The Tailored Biosimilar Approach: Expectations and Requirements

Abstract

Regulatory approval of biosimilar medicines currently requires a combination of physicochemical and functional testing, pharmacokinetic data, and clinical efficacy studies (CES). In this article, we discuss the tailored biosimilar approach, which represents an evolution in regulatory thinking by moving away from the default requirement for CES in biosimilar approval. We explore how physicochemical and functional data can be predictive for clinical performance and address the limitations of CES for regulatory decision-making. We argue that, in most cases, the combination of a robust package of physicochemical and functional testing, with appropriately designed pharmacokinetic studies provides sufficient evidence to establish biosimilarity. Additionally, we provide our perspective on the requirements, expectations, and exceptions for future biosimilar approvals, outlining specific scenarios where additional clinical evidence may be necessary. These include cases where the mechanism of action is unknown or poorly characterized, when product heterogeneity cannot be adequately characterized, or where relevant safety or immunogenicity concerns arise with the reference product or biosimilar candidate. Finally, we aim to clarify the remaining concerns surrounding the tailored biosimilar approach, providing insights into the potential to streamline biosimilar development and regulatory approval.

Key Points

Evolution of the regulatory framework towards a tailored biosimilar approach allows a shift away from clinical efficacy studies while maintaining the high standards for biosimilarity.

Clinical efficacy studies are not required by default for biosimilar development but rather in situations where these data provide crucial information for the approval of a biosimilar candidate.

Supportive safety and immunogenicity endpoints can be included in prospective pharmacokinetic studies.

Introduction

For over a decade, regulators, clinicians, and drug developers have engaged in discussions on ways to streamline and refine the regulatory approval process for biosimilars to improve patient access to affordable therapies [1–3] while maintaining high scientific standards. Initially, the biosimilar framework followed a conservative approach because of limited analytical abilities to fully characterize complex biologicals and limited regulatory experience with biosimilars. Clinical efficacy and safety data were generally required to confirm similar clinical performance. Since the earliest guidelines on biosimilars were published, it has been acknowledged that the cornerstone of biosimilarity is the functional and physicochemical characterization and that clinical confirmatory data, such as clinical efficacy studies (CES), may not always be needed. Subsequently, significant progress has been made, particularly regarding the limited value of in vivo studies for biosimilar development as well as the necessity of CES for first-generation biosimilars or those with accepted pharmacodynamic markers as surrogates for efficacy, for example, insulins, low-molecular weight heparin, and growth factors [4–6]. Current development programs and product-specific regulatory guidelines for these cases are clear and straightforward, acknowledging the limited scientific value of CES in the totality of evidence approach [6, 7]. Moreover, there are ethical concerns associated with conducting unnecessary clinical studies. However, opinions continue to differ about the necessity of CES in the development of more complex biologics. This ongoing debate highlights the need for a re-evaluation of regulatory frameworks to ensure the efficient and scientifically sound development of high-quality biosimilars.

What is Known About the Role of CES in Biosimilar Development

In two recent publications [8, 9], we discussed critical aspects of the characterization and regulatory approval of monoclonal antibody (mAb) and fusion protein biosimilar candidates. Both publications provide a data-driven analysis of marketing authorization applications (MAAs), European public assessment reports, and withdrawal assessment reports for mAb and fusion protein biosimilar candidates assessed by the European Medicines Agency (EMA) up to November 2022. They offer an in-depth analysis of the quality and clinical packages for a representative set of 23 mAbs with different indications and mechanisms of action, including applications that were withdrawn.

These studies concluded that the physicochemical and functional data package is predictive for the authorization of complex biosimilars. There were no instances where the evaluation of the quality dossier and the outcome of the marketing authorization procedure were not aligned. Most importantly, CES did not play a critical role in the final decision for regulatory approval. Uncertainties regarding the physicochemical and functional comparability between a biosimilar and its reference product could not be resolved by clinical efficacy data. Instead, further evidence of similarity was obtained by additional functional assays or clinical pharmacokinetic data. Furthermore, uncertainties regarding clinical comparability were resolved through the robustness of the physicochemical and functional quality data package along with pharmacokinetic comparability in sufficiently powered studies. Alternatively, they were justified based on the general limitations of clinical data in the biosimilar development programs, such as immature secondary endpoints, imbalances between trial arms, or chance findings. Notably, there were even cases of successful biosimilar programs despite formally failed CES [10–12].

