Annals of Hematology

, Volume 92, Issue 6, pp 831–839

Second- versus first-generation azoles for antifungal prophylaxis in hematology patients: a systematic review and meta-analysis

Authors

    • Baohong Ping, Department of HuiqiaoNanfang Hospital
  • Yangmin Zhu
    • Baohong Ping, Department of HuiqiaoNanfang Hospital
  • Ya Gao
    • Baohong Ping, Department of HuiqiaoNanfang Hospital
  • Chunyan Yue
    • Baohong Ping, Department of HuiqiaoNanfang Hospital
  • Bin Wu
    • Baohong Ping, Department of HuiqiaoNanfang Hospital
Original Article

DOI: 10.1007/s00277-013-1693-5

Cite this article as:
Ping, B., Zhu, Y., Gao, Y. et al. Ann Hematol (2013) 92: 831. doi:10.1007/s00277-013-1693-5

Abstract

Second-generation azoles may be more effective than first-generation azoles in the prevention of fungal infections in hematology patients. We performed a systematic review with meta-analysis of randomized controlled trials comparing second- with first-generation azoles in hematology patients with respect to proven or probable invasive fungal infections, invasive aspergillosis, receipt of empirical antifungal therapy, overall mortality, and withdrawal from the studies due to the development of adverse effects. We searched the Medline, Embase, and Cochrane Registry of Controlled Trials electronic databases as well as conference proceedings from 2002 to 2012 for randomized controlled trials comparing second-generation azoles (voriconazole, posaconazole) versus first-generation azoles (fluconazole, itraconazole). Treatment effect measures for all outcomes were expressed as odds ratio with 95 % confidence interval. Meta-analysis was performed using Review Manager, version 5.1. Data from four randomized clinical trials representing a large population of patients demonstrated that antifungal prophylaxis with second-generation azoles reduces proven or probable invasive fungal infections, invasive aspergillosis, and receipt of empirical antifungal therapy in high-risk hematology patients, while there were no differences between second- and first-generation azoles with regard to overall mortality and patients or withdrawal from the studies due to the development of adverse effects. In conclusion, antifungal prophylaxis with second-generation azoles can significantly reduce the incidence of invasive fungal infections and invasive aspergillosis but with no risk of an increase in adverse events.

Keywords

VoriconazolePosaconazoleAntifungal prophylaxisHematology

Introduction

The incidence of invasive fungal infections has increased over the past decade, largely because of the increasing size of the population at risk. The major risk groups for invasive fungal infections are hematopoietic stem cell transplant (HSCT) recipients and patients with hematological malignancies treated with cytotoxic chemotherapy. The development of invasive fungal infections in hematology patients is associated with substantial clinical consequences, including increased morbidity and mortality, in addition to delays in potentially curative treatment. Prophylaxis is a commonly used treatment strategy; many prophylactic approaches, which appear to be effective, have been recommended [1]. Both azoles and polyenes have been used as prophylactic regimens in this high-risk population. First-generation azoles, fluconazole, and itraconazole have been used most frequently as they are administered both orally and intravenously (IV) and are associated with fewer adverse effects. However, there are a number of clinically important limitations related to their range of activity, the development of resistance, and some toxicity [2]. To overcome these limitations, next-generation azole antifungals have been developed. These second-generation azoles (voriconazole, posaconazole, ravuconazole, etc.) exhibit favorable pharmacokinetic and toxicity profiles and show high levels of activity against resistant and emerging pathogens. However, only voriconazole and posaconazole have been adequately investigated in phase III studies and approved by the regulatory agencies for the treatment and prophylaxis of invasive fungal infections [3]. Several randomized control trials (RCTs) [47] and a retrospective study [8] showed that second-generation azoles were more effective than first-generation azoles in preventing the development of invasive fungal infections in hematology patients. It appears that first-generation azoles should no longer be considered as the standard prophylactic regimen. Although narrative reviews are available, to our knowledge, no systematic review and meta-analysis has addressed this issue. Herein, we report the results of a systematic review and meta-analysis of the available randomized control trials examining the impact of second- versus first-generation azoles for antifungal prophylaxis in high-risk hematology patients on proven or probable invasive fungal infections, invasive aspergillosis, receipt of empirical antifungal therapy, overall mortality, and withdrawal from the studies due to the development of adverse effects.

