Setting specifications for potency assays--basic principles

To develop any biological therapeutic successfully, it is necessary to characterise the product thoroughly, both physicochemically and biologically. To ensure that consistency of production is maintained, some level of control is required to limit the variability of the product from batch to batch. This premise forms the basis of specification setting. The biological activity or potency of the product must be appropriately assessed through a functional assay unless specifically justified otherwise e.g. for binding proteins where binding has been correlated with biological activity. Biological assays are particularly prone to assay variability and therefore it is necessary to design and execute the assay to reduce variability as much as possible while providing a statistically valid measure of the reproducibility of potency estimates. Therefore, specifications for potency should only be derived following establishment of a well validated and controlled bioassay and should include some measure of the variability of the estimate in addition to the actual potency estimate itself. The limits applied to those specifications need to reflect the true batch to batch consistency of the product, its nature and toxicity as well as its intended use.