Moving Beyond the Batch

Jan. 14, 2008
Cost savings, FDA endorsement and the willingness of drug manufacturers to collaborate to find new methods might facilitate the move to continuous processing.

Changing the status quo is always difficult. In pharmaceutical manufacturing, history is on the side of batch processing, which explains why continuous processing continues to face high barriers to acceptance.

Not every pharma process can, or should, run continuously, progress is being made in areas such as mixing and crystallization. “Not a lot of people in the pharmaceutical industry are experienced with continuous processing. They know and understand fluid bed drying and high shear granulation,” says Carl Anderson, PhD, assistant professor of pharmaceutical sciences at Duquesne University.

“They also perceive that regulatory problems and issues will be associated with the continuous process.” However, batch processing at some plants can be characterized by high labor costs and an excess inventory, leading to higher production costs.

“Continuous manufacturing promises efficiency because it is a wellcontrolled and flexible process,” says Bernhardt Trout, a Massachusetts Institute of Technology (MIT) associate professor of chemical engineering. “There is less waste and higher quality. You won’t have to throw out any batches, because they must adhere to tight specifications that are built into the process.” There are further advantages to a continuous process.

According to Lee Proctor, technical director at the pharmaceutical intermediate manufacturing company, Phoenix Chemicals (Bromborough, U.K.), utilizing continuous processing and flow reactor technology is a strategy for achieving improved efficiency. It delivers lower raw material and waste costs; it reduces environmental emissions and energy consumption, as well as overall unit operations and operational costs.

In addition, FDA has championed process analytical technology (PAT), Quality by Design (QbD) and the Critical Path Initiative, clearing the regulatory hurdles for continuous processing. These approaches identify critical processing variables and their effect on performance and quality, with the goal of improving product quality by designing it into the process.

This is exactly what continuous processing aims to achieve.

The $65 Million Investment

Last year, Novartis announced it would invest $65 million over the next 10 years to fund research activities at MIT and establish the Novartis-MIT Center for Continuous Manufacturing. The collaboration is enough to support research efforts for seven to 10 faculty members and dozens of graduate students, postdoctoral fellows and staff scientists.

Each researcher’s specific skill set will be needed. The goal is not to reinvent the batch process but to establish a new, never-tried process with all of the benefits of continuous manufacturing. “We are trying to make a quantum leap in the drug industry and transform how drugs are developed and manufactured,” says Walter Bisson, Novartis’ program leader for continuous manufacturing. “This means reducing cycle times and development times as well as looking at the efficient use of equipment, building costs, energy savings and raw material generation.” However, making this leap will not be easy.

The “blue sky” approach this collaboration is taking means throwing out existing processes and starting from scratch. It calls for inventing new methodologies and a new set of technologies. “We are investigating all aspects of the pharmaceutical process – from synthesis to the final drug form,” says Trout. “We hope to come up with a fully integrated, fully controlled system.” One area of particular concern to Trout is crystallization, a chemical solid-liquid separation technique in which mass transfer of a solute occurs, from the liquid solution to a pure solid crystalline phase. Factors such as impurity level, mixing regime, vessel design and cooling profile can have a major impact on the size, number and shape of crystals produced.

Poor control of the crystal size can result in longer process times and an inefficient process. The success of a crystallization step often depends upon how the crystals were designed and characterized in the lab. Analytical tools have already been developed to help measure the composition and particle size on-line.

The Insitec by Malvern Instruments (Southborough, Mass.) can make these measurements through a laser diffraction-based particle size analyzer, lending itself to a continuous process. “Particle size is determined from the pattern produced as light is scattered by particles in a sample,” says Malvern project engineer Alon Vaisman, project engineer. “The instrument is already used in chemical processing.”

The Novartis-MIT partnership is piloting easy to synthesize, small molecule drugs that the company already manufactures, as well as brand new chemical drugs in the R&D phase. The researchers are not yet ready to tackle the biologic drugs. Not everyone is taking as radical approach to continuous manufacturing as the MIT-Novartis project. Many companies are focusing on bringing continuous manufacturing to some of their unit operations and then fitting the pieces together.

This is often called semicontinuous manufacturing. Industries such as food, chemical, automotive and electronics have used continuous manufacturing systems for years. However, there is a certain inherent risk involved with the process. Continuous processing does not solely rely on out-of-process measurements. In a batch environment, a product recall is easy because each one has lots and batch numbers. In a continuous environment, product would have to be defined by the time when it was produced.

“How do you know which tablets are manufactured from the same batches of individual components as the bad product?” says Ron LeBlanc, engineer from the Little Island, Ireland engineering firm, PharmEng Ltd. “Risk is a major consideration as well as process knowledge and understanding.” However, FDA experts have said that using time rather than batch number, there would be little difficulty tracing recalled materials in a continuous environment.

