A network meta- analysis of primary prophylaxis for invasive fungal infection in haematological patients

Transcription

1 Received: 3 April 207 Accepted: 9 May 207 DOI: 0./jcpt.2579 REVIEW ARTICLE A network meta- analysis of primary prophylaxis for invasive fungal infection in haematological patients L. P. Leonart MSc F. S. Tonin MSc V. L. Ferreira MSc S. T. S. Penteado MSc A. Wiens PhD F. A. Motta PhD 2 R. Pontarolo PhD Department of Pharmacy, Universidade Federal do Paraná, Curitiba, Brazil 2 Faculdades e Instituto de Pesquisa Pelé Pequeno Príncipe, Hospital Pequeno Príncipe, Curitiba, Brazil Correspondence R. Pontarolo, Department of Pharmacy, Universidade Federal do Paraná, Curitiba, PR, Brazil. pontarolo@ufpr.br Summary What is known and objective: Antifungal prophylaxis is an option to reduce the incidence of invasive fungal infection (IFI) in haematological patients. To date, no network meta- analysis (NMA) of high- quality evidence (double- blind randomized controlled trials) has been performed on this subject. This systematic review and NMA aimed to evaluate the safety and efficacy of different antifungal agents used for prophylaxis of IFI in patients with haematological disorders. Methods: A systematic review was performed according to PRISMA and Cochrane recommendations. The search for articles was conducted on PubMed, Scopus and the Web of Science. We searched for double- blind randomized clinical trials comparing antifungal agents for IFI prophylaxis head- to- head vs placebo in patients with any blood cancer. Network meta- analyses were conducted using Addis version.6.6. Evaluation of the quality of included RCTs was also performed. Results: Twenty- five trials were included in the qualitative and quantitative analyses. Posaconazole stood out as the best IFI prophylaxis option and for avoiding IFI- related mortality. For the incidence of candidiasis outcome, the azoles were superior to placebo. Voriconazole and posaconazole were, respectively, the first and second best options. For the incidence of aspergillosis outcome, the probability rank suggested that voriconazole followed by liposomal amphotericin B is, possibly, the best choice. The quality of studies was considered good, with a mean Jadad score of 4.0. What is new and conclusion: The results of our work support prophylaxis with antifungal agents as reducing the risk of IFI in haematological patients. Overall, the secondgeneration azoles were found to be the best option for preventing IFI in this population. KEYWORDS antifungal agents, haematologic diseases, invasive fungal infections, network meta-analysis WHAT IS KNOWN AND OBJECTIVE Invasive fungal infections (IFIs) represent a leading cause of morbidity and mortality in immunocompromised patients, especially those with haematologic malignancies, undergoing haematopoietic stem cell transplantation or with human immunodeficiency virus infection. -6 The mortality can exceed 50% in some patient populations, depending on the pathogen and the underlying condition or disease. Candida albicans and Aspergillus fumigatus are some of the most well- known causes of opportunistic infections. 7 Candidemia- related mortality ranges from 20% to 54%, 8,9 and can be as high as 89% for invasive aspergillosis John Wiley & Sons Ltd wileyonlinelibrary.com/journal/jcpt J Clin Pharm Ther. 207;42:

2 53 Despite the improvement in diagnostic procedures for fungal infections, early and accurate diagnosis of IFI remains elusive. Moreover, because of the high- risk condition of immunocompromised patients with haematological malignancies, early treatment strategies and antifungal prophylaxis are options for mitigating the impact of IFI in this population and hence reducing mortality.,2-4 In recent decades, a variety of antifungal agents from the polyene, azole and echinocandin classes have been developed and are available in most countries as treatment strategies. 3,5 However, because of the complexity of handling IFI and considering that these drugs have different pharmacological profiles, costs, indications and adverse events, selecting the best therapeutic option remains difficult. In addition, recommendations for the management of these infections are often conflicting and controversial, which creates variability in clinical practice. 6-2 Recent mixed treatment comparisons on antifungal agent prophylaxis against IFI in patients with haematological malignancies showed how this method can represent technical advantages and support clinical decision-making. 3,22,23 However, these comparisons were not sufficiently extensive in terms of the included population, evaluated drugs and the language of the trials (limited to English). Two of them assessed only the azole class 3,23 in transplanted patients or subjects with haematological diseases, whereas the other study did not evaluate reference drugs such as amphotericin B. Furthermore, other systematic reviews with pairwise meta- analyses have also already been published. 24,25 However, they mixed any kind of cancer, as well as prophylaxis, empirical and curative treatments. This lack of evidence synthesis may affect future decision- making. Thus, to extend and improve previous evidence on the efficacy and safety of antifungal agents, we conducted a systematic review and network meta- analysis of RCTs evaluating amphotericin B (conventional and liposomal formulations), ketoconazole, fluconazole, itraconazole, micafungin, miconazole, nystatin, posaconazole or voriconazole as IFI primary antifungal prophylaxis in patients with haematological disease. 