Optimized Administration of Voriconazole and Therapeutic Drug Monitoring in Children and Adolescents: A Single-Centre Retrospective Experience from China

Background Voriconazole (VRC) is a triazole anti-fungal agent and a rst-line treatment for invasive fungal infection (IFI) generally. The purpose of our study was performed to explore the factors affecting voriconazole trough concentration (C trough ) and to show VRC dose adjustment experience in children. Methods The demographic information, concentration data, CYP2C19 genotypes and clinical outcomes of eligible children from January 1th, 2016 to December 31th, 2018 were collected. Factors affecting the voriconazole trough concentration were statistically analyzed. Results A total of 145 trough concentrations in 94 patients were included in this study, 54.5% of which achieved the target concentrations; however, 35.9% and 9.6% of which were sub-therapeutic and super-therapeutic post-multiple dosing. For children ≤ 2, 2–6, 6–12, and 12–18 years, the median VRC maintenance doses of 5.7, 6.7, 5.0 and 3.3 mg/kg twice daily respectively had been required in order to achieve therapeutic level (P < 0.001). Co-administration of proton pump inhibitors affected VRC target trough concentration signicantly (P = 0.001). early


Introduction
2017 Updated guideline has demonstrated that children are at a high risk of invasive fungal diseases (IFD) undergoing allogeneic hematopoietic stem cell transplants (HSCT), leukemia, prolonged neutropenia and receiving high-dose corticosteroids 1 , which had made an important contribution to the proportion of morbidity and mortality 2 . Voriconazole is a second-generation triazole antifungal agent with broad antifungal spectrum including Candida, Aspergillus, Fusarium, Cryptococcus neoformans and Scedosporium 3,4 . It was recommended by Aspergillus guidelines for rst-line treatment of invasive aspergillosis and being widespread used in clinic 5 . However, voriconazole dosing regimen has been particularly challenging for children.
There are signi cant individual variations in the pharmacokinetics of voriconazole in children and adults.
Several studies showed a signi cant concentration-dependent response between high steady-state plasma concentration and adverse events, between inadequate concentration and treatment failure 3,6,7 .
The steady-state voriconazole trough concentration was affected by various factors, such as age, ethnic, dosage, drug combination, genetic polymorphism of CYP2C19 enzyme and enzymes' activity. Hence, in order to ensure the e cacy of antifungal and reduce adverse reactions of voriconazole, routine and repeat therapeutic drug monitoring (TDM) of the steady-state voriconazole trough concentration was strongly recommended by the European Conference on Infections in Leukemia (ECIL) guidelines 8 . To date, no recommended voriconazole dosing is applicable to Asian children. With published narrow therapeutic range (1.0-5.5 mg/L) 9 , to achieve VRC exposures, 7 to 8 mg/kg intravenous twice daily of dose ranges were recommended in Caucasian children 7 . However, owing to lack of continually VRC TDM, a maintenance dose lower than the package insert manner was commonly prescribed in children patients in our hospital.
TDM of voriconazole in Asian children are not always available for clinician due to very limited studies.
For now, no de nite guidelines are available for voriconazole dose adjustment in pediatric patients. Therefore, we described our experience with voriconazole administration, dose adjustment and therapeutic drug monitoring. In this research, we retrospectively collected clinical outcomes of pediatric patients included through the electronic medical system at our hospital. The present study aimed (i) to investigate the relationship between steady-state voriconazole trough concentration and clinical response and adverse reactions, (ii) to explore the affecting factors on voriconazole trough concentrations and (iii) to study the impact of CYP2C19 genetic polymorphism on voriconazole metabolism in the growing child.

Setting
This single-center retrospective study was performed at the Second Xiangya Hospital of Central South University after approved by the ethics committee of the Second Xiangya Hospital of Central South University (ChiCTR.org Registration number: ChiCTR1900025821, 09/09/2019), all patients or their legal guardian provided informed consent for the usage of their clinical data and/or samples.
The full trial protocol can be accessed on the website of http://www.chictr.org.cn/index.aspx if available.

