Successful pharmaceuticals come of age much as humans do. Discovery is all about wonder and excitement. Painful, plodding development teaches valuable lessons and is then eclipsed by the euphoria of approval and scaleup. Each stage, from the synthesis of milligrams to grams, to the manufacture of hundreds of grams to tons, adds an order of magnitude to the complexity, investment and commitment required to launch a drug. When should you, as manufacturing experts, be involved in a pharmaceutical development program? The answer will depend on your organization's resources and its experience and comfort level with similar manufacturing processes. However, early is typically better. Because of the high attrition rate for Phase I/II medications, for most products it would be a "monumental waste of time" for manufacturing to get involved during Phase I, according to Peter Green, Michigan-based VP of pharmaceutical sciences for Pfizer Inc. (New York City). "If everyone gets off track and starts worrying too early about manufacturing, R&D can actually be delayed," says Eric Musser, chief technical officer at Atlanta-based software supplier Ross Systems. However, most observers agree that manufacturing should get involved in the development process as soon as possible" before drugs enter human trials, or once developers begin discussing dosage forms, says Tarun Patel, Ph.D., vice president of technical operations for Cardinal Health, Somerset, N.J. "Too often, dosage form decisions are made by marketing or scientific personnel who have only a vague notion of what it takes to make the new product on a large scale." When a dosage form is altered, Patel points out, recipe changes often require new machinery and equipment. "A change as simple as a different color tablet could make an existing machine vision system obsolete," he says. Manufacturing helps distinguish between the frivolous and the necessary changes, whether they're easy or difficult. Can manufacturing ever be called in too early? Patel thinks not, provided production's involvement is appropriate for a particular stage of development. "It's never too early to begin the discussion and technology transfer." Late is always more expensive than early, agrees Dr. Michael Breggar, national leader, life sciences regulatory at Deloitte, Philadelphia. "If you start something early, the worst thing that happens is you have too much information. When you're late, you often spend a lot of time catching up." Phase III, The Great Divide Because of regulatory issues and high material requirements, Phase III studies form a natural boundary between development and manufacturing. Counting backwards from the anticipated start of Phase III trials, companies can more or less estimate when they need to start thinking about a robust, transferable process, new capacity and human resources, and vendor/supplier relationships to support a product launch. However, timing is everything. By Phase III, a process is more or less set, notes Warren Levy, Ph.D., CEO of Unigene Laboratories, Fairfield, N.J.. However, if two to three years of development are required, the real dividing line comes much sooner, during early clinical trials or even late in preclinicals. "Unfortunately many companies wait until much later to address critical manufacturing concerns," he says. Chemical, Manufacturing and Controls (CMC) approvals often compete with clinical trials as the rate-limiting development step. Companies would do well not to dismiss CMC as a mere "engineering project," according to Michael Glembourtt, vice president of business development at Nektar Therapeutics, San Carlos, Calif. "Nonstandard products or drugs that require a device or unusual delivery system often entail substantial manufacturing equipment development. In those situations the CMC section may run neck-and neck with clinical studies." Nektar, which is developing inhalable insulin with Pfizer, only works on "unusual" drugs, which explains Glembourtt's partiality to very early manufacturing involvement. Nektar's other formulation/delivery technology, PEGylation, chemically modifies drugs with polyethylene glycol residues to improve the pharmacokinetics and absorption. "We wouldn't be involved if standard formulations, dosage forms, or delivery methods were appropriate," Glembourtt admits. Codevelopment Prevents Shortfalls "There are many horror stories of sponsors struggling to satisfy demand for approved drugs," says Pfizer VP Green. Sometimes products simply sell better than anticipated, but even that situation should not arise, says Green. "Usually material shortfalls occur because the interface between development and manufacturing wasn't properly handled, or the process wasn't robust enough. Even if you've sorted out your manufacturing problems, you've already developed a bad reputation and most likely someone else has grabbed your market." Pfizer emphasizes early manufacturing group involvement through a process it calls "codevelopment." By the end of Phase II, Pfizer has a pretty good idea how its product is doing, which is when manufacturing is summoned. "There are cases where we call them earlier," says Green, "especially if the product involves some kind of technical complexity, for example a device-based drug or novel delivery system. But if we're working on a tablet with no special chemistry or formulation, we may go a bit beyond late Phase II." Pfizer employs manufacturing-development teams consisting of two or three people from R&D and five or six manufacturing specialists. The team is initially headed by R&D, but leadership gradually transfers to manufacturing. The research/development group generates a dossier describing the product and process, preliminary expectations for demand and economic forecasts. At this point, the product and process are usually ready to move into a manufacturing plant for validation batches and production of Phase III material. Development and manufacturing work together to refine the process until it is deemed robust enough for launch and any realistic market expectations. "If we get a better-than-expected label, we want to make sure we're in place to make enough material to satisfy demand," Green says How Robust Is The Process? Determining how "robust" a process is can help smooth the transition from development to manufacturing. Pfizer uses an objective, statistical definition of robustness based on a Six-Sigma process capability index (PCI). If the process meets a predefined PCI after a number of batches, the company considers it suitable for commercial manufacturing. When Pfizer lacks data for a rigorous statistical analysis, Green