Design and methods

Search methods and selection of studies

We searched the Medline, Embase, and Cochrane Registry of Controlled Trials electronic databases in April 2012. We performed a manual search of abstracts from the annual meetings of the American Society of Hematology, American Society of Clinical Oncology, the European Hematology Association, and European Group for Blood and Marrow Transplantation for the years 2002 to 2012. Reference lists of all identified studies and relevant articles, including review papers, were screened. We used the following search terms: (voriconazole or posaconazole) AND (prophylaxis or prevention). Randomized and prospective controlled trials comparing second-generation azole antifungal agents (voriconazole, posaconazole) with regard to antifungal prophylaxis to first-generation azole antifungal agents (fluconazole, itraconazole) in hematology patients who were neutropenic following cytotoxic chemotherapy or HSCT, or receiving immunosuppressive therapy were included. The administration of intravenous azole antifungal agents was not permitted unless an invasive fungal infection was proven or suspected. Prospective studies with a historical cohort as a control arm and ongoing trials were excluded from our systematic review. Only studies written in English with available data were reviewed and analyzed. Two reviewers (C.Y.Y. and B.W.) independently screened the titles and abstracts of all identified studies to assess their eligibility for inclusion.

Data extraction and quality assessment

Two reviewers (Y.G. and Y.M.Z.) independently extracted data in a standardized format from the included trials. The date collected from each study included: first author, study ID, publication year, study population, total number allocated to treatment, mean patient age, type of antifungal drug, duration of prophylaxis, and study outcome. Any disagreement between these reviewers was resolved by a third reviewer.

All included trials were assessed by the two reviewers independently for quality parameters, such as randomization procedures, concealment of allocation, blinding, completeness of follow-up, selective reporting, and intention-to-treat analysis. We graded all quality parameters as low risk of bias, unclear, or high risk of bias according to the criteria specified in the Cochrane Handbook.

Definition of outcomes

The main outcomes of the study were proven or probable invasive fungal infections and invasive aspergillosis, which were defined in accordance with the European Organization for the Research and Treatment of Cancer/Mycoses Study Group criteria [9]. Secondary outcomes were receipt of empirical antifungal therapy, overall mortality, and withdrawal from the studies due to the development of adverse effects.

Statistical analysis

Treatment effect measures for all outcomes were expressed as odds ratios (OR) with 95 % confidence intervals (CI). Heterogeneity was assessed by Cochrane χ2 analysis and by measuring I2. Meta-analysis was performed using the fixed-effect model (Mantel–Haenszel method). For analysis with unexplained statistical heterogeneity (P < 0.1 or I2 > 50 %), a random effect model (Mantel–Haenszel method) was used. A linear regression test for small trial bias was not performed as the number of trials included was limited. Analyses were performed using Review Manager, version 5.1. All statistical tests were two-sided. The threshold of significance was p < 0.05 for all tests except for heterogeneity. We performed sensitivity analysis to assess the effects of quality parameters on effect estimates, including concealment of allocation and blinding. We also explored our findings further by subgroup analysis to assess the effects of the type of drug used.

