Economy by design

Published on: 

Pharmaceutical Technology Europe

Pharmaceutical Technology Europe, Pharmaceutical Technology Europe-12-01-2007, Volume 19, Issue 12

Incorporating quality and economy into downstream purification processes can expedite first-in-human clinical trials, product licensure and technology transfer. Experience in the chromatographic purification of biopharmaceuticals enables the use of downstream processing heuristics to produce target molecules in cost-effective processes suitable for regulatory scrutiny.

One of the latest trends in biotechnology is the application of the concept of quality by design (QbD). QbD entails designing therapeutic products to meet patients' needs and then consistently producing them. Consistency is achieved by understanding the impact of starting materials, critical process parameters and product quality attributes, and controlling variability. In today's marketplace, concomitant with designing in quality, companies producing biotherapeutics need to consider economy. "Quality" does not have to mean "expensive" because economy and quality can be built into a process at the same time. It isn't logical to design a high-quality biotherapeutic that no one can afford.

Fortunately, today's technology enables reasonable production costs. For example, in the manufacture of monoclonal antibodies (mAbs), the cost for producing one gram is estimated to be $100–300 (€68.32–204.95), depending on scale of operation. Lower costs have been published based on a production estimate of 10 tons per annum. More specifically, the cost for Protein A resins used in a capture step for mAbs is normally 3% of the total costs when used for more than 30 batches. The cost for ion exchangers in antibody purification is in the order of $0.5–1.0 (€0.34–0.68) per gram.1

More than 15 years of successful production in the biotechnology industry have resulted in downstream processing heuristics (i.e., rules-of-thumb) that point towards economical purification strategies (Table1). We will look at these individually.

Table 1 Six heuristics for economy by design.

Address current and future costs in development

For companies with a history of producing biopharmaceuticals, the cost-effectiveness of various production tools is usually known. However, this is not the case for many start-ups, especially those in which the downstream process developer has little or no industrial experience. (For information on protein purification heuristics, see Protein Purification Handbook 18-1132-29 [GE Healthcare, Sweden.]) Cost-effectiveness is designed by understanding processing needs.

In some downstream processes, utilization of reusable chromatography media and filters may be necessary to obtain good process economy. This is particularly true for large-scale production as multiple batches per year are required to meet market needs. In one economic evaluation of reuse, media and validation costs were calculated. The greatest cost for establishment of reuse was related to its validation. Incremental annual savings were approximately $113 million (€77.18)/year for 10 lots, $15 million (€10.25)/year for 30 lots, but by 90 lots the savings were down to $0.7 million (€0.48) and the validation costs had risen significantly.2

For large volume processes, throughput requirements are often critical, especially to reduce initial volume and minimize contact with proteases. In this case, chromatography media with higher capacity and higher flow properties will enable a more economical process design. By reducing initial volume, buffer, including costly water-for-injection (WFI), consumption needs will be lowered. A reduction of buffer consumption by one-third has been estimated to reduce cost by 6%.1

For other processes, disposables afford better economy as they can provide sanitary, ready-to-use processing equipment that reduces time to first-in-human (FIH) studies. Replacing chromatography and filtration media after each use may be less costly than trying to validate cleaning routines, particularly early in process development. Disposables are often the most economical solution for multiproduct facilities.

Storage of dilute buffers can also increase costs, but these can be minimized by using an automated in-line dilution of concentrated solutions. Clearly, there are many cost-saving measures that can be taken. During development, evaluate the process transfer capabilities of the technologies intended to decrease processing costs. If they cannot be transferred and new tools are needed, there could be a significantly negative impact on costs.

Evaluate the costs for buffers and other processing agents; column re-use, repacking production columns; storage; cleaning; extractable studies for disposables; automation and personnel costs. For example, packed column storage costs include facility space, storage solutions, removal of storage solutions and their disposal, and testing for bioburden and column integrity after storage.

Design a process that is suitable for intended use and anticipated dose of product

An economical process should be capable of producing a quality product. Defining the level of acceptable quality is certainly a challenge in initial development. However, industry experience and regulatory expectations dictate some commonalities for purification of therapeutic proteins and other biological molecules; for example, DNA plasmid vaccines. Target values for removal of host cell proteins and DNA; cell culture media and processing agents should be dictated by product indication and its maximum predicted dose.3 Over-specifying product attributes can be very costly. Keep in mind that purer is not always better, and a more highly purified product may alter safety, efficacy and potency.

