Validation of the SAS System

Validating a new version of SAS on a production server used to be a daunting task. The SAS System version 9.1.3 and 9.2 ships with user friendly installation qualification tools. This is coupled with existing tools that make it easier to validate SAS. Besides qualifying the installation, there are other tasks and components of the system that need to be validated or verified. Some of these components include:

1. Backward compatibility issues with older versions of datasets and format catalogs.


2. Validating multi use macros and standardized code templates.


3. Verifying stand alone or project specific programming and output.


4. Effects on standard operating procedures and programming practices.

The interconnectedness of the SAS computing environment does require considerable efforts in validating a SAS System. However, if this is executed successfully, it can allow for greater traceability between output, programs and source data. The performance qualification also sheds light on ways of optimizing the work and data flow of your computing environment. The many benefits of performing validation of the SAS System will outweigh the costs. In addition, it is a requirement within a regulated environment so it is recommended to be prepared.

Validation Overview
Validation of a SAS system most commonly occurs during an upgrade from an older version of SAS or moving to a new platform. The examples used in this paper include migrating from SAS 8.2 to SAS 9.1.3 and moving from a legacy operating system to the windows platform. In either case, similar validation challenges are confronted. It is recommended that you first acquire a global view of the system and identify the architecture. Only after gaining this perspective would it be useful to then zoom in on individual components. This allows you to access the scope and interconnectedness of each component so that your validation efforts are balanced and thorough. Once the architecture is clearly understood, the requirements and functional specifications of each component are documented. These functional specifications then drive the validation testing.
It is important to follow these steps in a systematic and orderly fashion since they are interdependent. Documentation of each step in the validation process is also essential in capturing and proving that the validation effort was done properly. Besides documenting each step, it is also important to capture the traceability of each validation task. For each test case that is performed, there is an associated functional specification which then is connected to the requirements for a particular component of the system as a whole. The map or traceability matrix that ties all these validation components together is pivotal to an auditor. Proper documentation will make the difference between a successful validation audit and a complete failure.

The main goal of the validation effort is to ensure that the installation and implementation of the SAS system and its associated tools function as intended by the vendor (SAS Institute) and your organization. The validation will ensure this success. In addition to this goal, the documentation of your validation effort will also ensure the integrity of your computing environment and be in compliance with regulatory requirements such as the CFR Part 11 within the biotechnology and pharmaceutical industry.

System Components
The first step in your validation effort is to understand what it is that you are working with. The SAS System, as delivered to you in a series of CDs, is a system which contains modules such as Base, Stat, Graph and other components of SAS. This however only makes up part of the system that you are implementing in your organization. The SAS software fits into a computing environment that interacts with other software and hardware. If you were to take into account all the associated hardware and software that SAS interacts with, this is what is considered the “SAS System” from a validation perspective. It is therefore important for you to take the right steps to identifying and documenting all these components.

Step 1: Identify all the hardware components of your computing environment. For example:

Hardware Component	Name
SAS Application Server	SASAPPSRV
SAS File Server	SASFILESRV
Client Desktops	CLIENTDSK

Keep It Clean

By Gina Shaw

Challenges in biopharma contamination control

In June, biotech giant Genzyme was forced to temporarily shut down its plant in Allston, Mass., when a virus was discovered in one of the plant’s six bioreactors. Genzyme said that although the virus could not infect humans, it could impede cell growth and slow production of the drugs manufactured there—Fabrazyme (agalsidase beta), which treats Fabry disease, and Cerezyme (imiglucerase), a therapy for Gaucher disease. The production suspension meant temporary rationing of the drugs, with patients asked to skip doses.

The suspension was projected to last until the end of July, and the rationing left patients taking the drugs—about 8,000 worldwide—worried about how they would get treatment. Other companies hurried to fill the gap; Shire PLC filed a new drug application with the Food and Drug Administration (FDA) for its Gaucher drug, velaglucerase alfa, under a treatment protocol that would allow the company to market the drug before approval, requiring that it be initially provided free of charge.

Meanwhile, Genzyme was expected to lose between $100 and $300 million in manufacturing revenues as a result of the shutdown. This was actually the second time the virus—called Vesivirus 2117—had hit Genzyme’s production facility; the first incident occurred in 2008 and caused declines in cell productivity. The 2009 contamination was detected by Genzyme’s own monitoring system; the company had recently developed a highly specific assay for Vesivirus.

Lessons Learned

The Genzyme incident underscores the importance of contamination control and monitoring in the biopharmaceutical industry, said Ken Christie, senior director of consulting services for VTS Consultants Inc. “Biopharmaceuticals are primarily sterile products by nature, and anything labeled a sterile drug is something that the FDA, or any European regulatory agency, is going to see as possessing the highest level of risk to the public. As a result, the control of potential sources of contamination becomes of primary importance. Because biopharmaceutical companies are growing cells and working with reactor processes, the challenges are even greater.”

Multiple elements must work effectively together to ensure good contamination control, said Christie. They include:

• facility and equipment design;
• environmental systems;
• maintenance;
• personnel;
• monitoring; and
• cleanup.

Of these areas, the facility and equipment design element remains one of the most significant challenges for biopharmaceutical companies today, according to Rebecca Brewer, director, consultancy services, validation and GMP compliance with Dober Group. “The reactor design is significantly more complex than in traditional pharma, and the bioreactors, in particular, become challenging if the cleaning-in-place (CIP) and sterilization-in-place (SIP) systems have not been designed to properly access and provide cleaning to all surfaces.”

That happens fairly often, Brewer said, because of the complexity of the equipment. People who design bioreactors are not experts in CIP, and people who design CIP are not experts in bioreactors. And they talk to each other, she said, “less often than you would hope.”

There’s no such thing as a perfectly cleanable system unless it’s “an empty tank with no features in it,” Brewer said. “Baffles, dip tubes, bottom-mount agitators—reactors have all sorts of things that make them difficult to clean,” she explained. “Because of the conditions that the product sees during processing, including heat or foam generation, there are difficult-to-remove soils in challenging locations.”

Biotech facilities today are seeking to augment CIP design by modifying factors such as spray ball type or spray ball position, or by changing them entirely. “In cases where more simple engineering fixes can’t be accomplished, you end up having to augment CIP with additional manual cleaning, which is never anybody’s favorite,” said Brewer. “If you can’t overcome a shadowed area or a blind spot, you’re left with very few choices in terms of how to manage it. But if you don’t look after those issues, you will develop buildup in [the] system, with contamination potentially spoiling batch after batch of product.”

Brewer suggested that the best approach to CIP design involves due diligence during engineering of the system, testing for coverage as it’s designed and built. “Typical coverage testing tests one spray device or one flow circuit at a time. And in the real world, to optimize the cleaning cycle, people want two or three fluid paths on at the same time. If you’re going to do that, you need to be sure that’s how they did the original testing, in order to get coverage without interference of one flow with another or cancellation of spray from having them strike each other in the middle of the vessel and miss their target.”

Source: Al Smith, PhD

On the left is a cluster of six calciviruses (genus vesivirus), the type of virus that recently caused contamination at a Genzyme plant. On the right is a magnified view of the surface of a single vesivirus obtained using cryomicroscopy and having a resolution of about 20 angstroms.

Cleaning, Sterilization Interface

The interface between cleaning and sterilization is also important. “If you have a microbiological contaminant in the system, biotech firms find it very difficult to eradicate it once it’s taken hold and perhaps formed a biofilm,” Brewer said. “Some of this may be due to construction and CIP issues and some due to inadequate SIP to conquer the problem after a CIP process has left residue behind. You have to look at CIP and SIP as partners in the same goal—microbiological cleanliness.”

Another key challenge to maintaining a sterile environment in a biotech facility is the very people who run it. “You can build a facility and put in systems that would give you a ‘cleanroom’ environment, but once you bring operators into that picture, the potential for contamination doubles. People are the biggest source of contamination,” said Christie. That’s where design of the facility itself comes in—systems literally have to be protected from their operators.

The FDA prefers isolator technologies when possible and, when not, restricted access barrier systems (RABS). Barrier isolators enclose the system and do not require a separate cleanroom, while RABS systems must be placed in a cleanroom. They cost less than isolators and appear to achieve the same results, but the agency does indicate a general preference for isolators.