and colleagues scrutinize the process with a checklist of questions: Is it scalable? Do batches meet expectations for reproducibility? Can the process be validated? Have conditions been varied sufficiently to convince everyone that it is producing a reliable product? A process has to be production-worthy in order to demonstrate robustness, says Warren Levy of Unigene Laboratories. "A manufacturer may need to perform three million-tablet runs to demonstrate robustness when all they need is a few thousand tablets for a Phase III study." At some point, Pfizer manufacturing weighs in with a detailed analysis of cost of goods (COG) and any capital investment required to satisfy potential upside demand. Like CMC, COG can be eye-opening when done early in development. "For some molecules, unfavorable COG can dramatically reduce a potential product's attractiveness," says Michael Glembourtt of Nektar. "Getting an accurate handle on cost of goods is relatively easy for drugs similar to others you've worked on before, but not so straightforward for novel products." At Pfizer, the transition from R&D to manufacturing sometimes occurs completely within the company. But, increasingly, outside manufacturers are brought in for difficult products such as those with multiple chiral centers or dangerous chemistries, or for state-of-the-art separations. Pfizer, for example, outsourced simulated moving bed chromatography for one product before acquiring the expertise in-house. Is Biotech Different? Biologics manufacturers seem to have a harder time grappling with capacity issues than makers of small-molecule drugs. Even if biotech's capacity crunch were nothing more than propaganda, the fact remains that many biotech companies have little or no internal manufacturing capacity. For this reason, technology transfer requires communicating just enough about a process to assure quality, but no more, according to Ashish Singh, vice president at consulting firm Bain & Co., Boston. "Because biologics manufacturing is so horribly expensive, companies are forced to begin thinking about manufacturing earlier and earlier," Singh notes. "Involvement with manufacturing groups is two to three times as intense in biotech as in chemical drug development, and remains so throughout a product's life cycle." Considering the natural divisions between manufacturing and R&D, devising the right structure for moving forward is even more critical in biotech, he says. Since most capital is tied up in manufacturing, companies seeking high return on investment (ROI) will want to keep as much manufacturing under their control as possible. "Especially in biotech," says Bain's Singh, "you need to own manufacturing capacity to maximize ROI." Better Knowledge Management Needed Creating a manufacturing team for the development work is the easy part. The difficulty, Breggar says, is culling and passing on the critical knowledge collected and processed during drug development. Even experienced companies may spend months getting the manufacturing right, then lose time as people enter and leave the company, and the ball is handed off from preclinical to clinical to manufacturing. "I have never seen this done really efficiently, where data weren't subsequently questioned by people receiving it. The problem goes well beyond planning and doing things early," he adds. Breggar recommends that companies get serious about applying knowledge management to clinical and manufacturing efforts. But results will not be immediately apparent, he warns Companies often underestimate formulation and dosage forms, especially when working with contract manufacturers. Darlene Ryan, CEO of contract formulator PharmaFab, Grand Prairie, Texas, recalls customers that waited too long because they thought their formulation needs were trivial. "They tell us "it's already formulated, all we want you to do is put it on your tablet press," she says. Unfortunately, they didn't understand how the machines worked at full scale, or the effect that excipients could have on a production run. Assess and Communicate Needs, Or Suffer Costly Delays Because busy contract manufacturers are committed to many customers, when one customer's formulations don't work out within a reasonable timeframe, the work must be pushed back.. And problems needn't be catastrophic to result in dire consequences. During one "easy" encapsulation project at PharmaFab, the active plug fell apart --and out pf--the capsule as the gelatin top was being inserted. "We spent three days trying to make that project work, then had to pull it for scheduling reasons," explains Ryan. "That incident delayed that customer's launch by a month because we were committed to working on other products." Some products require specialized equipment " a part change or custom-made tool, for example, which may take six months to order from a machine shop in Korea. "Marketing and development people don't always understand that," Ryan says. "They can't believe their product won't fit on our capsule filler. If we'd known of the problem earlier we could have told the customer about the delay, or tweaked the formulation to avoid the cost of new parts and time lost." It's also critical to inform any contract manufacturer when products require segregation or special handling. "We can often reformulate using different materials, but only if we know early on," says Ryan. To play it safe, contract manufacturers should be involved as early as possible in the development process, and sooner than inhouse manufacturing. Generally, it's best to involve contract manufacturers well before any commitments have been made for a clinical program, cautions Matt Coffey, Ph.D., vice president of product development for Oncolytics Biotech, a discovery- and development-stage company based in Calgary, Alberta. "Development programs have been shelved or cancelled because manufacturing could not keep pace with other developments," he says. Coffey believes that inhouse manufacturing should enter the picture when compounds are ready for animal testing. "Animal studies cannot begin without sufficient product." Deciding when to get involved in development is as much art as science. Although drug-makers tend to decide based on specific processes and product complexity, it's not always that straightforward, especially when economics or a drug's competitive environment come into play. Perhaps the best answer comes from Warren Levy of Unigene, who views manufacturing as a product-by-product issue. Like the response, "It depends," that doesn't offer much guidance. But what do you expect when you ask ten experts the same question?