chart of the study selection process.

Figure 2.

A schematic of the network of evidence used in MTC analysis. The links between the nodes indicate direct comparisons. The number of RCTs between direct comparisons is represented by the thickness of the links and by the numbers next to them. The size of the nodes represents the total number of patients for that treatment. Itraconazole is split into two categories: itraconazole solution (solid pattern) and itraconazole tablets (chequered pattern). P, posaconazole; V, voriconazole; F, fluconazole; I, itraconazole; M, micafungin; C, caspofungin; A, liposomal amphotericin B; Pl, placebo.

Figure 2.

A schematic of the network of evidence used in MTC analysis. The links between the nodes indicate direct comparisons. The number of RCTs between direct comparisons is represented by the thickness of the links and by the numbers next to them. The size of the nodes represents the total number of patients for that treatment. Itraconazole is split into two categories: itraconazole solution (solid pattern) and itraconazole tablets (chequered pattern). P, posaconazole; V, voriconazole; F, fluconazole; I, itraconazole; M, micafungin; C, caspofungin; A, liposomal amphotericin B; Pl, placebo.

Risk of bias

All studies included were RCTs (Table S1, available as Supplementary data at JAC Online). High risk of bias regarding sequence generation and allocation concealment was not found in any of the studies, but 48% and 40% of the studies, respectively, were characterized as having an unclear risk of bias because detailed descriptions of the randomization process were not provided. High risk of bias relating to blinding of participants and personnel was found in 36% of the studies. However, no studies were un-blinded to the outcomes. The majority of the studies relied on per-protocol analysis while others provided unclear information regarding the type of analysis conducted. Finally, the older studies did not use the EORTC/MSG definition for proven or probable IFIs.

Subject characteristics

The selected RCTs comprised a total of 7062 patients receiving one of nine alternative prophylactic strategies. The main patient characteristics of the included studies are presented in Table S2 (available as Supplementary data at JAC Online). Patients were on average 41.9 years of age, 53.5% of them were male and 54.4% were undergoing HSCT or bone marrow transplantation (BMT) at baseline. The most common underlying malignant disease was AML. On average, patients were followed for a maximum of 951 days and the average prophylaxis period was 33.8 days. Dosage regimens varied within studies for itraconazole tablets and solution, fluconazole and liposomal amphotericin B. Additionally, differences across studies in the formulation administered were observed for fluconazole, itraconazole solution and voriconazole. Micafungin, caspofungin and liposomal amphotericin B were only administered parenterally.

The main clinical outcomes reported in each of the included studies are presented in Table 1. An overall proven or probable IFI was identified in 6% of the followed individuals. Of these IFIs, 46.3% were caused by Candida, 38.4% by Aspergillus and 15.3% by other species. Approximately 11% of the followed individuals died during follow-up and around 38% of the patients with an IFI died as a result of the infection.

Table 1.

Main clinical outcomes reported by each of the included studies

Study name Treatment Follow-up (days) Sample size IFI
 
Mortality
 
overall IC IA other overall IFI non-IFI IC IA 
Goodman54 fluconazole 84 179 55 54 
 placebo 84 177 28 25 46 10 36 
Winston36 fluconazole >90 123 26 NR NR NR NR 
 placebo >90 132 10 24 NR NR NR NR 
Schaffner40 fluconazole >40 75 
 placebo >40 76 
Slavin41 fluconazole 110 152 10 31 11 20 NR NR 
 placebo 110 148 32 28 52 19 33 NR NR 
Rotstein45 fluconazole 60 141 15 13 
 placebo 60 133 32 20 11 15 10 
Yamac57 fluconazole NR 41 NR NR NR NR NR 
 placebo/control NR 29 NR NR NR NR NR 
Vreugdenhil38 itraconazole (tablets) 180 46 10 NR NR 
 placebo 180 46 10 14 NR NR 
Menichetti43 itraconazole (solution) 56 201 15 14 
 placebo 56 204 18 13 
Nucci46 itraconazole (tablets) 40 104 
 placebo 40 106 
Tollemar37 liposomal AMB 100 36 16 15 
 placebo 100 40 17 14 
Kelsey42 liposomal AMB 45 74 11 11 
 placebo 45 87 12 11 
Penack51 liposomal AMB 30 75 
 placebo 30 57 20 10 10 
Annaloro55 itraconazole (tablets) 70 31 
 fluconazole 70 28 
Ito56 itraconazole (tablets) NR 103 NR NR NR NR NR 
 fluconazole NR 106 NR NR NR NR NR 
Morgenstern44 itraconazole (solution) 120 288 
 fluconazole 120 293 10 
Winston39 itraconazole (solution) 180 71 32 26 
 fluconazole 180 67 17 28 12 16 NR NR 
Oren50 itraconazole (solution) 90 96 16 16 
 fluconazole 90 99 17 14 
Glasmacher47 itraconazole (solution) 56 248 25 23 
 fluconazole 56 246 28 25 
Hiramatsu52 micafungin 100 50 
 fluconazole 100 50 
van Burik16 micafungin 70 397 18 17 
 fluconazole 70 433 11 26 24 
Wingard20 voriconazole 180 303 14 NR NR NR NR 
 fluconazole 180 289 24 17 NR NR NR NR 
Cornely19 posaconazole 100 239 36 32 NR NR 
 fluconazole 100 240 26 18 55 13 42 NR NR 
 posaconazole 100 65 NR NR 
 itraconazole (solution) 100 58 NR NR 
Mattiuzzi48 caspofungin 38 106 
 itraconazole (solution) 38 86 
Mattiuzzi49 voriconazole 35 71 
 itraconazole (solution) 36 52 
Marks53 voriconazole 180 224 41 
 itraconazole (solution) 180 241 45 
Study name Treatment Follow-up (days) Sample size IFI
 
