In Search of Consistency

March 19, 2003
Drugmakers and the FDA see greatest payback potential in new manufacturing technologies designed to reduce process variability

With the possible exception of the semiconductor industry, the pharmaceutical business is the most scientifically innovative industry at the R&D stage. By contrast, pharmaceutical manufacturing relies on processes and information technologies that are decades--in some cases centuries--old.

Because of perceived high validation and regulatory overhead associated with new technologies, particularly process operations, pharmaceutical manufacturers have been skeptical of innovation. And, until recently, high margins have shielded drugmakers from the consequences of not adopting modern manufacturing methods.

But the U.S. Food and Drug Administration (FDA) is encouraging the application of new process analytical technologies, which happen to offer great payback potential.

At Biogen's manufacturing facility in Research Triangle Park (RTP), N.C., new technology deployments are allowed when they provide significant productivity improvements. Everywhere else, the firm prefers proven methods that minimize the risk of failure and delays.

"Process variability affects your success rate for new drugs and usually means longer approval times," says Nicolas Barthelemy, vice president and general manager of the RTP site. "The advantages of reducing variability and batch failure rate dwarf any potential savings from acquiring a new piece of equipment or re-inventing a process."

And while the high quality and safety of the pharmaceutical industry's end products remain unquestioned, 20 to 30 percent of unit costs typically are tied up in non-value-adding activities needed to ensure that quality, adds Velumani Pillai, global technology leader for automation and plant solutions, Pharmacia (Kalamazoo, Mich.).

"On average, the pharmaceutical industry operates at one-sigma quality levels," he says, in reference to the popular Six-Sigma standard for limiting process variability to within acceptable standards (3.4 times out of a million tries).

"An electronics manufacturer who operated at one-sigma would be out of business in no time," Pillai explains.

In all fairness to drugmakers, changing a validated process is no easy undertaking. Biotechnology companies in particular, which spend years optimizing fermentations that produce labile, high-value products, are understandably wary of changing anything. Even at innovative biotech firms, new technologies may be impractical or simply unavailable.

Regulatory Excuses?

Innovation woes often are blamed on the FDA and its labyrinthine drug development regulations (formulated, more often than not, with industry's blessing or at its request). Recognizing this, the agency has begun to encourage innovation through its Process Analytical Technology Committee (PAT) and Advisory Committee for Pharmaceutical Science (ACPS) with the hope of creating for manufacturers a win-win environment for innovation.

"There is a significant room for improvement in pharmaceutical manufacturing operations," says Ajaz Hussain, deputy director, Office of Pharmaceutical Science at the FDA's Center for Drug Evaluation and Research (CDER). "With today's emphasis on drug discovery, development and marketing, manufacturing has become a stepchild. Companies believe their opportunity for improving processes is limited, so they tend to stick with what they are already doing."

According to Hussain, industry has adopted a "hear no evil, see no evil" approach to manufacturing innovation, meaning they are inclined either not to use new technologies or, if they use them, not to tell the agency.

"There's the perception that regulatory uncertainty is holding back new technologies, and that NDAs [new drug applications] would be delayed or denied because FDA reviewers may not understand the underlying science," Hussain says. "We hope to address that through our PAT initiative."

By encouraging the application of process analytical technologies, the FDA hopes to quell the oft-stated sentiment that it is to blame for technologic stagnation. With the help of university-based pharmaceutical manufacturing experts, the FDA is training teams on what to look for in new process technologies and to encourage their use. Eventually the FDA plans to issue guidances based on its experiences.

Bringing the Lab on Line

Techniques for improving first-time quality are currently available. The aim is to assure that companies "manufacture right the first time"--a phrase used by the FDA's PAT subcommittee. Rapid, real-time, in-line analysis is one way to achieve this.

Example 1: Where research mass spectrometry (MS) analyzes complex chemical mixtures through hybrid separation/mass measurement techniques, process MS keeps it simple, concentrating on low-molecular weight off-gases.

"QA/QC [quality assurance/quality control] and process engineers don't understand how you can have an in-process MS without liquid or gas chromatography up front," says Tony Slapikas, product manager for process MS at Ametek (Paoli, Pa.). Process MS derives the mass balance in a biochemical reactor from the composition of the feed air going in and the air coming out of the process. According to Slapikas, off-gases provide simpler, more accurate indications of what is going on in the process than sampling and analyzing offline.

