Quantity of Research Available
A total of 580 citations were identified in the literature search. Following screening of titles and abstracts, 541 citations were excluded and 39 potentially relevant reports from the electronic search were retrieved for full-text review. Fifteen potentially relevant publications were retrieved from the grey literature search. Of these 54 potentially relevant articles, 33 publications were excluded for various reasons, while 21 publications, including five systematic reviews, three randomized controlled trials (RCTs), eight non-randomized comparative studies, and five evidence-based practice guidelines met the inclusion criteria and were included in this report. Appendix 1 describes the PRISMA flowchart of the study selection.
Summary of Study Characteristics
The body of evidence included five systematic reviews and meta-analyses,5,16–19 three RCTs,20–22 eight non-randomized studies,23–30 and five evidence-based practice guidelines6,8,31–33 addressing the effectiveness of posaconazole in the prophylaxis and treatment of invasive aspergillosis. General characteristics of the included studies are summarized below. More details are available in Appendix 2: Table 2, Table 3, Table 4, and Table 5.
Study Design
All of the five identified systematic reviews5,16–19 performed comprehensive searches in two or more electronic bibliographic databases and the grey literature. The latest end date of the literature search was November 2014 for the study by Zhao et al.16 All five systematic reviews included RCTs of the primary antifungal prophylaxis with posaconazole and other antifungal agents, and used meta-analyses to synthesize data from their included studies. Mixed comparison (network meta-analysis) methods were used in three of the systematic reviews.16,17,19
All of the three included RCTs20–22 were multi-centre trials that used a parallel group RCT design. One trial22 blinded both study participants, and the researchers to the study interventions and outcomes, using a double-dummy, double-blind technique. In one trial21 only the evaluators of the study outcomes were blinded to the study interventions. The third trial did not use a blinding technique.20
Of the eight included non-randomized studies,23–30 two were open-label multi-centre studies,23,24 four were single-centre retrospective cohort studies,25–28, and two were single-centre prospective cohort studies.29,30
Five evidence-based guidelines were identified,6,8,31–33 all of which used an expert panel consensus method informed by structured literature searches to make clinical recommendations.
Country of Origin
The included systematic reviews originated from Singapore (n = 1),16 Canada (n = 2),17,19 the United States (n = 1),18 and China (n = 1 ).5
Two of the included RCTs were performed internationally in various countries in Europe, North and Latin America, Asia, Australia, and Africa,21,22 and one was originated from China.20
One of the included non-randomized studies was a multinational study conducted in Europe, the United States Australia and Latin America,24 two originated from Germany,26,30 and one each from the United States,23 Korea,25 Italy,27 Austria,28 and France.29
The included practice guidelines were produced in the United States (n = 2),6,31 Canada (n = 1),8 or Europe (n = 2).32,33
Patient Population
All of the included systematic reviews5,16–19 focused on the antifungal agents as a prophylaxis in patients with hematological malignancies (including AML and myelodysplastic syndrome) and those undergoing HSCT. The patient population was limited to adult patients (definition of “adult” not provided) in two of the systematic reviews,16,19 and the remaining three reviews did not report any age restrictions.5,17,18
Two RCTs included neutropenic patients with AML or myelodysplastic syndrome who underwent chemotherapy,20,21 and one RCT recruited patients with acute graft-versus-host disease (GVHD) after HSCT.22 All RCTs examined prophylactic effects of antifungal interventions on their populations of interest. Two of the three RCTs included patients 13 years of age or older,21,22 while one RCT focused on patients between 18 and 70 years of age.20
Two non-randomized studies focused their assessments on salvage therapy with posaconazole in patients hematologic malignancies who had a confirmed diagnosis of IA.23,24 The patients who were included in these two studies were refractory to conventional antifungal treatments due to renal failure, severe drug toxicity, or other organ dysfunction. Five of the remaining six non-randomized studies used posaconazole and its comparators as a primary antifungal prophylaxis in adult patients with AML or myelodysplastic syndrome undergoing remission or induction chemotherapy.25,27–30 One study in this group also included patients who underwent HSCT.28 The sixth non-randomized study included pediatric patients with hematologic malignancies and prolonged chemotherapy-induced neutropenia.26
The included practice guidelines focused their recommendations on the diagnosis, prevention and management of IFIs, including IA, in patients with hematological malignancies (AML and myelodysplastic syndrome) undergoing chemotherapy, those receiving HSCT, and other immunocompromised patients at risk of IA.6,8,31–33 One guideline focused on the use of antifungal agents in children (patients one month to less than 19 years of age),8 while the other four considered adult and adolescent patient populations (aged 13 years and older).6,31–33
Interventions and Comparators
All of the systematic reviews used posaconazole as an intervention in at least one of their included studies. The comparators included: fluconazole,5,16–19 itraconazole,5,16–19 voriconazole,5,16–19 conventional (deoxycholate)18 or liposomal18,19 formulations of amphotericin B, micafungin,18,19 caspofungin,18 and anidulafungin.18
In the majority of included RCTs and non-randomized studies, posaconazole was administered as a study intervention to adults in the form of oral suspension in doses of 200 mg three times daily (for antifungal prophylaxis)20–22,25,27–30 and 200 mg four times daily (for the salvage therapy of IA).23,24 One non-randomized study administered posaconazole to children in 4mg/kg doses three times daily, up to a maximum dose of 200 mg three times daily.26 The study comparators included: fluconazole,20–22,25,26,28 itraconazole,26,28 high-dose lipid formulation of amphotericin B,23 conventional or liposomal amphotericin B,27 topical polyenes,30 best available standard of care (including the aforementioned drugs),24 or no prophylaxis.29
Outcomes
The majority of the included studies reported on the incidence (or time to occurrence) of proven or probable IFIs,5,16–28,30 incidence of IA,21,22,26–30 overall mortality,20,22,23,25,27–30 IFI-related mortality,25,27,30 and fungal-free survival.25,30 The use of pre-emotive and empiric antifungal therapy,25,27 and safety measures21,22,24,26,27 were also reported.