2 METHODS 2. Systematic review and eligibility criteria A systematic review was performed according to Preferred Reporting Items for Systematic Reviews and Meta- Analyses (PRISMA) and the Cochrane Collaboration recommendations. 26,27 All steps were conducted by two reviewers independently, and discrepancies were resolved by consensus with a third author. Only relevant double- blind randomized controlled trials were identified via PubMed, Web of Science, Embase, Science Direct, Cochrane Library, Scopus, Lilacs, International Pharmaceutical Abstracts and Scielo with a timeframe from inception to January 207. The following descriptors were used: clinical trial, random, fungal infection*, mycose*, mycosi* and antifungal agent* combined with the Boolean operators AND and OR (an example of the search strategy used is in Appendix S). Studies were included if they met all eligibility criteria: (i) patients with haematological diseases and with expected neutropenia who received antifungal agent as prophylaxis for IFI regardless of its pharmaceutical form, class and therapeutic regimen; (ii) patients of any age, ethnicity and gender; (iii) therapy should have as a comparator placebo or another antifungal used as a systemic drug (head- to- head studies); (iv) studies should report the outcomes of interest related to efficacy and safety (mentioned forward). Although some of the drugs included in our study are no longer used in clinical practice (eg ketoconazole, nystatin), we selected studies that also involved them in order not to lose any evidence from the literature and increase the strength of the networks through indirect comparisons. Studies that did not address the outcomes of interest, other types of studies (such as cohorts, case reports and reviews), non- doubleblind, non- RCTs, interventions that used an association of drugs and articles published in non- Roman characters were excluded. 2.2 Data extraction and quality assessment The relevant data extracted from each study included the following: baseline characteristics (authors names, year of publication, number of patients, sex, age, characteristics of antifungal agents, haematological condition); efficacy outcomes as: (i) incidence of proven and suspected IFI, (ii) incidence of infection caused by any species of Candida spp. (candidiasis) and (iii) incidence of infection caused by any species of Aspergillus spp. (aspergillosis); and safety outcomes such as IFI- related mortality. Overall, IFI was considered proven when fungal moulds or yeasts were histopathologically or cytopathologically demonstrated in deep tissue with evidence of associated tissue damage or a positive culture from a normally sterile site with clinical or radiological abnormalities consistent with infection. IFI was considered suspected when persistent fever refractory to broad- spectrum antibiotics was observed. These definitions were chosen as they were the most commonly reported ones in the RCTs included in this study. The methodological quality of included articles was assessed by two different tools: the Jadad scale 28 and the Cochrane Collaboration s tool for assessing the risk of bias Data analysis Network meta- analysis is recommended by the International Society for Pharmacoeconomics and Outcome Research and is based on Bayesian methods to compare multiple treatments simultaneously. 29 It permits a multiple treatment comparison among studies arms, including direct and indirect ones. A random- effect model performed by the Markov chain Monte Carlo simulation method was used to obtain pooled effect sizes. A consistency model was built, and treatment effects could be presented as odds ratio (OR) with associated 95% credible intervals (CrIs). The results of the consistency analysis allowed estimating probabilities for the interventions to be the best, second best and so on.

3 532 Node- splitting analysis was used for the inconsistency evaluation. In this analysis, it is tested whether results of direct and indirect evidence on a specific node (the split node) are in agreement or not (P- values <.05 indicate significant inconsistencies). 30 The analyses were performed using Addis version.6.6 ( org/addis). 3 RESULTS Through electronic searches, 729 studies were retrieved after the exclusion of duplicates. During the screening process (reviewing of titles and abstracts), 6904 records were excluded. After full- text analysis, 290 articles were excluded, resulting in 25 studies amenable to performing qualitative and quantitative analyses (Figure ). For the complete list reference of included records, see Appendix S2. 3. Studies characteristics The majority of studies (n=9) were conducted in a single country, with the United States of America and the Netherlands being the most mentioned ones. The most common treatment was fluconazole ( studies), followed by itraconazole (five studies). From the 25 eligible studies, 9 presented placebo as a comparator for one of the antifungal drugs under analysis and the other six were head- to- head studies. Other baseline data can be seen in Table. 