Discussion
Although the pharmacokinetics of PCr has already been investigated in adults, its pharmacokinetic characteristics in children is unclear and its metabolite Cr was ignored. To the best of our knowledge, this is not only the first PK study of PCr in children but also the first comprehensive population pharmacokinetic model of PCr and its metabolite Cr. In this study, the pharmacokinetic characteristics of PCr and Cr in children were studied by population pharmacokinetic method and the established joint model fitted the concentration data well. This study could be helpful to further understand the PCr and Cr PK characteristics in children and provide reference for pediatric clinical application.
In the base model development process, the structural model of both PCr and Cr were tested with one-compartment model and two-compartment model respectively. The PCr data was best fitted by a two-compartment model which is in accordance with a previous report [10, 11]. The structural model of Cr was best described by a two-compartment model with an estimation of baseline level of Cr level (baseCr).
There is only PCr dose for each subject and Cr was generated by PCr metabolism. The distribution volume of Cr and Fm of PCr metabolized to Cr couldn’t be estimated at the same time. It has been reported that large number of PCr were metabolized to Cr in animals after intravenous administration of PCr, and the conversion rate was about three-quarters [10]. In the concentration–time plots (Supplemental Material 3 ), Cr concentration increased rapidly after PCr infusion. In our study, we were more interested in knowing the distribution of Cr in children. Therefore, the apparent distribution volume of Cr was estimated and the fraction of PCr metabolized to Cr was fixed so as to avoid identifiability problem in the model building. There is no evidence that showed the exact ratio fraction of PCr to Cr in humans. The Cr concentration increased rapidly with the infusion of PCr seen in the concentration profiles and a general view is that a large amount of PCr is metabolized to Cr to provide ATP after PCr injection [5]. In this study, based on the results of animal experiment from the reference [10], the Fm of PCr converted to Cr was assumed to be 0.75. This setting could ensure the estimation the PK parameters of Cr in children and it didn’t affect the estimation of PCr parameters. To compare the estimation results of different fraction assumptions, we developed a model based on the fraction fixed 1 and the estimation results showed PCr parameters did not change much.
BW was incorporated in the base model with an allometric exponent of 0.75 for the clearance parameters and 1 for the volume terms which result in a large decrease in OFV. The addition of allometric scaling made the base model more stable and accelerated the speed of modeling. In this study, to find more factors that affect the in vivo behavior of the PCr and Cr as much as possible, a large amount of information of study sample were collected as covariates. After forward inclusion and backward elimination, GFR was identified as key covariates to the clearance of Cr. The influence of GFR on CLCr was estimated with exponent of 0.311, which means high level of GFR level lead to increase of Cr clearance. As GFR increased by 100%, Cr clearance increased by 24%.
The final PCr and Cr model was evaluated by bootstrap and VPC method, respectively. Results showed that the established model had good stability and predictability. The parameters estimated by the final model showed that distribution volume of the PCr central compartment (VcPCr) was 8.22 L, distribution volume of the PCr peripheral compartment (VpPCr) was 3.07 L, total distribution volume was 11.29 L, and the clearance was 1.33 L/min in children with 20 kg. The distribution volume of the Cr central compartment (VcCr) was 2.39 L, distribution volume of the Cr peripheral compartment (VpCr) was 2.9 L, and the clearance was 0.0825 L/min in children with 20 kg. The VpCr is larger than VcCr, suggesting that Cr can be widely distributed in muscle and other tissues. This study first reported the distribution characteristics of Cr in children under the Fm set to 0.75. The typical value of Cr base value (baseCr) is 66.6 μmol/L, which is about 8.73 mg/L, basically consistent with the normal content in human body (7-13 mg/L) [20]. The evaluated model was applied to simulate different regimens for subgroups with different age. The simulation results suggested that PCr did not accumulate in vivo.
In conclusion, the joint population pharmacokinetic model for PCr and its metabolite Cr in pediatric patients was successfully developed for the first time. The model adequately described the pharmacokinetics of both parent drug and metabolite with individual predicted value close to the measured value. This model may be helpful for pediatric clinical application of phosphocreatine.
Acknowledgements
The authors acknowledge the support of Haerbin laibotong pharmaceutical co. LTD for the clinical trial from which the Phosphocreatine and Creatine pharmacokinetic data were derived.
Competing Interests
The authors declare no conflicts of interest.
Contributors
All authors contributed to the design and implementation of the study. L.B.Z., Y.Y. and X.L.W. designed the study. N.S. and H.H. acquired the data. M.Z., Y.Z. and H.H. analyzed the data and develop the pharmacokinetic model. H.H. wrote the manuscript and L.B.Z., Y.Y. and X.L.W. revised it.