Biosimilars: Are Comparative Efficacy Studies Required?

It is far from a controversial statement that biologic therapeutics – such as monoclonal antibody and recombinant protein therapies – are complex and expensive to develop. As one might expect, therefore, the results of that development are also expensive when provided to patients, with per-patient annual costs for some commonly-used biologic therapies exceeding $20,000 in the US. By providing alternatives to originator therapeutics, biosimilars are crucial for expanding access to powerful therapeutics and providing competition which serves to lower expenses. In a recent draft update to regulatory guidance, the requirements for licensing a biosimilar candidate were simplified. Specifically, a clinical study is no longer required to demonstrate the efficacy of the biologic under most circumstances. Here, we’re going to explore this change, and discuss how this eases the path to market for many biosimilar products.

Biosimilars are therapeutics which are developed to be equivalent to an existing biologic, known as an originator or reference product. This means that they contain an active substance which is highly similar to that of the reference, with the intention that it has as close a clinical effect as the reference.

As a result, developing a biosimilar can provide an accelerated path to market compared to developing a completely novel therapeutic. By “piggy-backing” off the regulatory approval of the reference product, the focus of development and testing of a biosimilar moves from proving that the product is clinically effective in its own right to demonstrating that it is at least as effective as the reference.

An example of a therapeutic with several biosimilars is Adalimumab, which is a monoclonal antibody therapy targeting a wide range of auto-immune diseases, including psoriasis, rheumatoid arthritis, and Crohn’s disease. The originator therapeutic is marketed under the brand name Humira, while biosimilar therapeutics – including those marketed as Abrilada, Hyrimoz, and Yuflyma, among several others – have been approved.

The role of biosimilars in small molecule therapeutics is played by generics and bioequivalents. Generics are medical products which contain the same active ingredient at the same concentration and are delivered in the same way as an originator, and can thus be considered clinically identical. A generic paracetamol tablet is a generic form of Tylenol, for example. Bioequivalence extends this to include new delivery methods of already licensed products: if the active ingredient is the same, then researchers need only demonstrate that key pharmacokinetic (PK) parameters of that ingredient are similar under the new delivery method.

Biologics add a further complexity into this mix, namely that for larger molecules such as proteins it is very difficult to demonstrate that two molecules are truly identical. That means that the standard approach used for small molecule drugs – where specific molecules are easy to construct and their identity simple to verify – cannot be used for biologics. The WHO states that “The generic approach is not suitable for the licensing of biosimilars since biological products usually consist of relatively large and complex proteins that are more complicated to characterize and manufacture than small molecules”.

As a result, the sense of two biologic products being identical is different to that for small molecule drugs. In the US, biosimilars are licensed under section of the Public Health Service (PHS) Act (42 U.S.C. 262(k)), which defines the studies required to demonstrate biosimilarity as:

• “Analytical studies that demonstrate that the biological product is highly similar to the reference product notwithstanding minor differences in clinically inactive components”. These provide evidence that the proposed biosimilar product has a similar structure to the reference product, and has been shown to behave similarly in a laboratory setting.
• “An assessment of toxicity”. The biosimilar product formulation should be assessed for any safety concerns. This can be incorporated partially or wholly within the other two components described.
• “A clinical study or studies (including the assessment of immunogenicity and pharmacokinetics or pharmacodynamics [PD]) that are sufficient to demonstrate safety, purity, and potency”. A clinical evaluation of the biosimilar demonstrates that it performs similarly to the reference product when deployed against the target condition. This ensures that similarities in analytic and safety profiles translate to similarities in its clinical profile.

In the autumn of 2025, updated draft guidance for aspects of the clinical evaluation of a biosimilar was published by several regulatory bodies, including the FDA. This specifically involves a study known as a Comparative Efficacy Study (CES). A CES is, as the name suggests, a trial which directly compares a proposed biosimilar to its reference product in a clinical setting. They are intended to resolve ” after other stages of the evaluation are accounted for. These were not mandatory under previous guidance, but the onus was on the sponsor of the study to provide scientific justification for omitting a CES from the testing of a biosimilar.

