Pharmacokinetics and Safety of Voriconazole Intravenous-to-Oral Switch Regimens in Immunocompromised Japanese Pediatric Patients

Abstract

The aim of this study was to investigate the pharmacokinetics, safety, and tolerability of voriconazole following intravenous-to-oral switch regimens used with immunocompromised Japanese pediatric subjects (age 2 to <15 years) at high risk for systemic fungal infection. Twenty-one patients received intravenous-to-oral switch regimens based on a recent population pharmacokinetic modeling; they were given 9 mg/kg of body weight followed by 8 mg/kg of intravenous (i.v.) voriconazole every 12 h (q12h), and 9 mg/kg (maximum, 350 mg) of oral voriconazole q12h (for patients age 2 to <12 or 12 to <15 years and <50 kg) or 6 mg/kg followed by 4 mg/kg of i.v. voriconazole q12h and 200 mg of oral voriconazole q12h (for patients age 12 to <15 years and ≥50 kg). The steady-state area under the curve over the 12-h dosing interval (AUC_0–12,ss) was calculated using the noncompartmental method and compared with the predicted exposures in Western pediatric subjects based on the abovementioned modeling. The geometric mean (coefficient of variation) AUC_0–12,ss values for the intravenous and oral regimens were 51.1 μg · h/ml (68%) and 45.8 μg · h/ml (90%), respectively; there was a high correlation between AUC_0–12,ss and trough concentration. Although the average exposures were higher in the Japanese patients than those in the Western pediatric subjects, the overall voriconazole exposures were comparable between these two groups due to large interindividual variability. The exposures in the 2 cytochrome P450 2C19 poor metabolizers were among the highest. Voriconazole was well tolerated. The most common treatment-related adverse events were photophobia and abnormal hepatic function. These recommended doses derived from the modeling appear to be appropriate for Japanese pediatric patients, showing no additional safety risks compared to those with adult patients. (This study has been registered at ClinicalTrials.gov under registration no. NCT01383993.)

INTRODUCTION

Voriconazole is a triazole antifungal drug that has been recommended in guidelines for the treatment of invasive fungal infections in adults in Japan (1) and the rest of the world (2–5). A 2009 nationwide survey regarding treatment for pediatric patients with invasive fungal infections in Japan found that antifungal agents not approved for pediatric use, including voriconazole, were frequently being used in clinical practice at a wide range of doses (6).

Voriconazole is metabolized by the human hepatic cytochrome P450 enzymes, primarily CYP2C19 but also CYP3A4 and, to a lesser extent, CYP2C9, to its main circulating N-oxide metabolite (UK-121,265) (7, 8), with a significantly faster clearance of voriconazole in children than that in adults (9, 10). The CYP2C19 enzyme exhibits genetic polymorphisms, and the CYP2C19 genotype can be a key factor in the pharmacokinetics of voriconazole in individual patients. A marked interethnic difference in the frequency of CYP2C19 poor metabolizers (PM) has been noted (11). The frequencies of CYP2C19 PM have been estimated at 2.2% of Caucasian compared with 15.8% of Asian populations, with a higher frequency in Japanese than in Chinese subjects (21.3% versus 13.7%, respectively) (12).

The pharmacokinetics of voriconazole in immunocompromised children was investigated in a number of studies using different intravenous (3, 4, 5, 6, 7, and 8 mg/kg of body weight every 12 h [q12h]) and oral (4 and 6 mg/kg q12h and 200 mg q12h) maintenance doses (10, 13–15). These studies in largely Caucasian populations demonstrated that a higher intravenous dose was required in children to match the adult exposures and that children had higher intra- and interpatient variabilities than did adults in the pharmacokinetics of voriconazole (10, 13–15). Whereas intravenous voriconazole (at 3 and 4 mg/kg q12h) demonstrated nonlinear pharmacokinetics in adults, children receiving the same dosages demonstrated pseudolinear pharmacokinetics (10), with nonlinear pharmacokinetics at a higher dose of 7 mg/kg q12h, possibly due to saturation of its metabolism, as in adults (14). Body weight was found to be more influential than age in accounting for the observed differences in the voriconazole pharmacokinetics between children and adults (10, 15), as young adolescents with low body weight required higher doses to match the adult exposures (16). In order to further optimize dosing with voriconazole in the pediatric population, an integrated-population pharmacokinetic analysis was conducted using pooled data from children, adolescents, and adults from 5 studies, which derived the newly recommended doses for pediatric patients to match the adult exposures at approved doses (i.e., 6 mg/kg followed by 4 mg/kg of intravenous [i.v.] voriconazole q12h and 200 mg of oral [p.o.] voriconazole q12h) (17).

This study (registered at ClinicalTrials.gov under registration no. NCT01383993) aimed to evaluate the pharmacokinetics, safety, and tolerability of voriconazole in Japanese pediatric patients following these recommended intravenous-to-oral switch regimens (17). This study aimed to confirm whether the proposed doses based on population pharmacokinetic modeling were appropriate for pediatric patients. In addition, no robust pharmacokinetic data for voriconazole in Japanese pediatric patients were available. This study aimed to provide data evaluating whether there are racial differences in exposure to voriconazole between Japanese and Western pediatric patients.

