Sarfaraz K. Niazi, PhD, is an adjunct professor of biopharmaceutical sciences at the University of Illinois and the University of Houston, and founder of biosimilars companies Karyo Biologics and Adello Biologics. He also founded the biosimilar advisory company PharmSci.
The final guidance on interchangeability resolves some issues, creates many new issues and inquiries, and offers a path—albeit not a clear one—that is filled with scientific and clinical challenges.
The FDA has issued its final guidance on demonstrating interchangeability of a biosimilar with its reference product to assist sponsors in showing that a proposed therapeutic protein product is interchangeable with a reference product for submitting a marketing application or supplement under section 351(k) of the Public Health Service Act (PHS Act) (42 U.S.C. 262[k]). In formulating the final guidance, the FDA had considered the comments provided in response to the invitation by the FDA when it posted draft guidance.
There were 53 comments posted by the deadline of March 20, 2017, and these came from diversified sources, including a private consumer who suggested that the Affordable Care Act not be repealed to a physician suggesting the use of suffix to biosimilars, to organizations like Coalition of State Rheumatology Organizations, which suggested that a case-by-case approach is not satisfactory and that no extrapolation should be allowed unless a product is tested in patients for all conditions. The Parenteral Drug Association emphasized the role of human factor studies and need for new regulations to provide comparison; Johnson & Johnson opposed extrapolation and recommended that testing be done in each condition, and that the labeling of biosimilar products declare when they are not interchangeable. Apotex sought clarification on the meaning of immunogenicity and if a product is given at 2 doses; would it mean 2 separate studies? The Biosimilars Council questioned whether studies will be required in conditions that are not claimed or how extrapolation will be allowed; AbbVie and other reference product companies disagreed with almost everything in the draft guidance and suggested that the FDA does not have the authority to go around the Biologics Price Competition and Innovation Act (BPCIA) requirements.
In addition to the above comments from stakeholders, there were several publications that pointed out the shortcomings of the FDA guidance in establishing biosimilarity and interchangeability. There were 2 citizen petitions and several testimonies that questioned the FDA guidelines on biosimilars.
The final guidance on interchangeability took into account several recommendations that were made by me, including using biodistribution in pharmacokinetic (PK) and pharmacodynamic studies (PK) studies, de-emphasizing clinical efficacy studies, and allowing sponsors to propose novel approaches to licensing of biosimilars and interchangeable biologics.
It is important to reiterate that the FDA guidelines are not binding, and for the same reason, they do not preclude a sponsor from making an alternate proposal to the FDA, even though most sponsors would hesitate to do so. In this paper, I am suggesting that sponsors exercise their privilege to question the guidance and fully exploit the opening provided by the FDA for alternate proposals.
Section 351(k)(4) of the PHS Act provides that an interchangeable product must be “biosimilar to the reference product” and additionally, it “can be expected to produce the same clinical result as the reference product in any given patient.” For a biological product that is administered more than once to an individual, the risk in terms of safety or diminished efficacy of alternating or switching between use of the biological product and the reference product is not greater than the risk of using the reference product without such alternation or switch.
A sponsor may choose to file a new application for approval as a biosimilar or an interchangeable biosimilar; there is no requirement that the product must be first approved as a biosimilar.
To demonstrate biosimilarity, the sponsor can refer to previously submitted data, but only after consultation with the FDA, since the FDA now requires that, in addition to the data required for demonstration of biosimilarity, developers should also include analyses of any differences in the expected PK and biodistribution, toxicities of the product, immunogenicity risk of the product, and any other factor that may affect the safety or efficacy of the product in each condition of use and patient population for which the reference product is licensed.
The use of “differences in the expected” in the above description creates a dilemma for the sponsors: there is no such expectation, since the sponsor knows that the product is a biosimilar and therefore presents "no clinically meaningful difference" from the reference product in a situation in which the sponsor is filing for a change of status to the interchangeable product.