Results

Included trials

The electronic database search and manual search screened 168 citations and 18 potentially relevant citations were retrieved as full-text or identified for more detailed information (Fig. 1). Of these, 14 were excluded for the following reasons: non-RCT studies, address other issues, not hematology patients, administration of IV azole antifungal agents, or no available data. Finally, four trials with 2,165 patients from four publications were included in this meta-analysis. All were full-text publications. All trials (Table 1) were multi-center, two-arm, parallel prospective RCTs. One trial included recipients of allogeneic stem cell transplantation (Allo-SCT) with graft-versus-host disease (GVHD) who were receiving immunosuppressive therapy [5]. One trial included patients with prolonged neutropenia resulting from remission induction chemotherapy of acute myelogenous leukemia (AML) or myelodysplastic syndrome (MDS) [4]. Two trials included patients undergoing Allo-SCT after a myeloablative or reduced-intensity conditioning regimen [6, 7]. Two trials examined the effects of voriconazole [6, 7] and two evaluated posaconazole [4, 5]. The control arm consisted of fluconazole in two trials [5, 6], itraconazole in one trial [7], and either fluconazole or itraconazole in one trial [4]. Of four trials, recruitment ranged from March 1999 [5] to February 2009 [7]. The sample sizes ranged from 465 [7] to 602 [4]. The mean age ranged from 40.4 [5] to 50 years [4]. One trial did not report mean age with a median age of 43 years in both intervention and control arms [6]. Primary outcomes were reported by all of four trials. Secondary outcomes were reported by three trials. The quality of the trials in the analysis was acceptable. None of the four trials described randomization procedures. In one trial, allocation concealment was considered to be adequate [4] and in two trials, not stated [5, 6]. Patients and physicians were blinded in all trials except for one open-label trial [7]. All trials reported the loss to follow-up and methods used to deal with the loss to follow-up. Intention-to-treat or modified intent-to-treat analyses were performed by all trials. There was no selective reporting for any of the trials.
https://static-content.springer.com/image/art%3A10.1007%2Fs00277-013-1693-5/MediaObjects/277_2013_1693_Fig1_HTML.gif
Fig. 1

Study selection flow diagram

Table 1

Study characteristics and quality assessment.

First author

Ullman

Cornely

Wingard

Marks

Publication year

2007

2007

2010

2011

Study ID

NCT00034645

NCT00044486

NCT00075803

Not reported

Enrollment period

March 1999–February 2003

August 2002–April 2005

November 2003–September 2006

March 2006–February 2009

Study population

Patients with GVHD who were receiving immunosuppressive therapy

Patients with prolonged neutropenia resulting from remission induction chemotherapy of AML or MDS

Patients undergoing allo-SCT after a myeloablative conditioning regimen

Patients undergoing allo-SCT after a myeloablative or reduced-intensity conditioning regimen

Number

600

602

600

465

Mean age (range)

42.2 (13–72) versus 40.4 (13–70)

49 (13–82) versus 50 (13–81)

Not reported

43.3 (11–70) versus 42.3 (13–70)

Azole prophylaxis

Pos 200 mg, PO, 3/d versus Flu 400 mg, PO, 1/d

Pos 200 mg, PO, 3/d versus Flu 400 mg, PO, 1/d or Itr 200 mg, PO, 2/d

Vor 200 mg, PO, 2/d versus Flu 400 mg, PO, 1/d

Vor 200 mg, PO, 2/d versus Itr 200 mg, PO, 2/d

Duration of prophylaxis

112 days

Till recovery from neutropenia and complete remission

100–180 days

100–180 days

Duration of Follow-up

112 days

Randomization to 7 days after the last dose of the study drug had been administered during the last cycle of chemotherapy