MATERIALS AND METHODS

Study design.

This was a phase 2 open-label multiple-dose pharmacokinetic study conducted with immunocompromised pediatric subjects (age 2 to <15 years) at high risk for systemic fungal infection. The study was conducted at 6 centers in Japan. The study was approved by the local institutional review board (IRB)/independent ethics committee at each center or by a central IRB (1 center) and conducted in accordance with the ethics principles originating in or derived from the Declaration of Helsinki and in compliance with all International Conference on Harmonisation Good Clinical Practice guidelines. Written informed consent was obtained from all participating patients or their legal representatives prior to their enrollment in the study.

In this study, the patient population and study design were similar to those in a previous pediatric pharmacokinetic study conducted in the United States (13). The study consisted of a screening observation period, a pharmacokinetic period, an optional postpharmacokinetic period, and a follow-up period. The pharmacokinetic period extended from the start of intravenous treatment (7-day treatment) through the 7th day of oral treatment (6.5-day treatment; only the morning dose was taken on day 14). If the patient was unable to take oral medication on day 8, intravenous treatment could be continued for a maximum of 20 days before switching to the oral treatment. The optional postpharmacokinetic period was based on clinical need (persistent neutropenia [defined as an absolute neutrophil count, <500 cells/μl] and continued high risk for the development of systemic fungal infections) and was to continue for up to 30 days following the pharmacokinetic period. Treatment beyond 30 days was allowed after consultation with a Pfizer clinician. The follow-up period was 30 (±7) days after the last dose of the study drug.

The primary objective of this study was to characterize the pharmacokinetics, safety, and tolerability of voriconazole following recommended intravenous-to-oral switch regimens based on earlier models (17) in immunocompromised pediatric patients (age 2 to <15 years) who were at high risk for systemic fungal infection. The secondary objective was to explore the effect of potential covariates (e.g., CYP2C19 genotype status, age, and body weight) on voriconazole exposure.

Patients.

Pediatric patients (age 2 to <15 years) requiring treatment for the prevention of systemic fungal infection who were expected to develop neutropenia (defined as absolute neutrophil count, <500 cells/μl) lasting >10 days and could tolerate being switched to oral therapy after 7 to 20 days of intravenous voriconazole were eligible for the study. The exclusion criteria included serum aspartate aminotransferase (AST) or alanine aminotransferase (ALT) levels of >5× upper limit of normal (ULN), or total bilirubin of >2.5× the ULN; history of hypersensitivity or intolerance to azole antifungal agents; history/presence of arrhythmia (e.g., caused by past cardiotoxic chemotherapies, related to QT prolongation); treatment with inhibitors of CYP450 enzymes that affect voriconazole exposure (e.g., fluconazole and itraconazole) ≥24 h prior to the study start; or treatment with inducers of CYP450 enzymes (e.g., rifampin, carbamazepine, and long-acting barbiturates) within 14 days prior to the study start. Subjects were excluded if they would need to receive any other drugs prohibited in the voriconazole product label (e.g., sirolimus and ergot alkaloids) during the study. Subjects were also excluded if they had any other condition that, in the opinion of the investigator, made the subject unsuitable for enrollment.

Voriconazole administration.

Voriconazole (Pfizer, Inc.) was administered according to dosing regimens derived from the recent population pharmacokinetic modeling (17), as shown in Table 1. The intravenous formulation was prepared at a final concentration of 0.5 to 5 mg/ml and infused at rate of 3 mg/kg/h (i.e., the doses of 9, 8, 6, and 4 mg/kg were administered over 180, 160, 120, and 80 min, respectively). Voriconazole powder for oral suspension was reconstituted at a concentration of 40 mg/ml and administered ≥1 h before or after a meal.

TABLE 1.

Voriconazole dosing schedule

Patient age group (yr [wt]):	Loading i.v. dose (mg/kg) q12h on day 1a	Maintenance dose (mg/kg) administered q12h by route, days:
		i.v., 2–7	Oral, 8–14b
2 to <12	9	8	9 (up to maximum, 350 mg)
12 to <15 (<50 kg)
12 to <15 (≥50 kg)	6	4	200 mg

^ai.v., intravenous; q12h, every 12 h.

^bThe oral regimen was administered ≥1 h before or 1 h following a meal. The allowance time for dosing at 12-h intervals was ±1 h. If patients were unable to switch to the oral regimen on the 1st day of oral treatment (day 8), the intravenous regimen could be continued for up to 20 days (day 20) before switching to the oral regimen. On the 7th day of oral treatment (day 14), only the morning oral dose was given. If clinically indicated, voriconazole could be continued for up to 30 days (day 30), including the pharmacokinetic period. Treatment after 30 days was allowed in consultation with a Pfizer clinician.

Pharmacokinetic evaluation.

