Article Text

Download PDFPDF

Original article
How to select a biosimilar
  1. Niels Boone1,
  2. Hugo van der Kuy1,
  3. Mike Scott2,
  4. Jill Mairs2,
  5. Irene Krämer3,
  6. Arnold Vulto4,
  7. Rob Janknegt1
  1. 1Department of Clinical Pharmacy and Toxicology, Orbis Medical Center, Sittard-Geleen, The Netherlands
  2. 2Department of Clinical Pharmacy, Antrim Area Hospital, Antrim, UK
  3. 3University Medical Center, Johannes Gutenberg University, Mainz, Germany
  4. 4Department of Clinical Pharmacy and Toxicology, Erasmus Medical Center, Rotterdam, The Netherlands
  1. Correspondence to Niels Boone, Orbis Medical Center, Department of Clinical Pharmacy and Toxicology, PO Box 5500, Sittard-Geleen NL 6130 MB, The Netherlands; nwboone{at}


In the past few years biosimilars have penetrated the market following the expiry of patents of originator variants. This offers the opportunity to apply high-tech protein products at a lower cost. In contrast to small-molecule generics, clinicians and pharmacists have found it difficult to judge the efficacy and safety profiles of complex protein products. In recent years, the European Medicines Agency (EMA) has gained knowledge on assessing comparability between biosimilars and originator products in scientific and legal areas. This article provides an overview of an extensive set of 31 previously drawn biosimilar selection criteria and describes how several of these criteria are covered by EMA regulations and guidelines. A panel of experts (authors) reviewed the criteria and produced a shortlist of 10 criteria relevant for clinicians and pharmacists.

  • Clinical Pharmacy
  • Pharmacotherapy

Statistics from


Selection of biosimilars in hospitals is a relatively new issue. Biosimilars have also been referred to as follow-on protein products, follow-on biologicals, bio-therapeutic similars, similar biological medicinal products, similar bio-therapeutic products, etc. Patents of several expensive biopharmaceutical products have expired or are due to expire shortly. This is important due to the continuously increasing costs in healthcare and the numerous oncology drugs under development. The introduction of biosimilars was expected to reduce the costs of these medicines due to the enhancement of competition in biological medicine markets. Although costs have reduced, they have not decreased to the same extent as occurred in the generic small-molecule market.1 Costs of biosimilars are about 15–30% less than the originator prices in the European Union (EU) and the USA.2 Depending on the type of reimbursement system, cost reductions of medicines could improve patient access to biotechnological products in some European countries (eg, The Netherlands).

In contrast to generic small-molecule medicines, biosimilars are not identical to their reference product. New biosimilars have to be evaluated for inclusion into a hospital formulary. In this evaluation of biosimilars, it is necessary to have a complete and rational set of criteria to determine clinically and non-clinically relevant differences between biosimilars and to compare these to the reference product.3 ,4 See box 1 for an explanation of terminology.

Box 1

Biosimilar definitions

European Medicines Agency

A biosimilar medicine is a medicine, which is similar to a biological medicine that has already been authorised (the ‘biological reference medicine’). The active substance of a biosimilar medicine is similar to the one of the biological reference medicine. Biosimilar and biological reference medicines are used in general at the same dose to treat the same disease.

A biological medicine is a medicine whose active substance is made by or derived from a living organism.

(London, 22 October 2008, Doc Ref EMEA/74562/2006 Rev 1).

Food and Drug Administration

A biosimilar is a biological product that is highly similar to a US-licensed reference biological product notwithstanding minor differences in clinically inactive components, and for which there are no clinically meaningful differences between the biological product and the reference product in terms of the safety, purity, and potency of the product.

Biological products can include a wide range of products including vaccines, blood and blood components, allergenics, somatic cells, gene therapy, tissues, and proteins. Unlike most traditional, small-molecule prescription drugs that are made through chemical processes, biological products are generally made from human and/or animal materials. Biological products are usually larger than and have a more complex structure than small-molecule prescription drugs. Such products may be manufactured through biotechnology, derived from natural sources, or, in some cases, produced synthetically ( 11 March 2011).

A different generic approach

Non-protein drugs are typically organic molecules of low molecular mass and well defined molecular structure. Because the molecular structure of such a small-molecule drug can be fully analytically characterised, it is fairly easy for a generic drug manufacturer to produce a bio-equivalent medicinal product with the same drug usage form containing the same active ingredient as the innovator's drug product.

A protein product is a heterogeneous mixture of large molecules based on a sequence of amino acids folded in secondary and tertiary three-dimensional structures, which undergo post-translational folding processes to ultimately fold into a complex spatial structure. Post-translational modification is a function of host cells, which are not identical for the biosimilar and the originator medicinal product. This complex process is difficult to reproduce even in the production process of the originator drug. A full chemical characterisation of the product resulting from this process is a challenge using multiple analytical tools. However, it is not easy to decide which battery of chemical tests should be performed to counter certain clinical risks.5

In the hypothetical situation of using the same recombinant genetic construct, type of host cell, culture conditions and downstream processing like purification techniques, formulation and packaging there is no guarantee of bioequivalence or ability to substitute for the reference products. Companies use a unique master cell bank, which remains under patent, and proprietary techniques that may affect the efficacy and safety of a product.6 However, the concept of developing a biosimilar medicinal product is to reverse engineer the reference product whereby companies use their own state-of-the art process and production facilities to achieve the same quality profile as the originator medicinal product (J Windisch, personal communication).