Mortality
 
overall IC IA other overall IFI non-IFI IC IA 
Goodman54 fluconazole 84 179 55 54 
 placebo 84 177 28 25 46 10 36 
Winston36 fluconazole >90 123 26 NR NR NR NR 
 placebo >90 132 10 24 NR NR NR NR 
Schaffner40 fluconazole >40 75 
 placebo >40 76 
Slavin41 fluconazole 110 152 10 31 11 20 NR NR 
 placebo 110 148 32 28 52 19 33 NR NR 
Rotstein45 fluconazole 60 141 15 13 
 placebo 60 133 32 20 11 15 10 
Yamac57 fluconazole NR 41 NR NR NR NR NR 
 placebo/control NR 29 NR NR NR NR NR 
Vreugdenhil38 itraconazole (tablets) 180 46 10 NR NR 
 placebo 180 46 10 14 NR NR 
Menichetti43 itraconazole (solution) 56 201 15 14 
 placebo 56 204 18 13 
Nucci46 itraconazole (tablets) 40 104 
 placebo 40 106 
Tollemar37 liposomal AMB 100 36 16 15 
 placebo 100 40 17 14 
Kelsey42 liposomal AMB 45 74 11 11 
 placebo 45 87 12 11 
Penack51 liposomal AMB 30 75 
 placebo 30 57 20 10 10 
Annaloro55 itraconazole (tablets) 70 31 
 fluconazole 70 28 
Ito56 itraconazole (tablets) NR 103 NR NR NR NR NR 
 fluconazole NR 106 NR NR NR NR NR 
Morgenstern44 itraconazole (solution) 120 288 
 fluconazole 120 293 10 
Winston39 itraconazole (solution) 180 71 32 26 
 fluconazole 180 67 17 28 12 16 NR NR 
Oren50 itraconazole (solution) 90 96 16 16 
 fluconazole 90 99 17 14 
Glasmacher47 itraconazole (solution) 56 248 25 23 
 fluconazole 56 246 28 25 
Hiramatsu52 micafungin 100 50 
 fluconazole 100 50 
van Burik16 micafungin 70 397 18 17 
 fluconazole 70 433 11 26 24 
Wingard20 voriconazole 180 303 14 NR NR NR NR 
 fluconazole 180 289 24 17 NR NR NR NR 
Cornely19 posaconazole 100 239 36 32 NR NR 
 fluconazole 100 240 26 18 55 13 42 NR NR 
 posaconazole 100 65 NR NR 
 itraconazole (solution) 100 58 NR NR 
Mattiuzzi48 caspofungin 38 106 
 itraconazole (solution) 38 86 
Mattiuzzi49 voriconazole 35 71 
 itraconazole (solution) 36 52 
Marks53 voriconazole 180 224 41 
 itraconazole (solution) 180 241 45 

NR, not reported; AMB, amphotericin B.

Overall IFIs

Estimates from pairwise and network meta-regressions of the relative effectiveness of all agents against IFIs are presented in the upper and lower triangles of Table 2, respectively. Both fixed- and random-effects models were used and the goodness of fit after correction for various study-level covariates (percentages of AML and HSCT/BMT patients and average age) was assessed. Inference from goodness of fit measures revealed that the random-effects models with no explanatory covariates were the most appropriate model form (data not shown).

Table 2.

Meta-analysis (upper triangle) and MTC (lower triangle) RR estimates (and their credible intervals) of an IFI

PLA 0.33 (0.15–0.79) 0.79 (0.21–2.92)   0.14 (0.03–0.78) 0.56 (0.13–2.39)   
0.36 (0.22–0.61)a FLC 0.44 (0.13–1.21) 0.56 (0.29–1.05) 0.54 (0.09–3.29)  3.31 (0.39–28.03) 0.31 (0.14–0.67)  
0.22 (0.11–0.46)a 0.61 (0.35–1.07)a ITC (solution/iv) 0.21 (0.01–1.73)    0.76 (0.27–2.14) 1.14 (0.37–3.46) 
0.14 (0.05–0.34) 0.38 (0.14–0.83)a 0.62 (0.22–1.44)a VRC      
0.22 (0.06–0.73) 0.60 (0.19–1.75)a 0.98 (0.28–3.35) 1.59 (0.40–6.98) MCF     
0.17 (0.05–0.49)a 0.48 (0.12–1.50) 0.78 (0.18–2.75) 1.27 (0.28–5.61) 0.80 (0.13–3.96) LIP AMB    
0.62 (0.27–1.44)a 1.72 (0.671–4.26)a 2.81 (0.96–8.03) 4.54 (1.39–17.37) 2.89 (0.68–12.36) 3.59 (0.94–15.82) ITC (tablets)   
0.12 (0.04–0.35) 0.34 (0.14–0.83)a 0.55 (0.21–1.43)a 0.89 (0.28–3.41) 0.56 (0.13–2.51) 0.71 (0.17–3.66) 0.20 (0.05–0.72) POS  
0.23 (0.05–1.14) 0.64 (0.14–2.92) 1.05 (0.25–4.35)a 1.70 (0.34–9.94) 1.08 (0.16–7.16) 1.35 (0.2–10.01) 0.37 (0.06–2.18) 1.89 (0.34–10.44) CAS 
PLA 0.33 (0.15–0.79) 0.79 (0.21–2.92)   0.14 (0.03–0.78) 0.56 (0.13–2.39)   
0.36 (0.22–0.61)a FLC 0.44 (0.13–1.21) 0.56 (0.29–1.05) 0.54 (0.09–3.29)  3.31 (0.39–28.03) 0.31 (0.14–0.67)  
0.22 (0.11–0.46)a 0.61 (0.35–1.07)a ITC (solution/iv) 0.21 (0.01–1.73)    0.76 (0.27–2.14) 1.14 (0.37–3.46) 
0.14 (0.05–0.34) 0.38 (0.14–0.83)a 0.62 (0.22–1.44)a VRC      
0.22 (0.06–0.73) 0.60 (0.19–1.75)a 0.98 (0.28–3.35) 1.59 (0.40–6.98) MCF     
0.17 (0.05–0.49)a 0.48 (0.12–1.50) 0.78 (0.18–2.75) 1.27 (0.28–5.61) 0.80 (0.13–3.96) LIP AMB    
0.62 (0.27–1.44)a 1.72 (0.671–4.26)a 2.81 (0.96–8.03) 4.54 (1.39–17.37) 2.89 (0.68–12.36) 3.59 (0.94–15.82) ITC (tablets)   
0.12 (0.04–0.35) 0.34 (0.14–0.83)a 0.55 (0.21–1.43)a 0.89 (0.28–3.41) 0.56 (0.13–2.51) 0.71 (0.17–3.66) 0.20 (0.05–0.72) POS  
0.23 (0.05–1.14) 0.64 (0.14–2.92) 1.05 (0.25–4.35)a 1.70 (0.34–9.94) 1.08 (0.16–7.16) 1.35 (0.2–10.01) 0.37 (0.06–2.18) 1.89 (0.34–10.44) CAS 

PLA, placebo; FLC, fluconazole; ITC, itraconazole; VRC, voriconazole; MCF, micafungin; LIP AMB, liposomal amphotericin B; POS, posaconazole; CAS, caspofungin; iv, intravenous.