“All of these types of systems are designed to minimize the amount of interaction that an operator has around the critical areas of where a product is filled and stoppered,” Christie explained. “They restrict an operator’s interactions around the critical step where an open vial or syringe or plastic bottle is.”

Isolators and RABS have been around for quite some time, but recently their designs have become more efficient and better able to accommodate a variety of sizes in the filling lines they can encase. “You can have a small company down the road that may need to only fill several hundred vials for a clinical study, while larger companies will have lot sizes of hundreds of thousands of vials at a time,” said Christie. “Because of that, the overall size of the filling line gets to be rather large, and companies are coming up with much more efficient design, with access to motors and things that might break without jeopardizing the cleanroom environment in which these things are placed.”

A current RABS design, for example, allows access to the motorized components of the filling unit from the outside wall of the cleanroom. “This way, an electrician or mechanic does not have to go into the clean area and open the machine,” Christie said. “All access is from the wall on the non-controlled side of the equipment, so the environment where the filling actually occurs is not jeopardized. That’s involved in a lot of the new designs.”

If you have a microbiological contaminant in the system, biotech firms find it very difficult to eradicate it once it’s taken hold and perhaps formed a biofilm.
—Rebecca Brewer, Dober Group

Microbe Messes

Although the bane of Genzyme’s existence these days is a virus, many biotech facilities struggle with prevention, control, and cleanup of molds, according to Jim Polarine, a technical services specialist at STERIS Corporation who focuses on microbial control in cleanrooms and other critical environments.

“I was brought into a site in California where they had such a problem with molds that it got into the stainless steel and ductwork in the cleanroom, and they had to spend about $200 million to replace the stainless steel and the flooring,” he said. “By the time I got called in, contamination was running rampant. We tried all the usual chemistries and they could control it for a while, but they had let things get out of hand and the mold was innately in the surfaces. When it gets into the flooring and up into the gaskets of the HEPA filter, then you have a big problem.”

At another facility, this one in Boston, an operator dropped a glycol tube, which broke on the floor. Growing in the tube was a spore-forming bacterium called Bacillus polymyxa. “He tracked it all through the cleanroom and the entire biotech site, and it ended up getting into the product, on the walls, and on the floor,” Polarine said. “Needless to say, it became a very long, drawn out process.”

STERIS most often uses a liquid cold disinfectant called Spor-Klenz, a stabilized blend of peracetic acid, hydrogen peroxide, and acetic acid. “We’ve also been getting some accounts where we bring in vaporized hydrogen peroxide for room decontamination, to kill really resistant or recurrent spores. For regular routine decontamination of the cleanrooms, I see a lot of end users using phenolic disinfectants. Probably 80% of the industry in the U.S. and Puerto Rico uses phenols, which can kill higher end bugs such as mycobacterium and [have] surfactant technologies to remove particles, soils, and dirt.”

Justification Needed

Although cleaning validation methods are fairly standard, Brewer said that some in the industry still aren’t adequately justifying their decisions. “Particularly in biopharma, I still see companies relying on not more than 10 ppm [parts per million] in the next batch without adequate justification.” That’s the most common quality threshold applied to pharmaceuticals, but it’s been called a convenience value and may not have a lot of meaning.

“I also see people employ the purified water limits of 500 ppm carbon without adequate justification of what that particular limit means,” Brewer added. “Part of the reason that exists in biopharma, in particular, is that the determination of what would be a therapeutic dose-based limit or an appropriate contamination-based limit—particularly in upstream processing—is very difficult to do. As a result, they fall back to generic values. But that’s not adequate science for cleaning validation limits. They should be driven by the impact of carryover to the next drug they’re going to be producing.”

At Genzyme, Vesivirus 2117 was discovered by the company’s own monitoring, a still-evolving process. “Past practices had companies take periodic samples throughout the course of the filling process, which could last two shifts or 24 hours. The attitude was to sample at the beginning, middle, and end of a run, which might have occurred three times over a 16-hour period,” said Christie. “It varied. Now, the FDA wants to see more routine monitoring during the course of every shift, with their preference a continuous monitoring of the environment for both nonviable particulates and viable organisms.”

The technologies available to achieve that goal are becoming more user-friendly. “Technology is being developed where the counts or organisms can be downloaded to a database for trending of the results, which allows companies a better ability to detect potential sources of contamination before it adversely affects a filling process,” Christie said. “But don’t forget, the FDA wants to see you do something with that trend data. They like to see trends of environmental monitoring results—how do they vary between seasons? Is there one person who sticks out for routinely failing or having counts detected? But their next concern is what are you doing about it.”

Aspects of contamination control can’t be looked at in isolation. “Facility design, environmental systems, routine maintenance and calibration of those systems, the sensors used to monitor them, training of personnel—they all have to be controlled to give you a final ability to eliminate potential sources of contamination,” Christie said. “It’s really astonishing to somebody coming into biotech for the first time from, say, solid products or over-the-counter. The levels of regulation you’ll have to deal with are like nothing else you’ve ever encountered.”

Shaw is a freelance writer based in Montclair, N.J. Reach her at ginashaw@vagabondmedia.com.

IT Glue Holds Virtual Enterprise TogetherEFFECTIVE OUTSOURCING

By Kathy Pelley and John Postle

© DREAMSTIME.COM

IT Glue Holds Virtual Enterprise Together

Information technology eases dealings with contract organizations

Pharmaceutical and biotechnology companies have undergone tremendous change during the past dozen years. Until recently, with few exceptions, most operated in vertically integrated silos, with each company managing its own research and development, manufacturing, sales, and marketing under one roof.

Today, all that has changed. Biotech and pharma have accepted the compelling economic logic of outsourcing. Nearly every modern-day biopharm company has developed a network of contract research organizations (CROs), contract manufacturing organizations (CMOs), and other service providers to help manage clinical trials, submit regulatory filings to the Food and Drug Administration (FDA), and scale up for commercial production.

Managing a disaggregated enterprise is difficult under any circumstances; it is even more so when human safety, drug efficacy, and hundreds of millions in development costs are at stake. Fortunately, many biopharms realize that savvy managers, well-planned service level agreements (SLAs), and effective use of information technology (IT) can provide the glue that holds the virtual enterprise together-even when IT itself may be outsourced. Shortcomings in any one of these areas can cause the outsourced supply chain to fail.

The CMO Relationship

Consider a sponsoring biopharm's relationship with a CMO. The sponsor must evaluate and select the CMO that best suits its business model and goals, negotiate SLAs that anticipate and avoid problems before they occur, and employ IT to monitor and manage the relationship while ensuring regulatory compliance.

As supply chains get longer and go offshore, sponsors will have to continuously improve upon IT practices to manage the virtual enterprise. Outsourcing sponsors cannot depend on CMOs to establish IT good practices. In the end, the sponsor is responsible for every act of noncompliance that occurs in its supply chain. It is the sponsor's drug and patient-and the FDA will hold the sponsor accountable.

CMOs come in every flavor, from global manufacturers who manage extended supply chains of their own to boutiques making small batches to support clinical trials. Small and mid-sized biopharms with little outsourcing experience and no products on the market may do better outsourcing to an onshore boutique accustomed to supplying CROs with clinical batches. Large sponsors may want to leverage their global footprint and save money by offshoring to a group of related CMOs (see "Offshoring Risks and Rewards," p. 38).

Regardless of the supplier strategy chosen, it is important to winnow prospective CMOs to a manageable list based on their proposals, management experience, references, longevity, and industry track record.

Let's say that three CMOs have made the initial cut and the sponsor's immediate concern is to support clinical trials that have been outsourced to a CRO. One CMO is a boutique that specializes in making and managing batches for clinical trials. The other two are larger and offer clinical support through offshore commercial scale-up, as well as packaging, labeling, radio frequency identification tagging, commercial distribution, post-marketing surveil-lance, adverse event reporting, and other services.

Due diligence is critical. Ultimately, the FDA holds the sponsor responsible for the regulatory compliance of its suppliers. The sponsor must first visit the CMO and conduct a pre-audit walk-through of its manufacturing systems and processes-including IT-because those systems will be used by the sponsor to monitor and manage the CMO while providing an electronic audit trail for the FDA.

