Jennifer Markarian is manufacturing editor of Pharmaceutical Technology.

Jennifer Markarian

Digital twins of manufacturing processes are being explored in pharmaceutical manufacturing for the benefits that can be harnessed from a real-time simulation of the actual process. In 2020, GSK partnered with Siemens and Atos (a digital transformation company and a long-time Siemens partner) to pilot a digital twin for their vaccine manufacturing process that modeled and controlled the adjuvant particle manufacturing process. The partners succeeded with the project goal of proof of concept, and GSK is now seeking to replicate the digital twin concept on an industrial scale for its vaccine development and manufacturing platforms.

“Industry 4.0 concepts, such as digital twins, are generally thought to be a good idea, but this project created belief that it can really be used in vaccine development and manufacturing,” says Matt Harrison, head of Sciences, Digital Innovation, and Business Strategy at GSK Vaccines. “Now we are developing a vision for using it for multiple products and platforms, and we are deciding which opportunities will best deliver value and have the most impact.”

A significant benefit of the digital twin was that the model allowed developers to better understand the relationships between process variables to control variability.

“Controlling variability allows us to improve quality and make product ‘right first time’ every time,” says Harrison. “The model and the high degree of process control also allows us to better predict and optimize product yield. The aim was to build the twin during process development and then use it in manufacturing. The digital twin is an additional ‘deliverable’ of the development process.” Doing the work to understand and optimize the process “

” in the digital twin, rather than actual experiments with real material in a lab or test facility, saves a significant amount of time, materials, and cost.

Although the team has not yet gone through a tech transfer between manufacturing facilities using the digital twin, in theory, the ability to model any differences in scaling up or moving from one facility to another will reduce tech transfer risks, says Harrison.

Factors in the success of the pilot project include strong backing from GSK leadership and good teamwork both between GSK departments and with partners Siemens and Atos, notes Harrison.

Another key to the project’s success was that we started small and focused on our goal of demonstrating the concept. The next challenge is industrialization—using digital twins at scale in both development and operations throughout the organization. This effort will require continuous improvement, and we expect we will need to continue to make software and hardware changes as we further develop our capabilities,” says Harrison. “Although there are certainly technical challenges ahead, the aim is for digital twins to become routine.”

Harrison asserts that digital twins will be useful not only for process control in the operation of manufacturing facilities, but for the future of vaccine development. “When designing a vaccine, manufacturability is crucial. Having a digital twin to develop understanding of the manufacturing process is a key benefit,” he explains.

Articles

Modeling the manufacturing process benefits both development and operations. 

GSK Pilots Digital Twin for Vaccine Manufacturing 

Traditionally, compounding facilities—also called 503B outsourcing facilities in the United States in reference to the section of the US regulation under which they are registered (1)—have conducted drug manufacturing operations manually; “on demand” production of a wide range of drugs needed by hospitals and pharmacies as well as drugs on FDA’s drug shortages list did not lend itself to automation in the way that longer-run, scheduled production did. Robotic automation technologies available today, however, offer benefits of greater efficiency and higher quality even for small volumes.

“Robotics offer several significant advantages for aseptic compounding. Most notable is that the greatest source of contamination, the person, is removed from the process,” says Chris Smalley, an independent compounding pharmacist advisor. “There is a robust traceability to what is performed, including volume, drug barcode scan and more, and it does not matter what hour or day of the week the robot operates,” explains Smalley. He notes that, in some cases, a robot is set up to operate on its own overnight (i.e., “lights-out” manufacturing), and the completed products are unloaded and checked the next day.

“Minimizing risks associated with aseptic processing is crucial when working with compounding systems,” agrees Randy Fraatz, vice-president of North American Operations at Steriline, which produces aseptic manufacturing equipment. “The primary benefit of robotics and automation in general is safety, coming from the reduction of human-related mistakes in the entire process.” Other benefits, says Fraatz, include the accuracy, efficiency, and reliability that result from automation. “If errors are reduced, process reliability and quality improve,” he explains. 

Despite these advantages, uptake of robotics is slow, and many 503B outsourcing facilities continue to have technicians working in laminar flow hoods or biological safety cabinets to handle beakers and flasks for solution compounding and fill vials manually using syringes, notes Smalley. He points out that some large teaching hospitals have adopted robotics for large-volume parenterals and for oncologic compounding and filling. “Smaller hospitals that focus on a specialty, such as orthopedics or oncology, appear to be the second wave,” Smalley adds.