Study population
Children patients aged 1 to ≤ 18 years who was administered voriconazole enterally or parenterally and monitored at least one voriconazole plasma trough concentration from between 1st January 2016 and 31st December 2018 was eligible for study. Patients with multiple hospitalization would be a cumulative statistic data.
Exclusion criteria were: out-patients, incomplete medical record information and body weight of ≥ 50 kg when patients aged of 12 to 14 years.
From the electronic medical record information system, patients' demographic and clinical data were collected, including ethnicity, age, sex, body weight (BW), CYP2C19 genotype, underlying disease, treatment indication, site of infection, VRC dosing, route of administration, VRC trough concentrations, treatment duration, concomitant medications, adverse drug reactions, e cacy, liver function, and kidney function.

VRC administration and TDM blood sampling
Due to lacking of guidelines for the administration and monitoring of voriconazole in children, voriconazole dosing regimens were performed by clinicians considering recommending VRC package insert dose and the actual condition of patients in present study. Before the next dose, all TDM blood samples were collected at steady-state conditions. Repeat TDM was conducted by the clinician according to treatment response. The ideal target trough concentration of VRC was 1.0-5.5 µg/ mL 9 .
Analysis of VRC trough concentration VRC plasma concentration was measured by an automatic two-dimensional liquid chromatography method reported in our previous paper 10 . The linearity range was 0.35 to 11.26 µg/mL. The intra-day and inter-day precisions were 1.94-2.22% and 2.15-6.78%, respectively.
CPY2C19genotypes and phenotype assignment DNA was separated from the suspending white cells and was puri ed with the E.Z.N.A® SQ Blood DNA Kit II method. CYP2C19 genotypes was implemented by Sanger dideoxy DNA sequencing method with ABI3730xl-full automatic sequencing instrument (ABI Co., Carlsbad, California) from Boshang Biotechnology Co. Ltd. in Shanghai, China. According to de nition of the Clinical Pharmacogenetics Implementation Consortium (CPIC) 11 , CYP2C19 phenotypes were categorized as several types based on CYP2C19 *1, *2, *3, or *17 allele nomenclature, including ultra-rapid metabolizer (UM), rapid metabolizer (RM), extensive metabolizer (EM), intermediate metabolizer (IM) and poor metabolizer (PM). Extensive metabolizers were considered as a normal or wild-type metabolizer of CYP2C19.

Outcome and safety assessment
The de nition of IFI treatment was in accordance with the European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group and the National Institute of Allergy and Infectious Diseases Mycoses Study Group (EORTC/MSG) 12 . Clinical outcomes of target antifungal therapy were categorized as "clinical successful" (complete or partial remission of clinically signi cant signs and symptoms and improvement of imaging manifestations at the end of treatment) or "clinical failure" (no response, disease progression or death due to IFI). Prophylaxis and empirical antifungal therapy were classi ed as a 'successful' without breakthrough fungal infection after discontinuing voriconazole treatment.
Based on the Common Terminology Criteria for Adverse Events (CTCAE) de ned by the National Cancer Institute 13 , voriconazole-attributable adverse reaction such as visual disturbances, rash, vomiting, headache, elevated biomarkers of liver function, tachycardia and hallucinations at al. were categorized as a possible or stronger relationship event with voriconazole. Nevertheless, among them, increased liver biomarkers level, rash, and visual disturbances frequently resulted in voriconazole discontinuation.

Statistical analysis
Data was analyzed by SPSS (version 22.0; IBM Corp, Armonk, NY, USA). Continuous variables were expressed as median (range: quartiles). Categorical data was reported as frequencies and percentages. Using a Chi-squared test or Fisher's exact test and nonparametric Mann-Whitney U test or Kruskal-Wallis H test, categorical data and continuous variables were compared, respectively. A P value (2-side) of 0.05 was considered signi cant.