3.2 Quality assessment The quality assessment showed overall good quality, resulting in a mean Jadad score of 4.0 (ranging from 3 to 5). All the studies were scored at least on randomization, double- blinding, and properly reporting exclusions and dropouts. Concerning the risk of bias assessed by the Cochrane tool, around half of the studies were supported by pharmaceutical companies or presented conflict of interests. Most of the trials were not clear when reporting allocation concealment and blinding of outcome assessments (see Figure S and S2 in Appendix S3). 3.3 Network meta- analysis A network meta- analysis was built, including all arms comprising any of the ten antifungal agents or placebo (Figure 2). For the network meta- analysis of the incidence of IFI, all arms were included in the analysis, resulting in different comparators. Differences were observed among conventional amphotericin B, liposomal amphotericin B, fluconazole, itraconazole, posaconazole and voriconazole compared to placebo (see Figure 3 for OR with 95% CrI). These results demonstrated that these drugs are superior to control and more effective against IFI compared to placebo. Other comparisons (fluconazole vs ketoconazole, posaconazole vs ketoconazole and voriconazole vs ketoconazole) were also statistically significant and showed that ketoconazole treatment was associated with more IFI than the mentioned azoles. Included Eligibility Screening Identification Records identified through database searching (n=2,260) Records after duplicates removed (n=729) Records screened (n=729) Full-text articles assessed for eligibility (n=35) Studies included in qualitative synthesis (n=25) Randomized Clinical Trials included in quantitative synthesis (meta-analysis) (n=25) Records excluded (n=6904) Full-text articles excluded, with reasons (n=290) Addressed other outcomes: 25 Study type: 40 Non-randomized: 2 Non-double-blind: 98 Non-controlled: 6 Non-Roman characters: 9 FIGURE PRISMA flowchart diagram of study selection

4 533 TABLE Baseline characteristics of the studies included in the systematic review Author, year Country Population Treatment N Male (%) Mean age (y) Mean duration (d) Jadad score Benhamou, 99 France Children with haematological diseases, candidates for bone marrow transplantation Ketoconazole mg/d (oral) Placebo Brincker, 978 Denmark Patients with leukaemia or lymphoma Miconazole 500 mg 4 /d (oral) 5 53 NR Placebo Brincker, 983 Denmark Patients with acute leukaemia receiving chemotherapy Ketoconazole 400 mg/d (oral) NR 3 Placebo Goodman, 992 United States Bone marrow transplanted patients Fluconazole 400 mg/d (oral/intravenous) NR NR 4 Placebo Harousseau, 2000 >2 countries Patients with any haematological disease Itraconazole 2.5 mg/kg 2 /d (oral) Amphotericin B 500 mg 4 /d (oral) Huijgens, 999 The Netherlands Patients with any haematological disease Fluconazole 50 mg 2 /d (oral) 0 62 NR NR 4 Itraconazole 00 mg 2 /d (oral) 0 56 Kelsey, 999 United Kingdom and Ireland Patients with any haematological disease Lip Amphotericin B 2 mg/kg 3 /wk (intravenous) Placebo NR 4 Laverdiere, 2000 Canada Patients with any haematological disease Fluconazole 400 mg/d (oral) Placebo Menichetti, 999 Italy Patients with any haematological disease Itraconazole 2.5 mg/kg 2 /d (oral) Placebo Nucci, 2000 Brazil Patients with any haematological disease Itraconazole 00 mg 2 /d (oral) NR 4 Placebo Palmblad, 992 Sweden Patients with acute leukaemia Ketoconazole 200 mg/d (oral) 50 NR 5 NR 3 Placebo Rijnders, 2008 The Netherlands Patients with any haematological disease L ip Amphotericin B 5 mg ml d (inhalation) NR 5 Placebo Riley, 994 United States Bone marrow transplanted patients Low- dose amphotericin B 0. mg kg d (intravenous) NR 4 Placebo Rotstein, 999 Canada Leukaemia or bone marrow transplanted patients receiving chemotherapy Fluconazole 400 mg/d (parenteral) Placebo (Continues)

5 534 TABLE (Continued) Author, year Country Population Treatment N Male (%) Mean age (y) Mean duration (d) Jadad score Schaffner, 995 Switzerland Neutropenic patients with any haematological disease receiving chemotherapy Fluconazole 400 mg/d (oral/intravenous) 75 NR Placebo Slavin, 995 Australia Transplanted patients motivated by haematological disease Fluconazole 400 mg/d (oral/intravenous) NR 4 Placebo Tollemar, 993 Sweden Bone marrow transplanted patients L ip Amphotericin B mg kg d (intravenous) Placebo Ullmann, 2007 >2 countries Bone marrow transplanted patients with graft- versus- host disease Posaconazole 200 mg 3 /d (oral) Fluconazole 400 mg/d (oral) Van-Burik, 2004 United States and Canada Patients with any haematological disease Micafungin 50 mg/d (intravenous) Fluconazole 400 mg/d (intravenous) Vehreschild, 2007 Germany Patients with acute myeloid leukaemia Voriconazole 200 mg 2 /d (oral) Placebo Vreughdenhil, 993 The Netherlands Patients with any haematological disease receiving chemotherapy Itraconazole 200 mg 2 /d (oral) NR 3 Placebo Wingard, 987 United States Neutropenic bone marrow transplanted patients receiving chemotherapy Miconazole 5 mg kg d (intravenous) 97 NR Placebo 33 0 Wingard, 200 United States Patients with any haematological disease Fluconazole 200 mg 2 /d (oral/ intravenous) NR 5 Voriconazole 400 mg/d (oral/intravenous) Winston, 993 NR Neutropenic adults with acute leukaemia receiving chemotherapy Fluconazole 400 mg/d (oral) Placebo Young, 999 >2 countries Patients with leukaemia Fluconazole 200 mg/d (oral) Nystatin IU/mL 6 /d (oral) Data: NR, not reported; Lip Amphotericin B, liposomal amphotericin B.