An early example of a hospital compounding facility that adopted automation was the overall winner of the International Society for Pharmaceutical Engineering (ISPE) Facility of the Year Awards in 2019 (2). The Kantonsapotheke Zurich (KAZ) supplies oral, dermal, and parenteral formulations to the Canton of Zurich hospital system under current good manufacturing practice (CGMP) conditions. Exyte, the engineering firm for the project, said in the award announcement: “This facility raised the bar for quality performance to CGMP levels, which had never been done before in a hospital compounding facility” (2).

Although robotic systems have many advantages, one of the challenges is designing and programming for flexible operations. Since it takes time to set up the systems to manage different material combinations, speed can be limited. “For a compounding combination that requires a lot of flexibility, such as handling a wide range of materials to prepare one compound, the system will need more time to process materials,” says Fraatz. “In many cases, humans are typically able to handle different types of materials and combinations faster than robotics, but contamination risks are higher.” He explains that planning ahead is important for using robotics, so that the right tools, equipment, and programming instructions are ready for a specific process.

“The user requirements need to be clearly and completely defined,” agrees Smalley. “Robotics can only work with the tools that they are constructed with, for instance vacuum-assisted suction cups, pincers, barcode scanners, syringe operators, and the like. Additionally, they can only perform tasks that were included in their programming. Refitting and reprogramming can be problematic.”

Another challenge is the need to have barcodes that can be identified by the automated system. “Not all vials have barcodes, and new products or new brands of generic drugs result in barcodes not in the robot’s database. Software updates are constantly needed,” notes Smalley.

New developments are focusing on improving software and improving material handling. “Changes are being made to software to enable the robot to operate more smoothly and efficiently, as well as to make the user interface more intuitive. For material handling, some robotics manufacturers are focusing on loading cells, while others are looking at the ‘hands,’ as it were, trying to make the way the robot grasps items more securely without the danger of breaking the item,” reports Smalley.

A best practice for robotics users is periodic service by the equipment manufacturer that includes maintenance and adjustment, as well as software updates to address operational and security needs, notes Smalley.

“Routine maintenance and attention are absolutely required for robotic automation in the compounding space,” agrees Fraatz. In his experience, various hospitals tend to be familiar with using high-tech equipment such as robotics (as in some surgical settings), but the overall compounding industry is not yet as aware of how process automation can be used in compounding environments.

In particular, integrating robotics inside isolators is a relatively new technology for the compounding environment, says Fraatz. “True isolators have become very common in the pharmaceutical processing industry, and are the safest containment solution available for the compounding market as well,” he says. “From a best practice perspective, compounding organizations need to evaluate and adjust their quality risk management philosophy, including process and validation protocols for this type of automation. Evaluating [the use of isolators and robotics] is a multi-department effort, as there are also facility considerations in terms of what they may need, concerning cleanroom design or availability of utilities, for example.” Fraatz notes that it is also important to evaluate the process flow, including how materials are introduced to the compounding environment and handled after compounding. He points out that robotic equipment can be designed to be compatible with cleanroom conditions. “Cleanroom conditions require easy to clean tools, low particulate generation, and equipment designed to respect the clean airflow required,” notes Fraatz. 

A pharmaceutical manufacturing company adopting cutting-edge robotic technology for its CGMP 503B operations is Nephron Pharmaceuticals in South Carolina. The company, which specializes in producing generic respiratory medications using a fully automated process with blow-fill-seal (BFS) technology, launched its division for sterile compounded drugs in 2017 and began with manual operations, in which pharmacy technicians worked inside laminar-flow hoods to fill parenteral solutions coming from sterile filtration into intravenous (IV) bags or syringes. Now, Nephron is moving to robotic systems inside of the laminar-flow hood to perform these fill/finish operations. The company worked with the University of South Carolina (UofSC) and Clemson University in two separate projects to custom design robots for this application. In April 2021, Nephron validated the UofSC system and began commercial production, says Lou Kennedy, CEO of Nephron. A second robot is already being built, and more are planned. “We’re producing drugs on FDA’s drug shortages list, and this 503B space is growing,” she says. Robotic systems will improve productivity, reduce the burden of repetitive physical work for operators, and provide better accuracy and precision. Future projects will seek to increase speed of the robots to obtain higher throughput.