Even in those situations in which the sponsor is making its first filing and anticipates receiving an interchangeable status, there are no expectations of differences, and none can be scientifically anticipated. A more appropriate language would be "any clinically meaningful observed differences." I am suggesting that the sponsor make this assumption regarding the expectations of the FDA.
It is noteworthy that the FDA has recognized biodistribution in PK studies as a critical parameter; I had provided details of the clinical importance of the distribution volume parameter during a testimony given to the FDA recently. More details on this subject are provided in the section on Advice to the FDA, below.
The BPCIA states that an interchangeable product “can be expected to produce the same clinical result as the reference product in any given patient,” and this statutory language forms the basis of an allowance made in the final guidance. The statement, which cannot be modified by the FDA, can be construed to state that a biosimilar interchangeable product is “capable of demonstrating the same clinical result as the reference product in any patient”; the description of the test subject as “any patient” may also mean “not necessarily the patient treated for a disease for which the product has been approved.”
As a result, sponsors may use a patient population that is different from that used to support licensure of the reference product, or in healthy subjects, provided the sponsor also gives adequate scientific justification to support the fact that the study population is adequately sensitive to detect the impact of switching (eg, differences in clinical PK and/or PD, common adverse events [AEs], and immunogenicity). The use of healthy subjects may cause a lifetime autoimmune response—a risk to be evaluated and agreed upon by the FDA.
The final guidance allows several pathways, and studies associated with each pathway, to secure an interchangeable status for a biological product.
The FDA has long advised sponsors first to secure approval as a biosimilar product and then follow it up with a change of status to interchangeable; the final guidance details how the data submitted for approval as a biosimilar product will be referred to in an interchangeable application, what additional data will be required pertaining to the studies that were required for approval as a biosimilar, and how additional structural similarity testing and safety assessments can be used to support interchangeability.
The FDA will allow a sponsor to file its first application as an interchangeable application where a single clinical study may fulfil the primary requirement of biosimilarity and the secondary qualification as an interchangeable product.
Clinical experience with the reference product and comprehensive product risk assessments (eg, regarding immunogenicity) may also affect the data and information needed to support a demonstration of interchangeability. For example, products with a documented history of inducing detrimental immune responses may require more data to support a demonstration of interchangeability than products with an extensively documented history showing that immunogenicity does not impact clinical outcomes.
Some products have relatively low structural complexity, and their reference products have no history of inducing severe immune responses related to immunogenicity. For biosimilar products that have demonstrated a low incidence of serious AEs related to immunogenicity, similar in nature and frequency to those observed with the reference product as demonstrated in clinical studies, sufficiently extensive comparative analytical data and data derived from an appropriately designed dedicated switching or integrated study will be sufficient to support a demonstration of interchangeability.
For more complex products that may have a history of rare, life-threatening AEs related to immunogenicity, postmarketing data for the product as a licensed biosimilar, in addition to an appropriately designed switching study, may be required to support a demonstration of interchangeability.
Postmarketing safety monitoring for an interchangeable product should also have adequate pharmacovigilance mechanisms in place. Generally, postmarketing data collected from products first licensed and marketed biosimilars, without corresponding data derived from an appropriately designed, prospective, controlled switching study or studies, would not be sufficient to support a demonstration of interchangeability.
However, in certain circumstances, postmarketing data from a licensed biosimilar product may be helpful as a factor when considering what data are necessary to support a demonstration of interchangeability. For example, some sponsors may wish to submit postmarketing data describing the real-world use of the biosimilar product, including certain safety data related to patient experience with some switching scenarios.
Such data may reduce uncertainty about interchangeability and thus the data needed to support a demonstration of interchangeability. The FDA will evaluate proposals to include postmarketing data in applications to support demonstrations of interchangeability on a case-by-case basis. This scenario will apply to a licensed biosimilar product where the sponsor later files for interchangeable status.
A switching study is supposed to establish whether switching results in differences in immunogenicity and PK and/or PD (if available), as compared to not switching, as the primary end point. An application will otherwise be rejected.
A switching study may also incorporate the evaluation of efficacy end points as a secondary end point or as supportive data, as the FDA concludes that clinical efficacy end points would generally be less sensitive to detect changes in exposure and/or activity that may arise as a result of alternating or switching.