At this stage, the sponsor wants to know if the CMO can protect its intellectual property, manufacture sufficient product to support the clinical research organization, comply with the FDA's current Good Manufacturing Practices (cGMPs) and 21 Code of Federal Regulations (CFR) Part 11 regulations, and provide actionable data and reports in an electronic format that can be used to manage the business relationship. For example, the CMO's IT systems must be able to collect and report data on material handling processes, lab analyses of active pharmaceutical ingredients (APIs) and excipients, and the manufacturing controls (including sterilization procedures) that result in the finished product.

The IT Pre-Selection Audit Process

It is far too costly and time-consuming to conduct a complete field audit of each of the CMO finalists. But in most cases, an experienced quality assurance (QA) or quality control (QC) professional can audit enough critical IT systems to identify any weak technology and business processes that must be rectified if a CMO is to be hired. It takes more time to perform a pre-selection audit of a highly automated CMO. Depending on the size of the project being outsourced, a pre-selection audit should, as a rule of thumb, take from two to four days.

Conducting a pre-selection audit is a lot like "managing by walking around." An experienced QA or QC professional can spot a lot of problems just by wandering around the manufacturing floor and noting practices. Is the CMO using a 30-year-old Digital Equipment Corporation PDP 11 computer with no network connection to gather data on its manufacturing systems? Does the company even have a manufacturing execution system? How is manufacturing data shared with the enterprise resource planning system? What kind of IT infrastructure is available on the factory floor? Does everybody have password access to the Pentium server near the loading dock and the forklifts? Are there backup systems in case the server is flattened in an accident?

Some manufacturing environments are so chaotic that a sponsor won't need to interview the plant manager, review documentation, or query an IT system to know that a great deal of remediation and process improvement are going to be necessary.

Of course, the auditor will want to visit the CMO's IT department to interview key personnel and evaluate their expertise in managing IT systems in an FDA-regulated environment. What kinds of electronic reports does the department provide to other sponsors? Has it adopted industry standard data formats? Has it ever failed an FDA inspection? Does the department manage operational systems using QC methodologies like IT infrastructure library (ITIL) or Six Sigma? Or deploy validated and qualified systems using Project Management Institute (PMI) techniques? Is IT staff certified in any of these disciplines? Does the company support continuing education for IT staff? No CMO is going to be perfect, but sponsors should expect a good explanation for the answer to every one of these questions.

Most startups need basic infrastructure and applications until they reach the investigational new drug (IND) phase. Once they get IND approval and begin life as a viable pharma or biotech, data is their product, their most important intellectual property. Clinical trials require the introduction of outside partners such as CROs and CMOs; critical data is then generated and shared in multiple places. Control and protection of the data's integrity is key.

IMAGE COURTESY OF THE COURT SQUARE GROUP

Once the sponsor has addressed some of the IT-related business processes and IT systems of a prospective CMO, it's time to take a closer look at the company's ability to comply with FDA regulations.

First, of course, the sponsor must evaluate the CMO's approach to computer systems compliance. Does the company understand risk-based analysis and how it determines which systems receive more attention? Does the CMO understand the FDA's predicate rules governing software validation and hardware qualification? Has the company developed standard operating procedures for monitoring servers or other projects involving regulated systems? The sponsor should look at project documentation. It's easy to tell if the CMO really follows standard operating procedures or if everyone simply invents their own solutions on the fly. In the latter case, the CMO must rectify the problem to ensure reliable outcomes.

Offshoring Risks and Rewards

© ISTOCKPHOTO.COM

According to recent Congressional testimony, China is home to between 5,000 and 10,000 drug manufacturing firms, many of them contract manufacturing organizations (CMOs) serving U.S. drug companies. By its own admission, the Food and Drug Administration (FDA) currently lacks the resources, market access, and reciprocal agreements with China's regulatory agencies to inspect these factories. Low wages and favorable exchange rates also make China-and to a lesser degree, India-ideal for drug manufacture.

Some outsourcing consultants say that China enjoys a 60% cost advantage in drug manufacturing. But that figure does not take into account the hidden costs and risk of offshoring. Large companies sponsoring big production runs may achieve significant cost savings and may have the management talent to mitigate risks. But small and mid-sized firms would do well to consider the added costs that come with offshore production.

In the United States, commercial contracts and intellectual property are well protected by the courts. Developing nations may have strong laws on the books, but enforcement can be spotty. In China, strong commercial contracts are based on personal relationships. These require country managers on the ground and frequent visits by senior management. Small and mid-sized biopharms that outsource directly to China-based CMOs should consider the considerable cost of travel and the executive time it will take to establish and maintain profitable relationships. Travel time between New York and Shanghai is at least 24 hours each way when you consider both time in the air and time to recuperate. Several trips a year are required. Additionally, language and cultural gaps may make hiring Chinese nationals advisable.

Because regulatory compliance in China lags behind that found in the United States, companies will increasingly be forced to establish their own monitoring and testing programs. In many cases, sponsors will be required to employ an inspector during each production run by their offshore CMO partner.

Small and mid-sized biopharms determined to explore offshore production should investigate United States-based

It is absolutely critical that the sponsor understand that some CMOs are more efficient than others. One may use 15 kilos of an API to manufacture 8 kilos of a drug, while another will need only 10 kilos to accomplish the same result. The sponsor should insist that the CMO provide detailed data of its past and present production efficiency.

Remember, if the FDA audits the CMO's electronic records or its validation and qualification documentation and procedures, and the company fails, all associated product could be considered adulterated, forcing the sponsor to begin clinical trials anew. At the very least, the pre-selection audit process must make sure that the CMO maintains electronic records of the manufacturing controls and batch handling systems and methods that establish a safe drug-producing environment.

After Your Selection

After a sponsor has evaluated its CMO finalists and selected the best one, it's time to address any shortcomings, establish mandatory SLAs, provide for ongoing monitoring and improvement, and negotiate a contract. While this may seem like the least of a sponsor's IT management worries, it is in fact the most critical work any sponsor will accomplish during an outsourcing engagement. Ironclad SLAs are the only way to ensure that a sponsor will get the services it requires (see, "How to Manage CMO Shortcomings," p. 40).

During these negotiations, the sponsor must identify a CMO's specific shortcomings and insist on a remediation plan that the sponsor can monitor on an ongoing basis. For example, the sponsor may have long ago adopted ITIL and/or Six Sigma best practices, only to find that the CMO is deficient in these processes. The sponsor may insist that the CMO adopt the same industry good practices.

The contract should provide for periodic, unannounced audits by the sponsor for the life of the agreement. In addition, the sponsor should insist on timing ongoing audits prior to specific milestones such as regulatory submissions or product launches. The SLA should state that the CMO must maintain adequate supplies to ensure that clinical trial or other batch quantities are sufficient to meet demand. In addition, the SLA should establish detailed descriptions of the type and frequency of electronic data to be provided to the sponsor. Failure to establish this up front in the SLA often results in sponsors paying extra for electronic reports that should have been part of the basic agreement.

CMOs make their profits by scheduling production runs tightly. Whether manufacturing a batch of 10 kilograms or 10,000 kilograms, the CMO must clean and sterilize its manufacturing line before setting up a production run for another client. This process is expensive and time-consuming. A sponsor that finds itself without enough product to satisfy clinical needs or commercial demand may have to wait weeks and pay severe penalties to schedule another run. It is absolutely critical to understand that some CMOs are more efficient than others. One may use 15 kilos of an API to manufacture 8 kilos of a drug, while another will need only 10 kilos to accomplish the same result. The sponsor should insist that the CMO provide detailed data of its past and present production efficiency. API production and efficiency SLAs must be guaranteed by the CMO.

IT Best Practices Required

Outsourcing came late to biopharmaceuticals because drug makers were not certain they could guarantee regulatory compliance of operations they did not directly control. Managing extended supply chains in compliance with FDA's regulatory regime remains a challenge. But the economic benefits of outsourcing are undeniable.

How to Manage CMO Shortcomings

Outsourcing is attractive because it reduces costs and slashes personnel. Moreover, since most owner-operated batch manufacturing plants sit idle for 50% of the time, according to industry studies, outsourcing greatly improves returns on invested capital. The outsourcer pays only for the production time and equipment it uses, while the other customers share the full cost of the plant, personnel, and equipment.