The plasma voriconazole concentrations were measured on the 2nd day of intravenous treatment (day 2) just prior to the start of the morning dose, on the 7th day of intravenous treatment (day 7) just prior to the start of the morning dose, 1 h after the start of infusion, 2 to 20 min after the end of infusion, and 4, 6, 8, and 12 h after the start of infusion, and on the 7th day of oral treatment (day 14) just prior to the start of the morning dose and 1, 2, 4, 6, 8, and 12 h after dosing. If patients could not be scheduled for serial sampling on the 7th day of intravenous or oral dosing, they could remain on either treatment to allow sample collection on a later day.

Plasma voriconazole and the N-oxide metabolite concentrations were analyzed using a previously validated liquid chromatography coupled to tandem mass spectrometry (18) at Pharmaceutical Product Development (Richmond, VA, USA). The lower limits of quantification for voriconazole and the N-oxide metabolite were 10 and 20 ng/ml, respectively. The accuracies of the quality control samples for voriconazole and the N-oxide metabolite ranged from −2.25% to 1.53% and from −1.35% to 1.39%, with precision rates of ≤5.57% and ≤8.74%, respectively.

A noncompartmental pharmacokinetic analysis for voriconazole and the N-oxide metabolite was performed using the internally validated system eNCA version 2.2.1. The steady-state peak plasma concentration (C_max,ss), the time to reach the C_max (T_max), and steady-state trough concentration (C_min,ss) were estimated directly from the concentration-time data. The steady-state area under the curve over a 12-h dosing interval (AUC_0–12,ss) was estimated using the linear/log trapezoidal approximation. Samples below the lower limit of quantification were set to 0 ng/ml for analysis, and the actual sample times were used for the pharmacokinetic analysis. The ratios of voriconazole AUC_0–12,ss following oral treatment relative to intravenous treatment and steady-state pharmacokinetic parameters (AUC_0–12,ss, C_max,ss, and T_max) of the major N-oxide metabolite were also estimated.

CYP2C19 genotyping.

Two buccal swab samples (one from each cheek) were collected during the observation period (between day −30 and day −4, as early as possible after informed consent was obtained) and analyzed for CYP2C19 genotyping at Pfizer Pharmacogenomics Laboratory (Groton, CT, USA). DNA was isolated from the buccal swabs using the QIAcube (Qiagen N.V., Venlo, the Netherlands). Five single nucleotide polymorphisms for CYP2C19 (*2, *3, *4, *5, and *17) were evaluated using the TaqMan allelic discrimination procedure. An extensive metabolizer (EM) was defined as having the *1/*1 and *1/*17 genotypes, a heterozygous extensive metabolizer (HEM) was defined as having the *1/*2 and *1/*3 genotypes, and a PM was defined as having *2/*2, *2/*3, and *3/*3 genotypes.

Safety evaluations.

Safety laboratory tests and visual assessments (near-distance visual acuity test, color vision test, and assessment of visual symptoms in patients >5 years or as developmentally appropriate) were performed at screening, on the 7th day of intravenous treatment (day 7), on the 1st day of oral treatment (day 8, visual assessments only), on the 7th day of oral treatment (day 14), and at the 30-day follow-up visit. Vital signs and 12-lead electrocardiograms (singlet) were measured at screening, on the 1st day of intravenous treatment (day 1) just prior to the morning dose, on the 7th day of intravenous treatment (day 7) 15 min before the end of the morning dose, on the 1st and 7th days of oral treatment (days 8 and 14, respectively) at 2 h after the morning dose, and at the 30-day follow-up visit. When voriconazole treatment was prolonged (i.e., >14 days), these assessments were conducted every 7 days. Adverse events (AEs) were closely monitored throughout the study (any breakthrough of fungal infections would be reported as an AE).

Statistical analysis.

A target sample size of 20 patients (who received ≥1 dose of study treatment) was chosen for the safety analysis based on the number of patients expected to be able to enroll within 1 year; assuming a 20% dropout rate, this would ensure that ≥16 patients (who had at least one of the pharmacokinetic parameters of interest) were included in the pharmacokinetic analysis. All the data were summarized with descriptive statistics, and no formal inferential statistics were applied.

The pharmacokinetic parameters for voriconazole were evaluated by age, body weight, and CYP2C19 genotyping status. The patients were divided into 3 groups based on age and dosing recommendations: 2 to <12 years, 12 to <15 years weighing <50 kg, and 12 to <15 years weighing ≥50 kg. The correlation between AUC_0–12,ss and C_min,ss was also evaluated using a simple linear regression model (with the intercept) to calculate R², where R is the Pearson product-moment correlation coefficient. Given the large interindividual variability in voriconazole exposure and limited sample size, an informal statistical comparison of voriconazole total exposures (AUC_0–12,ss) in Japanese pediatric patients with the predicted exposures in Western pediatric subjects was performed. If there was a substantial overlap between the distributions of AUC_0–12,ss in the Japanese and Western pediatric subjects, the exposures were considered comparable.

RESULTS

Study population.

A total of 21 patients were enrolled and received study treatment between September 2011 and May 2013. Sixteen patients completed the study and five discontinued the study (two before switching to the oral treatment). The reasons for discontinuation included treatment-related AEs (2 patients), being no longer willing to participate (1 patient), and other reasons (2 patients required prophylactic treatment with commercial voriconazole during the follow-up, as they still had a high risk for fungal infection).