180 days

180 days

Randomization procedures

Unclear

Unclear

Unclear

Unclear

Concealment of allocation

Unclear

Low risk of bias

Unclear

High risk of bias

Blinding

Low risk of bias

Low risk of bias

Low risk of bias

High risk of bias

Completeness of follow-up

Low risk of bias

Low risk of bias

Low risk of bias

Low risk of bias

Selective reporting

Low risk of bias

Low risk of bias

Low risk of bias

Low risk of bias

Intention-to-treat analysis

Low risk of bias

Low risk of bias

Low risk of bias

Low risk of bias

Primary outcomes

Four trials including 2,267 patients reported proven or probable invasive fungal infections were included in the meta-analysis. Antifungal prophylaxis using second-generation azole antifungal agents significantly reduced proven or probable invasive fungal infections compared with first-generation agents (OR = 0.47, 95 % CI 0.32–0.69, I2 = 0 %, P = 0.0001) (Fig. 2). Data from the four trials including 2,267 patients were available for the analysis of invasive aspergillosis. Antifungal prophylaxis using second-generation azole antifungal agents significantly reduced invasive aspergillosis (OR = 0.28, 95 % CI 0.17–0.48, I2 = 28 %, P < 0.00001) (Fig. 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs00277-013-1693-5/MediaObjects/277_2013_1693_Fig2_HTML.gif
Fig. 2

Fewer proven or probable invasive fungal infections were observed after prophylaxis using second-generation azole antifungal agents

https://static-content.springer.com/image/art%3A10.1007%2Fs00277-013-1693-5/MediaObjects/277_2013_1693_Fig3_HTML.gif
Fig. 3

Fewer cases of invasive aspergillosis were observed after prophylaxis using second-generation azole antifungal agents

Quality of concealment of allocation (Low risk of bias or Unclear) and blinding (Low risk of bias) did not have a significant impact on the results for either proven or probable invasive fungal infections (OR = 0.46, 95 % CI 0.30–0.68, I2 = 17 %, P = 0.0001; three trials, 1,802 patients) and invasive aspergillosis (OR = 0.29, 95 % CI 0.17–0.50, I2 = 50 %, P < 0.00001; three trials, 1,802 patients).

When proven or probable invasive fungal infection was analyzed per drug type (Fig. 4), there were statistically significant differences for posaconazole (OR = 0.40, 95 % CI 0.19–0.87, I2 = 52 %, P = 0.02; two trials, 1,202 patients). However, there were no significant reductions in incidence rates proven or probable invasive fungal infections associated with voriconazole (OR = 0.56, 95 % CI 0.30–1.04, I2 = 0 %, P = 0.06; two trials, 1,065 patients). There were significantly fewer cases of invasive aspergillosis for both voriconazole (OR = 0.45, 95 % CI 0.21–0.96, I2 = 0 %, P = 0.04; two trials, 1,065 patients) and posaconazole (OR = 0.20, 95 % CI 0.06–0.65, I2 = 52 %, P = 0.008; two trials, 1,202 patients) (Fig. 5). Administration of second-generation azole antifungal agents resulted in significantly fewer proven or probable invasive fungal infections (OR = 0.47, 95 % CI 0.31–0.71, I2 = 0 %, P = 0.0003; three trials, 1,744 patients) and cases of invasive aspergillosis (OR = 0.31, 95 % CI 0.18–0.52, I2 = 44 %, P < 0.0001; three trials, 1,744 patients) when in comparison with fluconazole. Using second-generation azoles as antifungal prophylaxis also resulted in significantly fewer proven or probable invasive fungal infections (OR = 0.35, 95 % CI 0.14–0.87, I2 = 35 %, P = 0.02; two trials, 827 patients) and cases of invasive aspergillosis (OR = 0.11, 95 % CI 0.03–0.40, I2 = 0 %, P = 0.0008; two trials, 827 patients) in comparison with itraconazole.
https://static-content.springer.com/image/art%3A10.1007%2Fs00277-013-1693-5/MediaObjects/277_2013_1693_Fig4_HTML.gif
Fig. 4

Fewer proven or probable invasive fungal infections were observed after prophylaxis using posaconazole but not voriconazole

https://static-content.springer.com/image/art%3A10.1007%2Fs00277-013-1693-5/MediaObjects/277_2013_1693_Fig5_HTML.gif
Fig. 5

Fewer cases of invasive aspergillosis were observed after prophylaxis using voriconazole or posaconazole