Other data-driven analyses by various authors, which include different products, methodologies, and markets, consistently came to similar conclusions [1–3, 13]. Schiestl et al. [1] performed a retrospective review of 42 clinical trial results from biosimilar pharmacokinetic studies and CES from 38 development programs that were disclosed in European public assessment reports and US Food and Drug Administration drug review summaries from 2006 to 2019. The retrospective review of approved biosimilars in the EU and USA revealed that, in 100% (38 of 38) of biosimilar development programs, CES confirmed biosimilarity. The authors also concluded that, for 95% (36 of 38) of biosimilar development programs, CES provided no additional value to the scientific review process to approve a biosimilar. The remaining two cases (5%) involved biosimilar candidates that showed clinical differences in immunogenicity despite meeting efficacy outcomes. Neither product was approved in the EU or USA until their manufacturing processes had been optimized to improve product quality. The recurrence of these cases is unlikely today because of state-of-the-art assays and improved control of process-related impurities [1]. Bielsky et al. [2] performed an in-depth review of six classes of mAb and fusion protein biosimilars evaluated in the EU and identified no instances where efficacy trials added crucial information for the establishment of biosimilarity. Moreover, their analysis concluded that most differences observed were resolved by the robustness of the analytical and pharmacokinetic data package and the totality of evidence presented. Kurki et al. [3] comprehensively analyzed post-marketing surveillance data of all biosimilar mAbs and fusion proteins authorized in the EU up to July 2020. Analysis of post-marketing surveillance data covering up to 7 years of follow-up and more than 1 million patient-treatment years of safety data did not reveal any biosimilar-specific adverse effects. No product was withdrawn for safety reasons [3].

A limitation of these publications is that the evidence was derived from biosimilar MAAs and the low number of negative outcomes, which may be because programs with larger deficiencies are rarely submitted. Furthermore, the publications mentioned above are retrospective studies with limited capacity to inform on how to prospectively define clinically meaningful differences on the quality level. However, we believe these studies can help to address concerns regarding biosimilar approvals in the absence of CES. To date, there have been no instances where clinical data alone resulted in a negative decision to approve a biosimilar product.

Furthermore, there are concerns regarding the insensitivity of current CES that are part of biosimilar development programs. For example, there are cases where CES may not have the sensitivity to detect significant differences in potency between two similar proteins when the dose-response effect is unknown or shallow and they are given at doses higher than the maximum effective dose (i.e., on the flat part of the dose-response curve, for example, rituximab, infliximab, and pembrolizumab) [14–16]. As for the value of extensive analysis of secondary endpoints, in several cases they have been viewed as inconclusive or immature [2, 9]. In some cases, CES may introduce more concerns than they solve, producing results that are inconsistent with comparative analytical data in supporting biosimilarity. Moreover, CES have been shown to be an inefficient use of resources for establishing biosimilarity because they cannot resolve residual uncertainties arising from insufficient comparative quality or pharmacokinetic data of a biosimilar candidate [17].

We believe that the knowledge generated and the regulatory experience accumulated to date provide sufficient reassurance that CES are no longer warranted by default for the demonstration of biosimilarity. Conversely, the conduct of confirmatory CES should be carried out in specific situations where comparative clinical efficacy/safety/immunogenicity data provide crucial information for the approval of a biosimilar candidate. This does not mean that less evidence will be generated for the demonstration of biosimilarity but rather that the necessary evidence will be obtained through other means.

