OR WAIT null SECS
In this question-and-answer discussion of "generic" biologics, the authors contend small changes in manufacturing of biologics dramatically affect the safety and efficacy of the therapeutic molecule.
During the past few years, several people and organizations have stated public positions on the feasibility of creating a legal and regulatory framework for the facilitated approval or licensure of biologics, based on abbreviated applications (1-3). However, public debate on follow-on biologics is difficult because of the considerable complexity of the interrelated technical, legal, and regulatory issues. Inappropriate use of biologics terminology and misconceptions of the regulatory grounds for safe approval of biologics result from an incomplete understanding of the basis for approval of existing biologics within the same therapeutic category. We seek to clarify some of these issues in this article. We maintain, however, that the regulatory process for the approval of biologics must be shaped by three cardinal principles:
A biologic is a prophylactic, an in vivo diagnostic, or a therapeutic substance that can be made only by a living system and that has a large, complex, inherently heterogeneous molecular structure.
Biologics have been defined in disparate terms, depending on the scientific, regulatory, or legal context - no comprehensive and universally precise definition exists. For the purposes of discussion of follow-on biologics, we use the following operational definition:
A biologic is a prophylactic, an in vivo diagnostic, or a therapeutic substance that:
A "generic" drug is approved by reference to a strict definition of sameness to the innovator's product. Sameness, however, cannot be determined for biologics from different manufacturers because of the complex nature of the products and their manufacturing processes. Therefore, "generic biologics" produced by different manufacturers cannot exist.
The phrase "generic biologics" is often used to mean products that would be copies of, and be competitive with, biologics produced by pioneers (that is, innovative companies that developed and produced the first version of a particular product). FDA would approve these products by reference to the clinical data generated by the pioneer. This designation is derived from the regulatory process for approval of generic (chemical) drugs, which was created by the Hatch-Waxman amendments legislation of 1984 (4) and is codified in Â§505(j) of the Food, Drug, and Cosmetic Act (FD&C) (5). However, because of major technical differences between drugs and biologics, FDA stated that the regulatory process under Â§505(j) would not be appropriate for the approval of biologics because it would result in the approval of unsafe products (6). More recently, FDA Commissioner Mark McClellan stated that, "As a scientific matter, it is true that certain biological products, due to their inherent structural complexity, heterogeneity, and manufacturing process, do not currently lend themselves to being copied generically" (7).
Because the "sameness" standard under Â§505(j) defines the correct use of the word "generic," and the properties of biologics do not allow the standard of "sameness" to be met, there can actually be no "generic biologics."
"Follow-on biologics" are second and subsequent versions of biologics that are independently developed and approved after a pioneer has developed an original version. Follow-on biologics may, or may not, be intended to be molecular copies of the innovator's product. They do, however, depend on the same mechanism of action and are intended to be used for the same indication.
Currently, follow-on biologics are approved or licensed by submitting to FDA a full product license application - either a New Drug Application (NDA) or a Biologic License Application (BLA). Those products that have been approved or licensed are not copies of the original, innovative product. They resemble instead the innovative products in much the same way that members of a class of drugs resemble each other (statins, triptans, and H2 antagonists, for example). It is possible that some follow-on products could be approved safely on the basis of fewer or smaller efficacy studies than were required for the approval of the pioneer. Table 1 gives some examples of innovator and follow-on biologics.
The concept of "follow-on biologics" or "generic biologics" does not include the special circumstances in which a manufacturer may make changes to its production process without conducting new clinical trials (see the answer to Question 10 for further explanation).
With the success of the Human Genome Project and investments in genomics, bioinformatics, and proteomics by pharmaceutical companies and others, biotechnological medicines (therapeutics and preventives) present new opportunities for innovation and powerful, new therapeutic modalities. These therapies have already provided important new treatment options and, in some cases, revolutionized the standard of care for certain diseases. The promise of these modalities is even brighter. Patient confidence in the safety and efficacy of these products, however, must be maintained.
Another more subtle concern exists for regulators and the pharmaceutical industry. Chemical drugs have simple chemical structures, which can be completely characterized and specified. The Hatch-Waxman Act, therefore, established very rigorous standards of identity (sameness) of generic drugs to innovator drugs for the approval of those generics under Â§505(j) of the FD&C. Biologics and biotech drugs cannot be approved under Â§505(j) the same way, except by attenuation or violation of that sameness standard. Unlike chemical drugs, individual batches of biologics are heterogeneous at the molecular level as a result of the inherent random variability in the living processes by which they are made. These molecular differences are even greater between versions of the same molecule produced by different manufacturers. Therefore, approval of biologics under Â§505(j) would not only mean approval of products with uncertain clinical properties but would also set a precedent under which drugs significantly different from the innovator drug could be approved as generics. That would destroy the careful legal balance between the innovator and generic industries that was framed in the Hatch-Waxman Act and would have adverse consequences for the control of drug costs and public health.
