The plenary session early Monday set the overall tone of the conference: “We’re growing because we address what will be happening in pharma, not merely what was done previously,” said Tim Stevens, associate director, Bristol-Myers Squibb, pointing out that while most professional meetings were shrinking, IFPAC was growing, largely due to the agency presence and access to upper echelons of FDA personnel. The thrust of the meeting was the future of pharma — which will clearly be dominated by continuous manufacturing (CM).
The meeting, just like the Olympics, began before the opening ceremonies. A short course on quality risk management was offered Sunday morning, presented by the staff of 4Tune Engineering (Lisbon, Portugal). The afternoon symposium differed from previous years: It was a workshop on multi-attribute methods. Papers were presented by speakers from Amgen (Jette Wypych): “current state and application;” Pfizer (Jason Rouse): integrating MAM into early phase biotherapeutics development; Amgen (Michael Abernathy): regulating MAM from an industry perspective; FDA (Sarah Rogstad): regulatory considerations on implementing MAM; and a mathematical model was presented by Eric Deeds (U of Kansas). The lively panel discussion at the end of the papers implied that there is quite a bit of interest in the approach as a form of bio-QbD.
On Monday afternoon, the week of simultaneous sessions began with a vengeance. The presence of biopharma papers was nearly equal to those on small molecules. Lisa Graham (Alkemy Innovation) opined that many PAT and analytic tools from the brewing industry could be applied to bioprocess control. Chuck Miller (Merck) had suggestions on how to apply our experiences with small-molecule PAT instruments to vaccine manufacturing (I even looked at NIRS for that at Merck in 1992). Any number of Raman papers also addressed bioprocessing controls (20+) — a large increase over even last year’s meeting. Clearly, Raman is growing in stature, rapidly.
Legacy products were also more visible this year. Jose Menezes (4Tune Engineering) described a risk-based approach to implement post-approval changes in legacy processes. With so many blockbusters coming off patent and competition from biosimilars, the production costs of existing products will need to be brought down by moving to PAT and QbD methodologies. And, under new guidances from the agencies, lifecycle improvements go hand-in-hand with applying PAT/QbD to products, post-approval. Silvia Ribeiro (4Tune Engineering) detailed how risk assessment could be applied to product development and improvement over the lifecycle of the product. At least a dozen papers suggested ways for legacy products to be eased into PAT applications, on step at a time: Start with raw materials classification, move to mixing, then explore granulations. These points of monitoring have been addressed for decades and their implementation should be a “slam-dunk,” even for small to moderately sized generics and CMOs.
Of course, NIR remained the most applied spectroscopic method and was the focus of most presentations. Nowhere was this more obvious than in CM. All aspects, including the pre-mixes, were touched upon. Benoit Inge, facilities manager, Merck Group, addressed the fact that not all components were added simultaneously. Often the excipients were “pre-mixed,” followed by the addition of the API, and, finally, the lubricant is added. Each step needs to be monitored and controlled and this may be the most important part of CM. It was pointed out by several presenters that the chemometric equations for CM need be faster and less complex, so as to control the processes in real time.
The exhibition, while small, focused on process evaluation and control. The exhibitors were the usual suspects with the exception of Indatech (France), making its debut with real time NIR and/or Raman monitors capable of performing qualitative and quantitative measurements on (up to) 800,000 tablets or capsules/hour.
It was the only meeting where I wish I had a few partners with whom to share the event, since it was more than one person can see.