There are a number of strategic advantages that companies can achieve in pharmaceutical production by transferring production between sites
They can safeguard supply by producing at more than one site, and improve distribution by moving production closer to critical markets. They can also reduce programme costs and risks by moving production to sites that are better qualified, can produce more economically or are better positioned to meet the needs of regulators.
However, for several reasons, transferring production — and the technologies that support it — can be risky. The same product can behave differently in different equipment, resulting in low yields or even batch rejections; staff at a receiving facility may not have the proper technical skills to execute a specific process; or there may simply not be enough trained staff on site to ensure that the job will be done correctly. Also, even as companies may not take the time (or expend the effort) to create and provide good documentation of their product or process, their partners may not invest the necessary resources to ensure efficiency and quality.
If any of the above occurs, time and money will be lost in remediation and patients may not receive their medication. The net result of these risks is that companies do not transfer production as often as they probably should. However, Patheon has a great deal of experience in technology transfers and has developed processes and techniques to reap its rewards while reducing its risks.
Product X: The product was a registered intermediate of a novel first-in-class active pharmaceutical ingredient (API) that Patheon was producing for a client in a 10-step process. One high-temperature step was particularly complex, requiring good control to balance quality and yield. We were spending a disproportionate amount of time and effort getting this step right, and we knew that other manufacturers — with more experience in this specific technology — could likely do it better. Accordingly, we set out to find a manufacturer who could take over this production step for us, and we found one in China that could supply our plant in Austria. To get the process up and running on the supplier site in China, we sent two of our people to the Chinese plant to work with them for a week; from the first batch onwards, the product was within specification.
Including 4 weeks to evaluate suppliers, 4 weeks of tech transfer activities and 5–8 weeks for raw materials and lead times, the total time from project start to the first batch produced in China was about 4 months.
Product Y: This product was a recombinant-fusion protein entering Phase III clinical trials. Patheon’s client had a partially developed perfusion process and wanted production to commence simultaneously in Europe and Asia-Pacific. Their timeline was tight.
We transferred the process from our client to our plant in Groningen, the Netherlands, and our plant in Brisbane, in Australia, scaling-up production in both at the same time and writing half the batch records in each location. The total time from receiving the process information and ordering the raw materials to full production at both sites was 4 months — about 30–40% faster than the industry average.
Product Z: Product Z was a lyophilised small molecule product. Our client wanted to transfer commercial production to Patheon. Unfortunately, the technical data we received about the product and the lyophilisation cycle (the freeze-drying process) was not sufficiently robust to guarantee a consistently high-quality product.
In particular, we could see from the historical data that the lyophilising cycle was not challenged at the edge of the critical process parameters (CCP), and there were several cake appearance issues even under optimum conditions. Instead of replicating the process as it was, we worked with the client to improve it, thereby getting the product right first time, and improving both yield and quality compared with the prior process.
A shown in the examples of products X, Y and Z, we can see four important elements of our approach in operation.
1. Assessing risks and developing mitigation plans prior to transfer. To do that, we deploy a seven-step process we call the 7Ms. They are:
A comprehensive risk assessment will yield a predictable set of activities on which a plan can be constructed. Such a plan allows all stakeholders to begin on the same page, ensuring the alignment of both the sending and receiving teams before the actual transfer.
Furthermore, these steps generate thorough documentation of both process and product, making knowledge transfer easier and less vulnerable to informational gaps. Once identified, those can be filled, and steps that need to be improved can be identified (as was the case with Product Z’s lyophilisation cycle) and resolved before the transfer.
2. Fixing problems prior to transfer. As noted, our assessment of product Z revealed a weakness in the lyophilisation (lyo) cycle, notably cake appearance issues. Accordingly, and together with the client, we performed a systematic characterisation of the product and its formulation to better understand its behaviour during lyophilisation.
We mimicked the lyo process in the lab, identifying the critical process parameters that were likely to vary in production, such as equipment performance, temperature, humidity, and time and velocity ranges. Then, in the lab, we measured the impact of those factors on yield, quality and cosmetic appearance, and varied them to see which changes produced the highest quality across all measures.
In this way, we developed a more complete understanding of the factors that affected production and defined the envelope in which quality and yield would be optimal. During this process, we were also able to understand better the physical characteristics of the API, ensuring more reliable behaviour in production and limiting the impact of these variables on the drug product.
