Efficacy and safety of combination antifungal therapy in Korean haematological patients with invasive aspergillosis

Summary This randomised, double‐blind, placebo‐controlled trial assessed the efficacy, safety and tolerability of voriconazole+anidulafungin (combination) or voriconazole+placebo (monotherapy) for invasive aspergillosis (IA; NCT00531479). We present a post hoc analysis of Korean and non‐Korean patients with IA (including baseline positive serum galactomannan [GM]). Immunocompromised patients ≥ 16 years with IA were randomised 1:1, combination or monotherapy, for ≥ 2 weeks’ treatment. The primary endpoint was 6‐ and 12‐week all‐cause mortality (Korean modified intent‐to‐treat [mITT] population). Overall, 454 patients enrolled (Koreans: 56 [combination: 28, monotherapy: 28], non‐Koreans: 398 [combination: 200, monotherapy: 198]). The mITT population comprised 40 Koreans (combination: 23; monotherapy: 17) and 237 non‐Koreans (combination: 112; monotherapy: 125). Week 6 treatment difference in mortality rate between combination and monotherapy was −6.4% in non‐Koreans. This reduction was more marked in Koreans (−22.4%). Week 12 difference in all‐cause mortality between combination and monotherapy was −17.7% (Koreans) and −20.2% at Week 6 (Koreans; positive baseline GM). Week 6 mortality (Koreans [mITT]; baseline GM >0.5‐2.0) was 0/13 (combination) and 2/6 (monotherapy). Serious adverse events were numerically higher for combination than monotherapy (Koreans: 57.1%, 46.4%; non‐Koreans: 49.5%, 46.0%). In Koreans, combination therapy was associated with marginally better outcomes than monotherapy and more so than in non‐Koreans.


| INTRODUC TI ON
Invasive aspergillosis (IA) is an important cause of morbidity and mortality in patients with haematologic malignancies, as well as those who have undergone allogeneic haematopoietic stem cell transplantation (HSCT). [1][2][3] Risk factors for mortality in patients with IA following HSCT include the degree of compromised immunity and dissemination of infection, thus suggesting the importance of prompt diagnosis, initiation of therapy and reduction in immunosuppression after diagnosis. 3 Treatment with amphotericin B deoxycholate has been shown to be effective in the treatment of patients with IA but its use may be limited by nephrotoxicity. 4 In a prospective, randomised study, voriconazole led to better outcomes compared with amphotericin B deoxycholate as primary therapy for IA at Week 12 in the modified intent-to-treat (mITT) population 5 and is now the initial standard treatment for IA. 6 In vitro studies have demonstrated synergistic activity of voriconazole and echinocandins against Aspergillus species. [7][8][9][10] In patients with IA who experienced failure of initial therapy with amphotericin B formulations, a combination of voriconazole and caspofungin has been shown to improve survival compared with voriconazole alone, with fewer patients dying of IA 3 months after receipt of salvage therapy. 11 However, other studies have demonstrated no survival advantage among patients who received combination therapy with voriconazole and caspofungin compared with voriconazole monotherapy. 3,12 More recently, a randomised, double-blind, placebo-controlled trial assessed the safety and efficacy of voriconazole and the echinocandin anidulafungin, compared with voriconazole monotherapy in the treatment of IA. 13 14 Serial GM measurements have been investigated for monitoring treatment response or as a surrogate endpoint to assess clinical outcomes in patients with IA. [15][16][17] In the present study, we report the results of a post hoc analysis of data from the Korean sub-population of a primary double-blind, randomised study. 13 The rationale for evaluating this sub-group is that about 13%-23% of the Asian population inherit a cytochrome P450 2C19 (CYP2C19) poor metabolizer genotype compared with 3%-5% observed in Caucasians. [18][19][20][21][22] As voriconazole is mainly metabolised by CYP2C19, which contains known polymorphisms (eg poor metabolizer genotype) that reduce the activity of this enzyme, data on the response to treatment in this sub-population were explored. [23][24][25] Therefore, this post hoc analysis aimed to evaluate the efficacy and safety of voriconazole and anidulafungin combination therapy and voriconazole monotherapy in Korean and non-Korean patients with IA, including those with a positive serum GM at baseline.

