Parallelism for a Sigmoid Curve in Assay Development

• Activity Number: 293
• Type: Topic Contributed
• Date/Time: Tuesday, August 2, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Jason Liao*+
• Companies: Teva Branded Pharmaceutical Products R&D, Inc.
• Keywords: Bioassay ; Sigmoid curve ; Transformed data ; Relative potency ; Parallelism ; Equivalence limits

Abstract

For a meaningful comparison of the potency for different biological agents/products, a fundamental assumption is the curve from the test sample needs to be parallel to the curve of the reference standard sample. The common approach for assessing parallelism is to conduct either a significance test or an equivalence test on the parameters from the sigmoid dose response curve. However, there are many drawbacks for these approaches. The significance test can reward a lousy assay but penalizes a precise assay. It is very intractable to implement the equivalence test on multiple parameters of a nonlinear curve because of the difficulty to obtain the joint confidence region and the complexity of the joint confidence region. To overcome these drawbacks, an equivalence approach directly on the response difference, rather than on the curve parameters, is proposed to assess the parallelism in this paper. The advantages of the new approach are demonstrated through simulation and real data sets.