An important question is how the biosimilar product variation compares with the originator product batch variation. Different immunogenicity, differences in pharmacokinetic behaviour and distribution can be clinical consequences caused by variation in the tertiary structures of the protein drug products, for example caused by differences in post-translational modification.7

There are a number of issues that will make selection of biosimilars much more complicated and specific than the selection of generic drugs. Given the complexity of producing and regulating biological medicinal products, a checklist must be established, including all aspects that could affect the quality, efficacy and safety of biosimilars.

Most clinicians and pharmacists are uncertain about the evaluation of a biosimilar medicinal product due to the following factors:

  • The complex nature of a biological medicinal product and its production process. The whole concept of identity is not applicable to biological medicines, either to the different batches of the originator medicinal products or to the original products before and after changes in the manufacturing process.

  • Comparability trials could be conducted with different efficacy endpoints than those used in the initial trials with the originator product.8 The use of different endpoints makes it complicated for prescribing doctors or pharmacists who lack experience with the biosimilarity of medicines. Clinicians and pharmacists must realise that the aim of a biosimilar development programme and clinical trials is not to establish benefit; the aim is to prove biosimilarity.

  • Similarity and safety are studied in a specific patient population which makes it difficult to extrapolate the risks and benefits to individual patients in clinical practice. Due to the nature of biosimilar comparability trials a limited patient population is often studied in terms of indication, comorbidities, comedication, age and gender.

  • Important clinical information regarding long-term effects must be collected during the post-marketing phase according to the risk management programmes, which have to be implemented by the manufacturer to collect information about the efficacy and safety of approved biosimilar medicinal products that do not always comply with the authorised indication. In some cases the industry does not fully adhere to these programmes.9

Companies that develop biosimilar medicinal products have a quadruple mission to fulfil in comparison to the manufacturers of the originator product. They have to bring a complex product to the market that must be similar in numerous aspects to its originator counterpart but with lower acquisition costs and must meet US Food and Drug Administration (FDA) and European Medicines Agency (EMA) biosimilar regulatory standards that become higher and more clearly defined as they move away from the generic paradigm. During the latter half of 2012, Samsung and Teva-Lonza who were developing a rituximab biosimilar (the patent is due to expire in 2013) struggled with their research programmes due to regulatory issues and cost aspects.10

So far, the market uptake of biosimilar products has been relatively low in the EU. Market penetration is the lowest (below 5%; biosimilars vs total of biosimilars and reference products) in The Netherlands, Belgium, Norway and Austria for biosimilars of the above-mentioned reference products.11 However, the combined market share of biosimilar versions of the reference products Eprex, Genotropin and Neupogen have grown in recent years from 0.33% in 2007 to 15.52% in 2010.11 In fact, biosimilars may lead to lower prices for the originator products, without obtaining a relevant market share. This is advantageous from a societal aspect, but does not encourage other companies to invest in the development of biosimilars.

Substitution and therapeutic interchange

The differences between small-molecule medicines and biosimilar products have prompted questions relating to the substitution of biosimilars. It is important to note that selection criteria can be different for both types of medicinal products.

Substitution means that a pharmacist can substitute the originator product for a biosimilar without information from the prescribing physician. In this case both products are designated as interchangeable, which means that they can be alternated or switched. The Biologics Price Competition and Innovation Act, Section 351k, part of the American Affordable Care Act (2010), created an abbreviated FDA legislation pathway for similar biological products to become interchangeable with an FDA-licensed reference product. A designation of ‘interchangeable’ by the FDA is a stepwise process after proving biosimilarity but will be hardly impossible in a scientific way. Currently the FDA is considering which clinical and non-clinical information is necessary to delineate a biological as interchangeable.12 ,13 In the case of biosimilar medicinal products it is worth mentioning that the granting of approval does not mean that the biosimilar product can be automatically substituted for the reference product and vice versa. EMA does not make recommendations about interchangeability between a biosimilar medicinal product and its reference product and leaves this question to be answered based on to the judgement of the patient’s clinician and the pharmacist.14 Most clinicians and hospital pharmacists lack the knowledge to make a sound clinical judgement of all aspects of biosimilars. Several EU countries, such as France, Spain, Italy, Germany, The Netherlands, the UK and Sweden, have established legislative measures that forbid the automatic substitution of these products.15

However, the Dutch Ministry of Health has allowed health insurers to include generics in their preference policies. In this Dutch example insurers can choose to reimburse only a certain label of a generic medicine based on its costs. The physician can bypass this preference policy by specifying on the prescription that the patient has to use the originator product for a medical reason. In the same way, insurers in The Netherlands can now select a biosimilar for preferred reimbursement in the case of patients who are biological naïve. Some authors state that only a doctor can decide whether or not to switch a patient's treatment to a biosimilar product.16 But in daily clinical practice the individual physician will not have access to all the information that is mandatory to make a sound decision. According to the data available, the selection process of a biosimilar as a preferred product by healthcare insurers seems to be based only on cost. A full set of selection criteria used by insurance companies is not available.

To date, 13 biosimilars have been approved in Europe, covering only three reference products (Filgrastim, Epoetin, Genotropin). See table 1 for a list of these biosimilars. Some of these products are identical to each other because they are based on the same dossier and are comarketed. It is important to realise that the number of biosimilars and regulatory requirements will increase rapidly due to the expiration of patents, therefore constant review and assessment of the checklist is necessary.

