`
`Lessons learned from biosimilar epoetins and insulins
`
`Article in The British Journal of Diabetes & Vascular Disease · April 2010
`
`DOI: 10.1177/1474651409355454
`
`CITATIONS
`44
`
`2 authors, including:
`
`Martin Kuhlmann
`Vivantes Klinikum im Friedrichshain
`
`177 PUBLICATIONS 5,731 CITATIONS
`
`SEE PROFILE
`
`READS
`576
`
`Some of the authors of this publication are also working on these related projects:
`
`Dry weight measurement View project
`
`dry weight measurement View project
`
`All content following this page was uploaded by Martin Kuhlmann on 06 May 2016.
`
`The user has requested enhancement of the downloaded file.
`
`Exhibit 2071
`Page 01 of 10
`
`
`
`The British Journal of Diabetes & Vascular
`Disease
`
`http://dvd.sagepub.com/
`
`
`Lessons learned from biosimilar epoetins and insulins
`Martin Kuhlmann and Michel Marre
` 2010 10: 90British Journal of Diabetes & Vascular Disease
`
`DOI: 10.1177/1474651409355454
`
`The online version of this article can be found at:
`
` http://dvd.sagepub.com/content/10/2/90
`
`Published by:
`
`
`http://www.sagepublications.com
`
`
`
`Additional services and information for can be found at:The British Journal of Diabetes & Vascular Disease
`
`
`
`
`
`Email Alerts:
`
`
`http://dvd.sagepub.com/cgi/alerts
`
`Subscriptions:
`
`
`http://dvd.sagepub.com/subscriptions
`
`Reprints:
`
`
`http://www.sagepub.com/journalsReprints.nav
`
`
`
`
`http://www.sagepub.com/journalsPermissions.navPermissions:
`
`Citations:
`
`
`http://dvd.sagepub.com/content/10/2/90.refs.html
`
`>>
`
`Version of Record
`
`- Apr 12, 2010
`
`What is This?
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from by guest on October 11, 2013dvd.sagepub.com by guest on October 11, 2013dvd.sagepub.com by guest on October 11, 2013dvd.sagepub.com by guest on October 11, 2013dvd.sagepub.com by guest on October 11, 2013dvd.sagepub.com by guest on October 11, 2013dvd.sagepub.com by guest on October 11, 2013dvd.sagepub.com by guest on October 11, 2013dvd.sagepub.com by guest on October 11, 2013dvd.sagepub.com
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Exhibit 2071
`Page 02 of 10
`
`
`
`
`
`
`
`
`
`
`
`ACHIEVING BEST PRACTICE
`
`Lessons learned from biosimilar epoetins
`and insulins
`
`MARTIN KUHLMANN1, MICHEL MARRE2
`
`AbstractPatients with diabetes and renal failure may already
`
`be receiving biosimilar epoetin and may receive
`biosimilar insulin in the near future. Because these
`biosimilar pharmaceuticals (or follow-on biologics) are
`complex protein molecules manufactured in lengthy and
`inherently variable processes involving living organisms,
`they have the potential to induce an immunogenic, rather
`than a therapeutic, response. This response is dependent
`as much on the method of manufacture and formulation,
`as on the protein itself. Apparently small and innocuous
`differences in manufacture and formulation can lead to
`unforeseen clinical consequences. This article discusses
`two case studies illustrating this principle, that of three
`insulin formulations which were physicochemically simi-
`lar to comparator insulins, but with pharmacokinetic and
`pharmacodynamic profiles sufficiently different to have
`potentially serious clinical consequences and that of
`Eprex, for which an apparently minor change in one for-
`mulation caused an upsurge of cases of pure red cell
`aplasia which resulted in fatalities or complete transfu-
`sion dependence. Comprehensive and rigorous testing
`and long-term pharmacovigilance programmes are essen-
`tial to detect and forestall such consequences.
`Br J Diabetes Vasc Dis 2010;10:90–97.
`
`Key words: biosimilars, EMEA, epoetin, Eprex, insulin,
`Marvel, pure red cell aplasia.