For products derived from mammalian cell substrates, human or animal sources such as plasma or transgenic animals, demonstration of viral clearance is required. If the cell substrate is well-characterized and has no infectious viral particles, all raw materials controlled, and manufacturing protocols performed in a suitable environment with proper adherence to GMPs, then striving to achieve an excessively high overall log reduction value may not be warranted.4 Extra steps inserted into a process for virus inactivation and removal are usually very costly. Consider an inactivation step. There may be product loss as a result of aggregation; a subsequent unit operation may be required to remove an inactivating agent; increased analytical methods may be needed; more WFI and other raw materials might be necessary; processing time extended, and so forth. Regulatory authorities are not likely to request removal of a step so this step becomes "grandfathered" into the process, often at an unrealistic cost.

Of course, it is essential to consider patient safety during process design. Regulatory holds are very expensive, perhaps delaying shareholders' first-to-market expectations. Part of an economical design strategy is to perform a risk assessment that addresses impurities, potential adventitious agents, patient population and dose.


Evaluate regulatory compliance levels for early clinical trials through license application

During the last few years, there has been an increased emphasis on safety for patients enrolled in clinical trials. This is especially true in Europe where legislation was passed requiring inspections and a greater level of GMP compliance for manufacturing investigational medicinal products.5,6 However, in the US there has been a push to get more investigational products to patients in need of alternatives.7 One outcome has been a draft document on reduced compliance with cGMP for Phase 1 clinical studies.8 (Note, although the ruling for implementation of this approach was rescinded, the draft guidance still provides relevant information for phasing-in compliance.)

In a global economy, it is important to consider regulatory compliance expectations for all regions in which you want to market a biotherapeutic. Designing a process and product with quality attributes that only satisfy one region of the world could limit profitability by requiring expensive redevelopment, new clearance studies and even repeated clinical trials.

Minimize the number of processing steps

Every unit operation adds to production costs. By minimizing the number of intermediate steps (i.e., dilution, concentration and buffer exchange), these expenses can be reduced. The number of unit operations should be determined by a risk assessment and process capabilities. Consider current expectations for removal of impurities and control over adventitious agent contamination, and utilize analytical methods to design a process with as few steps as possible. Evaluate new analytical tools as they become available for process development; quantitative polymerase chain reaction (Q-PCR), for example, has enabled evaluation of DNA and virus removal at a reasonable cost during process development.

The optimal time to evaluate how many purification steps will be needed for a new process is during its development, as decreasing the number of steps for a licensed product may present a regulatory challenge. For example, recent advances in mAb purification have enabled a reduction in the number of purification columns after a Protein A capture step. Instead of two columns following Protein A, one multimodal chromatography media can be used to achieve a two-step purification process for many monoclonal antibodies.9 This approach can provide savings in both operating costs and time, yet still remove impurities (Table 2). While removal of dimers (and other aggregates) and Protein A is the same for both processes shown in Table 2, the overall removal of host cell proteins is greater in the three-step process. However, the level of 7.5 ppm achieved with the two-step process might be sufficient. That judgment would be based on preclinical and clinical data and a risk assessment.

Table 2 Comparison of a 2-step process using MabSelect SuRe and Capto adhere with a 3-step process using MabSelect SuRe, Capto S and Capto Q.

Design a process that is robust and transferable

Designing robust purification operations is an iterative process requiring both empirical work and design of experiment (DoE).10 A robust process requires an understanding of what each step accomplishes; for example, level of reduction of each impurity and preservation of product integrity. A robust process is one in which reasonable control parameters are utilized. As more sophisticated analytical tools become available, especially those that enable feedback with process control, the application of PAT will become more common in downstream processing, which is likely to enhance process robustness.

A robust process offers an economic advantage by preventing product batch failures that might occur in a process that is defined too narrowly to account for unavoidable variability. Take, for example, buffer preparation and column qualification measurements. In the case of buffer preparation, a reasonable range for pH might be ±0.1 pH unit. This is an achievable value for large-scale manufacturing. For column qualification, an acceptable height equivalent to a theoretical plate (HETP) value might be much larger for a capture step than for a final polishing step in which high resolution is needed.

During the design of a robust process, cleaning, sanitization and storage should be considered. Contamination with microorganisms during purification poses a significant economic risk to the product and a safety risk for its ultimate customer; a human patient. The selection of chromatographic media and raw materials is paramount to providing a high-quality product and an economical process. Ensuring minimal carryover from column runs is essential, and selecting materials that can be cleaned and sanitized in place allows the use of sufficiently stringent conditions.