Draft guidance which updates these recommendations was issued by the FDA in October 2025. This considers an approach where a CES may not be required for biosimilar licensing under specific named scenarios. This draft change is justified by advances in analytical technology and techniques since the previous guidance was issued in 2015, as well as increased experience on the part of the FDA itself in evaluating the effect of analytical differences between products on their clinical performance. The guidance states that a Comparative Analytical Assessment (CAA) – a study which evaluates biosimilarity on the basis of laboratory experiments – is typically more sensitive to differences between a biosimilar candidate and a reference product than a CES. This is due to several factors, including the chosen therapeutic dose range, the clinical study population, and the endpoints selected for the study. That means that there are fewer scenarios where a CES is likely to be required to prove biosimilarity.

The guidance outlines three checks which indicate that it is unlikely that a CES will be required to license a biosimilar candidate:

• “The reference product and proposed biosimilar product are manufactured from clonal cell lines, are highly purified, and can be well-characterized analytically”
• “The relationship between quality attributes and clinical efficacy is generally understood for the reference product, and these attributes can be evaluated by assays included in the CAA”
• “A human pharmacokinetic similarity study is feasible and clinically relevant”

The logic here is clear: if one can prove that the biosimilar product is structurally and behaviourally similar enough to the reference, and the connection between those properties and the clinical efficacy of the product is strong, then it is unnecessary to perform a clinical trial to reestablish that clinical efficacy for the biosimilar candidate. Previously, this was not possible thanks to the difficulty of establishing the complex structure and behaviour of biologic molecules with enough accuracy that robust assessments of biosimilarity could be made. The improved ability of researchers to make such measurements means that, when combined with appropriate human pharmacokinetic and immunogenicity studies to assess safety, a CAA is generally sufficient to demonstrate biosimilarity without the need for further clinical trials.

There are, nevertheless, cases where a CES might still be beneficial or even required under the new guidance. An example might be a product which acts only locally near the site of administration – one which is administered in the eye, for example – and for which, therefore, a comparative PK study is not clinically relevant. This means that a CES would be required to show the biosimilar candidate provides the same efficacy at its site of action given equivalent PK cannot be established.

By reducing the requirement for often-unnecessary clinical trials, the cost and duration of the biosimilars licensing process is expected to decrease noticeably. A CES is resource-intensive – FDA analysis indicates that the typical CES lasts 1-3 years and costs an average of $24 million – meaning the approval pathway for new biosimilar candidates becomes far more cost-effective to navigate with the updated guidance. This has the potential to vastly increase the number of biosimilar products on the market in the future, widening access and reducing costs for patients.

Non-parallelism and Biosimilars

A Statistical Understanding of Polyclonality in Monoclonal Antibody Production

Potency Assays for Cell and Gene Therapy: Statistical Challenges

As we’ve outlined earlier, the overall goal of the analytical component of biosimilar testing is to establish that the candidate is “highly similar” to the reference. The two main aspects of the candidate which must be shown to have sufficient similarity are:

• The structure of the candidate. Does the candidate have a similar molecular and physical structure as the reference?
• The functionality of the candidate. Does the candidate behave in a similar way to the reference?

Once these properties have been established, it must then be demonstrated that there is a strong connection between them and the clinical efficacy of the product. In essence, we seek to understand the shape of the product, how it interacts with its target, and how these properties translate to the real-world behaviour of the candidate when it is used in a clinical setting.

Even with modern analytical techniques, determining the molecular structure of a protein is a difficult proposition. A typical antibody might contain on the order of tens to hundreds of thousands of atoms. Compare this to a molecule of paracetamol, which contains 20 atoms, and we can appreciate the challenge of ensuring biologic products are well-characterised compared to small molecule therapeutics. And this is before we consider the shape, or conformation, of the protein. A protein’s conformation is crucial to its interactions with its target, meaning two proteins with identical molecular composition but different conformations might behave completely differently in a clinical deployment.

