Population pharmacokinetics of sirolimus in pediatric patients with kaposiform hemangioendothelioma: A retrospective study

Abstract

Numerous studies have established population pharmacokinetics (PPK) models of sirolimus in various populations. However, a PPK model of sirolimus in Chinese patients with pediatric kaposiform hemangioendothelioma (PKHE) has yet to be established; therefore, this was the purpose of the present study. The present study was a retrospective analysis that utilized the trough concentration data obtained from traditional therapeutic drug monitoring-based dose adjustments. A total of 17 Chinese patients with PKHE from a real-world study were characterized by non-linear mixed-effects modeling. The impact of demographic features, biological characteristics and concomitant medications was assessed. The developed final model was evaluated via bootstrap and a prediction-corrected visual predictive check. A one-compartment model with first-order absorption and elimination was used for modeling of data for PKHE. The typical values of apparent oral clearance (CL/F) and apparent volume of distribution (V/F) in the final model were 3.19 l/h and 165 liters, respectively. Age, alanine transaminase levels and sex were included as significant covariates for CL/F, while the duration of treatment with sirolimus was a significant covariate for V/F. In conclusion, the present study developed and validated the first sirolimus PPK model for Chinese patients with PKHE.

Keywords: kaposiform hemangioendothelioma; pediatric; population pharmacokinetics; real-world study; sirolimus.

Figure 1.

Goodness-of-fit plots of the base population model. (A) Observations vs. population predictions. The x-axis value of a point is the population prediction of sirolimus blood concentration and the y-axis value of a point is the observation of sirolimus blood concentration. The closer the x- and y-axis values of the same point are, namely, the closer the point is to the y=x line, the closer the population prediction value predicted by the model is to the observation of sirolimus blood concentration. The black solid line is the line of unity (the y=x line). The red smooth line represents the trend of the points. Hence, the closer the red smooth line is to the black solid line, the more predictive the model is. (B) Observations vs. individual predictions. The x-axis value of a point is the individual prediction of sirolimus blood concentration and the y-axis value of a point is the observation of sirolimus blood concentration. The closer the x- and y-axis values of the same point are, namely, the closer the point is to the y=x line, the closer the individual prediction value predicted by the model is to the observation of sirolimus blood concentration. The black solid line is the line of unity (the y=x line). The red smooth line represents the trend of the points. Hence, the closer the red smooth line is to the black solid line, the more predictive the model is. (C) |iWRES| vs. individual predictions. iWRES is the difference in the values of the individual prediction and the observation of sirolimus blood concentration. The x-axis value of a point is the individual prediction of sirolimus blood concentration and the y-axis value of a point is the iWRES of the corresponding individual prediction. The smaller the iWRES the better the model predictability. The red smooth line represents the trend of the points. Hence, the closer the red smooth line is to the black solid line (the y=0 line), the more predictive the model is. iWRES, individual weighted residuals.

Figure 2.

Goodness-of-fit plots of the final population model. (A) Observations vs. population predictions. The x-axis value of a point is the population prediction of sirolimus blood concentration and the y-axis value of a point is the observed sirolimus blood concentration. The closer the x- and y-axis values of the same point are, namely, the closer the point is to the y=x line, the closer the population prediction value predicted by the model is to the observed sirolimus blood concentration. The black solid line is the line of unity (the y=x line). The red smooth line represents the trend of the points. Therefore, the closer the red smooth line is to the black solid line, the more predictive the model is. (B) Observations vs. individual predictions. The x-axis value of a point is the individual prediction of sirolimus blood concentration and the y-axis value of a point is the observed sirolimus blood concentration. The closer the x- and y-axis values of the same point are, namely, the closer the point is to the y=x line, the closer the individual prediction value predicted by the model is to the observation of sirolimus blood concentration. The black solid line is the line of unity (the y=x line). The red smooth line represents the trend of the points. Hence, the closer the red smooth line is to the black solid line, the more predictive the model is. (C) |iWRES| vs. individual predictions. iWRES is the difference in values between the individual prediction and the observed sirolimus blood concentration. The x-axis value of a point is the individual prediction of sirolimus blood concentration and the y-axis value of a point is the iWRES of the corresponding individual prediction. The smaller the iWRES value, the better the model predictability. The red smooth line represents the trend of the points. Therefore, the closer the red smooth line is to the black solid line (the y=0 line), the more predictive the model is. iWRES, individual weighted residuals.

Figure 3.

Distribution of weighted residuals for the final model. (A) Density vs. weighted residuals. The black dashed line is the trend line of weighted residuals distribution for the final model. The closer the trend line is to the normal distribution, the more stable the model is. (B) Quantiles of weighted residuals vs. quantiles of normal. The x-axis value of a point is the quantiles of normal and the y-axis value of a point is the quantiles of weighted residuals. The black solid line is the line of unity (the y=x line). The closer the points are to the y=x line, the more stable the model is.

Figure 4.

Prediction-corrected visual predictive check for the final model. The solid red line represents the median of the prediction-corrected concentrations of the final model. The lower and upper red dashed lines are the 2.5th and 97.5th percentiles of the prediction-corrected concentrations of the final model, respectively, representing the lower and upper limits of the 95% CI of predicted values. The blue points are observed concentrations (measured concentrations). Theoretically, the higher the number of the measured concentrations included in the 95% CI of predicted values, the better the predictability of the model. CI, confidence interval.