Significant comparisons are presented in bold. For the upper (lower) triangle RRs lower than 1 favour the column (row)-defining treatment.

aDirect head-to-head RCTs exist for these comparisons.

Through the meta-analyses in Table 2 (upper triangle), it was observed that fluconazole and liposomal amphotericin B were significantly better than placebo/no prophylaxis in reducing proven or probable IFIs. Additionally, posaconazole prophylaxis was found to be more effective in comparison with fluconazole prophylaxis. Through the MTC regression models, all antifungal agents, with the exception of caspofungin and itraconazole tablets, were observed to have a significant prophylactic effect against placebo/no prophylaxis. In addition, posaconazole was found to be significantly more effective than fluconazole and itraconazole tablets in antifungal prophylaxis.

Invasive candidiasis/aspergillosis

When examining the risk of proven or probable IC, all agents, with the exception of micafungin, were observed to be more effective than placebo/no prophylaxis (lower triangle, Table 3). Additionally, itraconazole solution and caspofungin appeared to have a significant advantage against IC compared with fluconazole. With respect to IA infections, posaconazole was found to provide better prophylaxis when compared with placebo/no prophylaxis, fluconazole and itraconazole solution (upper triangle, Table 3). Liposomal amphotericin B and micafungin were significantly more effective against IA infections compared with placebo/no prophylaxis and fluconazole. Finally, voriconazole was found to be significantly more effective than fluconazole. No covariate was found to have a significant impact on the improvement of the model. Thus, an empty random-effects MTC model was applied in both cases.

Table 3.

RR estimates (and their credible intervals) from the MTC on IA (upper triangle) and IC (lower triangle) for all antifungal agents

PLA 1.13 (0.43–2.69)a 0.62 (0.19–1.82)a 0.34 (0.05–1.45) 0.08 (<0.01–0.8) 0.13 (0.01–0.91)a 0.59 (0.08–2.85)a 0.06 (0.01–0.34) 0.52 (0.03–8.64) 
0.14 (0.07–0.25)a FLC 0.55 (0.22–1.24)a 0.30 (0.06–0.96)a 0.07 (<0.01–0.58)a 0.11 (0.01–1.03) 0.52 (0.06–3.28)a 0.05 (0.01–0.25)a 0.47 (0.03–7.25) 
0.06 (0.02–0.16)a 0.44 (0.18–0.99)a ITC (solution/iv) 0.55 (0.11–1.99)a 0.13 (0.01–1.33) 0.20 (0.02–2.18) 0.95 (0.11–6.74) 0.09 (0.01–0.51)a 0.87 (0.06–11.93)a 
0.05 (0.01–0.18) 0.34 (0.08–1.14)a 0.77 (0.19–2.614)a VRC 0.24 (0.01–3.39) 0.37 (0.02–5.93) 1.76 (0.16–19.26) 0.16 (0.01–1.52) 1.58 (0.08–34.00) 
0.18 (0.03–1.07) 1.36 (0.28–7.09)a 3.17 (0.51–19.06) 4.07 (0.55–33.65) MCF 1.66 (0.06–110.69) 7.60 (0.36–418.57) 0.71 (0.03–35.73) 7.34 (0.19–586.24) 
0.19 (0.05–0.56)a 1.40 (0.32–5.09) 3.14 (0.62–15.2) 4.06 (0.66–27.32) 1.00 (0.11–8.06) LIP AMB 4.75 (0.27–82.16) 0.43 (0.02–8.68) 4.31 (0.12–175.57) 
0.26 (0.08–0.77)a 1.92 (0.58–6.23)a 4.35 (1.04–18.51) 5.57 (1.08–34.48) 1.40 (0.19–10.03) 1.37 (0.27–7.49) ITC (tablets) 0.09 (0.01–1.49) 0.94 (0.03–27.16) 
0.14 (0.04–0.58) 1.03 (0.3–3.573)a 2.40 (0.63–9.33)a 3.10 (0.57–18.57) 0.76 (0.1–6.01) 0.76 (0.12–5.07) 0.55 (0.10–3.32) POS 9.81 (0.41–297.5) 
0.02 (0.01–0.14) 0.14 (0.01–1.02) 0.32 (0.04–1.95)a 0.42 (0.03–3.81) 0.10 (0.01–1.30) 0.10 (0.01–1.08) 0.07 (0.01–0.65) 0.13 (0.01–1.19) CAS 
PLA 1.13 (0.43–2.69)a 0.62 (0.19–1.82)a 0.34 (0.05–1.45) 0.08 (<0.01–0.8) 0.13 (0.01–0.91)a 0.59 (0.08–2.85)a 0.06 (0.01–0.34) 0.52 (0.03–8.64) 
0.14 (0.07–0.25)a FLC 0.55 (0.22–1.24)a 0.30 (0.06–0.96)a 0.07 (<0.01–0.58)a 0.11 (0.01–1.03) 0.52 (0.06–3.28)a 0.05 (0.01–0.25)a 0.47 (0.03–7.25) 
0.06 (0.02–0.16)a 0.44 (0.18–0.99)a ITC (solution/iv) 0.55 (0.11–1.99)a 0.13 (0.01–1.33) 0.20 (0.02–2.18) 0.95 (0.11–6.74) 0.09 (0.01–0.51)a 0.87 (0.06–11.93)a 
0.05 (0.01–0.18) 0.34 (0.08–1.14)a 0.77 (0.19–2.614)a VRC 0.24 (0.01–3.39) 0.37 (0.02–5.93) 1.76 (0.16–19.26) 0.16 (0.01–1.52) 1.58 (0.08–34.00) 
0.18 (0.03–1.07) 1.36 (0.28–7.09)a 3.17 (0.51–19.06) 4.07 (0.55–33.65) MCF 1.66 (0.06–110.69) 7.60 (0.36–418.57) 0.71 (0.03–35.73) 7.34 (0.19–586.24) 
0.19 (0.05–0.56)a 1.40 (0.32–5.09) 3.14 (0.62–15.2) 4.06 (0.66–27.32) 1.00 (0.11–8.06) LIP AMB 4.75 (0.27–82.16) 0.43 (0.02–8.68) 4.31 (0.12–175.57) 
0.26 (0.08–0.77)a 1.92 (0.58–6.23)a 4.35 (1.04–18.51) 5.57 (1.08–34.48) 1.40 (0.19–10.03) 1.37 (0.27–7.49) ITC (tablets) 0.09 (0.01–1.49) 0.94 (0.03–27.16) 
0.14 (0.04–0.58) 1.03 (0.3–3.573)a 2.40 (0.63–9.33)a 3.10 (0.57–18.57) 0.76 (0.1–6.01) 0.76 (0.12–5.07) 0.55 (0.10–3.32) POS 9.81 (0.41–297.5) 
0.02 (0.01–0.14) 0.14 (0.01–1.02) 0.32 (0.04–1.95)a 0.42 (0.03–3.81) 0.10 (0.01–1.30) 0.10 (0.01–1.08) 0.07 (0.01–0.65) 0.13 (0.01–1.19) CAS 