Table 1. Some examples of innovator and follow-on biologics
There is no evidence to support this speculation.
Both CBER and CDER have been intimately involved in the creation of harmonized standards for biotechnological and biological products as part of the International Conference on Harmonisation (ICH) and under the FDA Modernization Act of 1997 (FDAMA) during the past decade. In announcing the recent move of CBER's Office of Therapeutics Research and Review (OTRR) to CDER, Lester Crawford, FDA deputy commissioner stated, "Current FDA policy on generic biologics will not be affected by this decision" (8). As quoted earlier, FDA Commissioner Mark McClellan has agreed that biologics do not lend themselves to being copied generically (7). This position has been common to both CBER and CDER for many years and is manifest in the fact that neither Center has approved any follow-on protein therapeutics under either the Public Health Service Act (PHS) or the FD&C Act. So despite the recent reorganization within FDA and the possibility of more in the future, there is presently no evidence that the standards by which biologics are approved are going to change.
Safe generic drugs have been approved with abbreviated data packages because it is possible to show that the generic product is chemically and pharmaceutically identical to the pioneer's product, to within very tight confidence limits. Therefore, it is scientifically reasonable to expect that the clinical properties of the pioneer's product will be shared by the generic. Because the follow-on biologic cannot be shown to be identical to the innovator product, this relationship does not hold for biologics.
The manufacturing process for each biologic defines, to a significant extent, the product because biologics are based on living cells or organisms whose metabolisms are inherently variable. Moreover, apparently small differences between manufacturing processes can cause significant differences in the clinical properties of the resulting products. For example, when the manufacture of tissue plasminogen activator (tPA) was changed from production in roller bottles to production in stirred bioreactors (tanks), the resulting product showed differences in its glycosylation pattern and degree of internal cleavage and possessed a different pharmacokinetic profile and dose response in humans.
There will always be differences between manufacturing processes designed by different manufacturers, especially those in competition with each other. As we stated, it is not possible to predict what the effects of these differences might be on the clinical properties of the two products. Therefore, it is not scientifically reasonable or safe to simply expect that the clinical properties of a pioneer's product would be shared by the follow-on product. The only way to characterize the clinical properties of the follow-on product is to evaluate them in appropriately designed clinical studies.
These products are not generic, because the approval or licensure of each product was based on evaluation of full sets of independent data.
Different manufacturers - as independent innovators - have developed each of these products and compiled complete regulatory applications, based on unique product, process, and clinical data sets. These applications received full and independent review by FDA. They cannot be presumed to be substitutable for each other (that is, they are not AB-rated, for bioequivalence, in FDA's Orange Book) (9), but they can represent therapeutic alternatives. A label claiming substitutability can only be supported by an appropriate clinical trial comparing the safety and efficacy of the product directly.
It is not accurate to say that there can be only one innovator for each biologic.
There are many instances of multiple manufacturers of therapeutically alternative biologic products on the market (growth hormone, insulin, Factor VIII, erythropoietin, and tPA, for example). Each of these alternative biologics were independently studied and demonstrated to be clinically safe and effective. Independent clinical trials were conducted for each. Interestingly, even though these products might appear to be similar, there can be differences between their clinical properties. For example, Betaseron (Berlex Laboratories) and Avonex (Biogen, Inc.) are both versions of beta-interferon used to treat multiple sclerosis. Avonex has 166 amino acids, whereas Betaseron has 165; Avonex is produced in mammalian cells, whereas Betaseron is produced in bacteria. Incidents ofpatient adverse events differ between the two products. These differences between the products are of the same order as differences between members of a drug class (see the answer to Question 3).
Likewise, the amino acid sequences of the alfa-interferons Intron A (Schering Corporation) and Roferon (F. Hoffmann-La Roche Ltd.) also differ by one amino acid. The licensure by CBER of these two products on the same day - on the basis of independent clinical trials - shows clearly that there can be multiple innovators for each biologic.
Forcalcitonin (Unigene Laboratories, Inc.) is not a true generic biologic.
Forcalcitonin is produced by expression of the gene for calcitonin from salmon, but the product is an oligopeptide containing only 32 amino acids, which does not satisfy our definition for a biologic (see the answer to Question 1). Forcalcitonin can be produced by means other than a living system; indeed, the pioneer's product, Miacalcic (Novartis), is produced by chemical synthesis. Forcalcitonin and Miacalcic are drugs (pharmaceuticals). The fact that Forcalcitonin was produced by recombinant technology is not relevant to the discussion of generic biologics.