High-quality raw materials should always be used. An inferior, potentially contaminated, or impure raw material can only lead to costly batch failures and possible product adulteration, which can incur a loss in product sales, and even a loss of reputation if that product gets into the marketplace.

Equipment requirements should also be designed into a downstream process. Leachables from resins and filters, as well as potential extractables from polymers in contact with product must be evaluated. It is more economically effective to use materials for which some data and analytical methods exist to address these issues. While the cost of raw materials with supporting documentation may be higher than others, the cost for designing and performing the analysis is usually even greater. Also consider the sourcing of all raw materials, as the concerns related to potential BSE contamination must be addressed.11,12 Variability in the strictness with which this issue is enforced varies worldwide and a risk assessment can be a cost-effective approach in demonstrating an insignificant risk.

As a process is being designed, think about how it will be transferred to pilot plant, full-scale production or contract manufacturer. If the process used in very early development requires that only one specific operator be present to make decisions, it is likely that scale changes and technology transfer will be problematic. A transferable process is one that has defined operating procedures and is understood as much as possible. Understanding a downstream process requires the use of orthogonal analytical methods that are qualified and reasonably robust. The application of in-process analytical methods that are transferable to manufacturing is clearly an advantage for smooth technology transfer.

Transferring a process to another company can be quite challenging; this includes transfer to a contract manufacturer. The product owner may not want to divulge too much confidential information, which often leaves the contract manufacturer with insufficient information to successfully run the process. The process development scientist should provide available protocols, descriptions of reagent quality and other necessary know-how. For example, packing large columns requires a certain level of skill. Transferring that skill often takes more time than anticipated and may lead to costly production delays. If the process developer has defined reasonable acceptance criteria for column packing, the technology transfer can be expedited. Other considerations include system wetted materials and configuration. When a different chromatography skid design is used, variation in sensitivity of in-line controls, such as those for UV and conductivity, may alter chromatographic performance.

Another technology transfer occurs when small-scale clearance studies are conducted at a contract testing laboratory; for example, for viral clearance. It is clear that scale changes will impact the technology transfer. The best approach is for the process developer to validate the scale-down at the site where the appropriate analytical methods are available. Otherwise, shipping of samples must be validated — further adding to the cost. For hazardous materials, such as virus or prion proteins, a mock spike (i.e., one without the actual agent being tested) should always be evaluated prior to the actual clearance study to determine the effect of volume, and any of the other components, on the spike material that may alter performance.

Communicate, plan for change and plan for validation

As process understanding increases during development, changes are made. This is particularly true for Phase 1 and 2 clinical studies. Once in Phase 3, making changes can be excessively costly — mostly because of the need for bridging studies and the ability to interpret clinical data when more than one product is being used for the same study. One of the greatest hurdles seems to be convincing upper management that process development costs and sufficiently allocated time are good business strategies. The costs for repeating a clinical study or having to implement a bridging study are huge compared with those for good process development that applies know-how based on industry experience with biopharmaceutical production.

Whenever changes are made during development, it is necessary to assess comparability, which requires extensive analysis, in-process and API retention samples, and working standards. For licenced products, change is inevitable as technology improves and expectations from regulatory bodies evolve. Changes in regulatory opinions often arise from the finding of a new risk factor or introduction of a technology that can enhance patient safety. New technologies may enhance quality, but they can also significantly add to production costs.

It is important to always evaluate new technologies, especially those that enhance patient safety, increase productivity and reduce overall cost while still maintaining production of a high-quality product. At the same time, consider the economics of making a change in conjunction with potential regulatory issues that might lead to delays in implementation. Clearly there is a balance between making a change in productivity and preventing release of a batch. In the US, comparability protocols provide a mechanism for implementing changes without excessive delays.13 ICH has also produced a guideline on comparability for biotech products,14 while in the EU there is a mechanism for making changes to a marketing authorization application (MAA).15,16 The importance of the application of risk management when production changes are made is illustrated by the EMEA guideline that came into effect in 2005, which requires that when an application for a biotech product involves a new manufacturing process, a risk management plan must be submitted along with the application.17

Wasted time as a result of poor communications is costly and communications within an entire organization are essential. Manufacturing capabilities need to be understood by process development, just as manufacturing must understand what was done in clearance studies at small-scale to avoid moving the process away from the acceptance criteria of those studies.

Take advantage of the regulatory agencies' willingness to communicate for developing projects. Both the US and EU have mechanisms for asking questions. FDA encourages pre-IND meetings, and even provides a list of frequently asked questions that can be used to prepare for the meeting.18 In Europe, EMEA now offers support for small- and medium-sized enterprises.19 Although EMEA charges, the advice could provide a significantly greater cost-savings derived from taking the right strategies to produce your biotech product.

