Issue Date: January 22, 2007
Breaking Down Barriers
IT MAY COME as a surprise to hear that the pharmaceutical industry stands out as having particularly inefficient manufacturing operations—far less efficient than those in other high-technology sectors, such as computers, or even in mature industries, such as soap and paint. That drug manufacturing is wasteful might seem odd to anyone who assumes that the scientific precision inherent in discovering drugs is reflected in elegant chemical processes that carry through to commercial production.
But contract service companies that serve the drug sector, as well as a handful of major drug companies themselves, attest to significant inefficiencies in drug manufacturing. Given the pharmaceutical sector's drop in profitability since the 1990s, they say, the situation is receiving increased attention.
Indeed, new cost-cutting programs across the industry are focused in part on improving manufacturing. Tellingly, three major drugmakers —AstraZeneca, GlaxoSmithKline (GSK), and Pfizer —have joined Britest, a U.K.-based consortium of manufacturers, contract service firms, academic chemical engineers, and equipment suppliers that is studying the problem. All agree that improving manufacturing means more than simply tightening ship at the plant; it means real change in research and business cultures. If the batch-heavy world of commercial-scale drug production reflects anything, they say, it is a lack of communication between laboratory science and plant engineering. Processes conceived by medicinal chemists to synthesize milligrams of pharmaceutical chemicals pass, basically unchanged, into ton-scale production —the realm of continuous process manufacturing in many other industries.
Part of the problem, they say, is a regulatory regime in which making changes becomes more costly and more disruptive the closer a drug candidate moves toward commercial approval. The Food & Drug Administration has, in fact, been a catalyst for change in the pharmaceutical industry with its Process Analytical Technology (PAT) program, a quality regime that expedites process and manufacturing adjustments to plants operating under the agency's current Good Manufacturing Practices guidelines.
Many of the industry's inefficiencies stem from a "batch-manufacturing culture" deeply rooted in the chemistry of pharmaceutical manufacturing. "There is a tradition of using standard, lab-scale batch processing," says Michael Matlosz, project director of Impulse, a Nancy, France-based consortium of manufacturers and equipment suppliers studying alternative process design. "This leads to very high inventories, very large amounts of side products, large inventories of solvents, and inefficient use of energy."
AS A RESULT, Matlosz says, the drug industry manufactures at a capital utilization rate of about 10-15%, compared with 95% in petrochemicals and other industries employing continuous processes. "This is because you're spending most of your time storing your reactants some place, putting them into stirred tanks, and taking them out of stirred tanks," Matlosz says. "It is a major issue in fine and specialty chemicals, but particularly big in pharmaceuticals."
Impulse is developing "process intensification" strategies aimed at deploying microcomponent reactors, thin-film contactors, heat exchangers, and other technologies to boost plant efficiency. These strategies, Matlosz says, may lead to fundamental change in manufacturing: a shift from batch to continuous processing.
Britest, the U.K.-based consortium, takes a more conceptual cut at the problem, developing analytical and procedural guidelines for process development intended to improve coordination among discovery chemistry, process development, and engineering, according to Sue Fleet, chief executive officer. The consortium's goal is to break down a debilitating cultural barrier in the industry.
"Companies have different cultures, but by and large, people work in groups," she says. The Britest tools, she says, are designed to bring these groups together to understand the process as well as the business imperatives. "Part of what we are doing is developing better ways of sharing knowledge we already have," Fleet says. "Some is about identifying areas where we have a fundamental lack of understanding."
Elaine Martin, an industrial statistician at the University of Newcastle, in England, points to a need for better management of the data-rich environment of chemical production. "Can we get information out of the data, turn it into knowledge, and use that in scale-up? Can we keep learning from experience with other products that have gone through the life cycle and transfer what we've learned through statistical methods to new products?"
FDA's PAT initiative is also focused on statistical analysis, and the program has been a key motivator among drugmakers, according to Martin.
Significantly, FDA is prodding drug companies to share information once assumed to be proprietary, she says. "Companies are becoming open to discussions of common problems such as new ways of doing crystallization. They aren't discussing their entire processes," she says.
PAT, which replaces a lengthy testing regime for validating system changes with an easier protocol for mid-operation adjustments, is encouraging the industry to pursue greater efficiency, says D. Christopher Watts, standards and technology team leader in FDA's Office of Pharmaceutical Science. "We want to allow companies to take advantage of the innovation that has taken place in other industries," he says.
BRITEST ITSELF gained standing as a gathering point for drug companies last October, when Pfizer became the third major drug company to join. Pfizer's presence bolstered the group, which is still largely U.K.-centered, with a significant U.S. presence.