6 535 NYS VOR MICA POS AB Lipo 3 FLU 6 PLA 3 KET MICO ITRA AB Conv FIGURE 2 Network of antifungal agent comparisons included in the analysis. Directly comparable treatments are linked with a line, the thickness of which corresponds to the number of trials that assessed the comparison. AB Conv: conventional amphotericin B; AB Lipo: liposomal amphotericin B; FLU: fluconazole; ITRA: itraconazole; KET: ketoconazole; MICO: miconazole; MICA: micafungin; NYS: nystatin; PLA: placebo; POSA: posaconazole; VOR: voriconazole 2 3 The network of the outcome incidence of aspergillosis was built with nine comparators: conventional amphotericin B, liposomal amphotericin B, fluconazole, itraconazole, micafungin, miconazole, posaconazole, voriconazole and placebo. The direct and indirect comparisons in the multiple treatment meta- analyses could not identify statistical differences among drugs (Figure S3 in Appendix S4). The rank shows that voriconazole, followed by liposomal amphotericin B and posaconazole, would probably be the best choice against aspergillosis, whereas miconazole is probably the worst in these patients (Figure S4 in Appendix S4). However, more RCTs are needed to strengthen this statement. The safety results regarding IFI- related mortality showed only one significant difference among drugs: placebo led to more death than posaconazole, showing the superior safety of this drug against control (see Figure S5 in Appendix S4 for OR with 95% CrI). No other statistical differences were observed. The network meta- analysis included both amphotericin B formulations, fluconazole, itraconazole, micafungin, placebo and posaconazole. In the ranking, posaconazole was the best treatment and liposomal amphotericin B and placebo were related to more deaths (Figure S6 in Appendix S4). Regarding the node- splitting technique, we did not identify any inconsistency factors (P values >.05 in all analyses), which demonstrates the robustness of our network. The multiple treatment comparisons also allowed for building ranks of the drugs for each measured outcome. Regarding the incidence of proven and suspected IFI, the rank order resulted in the probabilities of posaconazole followed by voriconazole being the best treatment. Otherwise, nystatin, placebo and ketoconazole were related to more IFI and therefore were the worst treatments (Figure S in Appendix S4). For the outcome incidence of candidiasis, we included arms. Differences were observed between fluconazole, itraconazole, miconazole, posaconazole and voriconazole vs placebo (see Figure 4 for OR with 95% CrI). Voriconazole and posaconazole were, respectively, the first and second best options against Candida spp. (Figure S2 in Appendix S4). 4 DISCUSSION Overall, the azoles posaconazole and voriconazole stood out as the most appropriate agents for IFI prophylaxis in this population. Of all antifungals, only posaconazole showed superiority to placebo concerning IFI- related mortality. Our findings are in agreement with recently published guidelines and recommendations. 4-20,3-33 Posaconazole was the most effective drug in reducing the incidence of IFI, occupying the first position in the ranking. This result is in agreement with the recommendations of the Infectious Diseases Working Party of the German Society for Hematology and Oncology regarding primary prophylaxis of IFI in AB Conv 0,77 (0,9, 2,4) AB Lip 0,35 (0,06,,22) 0,45 (0,2,,57) Ketoconazole,04 (0,28, 2,75),38 (0,52, 3,3) 2,99 (,0, 9,25) Fluconazole 0,83 (0,27,,79),06 (0,40, 2,9) 2,33 (0,80, 8,60) 0,78 (0,38,,78) Itraconazole,38 (0,8, 7,34),87 (0,30, 0,) 3,97 (0,64, 25,03),36 (0,29, 5,65),7 (0,30, 8,33) Micafungin 0,97 (0,6, 5,08),30 (0,28, 6,34) 2,82 (0,56, 6,54) 0,93 (0,23, 4,36),22 (0,26, 5,52) 0,7 (0,0, 6,33) Miconazole 0,26 (0,02, 2,0) 0,34 (0,03, 2,70) 0,75 (0,06, 6,98) 0,24 (0,03,,66) 0,3 (0,03, 2,28) 0,8 (0,0, 2,05) 0,26 (0,02, 2,62) Nystatin 0,35 (0,0, 0,84) 0,46 (0,2, 0,99),0 (0,39, 2,85) 0,33 (0,2, 0,58) 0,43 (0,22, 0,77) 0,25 (0,05,,24) 0,36 (0,08,,33),35 (0,9, 3,66) Placebo 2,09 (0,32, 8,47) 2,62 (0,58,,43) 5,80 (,27, 30,60),92 (0,62, 6,49) 2,5 (0,58, 9,39),42 (0,23, 0,) 2,07 (0,3, 2,93) 8,00 (0,83, 04,7) 5,76 (,55, 20,38) Posaconazole,79 (0,26, 6,36) 2,3 (0,45, 7,93) 4,98 (,0, 22,00),66 (0,47, 4,50) 2,9 (0,47, 6,69),26 (0,7, 6,93),82 (0,24, 9,55) 6,93 (0,69, 83,92) 4,99 (,30, 4,28) 0,88 (0,4, 3,7) Voriconazole FIGURE 3 Multiple treatment comparisons for incidence of proven and suspected invasive fungal infection consistency analysis based on network. Comparisons between treatments should be read from left to right, and the estimate is in the cell in common between the columndefining treatment and the row- defining treatment. Values are presented as odds ratio (OR) with credible interval (CrI). For this outcome, an OR higher than favours the incidence of IFI to the column- defining treatment. To obtain ORs for comparisons in the opposite direction, reciprocals should be taken. Significant results are in bold and underlined. Abbreviation AB Conv: conventional amphotericin B; AB Lip: liposomal amphotericin B