The clinical PK, PD, and immunogenicity assays must be developed and validated early in product development, and must show capability of detecting changes on the selected PK and/or PD end point or end points as a result of alternating or switching between products.
The validation study should demonstrate that the assay performs similarly for both the proposed interchangeable product and for the reference product. There may be situations in which the assays may be suitable for demonstrating biosimilarity but not interchangeability; sponsors are encouraged to seek FDA advice early in the program.
A study with a lead-in period of treatment with the reference product, followed by a randomized 2-arm period—with 1 arm incorporating switching between the proposed interchangeable product and the reference product (the switching arm) and the other receiving only the reference product (the nonswitching arm)—may be appropriate when designing a switching study.
The sample size of the switching study should generally be based on PK considerations. Inter-subject variability in AUCtau or Cmax as described for the reference product should be primary considerations, and prior information on product immunogenicity incidence and consequences should also be considered to justify the sample size. Study designers should anticipate the possibility of a considerable dropout rate for reasons unrelated to the study treatment.
The number and duration of switches between the reference product and the proposed interchangeable product should take into consideration the clinical condition to be treated, the therapeutic dosing of the product, and the duration of the exposure period to each product that would be expected to cause the greatest concern in terms of immune response and a resulting impact on safety and efficacy, if any.
To capture the full PK profile, intensive PK sampling should be performed during the final exposure period after at least 3 half-lives have elapsed following the last administration of the reference product in the switching arm.
Trough PK sampling should be conducted at an appropriate time point during each exposure period to ensure that a steady state is attained, when appropriate. The timing of PD and immunogenicity sampling should be appropriately justified.
For intravenous (IV) studies, AUCtau will be considered a primary study endpoint. For subcutaneous (SC) studies, Cmax and AUCtau will be considered as co-primary study endpoints. The statistical modeling follows bioequivalence testing, and for PD parameters, the sponsor will propose margins and statistical analyses.
For studies that have an integrated design that proposes to demonstrate biosimilarity and interchangeability at the same time, a 2-part design may be warranted. Following time points for the evaluation of biosimilarity, subjects in the reference arm may be randomized in the second part of the study to a nonswitching arm or a switching arm. Such a study would need to be powered to the appropriate endpoints, including PD and PK.
The population for these studies should be adequately sensitive to detect differences (and should be part of a licensed condition of use), and the patient population may be different from the one used to support licensure of the reference. If using healthy volunteers, sponsors should weigh the benefit and risks of exposing individuals to the product.
The population for switching studies should be adequately sensitive to allow for the detection of differences as a result of switching between the reference product and proposed interchangeable product in terms of PK and/or PD, common AEs, and immunogenicity between the switching and nonswitching arms.
Sponsors should use patients in switching studies because these studies are designed to assess the impact of switching as it may be observed in clinical practice. For treatments that have a long course of therapy, sponsors should anticipate dropouts in the study and should use a scientifically justifiable method to address the increased possibility of missing data.
Sponsors may also use a patient population that is different from the one used to support the licensure of the reference product, or in healthy subjects, provided the sponsors should also provide adequate scientific justification to support the fact that the study population is adequately sensitive to detect the impact of switching (eg, differences in clinical PK and/or PD, common AEs, and immunogenicity). The use of healthy subjects may cause a lifetime autoimmune response—a risk to be evaluated and agreed upon by the FDA.
A sponsor should select a route of administration that is more likely to challenge the patient's immune response, such as subcutaneous over intravenous, where such choices are available.
A “bridge” between the non—US-licensed comparator and the US-licensed reference product is required, similar to the requirement for biosimilarity testing, with more extensive testing. The comparator product (whether it is a non–US-licensed product or a US-licensed reference product) serves as a control against which the proposed product is evaluated. However, in a switching study, the comparator product plays a different role, as a control in both the active switching arm and the control nonswitching arm.