| PATIENTS AND ME THODS
The present investigation was a post hoc analysis of sub-groups of Korean and non-Korean patients who were included in a randomised controlled trial of the effectiveness and safety of a combination of anidulafungin and voriconazole, and voriconazole alone, for the treatment of IA. The authors confirm that the ethical policies of the journal, as noted on the journal's author guidelines page, have been adhered to and the appropriate ethical review committee approval has been received. The study was conducted in accordance with the International Conference on Harmonization Good Clinical Practice Guidelines and ethical principles originating in or derived from the Declaration of Helsinki, approved by the appropriate institutional review boards, and registered on ClinicalTrials.gov (primary study: NCT00531479).
Written informed consent was obtained for all patients.

| Patient eligibility (primary study)
Briefly, patients were included if they were male or female, aged ≥ 16 years and immunocompromised because of allogeneic HSCT after myeloablative/non-myeloablative conditioning or haematologic malignancies. 13 Patients had to be diagnosed with proven, probable or possible IA as revised by the modified EORTC/MSG consensus definitions 14 ; (Appendix S1: Table S1). Any patients enrolled with possible IA had to have a proven or probable diagnosis established within 7 days of enrolment.
Among the reasons for excluding patients were progressive haematologic disease that was unlikely to respond to treatment; systemic anti-mould antifungals for treatment of IA for ≥ 4 days; severe liver dysfunction (>5× upper limit of normal); or a Karnofsky score <20. Other criteria included anticipated death within 30 days due to non-infectious factors, those who required mechanical ventilation, were pregnant or lactating, had received interacting drugs (eg rifampin), or were allergic or experienced serious reactions to azoles or echinocandins. 13

| Design and endpoints (primary study)
The primary study was a randomised, double-blind, placebo-controlled multicentre trial in which patients were randomised 1:1 to either voriconazole plus anidulafungin (combination therapy) or voriconazole plus placebo (monotherapy). 13 The primary endpoint was allcause mortality at 6 weeks in patients with proven or probable IA in the mITT group. The mITT population included all randomised patients with proven or probable IA confirmed by Day 7 following enrolment, adjudicated by a data review committee (DRC), and who received at least one dose of study drug. Treatment randomisation (Appendix S1: Figure S1) and secondary endpoints are described in Appendix S1. 13

| Patients (post hoc analysis)
In the present analysis, a sub-group of patients from the primary study were included, with the objective of comparing the efficacy and safety of combination therapy or monotherapy in Korean and non-Korean patients. Patients were included if they had been randomised to receive voriconazole and anidulafungin in combination or voriconazole monotherapy (plus placebo) for a minimum of 2 weeks and met the criteria for the mITT population. Participants had to have a proven or probable IA diagnosis, based on EORTC/MSG criteria and confirmed by an independent DRC, and must have received therapy.

| Endpoints (post hoc analysis)
The primary endpoint was all-cause mortality at Week 6 in the Korean mITT population. In addition, sub-group analyses of all-cause mortality at Weeks 6 and 12 were performed in Korean and non-Korean mITT patients.
A GM screening assay has been evaluated as a potential diagnostic tool for early diagnosis of IA in patients receiving allogeneic HSCT. 15 In this analysis, we also examined all-cause mortality at

| Statistical methods
Statistical methods for the primary analysis, including a review of safety data, have been reported elsewhere. 13 Mortality rate was based on the Kaplan-Meier (KM) product limit estimator. Treatment difference (stratified) was based on a weighted difference in proportions (d), using the following formula: were based on a one-sided test.
The safety and tolerability of combination therapy and monotherapy in Korean and non-Korean sub-groups were summarised by descriptive statistics, including overall numbers of adverse events (AEs), patients with AEs, TEAEs, discontinuations due to AEs or requiring dose reduction.  (Table 1). Among Korean patients, the baseline demographic characteristics were comparable in the two treatment arms.

