Rising demand from aging, health-conscious baby-boomers promises to keep pharmaceutical manufacturing plants humming for many years to come. But what will these facilities look like? Genetics and biochemistry promise new generations of small-molecule and protein products, gene therapies, viral vectors, personalized medicine, and ever more-complex delivery and dosage forms. The more esoteric of these products may still be years from the marketplace, but the goal of safer, targeted, more effective medicines is already at hand, and with it, the goal of less-wasteful drug manufacturing.
Today's pharmaceutical engineers are borrowing ideas and technology from other process industries to assure that manufacturing keeps up with science. Facility designs routinely specify process integration, barriers or isolators to replace highly-classified space, and automation and controls to reduce human intervention. Like Japan's fabled automotive plants, larger pharmaceutical facilities today already operate, almost, in "lights out" mode, says Jeffrey Sarvis, director for facilities integration at Fluor Corp. (Greenville, S.C.).
If technical innovation continues at its present pace, tomorrow's processing suites may routinely be built in modular fashion hundreds or thousands of miles from their final locations, as some are already. Plants could be staffed by robots that see in the dark and never take a coffee break. Materials and products will be tracked continuously, automatically, flawlessly as they move through the supply chain. For biotech manufacturing, small, ultra-streamlined processing facilities might be built adjacent to rows of corn or soybeans, or above abandoned copper mines converted to farms, and fed by crude proteins provided by transgenic plants and animals.
Biotech to Challenge Manufacturing
As its importance increases, bioprocessing is nudging pharmaceutical manufacturing into the future. So far, the U.S. Food and Drug Administration (FDA) has approved 155 biotechnology drugs and vaccines, according to the Biotechnology Industry Organization. More than 70% of approvals occurred since 1998, and more than 370 biotech products targeting more than 200 diseases are now in clinical trials.
Monoclonal antibodies, biotech's fastest-growing product segment, also present its major manufacturing challenge. Administered in high doses, often for extended periods, the drugs require a constant supply flow, but the mammalian cell culture used to make the drugs is prohibitively expensive.
Nonetheless, significant strides have been made. During the past two decades, cultured cell densities have risen from about one million cells to between eight and 10 million cells per milliliter, according to Florian Wurm, Ph.D., chief scientist at Excellgene (Churm, Switzerland). At the same time, specific productivity has risen from one picogram of protein per cell per day to fifty picograms.
However, even this 500-fold increase in volumetric productivity can't keep pace with demand for cell culture capacity, and bioprocessors are scurrying for alternatives to large bioreactors. Inserting multiple protein-producing genes into cells promises to improve productivity, as does engineering cheaper and easier-to-grow bacterial or yeast cells.
Transgenic animals and plants will probably give biotech, and pharmaceutical manufacturing, its next boost. Proteins from transgenic plants and animals are untested in the marketplace and even more foreign to regulators. But their potential economic benefits are breathtaking: fermenters replaced by corn fields, hundreds of millions of dollars saved in upstream facilities and unlimited capacity. As for the naysayers, common wisdom circa 1975 held that FDA would never approve products made from hybridomas (engineered human cancer cells). Today, hybridomas are the workhorses of the monoclonals industry.
Transgenics' true test will not be whether corn or cows can churn out medicines--that has already been established. Companies can make protein for pennies per gram upstream, notes Cardinal Health's Brandon Price, but downstream purification still makes up 30% to 60% of total manufacturing costs in a traditional process. "If transgenics can't show at least a three- to five-fold reduction in overall manufacturing costs, they will be much less attractive, especially when you consider regulatory uncertainty," he says.
Whether change can occur quickly enough to resolve today's cost concerns is anyone's guess. Nevertheless, cost-containment, process integration, and a better appreciation for product segregation are manufacturing trends that are sure to be part of any facility design twenty years hence.
