March 2009

Cover Story

If you build it in ...

Building quality into pharma, biotech innovates, regulates, stimulates


  • Economy hits pharma, biotech manufacturing processes.
  • Process technologies, designing in quality innovative processes.
  • Communication between process development, manufacturing key.
By Ellen Fussell Policastro

The dwindling economy is not discriminating when it comes to pharmaceutical manufacturers and biotech firms. That is the prediction from a January ARC Advisory Group study, in which Janice Abel said pharmaceutical manufacturers will stay under the microscope. Industry will need to improve innovation to meet new initiatives in efficiency.


Process analytical technologies and quality by design are some of the key words pharmaceutical and biotech mavens use today as part of their daily vocabulary. "The fundamental point is not so much using the process analytical technology (PAT)," said Roddy Martin, vice president of industry strategy at AMR Research in Boston, Mass., "but the improvement of process performance as the basis of continuous improvement in the manufacturing process. The improvement of process performance lies in going back to the quality-by-design process rather than just looking at PAT as a capability; it's more fundamental than that."

Biotech moving faster

"If you look at biotech, they believe themselves to be much further along. And they are, strictly because real-time measurement in their processes has been online for a long time. It's the nature of their business," said Justin Neway, chief science officer at Aegis in Lafayette Col. "They have to do that because micro-organisms need to be coddled. In pharmaceuticals, that's less the case."

Applications for PAT

While companies such as Merck and Pfizer are seeing the benefits of PAT, it hasn't reached popularity in pharmaceuticals as much as in specialty chemicals, food, beverage and petrochemical said Alex Habib, an automation engineering consultant in Old Bridge, N.J.

From his experience in fermentation, Habib believes some of the most important technologies seeing use with PAT involve high performance liquid chromatographs (HPLCs) and mass spectrometers, as well as near infrared for product drawing and blending.

Processes that would benefit from quality by design practices include fermentation and reaction, especially batch, Habib said. "Most of the pharmaceutical industry is batch in nature. So batch reaction centrifuge sees use in batch processes," he said. "It's almost like a washing machine where final drying spins off liquid before you actually put clothes in the drier. Centrifuge is a way to help with drying. Instead of heating it to dry, you first get rid of most of the liquid."

Jose C. Menezes, Ph.D., professor at Technical University of Lisbon, agrees biotech is forging ahead more quickly because bioprocesses are "strongly influenced by the quality of inocula and raw-materials used," he said in his paper, "The Use of PAT in Biomanufacturing," co-authored by Licinia O. Rodrigues, Ph.D., senior PAT specialist at 4Tune Engineering Ltd., and Arthur Voogd, Ph.D., product manager in pharma, at Yokogawa Europe BV.

"The number of parameters measured in a bioprocess are often only those related to what can be controlled (e.g., cell density, pH, extracellular concentration of substrates)," Menezes said. "Sampling frequency is conservative since those parameters are manipulated at a low frequency." As such, process development is time consuming and complex mainly because a large number of experiments must generate enough process knowledge out of possible combinations of variables involved (strain or cell lines used, raw-material conditioning and media formulation, inocula scale-up, and fermentation operation). "Finally, any PAT strategy developed for a dynamic process whose trajectory strongly depends on the starting conditions (e.g., inocula quality or previous development steps) must take into account information from previous processing stages if a precise end-point is to be achieved consistently," he said.

Some possible and obvious strategies to cope with those difficulties would involve: monitoring intensification (more parameters measured more frequently in all relevant steps); tools to pull together the information of several fermentation batches (building and navigating a design space); and systems engineering tools to analyze several runs of the entire process (thus describing the interactions between process components), Menezes said. "We have successfully developed and applied such strategies in the biomanufacturing of several anti-infectives, and we believe in its general value to other bioprocesses, including those involving animal cells for monoclonal antibodies production."

Process technology means to an end

During conferences and while visiting users, Neway has seen more and more a deeper understanding that PAT is a means to an end. "The end we're really all striving for is to have quality designed into our processes so variability is under better control internally," he said.

"Users don't care whether you understand your process or not. They care that you make safe, efficacious product that's available and affordable. If you need to understand your process more today to do that, fine. If you need to employ online instrumentation to obtain real-time control, fine. That's what you'll do. There was this evolution that we all need to be in PAT; then we all need to be in process understanding. It's all the same - a means to an end. The end is quality by design."

Real-time quality by design

Talecris, a manufacturer of blood fractions, factor 8, and other blood-clotting factors (proteins, complex molecules) uses highly purified water in their process, Neway said. "They aren't measuring proteins; they're measuring water because when you inject water, you can't introduce contaminants," he said. Before introducing this online instrument, they had to take samples and put them through lab tests that took days, Neway said. This left them open to potential contamination in samples. "Once they had validated the inline measurement of water quality, they eliminated all that time and had closed-loop control on water quality." The beauty of this process was its straightforward simplicity of measuring water quality using known, simple technology. "It wasn't a measurement of the product itself; it was a measurement they had introduced into the stream of one of the components of the processing steps. It was the water they were using in the process," Neway said.