`
`Biosimilar (follow-on biologic): A therapeutic protein whose
`active substance is shown by appropriate testing to have simi-
`lar physicochemical, preclinical and clinical properties to an
`originator therapeutic protein. Because it is manufactured in
`living systems with inherent variability, biosimilar proteins can-
`not be identical to the corresponding originator protein.
`
`1 Medicine and Nephrology, Vivantes-Klinikum im Friedrichshain and
`Department of Internal Medicine – Nephrology, Berlin, Germany.
`2 Endocrinology and Metabolism, Xavier-Bichat Faculty of Medicine,
`University of Paris VII, Diabetes, Endocrinology and Nutrition
`Department, Hôpital Bichat and French National Institute for Health and
`Medical Research Unit U695, Paris, France.
`Correspondence to: Prof Dr med Martin K Kuhlmann
`Vivantes Klinikum im Friedrichshain, Landsberger Allee 49, D-10249
`Berlin, Germany.
`Tel: +49 30 13023 1004; Fax: +49 30 13023 2046
`E-mail: martin.kuhlmann@vivantes.de
`
`Abbreviations and acronyms
`
`AUC
`CHMP
`EMEA
`GMP
`HbA1C
`IGF
`PD
`PK
`PRCA
`
`area under the curve
`Committee for Medicinal Products for Human Use
`European Medicines Agency
`Good Manufacturing Practice
`glycated haemoglobin A1C
`insulin-like growth factor
`pharmacodynamic
`pharmacokinetic
`pure red cell aplasia
`
`Generic: A small-molecule drug whose active substance is shown
`by appropriate testing to have identical physicochemical, preclini-
`cal and clinical properties to an originator small-molecule drug.
`
`Introduction
`Biosimilar proteins are also known as follow-on biologics. These
`therapeutic proteins are expected to join originator biopharma-
`ceuticals in the marketplace when patent protection for the
`latter expires. Biosimilars already available in the European
`Union include epoetin alpha, growth hormone, interferon-β
`and factor VIII. Insulin analogues continue to have patent pro-
`tection until 2013 and beyond, but recombinant human insulin
`biosimilars will be available shortly. Clinicians may therefore
`expect that many of their patients who have renal failure asso-
`ciated with diabetes will be receiving both insulin and epoetin
`biosimilars within a few years. This article will use two case
`studies – the Marvel insulin dossier and PRCA associated with
`certain formulations of epoetin alpha – to illustrate some
`essential clinical issues raised by the advent of biosimilars.
`Biosimilar products should not be regarded as merely ‘bio-
`generic drugs’ analogous to traditional generic drugs, which
`are relatively small organic molecules. Biopharmaceuticals are
`generally proteins whose molecular weights are much higher,
`and whose three-dimensional structures are more complex,
`than those of traditional generics; hence, they are often cor-
`respondingly more fragile physically and chemically, and their
`formulation and storage requirements are more stringent than
`those of traditional generics. Moreover, biopharmaceuticals are
`manufactured using processes that are orders of magnitude
`more complex than processes used to manufacture traditional
`generics. Each of the steps in the process has some inherent
`variability, especially because living organisms are involved and
`the details of the processes developed by different companies
`are proprietary.1-3
`
`© The Author(s), 2010. Reprints and permissions: http://www.sagepub.co.uk/journalsPermissions.nav
`90
`
`10.1177/1474651409355454
`90
`VOLUME 10 ISSUE 2 . MARCH/APRIL 2010
`
`Exhibit 2071
`Page 03 of 10
`
`
`
`Figure 1. Post-fermentation steps in the manufacture of insulin glargine
`
`ACHIEVING BEST PRACTICE
`
`Reproduced with permission from sanofi-aventis groupe.
`
`The most important clinical consequence of this variability
`in manufacture is that an apparently small and innocuous
`change in any of the steps can have unforeseen consequences.