A validation master plan can prevent inadvertent omissions. The terminology used for qualification and validation is often inconsistent, even within one firm, and this can lead to costly delays. A PDA technical report on process validation of protein manufacturing and the ICH Q7A guideline on APIs provide a platform for consistent terminology that can be used worldwide to enable more rapid implementation of successful validation.20,21


The heuristics provided here offer a strategy for economical development and licensure of high-quality products. Designing in quality can also mean designing in economy. Economic processes can be "smart" processes that remove impurities, build in safety margins for adventitious agent clearance and maintain product integrity in a minimum number of steps. Such processes should also be robust and capable of being transferred. Both robustness and transferability require that the process is understood. That understanding, in turn, requires the use of multiple, reliable analytical methods — some of which should be capable of being placed in, or at, line in manufacturing.

A patient safety risk translates into a business risk. Product may be lost and, significantly, an adverse event because of poor product quality is likely to harm the value of a company. In this highly regulated industry, an economical process has to be one acceptable to regulatory bodies charged with making decisions about whether the product should be introduced into humans and whether is it manufactured consistently. Understanding and accounting for worldwide regulatory expectations in process design can enhance marketability of a biotherapeutic.


The Polymerase Chain Reaction (PCR) is covered by patents owned by Roche Molecular Systems and F. Hoffman-LaRoche. A licence to use the PCR process for certain R&D activities accompanies the purchase of certain reagents from licensed suppliers.

Gail Sofer is director of regulatory compliance, Fast Trak life sciences, at GE Healthcare Bio-Sciences (NJ, USA). She is a cochair of the PDA advisory board and also serves on several other advisory boards. She is co-author of a recently published book on process chromatography.

Guenter Jagschies is in his 22nd year with GE Healthcare Life Sciences (former Amersham) and has held senior management positions in sales, marketing, training and consulting within the bioprocess segment of the company business supplying the industry with downstream processing solutions for development and manufacturing. His current role is senior director biopharma technology working globally with industrial collaborations and as business advisor for the life sciences R&D team. He is based in Uppsala (Sweden).


1. L. Hagel, G. Sofer and G. Jagschies, Handbook of Process Chromatography: Development, Manufacturing, Validation and Economics (Elsevier, The Netherlands, 2007).

2. A.S. Rathore and G. Sofer, "Life span studies for chromatography and filtration media", Process Validation in Manufacturing of Biopharmaceuticals (CRC Press/Taylor & Francis, Boca Raton, FL, USA, 2005) pp 169–204.

3. H. Simmerman and R.P. Donnelly, BioProcess Int., 3(6), 32–40 (2005).

4. ICH Q5A — Viral Safety Evaluation (1997).

5. European Commission, Volume 4, Annex 13, Manufacture of Investigational Medicinal Products (2003).

6. EMEA/CHMP/BWP/398498/2005-corr, Guideline on Virus Safety Evaluation of Biotechnological Medicinal Products, Draft (2006).

7. FDA Guidance for Industry, Investigators, and Reviewers, Exploratory IND Studies (January 2006).

8. FDA Guidance for Industry, INDs: Approaches to Complying with CGMP During Phase 1, Draft (January 2006).

9. GE Healthcare, Capto adhere data file 28–9078–88AA

10. G. Sofer and M. Ahnfelt, BioPharm Int., February 2, Supplement (2007).

11. EMEA, Note for Guidance on TSEs EMEA/410/10-Rev. 2 (2002).

12. FDA, BSE/TSE Action Plan, Federal Register, 66(163), 44146–44149 (2001).

13. FDA, Guidance for Industry, Comparability protocols: chemistry, manufacturing, and controls information (2003).

14. ICH Q5E — Comparability of Biotechnological/Biological Products Subject to Changes in Their Manufacturing Process (2004).

15. EMEA Post-Authorisation Guidance for Users of the Centralised Procedure (2007).

16. The Rules Governing Medicinal Products in the European Union (2007).

17. EMEA/CHMP96268/2005 — EU risk management plan (EU-RMP) (2005).

18. Frequently Asked Questions on the Pre-Investigational New Drug (IND) Meeting (2005).

19. Addressing the needs of small- and medium-sized enterprises (SMEs).

20. PDA Technical Report 42 — Process Validation of Protein Manufacturing (2005).

21. ICH Q7A — Good Manufacturing Practices for Active Pharmaceutical Ingredients (2000).