Unfortunately, outsourcing production to a contract manufacturing organization (CMO) does not immunize the sponsor against Food and Drug Administration (FDA) regulations and enforcement. Managing a virtual enterprise is difficult at best, and it requires exceptional control using information technology (IT) best practices, because problems with FDA inspectors may surface when a CMO fails to document required processes such as software validation, hardware and network qualification, and maintenance of electronic records and audit trails per 21 CFR Part 11.

Outsourcing sponsors can protect themselves from substandard IT practices by insisting that their CMOs agree to service level agreements (SLAs) that require minimum levels of performance. FDA will not transfer responsibility for compliance to the CMO, but the sponsor can enforce the agreement with penalties if the CMO fails to deliver.

Keys to encouraging high performance with SLAs include the following:

• Insist that your CMO is aware of and has implemented good practices such as IT infrastructure library and Six Sigma;

• Enforce continuous improvement via periodic audits by your quality assurance/quality control staff;

• Establish standard operating procedures (SOPs) for validation and qualification for all CMO IT projects;

• Train CMO IT staff in risk-based analysis that meets your own standards and definitions;

• Create a remediation timeline for all substandard systems and practices;

• Establish documentation SOPs so that the CMO's IT staff is interchangeable;

• Agree on financial penalties to be shared by the CMO for failed audits by the FDA, other regulatory bodies, or independent audit bodies; and

• Develop a set of comprehensive change management procedures for the CMO.

As supply chains get longer and go offshore, sponsors will have to continuously improve upon IT practices to manage the virtual enterprise. Outsourcing sponsors cannot depend on CMOs to establish IT good practices. In the end, the sponsor is responsible for every act of noncompliance that occurs in its supply chain. It is the sponsor's drug and patient-and the FDA will hold the sponsor accountable. The recent tragedies surrounding heparin made with adulterated ingredients from China ensure the growth of FDA oversight of outsourcing. The only way to meet FDA's strict expectations of outsourced relationships is for biopharmaceutical sponsors to establish good practices of their own. �

Pelley is director of quality and Postle is vice president of life science enterprise practice at the Court Square Group. Reach them at pelley@courtsquaregroup.com and postle@courtsquaregroup.com or at (413) 746-0054.

Advanced Process Control

By Mojgan Moshgbar, PhD, and Steve Hammond
Get It Down PAT
Advanced process control can improve pharmaceutical manufacturing
Get It Down PAT
© DREAMSTIME.COM

The pharmaceutical industry continues to evolve. New International Conference on Harmonisation (ICH; Q8, Q9, and Q10) guidelines provide science- and risk-based approaches to development, risk management, and quality systems that can empower manufacturers to manage continuous improvement and technical innovation throughout the product life cycle. New manufacturing technologies are being introduced, and increasingly sophisticated approaches to process analytical technology (PAT) are being developed.

Key factors drive this evolution, including industry competition, cost containment, and quality and regulatory considerations. New technology initiatives provide capabilities with potential to enhance productivity by improving the process—capability, control, and robustness—reducing cycle times, and improving consistency, while at the same time ensuring compliance.

New technologies and advances in PAT-enabled process control can, in combination with strategic use of design space and implementation of quality risk management, enable quality by design (QbD) to achieve a desired state of manufacturing. These new manufacturing paradigms can provide opportunities for significant regulatory flexibility, including real-time release (RTR) and post-approval continuous improvement.
The Value-Added Role of PAT

Historically, the pharmaceutical industry has applied PAT to further process understanding. Over time, as the technology has grown and become more sophisticated, the potential for PAT-based applications to add value has increased.

Pharmaceutical companies such as Pfizer have applied PAT to:

* enable process understanding;
* identify and remove sources of variability;
* monitor processes on line to provide real-time data for information purposes; and
* determine process endpoints in chemical reactions, drying, and so on, to allow better timing of the off-line release samples.

All of these approaches add value by furthering process understanding, but, like the sophistication of the technology and implementation of PAT, value rises exponentially as the list descends. The jump from process understanding to determining process endpoints is certainly significant, but new PAT-based control strategies can overcome traditional process control limitations, bring new definition to process consistency and efficiency, and extend process capabilities beyond what is possible with conventional control approaches.

Newly developed and emergent PAT-based approaches are pushing the boundaries of process design and redefining strategies for process control. As the reliability and performance of PAT systems improve, its potential to serve an integral role in pharmaceutical processes will increase. Within this context, PAT is increasingly used to replace off-line final product tests with at-line or on-line PAT-based release tests, to provide the basis for process control strategy, and to enable continuous quality verification (CQV) and RTR.
Figure 1. Process Control Strategy: The Current State (Click to Enlarge)
All images courtesy of Pfizer Global Manufacturing
Figure 1. Process Control Strategy: The Current State (Click to Enlarge)
Traditional Process Control

Process control has traditionally been achieved through tight control of key process parameters at predetermined set points or ranges. The premise for this approach is the assumed or established relationship between process inputs—raw materials properties, process parameters such as temperature or pH, and so on—and critical and key product attributes such as process outputs. This control strategy, however, does not allow for mid-course correction to account for variation in starting materials or process upsets, nor does it allow flexibility within or between production runs to utilize the design space concept.

The set points for critical process parameters are commonly determined during development—typically in a design of experiment—and the process validated using a three-batch validation approach. Yet, in reality, after validation, the process will be subject to different sources of input variation that would be transferred directly to process outputs. Variability in quality attributes, therefore, is virtually inevitable.

Reducing common cause variation in such a traditionally controlled process can require significant effort. To ensure acceptable process capability, over-processing—over-drying, for example—is typically utilized, commonly resulting in increased costs and cycle times. Thus, control strategies that are based on fixed process parameters can result in higher variability of product attributes.
Figure 2. The Impact of Raw Materials: Separation of NIR Trajectories (Click to Enlarge)
Figure 2. The Impact of Raw Materials: Separation of NIR Trajectories (Click to Enlarge)
Advanced Process Control

No single definition of advanced process control (APC) exists in the literature, but the phrase as it is currently used describes mathematically advanced control algorithms that use predictive, adaptive, and optimization techniques to control multi-input, multi-output processes. A new concept in the pharmaceutical industry, APC is a mature technology that is commonly used in all other industrial sectors to improve quality, consistency, and process efficiency. Pfizer distinguishes APC as control strategies that utilize PAT, process models, or other techniques to manipulate process parameters (process inputs, Xs) within any required constraints, in order to actively control one or more active pharmaceutical ingredient (API) or drug product attributes (process outputs, Ys) at a set point or within a tight range.

PAT-based APC can overcome traditional control strategy limitations to enable capabilities that include:

* real-time monitoring of process outputs as well as process inputs;
* real-time prediction of process endpoints and product attributes to determine any potential deviation from desired range; and,
* calculation of the change needed in process inputs at each APC sample time to minimize the potential deviation in process outputs.

In a continuous process, outputs are typically controlled to reach and maintain steady state; thus, simpler steady state control strategies (time invariant) are sufficient. By comparison, in a batch process, outputs will typically follow a time-variant trajectory, necessitating more involved control strategies, including the use of multivariate controller models.

One PAT application uses near infrared (NIR) technology to monitor multivariate high shear wet granulation (HSWG) batch trajectory. (HSWG is a particle size enlargement process for maximizing powder handling and uniformity and minimizing dust hazards. Powder is mixed as the powder bed is simultaneously sprayed with binder solution.)

Figure 2 illustrates how the variation in raw materials and processing conditions results in separation of NIR trajectories. The performance of the granules during downstream processing can be predicted by the NIR trajectory, with the trajectory for Excipient 1 at X1 = M representing an optimal batch.

Figure 3, illustrates the use of PAT-based APC to control the real-time trajectory of a batch with non-optimal raw materials; it would follow the trajectory of the optimal “golden batch” that results in the required granule properties.
Figure 3. Control of NIR Trajectory: Drug Product Formulation Advanced Process Control (Click to Enlarge)
Figure 3. Control of NIR Trajectory: Drug Product Formulation Advanced Process Control (Click to Enlarge)
Benefits of PAT-Based APC

APC offers a new and promising paradigm for efficiency in pharmaceutical processes while providing tangible quality and business benefits. It can enable higher process capability, maintaining the process attributes close to specification. Improved quality, lower common cause variations, greater product consistency, improved yield, and cycle time improvements are among the benefits these advanced strategies can provide.

