FormalPara Key Points

A clear relationship between posaconazole exposure or voriconazole exposure and efficacy was not observed at the plasma exposures reached by the administered doses or drug formulations in the phase III randomized study.

The incidence of adverse events across plasma exposure quartiles was similar to those observed in the phase III ITT populations for both posaconazole and voriconazole.

1 Introduction

Patients who are immunocompromised —including those with prolonged neutropenia, those who underwent allogeneic hematopoietic stem cell transplantation or solid organ transplantation, those with inherited or acquired immunodeficiencies, and those who use corticosteroids—are at risk of invasive aspergillosis [1]. Guidelines recommend use of voriconazole or isavuconazole for the primary treatment of invasive aspergillosis [1,2,3].

Voriconazole is a triazole antifungal metabolized primarily by the hepatic cytochrome P450 enzymes CYP2C9, CYP2C19, and CYP3A4. CYP2C19 has genetic polymorphisms that cause large differences in the pharmacokinetics of voriconazole across patients, and CYP450-based therapies have known drug–drug interactions with voriconazole [4,5,6]. When administered orally, voriconazole reaches maximum concentration in < 2 h [6]. Steady state plasma concentrations are achieved after 5–7 days of treatment, but the time to achieve steady state can be reduced with a loading dose [6]. Voriconazole has non-linear pharmacokinetics in the therapeutic range, with maximal concentration and the area under the concentration–time curve value increasing disproportionately with dose increases, possibly because of the saturation of hepatic metabolism; the half-life is also dose dependent [6]. Because of concerns about voriconazole’s variable pharmacokinetics and potential drug–drug interactions, alternatives for the treatment of invasive aspergillosis are needed.

Posaconazole is a broad spectrum triazole approved as an injection, a delayed release tablet, and an oral suspension for the prevention of invasive fungal disease in selected patients who are immunocompromised [7, 8]. Posaconazole is also approved for salvage treatment of patients with invasive aspergillosis [8]. Once at steady state (by day 8), maximum concentration of posaconazole delayed release tablets is reached within 4 h of dosing, and bioavailability is superior to that of the posaconazole oral suspension (i.e., not impacted by food, gastric pH, or gastric motility) [9,10,11,12].

Adverse events (AEs) frequently reported for triazoles include nausea, vomiting, diarrhea, abdominal pain, and hepatotoxicity [13]. Voriconazole is also associated with specific drug-related AEs of visual disturbances, rashes, alopecia, periostitis, and QT prolongation; an exposure–safety relationship was observed between plasma trough concentrations (Ctrough) of voriconazole and the risk of developing hepatotoxicity [13, 14].

In a recent randomized double-blind phase 3 study of the primary treatment of invasive aspergillosis, posaconazole [intravenous (IV) or oral 300 mg twice on day 1, then 300 mg once a day thereafter] was non-inferior to voriconazole (6 mg/kg IV or 300 mg orally twice on day 1, followed by 4 mg/kg IV or 200 mg orally twice a day thereafter) for all-cause mortality at day 42 [primary endpoint; 15% of participants in the posaconazole group, 21% in the voriconazole group; treatment difference −5.3%, 95% confidence interval (CI) (−11.6, 1.0); p < 0.0001] (ClinicalTrials.gov, NCT01782131) [15]. The secondary endpoint of global clinical response in the full analysis set (FAS) population was similar for posaconazole and voriconazole at week 6 in 45% of participants who received posaconazole and 46% of participants who received voriconazole [treatment difference 0.6% (95% CI −11.2, 10.1)] [15]. The overall incidence of treatment-related adverse event (TRAE) rates in the intention-to-treat (ITT) population was 30% for posaconazole and 40% for voriconazole [treatment difference −10.2% 95% CI −17.9, −2.4)] [13]. This study used the approved formulations and dosages for both the IV and the tablet formulations of posaconazole and voriconazole. Considering the interpatient pharmacokinetic variability of both interventions, it is important to evaluate the relationship between clinical success and adequate levels of plasma exposure [6, 16]. Several previous studies used voriconazole and posaconazole plasma concentrations as surrogates for exposure to evaluate the relationship between plasma concentration and efficacy and safety [6, 16, 17]. The exposure response analyses for efficacy and safety from the current study is another key secondary objective that has not been reported.