Summary of Critical Appraisal
Detailed summaries of the critical appraisal of the included systematic reviews, RCTs, non-randomized studies, and evidence-based guidelines are provided in Appendix 3: Table 6, Table 7, Table 8, Table 9, respectively.
In all of the included systematic reviews, an a priori design was used; a comprehensive literature search was conducted in multiple databases, the studies were reviewed and selected by two independent reviewers, and the findings of the included studies were combined using appropriate methods.5,16–19 All of the systematic reviews provided a list of the included studies. The quality of the included RCTs was assessed and the results of quality appraisal were reported in four systematic reviews.5,16,18,19 It was not clear if a quality appraisal of the included studies was undertaken the systematic review by Bow et al.17 None of the systematic reviews presented a list of excluded studies, neither did they assess the likelihood of publication bias. All of the systematic reviews included a conflict of interest declaration, and stated the source of finding for the systematic review.5,16–19 However, data on the funding sources of the primary studies had been extracted in one of the five systematic reviews.16
The study objectives, eligibility criteria, interventions, main outcomes, and the characteristics of study groups were clearly described in all three included RCTs.20–22 All three RCTs used reliable and valid outcome measures, used appropriate statistical tests, and reported the point estimates of the effect alongside the estimates of random variability. In one RCT, both patients and evaluators were blind to the study interventions;22 one RCT blinded only the outcome evaluators;21 and in one RCT, intervention assignment and outcomes assessment were not carried out in a blind manner.20 Lack of blinding in RCTs could bias the study results by increasing the risk of differential co-interventions, administration and outcome measurement considerations, and patient drop outs. None of the three RCTs described the characteristics of the participants who were lost to follow up. Lost to follow up can result in missing or incomplete outcome data, and hence, introduce a systematic bias into the study because the patients who are lost may have poorer or better outcomes than those retained in the study. One RCT accounted for potential confounders in the analysis,22 while the other two RCTs did not use any analytical methods to adjust for the confounders. Confounding factors may be a threat to the internal validity of the study, as they can lead to an over- or under-estimation of the true association between the exposure and outcome variables, or change the direction of the effect. Two of the three RCTs reported their power calculation methods to determine the sample sizes required for the detection of a difference between groups at the desired level of significance in the statistical comparisons.20,22
All of the included non-randomized studies clearly described the study objectives, eligibility criteria, interventions, main outcomes, and the characteristics of study groups.23–30 Four studies used historical controls who were not recruited (or admitted to the research hospital) at the same time as the participants in the posaconazole group.23–25,27 In one study it is unclear if the participants in the posaconazole group and those in the control group were recruited from the same reference population.23 Selection bias may be introduced when the study comparison groups are sampled from different source populations. This is mainly because the selection of participants in one group may result in a different outcome than the selection in the other group. Two studies used matched controls to reduce the risk of selection bias.29,30 Michallet et al.29 chose comparison groups that were similar in terms of potential confounders such as age, sex, AML type, chemotherapy regimen, prognostic group, response to induction, etc. Vehreschils et al.30 included controls who were similar to the posaconazole group in terms of age, sex , and other confounding factors, except the type of chemotherapy regimen, length of hospitalization, and the use of granulocyte colony-stimulating factor. Estimates of random variability were not provided in two studies.26,28 Two of the non-randomized studies used a prospective cohort design,29,30 while the data was collected retrospectively in four studies.25–28 Two studies used a prospective data collection in the posaconazole groups, with a retrospective chart review in their external control groups.23,24 Retrospective studies are prone to a higher risk of bias due to the issues related to the quality (accuracy, reliability, and completeness) of pre-existing data. Six studies of prophylaxis with posaconazole were conducted in single university-based referral hospitals.25–30 This may limit the generalizability (external validity) of the study findings to other practice settings.
