Establishment,Evaluation And Application Of PPK Model Of Vancomycin In Critically Ill Patients

ObjectiveTaking vancomycin as an example,aiming at the common problems of antibacterial drug management.To establish a HPLC method for the detection of vancomycin concentration in serum of severe patients.To summarize and analyze the rate of reaching the standard of vancomycin concentration in critically ill patients.To evaluate the applicability of the population pharmacokinetics（PPK）model commonly used in China in the Intensive Care Unit（ICU）patient population.To establish a PPK model of vancomycin in critically ill patients and verify its predictive power.To establish a new model of vancomycin TDM and individualized drug administration in severe patients,improve the standard rate of vancomycin treatment and reduce the incidence of vancomycin related adverse reactions.Methods1.HPLC was used to detect the concentration of vancomycin in serum.The specificity,precision,recovery rate and stability of vancomycin were investigated.2.To concentrate vancomycin serum of ICU patients who receiving vancomycin from January 2019 to December 2020,and the concentration distribution and standard of vancomycin were analyzed.To evaluate the predictive performance of 4 common individualized dosing aid decision making systems in China in severely ill patients,they are Smart Dose,Pharm Van,Vancomycin dose recommendation and blood concentration prediction system,and Java PK?for Desktop（JPKD）.3.Nonlinear mixed effect model（NONMEN）is used to construct vancomycin PPK model.Using goodness of fit plots（GOFs）,non-parametric Bootstrap,prediction-corrected visual predictive check（pc-VPC）to conduct internal verification of the PPK_ICU,external verification using the prediction error test method.The external validation results of the PPK_ICU were compared with those of Smart Dose,Pharm Van,Vancomycin dose recommendation system and blood concentration prediction system and JPKD to evaluate the prediction stability and accuracy of the PPK_ICU.Finally,the established PPK model was used to adjust the administration schedule of vancomycin in severe patients.Results1.In this study,an HPLC method was developed for the determination of drug content in serum of critically ill patients.The detection limit of vancomycin was 0.50μg·mL^-1,the limit of quantilation was 1.00μg·mL^-1.When the concentration of vancomycin in serum was 1.56～100.00μg·mL^-1,and the linear relationship between concentration and peak area was good.RSD of intra-day and intra-day precision were lower than 15%,and the recoveries were between 97.81%and 100.64%.RSD of all the samples were lower than 15%and stable at room temperature within 72 h.RSD of all the samples were lower than 15%after five long freeze-thaw cycles.2.This study included a total of 87 blood concentration data of 49 patients treated with vancomycin in ICU from January 2019 to December 2020.The mean measured steady-state serum grain concentration(C_min)was（21.03±21.45）μg·mL^-1.Only 32patients（36.78%）achieved the target valley concentration of 10～20μg·mL^-1,33patients（37.93%）had C_min>20μg·mL^-1,and 22 patients（25.28%）had C_min<10μg·mL^-1.In severe patients,the compliance rate of vancomycin was not high,and the concentration of most drugs was higher than the normal value in patients who did not reach the standard.According to the current data set,the prediction performance of four models commonly used in China was evaluated.In the evaluation based on population prediction,the median value of absolute prediction error of all models was more than 30%,and the prediction performance of severe patients was poor.To analyse the reasons that may affect the prediction error of the four models are analyzed,the research shows that there is no specific reason affecting the prediction accuracy of the four models.3.The PPK model of critically ill patients was established with 87 blood concentration data sets of 49 patients receiving vancomycin in ICU from January 2019to December 2020.The one-compartment model of primary elimination was determined as the basic model,and the PPK_ICU was obtained by including creatinineclearance（CLcr）as the covariable.Three internal verification methods have confirmed that the model has good predictive performance and is more suitable for severe patients.External validation included 13 patients treated with vancomycin in ICU from January 2021 to December 2021,with 16 samples of serum trough concentration data.The median PE%values of individual and population in the final model were less than±20%,indicating that the prediction accuracy of the final model was good.The median APE%of individuals and the median APE%of groups were both below 30%,so it was considered that the PPK_ICUpredicted well.ConclusionsThe HPLC method established in this study is accurate and reliable for the determination of vancomycin in severe patients,and is suitable for the determination of vancomycin in severe patients.The available free Vancomycin individualized assisted decision model has poor extrapolation to our data set.In this study,a PPK model was established in severe patients,and its prediction accuracy and stability were verified by internal and external verification.