All 21 patients were analyzed for safety, and 20 patients were analyzed for pharmacokinetic parameters. One patient (11 years old, CYP2C19 EM) was excluded from the pharmacokinetic analysis because this patient had discontinued treatment on day 7 after having blood sampled at 1 time point only.

The baseline patient demographics and characteristics are summarized in Table 2. Of the 21 patients, 9 (42.9%) were male and 12 (57.1%) were female. Fifteen patients were between 2 and <12 years of age, four were between 12 and <15 years weighing <50 kg, and two were between 12 and <15 years weighing ≥50 kg. The underlying conditions that caused the immunocompromised status were acute lymphocytic leukemia (8 patients), acute myeloid leukemia (7 patients), Ewing's sarcoma (2 patients), and neuroblastoma, osteosarcoma, bone marrow transplant, and stem cell transplant (1 patient each). The CYP2C19 genotype status was categorized as EM in 9 patients, HEM in 10 patients, and PM in 2 patients.

TABLE 2.

Patient demographics and characteristics

Characteristica	All (n = 21)	2 to <12 yr (n = 15)	12 to <15 yr, wt <50 kg (n = 4)	12 to <15 yr, wt ≥50 kg (n = 2)
Sex (no. [%] male)	9 (42.9)	6 (40.0)	2 (50.0)	1 (50.0)
Age (no. [%]) (yr)
2 to 6	5 (23.8)	5 (33.3)	0	0
7 to 11	10 (47.6)	10 (66.7)	0	0
12 to <15	6 (28.6)	0	4 (100)	2 (100)
Mean (range)	9.2 (3–14)	7.7 (3–11)	12.5 (12–13)	14.0 (14–14)
Wt (kg)
Mean (range)	30.4 (11.5–55.2)	25.3 (11.5–43.0)	38.0 (31.5–44.7)	53.5 (51.8–55.2)
No. with CYP2C19 genotype:
EM	9	5	2	2
HEM	10	9	1	0
PM	2	1	1	0
No. with underlying condition:
ALL	8	4	2	2
AML	7	5	2	0
Ewing's sarcoma	2	2	0	0
Neuroblastoma	1	1	0	0
Osteosarcoma	1	1	0	0
BMT	1	1	0	0
SCT	1	1	0	0

^aEM, extensive metabolizer; HEM, heterozygous extensive metabolizer; PM, poor metabolizer; ALL, acute lymphocytic leukemia; AML, acute myeloid leukemia; BMT, bone marrow transplant; SCT, stem cell transplant.

The most common concomitant medications taken before and during the study were antibacterial agents, immunosuppressants, electrolyte replacements, and blood products. Antihistamines, corticosteroids, and antiserotonin agents were also common. Except for 1 patient who received omeprazole only for the first 2 days, the concomitant medications taken by the study patients had no known potential to affect voriconazole exposure. The concomitant medications that had the potential to be affected by voriconazole were well managed, in accordance with the clinical trial protocol. No AEs caused by drug-drug interactions were reported.

Pharmacokinetics.

The voriconazole pharmacokinetic parameters are summarized for all patients and by age-weight group in Table 3. The C_max,ss was generally observed at the end of intravenous administration (about 3 h) and about 1 h postdose following oral administration. For all patients, the geometric mean AUC_0–12,ss values following intravenous and oral administration were 51.1 and 45.8 μg · h/ml, respectively. The young adolescents with lower body weight (12 to <15 years weighing <50 kg, n = 4) had higher average exposure than that of the children (2 to <12 years, n = 14) with the same dosing regimens, especially following intravenous administration (e.g., AUC_0–12,ss, 83.4 μg · h/ml versus 51.9 μg · h/ml, respectively). These high average values reflect 1 of the 4 adolescents who had the highest exposure for intravenous dosing and was a CYP2C19 PM; the values for the 3 other adolescents were comparable to those for the children. Only 2 young adolescents weighing ≥50 kg (with CYP2C19 EM status) received lower voriconazole doses, and both had a lower AUC_0–12,ss for intravenous and oral dosing than did the other groups. Nonetheless, these low exposure values were still within the range observed in healthy adults receiving the same dosing regimens (e.g., median [range] of AUC_0–12,ss at 4 mg/kg of i.v. drug, 37.6 [13.7 to 104] μg · h/ml, and at 200 mg of p.o. drug, 12.8 [4.89 to 61.6] μg · h/ml) (13). The individual voriconazole AUC_0–12,ss values by age-weight and administration route are presented in Fig. 1 and demonstrate large interindividual variability, with percent coefficients of variation (%CV) of 68% and 90% following intravenous and oral administration, respectively. The results for voriconazole C_max,ss and C_min,ss were consistent with those observed for AUC_0–12,ss. In addition, there was a high correlation between the AUC_0–12,ss and C_min,ss (Fig. 2A, R² = 0.9474; Fig. 2B, R² = 0.9496).

TABLE 3.