`Even if currently available tests show that a biosimilar is physi-
`cochemically equivalent to its counterpart originator molecule,
`the two may not behave in an equivalent manner in a clinical
`setting.4 In the case of the Marvel insulin formulations, prod-
`ucts with identical amino acid sequences (primary protein
`structures) had PK and PD properties that deviated significantly
`from those of the chosen comparator product. Even small
`changes in the manufacturing process of an originator biomol-
`ecule can cause difficulties: for example, a change in one of the
`formulations of the epoetin alpha Eprex was responsible for the
`emergence of PRCA up to a year after exposure to the prod-
`uct.3 Moreover, unpredictable changes in PK or PD properties
`can arise from process-related impurities (derived from the
`manufacturing process, e.g. host cell proteins, media compo-
`nents), product-related impurities (e.g. precursors, degradation
`products) or contaminants (adventitious materials that were
`not intended to be part of the manufacturing process).5
`
`The EMEA and its CHMP lead the world in developing regu-
`latory policies for the marketing of biosimilars.6 CHMP has
`produced guideline documents on the general requirements for
`non-clinical, clinical and quality aspects of biosimilars,7-9 as well
`as several guidelines pertaining to specific products, including
`soluble recombinant human insulin10 and epoetin alpha prod-
`ucts.11 For example, manufacturing processes for biosimilars,
`like all human medicinal products, must be in compliance with
`GMP that addresses areas ranging from buildings and facilities
`to stability issues; in addition, specific guidance (which addresses
`areas such as removal of product and process-related impurities
`and contaminants, and appropriate acceptance criteria for
`these entities) has been laid out for products that are manufac-
`tured by cell culture/fermentation.12,13
`
`Biosimilar insulin: EMEA requirements
`The manufacture of recombinant human insulin (figure 1) is
`highly complex.14 First, the human insulin gene is isolated and
`attached to a vector, which is then inserted into a host cell
`(usually E. coli or a yeast species). Next the recombinant cells
`
`THE BRITISH JOURNAL OF DIABETES AND VASCULAR DISEASE
`
`91
`
`Exhibit 2071
`Page 04 of 10
`
`
`
`ACHIEVING BEST PRACTICE
`
`are screened, a master cell bank is established and the resulting
`cell lines are cultured and fermented. The recombinant insulin
`fusion protein thus produced must then be released from the
`cells, isolated, purified, folded to produce the required second-
`ary structure and enzymatically cleaved to yield the biologically
`active insulin. This product is further purified and concentrated
`in several adsorption and chromatographic steps. Eventually, it
`is crystallised, lyophilised and formulated by adding com-
`pounds to prevent protein aggregation and bacterial growth or
`to change the in vivo absorption characteristics of the insulin
`(e.g. adding protamine results in a longer-acting formulation).15
`Any variation in this long series of steps, ranging from the vec-
`tor chosen to transfect the host cells to the excipients added
`during formulation and stabilisation, has the potential to result
`in an insulin product whose amino acid sequence and structure
`may be identical to that of native insulin and of originator
`recombinant insulin, but whose clinical characteristics, e.g. PK
`or PD behaviour, differ subtly from those of the originator
`product.4,15
`The EMEA has developed specific prerequisites for market-
`ing authorisations for soluble insulin biosimilars. As for all
`biosimilars, both the drug substance and the drug product
`must be assessed with appropriate qualitative and quantitative
`analytical procedures for impurities and their impurity profiles
`compared with those of an appropriate reference product.8
`Required preclinical studies include in vitro PD studies, in vitro
`affinity bioassays, assays for insulin and IGF-1 receptor binding.
`Required comparative clinical studies include at least one sin-
`gle-dose PK crossover study using subcutaneous administra-
`tion, preferably in patients with type 1 diabetes, and a PD
`double-blind, crossover, hyperinsulinaemic, euglycaemic clamp
`study that demonstrates a time-effect hypoglycaemic response
`profile. Clinical efficacy trials are not required if the clinical PK
`and PD studies show that the biosimilar is comparable to the
`reference product. A further requirement is a clinical trial, of at
`least 12 months duration, to evaluate the immunogenicity of
`the product. Such a trial needs to include a comparative phase
`of at least 6 months. Finally, a pharmacovigilance programme
`must be proposed that allows the early detection of any clini-
`cally significant immunogenicity that may develop over the
`long term.10
`
`Case 1. Marvel Lifesciences Ltd.
