Economics of Perfusion Culture Media and Upstream Processing

Cell culture media and its development have been one of the most vital parameters influencing the performance and facilitating adoption of perfusion bioprocesses. The culture media used for perfusion can be used differently, usually at higher concentrations versus batch processing. The media may also be chemically and physically different.

Due to the recent growth in interest in perfusion, we prepared an economic analysis to assess the impact of culture media-related costs comparing perfusion versus batch processing. We estimated the relative costs of culture media used with perfusion versus comparable batch processing in single-use bioprocessing systems. Through interviews with bioprocessing decision-makers and managers, we identified the cost factors, and then defined typical data and ranges for variables that differentiate perfusion processes from comparable batch processes, with a focus on cell culture media-related costs.

Increasing Interest in Perfusion Technologies

According to BioPlan’s 17th Annual Report and Survey of Biopharmaceutical Manufacturing Capacity and Production (,1 there has been an increase in the percentages of respondents who are actively interested in exploring perfusion technologies in biomanufacturing. This year, we asked what novel bioprocessing systems/innovations are going to be evaluated or tested in facilities. Over 44% of respondents indicate “Upstream Continuous processing/perfusion” to be the leading systems to be evaluated or tested within the next year.

Although perfusion bioreactors have been in commercial use for biologics for decades, the recent resurgence in interest around these technologies is due to a variety of reasons. Christopher Brau, R&D Staff Engineer Scientist, at Thermo Fisher Scientific, provided a supplier’s perspective, “It is a combination of managing diverse portfolios, addressing downstream bottlenecking, as well as the need for more efficiency, such as automation and continuous processing.”

Developing an Economic Assessment

Our assessment of the economic factors impacting the perfusion versus batch decision making process was based on economic interviews with end-users and subject matter experts in August 2020. We sought to capture the factors that define the in-use economics. We then created an economic analysis model to support discussions regarding perfusion versus batch processing, including projecting culture media-related, and total upstream costs. Perfusion and comparable batch upstream processing costs are highly variable, and defining explicit parameters applicable across multiple facilities, scales, and applications is difficult.

The model created does not consider ‘indirect’ costs potentially associated with cell culture media. These costs can vary greatly and can be allocated differently, depending on the facility/company. In fact, many do not include ‘indirect,’ non-bioprocessing-specific costs. Others, however, do include various indirect costs as part of their cost analysis. Indirect costs can include those associated with general staff, facility, and company overhead and support expenses, general utilities, HVAC, QC/QA, regulatory support, IT support, costs of space/real estate, maintenance, management oversight, etc. Indirect costs are considered in the economics model as a percent of total direct costs related to media, much as overhead is calculated. Because many of these costs can be equivalent in perfusion or batch processes, we excluded these.

Selected Data—Novel Bioprocessing Systems/Innovations to Evaluate in Next 12 Months

We focused the comparative analysis on perfusion because users of perfusion technologies have increasingly expressed concerns and confusion about the costs of consumables, especially media.

Decision-makers considering a batch versus perfusion platform assessing consumables costs need to be careful to look at the big picture, and not just the media cost per gram of product, since these will be higher for perfusion versus fed-batch.

However, according to Brau, the potential of total cost of goods reduction per gram product and much lower capital expenditure can greatly favor perfusion in many new process cases. Brau notes that, “Even in current hybrid approaches with large stainless-steel fed-batch production there is a practical case to be made for N-1 perfusion in the seed train.” Although not all cell clones are well suited to continuous production, from a media consumption perspective, many manufacturing processes can benefit.

Economic Model of Costs: Perfusion vs. Comparable Batch Processes

Challenges in Batch Processing Resolved by Perfusion

Perfusion can address problems with toxic and unstable products because the constant product removal leads to a better system. Similarly, the downstream bottlenecks are reduced because the system can be designed for a more constant, smaller process, rather than having to handle a single, large protein harvest all at once. Less obvious challenges are also resolved such as permitting more flexible closed systems. Perfusion allows production reactors to shrink to more single use friendly sizes (e.g., under 1000L). This gives the production floor more flexibility. Some operations require a single-story work space; further, the smaller scale-up needed to produce relatively large amounts of product may allow for easier clinical material generation with less up-front investment. In addition, as the industry moves toward more automation, a steady-state process is generally easier to automate than a fed-batch run. This can also result in lower potential labor costs, and lower human-factor risks.