Fixing the lyo problem before the transfer enabled us to improve the process, strengthen the product’s regulatory submission with the additional information we obtained, and created a strong relationship with our client that served us both well in this transfer and in subsequent ones.
3. Standardising processes and equipment. For the recombinant-fusion protein, Product Y, Patheon’s plants in Groningen and Brisbane essentially had the same upstream and downstream equipment (such as chromatography columns and skids), so we could be confident that what worked in one would work in the other. And not only was the equipment the same, but the two plants also have very similar operating procedures.
For example, each plant operates the 500 L perfusion bioreactor (a finicky process) in the same way and has almost identical standard operating procedures for common downstream processes. Consequently, batch records written for one plant could be transferred easily to the other. We also made sure that the materials and supplies were already in both facilities’ systems, so neither they nor their suppliers needed to be requalified, saving time and ensuring consistency and quality.
4. Paying as much attention to people as to process. Typically, the staff at the sending plant have expertise in a particular product and process. That, of course, should be captured in the documentation. However, even if it is, a specific answer to a niggling problem may not be easy to find, and even with the best documentation, a nuance may go missing. The staff at the receiving plant may have more extensive expertise in the type of product, or technique, that can be applied to improve the existing process and streamline the transfer. However, sender and receiver can only help each other, and optimise the transfer, if their working relationship is close, collegial and congenial. The single most important ingredient that makes a relationship work is trust. And trust emerges most reliably from successful collaboration.
For the Fc-fusion protein, Product Y, the teams at each of the two plants communicated in real-time about progress, challenges, solutions and so on. Joint project team meetings were held regularly, and communication between the plant managers was excellent. Given that all the communication was done remotely, it certainly helped that many of the people in the Groningen and Brisbane plants knew and had worked with each other previously.
If teams have not worked together before, partners need to take the time and trouble to allow trust to develop naturally through proximity. For example, for the API production step on Product X that we transferred to China, we assumed the not-inconsiderable cost of sending knowledgeable people thousands of miles to the site was money well spent. It served to accelerate the knowledge transfer and when problems did surface (as they inevitably do), the receiving plant was not left to its own devices to try to solve or fix them without adequate support.
In the case of Product Z, we assembled a joint team with the client to work on improving the process, giving both parties time to get to know each other before the transfer. This subsequently paid off as the team was able to work together productively throughout the lifetime of the product, supporting each other in matters concerning strategy, supply, distribution and regulatory affairs.
As noted, technology transfers don’t happen as often as they might, or as often as perhaps they should. Companies have legitimate concerns about supply interruptions and the costs they may incur as a result. Then there is the expense. To plan one thoroughly and execute it carefully requires a significant investment. To minimise the cost, companies are sometimes tempted to shortcut the planning and rely on what’s worked before for similar products. This is a false economy.
Looking at our history of transfers, we have found that the chance of getting them right by mimicking the transfer of a similar product are no better than 50% — a coin toss. Conversely, beginning fresh, with a blank sheet, and analysing the process as if you have never done it before, the chance of right-the-first-time success is better than 90%. Yes, technology transfers require investment. However, the cost is less than that of fixing one that went wrong.
In addition, a technology transfer provides the opportunity to make product and process improvements that might otherwise be missed. In practice, production processes often run for years without being improved meaningfully. Companies always have other priorities, including new projects, which steal attention from ongoing products. But in making a transfer, a company is forced to make changes; there will always be differences at the receiving site that need to be accommodated. Those changes — properly planned for and executed — can realise enduring strategic advantages by lowering production or distribution costs, or by improving a company’s competitive position in a given market.
Transferring the production of sophisticated products from one site to another, or to several others, will always be challenging and will always involve risk. But not only are there more opportunities today for companies to improve how they have their products made, there are also more options to secure supply, produce closer to critical markets, become more agile when responding to changes in demand, and secure sophisticated help with technically challenging production steps. Specialist expertise, solid processes and standardised operating procedures and equipment all work together to minimise risk and costs.
A consistent risk-based approach, continually updated with new data from quality by design and design of experiments initiatives, will improve both Patheon’s and the industry’s ability to efficiently, effectively and seamlessly transfer production going forward. The rewards can be substantial, and transferring technology should be an option that every company places top of mind.