The objectives of the present study were to summarize the pharmacokinetics and assess the exposure response for posaconazole and voriconazole from this phase 3 study, including exposure efficacy and exposure safety relationships.

2 Methods

2.1 Ethics

All participants or their legal representatives gave written informed consent before initiation of any study procedures. The study was conducted in accordance with the principles of good clinical practice. The protocol and all amendments were reviewed and approved by the institutional review boards or independent ethics committees at all study sites before being initiated at each site.

2.2 Study Design and Participants

All participants were enrolled in the international (26 countries) 91-site double-blind phase 3 study being conducted to compare posaconazole 300 mg IV or 300 mg orally once daily (twice daily on day 1) with voriconazole 4 mg/kg IV twice daily or 200 mg orally twice daily (day 1 6 mg/kg IV or 300 mg orally twice daily) for the treatment of invasive aspergillosis (ClinicalTrials.gov, NCT01782131; Merck & Co., Inc., Rahway, NJ, USA protocol number MK-5592-069) [15]. Most participants were given initial therapy via the IV route; however, some were given initial therapy via the oral route. Azole therapy was switched from the IV route to the oral route when the participant was considered clinically stable and able to take oral medication.

The study design and inclusion and exclusion criteria are described in detail elsewhere [15]. Briefly, participants aged ≥ 13 years were eligible if they had features consistent with proven, probable, or possible invasive aspergillosis, as defined by Mycoses Study Group and European Organization for Research and Treatment of Cancer consensus criteria [18], with a likelihood to upgrade the diagnosis to proven or probable infection within the first 7 days of treatment. A subsequent protocol modification allowed for the inclusion of individuals with neutropenia of any duration as an acceptable host factor [15].

Randomly assigned participants who received at least one dose of study drug were included in the ITT population for evaluation of safety and the primary efficacy endpoint of all-cause mortality at day 42. The secondary objective of global clinical response of success at weeks 6 and 12 was evaluated in the FAS population, which included all participants in the ITT population with proven or probable invasive aspergillosis per independent adjudication. The pharmacokinetic-evaluable population included all ITT participants for whom at least one plasma concentration was available during the treatment period, and the pharmacokinetic evaluable Ctrough population included all participants with available plasma concentration data who met trough sample timing.

2.3 Pharmacokinetics Sampling

Steady state Ctrough samples were collected before each dose on day 7, week 2, week 4, week 6, and week 12 [end of treatment (EOT)]. If a participant discontinued early, a trough level sample (before dose, if possible) was collected at the time of study discontinuation, and the time of sampling was noted.

Plasma posaconazole concentrations were determined by PPD, Inc. (Richmond, VA, USA) using a validated high-performance liquid chromatography–tandem mass spectrometry (HPLC–MS/MS) method. The analytical range for posaconazole was 5–5000 ng/mL in human plasma [19]. Plasma voriconazole concentrations were determined by Syneos Health Clinique, Inc. (Quebec City, PQ, Canada) using a validated HPLC–MS/MS method [20]. The analytical range for voriconazole was 5–2000 ng/mL in human plasma.

Plasma Ctrough levels were defined as those for which the time since last dose was 16–32 h for posaconazole (administered once daily) and 8–16 h for voriconazole (administered twice daily) or for which the plasma sample of either drug was taken at the same time as the next dose. Summary statistics for the pooled posaconazole or voriconazole Ctrough sample for each study visit were determined for three groups of participants, defined as those who received (a) IV administration only, (b) oral administration only, or (c) both IV and oral administration because of switching at least once between these. Treatment blinding was maintained without therapeutic drug monitoring.

