Evaluating Bioequivalence of Two AUCs Obtained from Sparse or Incomplete Sampling in Parallel Designs

• Activity Number: 501
• Type: Topic Contributed
• Date/Time: Wednesday, August 3, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Steven Hua*+ and Jihao Zhou
• Companies: Pfizer Inc. and Pfizer Inc.
• Keywords: Bioequivalence ; Area under the curve ; sparse sampling ; pharmacokinetics ; clinical trials

Abstract

Pharmacokinetic (PK) parameter such as AUC is frequently assessed in establishing bioequivalence (BE) of two drugs. For many PK studies, the sparse or incomplete sampling technique is used to collect data. In sparse sampling, subjects are independently assigned to one of the two drugs and each subject can be sampled only once. For incomplete sampling, subjects are sampled more than once but not at all timepoints. Although confidence interval approaches have been given greater attention, we show that decision rules can be established using equivalence test procedures for either the ratio or difference (in logarithm) of two AUCs. To control the type I error rate, the involved degrees-of-freedom (df) is adjusted using the Satterthwaite approximation for sparse sampling data, although Kenward-Roger estimate is preferable for incomplete sampling data. For incomplete sampling data, a linear mixed-effect model with repeated-measures data can be used to generate estimates for BE assessment. We also discuss power functions for sparse sampling designs. Depending on the structure of variances, unequal allocation of subjects could improve results relative to equal allocation.