Secondary outcomes

Antifungal prophylaxis using second-generation azole antifungal agents resulted in significantly fewer patients receiving empirical antifungal therapy (OR = 0.62, 95 % CI 0.50–0.77, I2 = 0 %, P < 0.0001; three trials, 1,667 patients) (Fig. 6). Further subgroup analyses revealed that significantly fewer patients received empirical antifungal therapy for both voriconazole (OR = 0.66, 95 % CI 0.51–0.86, I2 = 0 %, P = 0.002; two trials, 1,065 patients) and posaconazole (OR = 0.56, 95 % CI 0.39 to 0.81, P = 0.002; one trial, 602 patients).
https://static-content.springer.com/image/art%3A10.1007%2Fs00277-013-1693-5/MediaObjects/277_2013_1693_Fig6_HTML.gif
Fig. 6

Fewer patients received empirical antifungal therapy after prophylaxis using second-generation azole antifungal agents

There was no difference in overall mortality between second- and first-generation azoles for antifungal prophylaxis (OR = 0.81, 95 % CI 0.64–1.01, I2 = 0 %, P = 0.06; three trials, 1,802 patients) (Fig. 7). There were also no significant differences in overall mortality for both voriconazole (OR = 0.90, 95 % CI 0.60–1.35, P = 0.61; one trial, 600 patients) and posaconazole (OR = 0.77, 95 % CI 0.59–1.01, I2 = 0 %, P = 0.06; two trials, 1,202 patients).
https://static-content.springer.com/image/art%3A10.1007%2Fs00277-013-1693-5/MediaObjects/277_2013_1693_Fig7_HTML.gif
Fig. 7

There was no difference in the rate of all-cause death between prophylaxis using first- and second-generation azole antifungal agents

No difference was detected in patients withdrawn from the studies due to the development of adverse effects with second-generation azoles when compared with first-generation azoles for antifungal prophylaxis (OR = 1.02, 95 % CI 0.81–1.30, I2 = 45 %, P = 0.84; three trials, 1,667 patients) (Fig. 8). Again, withdrawal from the studies due to the development of adverse effects was analyzed per type of drug, and no differences were observed for both voriconazole (OR = 1.41, 95 % CI 0.93–2.13, P = 0.10; one trial, 465 patients) and posaconazole (OR = 0.88, 95 % CI 0.66–1.17, I2 = 0 %, P = 0.37; two trials, 1,202 patients).
https://static-content.springer.com/image/art%3A10.1007%2Fs00277-013-1693-5/MediaObjects/277_2013_1693_Fig8_HTML.gif
Fig. 8

There was no difference in withdrawal from the studies due to the development of adverse effects between prophylaxis using first- and second-generation azole antifungal agents

Discussion

Invasive fungal infections are a primary cause of morbidity and mortality in patients with hematological malignancies. These hematology patients are particularly vulnerable as a consequence of prolonged neutropenia or immunosuppression following myelosuppressive chemotherapy or allo-SCT. Invasive aspergillosis is the most important invasive fungal infection in these populations and is the major cause of infection-related death in HSCT recipients. Recent data have shown that about 75 % of IFIs in HSCT patients are caused by molds, mostly in the form of invasive aspergillosis (59–71 %), with the remainder caused by Candida spp. [1012]. Invasive mold infections, especially those caused by Aspergillus, have become increasingly frequent in various hematology populations.

Azole antifungals have emerged as front-line drugs for the treatment and prophylaxis of many systemic mycoses. Antifungal prophylaxis with fluconazole reduces morbidity and mortality among recipients of allo-SCT and other patient populations with prolonged neutropenia with an acceptable adverse-event profile but lacks efficacy against molds, which have become increasingly frequent causes of infection in neutropenic patients. Itraconazole has a wider spectrum of activity than fluconzole, including activity against Aspergillus species but the poor tolerability of the cyclodextrin-containing oral solution and the erratic bioavailability of the oral-capsule formulation limited its clinical use [13]. The second-generation azoles, including voriconazole and posaconazole, had been investigated in several phase III studies and were shown to be more effective than early-generation oral azole agents for antifungal prophylaxis. To quantify the impacts of the different effects between second- and first-generation azoles for prophylaxis, data were pooled from available published trials in this meta-analysis.