| Safety endpoints
In the primary analysis, the safety profile of the combination of voriconazole and anidulafungin was similar to that of voriconazole monotherapy, except for a greater proportion of patients in the combination group had hepatobiliary AEs than in the monotherapy group (12.7% vs 8.4%). 13 The overall frequencies of TEAEs of all causalities in both populations are shown in Table 4 and by system organ class in Table 5.
In this post hoc analysis, the most common TEAEs in the Korean (Appendix S1: Table S2).
The mortality rate in Korean patients (safety population) at Week 6 was 14.7% (4/28) in the combination therapy arm and 32.1% (9/28) in the monotherapy arm, with a similar pattern at Week 12 (Appendix S1: Table S3). In the non-Korean population, the mortality rates at Week 6 were similar in the two therapy arms, respectively (21.0% and 22.6%) (Appendix S1: Table S3).

| D ISCUSS I ON
This study was one of the first large randomised, controlled trials to assess combination therapy for the treatment of IA. The results of the primary study showed that combination therapy with anidulafungin and voriconazole was associated with a numerically higher, but not statistically significant, reduction in overall mortality. 13 In the investigation, a relatively homogeneous population was enrolled, who were considered to have a lower risk for death because of their underlying conditions. The entry criteria were designed to measure treatment in patients with IA who had a low risk for death. 13 TA B L E 2 All-cause mortality at Weeks 6 and 12 in Korean and non-Korean patients (mITT population)  which can lead to increased trough voriconazole concentration, compared with Caucasians. [18][19][20][21][22]26 An antagonistic interaction has been observed between anidulafungin and voriconazole at higher exposures, suggesting that reduced efficacy may be observed in poor metabolizers receiving combination therapy compared with poor metabolisers receiving similar doses of monotherapy or non-poor metabolizers receiving combination therapy. 27 Koreans in our study had lower weight overall compared with non-Koreans. Although weight is generally thought to influence response to therapies, a pharmacokinetic-pharmacodynamic study of voriconazole in patients with IA did not determine weight as a predictor of efficacy. 28,29 As well as being used in the diagnosis of IA, a serum test for GM can help to predict clinical outcome in patients with probable or proven IA in patients with haematologic malignancies. 16  and hypokalaemia in Korean patients and pyrexia, diarrhoea and nausea in non-Korean patients. Temporary discontinuations or dose reductions were higher in the combination therapy group than in the monotherapy group and were higher in Koreans than non-Koreans in the combination therapy group. This could be due to a higher proportion of Korean than non-Korean patients with the poor metabolizer CYP2C19 genotype, resulting in higher trough voriconazole concentrations that may lead to increased AEs. 29 However, it is difficult to draw meaningful comparisons between the two groups due to the small number of Korean patients in the study. For the safety population, in Korean patients, all-cause mortality at Weeks 6 and 12 was lower in the combination therapy arm than in the monotherapy arm. In contrast, all-cause mortality at Week 6 was comparable in the two therapy arms in non-Korean patients.

Korean patients Non-Korean patients
The limitations of the primary study have been acknowledged. 13 The present study was a post hoc analysis and was not pre-planned in the primary study protocol; it therefore inherently has limitations related to both the post hoc analysis and the lower population. Numerical differences reported in the study may not represent statistical significance as overall numbers in the Korean sub-group were small, which also precluded multivariate analyses. Additionally, the post hoc data assessment was not blinded, although the original findings were obtained by independent blinded assessment.

ACK N OWLED G EM ENTS
The authors are grateful to the study investigators and patients.

R E S E A RCH DATA S H A R I N G
Upon request, and subject to certain criteria, conditions and exceptions (see https ://www.pfizer.com/scien ce/clini cal-trial s/trialdata-and-results for more information), Pfizer will provide access to individual de-identified participant data from Pfizer-sponsored global interventional clinical studies conducted for medicines, vaccines and medical devices (a) for indications that have been approved in the US and/or EU or (b) in programmes that have been terminated (ie development for all indications has been discontinued). Pfizer will also consider requests for the protocol, data dictionary and statistical analysis plan. Data may be requested from Pfizer trials 24 months after study completion. The de-identified participant data will be made available to researchers whose proposals meet the research criteria and other conditions, and for which an exception does not apply, via a secure portal. Data requestors must enter into a data access agreement with Pfizer to gain access.