2.4 Exposure Response Evaluations

Exposure response relationships for posaconazole and voriconazole were evaluated using summary statistics and by plotting data on a graph for efficacy and safety endpoints. For the efficacy endpoints (all-cause mortality through day 42 and global clinical response at week 6), exposure was characterized by the within-participant arithmetic mean of plasma Ctrough obtained through week 6 (when the endpoints were assessed). For each efficacy endpoint, participants were stratified into quartiles on the basis of the mean plasma Ctrough, and results were determined for each quartile. For the safety endpoints (TRAEs and specific categories of tier 1 AEs) (see Sect. 2.5), exposure was characterized by the within-participant arithmetic mean of plasma Ctrough obtained through week 12. In addition, the exposure–safety relationship for all AEs was evaluated for posaconazole using summary statistics and assessing number (percentage) of participants in each of the exposure quartiles with a particular AE or with any AE undergoing specific standard of care. For each safety endpoint, participants were stratified into quartiles on the basis of the mean plasma Ctrough, and results were determined for each quartile.

The endpoint responses were also determined in the group of participants for whom no plasma Ctrough was available. For the week 12 timepoint, plasma concentrations for participants who completed the week 12 visit were aggregated with those of participants who discontinued study therapy early and had an early EOT visit.

2.5 Safety

Detailed safety was previously reported [15]. Briefly, safety included the incidence of treatment emergent adverse events (TEAEs) and TRAEs and the incidence of serious AEs, summarized by treatment group. Safety analyses followed a tiered approach. Tier 1 events included hepatic laboratory changes (defined as elevated aspartate aminotransferase or alanine aminotransferase concentrations ≥ 3 times the upper limit of normal, elevated total bilirubin concentration ≥ 2 times the upper normal limit, and alkaline phosphatase concentration < 2 times the upper normal limit), central nervous system (CNS) and visual safety AEs (TEAEs related to visual and CNS disturbances), dermatological AEs (TEAEs, including rash and photosensitivity rash), and adrenal steroidogenesis and vascular AEs (TEAEs indicating adrenal insufficiency or temporally associated hypotension).

3 Results

3.1 Pharmacokinetic Analysis Population

The overall pharmacokinetic analysis population consisted of 506 of 575 participants (88%) in the ITT population who had at least one reported posaconazole or voriconazole plasma concentration at any time during the treatment period. Sixty-nine of 575 participants (24 treated with posaconazole and 45 with voriconazole) in the ITT population had no reportable plasma concentrations. A total of 437 of 506 participants (228 who received posaconazole and 209 who received voriconazole) in the overall pharmacokinetic analysis population had at least one plasma Ctrough. The remaining 69 of 506 participants (36 in the posaconazole arm and 33 in the voriconazole arm) had reported plasma concentrations that did not meet the timing criteria to be considered a trough concentration; of the excluded samples, > 80% could not be considered trough samples because they were drawn too soon after the next dose had been administered. The pharmacokinetic and exposure response populations are summarized in Supplementary Table S1.

3.2 Steady State Pharmacokinetic Summary

Pooled mean posaconazole plasma Ctrough approached steady state by the end of week 1. Geometric mean posaconazole Ctrough was approximately 1500 ng/mL through week 12 with high variability (percentage geometric mean coefficient of variation, approximately 70%–90%) (Table 1). Mean posaconazole plasma Ctrough for participants who started and remained on posaconazole IV were 3%–53% higher than for participants who either started and remained on posaconazole tablet or transitioned between posaconazole IV and posaconazole tablet before week 12 or EOT (Fig. 1). For the week 12 timepoint, posaconazole concentrations for participants who completed the week 12 visit were aggregated with those from participants who discontinued study therapy early and had an EOT visit before week 12; therefore, these results should be interpreted with caution.

Table 1 Summary statistics for trough plasma concentrations pooled across participants by drug
Full size table
Fig. 1
figure 1

Mean [standard deviation (SD)] posaconazole Ctrough for participants receiving only posaconazole tablets (grey circle, grey solid line), participants receiving posaconazole IV only (black circle, dotted line), and participants starting on posaconazole IV and switching between IV and tablet (open square, solid line). Week 12 includes EOT concentration data. Ctrough trough concentration, EOT end of treatment, IV intravenously

Full size image

Pooled mean voriconazole plasma Ctrough was approximately 4200 ng/mL at week 1, then decreased approximately 50% over the first 4 weeks of dosing (Table 1). Mean voriconazole plasma Ctrough for participants who started on voriconazole IV was 2.5-fold higher for the first 2 weeks of treatment than voriconazole plasma Ctrough for participants who started on oral voriconazole (Fig. 2). Voriconazole plasma concentrations were highly variable, with approximately 25% of participants having Ctrough > 4000 ng/mL. Similar to posaconazole concentrations, the voriconazole concentrations for the week 12 visit and the EOT visit were aggregated and should be interpreted with caution.