In addition to being used in commercial production at Nephron, Kennedy would like to see the robotic systems licensed to hospital compounding facilities. “What I love about both the Clemson and UofSC projects is that as we collaborate with both undergraduate and graduate students, we’re helping develop future industry employees,” says Kennedy.

Although a common fear is that robots will eliminate people’s jobs, Kennedy says this concern is unfounded. “Robotics is not eliminating people; it’s teaching people to have new skills to operate automation,” she notes. “Operators need to learn how to service the robot and work with it, to solve a jam or replenish components, for example. Operators are also needed for quality functions.”

Kennedy notes that Nephron sees a need for more pharmacy technicians, and the company works closely with local schools to help develop the future workforce. For example, Nephron built a sterile compounding lab on the nearby UofSC campus that is used to train pharmacy students at the university as well as two-year students from the local technical college in using robotics in sterile compounding.

Both developing an understanding of how robotics works and specific training with the equipment and its functions are key, adds Fraatz. “It is a new approach, which means trust needs to be gained, starting from education and understanding, so people can appreciate the purpose of robotic automation. Once they believe in the purpose, they can grow their familiarity and comfort with it.”

1. FDA, “Information for Outsourcing Facilities,” 

2. Exyte, “ISPE’s 2019 Facility of the Year Award for Operational Excellence & Facility of the Year Overall Award won by partners Exyte and Kantonsapotheke Zurich,” Press Release, Oct. 30, 2019.

Sterile manufacturing of small batches, in facilities including hospital compounding pharmacies and 503b outsourcing facilities, benefits from automation.

Robotic Automation Boosts Efficiency and Quality in Drug Compounding

When running manufacturing equipment continuously, rather than in batch mode, operators should consider what cleaning practices need to be adjusted. FDA’s draft guidance for continuous manufacturing of small-molecule, solid oral drug products notes time between equipment cleanings can depend on a variety of factors, such as running time or amount of product (1).

When considering equipment running continuously, cleaning might be automated with clean-in-place (CIP) elements or involve full disassembly with manual cleaning, say Paul Lopolito, senior manager, and Beth Kroeger, senior manager, in Technical Services at STERIS. “Using cleaning agents or cleaning tools requires a cleaning validation to demonstrate removal of these elements to acceptable limits. The calculation of accepted limits may utilize traditional uniform carry-over models or non-uniform residue or stratified residue models,” say Lopolito and Kroeger. These models are used because residue can become concentrated as it moves through the connected equipment (2). Other considerations with continuous manufacturing are addressing microbial issues and process intermediate degradant residue. “These residues may present a cross-contamination risk to the next lot or batch of product. If these hazards exist, then it is warranted to perform the appropriate level of cleaning and cleaning validation to mitigate the risk,” they conclude.

Most lines for continuous manufacturing of solid-dosage drugs today are cleaned in a “clean-out-of-place” mode, but a complete CIP solution would improve cleaning turnaround times, notes José Luís Santos, director of Hovione’s Continuous Tableting Center of Excellence, who suggests that end-users would need to work closely with equipment vendors to develop such a system for a full process train. Hovione’s contract development and manufacturing facility in New Jersey has been running continuous solid-dosage drug manufacturing equipment for a few years and working to streamline the manual cleaning process.

“The magnitude of the task of changing over a continuous manufacturing rig from one product to the next is very large,” explains Santos. “From a unit operation standpoint, there are no major differences from batch equipment, and in most cases the equipment is exactly the same at the unit operation level. The differences between continuous and batch have to do with the transition sections in between [the integrated] unit operations. Depending on the actual setup of the continuous rig, these transitions can be comprised of large pipe sections, in some cases with pass-through connections between floors. Also, such transitions might comprise large number of PAT instruments to measure, for example, powder level or quality attributes of the material being processed. Thus, continuous rigs have additional parts to be cleaned. If the continuous manufacturing line is entirely ‘clean-out-of-place,’ the extra equipment also poses the added challenge of keeping track of many equipment components of all different sizes as they move through the cleaning operation and subsequent reassembly; the learning curve associated with these operations may be much longer than comparable operations of individual batch manufacturing units.”

Santos notes that, “While in batch, each unit operation is operated independently, in separate rooms, and typically staggered in time; in continuous, the full set of equipment is used during manufacturing, typically with higher asset utilization. Hence, from a planning standpoint, the cleaning of continuous rigs requires significantly more resources, effort, and cleaning capacity (e.g., additional wash rooms and footprint for parts staging and storage) to address the full set of equipment without impacting productivity of the area or overall equipment effectiveness (OEE).”