PLA, placebo; FLC, fluconazole; ITC, itraconazole; VRC, voriconazole; MCF, micafungin; LIP AMB, liposomal amphotericin B; POS, posaconazole; CAS, caspofungin; iv, intravenous.

Significant comparisons are presented in bold. For the upper (lower) triangle RRs lower than 1 favour the column (row)-defining treatment.

aDirect head-to-head RCTs exist for these comparisons.

All-cause/IFI-related mortality

MTC estimates of the relative effectiveness of all antifungal agents on all-cause and IFI-related mortality are summarized in Table 4. Posaconazole was the only agent achieving a significant reduction in the risk of all-cause mortality in comparison with placebo/no prophylaxis. With respect to reducing IFI-related mortality, all agents except micafungin and caspofungin were found to be significantly superior to placebo in reducing IFI-related mortality. Additionally, posaconazole was found to be superior to fluconazole and itraconazole tablets. The relative effectiveness of voriconazole could not be determined due to the absence of IFI-related deaths in all voriconazole studies.

Table 4.

RR estimates (and their credible intervals) from the MTC on all-cause mortality (upper triangle) and IFI mortality (lower triangle) for all antifungal agents

PLA 0.90 (0.68–1.20)a 0.83 (0.56–1.23)a 0.78 (0.46–1.25) 0.67 (0.32–1.40) 0.85 (0.49–1.43)a 0.79 (0.41–1.51)a 0.56 (0.3–0.98) 0.63 (0.19–2.03) 
0.38 (0.16–0.83)a FLC 0.93 (0.68–1.25)a 0.87 (0.56–1.3)a 0.75 (0.38–1.46)a 0.95 (0.5–1.69) 0.87 (0.43–1.74)a 0.62 (0.36–1.01)a 0.70 (0.22–2.20) 
0.13 (0.03–0.37)a 0.33 (0.12–0.77)a ITC (solution/iv) 0.94 (0.62–1.41)a 0.81 (0.38–1.71) 1.01 (0.52–1.97) 0.94 (0.45–1.98) 0.67 (0.37–1.16)a 0.75 (0.24–2.30)a 
- (-) - (-) - (-) VRC 0.86 (0.4–1.91) 1.07 (0.52–2.28) 1.01 (0.45–2.24) 0.71 (0.36–1.33) 0.80 (0.24–2.69) 
0.13 (0.01–0.97) 0.34 (0.04–2.10)a 1.04 (0.11–8.55) - (-) MCF 1.26 (0.5–3.19) 1.17 (0.45–3.08) 0.83 (0.34–1.92) 0.95 (0.22–3.49) 
0.18 (0.03–0.70)a 0.47 (0.07–2.39) 1.39 (0.19–9.72) - (-) 1.37 (0.09–21.35) LIP AMB 0.94 (0.4–2.17) 0.66 (0.29–1.45) 0.74 (0.2–2.78) 
1.08 (0.32–3.93)a 2.84 (0.71–14.37)a 8.47 (1.74–63.09) - (-) 8.52 (0.80–127.7) 6.19 (0.98–54.67) ITC (tablets) 0.71 (0.29–1.68) 0.80 (0.19–3.02) 
0.09 (0.01–0.34) 0.22 (0.05–0.74)a 0.67 (0.15–2.58)a - (-) 0.63 (0.06–7.9) 0.47 (0.05–4.37) 0.08 (0.01–0.47) POS 1.12 (0.31–3.94) 
0.16 (0.01–1.51) 0.43 (0.05–3.66) 1.27 (0.18–10.00)a - (-) 1.26 (0.07–27.31) 0.94 (0.05–15.28) 0.15 (0.01–1.84) 1.90 (0.19–22.25) CAS 
PLA 0.90 (0.68–1.20)a 0.83 (0.56–1.23)a 0.78 (0.46–1.25) 0.67 (0.32–1.40) 0.85 (0.49–1.43)a 0.79 (0.41–1.51)a 0.56 (0.3–0.98) 0.63 (0.19–2.03) 
0.38 (0.16–0.83)a FLC 0.93 (0.68–1.25)a 0.87 (0.56–1.3)a 0.75 (0.38–1.46)a 0.95 (0.5–1.69) 0.87 (0.43–1.74)a 0.62 (0.36–1.01)a 0.70 (0.22–2.20) 
0.13 (0.03–0.37)a 0.33 (0.12–0.77)a ITC (solution/iv) 0.94 (0.62–1.41)a 0.81 (0.38–1.71) 1.01 (0.52–1.97) 0.94 (0.45–1.98) 0.67 (0.37–1.16)a 0.75 (0.24–2.30)a 
- (-) - (-) - (-) VRC 0.86 (0.4–1.91) 1.07 (0.52–2.28) 1.01 (0.45–2.24) 0.71 (0.36–1.33) 0.80 (0.24–2.69) 
0.13 (0.01–0.97) 0.34 (0.04–2.10)a 1.04 (0.11–8.55) - (-) MCF 1.26 (0.5–3.19) 1.17 (0.45–3.08) 0.83 (0.34–1.92) 0.95 (0.22–3.49) 
0.18 (0.03–0.70)a 0.47 (0.07–2.39) 1.39 (0.19–9.72) - (-) 1.37 (0.09–21.35) LIP AMB 0.94 (0.4–2.17) 0.66 (0.29–1.45) 0.74 (0.2–2.78) 
1.08 (0.32–3.93)a 2.84 (0.71–14.37)a 8.47 (1.74–63.09) - (-) 8.52 (0.80–127.7) 6.19 (0.98–54.67) ITC (tablets) 0.71 (0.29–1.68) 0.80 (0.19–3.02) 
0.09 (0.01–0.34) 0.22 (0.05–0.74)a 0.67 (0.15–2.58)a - (-) 0.63 (0.06–7.9) 0.47 (0.05–4.37) 0.08 (0.01–0.47) POS 1.12 (0.31–3.94) 
0.16 (0.01–1.51) 0.43 (0.05–3.66) 1.27 (0.18–10.00)a - (-) 1.26 (0.07–27.31) 0.94 (0.05–15.28) 0.15 (0.01–1.84) 1.90 (0.19–22.25) CAS 