Fig. 2
figure 2

Mean (SD) voriconazole Ctrough in participants receiving only voriconazole capsules (grey circle, grey solid line), participants receiving only voriconazole IV (black circle, dotted line), and participants starting on voriconazole IV and switching between IV and capsule (open square, solid line). Week 12 includes EOT concentration data. Ctrough trough concentration, EOT end of treatment, IV intravenous

Full size image

3.3 Exposure Efficacy

In addition to exploring efficacy relationships by quartile based on Ctrough (Table 2), efficacy endpoint data were summarized for participants with no available Ctrough data. For participants who received posaconazole and had evaluable Ctrough exposure data, there was no discernible trend across quartiles of posaconazole plasma Ctrough for the key efficacy endpoint of all-cause mortality through day 42 in the ITT population. Participants without evaluable posaconazole exposure data had higher mortality rates (Fig. 3a). Similarly, for clinical response at week 6, there was no clear trend across quartiles of posaconazole plasma Ctrough in participants with evaluable exposure data, whereas rates of clinical failure were two-fold higher in participants who did have evaluable exposure data (Fig. 3b).

Table 2 Ctrough quartile ranges for posaconazole and voriconazole (ITT population)a
Full size table
Fig. 3
figure 3

Posaconazole exposure data. a All-cause mortality through day 42 by quartiles of within-participant mean posaconazole plasma Ctrough (ITT population). b Global clinical response at week 6 by quartile of within-participant mean posaconazole plasma Ctrough (FAS population). Ctrough trough concentration, FAS full analysis set, ITT intention-to-treat, Q quartile

Full size image

All-cause mortality through day 42 was higher in participants in the highest quartile (Table 2) of voriconazole plasma Ctrough than in participants in quartiles 1–3 of voriconazole plasma Ctrough (Fig. 4a). The number of participants in the non-evaluable exposure group (approximately 28% of all voriconazole-treated participants) was higher than the number of participants in any of the exposure quartiles. As with posaconazole plasma Ctrough, efficacy outcomes were poorer in participants who did not have voriconazole plasma Ctrough pharmacokinetic data. A similar trend was observed for success versus failure in global clinical response at week 6 in the FAS population (Fig. 4b).

Fig. 4
figure 4

Voriconazole exposure data. a All-cause mortality through day 42 by quartiles of within-participant mean voriconazole plasma Ctrough (ITT population). b Global clinical response at week 6 by quartile of within-participant mean voriconazole plasma Ctrough (FAS population). Ctrough trough concentration, FAS full analysis set, ITT intention-to-treat, Q quartile

Full size image

3.4 Exposure Safety

Exposure–safety relationships of posaconazole and voriconazole by quartile of exposure were evaluated for TRAEs (Fig. 5). For posaconazole, although the proportion of participants with TRAEs was highest in the maximum quartile of exposure (Fig. 5a), no clear evidence of an exposure-related pattern was observed for most of the commonly reported AEs (those reported by ≥ 10% of participants) or system organ classes (Supplementary Table S2). The distributions of posaconazole Ctrough were similar for participants with and without tier 1 CNS and visual safety AEs.

Fig. 5
figure 5

Proportion of participants with TRAEs by quartile of within-participant mean a posaconazole Ctrough and b voriconazole Ctrough. Ctrough trough concentration, Q quartile, TRAE treatment-related adverse event

Full size image

The incidence of TRAEs was similar across all quartiles of voriconazole mean Ctrough (Fig. 5b). The distribution of voriconazole Ctrough was similar for participants with and without tier 1 CNS and visual safety AEs and with and without tier 1 dermatological reaction AEs (data not shown).