For example, it is possible that the reference product and the non—US-licensed comparator product have subtle differences in levels of specific structural features (eg, acidic variants or deamidations), process-related impurities, or formulation. However, in the context of switching between the products, multiple exposures to each product may potentially prime the immune system to recognize subtle differences in structural features between products. The overall immune response could be increased under these conditions. This immunologic response is highly dependent on the structural differences between the proposed interchangeable product and the comparator product used in the switching study, in addition to other potential differences between the products, such as impurities or formulation.
The FDA does give sponsors an opportunity to present a scientific argument to allow the use of a non—US-licensed comparator. In most cases, the non–US-licensed comparator may have been registered using essentially the same dossier as the US-licensed product, reducing the requirements of extensive testing; whether animal toxicology or other studies are required, besides analytical similarity, is a point for discussion in early development stages.
Primary end points will be PK/PD and distribution parameters, and secondary end points will be safety, immunogenicity, and efficacy assessed descriptively.
Regarding safety, it could be reasonable for a sponsor to focus on an evaluation of all serious AEs, immune-related safety events, and AEs of interest (eg, known cardinal adverse events previously described with the use of the reference product).
The immunogenicity assessment should include, but should not necessarily be limited to, an assessment of antidrug antibody (ADA) and neutralizing antibody (NAb) incidence, ADA and NAb titers, and an evaluation of the impact of the development of ADAs and NAbs on PK, PD, safety, and efficacy. Immunogenicity assays should be adequately sensitive to detect ADAs and NAbs in the presence of drug concentrations in study samples. Sponsors should discuss with the FDA their planned evaluation of safety and immunogenicity.
The product’s degree of structural and functional complexity may influence the extent of clinical data needed to support a demonstration of interchangeability. For example, clinical data needed to support a demonstration of interchangeability of a product expected to have a single target (eg, a receptor) may be more limited than the clinical data that may be needed for a product acting on multiple targets or on less-defined biological pathways.
In addition, the extent of clinical data needed may be affected by the presence of structural features that specifically impact interchangeability (eg, features that influence patient response to one product after exposure to another product).
Data sets that include highly sensitive analytics and/or sequential analytical methods that can identify molecules with different combinations of attributes (eg, charge variants and glycoforms), as well as a comprehensive assessment of the relationships between attributes, may provide information that reduces the uncertainty about interchangeability.
These approaches could be of greater importance for more complex products because these products would have a larger number of attributes and thus a potential for greater uncertainty regarding interchangeability. Advances in analytics may allow for extended analytical characterization that affects the extent of other data and information needed to support a demonstration of interchangeability, and may in certain circumstances lead to a more selective and targeted approach to clinical studies intended to support a demonstration of interchangeability.
If a sponsor of a proposed interchangeable product believes that data from a switching study is not necessary, the FDA expects the sponsor to provide a justification for not needing such data as a part of the demonstration. For biological products that are not intended to be administered to an individual more than once, the FDA expects that switching studies would generally not be needed. For products intended to be administered more than once, sponsors are encouraged to meet with the FDA to discuss the planned development approach, including any proposed justification of why data from a switching study is not needed.
While a sponsor may seek licensure for a proposed interchangeable product for fewer than all conditions of use for which the reference product is licensed, the FDA recommends that a sponsor seek licensure for all of the reference product’s licensed conditions of use when possible.
This statement should not be misconstrued as the FDA recommending switching studies in all indications, as demanded in several comments provided to the FDA by the reference product suppliers concerning the draft guidance.
A sponsor may conduct a study or studies in 1 or more indications and then create an argument to allow extrapolation of indications—the burden of proving that the extrapolation will not result in “the risk in terms of safety or diminished efficacy of alternating or switching between use of the biological product and the reference product is not greater than the risk of using the reference product without such alternation or switch.” This exercise is no different from the request for extrapolation made in the development of biosimilars.
However, the nature of this submission will likely be different from what is expected when a sponsor requests extrapolation of a biosimilar product:
• A sponsor may obtain licensure only for a condition or conditions of use for which the reference product is licensed.