Currently, at Pfizer, there are a number of APC applications, involving both drug product and API processes and focused on large-scale manufacturing, at different stages of development. They utilize a range of APC technologies, including various latent variable model-based batch control strategies, non-linear APC with hybrid process models, soft sensing, and multi-loop optimal control using quality and business cost functions. These applications aim to address some of the most technically challenging problems in our industry with regard to chemical and physical attributes of API and drug products.

CQV (ASTM E2537) is a science-based approach to process validation in which manufacturing process performance is continuously monitored, evaluated, and adjusted as necessary (see Figure 4, below, right). This science-based approach verifies that a process is capable of producing and will consistently produce product meeting its predetermined critical quality attributes. PAT-based CQV provides real-time quality assurance; the desired quality attributes are ensured through continuous assessment during manufacture. Data from each batch are used to validate the process.

RTR is the ability to evaluate and ensure acceptable quality of in process and/or final product based on process data, which include a valid combination of material attributes and process control (ICH Q8 [R1]). RTR requires a high level of process knowledge to understand the impact of process parameters and raw materials on product critical quality attributes, as well as to identify and control the sources of variations. Process data composed of larger numbers of continuously gathered samples as compared to samples taken at the beginning, middle, and end of a process serve as the basis for real-time release of the final product. Measurement may be indirect, e.g., blend uniformity using on-line NIR coupled with unit dose weight variation control versus United States Pharmacopeia testing of tablets. Enhanced process understanding, larger number of process samples, and effective process control provide an increased level of quality assurance with RTR.
Figure 4. Elements of Continuous Quality Verification (Click to Enlarge)
Figure 4. Elements of Continuous Quality Verification (Click to Enlarge)
Next Generation Manufacturing Initiative

Pfizer defines QbD as designing and developing formulations and manufacturing processes to ensure predefined product quality and understanding and controlling formulation and manufacturing process variables that affect the quality of a product.

Pfizer’s Next Generation Manufacturing Initiative extends QbD principles to achieve continuous manufacturing of a drug product. Design space and PAT are combined to demonstrate and achieve process control; CQV is used for process validation. The combination of these approaches enables the RTR of a continuously manufactured product.

Continuous manufacturing is the combination of multiple unit operations in a manufacturing process into a single integrated system. In a continuous process designed based on QbD principles, sources of variation are defined and controlled, and end product variation is minimized by controlling the process within the design space.

The benefits of continuous manufacturing include:

* a smaller equipment footprint;
* minimized scale up and innovative, science-based regulatory approaches that can reduce regulatory filings for scale changes;
* reduction in engineering lead time;
* shortened time to market;
* lower inventory; and
* safer processes.

Reduced waste and increased containment capabilities are among the environmental, health, and safety benefits afforded by continuous management.

Driven by cycle time reduction and capacity enhancement, as well as reduced off-line analytical testing and minimized change over time, resulting efficiencies may be captured in capital and operating cost reductions. The integration of on-line PAT tools supports the development of more advanced process control strategies and CQV that can lead to real-time release.

Among the PAT applications enabling CQV and RTR is a PAT-based system for real-time monitoring of blend potency and uniformity in new continuous drug product manufacturing processes. The application utilizes sophisticated PAT signal conditioning, advanced chemometrics, and multivariate calibration models to accurately measure the potency of flowing blends. Apart from enabling CQV and RTR, PAT supports process control strategy by providing fast, real-time measurement that is used by the supervisory control system to maintain the process and the product within the required limits.

As the reliability and performance of PAT systems improve, their potential to serve an integral role in pharmaceutical processes will increase. Newly developed and emergent PAT-based approaches are pushing the boundaries of process understanding at Pfizer and redefining process control strategies in batch and continuous processes. Within this context, PAT is increasingly used to replace off-line final product tests with at-line or on-line PAT-based release tests, to enable CQV and RTR, and to provide the basis for advanced process control. n

Dr. Moshgbar is senior manager and team leader in innovative processes, and Hammond is director and team leader in the process analytical support group, for Pfizer Global Manufacturing. Reach Dr. Moshgbar at mojgan.moshgbar@pfizer.com or (269) 833-9665 and Hammond at steve.hammond@pfizer.com or (973) 355-5619.


Get 100 Qualified Leads. Download pdf.

Spectroscopy Now
Click here to find out more!
Click here to find out more!