The tailored biosimilar development adjusts the extent of clinical data required for regulatory approval based on factors such as product class, indication, and mechanism of action, among others. Rather than mandating a confirmatory CES for every biosimilar by default, the tailored approach can be applied  where a CES may be replaced by an adapted pharmacokinetic trial design or omitted entirely [2, 17–19]. Therefore, tailored biosimilar development allows for flexibility while ensuring an appropriate data package is available to confirm biosimilarity and support regulatory approval, allowing a move away from clinical efficacy trials while maintaining the high standards for biosimilarity.

In this article, we offer our perspective on the remaining concerns regarding the tailored biosimilar approach, with the goal of providing clarity on this topic to all readers and stakeholders.

Ensuring Comparability When Comparative CES are No Longer Required

Expectations Regarding Quality Data

The extensive experience gained by developers and regulators, and the advances in analytical methods and techniques over two decades of biosimilar approvals, have already led to global discussions on re-evaluating the requirement for CES. This discussion is reflected in the revision of biosimilar guidelines from different regulatory agencies worldwide [7, 19–22] as well as several multi-stakeholder initiatives, events, and joint efforts that have taken place in the past 2 years focusing on this specific topic [17, 23]. At a recent international workshop, both regulators and industry experts acknowledged the limitations of CES. They agreed that CES are usually not sensitive enough to detect anything but large biological or physicochemical differences, which in any case would not be acceptable for a biosimilar candidate since high similarity to the reference product must be demonstrated [17].

A thorough analysis of the quality attributes (QA), including functional characterization, already builds the basis for the development program of a biosimilar candidate. Based on a good understanding of the mechanism of action, functional assays are designed to reliably analyze and compare the structure, potency, and purity of the biosimilar candidate and the reference product. Advancements in analytical methods mean that current assays are highly sensitive and provide crucial information on critical QAs (CQAs) that are linked to the mechanism of action.

The Basis of Quality Data Comparison

The methodology of demonstrating biosimilarity at the analytical level is rooted in decades of experience with approving manufacturing changes for originator products and therefore follows the basic principle outlined in the ICH Q5E guideline for comparability assessment of manufacturing changes [24]. Most biological medicinal products undergo numerous changes in the years after their initial approval. These changes can include significant alterations to the manufacturing process, the introduction of new manufacturing sites, and changes in formulation, to name a few. Because of the inherent heterogeneity of biological products, significant manufacturing changes can result in alterations to their quality profile. This could include differences in the impurity profile, changes in the oligosaccharide structure, or alterations in the charge profile, essentially creating a new “version” of the product.

To support such manufacturing changes, companies are required to submit comparability data to regulators, comparing batches of pre-change and post-change products at the physicochemical and functional levels. Any differences in the quality profile must be justified as having no potential impact on the efficacy and safety of the product. Relevant functional and biological tests, along with existing literature, should be used to provide evidence to support the lack of clinical significance of any difference noted in QAs. In most cases, regulatory approval for such manufacturing changes is based on physicochemical and functional data; this holds true even for major changes such as changing the production cell line. Clinical efficacy data are required only in very rare cases [25]. Therefore, approval of biosimilars without CES is conceptually similar to the regulatory process and scientific standard used for over 20 years for the approval of manufacturing changes to the reference products during their life cycle.

The Analytical Comparability Exercise

It is important to consider that variability is inherent to all biotechnology-derived medicines. High similarity of biosimilars to their reference product is ensured by the analytical comparability exercise [26] where multiple reference product batches are tested using a comprehensive panel of analytical methods, including orthogonal methods for many QAs. The efficacy and safety of the reference product has been demonstrated during the MAA procedure and by subsequent use in real-world settings. The data for the reference product batches are used to pre-define “similarity ranges” for each QA that serve as target ranges for the biosimilar. The approach to set the similarity ranges should be justified for each biosimilar as it highly depends on several factors, for example, the number of batches and the variability and criticality of the QA. If the QAs of the biosimilar fall within the ranges established based on characterization results from the reference product, it can be assured that the biosimilar will be highly similar to the reference product.

