By Paul Thomas, Senior Editor
Yesterday at the AAPS 2010 annual meeting in New Orleans, a half-day symposium was held on continuous manufacturing in the drug industry. Generically titled, “Continuous Manufacturing: Benefits and Challenges,” the session was in fact anything but ordinary and had a fluid mix (pun intended) of speakers and topics:
The University of Washington’s Brian Marquardt, PhD, spoke on his team’s and others recent efforts and new technologies in continuous drug synthesis; AstraZeneca’s James Kraunsoe, PhD, covered AZ’s cutting-edge work in continuous granulation and tablet manufacture; while James Evans, PhD, of the Novartis-MIT Center presented a blue-sky vision of what continuous manufacturing will become in the next five to ten years (from synthesis through tabletting and packaging); finally, FDA’s Christine Moore provided a regulatory context for all of this work going on—most importantly, Moore reiterated what we’ve heard consistently from the Agency lately about continuous manufacturing: please, by all means, get to it. (Some caveats to this, of course–see below.)
I’ll present summaries of each of these talks here in the next few days. While Moore’s was last on the docket, I’ll begin with hers to set the regulatory scene.
Is continuous manufacturing in pharma finally beginning to snowball? From these talks (and granted, the speakers are a bit biased), the answer is yes.
FDA’s Moore: The Future of Continuous Is, or Can Be, Now
Christine Moore, PhD, Deputy Director for Science and Policy for CDER’s Office of New Drug Quality Assessment, delivered a presentation whose mission was clear: to convey that the regulatory framework is in place to adopt continuous manufacturing methods, and that FDA is in full support of manufacturers who wish to do so.
- The Agency sees definite economic and quality advantages to continuous manufacturing.
- The science exists (and technologies to support continuous processes are on the market).
- There are “no regulatory hurdles” for industry to implement continuous practices, she said, though providing a caveat: “There is a lack of experience [within and without FDA]; therefore we really need to work together as we move forward with these approaches. . . . We recommend frequent and early conversations with our organization.”
Moore began her talk by qualifying the meaning of “batch,” pointing out that regulatory definitions of a batch include what we think of traditionally, but also those batches that are produced through continuous methods. 21 CFR 210.3 defines batch as “a specific quantity of a drug or other material that is intended to have uniform character and quality.” Said Moore, “Nowhere in this definition to you see any mention of mode of manufacturing. The key phrase is ‘specific quantity’.”
For continuous manufacturing, of course, it is still necessary to define a “batch” to meet various aspects of cGMP. But, Moore said, defining a batch is up to the manufacturer. It can be based on: a production time period; a variation in production (e.g., the use of a different feedstock); or another definition such as equipment or maintenance cycles.
Further, Moore reiterated that continuous manufacturing is consistent with FDA’s Quality by Design efforts: It is a more “modern manufacturing approach”; it has the potential to improve quality assurance and product consistency; and it enables quality to be directly built into process design.
The area of continuous manufacturing that is most different under the QbD paradigm is the control strategy (as defined by ICH Q10), Moore acknowledged. Control strategy can include such factors as parameters and attributes, facility and equipment operation standards, in-process controls, and so on.
Moore emphasized that manufacturers must understand the “dynamic system of response”, which includes residence time distribution (RTD), steady state operation, and process startup and shut-down. “I’m really talking about the mixing behavior of the system,” Moore said, “which can greatly affect product performance.” The typical method to analyze system dynamics is to look at an RTD, she said, which can be evaluated by tracer studies (e.g., a pulse through the system of a step input via dyes, changes in concentration, etc.). Further, she said, an RTD can provide insight on the effects of disturbances. Poor mixing patterns such as deadzones can cause “carryover issues” for product, so it is critical to minimize this risk.
Monitoring and Control
Next, Moore turned to specific in-process methods for monitoring and control of continuous process, citing several tested technologies:
- Spectroscopic (NIR, FTIR, Raman)
- laser light scattering (e.g., for particle size)
- NIR (e.g., for moisture)
- NMR (e.g., for weight)
- GC (e.g., for solvent detection)
Chemometric modeling based on these technologies is an essential part of control and tuning of continuous processes. Moore’s key considerations include:
- Establishing a good calibration data set, “including potential sources of variance”. Manufacturers must be able to “show a uniform distribution of spectra over the process’ analysis range”
- Including the “appropriate” data pre-treatment and spectral ranges in model development, and justification of the number of model factors used. It’s important to avoid “overfitting” these factors, she said.
- Performing thorough model validation
- Developing a “robust and representative” reference method. “You don’t want to go through all this effort and find out that there was something you missed in your method,” said Moore.
Surrogate models are becoming a more common means of establishing control, Moore noted, especially related to dissolution. A typical approach, she said, is for the manufacturer to build two models—one qualitative according to principal component analysis, which provides an operating range, and the other quantitative, which predicts via partial least squares where to generate a calibration curve. The considerations for building dissolution multivariate models are very similar to those for building chemometric models (listed above).
In regards to integrating these practices with QbD approaches, Moore recommends that manufacturers share their dissolution method early—in Phase 2, for example—to ensure that all parties are on the same page.
Lastly, Moore addressed Real Time Release Testing, a critical component of continuous manufacturing efforts. RTRT provides increased manufacturing flexibility and efficiency, she noted, and the potential to improve quality and cut costs. She offered up a few reminders on the approach:
- First, specifications are still required for RTRT.
- They should be “representative of actual measurements” (whether in-process or surrogate).
- Alternatives can be included for stability monitoring.
Is the future now? Moore asked, finishing her talk. Is the continuous manufacturing of drug products upon us? She didn’t answer, but made it clear that, if so, FDA is on board. Of continuous manufacturing she said, “All in all, it’s a more modern approach to manufacturing and control.”
- Continuous Manufacturing in Pharma: Myths and Missed Opportunities
- Despite many precedents from other industries, pharma resists continuous manufacturing and its financial benefits. Ali Afnan questions why. ...
- With Continuous Operations, Can Drug Manufacturing Become a Rock Star?
- At AAPS 2010, Novartis-MIT’s James Evans presented a blue-sky vision of continuous manufacturing, saying R&D shouldn’t get all the glory....
- Continuous Granulation at AZ: Where It Fits in Tablet Manufacture
- Continuous manufacturing opportunities must be “appreciated, evaluated, collaborated on, understood, and pursued!” says AstraZeneca's James Kraunsoe....
- Going with the Flow: Advancing Continuous Drug Synthesis
- At AAPS 2010, U Washington's Brian Marquardt highlighted progress in the use of flow reactors for the synthesis of drug...
- Pharmaceutical Quality and Lessons from Toyota: Thomas Friedli on Creating a Continuous Improvement Culture in Pharma
- "You can only achieve long-term savings if you have stable running equipment that enables stable and robust processes. This is...