Demographics and clinical characteristics
Ninety-four patients with a total of 145 VRC trough concentrations were enrolled in this retrospective study. Forty-nine of them were males. The median age was 7.5 years old. The primary diagnosis was hematological malignancy, accounting for 76.6%, followed by respiratory infection (69.1%) and bloodstream infection (38.3%). The major infection sites were lung (66.0%) and blood (25.5%). Among the enrolled patients, only 15 (19%) patients were diagnosed with con rmed IFI. Most of patients (n = 48) were administrated prophylactically, together with therapeutically (20.2%) and empirically (28.7%) respectively post-diagnosis. Totally, 80.8% of the patients achieved clinical success. Patients' demographic data and clinical characteristics are summarized in Table 1. Immunosuppressive agents 9(9.6%) 0 3 0 6 PPIs: proton pump inhibitors Voriconazole administration and therapeutic trough concentration For all patients, 5.2 mg/kg of voriconazole twice daily was the median level. The median dose of 7.1, 6.3, 5.2 and 3.4 mg/kg twice daily was administered for subgroup by age of ≤ 2, 2-6, 6-12, and 12-18 years respectively. For them, children aged 12-18 had the highest maintenance dose (P < 0.001) (Fig. 1). Except for children aged ≤ 2y, children had lower maintenance dose than the recommended dose levels. The medication duration and time of initial VRC TDM varied widely, median of 17.5 days (11-26.2 days) and 7 days (3.8-11 days), respectively. Patients' voriconazole administration showed in Table 2. In all blood samples, 54.5% of them achieved the target concentrations; however, 35.9% were subtherapeutic and 9.6% were super-therapeutic post-multiple dose. 62.8.% (59 in 94) of patients achieved one or more therapeutic level. As showed in Fig. 2, the disparate and weight-adjusted VRC dose was prescribed in age groups owing to achieve therapeutic concentrations (P < 0.001) (Fig. 2). However, we interestingly found that patients achieved sub-therapeutic concentrations (n = 9) had a higher maintenance doses compared with those achieved therapeutic level (n = 6) in group aged ≤ 2 years (7.1 vs 5.8mg/kg).
Oral administration of voriconazole (n = 83) was the most common route in our study. But no signi cant difference was found between the maintenance dose and administration route (P = 0.852) (Fig. 2).
Besides that, co-administration was also an important factor affecting VRC trough concentration. In our research, the group of combining PPI required higher dose in achieving therapeutic level compared with non-PPI group (P = 0.001) (Fig. 3). The inter-individual variability of the initial trough concentration in 94 patients was 120.6% (94 initial trough concentrations), and the intra-individual variability was 61.1% (33.1%-103.5%) (21 patients with the same dose, the same administration route and TDM frequency ≥ 2 times) (Fig. 4).

Outcome and Safety
Outcome data in patients with therapeutic, empirical or prophylactic treatment were analyzed. There was no signi cant difference in VRC dosing when compared clinical successful and clinical failure categorized by treatment indication (empirical/therapeutic: P = 0.867, prophylactic: P = 0.762) (Fig. 6). In patients with empirical/therapeutic course, VRC maintenance dose of 5.0mg/kg required for successful outcome (34/46;73.9%). For prophylactic treatment, 42 patients achieved successful outcome at median dose of 5.2mg/kg (4.0-6.4mg/kg). No signi cant difference was found between empirical/therapeutic group and prophylactic group.
Three of 94 patients occurred voriconazole-related adverse reactions, 2 of them with persistent elevation of alanine aminotransferase (ALT) and/or aspartate aminotransferase (AST) and vomiting just in one. One patient with trough concentration of 5.32 µg/mL gradually recovered to normal liver function after drug withdrawal and the other one also showed normal liver function by reducing VRC dosage. For one patient administered voriconazole tablets, the symptoms of vomiting disappeared after intravenous administration. All the 3 patients nally achieved clinical effectiveness.

