Continuous Manufacturing: Separating Hype from Reality

Published on: 
Pharmaceutical Technology, Pharmaceutical Technology-04-01-2017, Volume 2017 Supplement, Issue 2
Pages: s8-s11, s31

Continuous manufacturing will not work for all pharmaceuticals, but the right infrastructure, senior management support, and planning from the earliest stages of drug development could eventually allow up to 80-90% of small-molecule APIs to be made continuously, says Paul Sharratt, head of process science and modeling at Singapore’s Institute of Chemical and Engineering Sciences.

Continuous manufacturing has been around since the 18th century and is the mainstay of metal manufacturing, petrochemicals, food processing, and other process industries. For decades, some pharmaceutical processes, such as roller compaction, have been operated semi-continuously (1).

Within the past 10 years, however, through efforts that include work at leading universities; the Novartis/MIT Institute and its spinoff, Continuus Pharmaceuticals; and support from FDA and other regulatory agencies, continuous pharmaceutical manufacturing has begun to get more serious attention as a way to lower drug manufacturing costs and increase efficiencies. A turning point appeared to be reached in 2015, when FDA approved Vertex Pharmaceutical’s cystic fibrosis treatment, Orkambi (2), the first API manufactured via continuous processing.

In December 2015, FDA invited industry to comment on its first draft guidance document aiming to establish a modern manufacturing infrastructure, including continuous processing (3).  In 2016, FDA approved Janssen Pharmaceuticals’ switch from batch to continuous processing for manufacturing the HIV treatment, Prezista (4,5).  Together with the Biomedical Advanced Research and Development Authority (BARDA), the agency also funded more than $4 million for research into the use of continuous manufacturing within pharma (6).

Everyday economic realities

So far, however, companies that have made the switch from batch to continuous processing seem to be driven more by concerns about the future and a desire to innovate, rather than by practical, day-to-day economic realities. As observers have pointed out (7), given the existing manufacturing infrastructure, relatively few companies have a compelling reason to move headlong into continuous manufacturing.

Researchers at Singapore’s Agency for Science and Technology’s Institute of Chemical and Engineering Sciences (ICES) have been studying various aspects of continuous pharmaceutical processing, including the economics of switching from batch to continuous processing. In 2016, one team, working with GEA Pharma Systems (India), described a systematic process that process development teams could use to help them determine which processes could be most easily, and economically, shifted from batch to continuous mode (8).

Paul Sharrat, who is leading that research effort, shared insights from that work, and his opinions on the potential of continuous pharmaceutical processing, with Pharmaceutical Technology.

Is continuous processing in pharma being overhyped?

PharmTech: Do you think there has been too much ‘hype’ about continuous processing in pharma? Given the industry’s existing profit margins, costs for raw materials, and the way that manufacturing plants are set up, what percentage of small-molecule processes could reasonably be expected to be converted from batch to continuous?

Sharrat: For small-molecule API production, perhaps half of current processes could be converted from batch to continuous and made to work at least as well as the batch equivalent, and likely better.

The proportion of processes where there is a strong business case for making the change to continuous is likely to be much smaller. This is partly because a big financial benefit is needed to justify replacing and existing manufacturing capacity with a new continuous plant. Nevertheless, there are examples already in production where the swap was justifiable. If we look to the future and set out to deliver continuous options from day one of the development of any drug, then perhaps 80-90% continuous processes might be technically achievable, and most of those cases would be economically justified, too.

The hype might be seen as a problem and, certainly, plenty of exaggerated and even wild claims have been made for continuous manufacturing in pharma. However, I see [some of the hype] as inevitable, and even useful, to bring academic and innovation communities on board, to create awareness of the potential of continuous processing, and a broader movement for change. 

If we want [to establish] enthusiasm for change, then hype is hard to avoid. However, the strategic technical experts in many pharma companies know the subject well enough to be able to see through it all, and focus on reality.

PharmTech: Do you see a role for continuous processing of APIs in the future? What would need to change for this to be possible and practical?

Sharrat: Continuous is here to stay. I see arguments shifting from whether continuous should ever be considered [to] understanding when it will bring benefit. People are now thinking about how to embed continuous in process development and manufacturing activities.

