Validation of a limited sampling model to determine etoposide area under the curve

To validate the utility of a previously reported 3-point limited sampling model (LSM) for determining etoposide area under the curve to infinity (AUC(infinity)).

Secondary analysis of data from two clinical trials of etoposide.

University medical center clinical research center.

Thirty-four patients with different malignancies.

Etoposide was administered as a 2-hour infusion to 34 patients. Serial plasma samples were drawn over 24 hours after the infusion and analyzed for etoposide by high-performance liquid chromatography.

The 3-point LSM AUC was compared with a 14-point actual AUC calculated by the linear trapezoidal rule. Actual and predicted AUC(infinity) by the LSM were highly correlated (r=0.97, p<0.0001). The LSM predictions had a mean absolute error of 10.9% (95% CI -14.1, -5.3) and a mean error of -9.7% (95% CI 6.9, 14.9). Nine patients with poor AUC(infinity) estimations by the LSM (error > 12%) tended to have abnormally low or high peak concentrations.

Our findings suggest the development of more robust LSM using other techniques, such as pharmacostatistical models, that can accommodate a greater degree of pharmacokinetic variability.