Table 1

Biosimilars authorised by EMA

Due to the complexity of the process, it is not realistic for individual hospital pharmacists and clinicians to have enough time and knowledge to make a sound judgement of the properties of individual biosimilars. Therefore, it would be useful to determine a standard set of criteria for the judgement of biosimilars. Based on an extensive literature search, a panel of experts would establish the properties of the various biosimilars (and reference product). Individual hospital pharmacists (and clinicians) may assign a relative weight to each selection criterion, thereby making a personal but transparent and rational drug selection possible. The methodology is similar to that of the System of Objectified Judgement Analysis (SOJA).17 To ensure an adequate selection of biosimilars the authors used an evaluation checklist based on an established checklist proposed by Irene Krämer4 and the SOJA and InforMatrix selection criteria.17 ,18 The full list of selection criteria is shown in table 2.

Table 2

Full set of potential selection criteria

STEPSelect, which is a comprehensive quality, safety and efficacy based system, was proposed for the establishment of pre-qualification criteria.19 Then, all selection criteria were critically judged by an extensive panel of experts to distinguish those criteria that are adequately covered by EMA. Due to changes in the EMA regulations and more experience since the publication of Irene Krämer`s checklist, some of the questions are superfluous today. The remaining selection criteria are listed in table 3.

Table 3

Final set of selection criteria for evaluation of biosimilars


Pre-qualification criteria

Prior to the establishment of biosimilar selection criteria the following aspects relating to pharmaceutical licensing, reliability of supply, facilitation of a 24 h contact service with a medical division of the company to report adverse events, financial and insurance criteria, as well as public and product liability play a role. These aspects are supplemented with product-specific medication safety criteria in STEPSelect and together form important pre-qualification criteria to be used as a prerequisite.19

A: Production process/manufacturer

Risk management and quality assurance are integral parts of an effective pharmaceutical quality assurance system that can provide a proactive approach to identify, scientifically evaluate and control potential risks in quality. Such a system facilitates continuous improvement in process, performance and product quality throughout the product lifecycle20 and can also be applied to biotechnological products. The authors consider this the responsibility of EMA.

The International Conference on Harmonisation (ICH) of Technical Requirements for Registration of Pharmaceuticals for Human Use guidelines harmonise legislation between Europe, Japan and the USA. A medicine that complies with ICH guidelines can be marketed in Europe, Japan and the USA. The guidelines are based on good manufacturing practice (GMP) rules. Quality by design (QbD) is a part of the ICH guidelines and, although not mandatory, if a manufacturer complies with QbD this indicates the medicinal product is of high quality. QbD models can enhance scientific understanding and possibly predict the behaviour of a system under a set of conditions. These models indicate product quality and that the whole process is controlled, thus the risks associated with processes are decreased. Also, there is continuous process verification and improvement which minimises the risks associated with manufacturing failures and product recalls. QbD is explained in ICH Q8, 9, 10 and 11.21–24 In the case of biosimilars, EMA asks for compliance with the ICH guidelines.

Specifications for quality control are covered by the ICH Q6B ‘Guidance on test procedures and acceptance criteria for biotechnological/biological products’ guideline.25 A rationale has to be provided for the use of certain product-specific acceptance criteria based on data obtained from batches used during clinical and non-clinical tests supported by the applicant’s experience of the biosimilar product and testing of the reference product.

A1: Are all biosimilars approved in Europe produced under GMP conditions?

It is important to track down and check how EMA deals with the quality status of active pharmaceutical ingredients (APIs), excipients and finished products. This is relevant in when the manufacturing process is undertaken by a third party. Directive 2011/62/EU lays down considerations separately for APIs and finished medicinal products.26

Pharmaceutical companies in Europe are obliged to establish and implement GMP. When the marketing authorisation holder (MAH) asks for approval of a medicinal product, it must comply with GMP. When an applicant brings a new medicinal product to the market, EMA has to inspect the manufacturer for GMP and good clinical practice (GCP).27 ,28

This criterion is adequately covered by EMA and is not included in the final set of selection criteria.

A2: Is the exact production site of API and final product always known to EMA?

Since 2 January 2013 a new European requirement exists for importing APIs. All APIs must have been manufactured at a GMP site (GMP must be equivalent to EU’s GMP as a minimum). From 2 July 2013 onwards, according to article 46b2b of the EU Directive 2011/62/EU, the importer needs to request a document by the local competent authority in which it is written that the manufacturer complies with GMP standards to at least the same standard as EU GMP.26 The ICH Q7 document provides guidelines for manufacturing quality management of APIs and harmonises GMP standards in the EU, the USA and Japan.29

All sites involved in the production of the finished medicinal product and of the active substances must be described (name and detailed address, including building reference) in Module 1.2 of the application for a marketing authorisation together with a description of the steps performed. This should include, among other things, active substance and finished product manufacture.30

This criterion is adequately covered by EMA and is not included in the final set of selection criteria.

A3: Have all production facilities of API and final product been checked by EMA or other authorities?

The production site is subject to inspection by EMA inspectors or other authorities. A finished product can be formulated with APIs or excipients produced under Chinese and or Indian GMP (which is estimated to be 80% of bulk APIs). It is EMA’s role to review this process as a part of the registration procedure.

In relation to the document for third country manufacturers published by EMA, information about any previous EEA (European Economic Area) inspection in the last 2–3 years and/or any planned EEA inspections should be provided and should include details of the inspection dates, product category inspected and the name of the inspecting competent authority.30

Documents that should be attached to Module 1 of the application include the following.

For all sites (other than active substance manufacturers) located in third countries where a Mutual Recognition Agreement (MRA) or other relevant agreement is in place, a MRA certificate (not older than 3 years) from the local competent authority that carried out the inspection and/ or a GMP certificate from the EEA inspecting competent authority (if the site has been inspected by an EEA competent authority in the last 2–3 years) is required.