7 536 AB Conv 0.86 ( ) AB Lip 0.24 ( ) 0.24 ( ) Ketoconazole 2.29 ( ) 2.44 ( ) 9.79 ( ) Fluconazole.24 ( ).7 ( ) 4.74 ( ) 0.5 ( ) Itraconazole.07 ( ) 0.96 ( ) 3.94 ( ) 0.46 ( ) 0.85 ( ) Micafungin 2.23 ( ) 2.57 ( ) 8.84 ( ) 0.92 ( ).73 ( ) 2.0 ( ) Miconazole 0.65 ( ) 0.59 ( ) 2.26 ( ) 0.27 ( ) 0.49 ( ) 0.58 ( ) 0.29 ( ) Nystatin 0.27 ( ) 0.28 ( ).7 ( ) 0.2 ( ) 0.22 ( ) 0.26 ( ) 0.3 ( ) 0.44 ( ) Placebo 4.8 ( ) 5.07 ( ) 7.35 ( ).79 ( ) 3.53 ( ) 4.33 ( ) 2.06 ( ) 6.73 ( ) 5.53 ( ) Posaconazole 6.52 ( ) 2.96 ( ) 4.43 ( ).76 ( ) 3.42 ( ) 4.42 ( ).27 ( ) 7.40 ( ) 5.7 ( ).72 ( ) Voriconazole FIGURE 4 Multiple treatment comparisons for candidiasis consistency analysis based on network. Drugs are reported in alphabetical order. Comparisons between treatments should be read from left to right, and the estimate is in the cell in common between the columndefining treatment and the row- defining treatment. Values are presented as odds ratio (OR) with credible interval (CrI). For this outcome, an OR higher than favours the incidence of candidiasis to the column- defining treatment. To obtain ORs for comparisons in the opposite direction, reciprocals should be taken. Significant results are in bold and underlined. Abbreviation AB Conv: conventional amphotericin B; AB Lip: liposomal amphotericin B patients with haematologic malignancies. 33 The German group decided to decrease the strength of the recommendation for fluconazole compared to the previous guidelines because of its inferiority to posaconazole. In this document, posaconazole is recommended as IFI prevention in patients with chemotherapy- induced neutropenia with acute myeloid leukaemia or myelodysplastic syndrome and allogeneic haematopoietic stem cell transplant (HSCT) recipients with graftversus- host disease, based on two RCT. Fluconazole is recommended for invasive candidiasis before engraftment in allogeneic HSCT recipients, according to the Guidelines for Preventing Infectious Complications among Hematopoietic Cell Transplantation Recipients from the American Society for Blood and Marrow Transplantation, 32 and may be started from the beginning or just after the end of the conditioning regimen. Considering the prevention of infections specifically by Candida spp., our study showed that fluconazole is superior to placebo, ranked just after voriconazole and posaconazole. The first positions occupied by the third- generation azoles in the ranking may be related to the fact that fluconazole is not effective against Candida krusei and has variable activity against C. glabrata. Furthermore, posaconazole is already recommended as IFI prophylaxis against Candida infections, but only for post- engraftment at the date of publication of the ABSMT guidelines, because of the scarcity of RCTs. 6 Concerning the efficacy of systemic antifungals in preventing Aspergillus infections, the ranking suggests that voriconazole would be the most effective, followed by liposomal amphotericin B and posaconazole. According to the Infectious Diseases Society of America (IDSA) Guidelines for the Diagnosis and Management of Aspergillosis, 20 posaconazole, voriconazole and micafungin are indicated for invasive aspergillosis prophylaxis during prolonged neutropenia. The indication of posaconazole is supported by an open- label trial. 34 In our analysis, only the study of Ullmann et al. 35 included posaconazole, which was not considered as evidence in the IDSA guidelines for invasive aspergillosis prophylaxis. The use of voriconazole in the IDSA guidelines is supported by the work of Wingard et al., 36 and prophylaxis with micafungin is suggested according to Van Burik et al. 37 Both of these trials were included in our analysis. The IDSA guidelines 20 do not suggest the use of polyenes for invasive aspergillosis prophylaxis, claiming that no study has shown a reduction in the incidence of infections caused by Aspergillus spp. As a reference, they cited studies by Rousey et al., De Laurenzi et al. and Perfect et al The first two were not prospective randomized trials, and the latter was not double- blind. However, IDSA did not mention in its guidelines the studies of Kelsey et al. 4 and Tollemar et al., 42 which were both double- blind RCTs encompassing liposomal amphotericin B as intervention; these studies provided high- quality