"Pfizer cannot afford to keep doing things the way it has," says Rick McCabe, the firm's senior manager of global manufacturing. "To remain competitive, we must adapt to the changes happening around us. Using the Britest tools and methodology along with our own internal tools and methodology should help Pfizer incorporate some of the transformational changes it needs to do in a shorter time frame." Improved process design, he adds, promises to eliminate inefficiencies across Pfizer's supply chain.
McCabe claims that Pfizer has avoided the "historical pharmaceutical paradigm" of registering a batch-manufacturing process with FDA and living with it for the life of the product. The company, he says, has invested significant resources into developing second- and third-generation chemistries, although this has not eliminated batch processing. "Pfizer, like the rest of the pharmaceutical industry, still lives in a world of stirred tanks, centrifuges, and driers, while the rest of the chemical industry has moved into more efficient ways of carrying out specific unit operations," McCabe says. "I see Britest forcing us to consider the most efficient way of carrying out a unit operation even if it is using a piece of equipment or technology totally foreign to Pfizer."
Ian Shott, CEO of the pharmaceutical services firm Excelsyn, another Britest member, sees the group helping to foster a collaborative atmosphere among drugmakers. Shott, who heads an industrial advisory board on molecular engineering at the University of Newcastle, also sees a traditional lack of communication between chemists and engineers as a major cause for poor design of commercial-scale plants at the early stages of process development.
"In industries like petrochemicals, the engineers get involved quite early on in terms of design," he says. "In pharmaceuticals, chemists have driven the process pretty far down the pipeline, handing it over to engineers very late, who are able only to intercede in a hazard situation as opposed to finding ways to make the process efficient." The pharmaceutical industry, he adds, is highly averse to change once a project nears commercial production.
Streamlining existing manufacturing processes is a major focus for the Newcastle advisory board, Shott says. But the board is more interested in applying basic principles of engineering than in promoting radical shifts to continuous processes or novel production equipment. For example, "people talk about microreactors," he says, "but it doesn't always make sense to move the scale of reactors down. It might be a matter of getting more out of the reactor that exists or reducing the number of reactors." Nor is it a rule that batch processes must be replaced with continuous ones.
"The big point is that most of this isn't new," says Shott. "A lot of people look at it and say this is basically process engineering done as it should be done."
Another important front in the efficiency drive is chemical synthesis improvement. For example, Excelsyn's molecular development business unit has come up with improved transaminase technology to produce the unnatural amino acid tert-l-leucine, a compound used in the synthesis of a variety of drugs.
The technology, Shott says, eliminates one fermentation step in tert-l-leucine production, allowing a manufacturing process that traditionally required two or three reaction vessels to be carried out in one. Excelsyn, he adds, is working with pharmaceutical customers on three products in Phase III clinical trials that use the amino acid as a building block.
Gerard Kwant, principal scientist in DSM's advanced chemical engineering solutions division and a member of the Newcastle advisory board, points out that improvements in plant efficiency often result from prosaic steps, such as improving heat recovery. He notes, for example, that an optimized heat exchanger allowed DSM two years ago to convert a manufacturing process at its plant in Linz, Austria, from a batch operation using a 10-m3 vessel to a continuous process employing a 300-kg enclosure.
"It's all a matter of implementation strategy," Kwant says. "We are doing things step-by-step, gaining experience. As we become more confident in the benefits, we will change to much smaller plants with more continuous processes."
PROCESS DESIGN, however, begins with medicinal chemistry, and successful design hinges on translating that work to large-scale production. Contract service firms play a key role in the transfer, especially for small pharmaceutical and biotech drug companies that must contract out commercial and even clinical-scale production.
Mark Frishberg, president of Seres Laboratories, a California-based custom synthesis lab, says he must frequently challenge client's chemistry on the basis of problems that Seres chemists determine will likely come up once the projects move to large-scale production.
"In discovery, you often have your new Ph.D.s, right out of school, who are going to go with their academic training," he says. Thus, processes often involve low-temperature reactions or the use of chromatographic separations and other techniques not feasible for scale-up. Sometimes, Seres is handed a process that is unsafe at commercial scale, he says.
"We have a project now with a mid-tier pharmaceutical company that I'd be surprised if we don't lose," Frishberg says. "It is a five- or six-step process with about three safety aspects that really shouldn't be scaled up, one in particular that I think is dangerous." Seres presented the company with a price quote for kilogram-scale production that includes what it deems appropriate changes. "They said they don't want anything changed," Frishberg says.
What's needed is better management of the connections between the various drug development stages, says David Zembower, vice president of chemistry at DeCode Chemistry & Biostructures. The company offers contract research and small-scale manufacturing services from early medicinal chemistry through to Phase II production.
"In programs where we manage discovery or medicinal chemistry, it is our philosophy to engage process management teams earlier than has been done historically," Zembower says. Once discovery chemists are down to a handful of potential drug candidates, he says, DeCode brings in its process chemistry group to start looking at possible raw material concerns and safety issues. "We start thinking of those things ahead of time," he says.