Classified Floor Space Trending Down
Because of the high ongoing costs of cleanroom operation, designers of new facilities are turning to isolation technology to protects workers from toxic products and sensitive materials from contamination, according to David DiProspero, senior director for business development at Sear Brown (Melville, N.Y.). Fluor's Sarvis agrees. When barrier/isolator systems are employed, he says, the building envelope takes on much lower priority with respect to product quality. "The goal," he says, "is to divorce product quality from facility construction by enclosing processes, rather than installing a lot of classified space."
However, deploying barrier technology guarantees that the equipment itself will be more expensive to acquire and install, as will facilities. The savings, according to experts, is in long-term lower operating costs associated with having with less classified space.
Enclosed manufacturing processes are not cost-effective at all scales, however. In biotech, Sarvis believes disposable technology makes sense up to about the thousand-liter scale, or about 5,000 liters through multiple containers. However, with mid-sized biopharm processes, he says, it's easier to achieve substantial efficiencies of scale with automated enclosed systems based on stainless tanks and hard-piped clean-in-place/steam-in-place systems.
In other words, the ratio of cleanroom area to total facility area should go down dramatically as bioprocessing scale goes up. Certain early steps, at small scale, will probably still be done in cleanrooms, Sarvis notes. But as the scale increases, closed processing becomes more economical, he says, "liberating design from considerations of human intervention and the need for environmental quality control."
Integration by Design
Process integration, for example, in milling operations or in fluid handling and separations, is another trend. Traditionally, companies do a wet milling, transfer into a fluid bed dryer, and when the material is ready, they dry mill, DiProspero says. Newer processes are combining operations in line: the granulator connects directly to mill, which is interfaced to a fluid bed, which feeds directly into the dry mill. On the other hand, integrated systems present equipment stacking and facility height challenges.
"The big value variable is how systems are tied together," explains Fluor's Sarvis. "Equipment and square footage being constant, factors such as gaining adjacency advantages and stacking processes and operations is where the ultimate project cost variability comes in."
The integration trend also holds for facility design. Traditionally, plant design specifications are drawn for equipment and building, then pass on to the commissioning group. After they're done and all the light bulbs seem to work, the validation group takes over. "The three steps of design qualification, operational qualification, and process qualification are treated almost independently of each other," DiProspero observes.
More recently, to shorten timelines the separate documents are integrated: Design serves as the framework for commissioning, which in turn can be used for site validation. "Instead of throwing it over the fence, commissioning guys can start working before the design is even completed," DiProspero says.
Most facility design today is based on a fanatical preoccupation with segregation, which, according to Sarvis is a result of an ill-fated attempt to eliminate human error. "Primary segregation objectives should address the science or chemistry of the process: can you process live and dead cells side by side? How about two different processes for two different products, or upstream from downstream operations?"
Companies need to define their primary segregation objectives, understand which ones are absolutely essential, then interpret secondary segregation efforts that address human error independently, he says.
"If you design to meet the primary objectives successfully you have a lot of latitude on how you address the secondary objectives." The goal, he adds, is to focus on the three or four primary segregation objectives, then address the others procedurally---or engineer them out of the process and, subsequently, the design.
Busting the Blockbuster?
Preparing for tomorrow's complex, individualized medicines will require more science, more automation and more quality by design. And the time to start implementing these changes is now, says Ajaz Hussain, Ph.D., deputy director of pharmaceutical science for the FDA's Center for Drug Evaluation and Research.
Whether the move to genomics and personalized medicine will be a centralizing or decentralizing force in pharmaceutical manufacturing remains to be seen. Companies simultaneously value the cost savings of mega-facilities and the flexibility that smaller facilities bring---and that personalized medicine will demand. "The reality for production managers is that they will have to accommodate both cost-consciousness and rapid change," says Adam Bianchi, chief operating officer of the market research firm, Cutting Edge Information (Durham, N.C.).
But the smaller profits of the personalized medicines will take some getting used to. Consider Genentech's Herceptin antibody for breast cancer, with 2002 net sales of $385 million. Herceptin treats a subset of breast cancer patients who over-express the Her2/neu gene---as many as 30% of all women with breast cancer. Herceptin represents everything noble about the personalized medicine idea, but begs the question: with drug development costs as high as they are, how many $385 million products can big pharma support?