evidence on the subject. Possibly due to this difference in study selection between our analysis and the one by the IDSA, we reached a different conclusion. Based on our results, liposomal amphotericin B could be an option for prophylaxis considering its efficacy in preventing invasive aspergillosis, but more RCTs are needed to strength this evidence. Surprisingly, liposomal amphotericin B was associated with more deaths related to IFI according to the probability ranking. However, this result did not reach statistical significance and was only a tendency. This finding is possibly justified by the influence of the study by Rijnders et al. 43 For our analyses, we considered the intention- to- treat population of this study. In this trial, this population was defined as all patients who received at least one dose of liposomal amphotericin B, who were observed until 28 days after recovery of neutropenia and after the last cycle of chemotherapy, irrespective of whether the patient was able to continue the prophylaxis treatment. The death rate related to IFI was 4.5% and 3.6% in the placebo and liposomal amphotericin B groups, respectively. If we had considered the on- treatment population, that is patients who received weekly inhalations during all neutropenic episodes, the IFI- related mortality outcome would be considerably different: 6.2% in the placebo group and only.% in the liposomal amphotericin B group. 43 Hence, the polyene would not be linked to more deaths according to the probability ranking. Eighteen of the 26 RCTs (69%) included in this network metaanalysis were conducted before the year This shows that, despite the FDA approval of important antifungal agents such as caspofungin in 200, voriconazole in 2002 and posaconazole in 2006 (Drugs@FDA), fewer randomized double- blind studies are being

8 537 performed. This was reflected in our study by the inclusion of only two studies with voriconazole, one with posaconazole, and none with caspofungin. As these new azoles show promising results regarding primary prophylaxis of IFIs, more scientific investigation to assess their use is of paramount importance. Although some network meta- analyses have already been published on IFI prophylaxis in patients with haematological diseases, 3,22,23 our study differs considerably from these previous studies. Considering that these systematic reviews limited the antifungal classes evaluated and included open- label RCTs, the trials included in our review are mostly different. Bow et al. 3 included only five trials, of which two were also in our review. Zhao et al. 23 included 2 trials, and only nine were also present in our work. Pechlivanoglou et al. 22 included 25 RCTs, and only 2 were also present in our work, which included a total of 2 trials. Considering that our network meta- analysis used a different source of high- level evidence and reached similar conclusions to the previous studies, our research reinforces the practice of IFI prophylaxis with the last generation of azoles. 4. Limitations of the study Caspofungin and anidulafungin could not be evaluated by our network because none of the RCTs met the inclusion criterion of doubleblinding. We included only double- blind trials to reduce the risk of bias in our analysis, establishing a threshold of methodological quality during the selection of the studies to strengthen the clinical evidence. 27 We chose to include trials involving miconazole and ketoconazole because, despite not being recommended in clinical practice any longer for IFI prophylaxis, we did not want to lose any evidence from the literature. Considering the prevalence of older studies among the RCTs included in this network, the absence of a standardized definition of invasive fungal infection and outcomes was observed. Several studies considered persistent fever refractory to broad- spectrum antibiotics as an important outcome, referred to as suspected, possible or probable IFI. We decided to group proven and suspected fungal infections to gather more scientific evidence, as some articles presented these different categories of IFI, sometimes already aggregated. Another limitation was the difference in drug regimens and dosages among clinical trials, which makes it harder to determine the best therapeutic option. Some studies allowed the interchangeability between oral and intravenous regimens, which is highly necessary in clinical practice, but may complicate the estimation of the best regimen for each drug. We did not make any distinction regarding the haematological condition or the type of HSCT, to collect an adequate number of studies and include an appropriate number of patients, and thereby to perform