Summary of plasma voriconazole pharmacokinetic parameters

Parameter by administration routea	Result by patient group (age, wt)
	All	2 to <12 yr	12 to <15 yr, <50 kg	12 to <15 yr, ≥50 kg
Intravenous (n)	20	14	4	2
AUC0–12,ss (μg · h/ml)
Geometric mean (%CV)	51.1 (68)	51.9 (51)	83.4 (56)	17.3 (28)
Median (range)	59.3 (14.2–177)	60.2 (23.0–103)	70.5 (55.7–177)	17.6 (14.2–21.0)
Cmax,ss (μg/ml)
Geometric mean (%CV)	7.32 (51)	7.75 (38)	9.23 (55)	3.09 (42)
Median (range)	7.72 (2.32–19.6)	8.21 (4.62–12.6)	7.72 (6.24–19.6)	3.22 (2.32–4.12)
Tmax (h)
Median (range)	2.96 (0.950–4.20)	2.96 (0.950–4.00)	4.00 (2.92–4.20)	1.34 (1.00–1.67)
Cmin,ss (μg/ml)b
Geometric mean (%CV)	2.49 (112)	2.53 (96)	4.88 (60)	0.566 (26)
Median (range)	3.00 (0.471–10.4)	2.89 (0.596–9.36)	4.31 (3.09–10.4)	0.576 (0.471–0.680)
Oral (n)	18	14	3	1c
AUC0–12,ss (μg · h/ml)
Geometric mean (%CV)	45.8 (90)	48.2 (83)	59.4 (67)	NAd
Median (range)	45.6 (10.0–156)	45.6 (12.4–156)	49.4 (36.3–117)	10.0
Cmax,ss (μg/ml)
Geometric mean (%CV)	7.22 (59)	7.76 (50)	7.91 (45)	NA
Median (range)	6.48 (2.03–18.3)	6.70 (3.58–18.3)	6.21 (6.13–13.0)	2.03
Tmax (h)
Median (range)	1.04 (0.917–3.78)	1.09 (0.917–3.78)	1.00 (0.950–2.03)	1.00
Cmin,ss (μg/ml)
Geometric mean (%CV)	1.80 (173)	1.87 (175)	2.78 (112)	NA
Median (range)	2.06 (0.148–12.3)	2.06 (0.148–12.3)	3.00 (1.09–6.59)	0.306

^aAUC_0–12,ss, area under the curve over dosing interval at steady state; %CV, % coefficient of variation; C_max,ss, peak plasma concentration at steady state; T_max, time to reach C_max; C_min,ss, trough plasma concentration at steady state.

^bn = 21 (all), n = 15 (2 to <12 years), n = 4 (12 to <15 years, weight <50 kg), and n = 2 (12 to <15 years, weight ≥50 kg).

^cSummary statistics were not calculated if n = 1 (i.e., values are individual values).

^dNA, not applicable.

FIG 1.

Individual voriconazole AUC_0–12,ss for intravenous (A) and oral (B) administration by age-weight. AUC_0–12,ss, area under the curve over the 12-h dosing interval at steady state.

FIG 2.

Plot of AUC_0–12,ss against C_min,ss for intravenous (A) and oral (B) administration by CYP2C19 genotype. Treatment regimen 1 involved a loading dose on day 1 of 9 mg/kg i.v. q12h, a maintenance dose on days 2 to 7 of 8 mg/kg i.v. q12h, and a maintenance dose on days 8 to 14 of 9 mg/kg p.o. q12h (maximum, 350 mg). Treatment regimen 2 involved a loading dose on day 1 of 6 mg/kg i.v. q12h, a maintenance dose on days 2 to 7 of 4 mg/kg i.v. q12h, and a maintenance dose on days 8 to 14 of 200 mg/kg p.o. q12h. AUC_0–12,ss, area under the curve over the 12-h dosing interval at steady state; C_min,ss, trough plasma concentration at steady state; i.v., intravenously; p.o., orally.

The voriconazole exposures by CYP2C19 genotyping status are summarized in Table 4. Of the 2 patients categorized as PM, one had the highest exposure values for intravenous dosing, and the other had values among the highest. Although the average exposure values in the HEM group were higher than those in the EM group, there was a substantial overlap in the voriconazole exposures between these 2 groups. The CYP2C19 genotyping status had no impact on the correlation between the AUC_0–12,ss and C_min,ss values (Fig. 2).

TABLE 4.