PLA, placebo; FLC, fluconazole; ITC, itraconazole; VRC, voriconazole; MCF, micafungin; LIP AMB, liposomal amphotericin B; POS, posaconazole; CAS, caspofungin; iv, intravenous.

Significant comparisons are presented in bold. For the upper (lower) triangle RRs lower than 1 favour the column (row)-defining treatment.

aDirect head-to-head RCTs exist for these comparisons.

Ranking and inconsistency

The ranking distributions in Figure 3 represent the proportions of simulations in which each agent was ranked in each position (from ‘best option’ to ‘worst option’) based on its effectiveness against proven or probable IFIs and all-cause mortality. With respect to prophylaxis against IFIs, posaconazole was pinpointed as the most preferable agent in more than one-third of the simulations (35%), followed by voriconazole (25%), liposomal amphotericin B (20%), caspofungin (11%), micafungin (9%), itraconazole solution, itraconazole tablets, fluconazole and placebo (0%). With respect to reduction in all-cause mortality risk, posaconazole was again the most preferable prophylaxis agent (34%), followed by caspofungin, (33%), micafungin (18%), itraconazole tablets (8%), liposomal amphotericin B (4%), voriconazole (2%), itraconazole solution (1%), fluconazole (0%) and placebo (0%).

Figure 3.

Ranking distributions of the alternative prophylaxis strategies with respect to IFI (black bars) and all-cause mortality (grey bars) risk reduction from 100 000 simulations. The horizontal axes depict the order from best (first) to worst (ninth) prophylactic option.

Figure 3.

Ranking distributions of the alternative prophylaxis strategies with respect to IFI (black bars) and all-cause mortality (grey bars) risk reduction from 100 000 simulations. The horizontal axes depict the order from best (first) to worst (ninth) prophylactic option.

We identified no evidence of inconsistency in all MTC analyses and in any of the closed loops in the data, as inspected through the coherence plots. Finally, sensitivity analysis on the prior assumptions showed no significant deviations on the relative risk estimates from all MTC analyses.

Discussion

This study compared using an MTC method the effectiveness of nine IFI prophylaxis strategies for patients with neutropenia who were undergoing chemotherapy or HSCT. The results of the analysis indicate that IFI prophylaxis has a clear positive effect on IFI risk reduction, but its effect on all-cause mortality might not be as pronounced. The analysis has additionally supported posaconazole as potentially the most effective antifungal agent for prophylaxis against IFIs in neutropenic patients, especially those undergoing chemotherapy. Additionally, its relative effectiveness against Aspergillus was more pronounced compared with that against Candida infections. Pathogen-specific analysis also identified the inability of fluconazole to protect against Aspergillus species, as previously documented.24,59 Differences in effectiveness across itraconazole formulations were also observed in our study, with the solution formulation performing better than the tablets.

The results of this study are consistent with the recommendations made in previous studies.24,60,61 In particular, the order of antifungal preference identified through our model matches that recommended in a recent study of the German Society for Haematology and Oncology.24 Additionally, the importance of prophylaxis in patients with neutropenia became even more evident through our analysis, where prophylaxis with any of the antifungal agents that are recommended for use in the UK (based on the British National Formulary) was found to be significantly more effective than placebo/no prophylaxis.

Reduction in all-cause mortality is generally difficult to achieve in RCTs with patients with severe neutropenia, and the results of this study support this finding. Regarding IFI-related mortality, however, the benefit from prophylaxis has been more obvious in our analysis, with posaconazole achieving the highest effectiveness among antifungals. The results from the MTC were fairly robust and no severe problems of inconsistency were observed in the data. We have found no evidence supporting our literature-based hypothesis that studies with a higher proportion of AML patients would show a higher risk of acquiring an IFI, compared with other types of leukaemia.4

Limitations and future work

Some RCT studies included in this analysis suffer from inconsistent reporting; mean estimates for the covariates studied were not always present, forcing the use of medians, where available, as the best estimate of the mean. Lack of reporting was also observed at the outcome level, since four studies failed to report IFI-related mortality rates.20,36,56,57 In some studies the underlying disease status was also inconsistently reported. Despite the fact that the methods used have allowed us to utilize the information collected in the most efficient way, the collection of unreported data from the corresponding authors of the selected studies would be a more accurate solution.

Another major limitation in this analysis is that the underlying severity of the studied population makes isolation of the treatment effect difficult. As Rogers et al.62 pinpointed, a prophylactic effect on IFI incidence or IFI-related mortality might be related to difficulties with diagnosis of IFI because of the underlying severity rather than the effect of the prophylaxis itself. In addition to the last limitation, the variation observed in the dosing regimens between RCTs complicates the estimation of an aggregate prophylactic effect for each strategy. Also, no distinction was made between patients with neutropenia or immunosuppression due to chemotherapy or due to HSCT/BMT. Hence our results refer to a composite population which is potentially much more heterogeneous than we had assumed. A scenario analysis where results would be stratified depending on the underlying condition might be a more realistic alternative and a task for further research. From the current analysis, a summary and comparison of adverse effectiveness among different agents is missing as we were unable to synthesize this information quantitatively. Finally, the evidence synthesis process followed might have missed potential studies that would fit the inclusion criteria but were not written in English.

In conclusion, in the absence of head-to-head comparisons of the novel agents in an RCT setting, the estimates from this MTC analysis are supportive of prophylaxis with new agents in neutropenic haemato-oncology and HSCT patients, instead of a standard prophylactic option or placebo. However, this MTC is by no means a substitute for such RCTs. Head-to-head comparisons should be more intensively performed using both HSCT and chemotherapy populations in order to further facilitate decision making regarding the best prophylactic alternative. Finally, in addition to the establishment of the most effective prophylactic option(s), decision making should in parallel focus on the cost effectiveness of such antifungal agents.