Discussion
Hematological malignant tumor patients had a high-risk of fungal infection due to immunode ciency and chemotherapy drugs. Voriconazole is increasingly used for the prevention and treatment of IFI. The population pharmacokinetic characteristics of voriconazole in children were more complicated 14 and showed a large individual variation. Therefore, this retrospective study highlighted the reasons and factors affected VRC therapeutic trough concentration in children and discussed the dose adjustment strategy of voriconazole.
At the current maintenance dose for children younger than 12 years, voriconazole shows nonlinear pharmacokinetics 15,16 , warranting more cautious dose adjustment. Our study showed a large variability in voriconazole trough levels, with rate of 120.6% and 61.1% in inter-individual and intra-individual variability, similar to the published study 17 . More than 50% of children could not reach the target range (1.0-5.5µg/mL) at the initial trough concentration, which was consistent with many previous studies 9,18 , but was different with a European research 6 . The initial trough concentration in children aged 2-12 was signi cantly higher than that in children aged under 2 years, suggested a faster metabolism for infants. The EMA approved higher doses in 2 to 12 years old children, an 8 mg/kg intravenously twice daily (9 mg/kg day 1) or a 9 mg/kg orally twice daily. Our data suggested that underdose was so prevalent that Shima H et al 20 who recommended at least 8.5 mg/kg twice daily of VRC optimal dose for patients younger than 2 years, children with leukemia in our center might require higher doses and more frequent monitoring. The primary administration route of voriconazole in the present study was oral administration (88.3%) rather than the intravenous route recommended by the package insert for VRC. However, variability in oral bioavailability caused by meals and hepatic rst-pass effect might associate with lower drug exposure 3,21 , which could be the reason for sub-therapeutic trough concentration.
Similar to that reported (7-10 days) in other centers 19 , the average time of the rst TDM blood sampling in our hospital was 7 days (3.8-11days). In a randomized controlled trial 22 , the rst TDM blood sampling was done on the fourth day after the initiation of voriconazole, which was calculated based on data from the literature so that it coincided with the target trough concentration range (1.0-5.5 mg/L). However, Miyakis S et.al 23 showed that the initial trough concentrations of ≤ 0.35 µg/mL were signi cantly associated with increased mortality in pediatric with invasive candidiasis, so it was advised that the rst steady-state TDM should be done ideally as early as possible (day 3 of therapy) to allow prompt dose adjustment. Our ndings are consistent with previous study 24 , the co-administration of PPTs, mainly omeprazole, could signi cantly increase voriconazole plasma concentrations through CYP2C19 inhibition. However, in children the further exploration of the clinical implication of this drug interaction is absolutely imperative.
Another important factor affecting VRC therapeutic trough concentrations was polymorphism of CYP2C19. The CYP2C19 genotypic and phenotypic variability were extensive among different ethnic groups, but was controversial in relation to voriconazole treatment. Generally, the Caucasians or Africans have a lower proportion of PM metabolizers than Asians (2-5%, 6% and 13-23%, respectively). And for Caucasians and Africans, it is about 4 times more proportion in the Asian than them among the CYP2C19 * 17 allele [25][26][27][28][29]  Voriconazole, given orally or intravenously, is well tolerated in most of patients, with a rate of adverse reactions of 3.2%, which is much lower than previous reported (20.0%) 18 . Visual disturbances and photosensitive skin reactions were di cult to be detected and explained in pediatric may be one reason for that lower rate of adverse reactions.
Our study is a single-center retrospective study. There are a few limitations in study design and analysis.
First, too small sample size restricts the difference signi cance analysis.
Second, no uniform guidelines and standardized protocol for VRC dose adjustment based on TDM data, resulting in a few confusions in doctors when adjusting doses. Third, the number of CYP2C19 genotypic and phenotypic is too small to detect difference between trough concentration and genotypes.

Conclusion
In conclusion, there are large inter-individual and intra-individual variability with voriconazole trough concentration in children, and many factors could affect VRC therapeutic trough concentration, including age, dose, time of TDM sampling, PPI coadministration and CYP2C19 genotype. Younger pediatric patients might need to have a higher dosage regime. The early, routine and repeat TDM of trough concentration is extremely necessary in order to ensure safety and effective treatment. It is crucial to determine the initial dose of voriconazole based on CYP2C19 genotype data. Therefore, a large size, multi-center prospective studies is imperative to identify and validate dose-optimization strategies for pediatric patients in the future.

Declarations
Figures Figure 1 Distribution of voriconazole initial maintenance dose at different age groups (n=94), each point represents the dose for each patient. VRC maintenance doses required to achieve at least one therapeutic trough concentration (1.0-5.5µg/mL) categorized by age and route of administration (n=59).

Figure 3
Box plot of VRC maintenance dose requiring to achieve at least one therapeutic trough concentration (1.0-5.5µg/mL) categorized by age and PPI coadministration, which was represented by median, minimum, maximum, and interquartile range.

Figure 4
Distribution and interpatient variability of initial voriconazole trough concentrations at different weightadjusted maintenance doses (n=94).

Figure 6
Page 19/19 VRC maintenance doses required to achieve clinical successful and clinical failure categorized by treatment indication.