`The case of the Marvel LifeSciences recombinant insulin dossier
`illustrates both the potential effects of variations in manufac-
`turing processes and the application of the EMEA requirements
`to a submission for approval of a biosimilar product. In March
`2007, the MJ Group (Mumbai, India) submitted an application
`for a marketing authorisation for recombinant human insulin in
`three different formulations: a soluble rapid-acting insulin
`(‘Rapid’); a long-acting isophane insulin product (‘Long’); and a
`30:70 mixture of these two products (‘Mix’). The CHMP raised
`numerous concerns about the adequacy of the submission.16-18
`Regarding the quality of all three products, the CHMP noted
`that the information submitted was sparse in many critical
`
`aspects, notably: development, fermentation and purification
`processes; validation procedures; assays for impurities and sta-
`bility of the compounds; and in-process controls such as physi-
`cal separation and cleaning between the different products.
`Moreover, it was unclear whether the comparators used in stud-
`ies involving these insulins were actually valid reference prod-
`ucts. CHMP also noted that the dose-delivery properties of
`different presentations (vials and cartridges) had not been ade-
`quately tested and validated. Furthermore, in the case of the
`Long product, the protamine used to form the isophane crystals
`was not adequately characterised, and neither the manufactur-
`ing process nor the crystallisation process was documented in
`sufficient detail.17 In the case of the Mix product, there were
`also no details of formulation studies demonstrating a stabilised
`30:70 mixture. The CHMP concluded that none of the three
`products had sufficiently demonstrated its biosimilarity to a
`properly chosen reference product.18
`The CHMP also expressed concerns about the adequacy of
`the clinical data submitted in this dossier. For each of the for-
`mulations, the company carried out a single-dose, randomised,
`crossover, PD euglycaemic clamp study comparing the Marvel
`product with a reference originator insulin in 24 healthy male
`volunteers; however, these studies were not blinded and
`endogenous insulin secretion was not suppressed. PK data
`were derived from the euglycaemic clamp PD studies, but no
`independent PK studies were done. In particular, the single-
`dose crossover comparative study using subcutaneous injection
`recommended by CHMP was not carried out.
`The total AUC for glucose infusion rate for the Marvel Rapid
`product could be considered bioequivalent to that for the refer-
`ence product (Humulin S), because it fell within the classical
`interval of 80–125%; however, the AUCs up to 2 h after dosing
`were significantly higher, and the elimination half-life and mean
`residence time were significantly shorter for the Marvel prod-
`uct. In other words, it had a faster absorption, more potent
`effect and faster elimination than the reference product (figure
`2).16-18 The CHMP calculated that the Marvel soluble insulin
`could potentially induce a 45% greater glucose-lowering effect
`than Humulin S within the first hour after dosing, an unaccept-
`able degree of difference with obvious clinical relevance.16
`For both the Long and the Mix products, the mean PK and
`PD curves were largely superimposable on those of Humulin I
`and Humulin M3, their comparator products, but the study
`times were not long enough to obtain adequate data on
`elimination half-lives and clearance. The glucose infusion AUC
`was 27% lower for the Long product than for Humulin I, but
`23% higher for the Mix product (which contains 70% Long)
`than for Humulin M3. The manufacturer attributed this appar-
`ent inconsistency to batch-to-batch variability.17,18
`In addition, one efficacy and safety clinical trial was con-
`ducted in 526 patients with type 1 or type 2 diabetes. This
`consisted of a 6-month, double-blind, comparative phase test-
`ing all three Marvel insulin formulations against their respective
`reference products, followed by an open-label, 6-month, exten-
`sion whose results were not part of the dossier. In this trial,
`
`92
`
`VOLUME 10 ISSUE 2 . MARCH/APRIL 2010
`
`Exhibit 2071
`Page 05 of 10
`
`
`
`ACHIEVING BEST PRACTICE
`
`Figure 2. Marvel LifeSciences insulins: pharmacokinetic (PK) and pharmacodynamic (PD) data obtained in 24 healthy male volunteers16-18
`
`Pharmacokinetics of soluble insulin
`
`Pharmacokinetics of isophane insulin
`
`Pharmacokinetics of biphasic insulin
`
`Pharmacodynamics of soluble insulin
`
`Pharmacodynamics of isophane insulin
`
`Pharmacodynamics of biphasic insulin
`
`PK results (top three graphs): serum exogenous insulin (μIU/mL) versus time for soluble insulin (‘Rapid’), isophane insulin (‘Long’) and biphasic insulin ‘Mix’;
`a 30:70 mixture of Rapid and Long (all in black) compared to the reference Humulin products, Humulin S, Humulin I and Humulin M3, respectively (all in red).
