Objective: Develop and validate limited-sampling strategy (LSS) models for estimating the area under the plasma concentration versus time curve (AUC) of amlodipine, using data from a bioequivalence study.
Methods: Sixteen healthy volunteers received single 5-mg oral doses of amlodipine, as reference or test formulation, at a 14-day interval, in a randomized, crossover protocol. Plasma concentrations of amlodipine (n = 288), measured by mass spectrometry, were used to develop LSS models.
Results: Linear regression analysis of the AUC0-72 and a "jack-knife" validation procedure revealed that LSS models based on two sampling times (12 h and 48 h) predict accurately (R2 = 0.99; bias < 0.01%; precision = 0.03%) the AUC0-72 of amlodipine for each formulation. Validation tests indicate that the 2-point LSS model developed for the reference formulation predicts accurately (R2 > 0.90): (a) the individual AUC0-72 for the test formulation in the same group of volunteers; (b) the individual AUC0-72 for the same reference formulation in another bioequivalence study in Brazilian volunteers; (c) the average AUC0-72 reported in seven additional international studies performed under protocols similar to the present investigation; (d) the individual AUC0-72 corresponding to concentration data points provided by a first-order compartmental pharmacokinetic model, when the relative values of either the absorption rate (Kabs) or the bioavailability (F) model parameters were set at 0.85 or 0.6, of their respective original values.
Conclusions: The 2-point LSS models developed in the current study predict accurately the AUC of amlodipine under a variety of experimental conditions and, thus, may be valuable for exploring the relationships between the pharmacokinetics and pharmacodynamics of this calcium antagonist, at reduced costs of sample acquisition and analysis, and avoiding sampling at "unsociable" hours.