ABSTRACT
OBJECTIVE: To establish a population pharmacokinetics(PPK) model of vancomycin in adult patients and investigate the factors influencing vancomycin clearance.METHODS: The nonlinear mixed-effect model(NONMEM) was used to investigate the population characteristics of vancomycin in adult patients and the serum cystatin C was used as a marker of renal function. The final model was built by forward inclusion approach and backward elimination method. Fitting effect of the model was evaluated by the goodness of fit plots(GOFs). Nonparametric Bootstraps and normalized prediction distribution errors(NPDE) were performed to evaluate the robustness and predictive efficacy of the final model. External model evaluation was conducted using an independent dataset to evaluate the model predictability. RESULTS: Vancomycin PPK model was set up via 147 serum trough concentration data from 95 adult patients. The estimated population typical values of clearance rate and apparent volume of distribution were 3.57 L·h-1 and 63.30 L, respectively. The main factor influencing clearance was renal function. The GOFs showed that the final model was stable and effective, and the fitting degree of the final model was better than that of the base model. The robust rate verified by Bootstrap was 99.45%. All of the relative biases between the median of parameters validated by Bootstrap and the estimated parameters of final model were within ±3%, and the 95% confidence intervals of these validated parameters did not include zero. The NPDE followed the N(0,1) distribution with a global adjust P value of 0.334, which indicated that the model had a high predictive accuracy. External evaluation was performed via an independent dataset of 40 concentration data from 20 patients. The mean prediction error(MPE) and mean absolute prediction error(MAPE) based on population predictions(PRED) was -1.90% and 24.34%, respectively. CONCLUSION: Vancomycin PPK model established in the study is of as a good stability and high predictive accuracy, as a reference for developing individualized administration regimens.