For all sites other than active substance manufacturers, located in third countries with no MRA, a GMP certificate from the EEA inspecting competent authority (if the site has been inspected by an EEA competent authority in the last 2–3 years) is necessary. Alternatively, a reference can be made to the appropriate entry in the EudraGMP database.

In addition to the above, a copy of the registration or other document analogous to the manufacturing authorisation from the local competent authority must demonstrate that the site is authorised to manufacture a product or pharmaceutical form. Also details of any inspection performed by authorities other than the EEA authorities (eg, a GMP certificate or similar statement from the competent authority which carried out the inspection) must be provided.30

According to EMA an operational MRA provides assurance that equivalent GMP standards are applied by the parties of the MRA and removes the need for additional inspection and re-controls at import. This is applicable for production plants located in the following countries: Australia, Canada, Switzerland, Japan, New Zealand, the USA.30

Once validated, adding a new site or changing the steps of manufacture/batch release described under Module 1.2 (ie, application form) of the application is normally not allowed during the 210-day review period. Any additional site or change in the manufacturing or batch release arrangements should be submitted as a variation after the granting of the marketing authorisation.30

In the case of Omnitrope, during its development, several changes were introduced, such as a transfer to a production plant in the USA and a re-transfer to a production plant in Austria. Several improvements to the purification process were implemented at these manufacturing sites. Due to these major changes, additional similarity information had to be submitted and reviewed by EMA.31

During the process development of Ratiograstim, the site of fermentation and purification was changed when the manufacturing was transferred to a GMP production facility. In a public assessment report for Ratiograstim, EMA rapporteurs explained that the master cell bank and working cell bank complied with current GMP.32

This criterion is adequately covered by EMA and is not included in the final set of selection criteria.

A4: Are all recalls reported to EMA?

MAHs, manufacturers and importers are obliged to report to EMA whenever a centrally authorised product is not available because of a shortage of supply or when a recall has to be done. In addition, MAHs are requested to inform EMA of any prohibition or restriction imposed by the responsible competent authority of any country in which the medicinal product is placed on the market and of any other new information.33

The following FDA webpage has a section that shows current biological product shortages, including the reasons for the shortage and if they have been resolved: On the current version of the EMA website it is not possible to run a search on the criteria ‘show me all medicines with shortages’. This requirement has been noted and this type of functionality is expected to become available in the future. Currently, when a search is performed for a medicine by product name on the website, the medicine page returned will show any information available on shortages on the right-hand side of the page.

EMA only communicates on shortages if they concern a centrally authorised medicine. This is the case if the shortage is affecting more than one EU Member State and once the EMA has issued a recommendation. EMA asks the National Competent Authorities (NCAs) of individual member states for a more complete picture on shortages of nationally authorised products and centrally authorised products in a particular member state. This criterion is adequately covered by EMA and NCAs and is not included in the final set of selection criteria.

A5: Is the manufacturer experienced in the production of biopharmaceuticals?

This criterion is included in the final set of criteria.

Experience of biopharmaceuticals production: score 10% per year of experience. Therefore, the score is 40% if the company has been in production for 4 years.

A6: Which national authorities are responsible for inspecting the manufacturing site?

This information is covered in Module 1 of the EMA application where administrative information is requested. On the following link information can be assessed about the national authorities inside the EU:

This criterion is adequately covered by EMA and is not included in the final set of selection criteria.

A7: How long has the biopharmaceutical been on the market?

This criterion is included in the final set of criteria. The criterion is introduced as an estimate of the guaranteed safety of the biosimilar.

This is a standard SOJA criterion: 10% for every year on the market. This is used for a comparison of different chemical entities of small molecules. Because biosimilars are similar to the reference product, a flatter scale is used for biosimilars: 25% per year. If a biosimilar is on the market for 3 years, this product will score 75%.

A8: How extensive is the clinical experience with each individual biosimilar? Expressed as the number of patient days worldwide

This criterion is included in the final set of criteria. The criterion is introduced as an estimate of the guaranteed safety of the biosimilar.

This is a standard SOJA criterion: 1% for every million days of experience. This is used for a comparison of different chemical entities. Because biosimilars are similar to the reference product, a flatter scale is used for biosimilars: 10% per million patient days of experience. If the clinical experience with a biosimilar is 4 million patient days, this product will score 40%.

A9: Does the manufacturer guarantee active information about major changes of the manufacturing process?

Quality control and changes in the production process are included to the list of criteria set up by Krämer et al.4 The erythropoietin case has made clear that changes in the production process may affect the properties of the final product in a negative way.

Erythropoietin α (EPO α; Eprex) was used extensively for the treatment of anaemia associated with chronic renal failure. An increase in the incidence of pure red cell aplasia associated with EPO α occurred in 1998. After that, investigations were initiated to identify the cause of the increase in the immunogenicity of Eprex. Rubber stoppers were in use since the introduction of Eprex pre-filled syringes for subcutaneous administration. The replacement of human serum albumin with polysorbate 80 in 1998 appears to have resulted in a change in the leaching of potentially immunogenic compounds from the uncoated rubber stoppers.34 In addition, it was discussed that the new formulation was less stable and an increased aggregation rate caused immunogenic reactions. This example shows that it is necessary to inform EMA about any changes in the manufacturing process because it might affect the efficacy and safety.