However, given the inherent variability of products of biological origin, differences may occur in some QAs. Any such differences must be thoroughly justified to ensure they do not affect the product's safety or efficacy. Companies do this through additional testing, for example, on a functional level, and justifications based on current scientific knowledge. Moreover, process- and product-related impurities (e.g., host cell proteins, aggregates, non-human glycan structures), which could potentially increase immunogenicity, should be tightly controlled through the overall manufacturing control strategy, including release testing with acceptance limits for relevant product- and process-related impurities [27].

Like with any product, companies must also demonstrate that their manufacturing process will consistently produce batches of the required quality. This will be further ensured by in-process controls and release testing using stringent specifications, as for every other authorized medicinal product.

Any changes made to the biosimilar after marketing authorization must be assessed and approved by regulators in accordance with ICH Q5E [24], as described above.

Structure–Function Relationship

The knowledge gained in the past decades on structurefunction relationships, along with the advancements in analytical techniques, allows for highly sensitive physicochemical and functional characterization of molecules such as mAbs. QAs known to impact clinical efficacy and safety are closely monitored as part of the biosimilarity exercise. This was exemplified in our recent publication, where we discussed differences in glycosylation for some of the rituximab biosimilar candidates compared with the reference product [8, 9]. Functional analysis in cell-based assays demonstrated that these differences had no functional effect, so no impact on clinical performance is expected. Moreover, the clinical pharmacokinetic study found no significant differences, although the pharmacokinetic profile could, in principle, be influenced by the composition of sugar moieties. Another such example was the minor differences in binding results to several Fc-receptors found for adalimumab biosimilar candidates when compared with the reference product. Again, it was concluded that these differences would not have a clinical impact as the cell-based assays demonstrated no functional effect.

These examples highlight that appropriate physicochemical and functional analysis can clearly detect differences. Moreover, based on appropriate knowledge and understanding of the CQAs, it is possible to predict which differences may have a relevant clinical impact (i.e., differences that could potentially preclude a marketing authorization).

Quality Data Provide Sufficient Evidence of Biosimilarity

In our opinion, in most cases, sufficient evidence of biosimilarity can be generated via appropriate physicochemical methods and functional assays capable of detecting differences in functional attributes. This is complemented by an adequately powered pharmacokinetic study that can also provide supportive information on safety and immunogenicity. Comparable efficacy can be ensured throughout multiple levels of the analytical characterization, including physicochemical comparisons, binding assays, and cell-based bioassays (where relevant for the mechanism of action), together with pharmacokinetic data demonstrating equivalent exposure. Robust and convincing analytical data packages are a general regulatory requirement for biosimilar approval, and current physicochemical and functional assays are highly advanced and sensitive. In vitro potency assays can replicate the in vivo action of the biological medicine, providing a more precise evaluation of potency than could ever be determined by CES. Experience with biosimilars has shown that information on the immunogenicity of the biosimilar candidate can be largely deduced from the reference product, that is, the potential immunogenicity of the protein sequence itself can be derived from the information on the reference product [3, 28]. Developers should conduct an immunogenicity risk assessment that addresses relevant issues such as the known immunogenicity of the reference product, the sensitivity of the physicochemical and functional methods being employed, the potential interference from residual plasma drug levels, and other product- and patient-related risk factors.

Maintaining the Highest Standards of Scientific Evidence

When discussing biosimilar development without the requirement for CES, several aspects must be considered to guarantee safety and efficacy comparable to that of the reference product while maintaining the highest standards of scientific evidence.

An important consideration is whether the current approach for demonstrating physicochemical and functional biosimilarity needs to be adapted for tailored biosimilar development in the absence of CES. The comparability exercise at the quality level should be extensive, carefully designed with state-of-the-art orthogonal tests, and without any potential shortcuts. For some products, there may be a clear link between CQAs and clinical outcomes, whereas, for others, this link may not be as evident. CQAs need to be in the range of the reference product, or biosimilar developers may need to provide additional information, including appropriate studies to increase the overall knowledge of the product’s CQAs.