The Challenge of Outsourcing Manufacturing

By Christopher Abbott and Christian Phillips

ISTOCKPHOTO.COM What to look for in a CMO for process scale-up Selecting a contract manufacturing organization (CMO) to outsource parenteral product manufacturing to should rely on much more than the CMO's ability to fill a product aseptically. Consideration should be given to whether or not the CMO has had experience in the formulation of similar products. Drug companies often experience unexpected challenges with formulation scale-up. As Richard Pariza, PhD, chief scientific officer of Cedarbug Pharmaceuticals (Grafton, Wis.), said in an interview, "Only the simplest of reactions behave the same way on both small and large scale. There are always some engineering problems. However, an experienced medicinal chemist will have seen enough processes go to larger scale to design the chemistry, from the start, to avoid many issues. There is no substitute for the wisdom that comes from experience."1 It is imperative that a drug company select a CMO with experience and expertise in its specific type of formulation. This article is written from a process engineer's perspective. By looking at some of the challenges and solutions involved in process scale-up and tech transfer, it provides some criteria to consider when selecting a CMO. LIPOSOMES Challenge: Targeted drug delivery via lipid encapsulation is a cutting edge technology that employs the use of a phospholipid to suspend hydrophobic and hydrophilic components in an aqueous phase. Typically, the active pharmaceutical ingredient (API) is a soluble small molecule encapsulated within the lipid bilayer. This novel drug delivery mechanism is gaining momentum and popularity, as more drug companies come up with unique strategies for phospholipid formulation and targeted delivery. Successful liposomal formulation is often marked by a specified particle size or distribution, percent encapsulated drug, and/or the adequate removal of the organic phase. Upon scale-up, the latter can be particularly challenging from an operational perspective due in part to equipment size and footprint. Other critical parameters that should be considered include the infusion rate of the organic phase containing the lipid, the aqueous phase, the pH, and the temperature. There are a variety of different techniques that will achieve success. For example, a process may utilize an extrusion method versus homogenization via high shear mixing or high-pressure nozzle homogenization versus direct-stream infusion. These techniques dictate particle size and percent drug encapsulation. Pictured in the inset is a laboratory tabletop version of a microfluidizer (homogenizer). The other image is of a commercial size microfluidizer from Microfluidics International Corporation. This unit is in the 7000 series and is the largest model made by the manufacturer. It is used in the emulsion process. IMAGES COURTESY OF HYALURO CONTRACT MANUFACTURING. Solution: Process risk can be mitigated by investing in development and by performing engineering and demonstration runs. Liposomal formulation often involves so many critical process parameters with narrow tolerances to control, a situation requiring meticulous and gradual scale-up efforts, that it is wise to seek out industry expertise from experienced sources to devise an informed plan for development and scale-up. Because liposomal drug delivery technology is complex and relatively new, it may take some time to find someone experienced in the nuances of formulation. By the time a lipid formulation technology is transferred to a CMO, it has typically undergone extensive development at the bench prior to scaling up to pilot or clinical scale. Because these types of formulation do not scale in a linear manner, sponsors often need to invest in feasibility studies "at scale" to render worthwhile data. PROTEINS, OTHER LARGE MOLECULES Challenge: Therapeutic proteins are often used in enzyme replacement therapies for the correction of a clinical deficiency in the body and as antibodies to target harmful cells and substrate such as cancer cells and cytokines. Peptides are another macromolecule manufactured for a number of indications, including adjunct therapies. Large molecules, as these are called, can present unique challenges in their parenteral formulation and often include an array of excipients, including buffer, surfactant, and various forms of glycol, aimed at supporting the stable configuration of the molecule. Large molecules are sensitive to degradation and may have limited stability; strict instructions concerning proper storage conditions need to be followed. Unprocessed bulk API often arrives frozen at the CMO and might require freezing as a final product to maintain stability. This necessity can present further challenges. Both freezing and thawing must be done carefully to avoid gradients from forming in the bulk; these may lead to protein rearrangement and subsequent inactivation. Disposable biobags are often used to store bulk. Biobag polymer layers become susceptible to tearing upon freezingor thawing; precautions must be taken to prevent this problem. Often, there is a "time of exposure limitation" with respect to how long bulk product can be left at ambient room temperature. The window of exposure can be as little as 12 hours, a limit that can present complex logistical challenges for executing formulation and fill. Care must also be taken to limit solution surface-to-volume ratios, because protein aggregation and rearrangement may occur at the liquid-gas interface. Finally, large molecules, in the form of recombinant proteins, take on a different, more challenging registration process compared to other drugs, because they are governed by unique regulations in both the United States and Europe. Choosing a CMO that has the infrastructure and experience to support the compliant manufacture of biologics is critical. Solution: Proteins, while complex in their formulation, work well with disposable technology due to their overall compatibility with disposable product contact layers, including low-density polyethylene, ethylene vinyl acetate, silicone, polypropylene, and others. Disposable technology is more amenable to the tech transfer process when working with a CMO that operates a multi-product facility. Additionally, disposable technology allows for greater flexibility and even scalability because the disposable industry has widely recognized the need to provide scaleable formulation and compounding solutions. By the time a lipid formulation technology is transferred to a CMO, it has typically undergone extensive development at the bench prior to scaling up to pilot or clinical scale. Because these types of formulation do not scale in a linear manner, sponsors will often need to invest in feasibility studies "at scale" to render worthwhile data. Mixing systems like the disposable LevTech line offered by Sartorius-Stedim are good examples of this type of disposable technology. These systems allow for effective mixing, bulk sampling, and aseptic fluid transfer capability, while offering tight control over mixing profile to accommodate proteins sensitive to shear. An additional advantage is the ability to minimize the frothing or foaming at the liquid-gas interface that can lead to protein rearrangement and subsequent deactivation of API. While the compatibility profiles of large molecule formulations are often excellent for the use of disposables, a careful assessment must still be rendered. For example, there are simple, small-scale filterability studies performed with 45 mm filter membrane discs of various material types, including polyvinylidene fluoride, polyethersulfone, nylon, and cellulose acetate. This approach can help you to identify a filter membrane type that supports adequate flux and exhibits excellent chemical compatibility. A manufacturing operator adjusts the Ph while product is being transferred from one vessel to another using a peristalitic pump. Before making the decision to scale up to a disposable capsule that utilizes this membrane type, take into consideration the construction materials in the capsule housing. Taking this precaution can help you to avoid a situation in which the filtration, at scale, is working fine until the polycarbonate housing begins to crack due to incompatibility with the excipients being used. Special storage units are available for keeping and transporting frozen bulk in biobags. While these range in price and complexity, some of the higher end units offer not only protection for the frozen bulk but also uniform freezing and thawing to minimize gradient formation. For frozen bulk that arrives in bottles or carboys, you can establish procedures that provide a gentle swirl to the bulk periodically throughout the thawing event. Limiting exposure of bulk product to ambient requires careful consideration on the planning and scheduling end of production. But in order to be in a position to make informed decisions on orchestrating formulation/fill activities, engineers and operators must first gain experience with the formulation/fill process using bulk product-often provided as a placebo-via demonstration and engineering runs. CMOs have different facilities, with unique personnel and material flows, that need to be carefully evaluated by the engineer and project manager in order to meet expectations for exposure limits. Contract a CMO with proven experience manufacturing biologics in order to avoid not just technical problems but also the unique current good manufacturing and regulatory pitfalls that are associated with biologic product registration. A competent CMO will provide direction to the client with regard to quality control sampling plans-for example, the need for bulk sterility testing post sterile filtration but pre-fill-that support biologics manufacturing. This guidance will demonstrate awareness of the identification, qualification, and receipt of raw materials that go into that formulation. Because the lower concentrations required in large molecule formulations result in volume dosages that tend to be larger than that of small molecule products, give careful consideration to endotoxin specifications for raw materials entering the formulation. This can prove challenging, because some of the common excipients included in large molecule formulations include carbohydrate solutions such as sucrose that can be difficult to source with a highly limited endotoxin specification. Again, a CMO experienced in these areas can best address such challenges, either by sourcing the right raw material or by utilizing technology and experience-including ionic filtration, ultrafiltration, and other, more complex, techniques-for polishing endotoxin from the raw material. SMALL MOLECULES Challenge: While large molecules are often used to target the surface of cells, small molecules typically use diffusion to enter the cell for therapeutic effect. While the small molecule bulk product formulation itself-excipient and API recipe-is often less complex than that of large molecules, the compounding and formulation may still involve a rather complex process, depending upon the characteristics of the small molecule API. Molecules that are oxygen sensitive, photosensitive, or hygroscopic need special handling processes and procedures that protect against product degradation. Furthermore, small molecules, which require formulations comprised of organic solvents, are not compatible with disposable technology. As a result, stainless steel or glass vessels and potentially expensive Teflon tubing may be required. Fixed and/or reusable vessels typically add complexity and scope to a project. Procedures for equipment cleaning, sterilization, and changeover must be developed and adhered to, as do specific validations for equipment cleaning and sterilization. Training and instructions for vessel assembly may also add complexity to the batch record. Solution: Experienced process engineers can custom design vessels to protect product integrity, optimize yield, and simplify handling and processing. Glove boxes can be used under dry inert gas sparge (i.e., nitrogen or argon) to weigh out hygroscopic powders. Head-space and the liquid bulk can be sparged with inert gas to displace dissolved oxygen. Processing areas can be lit with special lighting to prevent photodegradation. Glass vessels can be used for small-to medium-scale processes when disposable technology is not compatible. For larger scale processes involving more than 30 liters, however, stainless steel reusable vessels are often necessary. It's vital to choose a CMO with experience in stainless steel skid assembly, steam in place (SIP), clean in place (CIP), and general associated process flows. In addition, the facility itself needs a footprint that is compatible for large vessel mobility, transport, and cleaning. The use of stainless reusable vessels necessitates a great deal of validation; the CMO's ability to provide comprehensive validation protocols that fully support the process is critical. MICRONIZED EMULSIONS Challenge: A system of two immiscible liquids in which droplets of one are dispersed in the other is considered an emulsion. Typical emulsion processes require the use of either high-shear mixers with a rotor stator head or a high-pressure nozzle or valve homogenizer that uses extremely high pressure to force solutions through a small orifice to form the emulsion. The formulation of an emulsion is complex; it consists of formulating two phases, one aqueous and the other organic. The use of stainless reusable vessels necessitates a great deal of validation; the CMO's ability to provide comprehensive validation protocols that fully support the process is critical. Often, large homogenizers and tanks are used to create the emulsion after the two immiscible solutions are introduced within a single vessel. Large equipment may come with significant floor space and facility infrastructure requirements. Large valve or nozzle type microfluidizer processors, such as those manufactured by Microfluidics Corp. (Newtown, Mass.), require the plumbing and installation of cooling loops and have significant electrical requirements. As tanks and equipment sizes increase, CIP and SIP systems are the only practical means to sterilize and clean them. Batch sizes can be limited by the receiving vessel's capacity. Solution: A large amount of effort is required, depending on process scale, to coordinate both the installation of the microfluidizer processor and the operation of the equipment. Because the scale-up of microfluidization processes leads to large footprints, a CMO should be assessed for its ability to meet utility requirements. The necessity of build-outs and expansion will mean that the facility and organization should be evaluated for flexibility and expertise in this area. The CMO will need strong facilities engineering and process engineering groups for managing both the facility and process design scope of the project. Large-scale micronization processes may demand operation in a grade A environment. A cleanroom may need to be fully dedicated to the process, or the CMO may opt to design a space around the equipment. Often, in development phases, homogenization processes are defined by a number of serial passes through the system to produce the desired particle size. On a larger scale, discrete passes are not easily obtained, because transferring the large volume bulk becomes cumbersome, requires multiple large holding vessels, and introduces areas of concern around potentially compromising the aseptic integrity of the system. Continuous pass homogenization is therefore opted for in large scale; because continuous pass is not as efficient as discrete, however, significantly longer processing times may result. HANDS-ON EXPERIENCE When it comes to formulation/fill/finish operations that will successfully carry clients from early clinical through commercial registration, it is imperative to select a CMO with a wealth of hands-on experience in the particular formulation being processed. Because of the complex and dynamic nature of these formulations and the way they are processed, an abundance of experience in just one particular formulation or process will likely not be adequate for taking on new formulations, even those that seem similar. This article has attempted to elucidate a number of tech transfer challenges, particularly with respect to scale-up, in the areas of lipid, small and large molecule, and micronized formulations. The CMO that is a multi-product facility will need effective procedures and a design for handling other formulations to protect personnel and control cross-contamination. Viscous product processing requires a CMO with specialized formulation and filling equipment and experience with both Newtonian and non-Newtonian fluids. Proper equipment and experience assist in overcoming technical challenges such as sterile filtration, in maximizing product recovery, and in maintaining adequate filling speed. There are dozens of nuances and unique challenges that accompany almost any new project, especially when it comes to scale-up. A CMO that can anticipate these types of challenges can make the difference between timely success or missed timelines and failure. A CMO should be assessed on its facility/infrastructure, its experience within a particular category of formulation, and an academic knowledge of the formulation chemistry and critical attributes. It is also critical to look at experience from an operational and quality standpoint. A CMO with both can overcome technical hurdles and ensure that all decisions are made in the context of current good manufacturing processes and overall compliance. This will support client objectives for successful manufacturing and product registration of their therapeutic product. ¦ Abbott is supervisor and Phillips is director of process engineering at Hyaluron Contract Manufacturing. Reach them at (781) 270-7900, ext. 165, or cabbott@hyaluron.com or at (781) 270-7900, ext. 181, or cphillips@hyaluron.com. REFERENCES 1. Roth G. Engineering and quality: process development. A CMO talks about scale-up issues [interview]. Contract Pharma. March 2007. Available at: http://www.contractpharma.com/articles/2007/03/engineering-quality. Accessed September 26, 2008.

