Effect of the number of samples on Bayesian and non-linear least-squares individualization: a study of cyclosporin treatment of haematological patients with multidrug resistance

Evaluation of Bayesian Estimation of Pharmacokinetic Parameters

The validity of pharmacokinetic parameters estimated by the maximum a posteriori probability (MAP) Bayesian method was investigated by simulation studies. A 1-compartment model with bolus intravenous administration was used as a pharmacokinetic model, and the coefficients of variation for the parameters and residual error were set at 30% and 10%, respectively. The accuracy of the posterior modes of pharmacokinetic parameters estimated by the MAP Bayesian method was assessed by the difference between the true value and the estimated value. The results showed that the accuracy of the Bayesian estimation depended on sampling times and on the differences between the prior means and individual true parameter values. For assessing the reliability and accuracy of the Bayesian estimation, the authors suggest using the whole posterior distribution of the pharmacokinetic parameters to describe the 95th percentile range for predicted blood concentration profiles. The authors believe that the proposed procedures provide helpful information for evaluating the Bayesian estimation of pharmacokinetic profiles.

Limited sampling strategies for estimating cyclosporin area under the concentration-time curve: review of current algorithms

Cyclosporin, the drug of first choice in transplantation surgery, is characterized by a low therapeutic index and variable absorption, so close monitoring of the drug is required to optimize the dosing. Predose blood cyclosporin levels are measured routinely for therapeutic monitoring, but this approach is not optimal because the area under the concentration-time curve (AUC) correlates better with clinical events. However, conventional methods of measuring AUC require many blood samples, which is not viable in a routine clinical setting. AUC monitoring can be simplified for use in a clinical setting by using a limited sampling strategy (LSS) that allows AUC to be estimated using a small number of blood samples collected at specific times. This article reviews the current literature on estimating cyclosporin AUC using LSS. Thirty-eight papers suggesting the use of specific time points were found. LSS has been developed for different transplant types, with different dosing regimens, and with different assays. Most authors suggested either two- or three-sample equations. Results from authors who validated their models suggest that equations defined on one transplant type may be applicable to other transplant types, to both adults and children, and to early or late after transplantation. Moreover, it seems that there is flexibility in the choice of equations available to clinicians. The number of samples to collect for accurate estimations is a matter of debate, but a wise choice can minimize the number. The choice of the optimal LSS and validation are discussed.

Methods for clinical monitoring of cyclosporin in transplant patients

Cyclosporin was introduced into clinical practice in the early 1980s and has since been shown to prolong survival for transplant recipients. Because cyclosporin is a narrow therapeutic index drug and there are significant consequences associated with 'subtherapeutic' and 'supratherapeutic' concentrations, cyclosporin therapy is monitored as part of routine patient follow-up. However, the optimal method for the therapeutic drug monitoring of cyclosporin has yet to be defined. Currently, the most common method involves monitoring pre-dose trough concentrations, but this method is less than ideal. Other methods of monitoring cyclosporin therapy include monitoring the area under the concentration-time curve, limited sampling strategies, monitoring of single concentrations other than troughs and pharmacodynamic monitoring. Bayesian forecasting has been used successfully in clinical practice with other drugs with narrow therapeutic indices. However, few studies are available regarding Bayesian forecasting and cyclosporin. Existing studies are preliminary in nature and involve the old Sandimmun formulation rather than the Neoral formulation. Although these methods show promise, they have not gained widespread acceptance. This is because of their impracticality and the lack of prospective studies comparing other monitoring methods with trough concentration monitoring. Further comparative studies evaluating the impact of the specific monitoring method on definite patient outcomes are warranted.

Prediction of serum vancomycin concentrations using one-, two- and three-compartment models with implemented population pharmacokinetic parameters and with the Bayesian method

Although previous studies have shown that vancomycin has a complicated pharmacokinetic profile requiring description using a two- or, better, three-compartment model, until recently predictions of serum vancomycin concentrations have been mainly based on one- or two-compartment models using computer software packages. In this study, we have predicted serum vancomycin concentrations in 59 patients using one-, two- and three-compartment models with implemented population pharmacokinetic parameters in the Abbott PKS program and by use of the Bayesian method. The percentage errors of predictions made using the one-compartment model were smaller when either the Bayesian method or implemented population pharmacokinetic parameters were used (medians of -8.61% and -9.49%, respectively). Predictions using the one-compartment model with the Bayesian method were less biased (median of -1.52 microgmL(-1). The best predictions were those made using the three-compartment model with the Bayesian method-they were most accurate (median of 3.40 microgmL(-1) and highly precise (median of 11.53 microg(2)mL(-1)). The results suggest that predictions made using the one-compartment model with implemented population pharmacokinetic parameters are preferable if no samples are available, otherwise predictions made using the three-compartment model with the Bayesian method are preferable. The results also supported our previous argument that the greater the number of compartments involved in individualization, the better the predictions obtained using the Bayesian method.