The ICH guidance (ICH-Q5E) concludes that products must have highly similar quality characteristics before and after changes of the manufacturing process and that safety and efficacy of the drug product is not affected.35 EMA published a document that underlines the importance of performing a proper comparability exercise in its guideline on comparability of medicinal products containing biotechnology-derived products as active substance non-clinical and clinical issues.8

The goal of the comparability exercise is to ensure the quality, safety and efficacy of a drug product after changing the manufacturing process. Therefore all relevant data must be registered and evaluated that might have any impact on the drug product characteristics. Comparability can be assessed on a combination of analytical testing, biological assays, and in some cases, non-clinical and clinical data. The manufacturer should confirm that the specifications after the process change are appropriate to ensure product quality. Of course, this is also true for the reference product. Schiestl et al36 investigated the variability in quality characteristics of different batches of Aranesp, Mabthera and Enbrel. Major structural changes in protein composition that can alter the pharmacokinetics and biological activity were found by chemical analysis of these products. The observed changes occurred because of changes in growth conditions or purification steps. All batches of the products remained on the market during the investigation period. The health authorities did not reason that the observed changes would affect the safety and efficacy profiles of these originator products. Therefore it can be concluded that the considerations to maintain a product on the market after production process changes are multifactorial and are not solely based on the chemical characteristics of the questioned protein products. The extent of the comparability exercise depends on what is known about its possible impact on safety and efficacy of the product. Sometimes additional comparative information is needed at the preclinical and clinical level. The objective of the ICH Q5E document is to provide principles for assessing the comparability of biotechnological or biological products before and after changes that are made in the manufacturing process for the drug substance or drug product.35

For hospital pharmacists and clinicians it would be useful to check whether there is a causal relation between an observed adverse drug event and a product change. The EMA webpage provides information about changes in the manufacturing process of specific biological medicinal products .37

This criterion is adequately covered by EMA and is not included in the final set of selection criteria.

A10: Is the reference medicinal product authorised in the European Community?

At the moment this is a prerequisite according to the biosimilar authorisation legislation. The reference medicinal product must be a medicinal product authorised in the European Community. However, changes are under discussion.

Because this is a valid prerequisite, this criterion was not included in the final set of selection criteria.

A11: How has it been validated that the production of the API and final product is always performed under the same conditions?

The EU has introduced new rules as an amendment of Directive 2011/62/EU for importing APIs to stop the import of substandard or counterfeit pharmaceutical ingredients.26 Competent authorities of the exporting country must confirm to the broker that the production site of an API complies with EU GMP rules and that the production plant is subject to control. Manufacturing standards for APIs are those of the ICH Q7.29

The manufacturing and purification processes have to be properly described and the process validation studies and the in-process control system must be considered acceptable to EMA. Product European public assessment reports (EPARs) explain that validation programmes were developed or implemented and have limits set. In these documents individual physicians and pharmacists cannot access further information, but there is a general view of the validation.

This criterion is adequately covered by EMA and is not included in the final set of selection criteria.

A12: Has EMA checked quality control systems of the producers of API and final product?

These data are described in the EPARs of the individual products. See Filgrastim Hexal and Nivestim documents as examples.38 ,39

This criterion is adequately covered by EMA and is not included in the final set of selection criteria.

A13: Are data concerning batch reproducibility (inter-batch and intra-batch variability) available at EMA before and after approval?

These data are described in the EPAR scientific discussions of the individual products. See Nivestim scientific discussion as an example.39

This criterion is adequately covered by EMA and is not included in the final set of selection criteria.

A14: Are specifications of quality control of the MAH/manufacturer known to EMA and assessed by EMA?

Analytical aspects of validation are covered by the EMA document ‘Guideline on bioanalytical method validation’.40

The manufacturer has to present quality data to EMA for a marketing authorisation application. The registration dossier includes data to establish the quality of the product, including

definitions and descriptions of the manufacturing process, and associated quality checks and standards;

data on consistency of manufacturing (quality control of the process); data on analytical tests (molecular structure, potency and purity profile); data on stability of the product. Most of these data are to be presented as a comprehensive comparability exercise with the reference product.41

It may be the case that these data are protected by EMA as confidential information. Even a company developing a biosimilar monoclonal antibody usually has no data on critical in-process controls and intermediates in the manufacturing process of the innovator.

This criterion is adequately covered by EMA and is not included in the final set of selection criteria.

A15: Is it taken into consideration whether materials of animal origin or potentially allergenic materials are used during the production or formulation?

If an application relates to a medicinal product which contains or uses in the manufacture materials of animal origin or these are used during the production process, applicants are requested to complete tables and information established on ‘Note for guidance on minimising the risk of transmitting animal spongiform encephalopathy agents via medicinal products’ and this information will be included in Module 3.2.R of the common technical document (CTD).42

For materials of animal origin other than those covered by the ‘Note for guidance on minimising the risk of transmitting animal spongiform encephalopathy agents via medicinal products’ applicants are requested to complete table B on ‘Other materials of animal origin’. If an application relates to a medicinal product which contains materials of human origin or these are used during the production process, applicants are requested to complete table C on ‘Albumin and other human tissue derived products’.30

When cell cultures are used for the production of the API, the information is given in the EPAR. The type of cells (expression systems) that are used to produce these products are mentioned. For instance, in the case of EPO the originator used mammalian Chinese hamster ovary cells as host cells and the biosimilar product Dynepo was produced by human cell lines.

This criterion is adequately covered by EMA and is not included in the final set of selection criteria.

A16: Does EMA take into consideration where clinical studies were performed or is only the quality of the study an item of discussion?