Phoenix Chemicals’ Proctor thinks it is important to consider the entire process as a candidate for continuous processing (much as Novartis-MIT is doing) and not just the reaction stage. For example, many chemical processes rely on a liquid-liquid extraction to isolate a product or remove a by-product. “The use of off-the-shelf technologies such as centrifugal extractors can be readily implemented into an existing process with immediate benefit,” says Proctor. “This technology can significantly reduce the quantity of extraction solvent required reducing raw material costs, waste costs and environmental emissions.”

Another example is the use of UV oxidation reactors that can be easily incorporated into a process to treat toxic waste streams and reduce disposal costs. While these ancillary continuous process elements are often overlooked by the industry, familiarization with these technologies and the “willingness to adapt and be flexible” can have an immediate benefit, according to Proctor.

Continuous Chemical Reactions

Phoenix Chemical products are made using one or more continuous processing elements. For example, the company is the largest supplier of products derived from diazomethane, a highly toxic, reactive and explosive gas. The diazomethane production unit is part of an intecontinuous grated multi-stage process that produces 200 metric tons of intermediates for the manufacture of HIV drugs and other pharmaceutical applications per year.

The process continuously operates four chemical steps including two washing stages, an integrated waste treatment stage and an evaporation and environmental abatement and solvent recovery system. The company is also the principal supplier of a key chiral intermediate for the world’s largest statin drug. This intermediate is produced using a multi-stage flow process including flow reactors, continuous extraction and a continuous wiped film distillation technology.

A continuous crystallization process was used by Bristol- Myers Squibb to produce small crystals of two compounds to improve their dissolution rate and bioavailability. A crystallization technique using a high-shear rotor-stator chamber to crystallize compounds at a high degree of supersaturation was effective in producing the desired small particles and was scaled up in the pilot plant.

These new processes do not spell the end for batch equipment, yet some changes may result. “There will always be a requirement for drugs to be manufactured on a batch basis. However, mitigating to continuous manufacturing will force the batch manufacturing process and equipment to become efficient to be competitive, and this is not a bad thing,” says PharmEng’s LeBlanc.

“We might end up with some new types of equipment that don’t currently exist,” says Bisson.

Industry Collaboration

Research into continuous pharmaceutical manufacturing is also underway in Europe, as part of IMPULSE, an industry-led project, with active participation from three different major supply-chain sectors: pharmaceuticals (GlaxoSmithKline), fine and specialty chemicals (Degussa) and consumer products (Proctor & Gamble).

In addition, four other industrial partners contribute their expertise in novel solvent utilization, process innovation, process design and control and project management. Much like the Novartis-MIT partnership, the IMPULSE approach is one of structured multiscale design, matching the length and time scales delivered by processing equipment to the needs of the process.

This approach lends itself well to the continuous manufacturing process, which IMPULSE researchers are studying. The physico-chemical characteristics of the process are considered, and from these, the best process conditions are devised to deliver the best process. Once the process has been developed on an equipmentindependent basis, then the most appropriate equipment can be designed or selected. IMPULSE researchers’ vision is one of smaller, safer, cleaner and more responsive manufacturing facilities with novel processing capabilities.

However, inventing new processes is not a quick fix for manufacturing problems. “Inherently, a badly run process, whether it uses a batch or continuous method, will end up unsatisfactorily,” says Anderson. “Typically, continuous manufacturing has the advantage that quality is already built into its creation.”

FDA Framework

FDA’s PAT and Quality by Design initiatives advocate that quality be built into the product throughout the process, not tested for in the final product. Since building in quality is also the basis for continuous manufacturing, adopting continuous processes should remove one roadblock to process innovation.

“You can do batch processing without PAT, but a properly run continuous process is not even possible without the concepts of PAT,” says Anderson. According to Bisson at Novartis, they are working with FDA on their new methods. Before the collaboration started, they presented a white paper to FDA and were “encouraged to continue” their aspirations. They are keeping FDA updated on their progress from the very beginning. It might be a cliché, but “thinking outside the box” is necessary for a continuous processing transformation.

“A continuous process requires the ability to think laterally and have a proactive mindset across the entire team, from lab development through to chemical production,” says Proctor. “There really is no ‘one box fits all’ solution to continuous manufacturing.” A more efficient and controllable process is the end result of using a continuous method. According to LeBlanc, the start-up and shut-down processes are the most difficult to control. “It is easier to keep a process within its operating boundaries when it is moving steadily,” he says.

Continuous processing also offers smaller equipment and plant footprints, reduced utility requirements, less product variability, less manpower, less testing and higher levels of equipment utilization and overall equipment effectiveness. “At the very end, we are looking to create a new paradigm. We hope that it will be the gold standard in the future,” says Bisson. “It might take 10 years or it may take 100.”

About the Author

Bill Swichtenberg | Senior Editor