Establishment Of Population Pharmacokinetic Model Of Vancomycin In Preterm Neonates And Recommendation Of Optimal Dosing Scheme

Objective:(1)The aim of this study was to establish a population pharmacokinetics(PPK)model of vancomycin in preterm neonates;(2)Based on the established PPK model,the optimal dosing scheme was developed to reach the AUC/MIC to 400～600(assuming MIC=1 mg/L).Methods:(1)Preterm neonates(gestational age<37 weeks)who received intravenous vancomycin and had at least one blood concentration were enrolled by retrospective study method.The essential data of premature newborns were obtained from the hospital’s electronic medical record system.The PPK model was established by nonlinear mixed effect model(NONMEM).The final model was internally evaluated by Goodness of Fit Plots(GOFs),Visual Prediction test(VPC),Non-parametric Bootstrap and Normal Prediction Distribution Error(NPDE).External evaluation was conducted in a set of data that was distinct from the modeled data.The stability and predictability of the final model were investigated.(2)Monte carlo simulation method was used to simulate the optimal dosing regimen of targeting AUC/MIC to 400～600(assuming the MIC=1 mg/L)based on the established PPK model.The probability of target attainment(PTA)was calculated and the best administration regimen was developed.Results:(1)52 patients with 104 vancomycin serum concentrations were included.A PPK model of vancomycin in premature neonates was developed.The pharmacokinetic parameters were estimated by using First Order Conditional Estimation Interaction(FOCEI).The final model was a one-compartment model which was formulated as:CL(L/h/70kg)=2.63×PMA12/(PMA12+31.212)×(WT/70)0.75×e0.114;V(L/70kg)=37.1×WT/70.GOFs showed that the predicted values of the final model had a good correlation with the measured values.The condition weight residues(CWRES)were uniformly distributed symmetrically within Y=±2.In the VPC figure,more than 90%of the measured concentration points were distributed within the prediction range of 90%.Bootstrap showed that the robustness rate of the model was 98.3%,95%CI contained the target value and did not contain 0.The mean of the estimated pharmacokinetic parameters were close to the final model’s parameters.The NPDE results showed that Wilcoxon test p=0.915,Fisher test p=0.463,shapirowilks test p=0.749,comprehensive test p=1,all p values were>0.05,numerical results showed that NPDE was subject to standard normal distribution.External evaluation results showed that MAE and MPE were small,respectively 2.59 and 0.65.The MPE%and MAP%were-6%and 29%respectively.MAP%was<30%and the absolute value of MPE%was<20%.GOFs showed that the predicted values of the model were close to the measured concentration values.Both internal and external evaluations showed that the final model had good stability and predictability.(2)The model was applied to simulate the recommended administration regimen and the average PTA was 50.83%.And the average PTA was acceptable.Conclusions:(1)The main factors that affected the Clearance rate(CL)of vancomycin were postmenstrual age and weight.No factors affecting apparent volume of distribution were found.The PPK model established in this study has good predictability and stability.The model could be considered to guide the clinical application of vancomycin.(2)The recommended regimens from the simulation of PPK model were acceptable,which could provide a reference for clinical decision making.