论文部分内容阅读
目的:用Monte Carlo算法编制群体药动学分析程序并认证该方法估计药动学参数和预测血药浓度的能力.方法:用阿米卡星作为模型药物,对来自42名新生儿共142对血药浓度时间数据进行分析;根据Sheiner等提出的群体药动学思想,我们编制了估计群体参数和个体参数的程序,目标函数最小值以Monte Carlo算法求得,方法的认证采用经典药动学 程序3p87作为对照,预测能力通过计算预测血药浓度的均方根误差(RMSD)和偏性(BIAS)来考察.结果:我们自编的程序运行稳定;本法提取的群体参数与3p87得到的一致,学习样本与认证样本的预测浓度与实测浓度显著相关(相关系数分别为0.995和0.990),预测误差大多数小于1 mg/L,认证样本RMSD和BIAS分别为0.58和-0.07 mg/L.结论:本法估计参数准确,预测血药浓度能力令人满意.
OBJECTIVE: To develop a population pharmacokinetic analysis program using Monte Carlo algorithm and to validate the ability of the method to estimate pharmacokinetic parameters and predict plasma concentration.Methods: Amikacin was used as a model drug in a total of 142 pairs of 42 neonates According to the group pharmacokinetic theory proposed by Sheiner et al., We developed a procedure for estimating population parameters and individual parameters. The minimum value of the objective function was obtained by Monte Carlo algorithm. The method was certified by classical pharmacokinetics Using program 3p87 as a control, the predictive power was assessed by calculating the root mean square error (RMSD) and bias (BIAS) of the predicted plasma concentrations.Results: Our self-compiled program ran stably.The population parameters extracted from this method were similar to those obtained from 3p87 The predicted concentrations of learning samples and certified samples were significantly correlated with the measured concentrations (correlation coefficients of 0.995 and 0.990, respectively). Most of the prediction errors were less than 1 mg / L. The certified samples RMSD and BIAS were 0.58 and -0.07 mg / L, respectively. Conclusion: The method is accurate in estimating the parameters and predicting the plasma concentration is satisfactory.