4 Discussion

This was an exposure response analysis from the phase 3 double-blind, double-dummy study that previously showed that posaconazole was non-inferior to voriconazole with regard to the primary endpoint of all-cause mortality up to day 42 in the ITT population, and the number of TRAEs with posaconazole was 10% lower than with voriconazole [15]. In the exposure response analyses presented here, a clear relationship was not observed between posaconazole exposure with efficacy or safety at the plasma Ctrough reached by the administered doses or drug formulations. For voriconazole, there was also no notable relationship between exposure and efficacy outcomes for quartiles 1–3, whereas participants in the highest Ctrough quartile of voriconazole had higher all-cause mortality through day 42 than participants in the lower three quartiles. There was no clear relationship between voriconazole exposure and safety.

Posaconazole mean plasma Ctrough was within the concentration ranges observed previously for posaconazole IV (mean Ctrough 1320 ng/mL) and tablet (mean Ctrough 1720 ng/mL) formulations [9, 21]. The overall distribution of voriconazole mean plasma Ctrough was higher than in prior clinical studies, with about 25% of participants having Ctrough more than ~ 4000 ng/mL in the highest quartile [22].

The efficacy outcomes of all-cause mortality and global clinical response for the individual Ctrough exposure quartiles for posaconazole and voriconazole were mostly similar to the efficacy outcomes observed for each ITT population [15]. Seeing no discernible relationship between Ctrough exposure quartiles and efficacy outcomes for posaconazole suggests that posaconazole exposures generated from the administered doses and formulations used in the primary phase 3 study were on the plateau of the exposure–efficacy curve, where efficacy is relatively insensitive to exposure (e.g., the lowest quartile of exposure was at or above the concentration threshold needed for efficacy).

For voriconazole, higher all-cause mortality and global clinical response failures were observed in participants in the highest quartile of voriconazole plasma Ctrough, which may be attributable to more seriously ill participants needing to be on IV voriconazole and as a result having higher voriconazole concentrations. Additionally, genotypic and phenotypic variability of hepatic enzymes that metabolize voriconazole can lead to higher than normal voriconazole plasma concentrations and toxicity [5]. An exploratory pharmacogenetic analysis indicated a low likelihood that the CYP2C19 poor metabolizer phenotype contributed to variability in the voriconazole efficacy or safety findings, but the sample size was relatively small [15].

The incidence of TRAEs across Ctrough exposure quartiles was mostly similar to observed TRAEs from the ITT populations for both posaconazole and voriconazole [15]. For the posaconazole exposure safety analysis, the relationship of posaconazole to safety by quartile of exposure was evaluated for all AEs and for drug-related AEs. A higher incidence of TRAEs was observed in the highest exposure quartile, but there were no observed relationships between exposures and the incidence of individual-reported AEs (regardless of investigator-reported relationship), including tier 1 AEs. Similarly, no exposure–safety relationships were observed in prior studies of posaconazole in the IV or tablet formulation, in which similar or higher exposures were achieved [9, 21]. An exposure–safety relationship was also not observed with posaconazole oral suspension, although exposures with the oral suspension were generally lower than those achieved with the IV or tablet formulations [7]. A potential contributing factor to the observed association in the current study is that a larger proportion of participants who were more seriously ill were likely to have been receiving the IV formulation rather than the tablet formulation and therefore would likely have had higher exposures. Furthermore, the severity of illness would have led to a higher incidence of AEs.

One limitation to this analysis is that plasma Ctrough data for the exposure efficacy analyses were not available for 21%–25% of participants, and there was a marked difference in mortality rate and clinical response rate for participants with and without exposure data. Efficacy was lower in participants who did not have evaluable exposure data for posaconazole and voriconazole, suggesting that participants who died or who remained seriously ill were more likely not to have evaluable exposure data, potentially introducing bias in the exposure efficacy analysis. Additionally, the enrollment of adolescents was limited to five participants; therefore, generalizing study conclusions to the adolescent population should be done with caution. A separate clinical study is ongoing to evaluate posaconazole for the primary treatment of invasive aspergillosis in pediatric patients.

5 Conclusion

No significant associations were observed between efficacy or safety and the plasma levels of voriconazole or posaconazole reached from the administered doses and formulations evaluated in this phase 3 trial for the primary treatment of invasive aspergillosis.