Biological medicines are complex and consequently do not easily lend themselves to the development of generic versions, such as is the case for small molecules. Most biologics are proteins range from 30-fold for small proteins such as insulin to over a 1000-fold the mass of, for example, aspirin with associated orders of magnitude increase in complexity. Further, proteins exist as many different variants due to modifications to their amino acid side chains. It has been calculated that for a monoclonal antibody, up to 100 million variants could potentially exist. All this makes for complex challenges when comparing the structure of a follow-on protein medicine or biosimilar to its reference product. In fact, one batch of the reference product will differ from the next, and sometimes the difference can be quite significant following a change in the manufacturing process.
Nevertheless, over the past two decades of biosimilars development, there has been a vast improvement in assay technology and a greater understanding of the impact of variants on efficacy and safety. Thus physico-chemical and biological testing data can go a long way to demonstrating that a follow-on biological product is similar to its reference product.
The next step in the development of a biosimilar is to show pharmacokinetic equivalence to the reference product; this can often be done in healthy volunteer studies. These studies provide an accurate comparison of the plasma profile of a biosimilar compared to the reference product.
It is, clearly, possible to demonstrate that the putative biosimilar looks the same, acts the same, and circulates in the blood at the same concentration as the reference product, so it seems inconceivable that in these circumstances, the biosimilar could display an efficacy-safety profile which is different from the reference product. However, there is one area of potential uncertainty, and that is differences in immunogenicity. Circulating antibodies can impact the PK profile, reduce efficacy, and impact safety in the worst case.
Current approach by the major regulatory agencies
Regulatory guidance from the FDA and EMA requires that in the absence of surrogate markers for efficacy, it is usually necessary to demonstrate comparable clinical efficacy of the biosimilar and the reference product in clinical studies in patients. Such studies need to be adequately powered by randomized parallel-group comparative clinical trials and use appropriate efficacy endpoints. Since many biological products, particularly monoclonal antibodies, do not have adequate surrogate markers, large clinical trials are required, often in 500 to 1000 patients.
Are clinical studies really needed?
But are these studies really needed? In many cases, such trials represent a blunt instrument with very limited sensitivity to detect differences compared to the application of in-vitro methods. In some cases, the recommended dose is on the flat part of the dose-response curve. This means that a lower or higher dose or potency may still show equivalence to the reference product, whereas in-vitro testing could detect differences.
Trials are fraught with difficulties, and one can never demonstrate equivalence with 100% certainty; even two identical products can show response differences because of inherent variation within the trial population. It is often not possible to show with reasonable certainty preservation of much more than 50% of the therapeutic effect of the reference product, whereas in-vitro potency can often be demonstrated with greater than 80% equivalence. In other circumstances, it can be impractical or unethical to design trials with adequate sensitivity. There is also the need to consider the amount of resource channelled into the clinical development of biosimilars which could more usefully be applied to the development of new and better medicines. Therefore, it is important to look at the need for clinical data critically.
In 2020 Schiestl et al. (1) examined 42 biosimilar programs and determined that four did not receive approval due to quality issues but none due to purely efficacy differences. Since then, one further biosimilar has not been approved due to quality issues. While three other programs initially failed to meet their primary efficacy endpoint, these were subsequently approved based on sound quality rationale for the observed differences in efficacy.
The UK MHRA takes a new position
While the major regulatory agencies remain adherent to a strict position of requiring clinical data in circumstances where no suitable pharmacodynamic marker exists, the UK MHRA has taken a different view (2). Whereas MHRA guidance states that it takes into consideration the EMA guidance documents, it also takes account of scientific and regulatory experience gained since the first biosimilar product was licensed in 2006. Like the EMA and FDA, the UK guidance states that clinical comparability should always include a pivotal comparative PK trial, which should consist of the measurement of pharmacodynamic markers if available. In contrast, the guidance also states that in most cases, a comparative efficacy trial may not be necessary if sound scientific rationale supports this approach. The guidance states that the efficacy of the reference product can usually be related to the biological events triggered by the binding of the active product to its known targets, which means that in most cases for UK approval of a biosimilar, comparative clinical efficacy trials will not be required. However, the assessment will be based on the total package, and therefore final acceptance of this approach can only be considered after submission of the clinical data package.
With regards to safety, the MHRA explains this can be predicted from the on-target effects of the reference product. Other common adverse drug reactions might include injection-related reactions triggered by various mechanisms, some mediated by anti-drug antibodies. These, to some degree, can be predicted by the quality attributes, including drug product characteristics such as aggregation and impurities and the formulation of the biosimilar candidate. On top of this, the healthy volunteer studies will provide immunogenicity data which can further support similarity.
The guidance concedes that there may still be cases requiring clinical trials in patients. These would mainly be required in circumstances where there is a lack of understanding of the biological functions of the reference product related to its clinical effects or where the relevant critical quality attributes may not be sufficiently characterized due to analytical limitations. Exceptionally additional clinical safety data may be required where safety uncertainties cannot be resolved without patient exposure pre-license, for example, where serious adverse reactions to the reference product have unpredictable root causes. A classic example of this is the pure red cell aplasia associated with immunogenicity following erythropoietin treatment (3). Exposure of a significant patient cohort to the biosimilar candidate would be considered the most appropriate approach to resolve any residual uncertainty around safety and immunogenicity.
What the future holds
In conclusion, as treatments become more tailored and complex and evolve to become different from the settings in which the reference products were originally developed, the potential to perform meaningful biosimilar studies will decline. Further, with the rapid advance in our understanding of therapeutic medicine, more novel therapies will be developed, making the conduct of therapeutic equivalence trials with biosimilars more challenging and less attractive. This could potentially drive up the cost of biosimilar development, defeating the basic objective of expanding access to expensive medicines to broader populations. The guidelines published by the UK MHRA represent a significant step towards resolving these challenges. It is anticipated this pioneering approach will, in the long run, be adopted by all major regulatory agencies.
1) Martin Schiestl et. al. The Path Towards a Tailored Clinical Biosimilar Development BioDrugs. 2020 Jun;34(3):297-306. doi: 10.1007/s40259-020-00422-1. PMID: 32266678