From a media consumption perspective, however, perfusion platforms can require planning for the logistics of larger volumes of medium for the process. Brau notes that, “Estimating medium exchange rates to support your cells means you will initially be running in a more wasteful manner until you narrow down the actual process limits.” Other challenges with perfusion involve cells lines with relatively poor production stability once selective pressure is removed, or with cell lines that are not robust and require a high bleed (a lot of cell and product removal) to sustain acceptable target cell % viability for the process. “Either of those can really eat away at your cost effectiveness,” says Brau.

Adoption of Perfusion Not Exclusively Cost-Driven

These direct cost-savings are often not the primary reason for adoption of perfusion. Besides some more sensitive APIs/proteins requiring the gentler bioprocessing associated with perfusion, perfusion provides some difficult-to-quantity cost savings. These are primarily related to the bioreactor and process lines being smaller. Perfusion allows single-use manufacturing, e.g., with repeated cycles with a 500 L perfusion single-use bioreactor being comparable over time to repeated use of a ≥4x larger batch mode bioreactor, with bioreactors over 2,000 L almost always stainless steel.

Subscribe to our e-Newsletters
Stay up to date with the latest news, articles, and events. Plus, get special offers
from American Pharmaceutical Review – all delivered right to your inbox! Sign up now!

Thus, perfusion enables commercial scale manufacturing (above the limits of batch 2,000 L bioreactors) using single-use systems, with single-use now preferred over stainless steel by many end-users, facilities, and companies. Also costs for process line and facility space, cleanroom sizes, utilities, plumbing, and particularly up-front and long-term capital expenses will generally be significantly lower with perfusion versus comparable batch processing. The lower processing intensity and smaller scales with perfusion can also allow reductions in bioprocessing staff. Perfusion, as noted, is also considered a better option for automation and related cost savings versus batch processing.

One particular advantage of producing a product in a perfusion process is the achievement of high(er)-quality protein. The continuous removal of toxic wastes and spent media from the cell culture and replenishment with fresh culture medium significantly enhances the quality of the final product. For example, fewer proteins with altered structures, such as glycosylation, can be attained.

Costs of Perfusion

The cost of perfusion culture medium per gram of drug substance/API in perfusion is typically 20 – 30% higher than that of a comparable batch process as more media is used per gram of drug product produced, with perfusion media also generally more concentrated versus batch.

Perfusion systems currently in use rarely exceed 500 L, which is also effectively the current upper limit for single-use perfusion systems. Very few facilities are at over 500 L scale, such as with the largescale manufacture of Factor VIII and other bioreactor condition-sensitive hemophilic factors, which have long used perfusion for commercial manufacturing.

The major costs with perfusion processing are culture media related. This includes the costs of the media, WFI and labor involved. Often at a comparable level, the major costs with batch processing are the single-use equipment and other consumables costs.

In terms of productivity, the cost/gram, the costs with perfusion are generally higher versus ‘comparable’ batch processing (making the same amount per period, such as annually). For example, a typical perfusion process at 250 L scale was estimated to result in total upstream costs of ~$50/g, while comparable batch processing was estimated to have total upstream costs of ~$38/g or 24% lower.

Cell Culture and Proteins

Depending on the cell line characteristics and protein being produced, dramatically higher quantities of protein can be obtained from a perfusion process that is stably run over a period of time (versus a ‘comparable,’ making the same amount in a time period, e.g., per year, batch process). Cell densities reaching 80 -120 million cells per milliliter can be obtained in perfusion compared to 20 – 30 million cells per milliliter in a typical batch process. And much more active agent can usually be manufactured per process run, but with perfusion processes run significantly longer than batch processes (and also using more culture media).

In a typical perfusion process, many volumes-worth of cell culture media are required over time be passed through the perfusion reactor to achieve a constant cell steady state that will be able to yield high volumetric productivity and optimum quality protein product. For a successful steady state perfusion process, it is required to determine the ideal cell line-based specific perfusion rate (CSPR) or perfusion rate (vessel volume per day, VVD) along with the bleed rate, i.e., the rate at which the cells are removed from the process stream.