In small-molecule manufacture, process MS can measure moisture or solvent levels in batches by monitoring masses corresponding to the solvent of interest. Slapikas estimates that MS can reduce drying time by at least 10 percent by eliminating the need to pause the operation, remove a sample and analyze it offline. For a 20-hour drying process used a hundred times a year, process MS pays for itself through higher productivity, lower electricity use and less wear and tear on equipment.

Example 2: Traditional Bragg-Brentano X-ray diffractometers require precise positioning of source, sample and detector. These limitations, along with large size, make them unsuitable for in-line measurements. Low-density materials and samples with curved surfaces (e.g. pills) or surfaces displaced by as little as one-tenth of a millimeter are generally inaccessible to X-ray analysis.

The solution--parallel beam x-ray diffractometry (PBXD)--uses optics that capture a large angle from the x-ray source and shape it into a parallel beam, making sample alignment non-critical. PBXD also allows less powerful, less expensive radiation sources that are about the size of a flashlight.

In-line X-ray measurements have tremendous utility for QA/QC engineers interested in a drug's polymorph composition in real-time, according to David Gibson, president of X-Ray Optical Systems (East Greenbush, N.Y.). PBXD's low power--5 to 50 watts versus 2 kilowatts to 18 kilowatts for conventional diffractometers--reduces radiation exposure to negligible levels, according to Mr. Gibson.

Track Those Records

Bringing traditionally laboratory-bound analytical techniques into real time is only one step toward helping to reduce the cost of quality in pharmaceutical manufacturing. Streamlined, automated tracking of electronic batch records from raw materials through distribution also can avoid non-value-adding efforts and mistakes according to Charlie Norz, marketing engineer for process control at Milwaukee-based Rockwell Automation.

"It's not unusual to have four batch records attached to a product lot: one each for raw materials, solution prep, forming and packaging," Norz explains. "How do you manage all that information? With electronic batch records, one record contains it all,and that reduces paperwork and quarantine time."


New Separations Technologies Poised to Improve Process Throughput

First the bad news: Large bioprocesses are indeed getting larger, less agile in design and operation.

The good news is that most biotech companies are still small enough, and flexible enough, to try out new process technologies,especially in the realm of separations.

"There's not a lot of new technology in fermentation, either upstream or downstream, to warrant changing an existing process," notes Tim Gregory, senior director of process sciences at Genentech (South San Francisco, Caif.). But new separation technologies that interest Gregory are tangential flow filtration (TFF) and a new wrinkle: high-performance tangential flow filtration (HPTFF), which reduces downstream operations by combining chromatography and filtration steps. Genentech, which invented HPTFF, is jointly developing it with Millipore (Bedford, Mass.) but has no immediate plans to adopt the technology.

High-performance, ambient-temperature resins also are enabling shorter, more convenient separations.

"Ten years ago, you were looking at a linear velocity in chromatography systems of 30 centimeters per hour. Today, 500 centimeters per hour is not unusual," says William Hensler, CDI Engineering Solutions (Phoenix). Purification suites designed around high-performance resins are less expensive to build and operate than cold rooms. In addition, workers can leave their winter coats at home.

Expanded bed chromatography is another innovation that allows relatively "dirty" process streams to be loaded directly onto chromatography columns, circumventing extensive harvesting operations such as centrifugation, diafiltration and ultrafiltration. Expanded bed, which works only for secreted proteins, requires a simple depth filtration step to remove cells. Expanded bed runs at high velocities, expanding the space between chromatography beads. Consequently, columns are less prone to clogging and dirtier feedstocks can be applied.

Conventional chromatography media have about 33 percent interstitial space, whereas in expanded bed systems it may reach 50 to 70 percent. Expanded bed resins do not necessarily cost more, but the columns are more difficult to run, do not scale up as easily and are generally less forgiving.

Hensler believes that transgenic technologies will one day make traditional upstream bioprocessing--fermentation and cell culture--obsolete for many products. Eventually, "pennies-per-pound" proteins will affect downstream processing as well. Although a few companies continue to chip away at the technical and regulatory hurdles of transgenics, big-name bioprocessors are staying home from this particular revolution.