`PD results (bottom three graphs): smoothed glucose infusion (mg/min/kg) versus time for soluble insulin (‘Rapid’), isophane insulin (‘Long’) and biphasic
`insulin (‘Mix’) (all in black) compared to the reference Humulin products, Humulin S, Humulin I and Humulin M3, respectively (all in red).
`Key: RHI = recombinant human insulin.
`
`THE BRITISH JOURNAL OF DIABETES AND VASCULAR DISEASE
`
`93
`
`Exhibit 2071
`Page 06 of 10
`
`
`
`ACHIEVING BEST PRACTICE
`
`Table 1. Effects on glycated haemoglobin A1C (HbA1C) of Marvel
`LifeSciences insulins and Humulin insulins after 24 weeks
`(mean HbA1C (%) after analysis of covariance adjustment for
`screening value)16-18
`
`Test Humulin Difference (95% CI)
`
`
`Type 1
`diabetes –
` fixed regimen
` fixed regimen
`Type 2
`diabetes –
` fixed regimen
`7.73 7.52
` flexible regimen 7.33 7.68
`
`8.43 8.16
`8.53 8.30
`
` 0.28 (-0.54–1.09)
` 0.22 (-0.15–0.60)
`
` 0.21 (-0.04–0.47)
`-0.35 (-0.85–0.15)
`
`Key: 95% CI = 95% confidence interval.
`
`patients received either soluble and isophane insulin in flexible
`doses or the fixed-dose combination. The primary efficacy end-
`point of the study was the HbA1C level at 24 weeks. Safety
`endpoints included treatment-emergent adverse events and the
`development of IgG anti-insulin antibodies.16
`There were significantly more withdrawals from the study
`in the Marvel groups than in the comparator groups (12% vs.
`7%). Dosing data were not analysed in enough detail to con-
`clude that patients in the test and comparator groups had
`actually received comparable doses of insulin. While HbA1C
`levels at 24 weeks did not differ statistically between the two
`groups, the trend was in favour of the Humulin comparator for
`all but one of the comparisons (table 1). The Marvel insulins
`performed similarly to the Humulin insulins for the first 12
`weeks, but worse during the second 12 weeks of the trial.
`Finally, the CHMP held that these clinical efficacy data could
`not substitute for euglycaemic clamp studies to demonstrate
`biosimilarity; the latter type of study is much more sensitive to
`differences between products than the former and the Rapid
`and Long formulations were administered together in many of
`the patients, thus invalidating PK comparisons for the individ-
`ual formulations.16
`Regarding safety outcomes in this trial, the Marvel products
`and the Humulin comparators were associated with similar
`rates of adverse events in patients with type 2 diabetes (25%
`vs. 31%, respectively) and with similar rates of new antibody
`formation in the first 24 weeks (10.7% vs. 12.5%, respec-
`tively), but those with type 1 diabetes had substantially higher
`rates of adverse events (24% vs. 12%) and of new antibody
`formation (21.9% vs. 14.0%) with the Marvel products. CHMP
`also concluded that immunogenicity was not fully evaluated;
`for example, the assay and its validation were not described,
`treatment-naïve patients were excluded, only new antibodies
`were considered and the impact of antibodies on safety and
`
`efficacy was not analysed. Finally, the pharmacovigilance sys-
`tem and the risk management plan submitted in this dossier
`were not considered to fulfil the requirements of EMEA guid-
`ance documents.16
`In January 2008, the EMEA announced that Marvel
`LifeSciences had withdrawn its applications for all three of its
`insulin formulations.19
`
`Epoetins, immunogenicity and PRCA
`Most biopharmaceuticals eventually induce an immune
`response. Often this response is without significant clinical
`consequences; for example, many patients receiving recombi-
`nant human insulin develop antibodies without any clinical
`sequelae. Sometimes, however, the immune response may
`decrease the efficacy of the product, or more rarely, produce a
`range of systemic immune effects, including local allergic reac-
`tions, serum sickness, anaphylaxis and other severe or fatal
`reactions such as the PRCA encountered with certain formula-
`tions of epoetin alpha.
