4 Discussion
The factors affecting voriconazole Cmin and PPK date of voriconazole in pediatric patients with hematologic malignancies are limited. we investigated the factors affecting voriconazole Cminand successfully developed a PPK model of voriconazole. Most notably, the optimal regimes were recommended based on the final model.
Regarding the influencing factors of voriconazole Cmin, as reported in adult patients with hematologic malignancies and other pediatric patients24, 25, voriconazole exhibited a non-linear pharmacokinetic profile. A report of adult inpatients26 showed no difference in voriconazole concentrations between intravenous and oral administration, which is inconsistent with our results. Karin et al27 argued that children have a greater metabolic capacity. Another possible explanation is that patients with malignant hematological diseases are prone to nausea and vomiting after chemotherapy. We also found a negative correlation between ALB and voriconazole concentration as reported by previous study28. Accordingly, we should be more alert to the effect of ALB levels on voriconazole concentrations, especially in patients with low serum albumin. In terms of concomitant administration, the results of this study showed that combination with PPIs resulted in higher concentrations while Co-administration of GLU resulted in lower concentrations, which is in line with previous studies29, 30. Some studies have documented that gender31, SCR32, liver function indicators (ALP, AST, ALT, etc.)28, 33and CRP levels34, 35 could affect voriconazole concentrations. However, no significant correlation between these factors and voriconazole concentration were found in this study. The explanation for this may be the difference of race, age and population and further investigation are necessary in the future.
In the part of PPK analysis, the results show that a one-compartment model fits the data best. However, the other studies of children were two-compartment models36-38, probably due to the fact that pediatric studies are mainly preclinical studies and multicenter large sample studies with a dense sampling strategy, while the sampling strategy of this study was sparse. Fortunately, Farkas et al.39 compared the accuracy and precision of voriconazole linear, nonlinear, and mixed linear models for prediction. The results suggested that the linear model was slightly more accurate than the other two models, but the differences were minor from a clinical point of view, making all three models suitable for voriconazole.
As summarized in a review of PPK of voriconazole40, the median of central compartment volume of distribution (VC) was 1.07 L/kg (0.81-3.26 L/kg), and the interindividual variability of V or V1 and CL was 14.2% (13.6 to 45.4%) and 69.6% (66.5% to 117.4%). These shows that the pharmacokinetic parameters of voriconazole vary significantly among the different pediatric populations, further demonstrating the necessity for individualized dosage in pediatric populations with hematologic malignancies. The population typical values of CL of voriconazole in another pediatric one-compartment model were 2.94 L/h and the apparent volume of distribution was 6.17 L and 7.67 L.40However, the typical values of CL and V in this study were only 1.52 L/h and 35.7 L. The possible reasons are as follows: the patients included in this study are patients with hematologic malignancies, who had different degrees of thrombocytopenia and liver dysfunction due to myelosuppression after receiving chemotherapy. However, several studies demonstrated that the reduction of voriconazole elimination is significantly related to liver function16, 41. It was also documented in the previous studies that platelet counts were associated with the severity of liver function, where the CL was only 0.88 L/h and 0.58 L/h12, 42.
The oral bioavailability in this study is 90.9%, higher than 44.6% ~ 73% in other study of children,22, 23, 43, 44 but lower than 96% and 94.2% in adult.45, 46 Numerous literatures reported that the oral bioavailability is affected by CYP2C19 genotype, dose, adverse reactions of chemotherapy, such as nausea and vomiting, disease status, gastrointestinal function and diet.45, 47, 48 A study written by Scholz et al48 indicated that after intravenous and oral administration of 400 mg voriconazole, the bioavailability of PM was 94.4% (78.8%, 109.9%), while EM was 75.2% (62.9,87.4%). Most notably, the proportion of PM patients in this study is 14.3%, which is much larger than that in Japan and Europe by 2.0% ~ 9.5%.49, 50 Besides, the bioavailability of 50 mg and 400 mg voriconazole was 39% and 86% respectively from the study by Hohmann et al6. These may be the reason why the oral bioavailability of this study is higher than that of other children’s studies.
The covariates included in the final model of previous pediatric PPK studies22, 49, 50 were body weight, CYP2C19 genotypes and ALT. In contrast, the EM and ALB were the significant covariates for the CL of voriconazole, and there were no covariates that had a significant effect on V in the present work. The results of adults with hematological malignancies 19 showed that CYP2C19 genotype significantly affects CL of voriconazole and AUC0-12h of PM is 2.5 and 1.8 times of EM and IM respectively. In this study, the CL of voriconazole in children with non-EM is reduced by 29.6% compared with EM, which is similar to the 35.5% of another pediatric study.50 The Wei et al51 study showed that Low ALB level is significantly correlated with the blood concentration of voriconazole > 5.5 mg/L. Moreover, the study conducted by Dote et al52 showed that hypoproteinemia (ALB < 30 g/L) is associated with the low CL of voriconazole. Therefore, attention should be paid to monitoring voriconazole blood concentrations in patients with low protein levels and to developing individualized dosing regimens based on ALB levels in conjunction with PPK models to improve treatment efficacy and reduce adverse effects associated with increased blood concentrations.
Currently, there is no recommendation of individualized drug administration for children with hematologic malignancy. In the present study, the dosing was simulated according to CYP2C19 genotype and ALB level. The recommended doses (mg/kg) were4, 8 and 12 mg/kg intravenously o orally twice daily for ALB levels of 20.0~35.0 g/L, 35.1~45.0 g/L and 45.1~55.0 g/L for patients with EM, and were 2, 4 and 7 mg/kg intravenously or orally twice daily for patients with non-EM. The results indicate that higher dose (mg/kg) should be recommended for extensive metabolizer, which is similar to the study of Lin et al53. Besides, as described by Prawat et al54, patents with lower ALB levels were more likely to achieve the target trough concentration. Nevertheless, when the ALB level of EM is 45.1 ~ 55.0 g/L, the proportion of the concentration < 0.5 mg/L is still high even if the dosage reaches 12 mg/kg in oral or intravenous administration. Hence, whether such patients choose to give high-dose voriconazole or switch to other antifungal drugs in clinical medication remains to be further considered for its efficacy and safety. Overall, the recommended regimen significantly improved the probability of reaching the target trough concentration of 0.5 – 5 mg/L and minimized the probability of the concentration < 0.5 mg/L and > 5 mg/L, which could maximize the efficacy and minimize the adverse reactions of voriconazole in the treatment process.
This study has several limitations: Firstly, due to the limitations of clinical sampling and testing techniques, only a few patients had MIC values, and it was not possible to combine PK/PD parameters AUC0-24h/MIC for dosing regimen development. Secondly, the effects CYP2C9 and CYP3A4 genotype on the pharmacokinetic parameters of voriconazole were not investigated. Finally, this study was a single-center study, whether the findings are applicable to other centers needs to be verified in the future.