Three things will need to happen to move adoption forward. First, senior management must be convinced that there is sufficient potential benefit, in order to want to build a capability. If we look across the big multinational pharmaceutical companies, that case has clearly been made at some level, and most Big Pharma companies have significant research efforts in place.

In the early stages, this research tends to be a matter of a few people getting used to how flow processing works. Then companies need to think about how to translate their own chemistries into flow and to see it working.

To make continuous a ‘normal’ way of doing things, the capability has to be able to take a process through to manufacture, so access to new engineering capabilities is needed.

Secondly, continuous options need to be readily available to technologists during process development. Since process development is carried out hand-in-hand with making materials for clinical trials, there is a barrier if the material used for trials is made by batch processing.

Today, we are seeing significant numbers of continuous processes being developed by big pharma companies. Assuming that these processes succeed in clinical trials, they are very likely to progress into continuous manufacturing.

Finally, continuous will have fully arrived when company infrastructures have adapted to leverage its potential business benefits fully. Current operating models in pharma are based on batch processing, and the infrastructure that supports batch isn’t well suited to continuous. This includes not only the manufacturing facilities but also staff skill sets, supply chains, accounting practices, and more.

Companies are only starting to figure out the best changes that will need to be made to promote adoption of continuous processing. It is likely that manufacturing facilities will be physically smaller with more compact utility systems, more and more modern control/data systems, and more modularity. In some cases, pilot plants and manufacturing facilities may become very similar or even merge.

How other aspects of the infrastructure will change is less clear, although change is inevitable if the full benefits of continuous processing are to be felt.




PharmTech: How long has your research group been looking into continuous processing?

Sharrat: I have been working on continuous with my team at ICES for nearly nine years, but started long before that while I was an academic at the University of Manchester Institute of Science and Technology in the United Kingdom. At ICES, we are already working on ‘second-generation’ technology, building on our earlier successes.

PharmTech: In your research, what findings surprised you most?

Sharrat: For me, the biggest surprise has been how broadly we have been able to apply continuous processing. For example, most people start tentatively with all-fluid systems and assume that, if there are solids, then batch is better. We decided to take on more challenging problems requiring solids handling and solids in flow. Despite the perceived risks, we were successful more often than not, and have expanded the range of processes for which flow could credibly be considered.

Time and cost of process evaluation

PharmTech: How easily can the approach that your team outlined be adopted by commercial companies, and how would that add to the drug development time frame or cost, or would it actually save time and money?

Sharrat: Our approach could readily be adopted either as is or more likely adapted to fit with other company workflows. We find the approach pretty intuitive and logical, so we believe that others could also readily learn how to do it.

PharmTech: Can you please describe, step by step, the general process that you devised for assessing reactions and equipment, and how you mapped out the processes?

Sharrat: We start by looking at the key processing characteristics of the system (e.g., rates of reaction, heat generation, number of phases, and mass transfer requirements). The data required needn’t be very precise but they help us to understand what the process needs from its equipment. For example, a fast, exothermic reaction will need intensive heat transfer if a serious exotherm is to be avoided. Typically, the information required comes from experiments, literature, and estimation methods. This allows a ‘sketch’ of the process to be developed, capturing the sequence and main processing requirements.

We use the Britest tool set for process representation (9) because we find it to be effective and intuitive, but other approaches could also be used.  The result of this first stage is an assessment of potential benefit and feasibility, and a decision on whether to proceed with the project.

The second step moves from the general to the specific. For each processing step, we look to identify the equipment that best matches the needs of the process. This is pretty much the opposite of traditional batch design, in which the process is modified and often downgraded to match the limited capabilities of a batch reactor.

This activity starts to turn the process into a ‘flowsheet’ that would be much more recognizable to chemical engineers. We also refine our estimate of overall process performance, leading up to a more accurate costing and process understanding to enable the decision to continue towards implementation.

The final stage is nothing more than a standard chemical process implementation project with detailed design, hazard assessment, and implementation.

PharmTech: How did you factor in the potential costs of moving to continuous processing?