`Human erythropoietin, first isolated and characterised in
`the mid-1980s,20,21 has been used to treat anaemia associated
`with a variety of conditions such as cancer, HIV and renal fail-
`ure.22-24 The first recombinant human erythropoietin (epoetin
`alpha) was developed by Johnson & Johnson. The stability and
`clearance of this glycoprotein depend critically on its degree of
`glycosylation25 and thus, unlike insulin, it must be manufac-
`tured in eukaryotic host cells such as Chinese hamster ovary or
`baby hamster kidney cells. (Its production in the yeast P. pasto-
`ris has also been investigated.26,27) Different commercially pro-
`duced follow-on epoetin alpha products share an identical
`amino acid sequence, but differ greatly in their glycosylation
`profiles, and hence in their biological activity.28 These differ-
`ences can be dramatic: one study using isoelectric focusing to
`compare 11 epoetin alpha products from manufacturers in
`Korea, Argentina, China and India found that their molecular
`weights varied substantially (figure 3).29 Additional studies have
`reported that epoetins from different sources vary widely in
`biological activity from 70% to over 200% of their own stated
`specifications.28,30 PRCA is a potentially fatal condition in which
`anti-epoetin antibodies neutralise both the exogenous product
`and endogenous erythropoietin, completely abolishing erythro-
`poiesis. Red cell precursors in bone marrow are absent, and the
`resulting anaemia is severe and intractable, while platelet and
`granulocyte counts remain normal.31,32
`
`Case 2. Eprex
`Before 1998, only three cases of PRCA had been published, but its
`incidence in patients with chronic kidney disease increased sharply
`between 1998 and 2002 to approximately 250 documented
`cases. Fatality rates were high, and many patients became
`completely transfusion dependent. Most cases involved subcu-
`taneous administration, but in some countries PRCA was
`absent despite the popularity of this route. A large proportion
`of the cases involved the original Eprex brand produced outside
`
`94
`
`VOLUME 10 ISSUE 2 . MARCH/APRIL 2010
`
`Exhibit 2071
`Page 07 of 10
`
`
`
`Figure 3. Epoetin alfa products from different manufacturers: variations
`in molecular weight and degree of glycosylation. Reproduced
`with permission from Oxford University Press 29
`
`Figure 4. Incidence over time of PRCA cases associated with Eprex and
`other epoetin alpha products37
`
`ACHIEVING BEST PRACTICE
`
`Key: E = epoetin alpha control; IA = 2,000 IU/mL (sample from Korea);
`IB = 4,000 IU/mL (Korea); IIA = 2,000 IU/mL (Korea); IIB = 10,000 IU/
`mL (Korea); IIIA = 2,000 IU/mL (Korea); IIIB = 10,000 IU/mL (Korea);
`IV = 2,000 IU/mL (Argentina); V = 10,000 IU/mL (Argentina); VI = 4,000
`IU/mL (India); VII = 10,000 IU/mL (China); VIII = 6,000 IU/mL (China).
`
`the USA. It is notable that in 1998, in response to concerns
`about transmitting bovine spongiform encephalopathy (‘mad
`cow disease’), the formulation of Eprex had been changed in
`Europe and human serum albumin had been replaced by poly-
`sorbate 80 and glycine.