This systemic review and meta-analysis of data from four randomized clinical trials representing a large population of patients demonstrated that antifungal prophylaxis with second-generation azoles reduces proven or probable invasive fungal infections, invasive aspergillosis, and number of patients received empirical antifungal therapy in high-risk hematology patients, while there was no difference between second- and first-generation azoles with regard to overall mortality. In addition, there was no difference between the two prophylactic regimens with respect to patient withdrawn from the studies due to the development of adverse effects. To our knowledge, this is the first comprehensive review focusing on high-risk hematology patients to compare new-generation azoles with early generation azoles for antifungal prophylaxis.

Current guidelines recommend dividing high-risk hematology patients into three groups to assign the most appropriate agent for primary antifungal prophylaxis [14]. During induction chemotherapy for acute leukemia, the first-line agent of choice is posaconazole. The second patient group consists of allo-SCT recipients during the neutropenic phase, for whom voriconazole and fluconazole are most strongly recommended. For allo-SCT patients with GVHD, posaconazole and voriconazole are the most strongly recommended agents for the primary prophylaxis. The reduction in invasive fungal infections, invasive aspergillosis, and empirical antifungal therapy demonstrated in our meta-analysis supported these recommendations. A recent retrospective study also supported our results [8].

Our pooled results suggested that prophylaxis with second-generation azoles is more effective than prophylaxis with first-generation azoles in prevention of invasive fungal infections and invasive aspergillosis. These results were underlined by the results of subgroup analyses, which indicated that second-generation azoles prevented the development of invasive fungal infections and invasive aspergillosis when compared with fluconazole or posaconazole. Second-generation azoles had been demonstrated to have expanded antifungal activity compared with first-generation azoles in vitro and in vivo [3]. Fluconazole has no activity against Aspergillus spp., Scedosporium spp., Fusarium spp., or Zygomycetes and no activity against Candida krusei and some strains of Candida glabrata. Itraconazole has activity against most Candida spp. and Aspergillus spp., with limited activity against Zygomycetes. Voriconazole shows activity against C. krusei, some but not all fluconazole-resistant C. glabrata, Fusarium spp., and Scedosporium apiospermum, but varying activity against Scedosporium prolificans. Posaconazole showed the broadest spectrum, including Candida spp., Aspergillus spp., and activity against many Zygomycetes [2]. It should be noted that the reduction in proven or probable invasive fungal infections seen with posaconazole was not observed with voriconazole. Although voriconazole and posaconazole share a number of similarities as primary antifungal prophylaxis in allo-SCT recipients, both are available as oral formulations, both are clinically active against molds as well as yeasts, and both share gastrointestinal and hepatic side effects as some of their key treatment-related adverse events; there are also a number of differences between the two agents. As listed above, voriconazole is not active against the Zygomycetes and may have reduced activity against certain strains of C. glabrata and Candida albicans that have acquired fluconazole resistance [2, 15]. This may have contributed to the difference in their impact on invasive fungal infections. The limited number of trials and low incidence of fungal infections could also account for the differences in outcome between the two agents.