This has been published on the EMA web page. On 16 April 2012, the document published is based on ethical and GCP aspects of clinical trials of medicinal products for human use conducted outside of the EU/EEA and submitted in marketing authorisation applications to the EU regulatory authorities.

This document strengthens existing processes to provide assurance to regulators and stakeholders that clinical trials meet the required ethical and GCP standards, no matter where in the world they have been conducted. The reflection paper is part of the Agency's strategy developed to address the challenges arising from the increasing globalisation of clinical research: no matter where you stand today, most clinical trials are being conducted somewhere else in the world, under a different regulatory framework and in a different cultural setting. Yet regulators, healthcare professionals and patients worldwide all rely on the same trial data when making decisions on whether to authorise a medicine to come to the market or not and on whether to use a medicine or not.43 Nowadays, applicants and/or MAHs will be inspected (GCP inspection) within routine inspections required to ensure that information is available. Moreover, the conduct of clinical trials in accordance with ethical requirements will be checked.

When a GCP inspection is performed, the reasons for inspection should be described. Relevant information from the inspection report may be made publicly accessible. The actions taken should be reflected in the public assessment reports.43

This criterion is adequately covered by EMA and is not included in the final set of selection criteria.

A17: How is batch consistency ensured by the manufacturer and/or by third party?

Small variations are an indicator of a consistent manufacturing process. In some cases it might be that the manufacturer lacks experience in batch conformity because of a limited number of batches produced. Access to batch certificates of three recently produced batches by the MAH allows EMA to check and document pharmaceutical quality and also strengthens mutual trust.

This criterion is adequately covered by EMA and is not included in the final set of selection criteria.

B: Product specifications

B1: Are there any differences in isoform pattern in comparison to the reference product or other biosimilars?

The therapeutic relevance of a difference in isoform patterns between a biosimilar and the originator product is addressed by ICH and EMA guidelines. See EMA and ICH guidelines.8 ,35

This criterion is adequately covered by EMA and is not included in the final set of selection criteria.

B2: Are there any differences in drug formulation and administration in comparison to the reference product or other biosimilars?

The packaging material may affect the integrity of a biological medicinal product. Protein instabilities caused by adsorption from equipment surfaces or due to shed particles, such as plasticisers, can be caused by operational or procedural changes or by changing the surface or equipment type itself (eg, membranes, pumps and tubing). Similarly, compatibility studies can identify problems with particular container closures and delivery devices.44 In STEPSelect tenders such an assessment is undertaken based on the National Quality Control and Risk Assessment of licensed medicinal products to assess and score individual entities and this methodology would be applicable.19

This is included in the final set of criteria, as separate criteria. The first four criteria are relevant from a patient’s perspective; the last two are relevant from a pharmacy perspective.

The scoring is dependent on the product classes. General scoring is not possible.

  1. Dosage form or route of administration. A patient intuitive device will score higher than a standard prefilled syringe or injection vial.

  2. Route of administration (higher score when more routes of administration are possible, eg, subcutaneous, intramuscular, intravenous).

  3. More strengths available. The higher the number of strengths available and the more patient friendly, the higher the score.

  4. Dosage frequency: higher score for less frequent administration.

  5. Room temperature>fridge>frozen: decreasing scoring.

  6. Extended stability: higher score.

B3: What is the number of registered indications for the biopharmaceutical/biosimilar?

This criterion is included in the final set of criteria.

This is a standard SOJA criterion and is relevant from a pharmacy perspective. The higher the number of approved indications and the higher the applicability, the higher the scoring. This depends on the drug class in question and no general scoring is possible.

C: Reliability of supply

C1: Does the supplier reliably guarantee the supply of the biosimilar over a long time period?

In this section, we have included the lead time that the supplier guarantees and the actions taken after stock outs to evaluate how the manufacturer and supplier manage these problems. Availability of the drug product is important because otherwise unwanted switches are necessary.

A continuous and reliable supply is a prerequisite. Do not consider formulary inclusion if the local supplier cannot give guarantees about timely delivery for the contract period.

This criterion is part of STEPSelect where shortages of stock clauses and penalties for failure to supply are included in the special terms and conditions of the contract for the purchase of pharmaceuticals to which companies must comply and hence is a prerequisite.19

This criterion should be a prerequisite for individual contracts between hospitals and suppliers and is not included in the final set of selection criteria.

D: Good handling practice

D1: Is the biosimilar delivered according to good storage (GSP) and distribution practice (GDP)?

To maintain the original quality of pharmaceutical products, every party active in the distribution chain has to comply with the applicable legislation and regulations. Every activity in the distribution of pharmaceutical products should be carried out according to the principles of GSP and GDP as applicable.45 ,46 Manufacturers should apply principles of GDP in the distribution chain and in addition should select a shipping company that achieves quality requirements. Therefore the probability of supply chain related detriments, which could, for example, affect the immunogenic profile or shelf life, is reduced. Public health considerations demand that pharmaceutical products should not be treated in the same way as ordinary commodities. Their manufacture and subsequent handling within the distribution chain, nationally and internationally, must meet WHO guidelines and should be rigorously controlled. These precautions serve to assure the quality of products, and to prevent the infiltration of illicit products into the supply system.

