Statistical Considerations for Defining Cut Points and Titers in Anti-Drug Antibody (ADA) Assays

• Activity Number: 293
• Type: Topic Contributed
• Date/Time: Tuesday, August 2, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Ken Goldberg*+ and Sheng Dai and Allen Schantz
• Companies: Johnson & Johnson Pharmaceutical R&D, LLC and Johnson & Johnson Pharmaceutical R&D, LLC and Johnson & Johnson Pharmaceutical R&D, LLC
• Keywords: Immunoassay ; Anti-drug antibody ; Cut point ; Titer

Abstract

Biologic drug products can induce immune response, which in turn can adversely affect safety or efficacy so it is important to develop reliable laboratory test methods. Two ADA assays are presented. Nonlinear regression is used in each to help choose a screening or a titer cut point. The "Screening cut point" is the response above which undiluted samples are classified as reactive. Dilution of a reactive sample above its titer drops its signal below the "Titer cut point". In an RIA to detect anti-drug antibodies, our problem is to choose an adjusted signal (%binding) definition from many functions of various assay controls. The screening cut point has 5% false positive rate by definition. We choose the %binding formula with the lowest limit of detection. In a separate titration assay, the mean response for pure assay diluent is greater than for naïve negative samples. Since high titration samples approach pure diluent and high signal, the titer cut point must be greater than the screening cut point. A method for choosing the titer cut point is developed based on the titer's CV so that intra-assay variability adds negligibly to inherent uncertainty due to the titer's discreteness.