Pain-Free System Transition

By Dimitrios Saponas and Demetre Kolokotronis

© PHOTOGRAPHER: MYLIGHTSCAPES | AGENCY: DREAMSTIME.COM

IT Rides Off Into the Sunset

Retiring an information technology system doesn't have to be a headache

Retirement is the last phase in an information technology (IT) system'S life cycle (SLC) and must be performed to satisfy U.S. Food and Drug Administration (FDA) requirements for record retention and system validation. Retirement, or decommissioning, means the cessation of the use of hardware (HW) and/or software (SW), along with the migration or conversion of data. Unfortunately, this last phase, if it is performed at all, is usually done in a haphazard manner. When this occurs, both availability of records and compliance with regulatory requirements may be compromised.

Why and when a system is retired is up to the business owner and is, in essence, a business risk decision. Some considerations that influence decommissioning may stem from the following:

• Are the systems in question no longer supported by the vendor, or is the vendor no longer in business?

• Is upgrading more profitable in the long run? Will it improve throughput and output and/or reduce required resources and cost?

• Will the change further ensure part 11 requirements or enhance the quality of the end product or, at the very least, maintain acceptable quality and safety?

What the competition is doing can also fuel the need for change, potentially resulting in a loss of market share. The question is, therefore, whether or not to change. Retiring, modernizing, improving, or moving SW or HW used by IT or in production that is affected by good manufacturing, clinical, and laboratory practices (GxPs) must be a decision that respects the predicate rule requirements. There can be no compromise allowed. This discussion will focus on the IT requirements as they pertain to the decommissioning of electronic records, SW, and HW, including firmware (a computer program embedded in a HW device) and their SLC documentation.

All About the Records

All businesses in the medical device, pharmaceutical, cosmetic, and food industries must account for any activity related to the manufacture, production, and distribution of human and animal products destined for implant or consumption. The requirement to document all related activity is governed by the predicate rules, namely, the Code of Federal Regulations (CFR). The CFR is law in the United States and applies to all GxP-impacted products destined for the U.S. market.

These records are the hard copy and/or electronic copies that are generated and, at a minimum, are required to demonstrate adherence to the stipulated requirements set forth by the FDA. They also demonstrate that due diligence is exercised by all parties at every stage of the product'S manufacture through its receipt by the patient or end user. A company'S internal policy should clearly identify any records that are regulated, how they are to be maintained, and the length of time they are to be retained. The paper printouts or electronic records must meet all of the requirements of the applicable predicate rules.

The FDA may take a company'S business practices and decisions into account when determining whether predicate regulation and internal policy meet Part 11 requirements and record retention periods as specified in agency regulations (see "Part 11 and Record Retention," p. 49). The predicate rules imply that all SW-generated data and records must be readily available to an inspector. If the SW can no longer operate and be validated on current computer HW and operating systems (OS), it is a generally accepted-and perhaps easier-practice to maintain a validated version of the last computer system the given application SW and OS were run on. Effective and representative data migration and data conversion are expensive, time consuming, and, in most cases, not necessary if the requirements can be met by maintaining a working validated version of the original computer equipment, as discussed below.

The consideration of the validated state brings into question the validity of acquired GxP data. In order to authenticate and document the validated status of equipment at the time of its retirement, it must be proven beyond any doubt that the equipment was in its intended working state. This proof, in turn, provides confidence that at the time the equipment was removed from service, it was decommissioned with sound scientific evidence and that all results generated with the unit in question for the release of product were trustworthy and reliable. If this evidence is not gathered, all output from the unit since its last documented validated state can be called into question. As appropriate, final calibration and functional verification should be conducted at the time of system retirement.

When the time to decommission does arrive, a company must (at a minimum) follow an outline or plan similar to the one presented below. The detailed outline or plan must be explained in a company'S retirement policy.

Retirement Policy and Plan

It is good practice to have a company or corporate policy that addresses system retirements. Such a policy can serve as a road map to ensure that FDA requirements for the conservation and retrieval of GxP data are met. Moreover, the policy must consider activities for the retention of system documentation. These must demonstrate, at the time of retirement, the preservation of the validated state and proper operation with respect to intended use throughout the life of the system.

The system retirement policy must require that a retirement plan be generated for each GxP system. The policy must indicate how the retirement process is initiated-through a controlled notice or a change request/control process-as well as how the process is to be closed upon completion of all required activities.

The FDA may take a company'S business practices and decisions into account when determining whether predicate regulation and internal policy meet Part 11 requirements and record retention periods as specified in agency regulations.

The retirement plan (RP) is a written document that must be approved by all departments affected or be required in the company'S retirement policy. The effort that will be put into the activities outlined in the RP must be based on a documented risk assessment. At a minimum, the following topics should be considered when generating a RP.

1. Introduction

This section briefly explains the use of the system at the time of retirement and includes a brief description of its interfaces with other systems, SW, and components. A reason/rationale for conducting the system retirement should also be provided here.

2. Roles and Responsibilities

This section outlines the tasks that each participating department (system owner, IT, engineering, quality assurance, and so on) has to accomplish for the decommissioning/retirement of the system. The roles and responsibilities are determined on a per project basis and depend on a multitude of factors such as the type and complexity of the system and other specifics discussed below.

3. Record Requirement

This section documents what electronic records and data stored in the system must be retained in order to satisfy predicate rule record retention requirements. Working together, the system'S owners, users, and quality assurance (QA) can identify the records used to support GxP activities. The vendor and/or IT can identify the database tables containing the data required to generate the reports. As a note of caution, it is often easier to keep the original data. Another important consideration is the retention of associated audit trail data and meta-data-data about the data-as well as other information. As stated previously, record retention is required by various predicate rules and must indicate how long the records will be maintained before they are allowed to be destroyed.

4. Documentation

Each system has a certain amount of documentation that comes with it. This includes vendor user manuals, system standard operating procedures (SOPs), commissioning and validation protocols and reports, specifications, logbooks, and more. Documentation found with the equipment files, such as calibration and maintenance reports, should also be considered. A list of all system documentation must be established, with an associated action to be performed for each. SOPs will have to be retired and removed from circulation, and all other documentation will be archived for a period at least equivalent to the duration of the records requirement. Documentation can be archived either on or off site but must be readily available for any auditor who asks to review it. The documentation can only be disposed of at the end of the required retention period; the manner in which this will be accomplished must be clearly explained.