Chemistry done in the discovery stage is fundamentally different from downstream development chemistry, Zembower points out. "In discovery, you're trying to understand the structural relationship activities of your compound series with respect to some type of biological response or pharmacokinetics," he says. At this stage, there is little incentive to think deeply about how to manufacture kilogram volumes of the compound.
If there is no coordination of medicinal chemists and process chemists, however, problems can mount. "Then you do have this kind of shock period," he says. "And typically, we don't see too many clients that are all that concerned about what the final validated manufacturing route is going to look like."
Zembower notes that quirky raw materials are a persistent problem. "We've started to see a carry through of some exotic reagents developed back in the 1990s," he says, referring to the heyday of combinatorial chemistry, when a "cottage industry" sprang up, producing small quantities of specialty reactive agents such as amines, carboxylic acids, and sulfonyl chlorides.
"A lot of these unusual fragments are now part of drug candidates that are moving forward," he says. "Some of these items were made in small quantities by companies on the other side of the world who produced 20 g of the material, and that's all they ever made." The sooner process chemists go to work on finding alternatives, he says, the better.
"RIGHT UP FRONT, we look at every step and show where we can reengineer it," he says. "Sometimes we see an eight- to 10-step synthesis, 60% of it involving chromatography. That's absolutely untenable when you're trying to move forward. So what we do is lay out a detailed proposal of what we see as the issues and how we would go about fixing those."
Many process chemistry problems result from time constraints, according to Frishberg and Zembower. But coordinating medicinal chemistry and process development can save time compared with forcing initial processes forward into development, according to John Nuss, vice president for medicinal chemistry at biotech drug firm Exelixis.
"Our medicinal and development chemists get together on a regular basis to talk about the chemistry," he says. "At the end, our development chemists will scale the compound up and look at a few routes on their own because they know the basic chemistry."
Coordination can cut the time it takes to get to an Investigational New Drug filing with FDA from the standard two years down to one year, Nuss says. Exelixis has filed 10 INDs over the past four years, reaching its goal of three per year. "Breaking down the barrier between process and discovery chemistry has primed the pump," Nuss say. "It's really helped us out a lot."
Some of the largest drug companies are similarly priming the pump. "There is a lot going on in big pharma right now," says Paul Richardson, engineering technical director at GSK. "We are looking at all the options: new chemistry, new engineering at the plant, efficiency improvements at existing facilities." Richardson says GSK has been working for several years on improvements to technology, work processes, and research culture under an operational excellence program spearheaded by Ray Scherzer, the company's head of engineering.
The company is also involved in industry consortia. At the Impulse consortium, where GSK is the pharmaceutical industry sector leader, the firm is testing continuous-flow reactors for hydrogenation and investigating processes that integrate pharmaceutical chemical manufacturing and formulation at a single site.
As for Britest, Richardson says it has been most successful at bringing chemists and engineers from several companies together to brainstorm and pool their know-how. "Over the past five years," he says, "companies have realized that to make the necessary changes, they are going to have to work with other people."
Neville Brewis, director of process engineering at AstraZeneca, agrees, adding that Britest has developed a useful regimen of analytical tools. "It provides a more structured approach to process development, bringing in process engineers much earlier than has traditionally been done," he says. "There is none of this over-the-fence business where process chemists hand things over to manufacturing, and manufacturing has to make the best of what they've got."
Brewis says the Britest tools provide a common language that allows engineers and chemists to "tease out" problems early in development. He agrees with Excelsyn's Shott that Britest offers a structured approach to standard techniques for an industry in need of structure. "We hope it will lead us away from the traditional batch reactors that have been the bread and butter of the fine chemicals and pharmaceutical industry for the 30 years I've been working in it, and 20 years before that as well," he says.
CHANGES TO the pharmaceutical industry's manufacturing culture will likely be evolutionary rather than revolutionary. Indeed, most major drug companies have operational improvement programs similar to GSK's, and many are unwilling to concede to anything like institutionalized inefficiencies in production.
Mauricio Futran, vice president for process R&D at Bristol-Myers Squibb, says collaboration between medicinal chemistry and process development has a long history at his company. "We have a close community of practice in chemistry that spans discovery and process development," he says. "And we have formal and informal ties in the business process by which we always look at the chemistry that is coming on a three- to six-month horizon."
Futran says there is no single formula for how and when to bring discovery and development chemists together, whether they be in-house groups, service providers, or outside research partners.
"Most big drug companies are focusing on efficiency. We certainly are," he says. "And things that FDA has been talking about in terms of quality-by-design will naturally lead us to more efficiency in continuing from R&D to manufacturing. This is not a static industry by any stretch of the imagination."
- Chemical & Engineering News
- ISSN 0009-2347
- Copyright © American Chemical Society