It is not true that an innovator can always introduce process changes without doing clinical trials. Whether that can be done depends on the nature of the change and the results of comparative product characterization studies. FDA still requires clinical validation of process changes in some cases.
The innovator has a comprehensive database of every step of the manufacturing process and key intermediates and has established many in-process controls and reference standards. Generally, changes are minor alterations to a well-understood, extensively validated and licensed process, with all other aspects of production remaining unchanged. The innovator is able to compare product made before and after the process change and show that the change has had essentially no effect on the product made. By contrast, a follow-on manufacturer does not have access to the pioneer's data, key intermediates, reference standards and reagents, and the complex or unique validated methods (such as bioassays) that would enable a comparison of the follow-on process with that of the innovator's. Even if all the pioneer's manufacturing data were available, the inherent differences in production and analytical methodologies, facilities, and reagents between manufacturers would render the data nontransferable (see the answer to Question 6).
Comparability of a product made before and after a process change (intramanufacturer) is a different concept from establishing the equivalence of two products made by different manufacturers by necessarily different processes (intermanufacturer). On this basis, the transfer of a manufacturer's process between facilities can be supported by a demonstration of comparability, whereas the establishment of a (new) process by a different manufacturer cannot.
The circumstances of the development and approval of Avonex were unique and do not support a general case that transference of clinical data between products made by differentbiologic manufacturers is possible.
In the Avonex case, the two manufacturers were not in competition with each other. The consequence of this was an unrestricted flow of manufacturing information and key intermediates between contractually bound companies. It was this exchange that enabled the second manufacturer (Biogen, Inc.) to show that its product was indistinguishable from that of its partner Bioferon (a German company in which Biogen held 50% of the equity), which had manufactured the earlier batches of the product. FDA allowed the clinical data generated with Bioferon's product to be transferred to Biogen's product, as process comparability had been demonstrated. Interestingly, Biogen's first attempt to manufacture the same product that Bioferon produced failed, in spite of Biogen's advantageous relationship. The demonstration that the two versions of the product were the same occurred only during the second manufacturing attempt.
Clearly, the relationship between the parties in the Avonex case is completely different from the competitive relationship that would exist between a pioneer and a follow-on manufacturer.
Manufacturers do not test for all the impurities in a biologic - they test for those impurities that they know to be relevant to the process used and for which analytical tools (usually dedicated and specially developed) are available.
Methods to detect and quantitate impurities in the molecules of a biologic are not well-developed, particularly for those impurities that are process-dependent. This is another key distinction between biologics and chemical drugs, in which impurities can be more sensitively identified and monitored. The ultimate demonstration of the safety and efficacy of the product is in the human clinical trials. If the outcome of these clinical trials were predictable, even the innovator of the product might not need to perform these tests. For example, manufacturers could demonstrate only that there were no physicochemical differences between their products and the natural substances that they are intended to replace, and this would be sufficient.
Reagents, test methods, and reference standards are developed specifically to suit the process being used and the product being created and are not generally applicable to any process or product.
Specifications for one product and process are idiosyncratic and unlikely to be applicable to a process using a different cell line, recombinant construct, purification process, and validation profiles or being manufactured in different facilities and evaluated by different analytical methods (10). Because of the inherent heterogeneity of biologics, it is not possible to analyze and characterize them with such precision that no clinical differences would result from two products that are indistinguishable by in vitro tests. If this were not so, the present regulatory rigor that innovators encounter when introducing process changes would not be warranted. Recognizing these points, the leaders of the international regulatory community (in the United States, the European Union, and Japan) have agreed to the ICH guidance document (Q6B), which specifically states that a set of specifications is unique to each product and to the process used to generate the product (including its validation) (11).
Establishing bioequivalence (the same rate and extent of absorption) is relatively complicated for biologics, and for many biologics, it may be impossible to establish bioequivalence at all because:
Sometimes, the use of pharmacodynamic or surrogate end points can facilitate bioequivalence evaluations of biologics. However, in many cases there is no alternative to an appropriately designed clinical trial for establishing therapeutic equivalence. It is also important to remember that even when bioequivalence of two versions of a biologic can be established by conventional techniques, these data provide no information about the safety of the products (see the answers to Questions 6 and 15).
Safety is paramount and can only be established if a statistically relevant number of patients are tested. Similarly, demonstrating efficacy requires a statistically relevant number of patients, which may be greater or less than the number needed for the innovator product.