The advantages of second-generation azoles over first-generation azoles include higher oral bioavailability, fewer drug-drug interactions, and higher activity against resistant and emerging pathogens [2]. Our analysis also demonstrated that prophylaxis using second-generation azole antifungal agents resulted in significantly fewer patients receiving empirical antifungal therapies. It is easy to explain why fewer suspected or proven invasive fungal infections could lead to fewer recipients of empirical antifungal therapy. The most relevant endpoint of antifungal prophylaxis is the reduction in mortality. However, death attributable to invasive fungal infection is difficult to prove. In our analysis, there was no difference observed in overall mortality between the uses of second- and first-generation agents for antifungal prophylaxis. It should be noted that, although a reduction in overall mortality is a desirable endpoint of any clinical decision, it is difficult to achieve in the context of multiple competing illnesses in a severely immunocompromised host [16]. In addition, the overall mortality was assessed at a maximum of 180 days, and the effects of a delayed treatment because of invasive fungal infection may result in increased morality after this period. Therefore, the effect of second-generation azoles on overall survival must be interpreted with caution. In addition, no difference was observed regarding withdrawal from the studies due to the development of adverse effects with second-generation azoles when compared with first-generation azoles. Adverse effects associated with voriconazole include transient visual disturbances, fever, skin rash, nausea, diarrhea, liver function test abnormalities, and prolongation of QT interval [17]. Voriconazole treatment-related visual disturbances are the most common adverse events reported by more than 30 % of patients in clinical trials [15]. However, the visual disturbances were generally mild and rarely resulted in discontinuation of treatment. Data from previous studies indicated that posaconazole is well tolerated, even upon long-term administration. The most commonly reported adverse events were fever, nausea, diarrhea, vomiting, and headache [18, 19]. Other notable adverse events included hypokalemia, skin rash, thrombocytopenia, and abdominal pain. The incidences of these adverse events were similar to those with itraconazole and fluconazole.

Antifungal prophylaxis in hematology patients is important and reduces the use of antifungal therapy for suspected or proven invasive fungal infections, reduces all-cause mortality, fungal infection-related mortality, and avoids the costs of management of suspected or proven invasive fungal infections. Our analysis demonstrated that prophylaxis with second-generation azoles is more effective than that with first-generation azoles in prevention of invasive fungal infections and invasive aspergillosis. However, there is a considerable risk that other hazardous species, such as Zygomycetes, Rhizopus, Mucor, etc., may be selected for. Antifungal prophylaxis with second-generation azoles could also lead to further spread of azole resistance and become a major problem in the future. Hamprecht et al. reported the first culture-proven case of invasive aspergillosis caused by azole-resistant Aspergillus fumigatus in a patient with AML in Germany and invasive aspergillosis presented as breakthrough infection under posaconazole prophylaxis [20]. Resistance to azole antifungals in clinical A. fumigatus isolates has also been reported in several countries [2124]; however, it is interesting to note that patients with hematological malignancies have rarely been reported to show azole resistance [25].

The present meta-analysis had several limitations that should be taken into consideration when interpreting the results. First, heterogeneity is a potential problem when interpreting the results of meta-analyses. However, it should be noted that significant heterogeneity was detected in only some subgroup analyses. There was no significant heterogeneity in our main analysis. Second, only four RCTs with available data were included in the analysis. Only published studies were included in this meta-analysis, and non-significant or negative results may be unpublished. Third, the duration of administration of antifungal drugs and follow-up period varied between studies. Especially, in one RCT [4], the administration of azoles was continued until recovery from neutropenia and complete remission, and the duration of the follow-up was considered to be the period from randomization to 7 days after the last dose of the study drug had been administered during the last cycle of chemotherapy. However, in the remaining studies, the administration of azole prophylaxis was continued for 100–180 days, and the duration of the follow-up ranged from 112 to 180 days. Fourth, our analysis referred to specific populations, i.e., patients undergoing allo-SCT as well as those with GVHD who were receiving immunosuppressive therapy or with prolonged neutropenia resulting from cytotoxic chemotherapy, and we did not perform a subgroup analysis according to study population. Finally, antifungal resistance in such high-risk patients requires further attention.

In conclusion, this meta-analysis demonstrated that antifungal prophylaxis with second-generation azoles (voriconazole, posaconazole) can significantly reduce the incidence of invasive fungal infections and invasive aspergillosis but with no risk of increase in adverse events.

Conflict of interest

The authors declare that they have no conflict of interest.

Copyright information

© Springer-Verlag Berlin Heidelberg 2013