The Clinical Trials Landscape Is Evolving in Biosimilar Development

Among the hottest topics in biosimilar development is the future of clinical trials, and that fact was evident during the opening session of the fifth DIA Biosimilars Conference, held October 22 to 23 in London, United Kingdom. During the presentation, a review of the biosimilar regulatory framework, 4 experts explained the ways in which the landscape for trials is evolving for biosimilars.

Alex Kudrin, MD, MBA, PhD, an independent consultant who works with such biosimilar developers as Celltrion and Fresenius Kabi, started his presentation by reviewing the well-known pyramid diagram that depicts the biosimilar development process with a large analytical base, followed by nonclinical data, pharmacokinetic (PK) and pharmacodynamic (PD) data, and finally by clinical data at the apex. In the current context, Kudrin said, “We’re not feeling that comfortable on top of that pyramid.”

Reasons for discouragement with the current paradigm, he said, hinge on the time and cost of clinical programs. According to his data, a 600-patient phase 3 clinical study can cost $50 million, though costs can quickly exceed that threshold; several biosimilar trastuzumab studies, for example, have enrolled more than 800 patients each.

Studies may also extend for a long period; Kudrin noted that Sandoz’s ASSIST-FL trial for biosimilar rituximab ran for 7 years. “Why does it take us 10 years to develop a biosimilar? That clearly is not a good situation,” he said.

Furthermore, “Sometimes the endpoints we use are not perfect. Yes, they are valuable from a marketing perspective,” he said, but they do not typically generate useful additional information. Kudrin used the example of 2 trastuzumab biosimilars approved in the European Union; both ABP-980 and SB3 failed to show equivalence with the reference Herceptin based on clinical studies that used the European Medicines Agency (EMA)'s recommended equivalence margin, but both products gained approval based on their analytical characteristics.

Such cases raise the question of whether analytical data could be adequate for the purposes of regulatory approval. Kudrin suggested that if a developer is working on a first-in-class product, it is unlikely that a regulatory body will approve the product without a clinical study, but a third or fourth product in its class (such as another trastuzumab or anti—tumor necrosis factor therapy) could potentially be approved under a tailored path that does not require the use of phase 3 studies.

Tools that could be used in such a tailored path, one that uses a stepwise approach to determining the need for clinical trials, include refined analytical similarity studies, in silico methods for assessing immunogenicity, and studies designed to show PK similarity and a descriptive assessment of immunogenicity and safety in healthy volunteers or patients.

Christian Maasch, managing consultant at Xendo, also addressed the question of how far analytical data can take a biosimilar development program, and went a step farther. In the current context, he said, the science of development has advanced considerably, and is even outpacing regulatory knowledge. According to Maasch, development may have reached a stage at which even nonclinical in vivo studies may no longer be useful, and may only be providing a “comfort factor.”

Maasch explained that, in 2014, a working group from the United Kingdom’s Medicines and Healthcare Products Regulatory Agency assessed the available data and found that there was no case in which a nonclinical in vivo study provided useful information to the development program for a monoclonal antibody. With those findings in mind, the European Union has, he said, been promoting initiation of clinical trials based on in vitro data only (the FDA typically requires at least 1 nonclinical in vivo study, however).

Acknowledging that clinical investigators may be hesitant to undertake a phase 3 study without in vivo data, Maasch said the EMA’s pilot program for providing tailored scientific advice to developers could help.

However, Inger Mollerup, MS, MSc, an external consultant for Novo Nordisk, explained that in the world of follow-on drugs, clinical trials may be indispensable.

In the European Union, some peptides and smaller proteins can be approved after development as biologics (resulting in biosimilars) or development through chemical synthesis (resulting in new drugs, hybrids, or generics).

Even when chemically synthesized, such peptides and smaller proteins have the potential for immunogenicity, similar to larger proteins. For example, synthetic liraglutide contains potential T-cell epitopes. Furthermore, immunogenicity can be impacted by different aspects of manufacturing, including container and closer, impurities, and fibrillation and aggregation. Crucially, “A chemically synthesized protein using a biological medicinal product as its reference product is not the same active substance as the reference product.”

Thus, these follow-on drugs need clinical studies to confirm that there are “no surprises” between a synthetically synthesized follow-on versus a biological reference.

Speaking from the regulator’s perspective, Thomas Lang, MSc, senior statistical assessor for the Austrian Agency for Health and Food Safety, said that he was first approached by a biosimilar developer asking for scientific advice on how to complete a development program without phase 3 clinical trials evidence. At the time, he said, “We are not fit to answer that question.”

For Lang, the question that needed to be answered was, “What would be the evidence for us as a decision maker?”

As a result, the EMA undertook a draft of a reflection paper on the topic that could serve as a platform for interaction between quality experts and statisticians. The paper was later adopted by the EMA’s Committee for Medicinal Products for Human Use in 2017 and discussed in a comment period and an EMA workshop in 2018.

Among submitted comments, he noted, approximately 60% of the stakeholders said that statistical testing should not become a pass-or-fail criterion without reflection on context or without the involvement of chemistry, manufacturing, and control experts. Furthermore, stakeholders were concerned that reliance on analytical data could endanger the totality-of-the-evidence paradigm for development.

One of the learnings from the paper were that if stakeholders want to see a leaner development process, there will be a need for adequate control of the risks for decisions made based on false positives or false negatives in analytical testing.

The reflection paper will be revised, he said, taking into consideration comments and the proceedings of the workshop, but the EMA’s upcoming move to Amsterdam has put this work on hold, he noted.