`The precise mechanism by which the manufacturing
`change caused Eprex to become immunogenic remains a mat-
`ter of debate.33 Variation in the structure of the epoetin mole-
`cule itself was ruled out as a major factor. It has been suggested
`that polysorbate 80, a surfactant that tends to form micelles,
`might have promoted the formation of epoetin-containing
`micelles that resembled foreign pathogens, thereby unleashing
`the breakdown of immune self-tolerance.34 An alternative
`explanation is that polysorbate 80 might have reacted with the
`uncoated rubber stoppers of the prefilled syringes to produce
`leachates that could act as adjuvants, provoking an antibody
`response.35,36 Otherwise, the new formulation might have been
`less stable and therefore more vulnerable to denaturation or
`aggregation under inadequate storage or improper handling
`conditions. In 2003, Johnson & Johnson replaced uncoated
`rubber stoppers with Teflon-coated stoppers in the prefilled
`syringes, enhanced control of the product’s cold chain and
`issued detailed guidelines and instructions that emphasised the
`importance of storage at between 2 and 8 ºC. In addition,
`intravenous administration was recommended over subcutane-
`ous administration. After these changes were effected, the
`
`Key: HSA = human serum albumin; PRCA = pure red cell aplasia.
`
`incidence of PRCA subsided dramatically to a level similar to
`that seen prior to 1998 (figure 4).33,37-39
`Immunogenicity is the most important safety issue associ-
`ated with biosimilars. As this story amply illustrates, immuno-
`genicity in vivo cannot be predicted, can result from minor
`changes in manufacturing, and sometimes emerges only after
`long-term exposure to the product. It is essential, therefore,
`that a long-term pharmacovigilance programme be instituted
`for any biopharmaceutical made available for clinical use.
`
`Summary and conclusions
`To obtain marketing authorisation in Europe, makers of biosim-
`ilars must submit detailed descriptions of the manufacturing
`process and its controls (including processes for reduction of
`impurities and contaminants), preclinical and clinical studies
`and a risk management programme that includes measures to
`detect late-emerging immunogenicity. The two case studies
`above illustrate why the EMEA prerequisites for biosimilars
`must be more stringent than those for traditional generics.40,41
`The PRCA experience strikingly illustrates how an apparently
`relatively minor change in the originator manufacturer’s own
`process had serious and unpredictable clinical consequences
`whose exact causes have been difficult to unravel, because
`they emerged after relatively long exposure to the product. The
`physicochemical properties of the Marvel insulins were similar
`to those of the reference products, but PK, PD, clinical efficacy
`and immunogenicity data revealed unexpected differences
`with potentially serious clinical implications. Over the past
`decade, the availability of a wide range of insulin formulations
`and the introduction of insulin analogues have greatly improved
`the management of diabetes, allowing people with diabetes to
`adjust their insulin regimens to their lifestyles, rather than vice
`
`THE BRITISH JOURNAL OF DIABETES AND VASCULAR DISEASE
`
`95
`
`Exhibit 2071
`Page 08 of 10
`
`
`
`ACHIEVING BEST PRACTICE
`
`Key messages
`
`
`● Biosimilars should not be regarded as simply analogous
`to generic versions of traditional pharmaceuticals
`● The Marvel insulin formulations were physicochemically
`similar, but had PK/PD and clinical properties that
`differed substantially from comparator insulins
`● An apparently minor change in the formulation of
`Eprex was responsible for an upsurge of PRCA
`● The Marvel and Eprex cases illustrate that changes in
`manufacturing processes can have important clinical
`consequences
`
`versa. The benefits of these existing insulin products should be
`borne in mind when evaluating the new insulin biosimilars.
`
`Acknowledgements
`Editorial support for this article was provided by the medical
`writing agency PHOCUS and by sanofi-aventis groupe. The
`opinions expressed in the current article are those of the
`authors. The authors received no honoraria or other form of
`financial support related to the development of this manuscript.
`MK reports consultancy for Roche and sanofi-aventis
`groupe; is a clinical investigator for MedaSorb; participates in a
`speakers’ bureau for Roche; and receives speaker’s honoraria
`from Amgen, Fresenius, Genzyme, Roche, sanofi-aventis
`groupe and Shire. MM reports grants/research support from
`Novo Nordisk, sanofi-aventis groupe and Servier; is a consultant
`for Lilly, Merck, Novo Nordisk, sanofi-aventis groupe and
`Servier; is a clinical investigator and participates in a speakers’
`bureau for Merck, Novo Nordisk, sanofi-aventis groupe and
`Servier.