5. Data Migration and Archiving

Data migration is needed when data from the retired system must reside on the new computer systems and the database management system (DBMS) or if its configuration is incompatible with the new systems. If the DBMS or the records file is compatible, simply copying and moving the data from one system to the other may be sufficient.

Data migration can be complex and requires planning with IT personnel knowledgeable in the specific tasks needed to ensure complete and correct migration of the data mapping. Data mapping reads data from the system to be retired, formats it to the new system'S formats and requirements, and then writes it to the new system'S databases. The complexity of the mapping activity depends on many factors, including types of DBMSs, ability to export data in a readable format, the need for conversion, and the number of tables or files to migrate. Data migration is usually executed using customized programs or scripts that automatically transfer the data.

Data archiving involves copying files and data to a storage device for long-term storage, with the intention of deleting or removing the data from the source media after it has been copied. Data archiving is usually performed when the data has record retention requirements but will not be required for day-to-day activities and/or is not likely to be asked for by an auditor/inspector. Data and records that are migrated or archived must maintain their integrity, so it is important to include all date-time stamped audit trails and associated metadata and to retain electronic signatures where applicable.

The company should also consider the need to conduct parallel operations for a period of time, utilizing the retiring system for official results while the new system is checked for areas of disparity. This period of parallel operation can be used for operator training and confirmation of validation.

The RP for the system must include a statement explaining whether or not the data from the old system will be retained. If data is retained, explain the strategy-migration or archival-and include the data that will be retained or not retained, again with a rationale. Migration activities must include testing after loading to verify that the data was accurately mapped, that it supports functionality in the new system, and that it is complete. The company should also consider the need to conduct parallel operations for a period of time, utilizing the retiring system for official results while the new system is checked for areas of disparity. This period of parallel operation can be used for operator training and confirmation of validation.

If the data migration is complex, a system-specific data migration plan should be developed, but the details of such a plan are beyond the scope of this discussion. If data is archived, the RP must define the media on which the data will be archived (e.g., CD, DVD, or hard drive), the storage format (e.g., backup or image), and the final destination where the data will be securely maintained. Testing should be performed on the archived data to ensure that it is retrievable. Making two archive copies of the data should be considered in case one cannot be read. Finally, the destruction of data from the retired system must be performed when the archiving and restoration processes have been determined to be satisfactory.

6. Hardware and Software

If the data is successfully migrated, the system'S HW and SW may not require retention, and their disposition must be covered in the RP. Some companies keep all HW and SW, while others dispose of them or recycle and reuse the HW for another application. It is the company'S responsibility to assess what needs to be done with retired HW and SW and to document this decision in the RP.

PART 11 AND RECORD RETENTION

The FDA'S regulations set out Part 11 requirements and record retention periods. The agency may take a company'S business practices and decisions into account when determining whether the following policies have been met:

• Data handling, storage, and retrieval-cGLP 21 CFR 58.81 (b)(10);

• Retention of records-cGLP 21 CFR 58.195 (b)(1)(2)(3) for five years after submission of the application or two years after approval;

• General record requirements-cGMP 21 CFR 211.180 for at least one year after the expiration date of the batch or, in the case of certain over-the-counter drug products lacking expiration dating because they meet the criteria for exemption under 211.137, three years after distribution of the batch;

• Retention of records-cGCP 21 CFR 312.57 (c) A sponsor shall retain the records and reports required by this part for two years after a marketing application is approved for the drug; or, if an application is not approved for the drug, until two years after shipment and delivery of the drug for investigational use is discontinued and the FDA has been so notified; and

• General record requirements-cGMP 21 CFR 820.180 (b) record retention period. All records required by this part shall be retained for a period of time equivalent to the design and expected life of the device, but in no case less than two years from the date of release for commercial distribution by the manufacturer.

If the data is archived, the company may need to retain SW to process the data. Likewise, specific HW to run the application SW and other supporting programs may be required. The considerations that determine the SW and HW items to be retained are many and should be addressed by the retirement team, with input from IT and system vendors (see "Should it Stay or Should it Go?," below).

The easiest method to retain HW and SW is to keep the whole system intact. This is rarely used today because it takes a great deal of storage space and does not allow recycling of HW systems. Another approach is to retain all installation media and re-install the system when electronic data must be retrieved.

It is good practice to have a company or corporate policy to address system retirements. Having such a policy in place can serve as a road map to ensure that all required steps and considerations are taken to make certain that FDA requirements for the conservation and retrieval of GxP data are accomplished.

SHOULD IT STAY OR SHOULD IT GO?

When retiring one information technology system and migrating to another, a company needs to decide which software and hardware must be retained to meet future needs. Below is a non-exhaustive list of questions that may help determine what to retain and what to dispose of:

• Is an operating system (OS) or are other program requirements needed to run the application software (SW)?

• Can the system run with a more recent OS version?

• Will the required original OS version be available to run the application when and if data retrieval is required?

• Does the system have a configuration that must be retained?

• Are there special hardware (HW) requirements?

• Can the OS or the application SW run with other or more recent HW?

• What is the possibility of accurately retrieving data if the original HW is not available?

The major drawback with this approach is that system configuration is not retained. The third and most common approach is to perform an image of the system'S hard drive. An image is a clone of the disk that allows for an exact copy of the hard drive or partition. The OS and all SW, programs, and system configurations can be saved with an image. Additionally, data can also be imaged and used as backup.

7. Other Considerations

A system that is retired may have operational or documentation ramifications on other systems. The RP should list, where possible, the other systems that are affected, what is impacted, and how it will be addressed. All training material related to the system and the retired SOPs should be archived. Additionally, employee training modules should be reviewed to remove any scheduling that is no longer relevant. Typically regulated computer system activities such as backups, periodic reviews, and performance monitoring should be stopped. The site computer-system inventory should be updated, and any vendor support agreements should be reviewed to determine if they are still applicable.

Users should be notified that a system is to be removed from active support and use. Depending on the environment, this notification can take many forms, including a change control or a memorandum that is distributed before any activities begin. The RP should explain how the process is initiated and what information it must include. This notification must ensure that all users of the system are aware that it will be retired and must be distributed with enough lead time to make any arrangements that are necessary. Once all activities in the RP are completed, another notification memo should be distributed to inform all parties that the system'S retirement is complete.

The final notification memo should be signed by a QA representative and should confirm that all required activities outlined in the RP have been accomplished. The RP and final notification should be stored as per company documentation practices. �

Saponas is a process and utilities engineer and Kolokotronis is director of compliance and quality systems, both at Validapro Biosciences, Inc. in Laval, Québec, Canada. Reach them at d.saponas@validapro.com or (450) 668-1144, or demetre@validapro.com or (917) 941-0177.

RESOURCES

1. International Society for Pharmaceutical Engineering (ISPE). GAMP 4. Guide for Validation of Automated Systems. Tampa, Fla.; 2001:59.

2. Institute of Electrical and Electronics Engineers, Inc. (IEEE). Std 1074-1997. Standard for Developing Software Life Cycle Processes. Washington, D.C.: Institute of Electrical and Electronics Engineers, Inc.;1997:57-58.

3. U.S. Food and Drug Administration. Title 21 Code of Federal Regulations (21 CFR PART 11). Electronic Records; Electronic Signatures. Washington, D.C.: U.S. Food and Drug Administration; 2000.

4. U.S. Food and Drug Administration. Title 21 Code of Federal Regulations (21 CFR PART 58). 5. Good Laboratory Practice for Nonclinical Laboratory Studies. Washington, D.C.: U.S. Food and Drug Administration; 2001.

5. U.S. Food and Drug Administration. Title 21 Code of Federal Regulations (21 CFR PART 211). Current Good Manufacturing Practice for Finished Pharmaceuticals. Washington, D.C.: U.S. Food and Drug Administration; 2001.

6. U.S. Food and Drug Administration. Title 21 Code of Federal Regulations (21 CFR PART 312). Investigational New Drug Application. Washington, D.C.: U.S. Food and Drug Administration; 2007.

7. U.S. Food and Drug Administration. Title 21 Code of Federal Regulations (21 CFR PART 820). Quality System Regulation. Washington, D.C.: U.S. Food and Drug Administration; 2007.