Efficacy can be established in a small number of patients only if a precise end point or a validated surrogate for a clinical outcome is established and endorsed by FDA (for example, the viral load of HIV is more readily established than the morbidity or mortality for AIDS). For a pioneer sponsor, testing a relatively small number of patients in safety studies may be acceptable so that a breakthrough product is allowed to reach the market, especially if there is no therapeutic alternative. This consideration does not apply to a follow-on product because the risk-to-benefit balance is less compelling when patients are already being treated.
The presence of different impurities in a follow-on product would be particularly pertinent to safety and can only be established in human clinical trials. Another important consideration in showing the safety of a follow-on biologic for treatment of chronic conditions is the demonstration of its safety in patients who have already been treated with another similar product ("switching" safety). Therefore, the number of patients required to show the safety of a follow-on product will generally be a significant percentage of the number enrolled by the pioneer and may even exceed that number.
The efficacy of follow-on products must usually be determined by comparative clinical studies against the pioneer's product. Comparative studies should be the basis for any official recognition of substitutability. These studies are particularly relevant to products for which bioequivalence is not a validated predictor of efficacy and cannot be demonstrated formally.
Rationally, the quantity of data required to license follow-on biologics will vary for each product and will depend on each product's complexity and on how much is known about the product's mode of action. It would be important that this need for flexibility not be constrained by rigid legislative language.
Manufacture of biological products depends on rigorous controls and repeated in-process testing, so the facilities must be more integrated than is the case for drugs.
The potential for microbiological contamination of starting materials, the delicate nature of biological macromolecules, and the extreme sensitivity of the living cells that produce biologics impart complex requirements for fermentation, aseptic processing, storage, and testing. For example, a typical manufacturing process for a chemical drug might contain 40–50 critical tests. The typical process for a biologic, however, might contain 250 or more critical tests. These issues translate into very specialized engineering and process designs that transcend plant and machinery to encompass subtle human factors. Consequently, construction and validation of new facilities is disproportionately expensive and time-consuming. This is, in part, why there is, at present, a global shortage of biomanufacturing capacity.
Manufacturing changes for drugs, although still onerous, can be more readily accommodated than those changes for biologics. Drug batches are released according to specifications for the drug substance and the final product, without the emphasis on extensive characterization and critical testing required of biologics.
Understanding these issues, it is clear that approval or licensure of each follow-on biologic must rely on specific clinical data that support its safety and efficacy. Assessments cannot be based solely on examinations of data relating to other members of the same class of drugs.
Whether two or more biologics can be substituted for each other safely is a second order issue:
In most cases, the only satisfactory basis of assessment will be direct comparison in appropriately designed trials.
Just as manufacturing and quality data from comparability studies reduce risk when conducted to support changes to the product or process of a licensed biologic, data from clinical trials are needed to support independent approval and licensure of follow-on products or approval of the substitutability of similar products within a class of drugs. Absence of such data constitutes substantial risk.
(1) Bende, S., presented at the Hatch–Waxman Update 2002 conference (FDLI, Washington DC, December 2002).
(2) "Generic Biologics: Xigris Approval Suggests Precedent for FDA, GPhA Says," FDA Week, p. 32 (16 December 2002).
(3) Hatch, O., FDA Week, p. 16 (12 August 2002).
(4) "The Drug Price Competition and Patient Term Restoration Act of 1984," U.S. Code Title 35, Section 156 (Hatch–Waxman Act, S-1538, HR-3605, 24 September 1984).
(5) "Federal Food, Drug, & Cosmetic Act," U.S. Code Title 21, Section 355(j) (1938, revised 1999).
(6) Zoon, K., "CBER Chief: Generic Biologics a Problem from Scientific Standpoint," FDA Week, p. 18 (20 April 2001).
(7) McClellan, M., "FDA Generic Biologics Policy May Change Under McClellan," Pink Sheet, p. 7 (14 October 2002).
(8) Crawford, L.M., "FDA to Consolidate Review Responsibilities for New Pharmaceutical Products," FDA News (6 September 2002).
(9) CDER, Electronic Orange Book: Approved Drug Products with Therapeutic Equivalence Evaluations (FDA, Rockville, MD, updated 29 May 2003). Available at www.fda.gov/cder/ob.
(10) Copmann, T. et al., "One Product, One Process, One Set of Specifications: A Proven Quality Paradigm for the Safety and Efficacy of Biologic Drugs," BioPharm 14(3), 14–24 (March 2001).
(11) ICH, Q6B:Specifications for New Drug Substances and Products: Test Procedures and Acceptance Criteria for Biotechnological/Biological Products, CPMP/ICH/365/96 (Geneva, 1999). Also Federal Register 64(159), 44928–44935, Doc. 99-21352 (18 August 1999). This is a test.