`
`References
` 1. Roger SD. Biosimilars: how similar or dissimilar are they? Nephrology
`2006;11:341-6.
` 2. Schellekens H, Ryff JC. ‘Biogenerics’: the off-patent biotech products.
`Trends Pharmacol Sci 2002;23:119-21.
` 3. Locatelli F, Roger S. Comparative testing and pharmacovigilance of
`biosimilars. Nephrol Dial Transplant 2006;21(suppl 5):v13-v6.
` 4. Kuhlmann M, Covic A. The protein science of biosimilars. Nephrol Dial
`Transplant 2006;21(suppl 5):v4–8.
` 5. European Medicines Agency. International Conference on Harmonization
`Topic Q6B: Specifications: test procedures and acceptance criteria for
`biotechnological/biological products. Document # CPMP/ICH/365/96.
`September 1999. London: EMEA, 1999. www.emea.europa.eu/pdfs/
`human/ich/036596en.pdf. (Accessed 2 September 2009)
` 6. Wiecek A, Mikhail A. European regulatory guidelines for biosimilars.
`Nephrol Dial Transplant 2006;21(suppl 5):v17-20.
` 7. Committee for Medicinal Products for Human Use. Guideline on
`similar biological medicinal products. London: EMEA, 2005. www.
`emea.europa.eu/pdfs/human/biosimilar/043704en.pdf.
`(Accessed
`13 April 2009)
`
` 8. Committee for Medicinal Products for Human Use. Guidelines on
`similar biological medicinal products containing biotechnology-derived
`proteins as active substance: quality issues. London: EMEA, 2006.
`www.emea.europa.eu/pdfs/human/biosimilar/4934805en.pdf.
`(Accessed 13 April 2009)
` 9. Committee for Medicinal Products for Human Use. Guideline on similar
`biological medicinal products containing biotechnology-derived pro-
`teins as active substance: non-clinical and clinical issues. London: EMEA,
`2006. www.emea.europa.eu/pdfs/human/biosimilar/4283205en.pdf.
`(Accessed 13 April 2009)
`10. Committee for Medicinal Products for Human Use. Annex guideline on
`similar biological medicinal products containing biotechnology-derived
`proteins as active substance: non-clinical and clinical issues. Guidance on
`similar medicinal products containing recombinant human insulin. London:
`EMEA, 2006. www.emea.europa.eu/pdfs/human/biosimilar/3277505en.
`pdf. (Accessed 13 April 2009)
`11. European Medicines Agency. Annex to guideline on similar biological
`medicinal products containing biotechnology-derived proteins as active
`substance: non-clinical and clinical issues. Guidance on similar medicinal
`products containing recombinant erythropoietins. London, 22 March
`2006. Document ref EMEA/CHMP/BMWP/94526/2005 Corr. London:
`EMEA, 2006. www.emea.europa.eu/pdfs/human/biosimilar/9452605en.
`pdf. (Accessed 1 May 2009)
`International Conference on
`12. European Medicines Agency.
`Harmonization Topic Q7: Good manufacturing practice for active phar-
`maceutical ingredients. Document #CPMP/ICH/4106/00. November
`20, 2000. London: EMEA, 2000. www.emea.europa.eu/pdfs/human/
`ich/410600en.pdf. (Accessed 2 September 2009)
`13. European Medicines Agency. International Conference on Harmonization
`Topic Q5A (R1): Quality of biotechnological products: viral safety evalu-
`ation of biotechnology products derived from cell lines of human or
`animal origin. Document #CPMP/ICH/295/95. October 1997. London:
`EMEA, 1997. www.emea.europa.eu/pdfs/human/ich/029595en.pdf.
`(Accessed 2 September 2009)
`14. Walsh G. Therapeutic insulins and their large-scale manufacture. Appl
`Microbiol Biotechnol 2005;67:151-9.
`15. Joshi SR, Parikh RM, Das AK. Insulin: history, biochemistry, physiology
`and pharmacology. J As