Funding

This work was supported by an unrestricted grant from Merck & Co. The sponsor gave advice related to the framework of the analysis. The sponsor also reviewed and commented on this article. All conceptualization, drafting and analysis were conducted within the Unit of PharmacoEpidemiology & PharmacoEconomics, University of Groningen.

Transparency declarations

J. A. S. has previously received educational support from and attended advisory boards hosted by MSD (UK) and Pfizer. M. J. P. has received grants, honoraria and travel stipends from Jansen Cilag, GSK, Sanofi, Roche, SPMSD, Pfizer, Shire, BMS, GSK Bio, Wyeth, Ortho Biotech, Novartis, Alexion, Novartis, Roche Diagnostics, Novo Nordisk, Astra Zeneca, Boehringer Ingelheim, Teva, Biogen Idec, Gideon Richter, Servier, MSD, Ferring, Amgen, Baxter, Agendia and Pasteur Merieux. All other authors declare no conflicts of interest relevant to this manuscript.

Supplementary data

Tables S1 and S2 are available as Supplementary data at JAC Online (http://jac.oxfordjournals.org/).

Acknowledgements

We would like to thank Efthalia Nikoglou and Michaela Lion (Merck Sharp & Dohme Limited, UK) for their support in designing and drafting this manuscript.

References

1
Goodrich
JM
Reed
EC
Mori
M
, et al.  . 
Clinical features and analysis of risk factors for invasive candidal infection after marrow transplantation
J Infect Dis
 , 
1991
, vol. 
164
 (pg. 
731
-
40
)
2
Prentice
HG
Kibbler
CC
Prentice
AG
Towards a targeted, risk-based, antifungal strategy in neutropenic patients
Br J Haematol
 , 
2000
, vol. 
110
 (pg. 
273
-
84
)
3
Pagano
L
Caira
M
Nosari
A
, et al.  . 
Fungal infections in recipients of hematopoietic stem cell transplants: results of the SEIFEM B-2004 study-Sorveglianza Epidemiologica Infezioni Fungine Nelle Emopatie Maligne
Clin Infect Dis
 , 
2007
, vol. 
45
 (pg. 
1161
-
70
)
4
Pagano
L
Caira
M
Candoni
A
, et al.  . 
The epidemiology of fungal infections in patients with hematologic malignancies: the SEIFEM-2004 study
Haematologica
 , 
2006
, vol. 
91
 (pg. 
1068
-
75
)
5
Nosari
A
Barberis
M
Landonio
G
, et al.  . 
Infections in haematologic neoplasms: autopsy findings
Haematologica
 , 
1991
, vol. 
76
 (pg. 
135
-
40
)
6
Martino
R
Subira
M
Invasive fungal infections in hematology: new trends
Ann Hematol
 , 
2002
, vol. 
81
 (pg. 
233
-
43
)
7
Hamza
NS
Ghannoum
MA
Lazarus
HM
Choices aplenty: antifungal prophylaxis in hematopoietic stem cell transplant recipients
Bone Marrow Transplant
 , 
2004
, vol. 
34
 (pg. 
377
-
89
)
8
Pfaller
MA
Diekema
DJ
Epidemiology of invasive candidiasis: a persistent public health problem
Clin Microbiol Rev
 , 
2007
, vol. 
20
 (pg. 
133
-
63
)
9
Playford
EG
Marriott
D
Nguyen
Q
, et al.  . 
Candidemia in nonneutropenic critically ill patients: risk factors for non-albicans Candida spp
Crit Care Med
 , 
2008
, vol. 
36
 (pg. 
2034
-
9
)
10
Viscoli
C
Girmenia
C
Marinus
A
, et al.  . 
Candidemia in cancer patients: a prospective, multicenter surveillance study by the invasive fungal infection group (IFIG) of the European Organization for Research and Treatment of Cancer (EORTC)
Clin Infect Dis
 , 
1999
, vol. 
28
 (pg. 
1071
-
9
)
11
Garcia-Vidal
C
Upton
A
Kirby
KA
, et al.  . 
Epidemiology of invasive mold infections in allogeneic stem cell transplant recipients: biological risk factors for infection according to time after transplantation
Clin Infect Dis
 , 
2008
, vol. 
47
 (pg. 
1041
-
50
)
12
Marr
KA
Carter
RA
Crippa
F
, et al.  . 
Epidemiology and outcome of mould infections in hematopoietic stem cell transplant recipients
Clin Infect Dis
 , 
2002
, vol. 
34
 (pg. 
909
-
17
)
13
Lin
SJ
Schranz
J
Teutsch
SM
Aspergillosis case-fatality rate: systematic review of the literature
Clin Infect Dis
 , 
2001
, vol. 
32
 (pg. 
358
-
66
)
14
Upton
A
Kirby
KA
Carpenter
P
, et al.  . 
Invasive aspergillosis following hematopoietic cell transplantation: outcomes and prognostic factors associated with mortality
Clin Infect Dis
 , 
2007
, vol. 
44
 (pg. 
531
-
40
)
15
Erjavec
Z
Kluin-Nelemans
H
Verweij
PE
Trends in invasive fungal infections, with emphasis on invasive aspergillosis
Clin Microbiol Infect
 , 
2009
, vol. 
15
 (pg. 
625
-
33
)
16
van Burik
JA
Role of new antifungal agents in prophylaxis of mycoses in high risk patients
Curr Opin Infect Dis
 , 
2005
, vol. 
18
 (pg. 
479
-
83
)
17
Maertens
J
Buve
K
Anaissie
E
Broad-spectrum antifungal prophylaxis in patients with cancer at high risk for invasive mold infections: counterpoint
J Natl Compr Canc Netw
 , 
2008
, vol. 
6
 (pg. 
183
-
9
)
18
Pagano
L
Caira
M
Valentini
CG
, et al.  . 
Current therapeutic approaches to fungal infections in immunocompromised hematological patients
Blood Rev
 , 
2010
, vol. 
24
 (pg. 
51
-
61
)
19
Cornely
OA
Maertens
J
Winston
DJ
, et al.  . 
Posaconazole vs. fluconazole or itraconazole prophylaxis in patients with neutropenia
N Engl J Med
 , 
2007
, vol. 
356
 (pg. 
348
-
59
)
20
Wingard
JR
Carter
SL
Walsh
TJ
, et al.  . 
Randomized double-blind trial of fluconazole versus voriconazole for prevention of invasive fungal infection (IFI) after allogeneic transplantation
Blood
 , 
2010
, vol. 
116
 (pg. 
5111
-
8
)
21
Lu
G
Ades
AE
Combination of direct and indirect evidence in mixed treatment comparisons
Stat Med
 , 
2004
, vol. 
23
 (pg. 
3105
-
24
)
22
Jansen
JP
Crawford
B
Bergman
G
, et al.  . 
Bayesian meta-analysis of multiple treatment comparisons: an introduction to mixed treatment comparisons
Value Health
 , 
2008
, vol. 
11
 pg. 
956
 
