Paper by European Regulators Foresees "Continuing Reduction in Clinical Data Requirements" for Biosimilars

September 24, 2019
Kelly Davio

This month, Elena Wolff-Holz, MD, PhD, chair of the European Medicines Agency’s (EMA’s) Biosimilar Medicinal Products Working Party, together with coauthors from the EMA and the Federal Institute for Drugs and Medical Devices in Germany, published a new paper in which the team outlines the extent of clinical confirmation of biosimilarity, taken together with analytical and functional data, that is considered necessary for biosimilar drugs to be approved in Europe.

In recent months, the discussion about whether the biosimilar development paradigm can be made more efficient has gained new strength, with the publication of papers that argue that some studies generally undertaken to facilitate biosimilar approval are unnecessary or even unethical in nature.

This month, Elena Wolff-Holz, MD, PhD, chair of the European Medicines Agency’s (EMA’s) Biosimilar Medicinal Products Working Party, together with coauthors from the EMA and the Federal Institute for Drugs and Medical Devices in Germany, published a new paper in BioDrugs in which the team outlines the extent of clinical confirmation of biosimilarity, taken together with analytical and functional data, that is considered necessary for biosimilar drugs to be approved in Europe.

Up to July 2019, a total of 91 marketing authorization applications for biosimilars (including duplicate authorizations) have been submitted to the EMA, write the authors, and 61 products have been given positive opinions from the Committee for Medicinal Products for Human Use and have been approved by the European Commission. The foundation of demonstrating biosimilarity to the EMA, the authors note, is analytical and functional comparison of the biosimilar and its reference, followed by confirmation of biosimilarity using clinical data.

Generally, the authors write, 1 confirmatory randomized controlled trial will be conducted to assess efficacy, safety, and immunogenicity, though in orphan indications in which such a trial might not be feasible, exceptions may exist. They note, however, that analytical and functional in vitro assays are “generally more specific and sensitive than studies in humans” at detecting potential differences.

The extent of clinical work needed varies by molecule; for example, biosimilar teriparatide, a single-chain, nonglycosylated protein, does not require an efficacy or safety trial, while biosimilars of infliximab, adalimumab, trastuzumab, and bevacizumab, which are more complex, require safety and efficacy trials in sensitive populations.

Comparative pharmacokinetic (PK) studies, however, remain a bedrock of development, and in the presence of suitable pharmacodynamic (PD) endpoints, a PK/PD study “may be sufficient clinical work for marketing approval.” The authors point out that, in recent years, insulins, pegfilgrastim, and low—molecular-mass heparins have derived their pivotal evidence from PK and PD studies, and comparative efficacy trials were not deemed necessary. More complex products will still require comparative safety and efficacy trials, however, though “there should always be a sound scientific rationale for conducting clinical studies.”

With respect to efficacy end points, clinical end points should be sensitive to potential clinically relevant differences, and hard clinical end points—like overall survival—are “rather insensitive in this respect,” the authors say. Thus, PD endpoints are preferred to establish similar efficacy. For smaller proteins, PD parameters—like absolute neutrophil count for granulocyte-colony stimulating factor therapies or serum calcium levels for teriparatide—are preferred. For larger molecules, such as eculizumab, a potential PD marker could be serum lactate dehydrogenase level.

According to EU guidelines, comparability margins should reflect the largest difference in efficacy that would not matter in clinical practice; acceptable equivalence margins will depend on patient populations, end points, concomitant therapy, and estimated treatment effect; population differences should be considered when planning these margins.

The paper also provides case studies that show how the totality-of-the-evidence approach to biosimilar development is currently handled by regulators:

  • For infliximab, a biosimilar showed a small difference in its proportion of afucosylated forms. The difference was shown not to be clinically relevant in a large PK study in patients with ankylosing spondylitis, as well as in a phase 3 study in patients with rheumatoid arthritis. Also, for 1 biosimilar infliximab, there were numerical differences in anti-drug antibody (ADA) rates between the biosimilar and its reference. However, the numerically higher incidence of ADAs did not have a relevant impact on efficacy or safety.
  • In the cases of 2 adalimumab products, initial PK studies failed. In one case, a second, improved PK study succeeded, and in the other, an adapted study design to reduce intrasubject variability was performed, and it too demonstrated PK similarity.
  • Teriparatide’s clinical development program comprised 1 single-dose comparative PK study in 54 healthy volunteers. Mean exposure and peak serum concentrations were approximately 8% lower for the biosimilar than the reference, but the clinical impact was considered negligible.
  • Pegfilgrastim has notably high PK variability, but not high PD variability, making PK end points more sensitive to detect potential differences. Two of 6 marketing applications for biosimilars had failed PK trials, and both of these products later presented new PK studies and justification as to why the new studies should outweigh the results of the failed studies.
  • For one insulin glargine, subjects with low glucose requirements were excluded, and when they were included, PD similarity was not shown. However, biosimilarity was concluded on the basis of the totality of the evidence.
  • In the case of trastuzumab, 2 biosimilars did not formally meet the upper bound of the prespecified equivalence margin, but these programs were at least partially confounded by a shift in a critical quality attribute of the reference.
  • Rituximab, too, was subject to shifts in some critical quality attributes. Clinical comparability could be demonstrated, however, on the basis of PK trials in healthy volunteers together with efficacy trials and a PK bridging study.
  • Finally, for etanercept, a biosimilar had a lower incidence of ADAs than the reference, but the clinical impact of this difference was deemed negligible.

The authors write that guidance documents that put forth the requirements for the development and licensing of biosimilars must be considered living documents, and should be revised and updated as needed. They conclude that, as experience and analytical capabilities mature, and as understanding of structure—function relationships progresses, “a continuing reduction in clinical data requirements for biosimilar developments can be foreseen.”

Reference

Wolff-Holz E, Tiitso K, Vleminckx C, Weise M. Evolution of the EU biosimilar framework: past and future [published online September 20, 2019]. BioDrugs. doi: 10.1007/s40259-019-00377-y.