Another instance of simple, straightforward process to build in quality is Pfizer's focus on the risk analysis portion of process development to try to understand what portions of the process would need the most development and attention, "because they were likely to be the largest sources variability and the outcomes," Neway said. "Again it didn't require fancy online instrumentation but a methodical examination of risks. It turned out that product didn't do well in the clinic, but they still published their work in risk analysis and the kind of thinking that's required to move toward real time and building quality by design."

Neway also referred to MerckSerono's approach, achieving lower variability in manufacturing by doing process development in a way that designs those processes to go into manufacturing better. "They've set up an on-demand data access and collaborative analytics environment that allows them to have their process development group move more quickly through development steps as well as collaborate with the manufacturing group," he said.

Regulatory freedom

The PAT initiative provides pharmaceutical manufacturers a certain level of freedom, said Dan Matlis, president Axendia, a life sciences strategic advisory company in Yardley, Pa. "In the traditional process, license holders have provided FDA detailed information regarding process equipment and parameters to be used for the manufacturing of a new drug product in the chemistry, manufacturing and controls (CMC) section of an application," he said. "Under this paradigm, changes in the manufacturing process for an approved drug may require review by FDA, a process that can be lengthy and tedious. As a result, manufacturers would often forgo changes to the process and get by with less-than-optimal processes.

"PAT requires a much deeper and mechanistic understanding of the product and the process," Matlis said. "With PAT, manufacturers submit process schematics as well as a characterization of critical product and process parameters. Rather than locking the specific steps and equipment to be used in manufacturing, PAT give companies freedom regarding the how of manufacturing, while focusing on ensuring tight processes that maximize product quality and minimizing costs. In a well-designed, PAT-based process, documenting compliance becomes a natural by-product of the process, rather than its primary focus," he said.

Communication key

Quite a few changes manufacturers submit to the FDA for approval could result from poor communication between the development team and manufacturing team, Neway said. "Merck Serono and GSK recognized quality by design starts in process development. You shouldn't be designing the process in manufacturing," he said. "What you do in process development must be transferred to manufacturing." This includes not just data, but knowledge and experience as well.

What's learned in manufacturing has to come back to people in process development and help them to develop better processes next time. This has just started to happen in the last few years. Quality by design starts in process development and requires process development, along with quality assurance and manufacturing groups, to function as one single team. These companies are putting together a collaborative data access and analytics environment that spans process development and manufacturing.

Process analytical technology really involves communication and behaviors. So in effect, it is not so much a technology as it is a lifestyle, Neway said.

PAT is not about protocols and instruments but a way of life with three main goals:

  1. Make material for the clinical trials that comes from the process development efforts. Early and late stage trials
  2. Develop the process to see use in manufacturing
  3. Put together the scientific basis the FDA uses to approve the process (chemistry, manufacturing and controls (CMC), a section of the regulatory submission a company makes for their next new drug).

When people who have experience in biopharmaceuticals have to scramble to get data together from spreadsheets, it can slow submissions and getting approval. "If from day one you say 'my data matters, and everyone's working from the same data,' then putting the CMC together will be easier, take less time, and you've ended up with better science," Neway said. "But it's an ongoing operation. It has to be continuous, not episodic or crisis-based."


Ellen Fussell Policastro is the associate editor of InTech .  Her e-mail is efussell

Economic crisis looms

The financial crisis has only added to industry turbulence in pharmaceuticals and biotechnology, and investors are extremely risk-averse. Investment capital is scarce, resulting in a credit squeeze for cash-poor biotech companies, said a January ARC Advisory Group study. At the same time, big pharma has seen stock values plummet. The market for initial public offerings has all but disappeared.

The global market for pharmaceutical sales is predicted to be slightly lower than it was this year, with increases of one to two percent in 2009 to around $300 billion in the U.S.

Overall, the global market could achieve growth of between four to five percent, mostly as a result of expansion from emerging nations such as China, South Korea, India, Turkey, Russia, and Latin America. The pharmaceutical companies are targeting these potential high-growth regions. However, since the U.S. market accounts for the bulk of worldwide pharmaceutical sales and the majority of the industry profits, 2009 will not be a banner year for the industry, even with significant efficiency-increasing initiatives.

Near-infrared as process tool

By Jose C. Menezes

The potential of process analytical technology (PAT) in biomanufacturing has not reached its pinnacle of effectiveness mainly because of  insufficient use of intrinsically multiparametric monitoring tools, such as near infrared (NIR); disregard, or ineffective use of available process information, such as data on historical batches; data collected at different process stages, such as inocula development and raw-material lots; and lack of a process and plant-wide perspective for the proposed PAT strategy.

NIR spectroscopy can be a process monitoring and process supervision technique in the context of biomanufacturing. An industrial-scale bioprocess can illustrate the use of on-line in-situ NIR monitoring by means of an immersion transflectance NIR probe.

NIR calibration development must be performed carefully and should incorporate steps to obtain a properly validated model, which exhibits long-term robustness and is independent of process scale. In general, NIR can be as accurate as the reference methods employed and at least as precise provided sufficient spectral selectivity and sensitivity exists.

NIR can also see use as a direct technique for fast process monitoring and process supervision, thus enabling you to follow the trajectory of a process. This alternative to the indirect use of NIR through laborious calibration development with direct reference methods has seen little research. Since NIR is sensitive to both chemical and physical properties, the analysis of whole samples enables relevant process information to be captured and thus generates better process state estimates than by simply looking at defined process parameters one at a time.