From the desk of Henry Leng
Extrapolation, when used in the context of biosimilars, means their registration for all the indications of the reference product without the need to conduct clinical trials to support each of those indications. The assumption that is made by most regulatory specialists, whether from industry or the regulatory authority, is that extrapolation to other indications of the reference product depends on the demonstration of therapeutic equivalence (or non-inferiority) to the reference for one indication in a comparative clinical trial. If this can be shown, and provided that other criteria, such as the same mechanism of action or the same receptors, are involved in the other indications, then, and only then, will extrapolation be allowed. This is a false assumption.
The concept of “extrapolation” is common in statistics, and we use it quite often when we want to propose (Applicant) or allocate (Regulator) a (provisional) shelf life to a biologic drug substance or final product. In this instance, we use quantitative data generated over a limited time (12 months) at a specific storage condition (usually 2 – 8 °C) for critical quality attributes (CQA) such as potency or impurities (e.g., aggregates) and then use a statistical tool such as linear regression to predict future trends in the values for those CQAs. Thus, extrapolation is limited to forecasting future trends in data based on current data for a specific CQA. We do not extrapolate known data for one CQA such as potency to predict performance or outcomes in another CQA, such as percentage aggregates. This is in fact what we attempt to do when we extrapolate findings for one indication of the reference with the biosimilar, to all other registered indications of the reference product.
The most important criticism against extrapolation from one indication to other indications of the reference, is that the comparative study was not designed to demonstrate comparative safety and efficacy of the candidate biosimilar to the reference product, but to CONFIRM biosimilarity, which was already demonstrated through comparative physicochemical, structural and in vitro biological assays. Thus, “extrapolation” for want of a better term, should, and is based, on the knowledge that the candidate biosimilar is similar to the reference product in all critical quality attributes. Once biosimilarity has been established between the candidate biosimilar and the reference product, then extrapolation can be done from the reference product to the candidate biosimilar product – not from one indication to other indications. What the biosimilarity exercise ultimately shows is that the drug substance of a biosimilar product is just another variant of the drug substance of the reference product (Krendyukov and Schiestl, 2018). In other words, the drug substance of the candidate biosimilar will be indistinguishable from the drug substance of the reference product. We all accept that biologics are subject to batch-to-batch variability.
Extrapolation as explained above is not an unusual concept in the biopharmaceutical industry. Manufacturing changes to biological medicines occur several times during their life cycle and the regulator seldom, if ever, requires from an innovator manufacturer to repeat all the clinical studies to confirm that the post-manufacturing changed product performs the same as the pre-changed one. What is required is a comprehensive physicochemical and biological comparability exercise to demonstrate that the product after the change is essentially the same as before the change, i.e., that its CQAs remain within registered acceptable limits. I recently reviewed such an application where the applicant changed the product from a freeze-dried to a liquid formulation. The applicant performed comprehensive physicochemical and biological comparability studies between the old and new formulations and supplemented their findings of equivalence on those levels with a comparative biostudy that showed equivalent pharmacokinetic profiles as well. I could see no reason why SAHPRA would not approve this application. The EMA did.
References
Krendyukov A, Schiestl M. Extrapolation concept at work with biosimilar: a decade of experience in oncology. Published February 6, 2018. Br Med J. doi:10.1136/ esmoopen-2017-000319.