ARTHRITIS & RHEUMATISMVol. 63, No. 10, October 2011, pp 2848–2850DOI 10.1002/art.30505© 2011, American College of Rheumatology

COMMENTARY

Biosimilars: The Debate Continues

Robert A. Colbert1 and Bruce N. Cronstein2

There is no question that biologics have had amajor impact in rheumatology as in other areas ofmedicine, with many more potentially efficacious drugsin various stages of development. However, the high costof biologics is a growing concern for consumers andinsurers, particularly as new products become availableand their use for the treatment of autoimmune andautoinflammatory diseases continues to expand. World-wide, an estimated $130 billion was spent on biologics in2009, and $6.6 billion was spent on Rituxan (rituximab)alone in 2010 (1). Biosimilars (also referred to asfollow-on biologics) are intended to be subsequent ver-sions of reference biologics with comparable safety andefficacy profiles, analogous to generics of brand namedrugs. Biosimilars become potential cost-saving alterna-tives once patent and exclusivity rights on reference

biologics have expired. They also represent a multibil-lion dollar market for producers, and a threat to inno-vators who produce the reference biologics, as they willundoubtedly experience substantial losses in marketshare. The stakes are enormous.

In 1984 the US Congress approved the DrugPrice Competition and Patent Term Restoration Act(also known as the Hatch-Waxman Act), establishing anabbreviated new drug application (ANDA) pathway forgenerics. The main difference between a new drugapplication and an ANDA for a generic is that the latterdoes not require preclinical and clinical studies ifbioequivalence—the rate and extent of absorption—canbe demonstrated. This saves time and considerableexpense in the approval process. Generic equivalents ofmany drugs have proven to be relatively straightforwardto produce, and consequently the use of generics hasrisen dramatically in the last two decades, with consumersavings estimated to be �$10 billion a year in 2004 (2).Can a similar abbreviated approval pathway be used forbiosimilars to reap comparable benefits for consumersand insurers? This is where the debate begins.

In most instances, the term “biologics” refers to aclass of medications produced by living cells usingrecombinant DNA technology (3). The recombinantDNA is introduced into the cells, which have the ma-chinery to read and decode the DNA and producecomplex proteins that can be purified and used formedicinal purposes. Typical biologics include proteinsthat are intended to be almost identical to naturalproducts that the body makes, often used as replacementtherapy in the case of a genetic deficiency, or simply toaugment the body’s response to an endogenous product.Other biologics include monoclonal antibodies that bindto soluble or cell surface proteins and block overactivepathways or cells, or creatively engineered proteins thatmimic receptors or naturally occurring receptor antago-nists, but are modified to be soluble or more stable, andthus prevent activation of signaling pathways.

The problem arises when one tries to develop acopy, or biosimilar, of a reference biologic. The chal-

Dr. Colbert’s work was supported by the NIH (NationalInstitute of Arthritis and Musculoskeletal and Skin Diseases Intramu-ral Research Program grant AR-041184). Dr. Cronstein’s work wassupported by the NIH (grants AR-56672, AR-56672S1, and AR-54897) and by the New York University–New York City Health andHospitals Corporation (NYU-HHC) Clinical and Translational Sci-ence Institute, which receives NIH funding (grant UL1-RR-029893).

1Robert A. Colbert, MD, PhD: National Institute of Arthritisand Musculoskeletal and Skin Diseases, NIH, Bethesda, Maryland;2Bruce N. Cronstein, MD, New York University School of Medicine,New York, New York.

Dr. Colbert has a sibling who is employed by Abbott. Dr.Cronstein has received consulting fees from CanFite Biopharmaceu-ticals, Savient, Bristol-Myers Squibb, Novartis, Cypress, Roche, Cell-zome, TAP (Takeda), Prometheus, Regeneron (Westat, DSMB),Sepracor (Sunovion), Amgen, Endocyte, Protalex, Allos Therapeutics,CombinatoRx (Zalicus), and Kyowa Hakko Kirin (less than $10,000each); he owns stock or stock options in CanFite Biopharmaceuticals.Dr. Cronstein holds patents for the use of adenosine A2A receptoragonists to promote wound healing and the use of A2A receptorantagonists to inhibit fibrosis, which are licensed to King Pharmaceu-ticals and for which he receives licensing fees; he also holds or has filedapplications for the use of adenosine A1 receptor antagonists to treatosteoporosis and other diseases of bone, the use of adenosine A1 andA2B receptor antagonists to treat fatty liver, and the use of adenosineA2A receptor agonists to prevent prosthesis loosening.