Summary of plasma voriconazole exposure parameters by CYP2C19 genotype

Parameter by administration routea	Result by CYP2C19 genotypeb
	EM	HEM	PM
Intravenous (n)	8	10	2
AUC0–12,ss (μg · h/ml)
Geometric mean (%CV)	36.0 (61)	56.4 (50)	128 (49)
Median (range)	37.6 (14.2–70.0)	66.6 (23.0–103)	134.4 (91.8–177)
Cmax,ss (μg/ml)
Geometric mean (%CV)	5.32 (42)	8.12 (37)	15.7 (32)
Median (range)	5.97 (2.32–8.31)	8.75 (4.62–12.5)	16.1 (12.6–19.6)
Cmin,ss (μg/ml)c
Geometric mean (%CV)	1.83 (127)	2.61 (86)	7.82 (42)
Median (range)	1.88 (0.471–9.36)	2.95 (0.596–7.12)	8.14 (5.88–10.4)
Oral (n)	6	10	2
AUC0–12,ss (μg · h/ml)
Geometric mean (%CV)	31.2 (102)	49.3 (79)	99.1 (24)
Median (range)	39.6 (10.0–80.8)	45.6 (14.5–156)	101 (84.0–117)
Cmax,ss (μg/ml)
Geometric mean (%CV)	5.49 (70)	7.66 (49)	12.3 (8)
Median (range)	5.93 (2.03–11.0)	6.48 (4.45–18.3)	12.3 (11.6–13.0)
Cmin,ss (μg/ml)c
Geometric mean (%CV)	1.17 (170)	1.89 (184)	5.13 (36)
Median (range)	1.65 (0.283–4.50)	1.68 (0.148–12.3)	5.30 (4.00–6.59)

^aAUC_0–12,ss, area under the curve over dosing interval at steady state; C_max,ss, peak plasma concentration at steady state; C_min,ss, trough plasma concentration at steady state.

^bEM, extensive metabolizer; HEM, heterozygous extensive metabolizer; PM, poor metabolizer.

^cn = 9 (EM), n = 10 (HEM), and n = 2 (PM).

As shown in Fig. 3, the observed total voriconazole exposures (AUC_0–12,ss) by age-weight group in the Japanese pediatric patients were within the distribution of the predicted exposures in the Western pediatric subjects based on previous modeling, although the Japanese patients had higher average AUC_0–12,ss than did the Western subjects. This indicates that the voriconazole exposures were comparable between the Japanese and Western pediatric subjects at the same dosing regimens.

FIG 3.

Comparisons of observed exposures (AUC_0–12,ss) in Japanese children (2 to ≤12 years) and adolescents (12 to <15 years, weight <50 kg or weight ≥50 kg) compared with predicted exposures in Western pediatric subjects at intravenous doses of 8 or 4 mg/kg and oral doses of 9 mg/kg (up to a maximum of 350 mg) or 200 mg. In each diagram, the box represents the interquartile distance, with the median indicated by a solid line in the center of the box, the whiskers representing ≤1.5 times the interquartile range, and outliers are represented by points (open circles) outside the whiskers for the predicted exposures in Western patients. The actual exposures in the Japanese patients are represented by the solid circles. AUC_0–12,ss, area under the curve over the 12-h dosing interval at steady state.

Among the 18 patients who completed both intravenous and oral treatments, the average ratios of oral AUC_0–12,ss to intravenous AUC_0–12,ss were 1.0 for the 14 children (2 to <12 years), 0.6 for the 3 young adolescents weighing <50 kg, and 0.5 for the 1 young adolescent weighing ≥50 kg. Based on the CYP2C19 genotype, the average ratio of oral AUC_0–12,ss to intravenous AUC_0–12,ss was 1.0 for 6 EM patients and 10 HEM patients and 0.8 for 2 PM patients. The individual values for the ratio of oral AUC_0–12,ss to intravenous AUC_12,ss ranged from 0.21 to 2.15, which is also a reflection of the large interindividual variability observed. It is noted that no obvious reasons can be identified for the cases with the AUC_0–12,ss ratio of >1.1 (assuming 100% absorption).

For the N-oxide metabolite, the median T_max was longer after intravenous (4 to 5 h) than oral (2 h) administration, due to the relatively long infusion time (see Table S1 in the supplemental material). The time course of the plasma concentration of the N-oxide metabolite was relatively flat throughout the 12-h dosing interval for both intravenous and oral administration. For all patients, the geometric mean AUC_0–12,ss values of the N-oxide metabolite following intravenous and oral administration were 65.7 and 77.9 μg · h/ml, respectively. Exposure to the N-oxide metabolite (AUC_0–12,ss and C_max,ss) by age-weight group was generally similar to the exposure to voriconazole. When the N-oxide metabolite pharmacokinetic parameters were summarized by CYP2C19 genotype status, the geometric means of the AUC_0–12,ss and C_max,ss were highest in the EM group and lowest in the PM group (see Table S2 in the supplemental material).

A comparison of the N-oxide metabolite-to-parent compound AUC_0–12,ss ratio by route of administration showed that more N-oxide metabolite was produced after oral administration than after intravenous administration (arithmetic means, 1.43 and 1.90 for intravenous and oral administration, respectively). As expected, the CYP2C19 genotype significantly affected N-oxide metabolite production, with the highest levels of the N-oxide metabolite observed in patients with CYP2C19 EM, and the lowest levels observed in patients with the CYP2C19 PM genotype (see Table S3 in the supplemental material).

Safety.

In total, 18 patients (85.7%) reported 80 all-causality AEs, and 12 patients (57.1%) reported 19 treatment-related AEs. No AEs were reported by 3 patients. No apparent trend in the incidences of AEs was observed by age-weight or CYP2C19 genotype. No deaths or serious AEs were reported. Two patients discontinued treatment due to AEs (patients 15 and 18 in Table S4 in the supplemental material). One patient had severely abnormal hepatic function (≥2× the ULN at baseline and ≥5× the ULN on day 10) and the other had moderately abnormal liver function tests (≥2× the ULN at baseline and ≥5× the ULN on day 7 before switching to oral treatment); both events were considered to be related to the study treatment, required treatment, and were confirmed to have resolved after discontinuation of the study treatment. No breakthrough fungal infections were reported.

The most common all-causality AEs are shown in Table 5. The most common (incidence, ≥10%) treatment-related AEs were photophobia (9 patients [42.9%]) and abnormal hepatic function (3 patients [14.3%]). There was no correlation between exposure and treatment-related photophobia; however, patients with treatment-related abnormal hepatic function showed relatively higher voriconazole exposures (see Table S4 in the supplemental material). The majority of AEs were mild to moderate in severity, with 4 patients (19.0%) reporting severe AEs (abdominal pain, increased ALT levels, and oropharyngeal pain [1 patient], febrile neutropenia and epistaxis [1 patient], abnormal hepatic function [1 patient], and sepsis [1 patient]), and 1 patient (4.8%) reporting a treatment-related severe AE (abnormal hepatic function). Excluding the single case of increased ALT level, the severe AEs were confirmed to have resolved.

TABLE 5.

All-causality adverse events reported for ≥3 patients

Adverse event	No. (%) with AE gradea:
	All	Mild	Moderate	Severe
Febrile neutropenia	13 (61.9)	6 (28.6)	6 (28.6)	1 (4.8)
Photophobia	9 (42.9)	8 (38.1)	1 (4.8)	0
Rash	4 (19.0)	4 (19.0)	0	0
Alanine aminotransferase increased	3 (14.3)	1 (4.8)	1 (4.8)	1 (4.8)
Aspartate aminotransferase increased	3 (14.3)	1 (4.8)	2 (9.5)	0
Hepatic function abnormal	3 (14.3)	1 (4.8)	1 (4.8)	1 (4.8)

^an = 21 patients total.

As was to be expected from the underlying disease and concomitant treatment, the incidence of hematology test abnormalities was high. The most common nonhematological test abnormalities (incidence, ≥20%) were increases in C-reactive protein (76.2%), gamma-glutamyl transferase (38.1%), AST (28.6%), and ALT (28.6%). No vital sign or electrocardiogram abnormalities were reported as AEs.

DISCUSSION

This study confirms that the doses derived from the most recent population pharmacokinetic modeling (17) were appropriate for pediatric use. The observed voriconazole exposures in the Japanese pediatric patients were comparable to the predicted exposures in the Western pediatric patients receiving the same dosing regimens (17). No particular trends were observed between the voriconazole pharmacokinetic (PK) parameters and age-weight group, indicating that the recommended dosing regimen based on age and body weight is appropriate for achieving similar voriconazole exposure across these groups. A recent retrospective analysis of voriconazole trough concentrations in 24 Japanese children and adolescent patients receiving different dosing regimens also showed consistent results with those of our study in general, although the analysis approach was different (19).

In comparison with the voriconazole exposures in healthy adults, the average voriconazole exposures in the Japanese pediatric patients were higher (median [range] AUC_0–12,ss, 59.3 μg · h/ml [14.2 to 177 μg · h/ml] versus. 37.6 μg · h/ml [13.7 to 104 μg · h/ml]) at matching intravenous doses (8 mg/kg versus 4 mg/kg, respectively), and much higher (median [range] AUC_0–12,ss, 45.6 μg · h/ml [10.0 to 156 μg · h/ml] versus 12.8 μg · h/ml [4.89 to 61.6 μg · h/ml]) at matching oral doses (9 mg/kg versus 200 mg, respectively) (13). Nonetheless, there were substantial overlaps in the voriconazole exposures between these groups. It has been demonstrated that a 200-mg p.o. dose in healthy adults provides exposure similar to a 3-mg/kg i.v. dose, and a 300-mg p.o. dose provides exposure similar to a 4-mg/kg i.v. dose (20). In pediatric patients, a 9-mg/kg p.o. dose is expected to provide voriconazole exposure at least as high as that with a 200-mg p.o. dose in adults (17). The data showed a much larger variability in voriconazole exposure in the Japanese pediatric patients than in healthy adults, especially for oral administration, which may be at least partly due to the many concomitant medications and complicated background therapy administered to these pediatric patients.

In a recent phase 3 study in adult recipients of hematopoietic stem cell transplantation (HSCT) and patients with hematologic malignancies with invasive aspergillosis (IA), the estimated geometric mean (%CV) of voriconazole AUC_0–12,ss at 4 mg/kg i.v. q12h was 51.0 μg · h/ml (43%) (21), which was similar to that observed in our study (51.1 μg · h/ml [68%]). The estimated geometric mean (%CV) AUC_0–12,ss values at 200 mg and 300 mg of p.o. voriconazole q12h in adult patients with IA were 22.0 μg · h/ml (46%) and 32.9 μg · h/ml (45%), respectively (21), which were lower than those in the Japanese pediatric patients after oral administration (45.8 μg · h/ml [90%]). Again, due to very large interindividual variability in the Japanese pediatric patients after oral administration, there were substantial overlaps in voriconazole exposures between the Japanese pediatric and adult patients with IA.