In summary, it is paramount to ensure that any observed analytical difference will not affect clinical performance. If there is an understanding of how relevant CQAs impact clinical performance, and there are either no differences between the biosimilar and the reference product in physicochemical and functional data, or any minor differences can be justified as not affecting safety and efficacy, then, in our opinion, a combination of physicochemical and functional testing data, supported by pharmacokinetic and a subset of safety and immunogenicity data (with pharmacodynamic data as appropriate), should be considered sufficient for demonstrating biosimilarity.

Expectations Regarding Clinical Data

The safety and immunogenicity profile of the reference product, for example, across different indications or in combination with other medicinal products, provides important information and will build the basis for the planning and conduct of a comprehensive and well-considered clinical program. Although a comparative clinical pharmacokinetic study is considered a basic requirement, in our opinion, a CES is unnecessary if sufficient evidence of biosimilarity can be inferred from other parts of the comparability exercise. Most clinical safety and immunogenicity data, which are only secondary endpoints in current CES studies, can be partly derived from analytical similarity coupled with supporting data obtained as part of clinical pharmacokinetic studies.

Comparative pharmacokinetic studies should be performed in a sufficiently sensitive and homogenous population. Healthy volunteers are the preferred option, as the interindividual variation between healthy subjects is lower and there are fewer potential confounders compared with patients [26]. In cases where the biological carries safety concerns, the inclusion of healthy volunteers is unethical and an adequate patient population has to be selected.

In addition, depending on the safety and immunogenicity profile of the reference product, the design of the pharmacokinetic study could be adapted, for example, the dosing schedule and study duration. For instance, a pharmacokinetic study using multiple dosing and a study duration of sufficient length could be envisioned to capture the development of potentially delayed antibody responses. A comparative pharmacokinetic study could, in many cases, provide sufficient supportive information on safety and immunogenicity. The clinical safety and immunogenicity data derived from comparative pharmacokinetic studies, in combination with the analytical testing for CQAs related to immunogenicity, will guide the decision as to whether additional clinical studies are required to show biosimilarity.

Analyses from different authors, and the accumulated experience with biosimilar development and evaluation have substantiated the validity of safety and immunogenicity data derived from pharmacokinetic studies [2, 3, 8, 9]. In addition, our most recent publication showed that, where biosimilarity was established at the analytical level, this always translated into pharmacokinetic similarity between the biosimilar candidate and the reference product [9].

Other authors have reviewed the long-term post-marketing pharmacovigilance data for biosimilars across various substance classes. Their findings indicated that biosimilars have a safety profile that cannot be distinguished from the reference product during broader application in real-world use [3, 29, 30]. Although biosimilarity must be ensured as prerequisite for approval, the post-marketing pharmacovigilance requirements are the same for the biosimilar as for the reference product. This includes a robust pharmacovigilance system and risk management plan that follow current and evolved standards to allow the implementation of comprehensive safety monitoring and adequate risk minimization strategies in accordance with the respective originator.

When Will CES Still be Needed?

Although comparative CES studies may not be necessary in most cases, there remain specific scenarios where comparative clinical efficacy/safety/immunogenicity data are required to ensure similar clinical performance.

In our opinion, these scenarios include (but may not be limited to) the following.

(a) Biologicals with an unknown or poorly characterized mechanism of action. It is important to highlight that biological medicinal products can vary widely in complexity. In the future, new classes of biosimilars may emerge in areas such as advanced therapy medicinal products or antibody–drug conjugates. In some cases, the mechanism of action of the reference product may be unknown or poorly understood and/or in vitro tests to measure them may not be available or may be insufficiently sensitive. For example, with antibody–drug conjugates, the complex chemistry involving the cytotoxic payload, linker, and mAb may make it challenging to develop a comprehensive physicochemical and functional testing package to support biosimilarity. The mechanism of action of certain advanced therapy medicinal products, such as cell-based products, is currently incompletely understood because their potency depends on a combination of cellular properties. In such cases, where it may not be possible to fully characterize all clinically relevant functional and physicochemical attributes of the reference product, confirmatory CES might be necessary to provide sufficient evidence of biosimilarity.