consistent and reliable network meta- analyses. Beyond that, the idea of the analysis was to develop an overview about the best primary prophylaxis for haematological disorders. However, stratification into subgroups considering patient severity would be important to determine the best prophylaxis for each condition. For this to be possible, the initiation of high- quality RCTs considering different haematological conditions is necessary. ACKNOWLEDGEMENTS The authors would like to thank the Ministério da Ciência e Tecnologia, Ministério da Saúde, Conselho Nacional de Desenvolvimento Científico e Tecnológico, Coordenação de Aperfeiçoamento de Pessoal de Nível Superior and Secretaria da Ciência, Tecnologia e Ensino Superior for the financial support for laboratory infrastructure. CONFLICT OF INTEREST The authors declare that they have no conflict of interest. REFERENCES. Ziakas PD, Kourbeti IS, Mylonakis E. Systemic antifungal prophylaxis after hematopoietic stem cell transplantation: a meta- analysis. Clin Ther. 204;36: Fung M, Kim J, Marty FM, Schwarzinger M, Koo S. Meta- analysis and cost comparison of empirical versus pre- emptive antifungal strategies in hematologic malignancy patients with high- risk febrile neutropenia. Plos One. 205;0: Kontoyiannis DP, Marr KA, Park BJ, et al. Prospective surveillance for invasive fungal infections in hematopoietic stem cell transplant recipients, : overview of the Transplant- Associated Infection Surveillance Network (TRANSNET) Database. Clin Infect Dis. 200;50: Neofytos D, Horn D, Anaissie E, et al. Epidemiology and outcome of invasive fungal infection in adult hematopoietic stem cell transplant recipients: analysis of Multicenter Prospective Antifungal Therapy (PATH) Alliance registry. Clin Infect Dis. 2009;48: de Oliveira RB, Atobe JH, Souza SA, de Castro Lima Santos DW. Epidemiology of invasive fungal infections in patients with acquired immunodeficiency syndrome at a reference hospital for infectious diseases in Brazil. Mycopathologia. 204;78: Shen YZ, Qi TK, Ma JX, et al. Invasive fungal infections among inpatients with acquired immune deficiency syndrome at a Chinese university hospital. Mycoses. 2007;50: Wang JF, Xue Y, Zhu XB, Fan H. Efficacy and safety of echinocandins versus triazoles for the prophylaxis and treatment of fungal infections: a meta- analysis of RCTs. Eur J Clin Microbiol Infect Dis. 205;34: Cugno C, Cesaro S. Epidemiology, risk factors and therapy of candidemia in pediatric hematological patients. Pediatr Rep. 202;4:e9. 9. Colombo AL, Nucci M, Park BJ, et al. Epidemiology of candidemia in Brazil: a nationwide sentinel surveillance of candidemia in eleven medical centers. J Clin Microbiol. 2006;44: Wingard JR, Ribaud P, Schlamm HT, Herbrecht R. Changes in causes of death over time after treatment for invasive aspergillosis. Cancer. 2008;2: Cornely OA, Aversa F, Cook P, et al. Evaluating the role of prophylaxis in the management of invasive fungal infections in patients with hematologic malignancy. Eur J Haematol. 20;87: Freifeld AG, Bow EJ, Sepkowitz KA, et al. Clinical practice guideline for the use of antimicrobial agents in neutropenic patients with cancer: 200 update by the Infectious Diseases Society of America. Clin Infect Dis. 20;52:e56 e Bow EJ, Vanness DJ, Slavin M, et al. Systematic review and mixed treatment comparison meta- analysis of randomized clinical trials of primary oral antifungal prophylaxis in allogeneic hematopoietic cell transplant recipients. BMC Infect Dis. 205;5: Flowers CR, Seidenfeld J, Bow EJ, et al. Antimicrobial prophylaxis and outpatient management of fever and neutropenia in adults treated for malignancy: American Society of Clinical Oncology clinical practice guideline. J Clin Oncol. 203;3:

9 Hicheri Y, Toma A, Maury S, Pautas C, Mallek-Kaci H, Cordonnier C. Updated guidelines for managing fungal diseases in hematology patients. Expert Rev Anti Infect Ther. 200;8: Pappas PG, Kauffman CA, Andes DR, et al. Clinical practice guideline for the management of candidiasis: 206 update by the Infectious Diseases Society of America. Clin Infect Dis. 206;62:e e Cornely OA, Bassetti M, Calandra T, et al. ESCMID* guideline for the diagnosis and management of Candida diseases 202: nonneutropenic adult patients. Clin Microbiol Infect. 202;8(Suppl 7): Barberan J, Mensa J, Llamas JC, et al. Recommendations for the treatment of invasive fungal infection caused by filamentous fungi in the hematological patient. Rev Esp Quimioter. 20;24: Ruiz-Camps I, Aguado JM, Almirante B, et al. Guidelines for the prevention of invasive mould diseases caused by filamentous fungi by the Spanish Society of Infectious Diseases and Clinical Microbiology (SEIMC). Clin Microbiol Infect. 