For medicinal products imported from third countries, retesting of each batch within the EEA upon importation is required unless an MRA or other relevant agreement covering GMP for the product under consideration is operating within the country where the medicinal product is manufactured. If such an MRA is in operation, batch controls/tests carried out in the country where the product is manufactured are acceptable. It should be noted that MRAs cover batch control/testing and do not cover batch release. Batch release must take place in the EEA territory for every production batch released to market in the EEA, irrespective of whether an MRA with the exporting country is in place or not. Laboratories used for contract testing upon importation of medicinal products manufactured in third countries may be located in any EEA country.30

This criterion is adequately covered by EMA and is not included in the final set of selection criteria.

E: Clinical efficacy

E1: Is the clinical development programme consistent with the current regulatory requirement?

This is extensively discussed in the various EMA assessments. Various additional guidelines were developed during the last few years (product specific and non-specific).

In most development programmes for biosimilar medicines comparability is demonstrated in a stepwise process that starts with pharmacokinetics and pharmacodynamics studies followed by clinical efficacy and safety trials. In efficacy trials an equivalence design is used in most cases. Endpoints for trials are clearly defined in certain product-class-specific guidelines by Committee for Medicinal Products for Human use experts. These primary endpoints do not necessarily have to be the same as those used in the trials with the originator product but deviations from these guidelines have to be justified. The margins of clinical comparability should be pre-specified with respect to statistical and clinical aspects by using the data of the reference product.8

In our opinion this criterion is adequately covered by EMA and is not included in the final set of selection criteria.

E2: Which clinical trials in which patient populations with which designs, endpoints and results were performed?

The criterion is introduced as an estimate of the documentation of the biosimilar in randomised clinical trials.

For biosimilar products that are very similar to the reference product abridged clinical development programmes can be performed because experience is gained with the originator product. Nevertheless biosimilars have to go through the whole registration process and in the end the average time to positive opinion issued by the EMA is longer than for originator biological medicines.47

For products that are less similar to their reference products, more extensive clinical studies are to be done and extrapolation of indications will be more unlikely. An extended clinical study programme would favour products that had not proven close similarity to the reference product because that product would have to demonstrate efficacy and safety in a study with a higher number of included patients. Therefore, this criterion was not included in the final set of criteria.

E3: Are there different results in comparison to the reference product?

This criterion is included in the final set of criteria.

This is a standard SOJA criterion and is usually assigned the highest weight of all selection criteria. No general scoring is possible. The higher the efficacy in randomised clinical trials, the higher the score. If there are no documented differences in clinical efficacy, both drugs will have identical scores.

F: Clinical safety and tolerability

F1: Which (serious and mild) adverse events and in which frequency were they reported in clinical trials with the biopharmaceutical?

This criterion is included in the final set of criteria.

This is a standard SOJA criterion and is usually given a high relative weight. No general scoring is possible. The numbers of patients included in clinical trials will usually be insufficient to identify rare side effects and to make statements concerning safety. Only the judgement of tolerability is possible here. To make statements on safety, many more patients would have to be included and the duration of studies would have to be much longer. The better the tolerability in clinical trials, the higher the score for the drug in question.

Besides clinical trials, data from pharmacovigilance studies will also be taken into consideration in the judgement of relative safety (see section G).

F2: Are there any contraindications, precautions or warnings which are different compared with the reference product?

This criterion is included in the final set of criteria.

This is a standard InforMatrix criterion.18 A large number of contraindications or many warnings or precautions may limit the usefulness of the drug for the treating physician. No general score is possible; scoring depends on the specific contraindications and warnings for specific drugs. The more contraindications and precautions, the lower the score.

F3: Is immunogenicity, as far as known, caused by a homogeneous type of antibody or is there a high intra-individual or inter-individual variability? Is there a difference between biosimilar products regarding drug antibody homogenicity?

This criterion is included in the final set of criteria.

Immunogenicity can affect the safety and efficacy of biosimilars that may lead to clinical consequences such as adverse effects and reduction of efficacy. Immunogenicity is influenced by factors relating to the pharmacological active substance itself. Aggregation of the protein product can also induce immunogenicity, including manufacturing changes, impurities as by products (leachates or leachables=lixiviate which may migrate from the primary packaging into the drug product). Plasticisers, antioxidants and allergens can be leached from rubber in container closures causing product impurities. The presence of degradation products, host cell products and formulation changes and the use of a different packaging system (containers and closures) are other factors that influence safety and tolerability.48–50 In addition to these aspects, immunogenicity may be influenced by factors related to the individual susceptibility of the patient (eg, patients who are immunocompromised), disease, treatment, dose, route or schedule of administration. It is worth mentioning that testing strategies or analytical methods should guarantee a correct immunogenicity assessment. The assessment of immunogenicity is covered in an EMA guideline.51 These immunogenicity assays should guarantee the safety and efficacy of biopharmaceuticals.

The immunogenicity of biopharmaceuticals/biosimilars often cannot be fully predicted using preclinical (non-clinical) in vitro and ex vivo studies and clinical immunogenicity studies are usually required before approval and sometimes also after approval. Moreover, patient-specific immunogenicity may sometimes only emerge after extensive exposure and usage. Further systematic immunogenicity testing may also be required after gaining marketing authorisation. Assessment of immunogenicity may be part of post-approval risk management plans (RMPs) and pharmacovigilance activities.51

No general score is possible. The score will depend on the available data for each individual biological medicinal product. The more (severe) reactions, the lower the score for the specific biosimilar.

F4: Are there differences in the incidence and severity of drug interactions?

This criterion is included in the final set of criteria.

This is a standard SOJA criterion. No general scoring is possible. The lower the incidence and severity of drug interactions, the higher the score for the drug in question. However, it seems unlikely that there will be relevant differences between the reference drug and biosimilars regarding this criterion.