23
Higgins
J
Green
S
Cochrane Handbook for Systematic Reviews of Interventions
 , 
2011
 
The Cochrane Collaboration Available from www.cochrane-handbook.org, (6 August 2013, date last accessed)
24
Cornely
OA
Bohme
A
Buchheidt
D
, et al.  . 
Primary prophylaxis of invasive fungal infections in patients with hematologic malignancies. Recommendations of the Infectious Diseases Working Party of the German Society for Haematology and Oncology
Haematologica
 , 
2009
, vol. 
94
 (pg. 
113
-
22
)
25
Marr
KA
Carter
RA
Boeckh
M
, et al.  . 
Invasive aspergillosis in allogeneic stem cell transplant recipients: changes in epidemiology and risk factors
Blood
 , 
2002
, vol. 
100
 (pg. 
4358
-
66
)
26
Jantunen
E
Ruutu
P
Niskanen
L
, et al.  . 
Incidence and risk factors for invasive fungal infections in allogeneic BMT recipients
Bone Marrow Transplant
 , 
1997
, vol. 
19
 (pg. 
801
-
8
)
27
Boogaerts
M
Maertens
J
Clinical experience with itraconazole in systemic fungal infections
Drugs
 , 
2001
, vol. 
61
 
Suppl 1
(pg. 
39
-
47
)
28
Trifilio
S
Update on antifungal drug dosing and therapeutic drug monitoring
Curr Fungal Infect Rep
 , 
2011
, vol. 
5
 (pg. 
92
-
102
)
29
Cooper
NJ
Sutton
AJ
Morris
D
, et al.  . 
Addressing between-study heterogeneity and inconsistency in mixed treatment comparisons: application to stroke prevention treatments in individuals with non-rheumatic atrial fibrillation
Stat Med
 , 
2009
, vol. 
28
 (pg. 
1861
-
81
)
30
Cooper
NJ
Sutton
AJ
Lu
G
, et al.  . 
Mixed comparison of stroke prevention treatments in individuals with nonrheumatic atrial fibrillation
Arch Intern Med
 , 
2006
, vol. 
166
 (pg. 
1269
-
75
)
31
Spiegelhalter
DJ
Best
NG
Carlin
BP
, et al.  . 
Bayesian measures of model complexity and fit
J R Stat Soc Ser B Stat Methodol
 , 
2002
, vol. 
64
 (pg. 
583
-
639
)
32
Salanti
G
Higgins
JP
Ades
AE
, et al.  . 
Evaluation of networks of randomized trials
Stat Methods Med Res
 , 
2008
, vol. 
17
 (pg. 
279
-
301
)
33
Gelman
A
Carlin
J
Stern
H
, et al.  . 
Bayesian Data Analysis
 , 
2003
Boca Raton, FL
Chapman & Hall/CRC
34
Lunn
D
Thomas
A
Best
N
, et al.  . 
WinBUGS—a Bayesian modelling framework: concepts, structure, and extensibility
Stat Comput
 , 
2000
, vol. 
10
 (pg. 
325
-
37
)
35
Carrigan
G
Barnett
AG
Dobson
AJ
, et al.  . 
Compensating for missing data from longitudinal studies using WinBUGS
J Stat Soft
 , 
2007
, vol. 
19
 (pg. 
1
-
17
)
36
Winston
DJ
Chandrasekar
PH
Lazarus
HM
, et al.  . 
Fluconazole prophylaxis of fungal infections in patients with acute leukemia: results of a randomized placebo-controlled, double-blind, multicenter trial
Ann Intern Med
 , 
1993
, vol. 
118
 (pg. 
495
-
503
)
37
Tollemar
J
Ringden
O
Andersson
S
, et al.  . 
Randomized double-blind study of liposomal amphotericin B (Ambisome) prophylaxis of invasive fungal infections in bone marrow transplant recipients
Bone Marrow Transplant
 , 
1993
, vol. 
12
 (pg. 
577
-
82
)
38
Vreugdenhil
G
Van Dijke
BJ
Donnelly
JP
, et al.  . 
Efficacy of itraconazole in the prevention of fungal infections among neutropenic patients with hematologic malignancies and intensive chemotherapy. A double blind, placebo controlled study
Leuk Lymphoma
 , 
1993
, vol. 
11
 (pg. 
353
-
8
)
39
Winston
DJ
Maziarz
RT
Chandrasekar
PH
, et al.  . 
Intravenous and oral itraconazole versus intravenous and oral fluconazole for long-term antifungal prophylaxis in allogeneic hematopoietic stem-cell transplant recipients: a multicenter, randomized trial
Ann Intern Med
 , 
2003
, vol. 
138
 (pg. 
705
-
13
)
40
Schaffner
A
Schaffner
M
Effect of prophylactic fluconazole on the frequency of fungal infections, amphotericin B use, and health care costs in patients undergoing intensive chemotherapy for hematologic neoplasias
J Infect Dis
 , 
1995
, vol. 
172
 (pg. 
1035
-
41
)
41
Slavin
MA
Osborne
B
Adams
R
, et al.  . 
Efficacy and safety of fluconazole prophylaxis for fungal infections after marrow transplantation: a prospective, randomized, double-blind study
J Infect Dis
 , 
1995
, vol. 
171
 (pg. 
1545
-
52
)
42
Kelsey
SM
Goldman
JM
McCann
S
, et al.  . 
Liposomal amphotericin (AmBisome) in the prophylaxis of fungal infections in neutropenic patients: a randomised, double-blind, placebo-controlled study
Bone Marrow Transplant
 , 
1999
, vol. 
23
 (pg. 
163
-
8
)
43
Menichetti
F
Favero
AD
Martino
P
, et al.  . 
Itraconazole oral solution as prophylaxis for fungal infections in neutropenic patients with hematologic malignancies: a randomized, placebo-controlled, double-blind, multicenter trial. GIMEMA Infection Program. Gruppo Italiano Malattie Ematologiche dell’ Adulto
Clin Infect Dis
 , 
1999
, vol. 
28
 (pg. 
250
-
5
)
44