20;7: Patterson TF, Thompson GR 3rd, Denning DW, et al. Practice guidelines for the diagnosis and management of aspergillosis: 206 update by the Infectious Diseases Society of America. Clin Infect Dis. 206;63:e e Chau MM, Kong DC, van Hal SJ, et al. Consensus guidelines for optimising antifungal drug delivery and monitoring to avoid toxicity and improve outcomes in patients with haematological malignancy, 204. Intern Med J. 204;44: Pechlivanoglou P, Le HH, Daenen S, Snowden JA, Postma MJ. Mixed treatment comparison of prophylaxis against invasive fungal infections in neutropenic patients receiving therapy for haematological malignancies: a systematic review. J Antimicrob Chemother. 204;69:. 23. Zhao YJ, Khoo AL, Tan G, et al. Network meta- analysis and pharmacoeconomic evaluation of fluconazole, itraconazole, posaconazole, and voriconazole in invasive fungal infection prophylaxis. Antimicrob Agents Chemother. 205;60: Robenshtok E, Gafter-Gvili A, Goldberg E, et al. Antifungal prophylaxis in cancer patients after chemotherapy or hematopoietic stemcell transplantation: systematic review and meta- analysis. J Clin Oncol. 2007;25: Jorgensen KJ, Gotzsche PC, Dalboge CS, Johansen HK. Voriconazole versus amphotericin B or fluconazole in cancer patients with neutropenia. Cochrane Database Syst Rev. 204;(2):CD Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta- analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ. 2009;339:b Higgins JPT, Green S. Cochrane Handbook for Systematic Reviews of Interventions Version The Cochrane Collaboration; Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials. 996;7: Hoaglin DC, Hawkins N, Jansen JP, et al. Conducting indirecttreatment- comparison and network- meta- analysis studies: report of the ISPOR task force on indirect treatment comparisons good research practices: part 2. Value Health. 20;4: Higgins JP, Jackson D, Barrett JK, Lu G, Ades AE, White IR. Consistency and inconsistency in network meta- analysis: concepts and models for multi- arm studies. Res Synth Methods. 202;3: Walsh TJ, Anaissie EJ, Denning DW, et al. Treatment of aspergillosis: clinical practice guidelines of the Infectious Diseases Society of America. Clin Infect Dis. 2008;46: Tomblyn M, Chiller T, Einsele H, et al. Guidelines for preventing infectious complications among hematopoietic cell transplant recipients: a global perspective. Bone Marrow Transplant. 2009;44: Cornely OA, Böhme 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;94: Cornely OA, Maertens J, Winston DJ, et al. Posaconazole vs. fluconazole or itraconazole prophylaxis in patients with neutropenia. N Engl J Med. 2007;356: Ullmann AJ, Lipton JH, Vesole DH, et al. Posaconazole or fluconazole for prophylaxis in severe graft- versus- host disease. N Engl J Med. 2007;356: Wingard JR, Carter SL, Walsh TJ, et al. Randomized, double- blind trial of fluconazole versus voriconazole for prevention of invasive fungal infection after allogeneic hematopoietic cell transplantation. Blood. 200;6: van Burik JA, Ratanatharathorn V, Stepan DE, et al. Micafungin versus fluconazole for prophylaxis against invasive fungal infections during neutropenia in patients undergoing hematopoietic stem cell transplantation. Clin Infect Dis. 2004;39: Rousey SR, Russler S, Gottlieb M, Ash RC. Low- dose amphotericin B prophylaxis against invasive Aspergillus infections in allogeneic marrow transplantation. Am J Med. 99;9: De Laurenzi A, Matteocci A, Lanti A, Pescador L, Blandino F, Papetti C. Amphotericin B prophylaxis against invasive fungal infections in neutropenic patients: a single center experience from 980 to 995. Infection. 996;24: Perfect JR, Klotman ME, Gilbert CC, et al. Prophylactic intravenous amphotericin B in neutropenic autologous bone marrow transplant recipients. J Infect Dis. 992;65: 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. 999;23: Tollemar J, Ringden O, Andersson S, Sundberg B, Ljungman P, Tyden G. Randomized double- blind study of liposomal amphotericin B (Ambisome) prophylaxis of invasive fungal infections in bone marrow transplant recipients. Bone Marrow Transplant. 993;2: Rijnders BJ, Cornelissen JJ, Slobbe L, et al. Aerosolized liposomal amphotericin B for the prevention of invasive pulmonary aspergillosis during prolonged neutropenia: a randomized, placebo- controlled trial. Clin Infect Dis. 2008;46: SUPPORTING INFORMATION Additional Supporting Information may be found online in the supporting information tab for this article. How to cite this article: Leonart LP, Tonin FS, Ferreira VL, et al. A network meta- analysis of primary prophylaxis for invasive fungal infection in haematological patients. J Clin Pharm Ther. 207;42: