On 10th November at 9:00 AM - 9:30 AM CT, Sarang Oka, scientific fellow, Hovione, will do a talk titled "Addressing Practical Challenges in CM via Novel Hardware and Automation" at PharmSci360 in San Antonio, TX.
Hovione
1. To kick things off, could you give us a quick overview of what continuous manufacturing (CM) is and why it’s becoming such an important approach in pharma?
Pharmaceutical continuous manufacturing (CM) is an advanced manufacturing approach in which raw materials are continuously fed into an automated process, and the final drug product or drug substance is continuously and simultaneously removed. The method integrates all manufacturing steps in a single, highly automated process train, housed in a single process suite. This is in contrast to traditional batch manufacturing, in which a discrete quantity of product (batch) is processed one step at a time through the different manufacturing steps, often involving wait times, manual transfers, and separate quality checks between steps. Continuous Tableting (CT), which is the focus of my presentation, is continuous manufacturing of solid oral dose drugs, specifically, tablets.
There are multiple advantages to the technology, and in our experience, different industry segments resonate with the technology for own their unique value drivers. And as such there are several reasons why we are seeing increasing adoption of CM, and specifically, CT. The most important ones are below:
- Speed to Market: Scale independent nature of CM processes means that the scale-up step, typical for a batch process that is executed over the course of the product launch, is eliminated. For programs, where the CMC roadmap, rather than the clinical end point, is the rate limiting step, CM enables products to launch faster. This was the primary motivation for some early CM adopters and approvals and continues to be a very attractive value proposition.
- Supply Chain Agility: Scale independent nature of CM processes also means that users can quickly adapt to market demands, both, on the low and the high end. And importantly, this also means CM processes can respond to drug shortages during emergencies and supply chain disruptions.
- Efficiency and Speed: Production time is significantly reduced compared to batch processing. CM eliminates lengthy hold time between batch steps, shrinking production times from weeks to days or less.
- Sustainability and Cost: The facility footprint is meaningfully smaller compared to an equivalent batch facility, and besides costs, this translates to a lower carbon footprint. Additionally, based on the number of the operators supporting the effort, the overall OPEX costs can also be substantially lower for CM. One of the key motivations for the innovations that I will highlight during my talk was to reduce the overall numbers of people running the process during development and commercial manufacturing.
- Product Quality and Consistency: CM enables meaningful implementation of Process Analytical Technology (PAT), and this PAT driven process monitoring allows for continuous and enhanced quality verification. Accompanying process control results in more robust processes and superior product quality.
2. Your work focuses on the “practical challenges” of CM. What are some of the biggest bottlenecks or pain points teams usually face when implementing it?
There are certain challenges that if tackled have the potential to lower barriers for adoption. This is not a comprehensive list and do not touch on aspects such as regulatory and quality systems, but key challenges include higher complexity of automation software for CM machines, lack of robust and cheap containment solutions for handling potent materials and ease of cleaning, and spectroscopic-PAT solutions for low-dose formulations. I will expand on one of them where we see the biggest opportunity – complexity of the automation elements.
The multi-layered nature of the control system that runs CM processes can often be underappreciated in the early stages of design and installation. It originates from integration of the control architecture of hardware components, combined with integration of the control logic of the transition elements (for example, pneumatic transfer elements, rotary valves, transfer chutes, etc.). The classical hardware integration subsequently needs to be married to a material tracking and segregation system along with the PAT Management System (PMS). The broad range of features of the control architecture translates to more intricate system interfaces.
Besides the comparably higher effort needed to qualify CT automation system due to the above, operator training also becomes an important consideration. Unlike batch processes, where operators run single unit operations typically via a local HMI, the integrated nature of continuous processes results in an automation interface that controls multiple unit operations. The frequency of operator prompts, process warnings, and alarms that an operator typically must address in a continuous process is also found to be higher by virtue of multiple unit operations running simultaneously. Naturally, as the technology evolves and becomes simpler, the learning curve will become less steep, and we and others are working towards this simplification effort.
3. Speaking of your current work, you’ve developed an auto-injector for tracer material and an automated sampling device. How does this innovation reduce human error and increase efficiency?
Charactering the residence time distribution (RTD) of the process is an essential step during process development. Despite the need for accurate estimation of RTD parameters, characterisation experiments today are performed manually and require a high level of coordination and personnel involvement to introduce the tracer material, collect samples, and timestamp all events to match up with the process data, creating a significant risk for human-induced experimental error. Our work constitutes the development of an auto-injector for the tracer material that is situated above the inlet of each blender and is coupled with an automated sampling device for tablets; the entire apparatus is coupled with a programmable recipe. This enables the entire RTD experiment to be run automatically by a single individual from a single point of control.
4. The development of a control system that can run pre-configured multi-point DOEs automatically is a big step. What does that unlock for researchers and manufacturers?
Running multi-point DOEs with today’s control system involves an individual continuously supervising the process, manually inputting multiple condition setpoints, waiting for steady state, collecting data at steady state, then inputting the next condition and so on. We wanted to take away the heavily manual nature of this workflow. Our innovation enables the user to run pre-configured muti-point DOEs in a completely hands-off way. A feature in the recipe management system of the control software now allows the user to pre-configure a DOE as a semi-automated sequence. This includes setting the operating set points at different conditions, the length of operation for condition, and the length of transition between set points coupled with automated sampling. Once the user hits “GO’, the automation system takes over and runs the DOE, including automatically taking samples, and labelling them. The user only has to confirm they are ready to move to the next step. Not only does this reduce the number of users needed to run DOEs but it also minimises potential for experimental error.
5. Expanding on this, beyond efficiency, how does automation improve safety — especially when working with highly potent compounds?
The auto-injector device and the automated sampling device that I mentioned earlier are clear examples of how automation can tackle containment challenges. Addition of tracer material today is a manual process and involves compromising containment – the user has to manual insert the tracer into a running operation, and in the process breaks containment increasing risk of exposure. This is a non-starter when working with highly potent compounds. The auto-injector allows the user to preload the tracer to the injector device, and the user can deliver the tracer at the appropriate time at the push of a button from the HMI, all without breaking containment.
Subsequently, the automated sampling, packaging and labelling device takes tablet samples when called upon, packages them in PE bags, and appropriately labels the bags, again at the push of an HMI button. This process today is completely manual and exposes the user to the sample. You can imagine how our device provides superior protection.
6. You mention the ability to swap between high-throughput and low-throughput blenders — what does that flexibility mean for companies in terms of cost, efficiency, and scalability?
Continuous manufacturing process trains today can be characterised by the mass throughput that they can process. This includes a nominal range that is typically dictated by the rate limiting unit operation in the process train. While the product attributes and formulation will dictate the throughout range, generally the blender sizes will dictate the nominal allowable range for most continuous direct compression (CDC) units. The need to operate at throughputs outside (below or above) of this nominal range at various stages of product development normally requires switching over to a completely different sized machine. And this means purchasing a new machine rated for that throughput requiring additional investment and time. We wanted to address that. Innovation developed as part of this work constitutes a modular process train that enables the user to swap between “high-throughput” and “low-throughput” linear blenders. Additionally, modularity is built into film coating step that enables the user to “add” more coating units when there is need for additional capacity. The modular hardware is complemented by adaptable software that facilitates seamless switching between low throughput and high throughput modes of the process train. This means that a user is able to run their process, on the same machine, from 1 kg/hr to 200 kg/hr. Additionally, it is also possible to run the machine as a batch unit for batch requirements, or if there is a need to increase equipment utilization during times when CT is not required. This is made possible by allowing the user to dock a blending tote onto the tablet press, bypassing the loss-in-weight feeders and the linear blenders. The machine is called the CDC-Flex, and the name captures the Swissknife like flexibility that the machine provides.
7. If you could leave attendees with one key takeaway from your PharmSci 360 talk, what would it be?
We firmly believe in CM, specifically in CT as a technology standard of the future and are fully invested in it. We believe that the technology is a meaningful step forward for the industry and can deliver on its many potential benefits shared earlier. For the past couple of years, and in partnership with GEA, we have spent substantial effort on hardware and automation innovation that we hope will simplify CT, and in the process lower barriers to entry. And as we continue on this journey, we look forward to welcoming many more willing collaborators and partners to further adoption and transform pharmaceutical manufacturing of oral solid dose products.