Finding Biopharmaceutical PAT Success One Step at a Time

PAT is nothing new for Talecris, a maker of plasma-derived therapeutics that was spun off from Bayer Biological Products earlier this year. Many of its experts have worked on solutions for on-line monitoring and control as far back as the 1980s and 1990s. The key difference now is a more conducive regulatory environment since FDA released its PAT guidelines.We spoke with four key members of Talecris’ PAT team to find out more about the company’s approach: Vogel, Senior Engineering Manager, Process Control Systems; Joydeep Ganguly, Principal Process Control Systems Engineer and PAT lead; Gerold Mohn, Bioanalytics Director; and Peter Vandeberg, Principal Research Scientist, Analytical Development.Talecris will offer a more in-depth look at its PAT program and relevant case studies in the March 2006 issue of Pharmaceutical Manufacturing.Pharmaceutical Manufacturing: What is Talecris’ general approach to PAT?Joydeep Ganguly: A lot has been published about how difficult PAT is in the biologics industry. Our PAT approach, like many others in the industry, was to parse PAT efforts into smaller, more manageable tasks to begin with. We did not start by approaching PAT on a big-scale for entirely new processes; we tried to find small, quick wins. We also managed to identify and implement PAT in various areas — from packaging, to utilities, to production lines for yield improvements, to quality assurance. We used the early successes to build momentum for our bigger, more complex PAT efforts.Gerrit Vogel: We are taking a comprehensive approach. Bayer has a long history of PAT, but focused on the process analyzers and analytical instruments. Talecris had quite a history to build on. About the middle of last year, we requested the position of a PAT engineer to support the multidisciplinary effort and to focus on the implementation of PAT projects. This is now Joydeep’s position, which was approved and shows the management’s commitment to this initiative.P.M.: Where did you begin?Gerrit Vogel: We decided to start very small. FDA’s PAT guideline was helpful in our planning and implementation. It was perfect timing for us, and we came up with a multidisciplinary approach, a master plan on how we wanted to structure PAT, a core team with members from all different disciplines that would also work with a steering committee.When Talecris kicked off this initiative, we decided that we could not establish a huge PAT department and have a lot of projects that might bring a return on investment in a few years. We focused on things we either had already installed or were affordable to do in a reasonable time frame. We defined a quick win that was related to our water system, where we had already installed analyzers, and we have already tied them together. We implemented this in about two or three months.We have a high-impact project of vacuum detection of our fill lines. We also have strategic projects that really go in the direction of the PAT guidelines, which takes an approach to look at the entire bioprocess in one of our new products. We tried to do several clear projects that show potential and success rather soon, but without forgetting the long-term strategy. It worked well for us.Joydeep Ganguly: Gerrit and I have presented and wrote on the automation requirements of PAT. There’s a lot of interest in this from vendors and industry on issues such as data availability for PAT efforts. One of the first projects we looked into while implementing PAT was an IT concept that supported data transfer between different skids and analytical devices to allow for process understanding.P.M.: But your PAT program really goes back much further than that.Gerold Mohn: This whole thing started long before FDA’s guidelines and the PAT initiative. In R&D, we have always done a lot of things to characterize our processes, to make them more efficient and reliable. Our PAT initiative is actually a consequence of the groundwork that we laid out in the mid- to late 1990s.Gerrit Vogel: We put together an analytical study for our site back in 1998. We looked at what the key quality attributes are that we need to focus on for on-line measurement, and did many other things as well. So it had some history.Peter Vandeberg: In some ways, the biopharmaceutical industry is ahead of the chemical or small-molecule industry. For biological products, the process has always been important for defining your end product. So, it’s very common for biopharmaceuticals to be tested at multiple stages, for multiple intermediates. What we’re doing with PAT involved moving that testing from the lab to the production floor.P.M.: Not everyone in the biotech industry has embraced PAT. In your opinion, does PAT lend itself well to biopharmaceuticals?Gerold Mohn: From our experience, if you pull a sample somewhere in the process and bring it to the lab and wait for the results, it’s too late to make a decision in the process. You have the potential of losing a batch and losing a lot of money. There is always a need to have this on-line feedback in real-time, to make decisions and adjustments in the process. It helps to avoid problems in process.Joydeep Ganguly: What Gerold said is very important, because FDA has stressed that it does not consider PAT to be just another initiative to replace one test method with a more sophisticated method. Its logic is that there has to be some element of control or feedback such that you use decisions made during the process to ensure that real-time quality assurance in your product is met. It is important to note that you do not necessarily need feedback control for everything. What you do need is some sort of control (feedback/feedforward/administrative) or some sort of way to use in-process readings to ensure that with every batch you end up with a good, well-tracked product.Gerold Mohn: In biological processes this is more of a change than in small-molecule chemistry because you start with well-defined starting materials. But, we are making product out of human plasma and there is a lot of change over time. We have to enable our processes to deal with that variability.P.M.: Are there misconceptions about how useful PAT can be in the biopharmaceutical industry?Joydeep Ganguly: Yes. The biggest myth about PAT is that it’s going to be a quick fix for all problems, and that you have to spend a lot of money. It’s not about fancy toys and big, expensive analyzers. It can be something as simple as data that we used on past batches to find out what truly impacts processes, and using that control for future batches.Gerold Mohn: It starts with your strategy. You have to decide how to control a process, and then you select the method to make the measurement or control. We’re not starting the other way around, that we can measure something on-line and then see how we can start to control the process. You have to have the control strategy first, and then you select the right method to do it.Peter Vandeberg: Some people I talk to say, ‘What’s this PAT? We’ve been doing this all along.’ Certainly the technology is getting better, but for biotech products, column control, integrating UV detection and pH and other things with your control has been around quite a while. We’ve been trying to do this all along, and the guidance by FDA supports that.P.M.: What did the guidance mean to you, then?Joydeep Ganguly: Our approach was, we’ve always been doing PAT in some form, and we’re not so much looking for regulatory relief as we are improved process understanding, yield improvements and all the good things that PAT brings Talecris. The thing that the guidance gives us is the conduit to the Agency to support novel analytic methods and to go through submission strategies that are less cumbersome.Gerold Mohn: One conception about PAT in biological products may come from the fact that these are very complex mixtures of complex biological molecules. There is no easy method to measure, for example, the concentration or low-level contamination of a protein with on-line measurement. But if you look at the concept of PAT, it really means that you have to understand your process and learn what controls this process. You have to know how all of your parameters play together to impact the quality of your product.Gerrit Vogel: There are a lot of opinions out there about what PAT is — we said from Day One here is that it is not just putting in analyzers and instruments. We really want this to be an initiative with a structured approach. We created a mission statement, a vision statement, and we have a master plan, and a steering committee. We defined our approach to PAT.Because we have these complicated processes, we need a solid infrastructure even more to help us identify, monitor, analyze, and control, and also to apply statistics and having parameters ranges rather than fixed settings. There is a lot of opportunity in PAT that we are set to explore.P.M.: Have you had a problem finding enough qualified people in your organization to staff your PAT efforts?Gerrit Vogel: We have some excellent people at Talecris, and again, the approach we took was to bring a team together from different disciplines. We didn’t establish a PAT organization — our PAT team is really Joydeep and a few engineers that form a administration and implementation group, supported by various analytical, mechanical, IT, technology experts in their areas. This approach made the best people available to us. We need to keep the excitement up and define an approach, we have found an excellent set of experts who are also very motivated people.Gerold Mohn: If we have a problem at all, it’s a question of resources. All of these people who help us are very busy. We are a growing organization with a lot of production and development to support. That is the biggest challenge for Talecris.