All of the included guidelines described their objectives, clinical problems of interest, and the patient populations to whom guidelines were meant to apply.6,8,31–33 All of the guidelines were developed by guideline panels which consisted of experts from relevant professional groups, and combined expert consensus with structured literature reviews and evidence syntheses to formulate recommendations. All five guidelines were considered to be evidence-based and explicitly linked their recommendations to valid citations and provided levels of evidence. However, the details of the literature search strategies were reported by three guidelines.6,8,32 The search strategies used by these guidelines consisted of a search in two or more bibliographic databases along with the grey literature search in the websites of relevant professional organizations, as well as conference abstracts. Two guidelines did not provide a detailed information about their literature search criteria.31,33 Although all the major studies in the field of antifungal therapy for IA were discussed and linked to the recommendations in these two guidelines, further details of the literature search process used in development of the recommendations would have made the guidelines more reliable. Patients’ input was sought by two of the guidelines.8,33 One guideline specified its targeted users.8 The target audiences were not clearly described in the remaining four guidelines. Four guidelines were validated through internal and external review and approval processes.6,8,32,33 Implementation and cost issues were described in only one guideline.8 This guideline also provided the users with monitoring and auditing criteria.8
Summary of Findings
The overall findings of the review are summarized below. Additional details are available in Appendix 4: Table 10, Table 11, Table 12, Table 13.
Major conclusions from all studies are summarized in Appendix 5, Table 14.
What is the clinical effectiveness of posaconazole for the treatment of invasive aspergillosis in patients of any age?
No studies reporting on the effectiveness of posaconazole as a first-line treatment of IA were identified.
Two non-randomized studies used posaconazole for salvage therapy in hematologic malignancy patients with refractory IA.23,24
Treatment response rate
Overall response to salvage therapy was significantly higher in posaconazole-treated patients with refractory IA, when compared with high-density lipid formulation amphotericin B (40% versus 8%; P < 0.01),23 or with the best available standard care (42% versus 26%; P = 0.006).24
Mortality
When compared with standard care, posaconazole was associated with a significantly higher survival rate at 30 days after salvage therapy (74% versus 49%; P = 0.0003). This beneficial effected lasted until the end of the treatment (up to 372 days after initiation of the treatment; 38% versus 22%; P = 0.0003).24
When compared with high-density lipid amphotericin B, a 12-weeks overall survival rate was significantly higher in the posaconazole group. The Aspergillus-related mortality rate was reported to be statistically lower in the posaconazole group than that in the high-density lipid amphotericin B group (40% versus 69%; P = 0.008).
What is the clinical effectiveness of posaconazole for either prophylaxis of invasive aspergillosis in patients who have undergone stem cell transplants or patients with cancer at high risk of neutropenia?
Five systematic reviews,5,16–19 three RCTs,20–22 and six non-randomized studies,25–30 reported on the effectiveness of posaconazole a primary antifungal prophylaxis in HSCT-recipients or cancer patients.
Incidence of proven and probable IFIs
Three systematic reviews reported pooled estimates on this outcome from both direct and indirect posaconazole comparisons.16,17,19 The results from these systematic reviews were consistent in showing that, when used for prophylaxis, posaconazole was superior to placebo,16 fluconazole,5,16,17,19 itraconazole,5,16,19 and voriconazole16 in preventing IFIs.
Data from the included RCTs showed that during the treatment phase posaconazole was non-inferior to fluconazole in preventing IFIs.20,22 However, when compared with a group of controls who received either fluconazole or itraconazole prophylaxis, posaconazole provided a statistically better protection against IFIs (2% versus 8% incidence rate; P < 0.001).21
Five non-randomized studies reported on the incidence of IFIs after posaconazole prophylaxis.25–28,30 Three of these studies reported significantly fewer breakthrough IFIs in patients treated with posaconazole than with fluconazole,25 amphotericin B,27 and topical polyenes.30 One study that included pediatric patients reported statistically similar IFI incidence rates in posaconazole, fluconazole, and itraconazole groups.26 One study reported similar rates between posaconazole and a historical control group in terms of preventing IFIs in adult patients.28
Incidence of IA
Four systematic reviews,5,16,17,19 two RCTs,21,22 and two non-randomized studies29,30 reported on the incidence of IA.