Sharrat: The business case for a commercial company will depend on many factors--including the cost of making any changes. The critical issue is to have an idea of the feasibility and potential value early so the assessment can be made before a lot of effort is wasted. Our method allows a company to set out a high-level business case very early, in which the cost of making any changes can be accounted for.

PharmTech: Generally, were there any broad rules of thumb that you found regarding types of processes, and how well they would work in continuous mode?

Sharrat: In general, we found that it is easiest to convert processes that work in a single liquid phase only, and where reactions are relatively fast. Additional factors increase the complexity and difficulty, but additional fluid phases and heat transfer don’t cause much problem. When solids have to be processed, then that results in a step increase in difficulty, while difficult rheology (e.g., processes involving highly viscous or non-Newtonian fluids) and materials that are sticky or can foul equipment are almost always a serious challenge.  Another issue worth considering is the number of processing steps. A simple mix-heat-react-cool process would be much easier than a complex sequence of additions, reactions, temperature changes, and separations.



Cases where continuous approaches won’t work

PharmTech: Are there types of processes, or processes involving specific materials, that won’t work continuously?

Sharrat: There are some things that can be very difficult to accommodate in continuous processes, but very few absolute barriers. What tends to happen is that the cost of making ‘difficult’ systems work becomes prohibitive when compared with batch. Some things are indicative of likely problems (e.g., systems that deposit materials on surfaces are problematic). Reactions with multiple phases and very slow kinetics can be a lot more expensive in flow than batch. Continuous solid-liquid separation can be challenging, too.

PharmTech: Are there any plans to extend your research into additional problems/areas?

Sharrat: We are looking to extend the range of separation techniques that can comfortably be used at the small scale that is useful in pharma. Continuous separations are easier at larger scale (e.g., the large fractional distillation columns used in the petrochemical industry). Smaller-scale fractionation is challenging because performance can be affected much more by variables such as wall effects and heat loss.

PharmTech: What role do you see process intensification playing in pharmaceutical processes in the future?

Sharrat: Process intensification is a great concept, but it has been a struggle to have it be adopted widely, partly because plants that are smaller are not always the most cost-effective or relevant to the business need. I do think that we will see more intensive processes in pharma in future--with both individual plants and whole supply chains driving down size and inventory where it reduces cost and improves responsiveness. However, the driver will not simply be size reduction, but, rather, fitness for the business.

Our method actually provides a systematic approach to deliver intensified processes. Because we consider the economic viability throughout, we don’t run the risk of selecting an intensive but uneconomic design. In the end, systematic approaches to identifying good process options will prove the most effective route to intensification.


1. A. Pellek and P. Van Arnum, Pharmaceutical Technology, Vol. 32, (9), September 2008.
2. C. Challener, Pharmaceutical Technology, Vol. 40 (6), June 2016.
3. FDA Draft Guidance, Advancement of Emerging Technology Applications to Modernize the Pharmaceutical Manufacturing Base, Guidance for Industry,
4. “FDA Approves Tablet Production on Janssen Continuous Manufacturing Line,”, April 12, 2016, Accessed April 4, 2017.
5. Z. Brennan, “FDA Allows First Switch From Batch to Continuous Manufacturing for HIV Drug,”, April 12, 2016, accessed April 4, 2017.
6. “Continuus Pharmaceuticals Partners With FDA and BARDA on Continuous Processing,”, December 5, 2016, accessed April 4, 2017.
7. G. Malhotra, “Continuous Processing in Pharmaceutical Manufacturing: Considerations, Nuances, and Challenges,”, June 1, 2015, accessed April 4, 2017.
8. S.K. Teoh et al., Organic Process Research and Development, 20, 2016, pp. 414-431.
9. Britest website,

Article Details

Pharmaceutical Technology
Supplement: Solid Dosage Drug Development and Manufacturing
Vol. 41
April 2017
Pages: s8–s11, s31


When referring to this article, please cite it as A. Shanley, "Continuous Manufacturing: Separating Hype from Reality," Pharmaceutical Technology Solid Dosage Drug Development and Manufacturing Supplement (April 2017).