Proteins can be thought of as chains of smaller molecules, known as amino acids. In order for two proteins to be biosimilar, their amino acid sequence of the biosimilar ought to be highly similar, if not identical. Sequence variation can, however, occur due to transcription and translation errors in the biological process which produces the product. Indeed, biosynthesis processes are likely to also introduce conformational differences between the biosimilar and the reference product. Both the amino acid sequence and molecular structure of the biosimilar should, therefore, be assessed using in vitro assays as part of a CAA. They need not be completely identical, though any differences should be documented and controlled, and their clinical impact evaluated.

The biological activity of the biosimilar should also be assessed. Ideally, the mechanism of action of the reference product should be well understood, so examining the biological activity of the biosimilar will provide a link to its clinical behaviour. Thus, the choice of bioassay used to evaluate the biological activity of the biosimilar should be chosen to reflect that mechanism of action. Potency is the recommended metric for biological assay: we would expect that the potency of the biosimilar is similar to that of the reference product. The potency assay used should be sufficiently accurate and precise to detect small changes in the biological activity due to minor differences in the structure of the molecules.

Evaluating the potency of a biosimilar relative to a reference product can be more complicated than in many standard relative potency bioassays due to non-parallelism. We usually require that the dose-response curves for a test sample and a reference be parallel – i.e. their shapes be identical, with the only allowed difference being a shift in the x-direction – in order for a relative potency measurement to be valid. This may not be the case for a biosimilar. While a biosimilar is intended to behave as similarly to its reference product as possible, it will not necessarily be expected to behave as a dilution of the reference, meaning some degree of non-parallelism may be inevitable. Techniques exist to account for this non-parallelism and provide a well-defined characterisation of the relative potency of the biosimilar.

As previously noted, both the reference product and the candidate biosimilar are manufactured using biological processes, a noticeable degree of lot-to-lot variability is expected for both products, particularly when considering their biological activity. As a result, the FDA guidance regarding CAA studies recommends that the properties of both products are evaluated over several lots – at least 10 for the reference and 6-10 for the candidate biosimilar – to ensure both “adequate characterisation of the proposed product and understanding of manufacturing variability” and “adequate comparison to the reference product”. These lots should be chosen to represent the full range of variability of the two products, such as those with different expiry dates. For the candidate biosimilar, lots used for different stages of the development process should be assessed, including, for example, those used for engineering, those manufactured using clinical-scale processes, and those used for clinical trials (should any occur).

There are other properties of the candidate biosimilar which might be considered for investigation in a CAA. For example, the manufacturing process might be considered or the presence or nature of any impurities in the final product. The choice of which of these attributes are included in the CAA are left to the sponsor but, crucially, a risk-based assessment of these inclusions and exclusions should be part of the final plan for the CAA.

An overarching theme that emerges from these updates is one of growing confidence. As regulators gain experience with biosimilars, and analytical technologies continue to improve, there is reduced need to rely on resource-intensive clinical trials. Modern characterisation methods allow researchers to examine the structural and functional properties of biologic molecules with a level of resolution that simply wasn’t achievable when the original biosimilar frameworks were established. When this analytical understanding is paired with targeted human PK and immunogenicity studies, the evidence needed to demonstrate biosimilarity can often be assembled more efficiently, without compromising safety or scientific rigour.

Comparative efficacy studies have historically been one of the most expensive and slowest components of biosimilar development, yet in many cases they offer little additional information beyond what can now be captured by careful laboratory assessment. By reducing reliance on these trials, the updated guidance stands to shorten development timelines and lower overall costs, which in turn supports a more competitive and accessible market for biologic therapies.

In essence, the path to bringing a biosimilar to patients is becoming more proportionate to the scientific questions that truly need answering. That change not only benefits developers and regulators: it ultimately benefits patients, by encouraging a broader, more sustainable ecosystem of high-quality biosimilar medicines.