Address correspondence to Robert A. Colbert, MD, PhD,Pediatric Translational Research Branch, National Institute of Arthri-tis and Musculoskeletal and Skin Diseases/NIH, 10 Center Drive,CRC, Room 1-5142, MSC 1102, Bethesda, MD 20892. E-mail:colbertr@mail.nih.gov.

Submitted for publication April 28, 2011; accepted in revisedform June 14, 2011.

2848

lenge is quite different from reproducing small chemicalcompounds to produce a generic drug. The molecularand structural complexity of most biologics makes itdifficult if not impossible to produce exact replicas. Evenif the same recombinant DNA sequence is used, thecomposition of the final product is influenced by avariety of factors, including the cells used to produce theprotein, culture conditions, posttranslational modifica-tions, purification methods, stabilization and storageconditions, and packaging. Manufacturers of referenceproducts may use proprietary growth and purificationconditions, and therefore knowing the exact sequence ofthe DNA construct used to produce the protein is notsufficient to produce the same biologic product. Whatmay seem to be subtle differences between a biosimilarand the reference biologic can affect efficacy as well assafety and immunogenicity.

In March 2010, the Patient Protection and Af-fordable Care Act, more commonly known as HealthCare Reform, was signed into law in the US. It outlinesan abbreviated approval pathway for biologic productsthat are demonstrated to be “highly similar” (biosimilar)to, or “interchangeable” with, a Food and Drug Admin-istration (FDA)–licensed biologic product. Known asthe Biologics Price Competition and Innovation Act(BPCIA), this legislation is similar to the 1984 DrugPrice Competition and Patent Term Restoration Actthat established an abbreviated approval pathway forgenerics. Currently, according to the FDA definition (4),a product is biosimilar if data from analytical studies,animal studies, and a clinical study or studies demon-strate the product to be “highly similar” to the referenceproduct, notwithstanding minor differences in clinicallyinactive components, and if there are no clinically mean-ingful differences in terms of safety, purity, and potency.To meet the higher standard of interchangeability re-quires demonstrating that the product can be expectedto produce the same clinical result as the referenceproduct in any given patient, and, if the biologic isadministered repeatedly, there is no greater risk in termsof safety or diminished efficacy from alternating orswitching than there would be from continued adminis-tration of the reference product. If interchangeabilityhas been demonstrated, then the pharmacist can substi-tute the biosimilar for the reference product without theintervention of the prescribing health care provider. Atissue are the nature and quantity of evidence needed toestablish biosimilarity and interchangeability for bio-logics. In the US the bar seems to be set pretty high.

In November 2010, the FDA held a public hear-ing to obtain input from various stakeholders on how to

implement the BPCIA (5). The results were not surpris-ing considering the financial and health care implica-tions. Innovators who have developed and produce thereference biologics think the bar should remain high,while companies that produce generics favor lowerclinical standards to demonstrate biosimilarity. Patientadvocacy organizations and medical specialty profes-sional societies insist, understandably, that safety andefficacy be top priorities. While it should be emphasizedthat everyone advocates safety and efficacy for biosimi-lars, the perceptions of how much testing should berequired before a drug is marketed vary dramatically.Other issues received attention at the FDA hearing,including the question of whether a biosimilar of areference biologic approved for more than one indica-tion would need to be tested in each of those indications,and whether the clinical target should be equivalence ornoninferiority.

While the debate goes on in the US, Europeanregulatory authorities have continued to move forward.The European Medicines Agency (EMA) was the first todevelop a comprehensive approach to licensing biosimi-lar agents (http://www.ema.europa.eu/). The EMA hasdeveloped a flexible case-by-case approach to licensingbiosimilars with clear and precise regulatory tracks forlicensing of specific agents such as growth hormone,insulin, and filgrastim. Preclinical data that establish thesimilarity of the biosimilar to the reference biologic withrespect to antigen or ligand binding, evidence of theidentity of the agents (but not the glycosylation state inthe case of larger proteins), and results of appropriatelydesigned preclinical pharmacokinetic and pharmacotoxi-cologic tests in animals are required. In addition, regu-lators scrutinize manufacturing standards. Require-ments for demonstration of pharmacokinetics andpharmacodynamics in humans are also necessary, butthese requirements are eclipsed by the need to demon-strate similar clinical efficacy and safety of the biosimilarcompared to that of the reference biologic. For anumber of proteins, regulatory pathways have beenestablished with clinical trial requirements ranging fromnone (for human insulin preparations) to trials compar-ing hundreds of patients taking either the original bio-logic or the biosimilar.