The CYP2C19 genotype is likely one of the factors contributing to the variability of voriconazole pharmacokinetics in both adult and pediatric patients. In this study, the 2 patients with the CYP2C19 PM genotype had among the highest exposures after both intravenous and oral administration. For intravenous administration, the second highest exposure was observed in a patient with HEM, and for oral administration, the 2 highest exposures were both observed in patients with HEM.

The safety profile observed in the pediatric patients in this study using the newly recommended voriconazole regimens raised no further concerns compared with the safety profiles in adults and in previous pediatric studies (10, 13–16). The intravenous-to-oral switch regimens of voriconazole were safe and well tolerated, showing no obvious differences from the existing safety profile. The majority of AEs were mild or moderate in severity. The frequently reported treatment-related AEs were photophobia and abnormal hepatic function. There was no breakthrough fungal infections during the study. Since most of the patients in this study had received prior chemotherapy for treatment of the underlying disease, the observed abnormalities in liver function tests may have been a consequence of the chemotherapy. It is noted that this study has the limitation that the patients received voriconazole treatment for a shorter duration than occurs in usual clinical practice.

The high correlation between voriconazole exposure and C_min,ss indicates that trough concentration is a good parameter for providing a rough estimation of voriconazole exposure. Many published studies have shown a positive association between voriconazole concentrations and treatment-related toxicity and/or clinical efficacy (22–31). The proposed target minimum values for voriconazole C_min,ss to improve clinical outcomes included 1 or 2 μg/ml, and the proposed target maximum values for voriconazole C_min,ss to minimize treatment-related toxicity included 6, 5, or even 4 μg/ml. Notably, most of the maximum target concentration proposals were based on the identification of the association between elevated concentrations and neurologic AEs but not hepatic AEs. The Japanese practice guidelines for therapeutic drug monitoring (TDM) of voriconazole also recommend a target trough level of 1 to 2 μg/ml in terms of clinical efficacy, and in patients with a trough level of >4 to 5 μg/ml, elevated liver function test results potentially attributable to voriconazole should be considered (32). On the other hand, 2 recently published retrospective analyses of large-scale voriconazole TDM data from real-world clinical settings showed no associations between the voriconazole concentrations and clinical outcomes or treatment-related toxicity (33, 34). An analysis of data from a recent phase 3 study in adult patients with IA also did not establish positive associations between the voriconazole concentrations and clinical outcomes or treatment-related toxicity when voriconazole was administered as a monotherapy (21). Until now, no formal consensus on the voriconazole exposure-response relationship has been reached, due to the complex clinical setting of fungal infections.

In our study, the correlation between voriconazole exposure and hepatic AEs was not evident, although the sample size was limited. Specifically, 6 patients had trough concentrations of >5 μg/ml (2 of these patients were 10 years old and the other 4 patients were 6, 11, 12, and 13 years old), and only one of them (patient 15, with a C_min,ss of 9.36 μg/ml; see Table S4 in the supplemental material) experienced a moderately abnormal liver function test (treatment related, which was then discontinued), while the other 2 patients with severe liver abnormalities (patients 8 and 18 [treatment related, which was then discontinued]; see Table S4) had trough concentrations of <4 μg/ml. In our study, 19.0% (4/21) of patients had voriconazole trough concentrations of <1 μg/ml following intravenous administration and 16.7% (3/18) following oral administration. Hence, the dosing regimens evaluated in this study are considered acceptable as starting doses for pediatric patients, although the interindividual variability in voriconazole exposure was large.

It is recognized that some pediatric patients may have much higher voriconazole exposure at the recommended dosing regimens due to large interindividual variability, especially with oral administration. Therefore, close monitoring of AEs is warranted, and flexible dose adjustment should be considered for children in clinical practice. For this purpose, TDM would be useful as an additional tool for voriconazole management in pediatric patients, although there is no definitive setup of a therapeutic range for voriconazole trough concentrations.

The strength of this study is that the pharmacokinetic parameters were obtained from noncompartmental analysis using intensive concentration data. However, its limitations include the small number of patients to extrapolate the data for adolescents with body weight of ≥50 kg and for the effect of CYP2C19 genotype on voriconazole clearance in pediatric patients. It was difficult to appropriately estimate the oral bioavailability of voriconazole in this study due to the large variability in the ratios of oral AUC_0–12,ss to intravenous AUC_0–12,ss (0.21 to 2.15).

In summary, the observed voriconazole exposures in this Japanese population were comparable to the predicted exposures in Western pediatric subjects at the same dosing regimens. Voriconazole was well tolerated, and no further safety concerns were raised for pediatric patients. The recommended doses derived from population pharmacokinetic modeling appear to be appropriate as the starting doses for Japanese pediatric patients, with no additional safety risks compared to those with adult patients.