As the current experience in biosimilars covers mainly therapeutic proteins, additional knowledge on these new therapeutic groups is required.

(b) Products with high intrinsic heterogeneity and/or insufficient characterization of structure. Characterization of complex biologics with significant structural heterogeneity (such as heavily glycosylated proteins or molecules with complex oligosaccharide profiles, such as certain enzymes) or other less characterized products (such as naturally derived mixtures of proteins or polynucleotides such as defibrotide) can be challenging, which complicates the process of establishing comparability at the physicochemical and functional level. If current analytical methods or prior knowledge cannot provide sufficient understanding of these heterogenicities and their significance, CES may be necessary to exclude a possible clinical impact and relevance.

(c) Situations where pharmacokinetic studies are not relevant. For products that are locally administered (e.g., intraocular, intrathecal, intraspinal administration) and are thus not reaching relevant systemic exposure, the safety and immunogenicity information cannot be derived as part of the pharmacokinetic studies. In these cases, further discussion and research may be needed to decide whether physicochemical, structural, and functional characterization studies are sufficient, or whether additional studies, such as CES, are required to confirm comparable efficacy and safety.

For biosimilars of reference products known to exhibit clinically relevant safety or immunogenicity concerns, such as those with high levels of anti-drug antibodies leading to reduced efficacy and/or where immunogenicity is associated with safety issues such as severe infusion reactions, additional clinical evidence may be necessary. However, this should not preclude the possibility of waiving a CES. Other options might be possible, for example, multidose pharmacokinetic studies with safety and immunogenicity endpoints, to identify potential differences in immunogenicity profiles and their impact on efficacy and/or safety and to implement respective risk management strategies.

Conclusions

Although biosimilars may exhibit some minor differences to the reference product, adequate and convincing demonstration of biosimilarity will, in our view, be possible by means of a totality of evidence approach without the need for confirmatory CES in most cases. If a CES is required, it should be specifically designed to address questions that cannot be answered through comparative physicochemical and functional characterization, complemented with a clinical pharmacokinetic study. In all cases, the best path to provide the necessary evidence for development and approval of a biosimilar must be considered. Although additional clinical data may provide confidence, the conduct of supplementary analytical methods or adequate modifications of the manufacturing or purification processes may be more able to ensure comparable clinical performance, given the lower sensitivity of clinical studies to detect differences. In general, adapting the manufacturing process of the biosimilar to more closely align with the quality profile of the reference product is preferred over conducting CES, as the aim of a biosimilar program is to produce a medicinal product that is as similar as possible to the reference product. It is critical that the biosimilar regulatory framework is further refined and harmonized globally to clarify how any remaining uncertainties from the analytical comparison can be resolved and in what cases CES would be required. Moreover, building confidence in biosimilar medicines and their regulatory framework among all stakeholders, including patients and physicians, is essential. Although comparative efficacy data can sometimes be easier to understand and may provide further confidence to clinicians and patients, educational efforts must be undertaken to improve understanding of the rigor and transparency of the entire analytical similarity assessment process.

Declarations

Author contributions

EG, NK-S, SB, NJ, and NE wrote the first draft of the manuscript and VK, JK-G, JM, MW, AL, RA and GvZ commented and critically revised previous versions of the manuscript. EG performed the literature analysis and interpretation. NK-S conceived and designed the manuscript and supervised the process. All authors read and approved the final manuscript.

Conflict of interest

Open Access funding enabled and organized by Projekt DEAL. This study was funded by the German Federal Ministry of Health based on a decision by the German Bundestag. The authors have no potential conflicts of interest or financial disclosures that are pertinent to this article. The views expressed in this article are the personal views of the authors and should not be understood or quoted as being made on behalf of or reflecting the position of the regulatory agencies with which the authors are affiliated.

Funding

This study was funded by the German Federal Ministry of Health based on a decision by the German Bundestag (Grant no. 2522FSB450).

Ethics approval

Not applicable.

Consent to participate

Not applicable.

Consent for publication

Not applicable.

Availability of data and materials

No data were used for the article.

Code availability

Not applicable.