Résumé
OBJECTIVE To establish the estimation model for the exposure of mycophenolic acid （MPA） in early renal transplant recipients [calculated by the area under the plasma concentration-time curve with 12 h （AUC0－12 h）]. METHODS Twenty kidney transplant recipients， who received triple immunosuppressive therapy of mycophenolate mofetil （MMF）+tacrolimus+ methylprednisolone， were selected and given MMF dispersible tablets （750 mg， q12 h） on the 15th day after the operation； the blood samples were collected from the patients before and 0.5， 1.0， 1.5， 2.0， 3.0， 4.0， 6.0， 8.0， 12.0 hours after the administration， respectively. The blood concentration of MPA was determined， and the pharmacokinetic parameters of MPA were calculated. The multivariate linear stepwise regression analysis method was used to fit an estimation formula for the finite sampling method suitable for MPA-AUC0－12 h of the recipients. Bland-Altman analysis was used to evaluate the agreement between the estimation formula and the classical pharmacokinetic method. RESULTS The main pharmacokinetic parameters of MPA in 20 renal transplant recipients： c0 was （1.53±0.84） μg/mL， cmax was （12.07±5.97） μg/mL， t1/2 was （5.41±3.67） h， tmax was （1.58±0.75） h， and the average AUC0－12 h calculated by the classical pharmacokinetic method was （33.95±13.40） μg·h/mL. MPA-AUC0－12 h was estimated with sampling points of “4.0， 8.0， 12.0 h”； the simplified calculation formula was AUC0－12 h＝12.058+2.819c4.0+7.045c8.0+ 3.879c12.0 （R 2＝0.934）. The predicted value had a good correlation and consistency with the measured value， and 95.0% of predicted values did not exceed the x±1.96SD （standard deviation） range. CONCLUSIONS The estimation model is established successfully for the exposure of MPA in early renal transplant recipients； the model has better prediction accuracy and fewer sampling points.
Résumé
AIM: To establish an ultra high performance liquid chromatography (UPLC) method for the determination of micafungin in plasma of critically ill patients. And to establish a model for estimating the area under the concentration-time curve (AUC) of micafungin by limited sampling strategy. METHODS: Patients with severe infection were administrated with micafungin once a day, 1 h for each infusion. The blood samples were collected before administration and 1, 2, 4, 8, 12, 24 h after administration and were measured by UPLC. The pharmacokinetic parameters were calculated by Phoenix winnonlin 6.4, and the drug concentrations at 2-4 blood collection points were analyzed with SPSS 22.0 to establish limited sampling models. RESULTS: The calibration curve was linear over a concentration range of 1.0 to 50 μg/mL (r
Résumé
Combining classical pharmacokinetic principles with statistical models, population pharmacokinetics (popPK) can effectively utilize sparse data for pharmacokinetic analysis. An optimally designed population pharmacokinetic study will balance the efficiency of a popPK study and the precision with which the parameters are estimated to ensure the unbiased estimation of pharmacokinetic parameters and facilitate the development of clinical and non-clinical trials. Sparse sampling methods have been developed for designing population pharmacokinetic experiments including random sampling method, limited sampling strategy, maximum a posteriori Bayesian method, Fisher information matrix method, and informative block randomized design, which have been widely applied in the uni-response and multi-response popPK sampling optimization. In recent years, population pharmacokinetics has been developed rapidly in Chinese materia medica (CMM), but few studies have been conducted to optimize sampling. By comparing the advantages and disadvantages of each sparse point sampling optimization method and the applicable conditions, this work provides a comparative review of optimal design methodologies and gives its application examples, which provides a reference for pharmacokinetic sampling optimization of CMM.
Résumé
OBJECTIVE: To establish limited sampling strategy to estimate area under the curve (AUC) of HSCT patients who had received busulfan (Bu) intravenous infusion. METHODS: Plasma samples were collected at certain time points from 22 HSCT patients who had received busulfan intravenous infusion. The LC-MS/MS method was used to measure blood concentration of busulfan. The classical method was used to calculate pharmacokinetic parameters. To estimate AUC0-6, measured blood concentration data was used to build a multiple linear regression model which was subsequently validated. Consistency between results produced by the traditional method and the proposed LSS respectively was assessed by the intraclass correlation coefficient (ICC) and Bland-Altman (BA) analysis. RESULTS: It is shown that model based on concentrations on two sampling time points (2 and 5 h) is able to predict AUC0-6 accurately (adjusted r2=0.917. MPE=0.2%, RMSE=4.48%). The 95% confidence interval of ICC is 0.914-0.984 while the limit of agreement in the BA is -0.9-8.8. Results produced by limited sampling strategy are nearly consistent with AUC0-6 produced by the classical method. CONCLUSION: The proposed limited sampling strategy to estimate AUC0-6 based on C2 and C5 can be used to busulfan blood concentration monitoring in clinical practice.
Résumé
OBJECTIVE: To establish limited sampling strategies for the estimation of exposure (AUC) to mycophenolic acid (MPA) in patients with autoimmune disease (AID). METHODS: The 24 AID patients were treated with MMF 0.75 g q12 h to steady state. Serum MPA concentrations were measured at pre-dose, 0.5,1,1.5,2,4,6,8 and 12 h after dosage and MPA AUC0-12 h were caculated with the trapezoidal rule. The resulted data were randomly devided into index set and validation set. Multiple liner regression analysis of the index set was used to determine the 1 to 4 sampling estimation models and the models were validated in the validation set. The prediction bias and precision combining clinical feasibility were used to choose the final models. RESULTS: The 1-point model based on the simple measurement of c12 h produced acceptable prediction for MPA AUC0-12 h, in which, r was 0.957, the prediction bias and imprecision were repectively-3.93 and 11.93, both <15%; 4-point model based on 0.5,1.5,6,8 h or 1.5,6,8,12 h sampling times were fitted to MPA AUC0-12 h, with highest correlation coefficients of 0.996, and prediction errors between ±15% for 9 patients; other 2 or 3-point models resulted somewhere in-between prediction performance. CONCLUSION: Considering the clinical feasibility, the accurate 4-point model AUC=3.19+0.49c0.5 h+1.76c1.5 h+2.95c6 h+5.46c8 h is applicable to MPA AUC monitoring in AID inpatients after morning dose; the simple 1-point model AUC=10.82+13.37c12 h is a good choice for AID outpatients after last-night dose. Bland-Altman analysis reveals that both the two models could effectively predict MPA AUC in Chinese adult AID patients receiving MMF concomitant methylprednisolone therapy.