In vitro evaluation of a continuous-sampling device for pharmacokinetic parameter estimation

Using an in vitro pharmacokinetic model, area under the curve (AUC) estimates from an osmotic continuous-withdrawal device were compared to AUC estimates by a conventional trapezoidal method. Ten experiments were done under two different conditions: (a) half-life of 1 h (n = 5) and (b) half-life of 2 h (n = 5). Sampling was done for 6 and 12 h, respectively. The AUC estimates from the two methods were highly correlated (r = 0.948, p < 0.0001). The mean coefficient of variation was 9.5% (n = 10), and mean AUC values were not statistically different by analysis of variance (p = 0.4). The sampling device sampled at 0.121 +/- 0.011 ml/h; the mean volume was 0.79 +/- 0.03 and 1.34 +/- 0.02 ml over 6 and 12 h, respectively. Continuous sampling provided a reasonable estimation of AUC, required only one analytical sample, and sampled at a consistent zero-order rate.