• C Balram, C Sivathasan, Yin Bun Cheung, Say Beng Tan, and Yong Seng Tan.
• Clinical Pharmacology Unit, Division of Clinical Trials and Epidemiological Sciences, National Cancer Centre, Singapore. ctebal@nccs.com.sg
• J. Heart Lung Transplant. 2002 Sep 1;21(9):1016-21.

BackgroundTherapeutic drug monitoring of cyclosporine in heart transplant patients is used to monitor therapy and prevent rejection. Of the various methods available for performing therapeutic drug monitoring of cyclosporine, the method of limited sampling strategy for area under the concentration-time curve profiling has been used most widely recently. The process of identifying sparse data points to predict area under the concentration-time curve is essentially a variable selection problem, with the variables being the drug concentrations at the various timepoints. Although fitting more variables into a model will typically allow for a better prediction of area under the concentration-time curve, improving the prediction has to be traded-off against the desirability of using as few timepoints as possible. The objective of this study was thus to formulate a model that would provide a good prediction of area under the concentration-time curve based on a limited number of sampling points.MethodsWe studied 15 stable heart transplant patients (11 Chinese and 4 Indians). All patients were receiving Neoral-based immunosuppression. Whole blood samples for area under the concentration-time curve analysis were obtained at the following timepoints: pre-dose (C(0h)) and at 1, 2, 3, 4, 6 and 12 hours (C(1h), C(2h), C(3h), C(4h), C(6h), C(12h), respectively) post-dose during the first dosing interval. The linear trapezoidal rule was used to calculate the area under the concentration-time curve (AUC) from time 0 h to 12 h. Various limited sampling strategies, as well as Keown's formula, which was derived in renal transplant patients and used C(0h) and C(2h), were compared based on their capacity for reducing total error squared.ResultsC(4h) was found to be the single most predictive timepoint and explained 95.3% of AUC(0-12) variation. C(0h) and C(12h) explained 60% and 75.7% of the variation in AUC(0-12), respectively. The best 2-variable model identified by stepwise selection procedures included C(1h) and C(4h) as predictors, explaining 97.3% of the variation in total area under the concentration-time curve from time 0 h to 12 h. Using Keown's algorithm, the R(2) was only 80.9%.ConclusionWe recommend using C(1h) and C(4h) as surrogate markers of area under the concentration-time curve from time 0 h to 12 h in our heart transplant patients. Because C(1h) and C(4h) represent timepoints within the zone of highest variability for Neoral's absorption phase, a model incorporating these timepoints would be able to explain a greater degree of variability associated with the Neoral absorption profile.

21 keywords...

hide…