Simultaneously, drugmakers must become more efficient in their production methods, particularly in quality control and assurance. "Quality control will be one of the battlegrounds of the ongoing healthcare cost control war," adds Cutting Edge's Bianchi.
But is the pharmaceutical industry up to the task of producing better products at lower cost? Pharmaceuticals find themselves in a difficult regulatory and technologic transitional phase, says Greg Page, Ph.D., life science practice leader at Deloitte (Jericho, N.Y.). "Industry is still trying to figure out what FDA wants, and the agency itself may not be sure beyond its position that it wants to see a greater effort towards innovation and new technologies."
Historically, regulations hinder rather than stimulate technical innovation. Based on recent FDA directives, particularly those that stress risk management and process analytics, a "thousand points" of innovative light may yet shine on pharmaceutical manufacturing. At this point, though, many quality, regulatory and legal experts in the pharmaceutical industry make judgments based on interpretations of regulations, instead of science or procedure, Page notes. Way back when, if a batch release required taking a hundred samples during filling, and some plates were dropped or a site was missed, it was treated as a deviation, not as a quality issue. By the late 1990s, perfectly good product was deemed unreleasable because only 98 samples had been collected. "Manufacturing had become more like a legal proceeding," he says.
No surprise, then, that our crystal ball doesn't see facilities becoming any less expensive to build in the future, although they may be less expensive, long term, to run. They'll also allow a wider range of products to be made, including some that were unthinkable a few years ago. " These are drugs that in the past," notes Sear Brown's DiProspero, "you simply couldn't make because of limited control of environment, product, and people."
Technology Promises to Help Contain Costs
At some point, drugmakers must address the high price of their products, and the outrageous costs and timelines associated with developing, launching and manufacturing them. The typical drug currently takes eight to ten years to bring to market, at a cost of $1.2 billion, according to Ajaz Hussain, Ph.D., deputy director of pharmaceutical science for the U.S. Food and Drug Administration's Center for Drug Evaluation and Research.
Drug discovery and development, which devour enormous resources, are an obvious place for pharmaceutical companies to begin their reformation. Concurrently, the manufacturing side must clean house---weeding out waste, purging their culture of quasi-regulatory fetishes, and learning from other advanced process industries.
Manufacturing professionals have their work cut out for them. Currently, only about five percent of what companies do adds value, defined as what the customer will pay for, says Darren Dolcemascolo, senior partner at EMS Consulting (Carlsbad, Calif.).
Compounding the waste is the tendency for companies to over-interpret FDA regulations, down to the color of ink used in the hand-written notes on batch records. FDA-phobia locks companies into equipment, protocols and operations, down to the tiniest detail, according to Richard Chua, executive vice president at the Juran Institute. "FDA regulations are no substitute for thinking," he says.
The move to individualized medicine and small-molecule therapies will require exotic delivery systems and dosage forms, and more sophisticated manufacturing. To meet this challenge, pharmaceutical manufacturing must evolve and become more efficient, particularly in documentation, quality-by design, and reducing cycle times, says FDA's Hussain.
Companies are too focused on documentation, he says. "Filling in, signing, and checking documents for accuracy, which causes long delays, often due to deviations that have to be investigated." Today's pharmaceutical companies, notes Hussain, rely too heavily on "testing to document quality" rather than "quality by design." He cites as an example powder blending, which is performed to a fixed time, at which point the process is stopped, a sample is tested and the process re-started if necessary.
In-line measurements could report on blend homogeneity to a performance end-point rather than a predetermined time point. "This approach, done right, can provide better management of raw material variability while improving both quality and manufacturing efficiency," Hussain says. Naturally, Hussain hopes that manufacturers will be inspired by FDA's process analytical technology (PAT) initiative to deploy real-time monitoring and other process innovations. Manufacturers also need to shorten batch cycle times, which, according to Hussain, may take as long as three months, of which only a fraction is spent on actual manufacturing.