• A sponsor should choose a condition of use to study that would support subsequent extrapolation of data to other conditions of use.
• A sponsor should use a condition of use that would be adequately sensitive to assess the risk of alternating or switching between the products, in terms of safety or diminished efficacy, in a switching study.
• A sponsor may not choose an indication that was approved under section 506(c) of the Federal Food, Drug, and Cosmetic Act and 21 CFR part 601, subpart E (accelerated approval) if the anticipated clinical benefit in that indication has not yet been verified in postmarketing studies.
• Advanced structural and functional characterization may provide additional support for the justification for extrapolation.
Sponsors are well aware that the FDA guidelines are not binding to the FDA, but they often do not fully realize that they, too, are not bound by this guidance; in this final guidance, the FDA has taken a major step in stating that a sponsor may present alternate protocols and studies, provided guidance on staging the development from biosimilar to interchangeable, clarified the status of simpler and safer molecules, and provided an experimental design for a dedicated switching study.
The FDA has also stated that efficacy studies are not a good predictor of interchangeability, and therefore, by extrapolation, of biosimilarity, as stated in the guidance: “Although assessments of efficacy endpoints can be supportive, at therapeutic doses many clinical efficacy endpoints would generally be less sensitive to detect changes in exposure and/or activity that may arise as a result of alternating or switching,” leaving the focus mainly on PK/PD studies, even allowing the use of healthy subjects. Such scientifically sound and practical approaches should allow more sponsors to pursue interchangeable products, either in a series of approvals or as the first approval.
However, there remain several key elements that have not been made clear, and these include:
• A non-US comparator that has been approved using essentially the same dossier as the US-licensed product and that meets all other requirements of BPCIA should be allowed without any additional testing, except for release, using the specification applied to the biosimilar candidate.
• For products administered intravenously, the FDA now states that only the AUC should be used as the comparative parameter, yet continues to require the statistical modeling applied to bioequivalence testing—a scientific glitch, since products given by intravenous routes are, by definition, bioequivalent. The only reason the AUC might vary is because of differences in the elimination rate constant and in the distribution volume. By using the AUC alone, the impact of receptor binding (distribution volume) and elimination rate constant (how the body sees the molecular structure) are heavily confounded. It will be a rare event if these studies ever fail, let alone fail in a situation where subtle differences in the PK parameters are sought. The use of individual parameters has been widely established, and this change will also remove the need for the statistical model for bioequivalence testing. The end result of this change will be a more robust model, with fewer subjects and data that are more clinically meaningful. I presented the concept of distribution volume in 1976, and the concept has been used widely to evaluate the safety and efficacy of drugs, more particularly the drugs that bind to receptors.1-3
• While the FDA has allowed sponsors to discuss the study protocols required for drugs that are not administered as repeated doses, it would have been more appropriate if the FDA were to define the available options. In my opinion, this is a simpler question; since a biosimilar product comes with an assurance that it has "no clinically meaningful difference" from the reference product, simpler side-by-side multiple dose PK studies should suffice to establish interchangeability.
The final guidance on interchangeability resolves some issues, creates many new issues and inquiries, and offers a path—albeit not a clear one—that is filled with scientific and clinical challenges. While the FDA has opened that path for the sponsors to offer alternate study designs to satisfy the requirements to meet the safety and efficacy standards set forth in the BPCIA, the FDA also opens up many possibilities for the sponsors to realize the complexities and the costs involved in seeking an interchangeable status.
1. Niazi S. Volume of distribution as a function of time. J Pharm Sci. 1976;65(3):452-4. doi: 10.1002/jps.2600650339.
2. Niazi S. Volume of distribution and tissue level errors in instantaneous intravenous input assumptions. J Pharm Sci. 1976;65(10):1541-1543. doi: 10.1002/jps.2600651034.
3. Wesolowski CA, Wesolowski MJ, Babyn PS, Wanasundara SN. Time varying apparent volume of distribution and drug half-lives following intravenous bolus injections. PLoS One. 2016;11(7):e058798. doi: 10.1371/journal.pone.0158798.