Morgenstern
GR
Prentice
AG
Prentice
HG
, et al.  . 
A randomized controlled trial of itraconazole versus fluconazole for the prevention of fungal infections in patients with haematological malignancies
Br J Haematol
 , 
1999
, vol. 
105
 (pg. 
901
-
11
)
45
Rotstein
C
Bow
EJ
Laverdiere
M
, et al.  . 
Randomized placebo-controlled trial of fluconazole prophylaxis for neutropenic cancer patients: benefit based on purpose and intensity of cytotoxic therapy
Clin Infect Dis
 , 
1999
, vol. 
28
 (pg. 
331
-
40
)
46
Nucci
M
Biasoli
I
Akiti
T
, et al.  . 
A double-blind, randomized, placebo-controlled trial of itraconazole capsules as antifungal prophylaxis for neutropenic patients
Clin Infect Dis
 , 
2000
, vol. 
30
 (pg. 
300
-
5
)
47
Glasmacher
A
Cornely
O
Ullmann
AJ
, et al.  . 
An open-label randomized trial comparing itraconazole oral solution with fluconazole oral solution for primary prophylaxis of fungal infections in patients with haematological malignancy and profound neutropenia
J Antimicrob Chemother
 , 
2006
, vol. 
57
 (pg. 
317
-
25
)
48
Mattiuzzi
GN
Alvarado
G
Giles
FJ
, et al.  . 
Open-label, randomized comparison of itraconazole versus caspofungin for prophylaxis in patients with hematologic malignancies
Antimicrob Agents Chemother
 , 
2006
, vol. 
50
 (pg. 
143
-
7
)
49
Mattiuzzi
G
Cortes
J
Alvarado
G
, et al.  . 
Efficacy and safety of intravenous voriconazole and intravenous itraconazole for antifungal prophylaxis in patients with acute myelogenous leukemia or high-risk myelodysplastic syndrome
Support Care Cancer
 , 
2011
, vol. 
19
 (pg. 
19
-
26
)
50
Oren
I
Rowe
JM
Sprecher
H
, et al.  . 
A prospective randomized trial of itraconazole vs fluconazole for the prevention of fungal infections in patients with acute leukemia and hematopoietic stem cell transplant recipients
Bone Marrow Transplant
 , 
2006
, vol. 
38
 (pg. 
127
-
34
)
51
Penack
O
Schwartz
S
Martus
P
, et al.  . 
Low-dose liposomal amphotericin B in the prevention of invasive fungal infections in patients with prolonged neutropenia: results from a randomized, single-center trial
Ann Oncol
 , 
2006
, vol. 
17
 (pg. 
1306
-
12
)
52
Hiramatsu
Y
Maeda
Y
Fujii
N
, et al.  . 
Use of micafungin versus fluconazole for antifungal prophylaxis in neutropenic patients receiving hematopoietic stem cell transplantation
Int J Hematol
 , 
2008
, vol. 
88
 (pg. 
588
-
95
)
53
Marks
DI
Pagliuca
A
Kibbler
CC
, et al.  . 
Voriconazole versus itraconazole for antifungal prophylaxis following allogeneic haematopoietic stem-cell transplantation
Br J Haematol
 , 
2011
54
Goodman
JL
Winston
DJ
Greenfield
RA
, et al.  . 
A controlled trial of fluconazole to prevent fungal infections in patients undergoing bone marrow transplantation
N Engl J Med
 , 
1992
, vol. 
326
 (pg. 
845
-
51
)
55
Annaloro
C
Oriana
A
Tagliaferri
E
, et al.  . 
Efficacy of different prophylactic antifungal regimens in bone marrow transplantation
Haematologica
 , 
1995
, vol. 
80
 (pg. 
512
-
7
)
56
Ito
Y
Ohyashiki
K
Yoshida
I
, et al.  . 
The prophylactic effect of itraconazole capsules and fluconazole capsules for systemic fungal infections in patients with acute myeloid leukemia and myelodysplastic syndromes: a Japanese multicenter randomized, controlled study
Int J Hematol
 , 
2007
, vol. 
85
 (pg. 
121
-
7
)
57
Yamac
K
Senol
E
Haznedar
R
Prophylactic use of fluconazole in neutropenic cancer patients
Postgrad Med J
 , 
1995
, vol. 
71
 (pg. 
284
-
6
)
58
Choi
SM
Lee
DG
Choi
JH
, et al.  . 
Itraconazole oral solution versus fluconazole syrup for prevention of invasive fungal infections in patients receiving hematopoietic stem cell transplantation: prospective, randomized, comparative clinical trial
Infect Chemother
 , 
2005
, vol. 
37
 (pg. 
71
-
8
)
59
Freifeld
AG
Bow
EJ
Sepkowitz
KA
, et al.  . 
Clinical practice guideline for the use of antimicrobial agents in neutropenic patients with cancer: 2010 update by the Infectious Diseases Society of America
Clin Infect Dis
 , 
2011
, vol. 
52
 (pg. 
e56
-
93
)
60
Robenshtok
E
Gafter-Gvili
A
Goldberg
E
, et al.  . 
Antifungal prophylaxis in cancer patients after chemotherapy or hematopoietic stem-cell transplantation: systematic review and meta-analysis
J Clin Oncol
 , 
2007
, vol. 
25
 (pg. 
5471
-
89
)
61
Mikolajewska
A
Schwartz
S
Ruhnke
M
Antifungal treatment strategies in patients with haematological diseases or cancer: from prophylaxis to empirical, pre-emptive and targeted therapy
Mycoses
 , 
2012
, vol. 
55
 (pg. 
2
-
16
)
62
Rogers
TR
Slavin
MA
Donnelly
JP
Antifungal prophylaxis during treatment for haematological malignancies: are we there yet?
Br J Haematol
 , 
2011
, vol. 
153
 (pg. 
681
-
97
)

Supplementary data