A cleaning best practice applied at Hovione was to allocate enough resources to address the manual cleaning process—including a large team of operators and enough space to do the cleaning—and then to optimize with shop-floor operational excellence tools, says Santos. “In our experience, the use of Lean [management tools] brought not just the acceleration of the operation, but also an increase of the comfort levels of the team members involved with the cleaning. An otherwise huge challenge could be decomposed into smaller, more manageable, blocks of work, with a clear visibility of how the work was progressing during each day of the operation,” he explains.

Another best practice is to maintain control of the organization of equipment components from disassembly through assembly. “For example, use specific bins to contain disassembled components from specific (predefined) sections of the line so that those components, which make up those specific line segments, stay together throughout the cleaning process. Organization is critical to reduce lost and mixed-up equipment components among thousands of such components,” Santos explains.

In biopharmaceutical manufacturing, process intensification can change the way the equipment is used and thus affects cleaning methods. Beth Kroeger and Paul Lopolito, senior managers for Technical Services at STERIS, shared some points to consider in an interview with Pharmaceutical Technology. 

Click to read: “Considerations for Cleaning Biopharmaceutical Process Equipment”.

Process analytical technology (PAT) sensors in the equipment are a crucial part of continuous manufacturing systems, but, in some cases, such as near infrared (NIR) probes, they may be fragile and require special handling during assembly and disassembly, notes Santos. He adds that it is important to use the PAT vendor’s procedures for proper cleaning and maintenance. “Having additional instruments to address concurrently with cleaning of the manufacturing equipment is logistically quite demanding, requiring close communication and planning in order to keep operations running efficiently. Developing and controlling standard procedures with the right level of details and mistake-proofing become even more critical in the context of preventing damage to such sensitive components during handling and cleaning.”

“When cleaning equipment with internal sensors, consideration should be given to the material of construction to ensure compatibility with the chosen cleaning agent. Typical substrates may include glass, titanium, or polymeric material,” note Lopolito and Kroeger. If using a CIP cleaning method, they recommend working with the PAT vendor to check compatibility to determine if there will be any impact to the sensors through chemical exposure, high-pressure steam, foaming, build-up of residue on the probes, or through any interaction of materials.

Another concern with sensors in a CIP process is determining how well the cleaning and rinse solution flows in and around the sensor and whether there is a significant change in the flow dynamics through the piping. “Coverage testing can be confirmed using riboflavin, and flow dynamics can be assessed through computer modelling, Reynold’s number calculations, or inspection with a borescope,” they explain.

It may be possible to use the existing PAT (which measures process variables when the process is running) to also monitor a CIP cleaning process, says Lopolito. “An example would be an ultraviolet (UV) or Fourier Transform Infrared (FTIR) spectroscopy sensor (to monitor drug active) that can also be used to detect trace levels of cleaning agent in rinse water and stop the rinse process when a target limit is achieved within a specified time,” he explains.

FTIR is also being investigated as an approach to cleaning verification, using a handheld instrument to detect and quantify surface contamination (3).

One of the challenges for manual cleaning is the difficulty of standardizing across a wide range of equipment components with different degrees of product exposure or adhesion, notes Santos. “New technologies such as handheld FTIR can certainly bring a level of simplicity to this process, either in terms of an in-process control to determine the endpoint of cleaning of a component or to eliminate dependence on analytical samples altogether,” he concurs.

Draft Guidance for Industry, Quality Considerations for Continuous Manufacturing

3. C. McSweeney et al., “New Approaches to Cleaning Verification for Pharmaceutical Manufacturing Using Handheld Mid-IR,” Presentation at IFPAC (online, March 2021).


Vol. 45, No. 6
June 2021
Pages: 30–32, 42

When referring to this article, please cite it as J. Markarian, “Cleaning Continuous Manufacturing Equipment ,” 

Consider best practices for manual or clean-in-place procedures.

Cleaning Continuous Manufacturing Equipment 

 (Sept. 10, 2020).
2. Exyte, “ISPE’s 2019 Facility of the Year Award for Operational Excellence & Facility of the Year Overall Award won by partners Exyte and Kantonsapotheke Zurich,” Press Release, Oct. 30, 2019.

When referring to this article, please cite it as J. Markarian, “Considering Robotics for Drug Compounding,” 

Automation offers benefits for sterile manufacturing in 503B outsourcing facilities. 