F5: Are there differences in the incidence of local reactions?

This criterion is not included in the final set of criteria but is taken into consideration under criterion F1.

G: Pharmacovigilance

In this section we have included questions about the risks associated with an unwanted immune response established and if there are any specific risk management strategies when the manufacturing process changes.

G1: Does the MAH have a 24 h phone number equipped with adequate personnel to report adverse events?

This criterion should be a prerequisite for individual contracts between hospitals and suppliers and is not included in the final set of selection criteria.

G2: How is pharmacovigilance controlled?

Pharmacovigilance is a continuous process as is the case for biosimilar medicinal products. Knowledge about the equilibrium between efficacy and safety from the pre-authorisation period can change over time after market access. The expanding number of patients exposed to the drug and their characteristics and subtle changes to a complex medicinal product can have an impact on this equilibrium. EMA considers the detection, assessment, minimisation and communication about the associated risks to be the responsibility of the MAH. The MAH must provide a RMP in the application as a part of the CTD. In addition, the MAH must have a qualified person who is continuously available to deal with pharmacovigilance issues. After approval, the MAH should promptly report to the competent authorities or directly to EMA for centrally authorised products in the case of pharmacovigilance issues.9

EU legislation about pharmacovigilance and reporting of non-serious unexpected adverse events is given in EC regulations EC726/2004 and EC 540/95, respectively.52 ,53 These regulations are supported by guidelines condensed in the Eudralex volume 9 ‘Pharmacovigilance’.54 This criterion is adequately covered by EMA.

Healthcare providers can access information about the focus areas of the RMP for the individual product during post-marketing surveillance by searching the website of the EMA where EPAR scientific discussion can be found.37

For healthcare providers, as part of the pharmacovigilance system, it is important to know that product identification is crucial when reporting an adverse event. Only the brand name and the batch number can identify the product in the case of biosimilar medicinal products.9 The use of international non-proprietary names (generic names) and misconceptions about interchangeability between biosimilar and originator medicinal products can stress problems of traceability of the incriminated product.


The checklist created by Krämer et al makes it possible to make informed decisions about the selection and evaluation of biopharmaceuticals and biosimilars. In this article we have added or introduced some issues and questions to complement Krämer's checklist. An expert panel reviewed all issues and to what extent they are covered in EMA laws and regulations. This resulted in a compact and practical set of questions useful for clinicians and pharmacists who want to make an informed decision when selecting biopharmaceuticals.

Since the first biosimilar guidelines were implemented some products have been authorised and additional applications are under review. The number of advice notes and revisions made by EMA on the assessment and development of biosimilars has increased enormously. As a consequence, some issues about the selection criteria might change or indeed other questions might be added. EMA proposed to re-evaluate some topics about biosimilars as a result of the experience gained during the last few years. In December 2012 an overarching guideline by EMA established the general principles to prove comparability between a biosimilar and a reference product.55

The education of healthcare professionals about the differences between biosimilars and originator products is another issue to take into account. We hope that this article contributes to the general knowledge about biosimilar and biopharmaceutical products to help healthcare professionals make well informed choices.

We have not taken the criterion acquisition cost into consideration. The present set of criteria is intended to allow a quick comparison of the reference product and available biosimilars. Usually, the reference product will show the highest score, based on its more extensive clinical documentation. It is up to the individual hospital to determine which difference in overall score is acceptable (for instance 10%). All products that show a score that is less than 10% lower than the drug with the highest score proceed to the second phase of the drug formulary evaluation. In that phase the (tendered or procured) acquisition cost will be the only criterion.

However, it is important to emphasise that the number of biosimilars may increase rapidly due to the expiration of patents, therefore constant review and assessment of the checklist and the regulatory requirements and revisions for biosimilars will be necessary. It is our intention to work together as European hospital pharmacists to form a team of pharmacists with expert knowledge on specific drug classes and their relevant biosimilars.i This expert team will judge the drugs using the criteria suggested in this paper and their assessments will be made available on the internet in an interactive manner. Hospital pharmacists from all over Europe can then enter their own relative weight to the selection criteria, thereby making a personal but evidence-based selection of biosimilars to be used in their own hospitals.

What this paper adds

  • What is already known on this subject

  • In the past few years regulatory bodies like the European Medicines Agency (EMA) and the US Food and Drug Administration (FDA) have built up extensive knowledge on evaluating comparability between a biosimilar and its originator market variant. This knowledge is reflected in a capacious set of guidelines and regulations.

  • When starting biosimilar therapy in patients it is difficult for individual clinicians and pharmacists to judge efficacy and safety because of the complex nature of biosimilars and their reference products.

  • What this study adds

  • A panel of biosimilar experts reviewed guidelines and regulations of EMA and FDA and judged what would be relevant for biosimilar selection criteria in clinical practice.

  • This article provides a set of 10 clinically relevant selection criteria to which individual weights can be assigned to subsequently score the originator and the biosimilar product.


The authors would like to thank Dr Joerg Windisch, Sandoz for his comments that helped to improve this manuscript.



  • Contributors All authors delivered a proportionate contribution to the creation of this manuscript.

  • Competing interests None.

  • Provenance and peer review Commissioned; internally peer reviewed.

  • i The authors of this article are creating a European panel of experts to evaluate biosimilars. Hospital pharmacists with expert knowledge of the following topics are invited to apply to join the panel: biosimilars in general; erythropoietins; filgrastim and other growth factors; somatropin. Please contact n.boone{at} or r.janknegt{at} for further information.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.