Because of the structural complexity of biologics,a considerable problem for establishing an abbreviatedregulatory pathway for biosimilars is immunogenicity.One example where this has had significant conse-quences is the development of pure red cell aplasia inpatients receiving recombinant erythropoietin. Althoughthis may be a rare occurrence, there is a formal EMA

COMMENTARY 2849

requirement for an extended pharmacovigilance planthat must be approved and in place. The requirement forpharmacovigilance dictates that a single generic namecannot be applied to the reference biologic and subse-quent biosimilar agents, since it would not be possible todetect a novel toxicity of the biosimilars. The issue ofapproval for single versus multiple indications andwhether additional clinical testing will be required alsoremains to be resolved. This will likely be handled on acase-by-case basis. Nonetheless, a number of biosimilaragents have already been licensed in Europe (althoughnone for rheumatology indications) following the regu-latory guidance offered by the EMA (1). In addition, theWorld Health Organization recently issued a guidancedocument for licensing of biosimilars, as have Korea andSingapore. Canada, Australia, and Japan are now final-izing regulatory guidance documents, and it is likely thatmost other countries will follow suit rapidly. Most of theregulatory pathways adopted to date are similar.

The EMA has recently taken another step for-ward with guidelines for biosimilar monoclonal antibod-ies (6). While the Committee for Medicinal Products forHuman Use (CHMP) of the EMA has said that com-parative phase III clinical studies demonstrating safetyand efficacy for biosimilars are necessary, they haveplaced the focus clearly on designing trials to establish“similarity” to the reference compound without neces-sarily demonstrating patient benefit. The guidelines ad-vise using selected patient populations where the bio-logic is known to be highly efficacious to look for similareffects of the biosimilar. Differences would be expectedto be more readily apparent in smaller numbers ofindividuals. The use of more homogeneous patientpopulations with similar disease severity and a lack ofcomorbidities would further facilitate these comparisonsand keep sample sizes down, thus minimizing cost. Increating these guidelines, the CHMP/EMA have estab-lished a lower, yet quite rational, hurdle for establishingmonoclonal antibody biosimilarity. The new guidelinesalso begin to address the issue of multiple indications.They outline situations where extrapolation of clinicalefficacy and safety data for another indication may bepossible, such as when the mechanism of action of the

biologic is the same. When the mechanism of action fortwo or more indications is different, then the situationwill be more complex and require additional testing. Itshould be noted that these guidelines are currently openfor comment, and are likely to be further revised prior totheir general acceptance.

The American College of Rheumatology hasdeveloped a position statement on biosimilars. In es-sence, it puts the patient first, emphasizing that safetyand efficacy are paramount. However, it can be arguedthat unless biosimilars are much less expensive thantheir reference biologics, the entire exercise is academic.Decisions about biosimilarity and interchangeabilityneed to be data driven, but the process needs to recog-nize the heterogeneity of different classes of biologics,with flexibility built into the guidelines. For biosimilarsto reduce the cost of health care and increase theavailability of effective treatments will require that theyare significantly less expensive than reference biologics,and that they are prescribed with confidence. Givenissues related to immunogenicity, interchangeabilitymay be one of the highest hurdles to clear. Stay tuned;the debate is far from over.

AUTHOR CONTRIBUTIONSDrs. Colbert and Cronstein drafted the article, revised it

critically for important intellectual content, and approved the finalversion to be published.

REFERENCES

1. Carey K. Biosimilars encircle Rituxan, US debates innovator exclu-sivity. Nat Biotech 2011;29:177–8.

2. Haas JS, Phillips KA, Gerstenberger EP, Seger AC. Potentialsavings from substituting generic drugs for brand-name drugs:medical expenditure panel survey, 1997–2000. Ann Intern Med2005;142:891–7.

3. Jelkmann W. Biosimilar epoetins and other “follow-on” biologics:update on the European experiences. Am J Hematol 2010;85:771–80.

4. US Department of Health and Human Services, Food and DrugAdministration. Docket no. FDA-2010-N-0477. Approval pathwayfor biosimilar and interchangeable biological products; public hear-ing; request for comments. Federal Register vol. 75, no. 192. 2010Oct. 5. URL: http://edocket.access.gpo.gov/2010/pdf/2010-24853.pdf.

5. Mullard A. Hearing shines spotlight on biosimilar controversies.Nat Rev Drug Disc 2010;9:905–6.

6. Similarity trials [editorial]. Nat Biotech 2011;29:1.

2850 COLBERT AND CRONSTEIN