Data from the included systematic reviews based on direct and indirect comparisons indicated that posaconazole had a greater effect on reducing IA infections, when compared with placebo,16,19 fluconazole,5,16,17,19 itraconazole,5,16,19 and voriconazole.16
Similarly, the included RCTs and non-randomized studies reported statistically lower IA incidence rates in posaconazole groups, as compared with either fluconazole or itraconazole prophylaxis,21,22 topical use of polyenes,30 or no prophylaxis.29
Mortality
Overall (all-cause) mortality was reported in four systematic reviews,5,16,17,19 three RCTs,20–22 and six non-randomized studies.25–30 However, two of the non-randomized studies26,28 did not report mortality rates in the control group. IFI-related mortality was reported in two systematic reviews,16,19 and two RCTs.21,22
The systematic review and meta-analysis results showed that posaconazole prophylaxis had a statistically greater effect on reducing mortality when compared with placebo,16,19 fluconazole,5,16, or itraconazole.5,16,19 The systematic review by Bow et al17 reported no statistically significant difference between the posaconazole and fluconazole groups in terms of overall mortality.
The included RCTs showed statistically similar overall mortality20,22 or cumulative survival rates21 between the posaconazole and fluconazole groups.
The results of the non-randomized studies showed that the effect of posaconazole in reducing overall mortality was comparable to that of fluconazole,25 amphotericin B,27 or topical polyenes.30 In one study, all-cause mortality rate was statistically lower in the posaconazole prophylaxis arm than the no-prophylaxis control group (P = 0.0023).29
Posaconazole was reported to statistically reduce the rate of IFI-related mortality, when compared to fluconazole,16,19,21,22 or itraconazole.19,21
What are the evidence-based guidelines associated with the use of posaconazole for the prophylaxis and treatment of invasive aspergillosis in any patient?
The guidelines of the Infectious Diseases Society of America (IDSA) recommend posaconazole as an alternative to voriconazole for salvage therapy of IA [strong recommendation; moderate-quality evidence], and for the primary antifungal prophylaxis against IA in patients at high risk of neutropenia (i.e., patients with AML or myelodysplastic syndrome) [strong recommendation; high quality evidence], and HSCT recipients with GVHD [strong recommendation; high quality evidence].6
The National Comprehensive Cancer Network (NCCN) guidelines recommend posaconazole for antifungal prophylaxis in neutropenic patients with AML and myelodysplastic syndrome receiving induction or re-induction chemotherapy, until resolution of neutropenia [weak recommendation; moderate-quality evidence], and in HSCT recipients requiring intensive immunosuppressive therapy for GVHD, until resolution of significant GVHD [weak recommendation; moderate-quality evidence].31
The European Conference on Infections in Leukemia (ECIL-6) guidelines recommend posaconazole for salvage therapy in patients with IA [Grade BII recommendation].32
The guidelines of the Infectious Disease Working Party of the German Society of Hematology and Medical Oncology Association (AGIHO/DGHO) recommend posaconazole oral suspension as standard of care for antifungal prophylaxis in patients with chemotherapy-induced neutropenia with AML or myelodysplastic syndrome, and allogeneic HSCT recipients with GVHD [Grade AI recommendation].33
The guidelines of the C17 council of pediatric hematology and oncology centres in Canada do not routinely recommend posaconazole in pediatric patients. However, the guidelines panel suggests posaconazole prophylaxis as an alternative to fluconazole in patients aged at least 13 years with AML or myelodysplastic syndrome during chemotherapy, especially for those considered at higher risk for IA [weak recommendation; moderate-quality evidence]. The guidelines also recommend posaconazole as a consideration in children at least 13 years of age with GVHD after allogeneic HSCT [weak recommendation; moderate-quality evidence].8
Limitations
No relevant evidence was found on the use of posaconazole as a first-line treatment option for IA. The review also found a paucity of primary studies that reported on the effectiveness of posaconazole in children at high risk of IA.
Most of the included systematic reviews used a mixed-treatment analysis method, which combined direct and indirect evidence to simultaneously compare multiple treatments. This type of meta-analysis tends to be valid for studies that are adequately similar in terms of methodological and clinical characteristics.34 Given the methodological heterogeneity (trial populations, type of the fungal infections for which the patients were treated, risk of IFIs, etc.) that existed between the RCTs included in the analyses, the possibility of potential bias, stemming from the use of inconsistent data, cannot be ruled out. Therefore, the results of the indirect comparisons should be interpreted with caution.
Most of the included non-randomized studies were conducted in single university-based referral hospitals, and used a retrospective data collection methods and historical controls. These issues can be considered a potential threat to the generalizability of the findings of these studies. None of the included RCTs or non-randomized studies were conducted in Canada. Therefore, it is not clear to what extent the results will be applicable to a Canadian setting.