An optimally stable cell line that can withstand long durations of the perfusion process is an essential component for perfusion processing. The growth rate of a perfusion cell line should be such that the bleed rate and CSPR do not exceed or work against the economic viability of the process. Therefore, a perfusion process is considered optimal when the cell density is high enough to achieve a high productivity, but the cells remain in a state where growth is controlled either by nutrient limitation or other environmental factors to minimize the bleed rate.

In our research, interviewees reported that to achieve yields similar to a perfusion process, ‘comparable’ (in terms of Kg output over time, such as per year) batch/fed-batch bioreactor sizes generally need to be 3X bioreactor. Currently, the sizes of most batch bioreactors are ~5X that of the perfusion used for commercial scale manufacturing and are larger than 2,000 L and can go up to 25,000L. These batch bioreactors at >2,000 L are generally stainless steel. Note, this study and model were restricted to use of single-use equipment, effectively limiting perfusion and batch bioreactor sizes to ≥2,000 L (although higher volumes can be input in the economics model; and model users can also calculate costs for scaling-out, running multiple bioreactors or process lines in parallel).

Media Volume and Preparation

In comparison to a batch process, wherein media volume is roughly equivalent to the bioreactor volume, a comparable perfusion process requires media over time in multiples of the bioreactor volume. For instance, if a 100L perfusion process runs for 30 days at 1VVD, the volume of media it will require is = (100 L) X (30 media exchanges) = 3,000 L versus 500 L media in a ‘comparable’ batch bioreactor process (here not considering for the dissolution factors or concentration of media dissolved per L). This significantly increases the costs of media in perfusion versus comparable batch processing. But the cost per gram (or Kg) of active agent and particularly overall processing-related costs are often relatively comparable or cheaper with perfusion versus batch processing. This is usually attributable to the smaller-scale (less expensive) bioreactors and facilities being used.

Frequency of culture media preparation for both perfusion and batch processing varies greatly among facilities. Multiple interviewees stated that media preparation for perfusion processing is repeated often enough such that it is viewed as a rather “continuous process.” Many end-users prepare batches of media days in advance of when needed and store them in bags. Most prefer to prepare media every two-three days to save on storage space. Some may even do it daily to support perfusion. Generally, with batch processing (and by definition), culture media preparation is only done once.

Storage of prepared media is among the main challenges for perfusion users (and comparably, storage of perfusates can be a challenge too). Facilities need to have large media holding tanks (beside large tanks for holding perfusates). For users storing media in storage bags or vessels, dedicated refrigerated rooms, freezers or other cooled GMP-grade space are needed. Despite the repeated work involved with culture media preparation, particularly for perfusion, most end-users prepare media as needed versus longer-term storage.

In terms of labor, nearly 70% interviewees reported that typically 1 – 2 staff work in 8 hour shifts to prepare the required media for perfusion process. They also concurred that the time needed (man-hours) to prepare the increased amount of perfusion media versus batch is at least three times (3X) higher in perfusion than a comparable batch process. Typical man-hours reported for a comparable batch media preparation range from 1 – 2 staff working in 3 – 5 hours shifts.

Concentrated Media Usage

To reduce the costs associated with the large amounts of media consumed in perfusion and to better support perfusion productivity, many perfusion users prefer to use concentrated media for cell culture. Employment of concentrated media saves storage space and costs as less volume of media needs to be prepared. However, the use of highly concentrated media can become challenging due to solubility or stability constraints of some media components. Decoupled media in the form of feeds has been traditionally used in fed-batch cultures and is finding a similar use in some perfusion processes as well. Proprietary media targeted for perfusion use are now available that claim to result in high cell densities, low CSPR, and high volumetric productivities.


Perfusion processes presumably use much more consumables compared to comparable batch processes, primarily in the form of TFF/ATF filters, single use bags, more and larger media and perfusates containers, connectors, tubing, etc. are additional in a perfusion process. This increases the per run cost of a perfusion process. At least 70% respondents agreed perfusion generally requires approximately 2-3X more consumables than a comparable batch process.

Productivity of protein expression is the primary differentiating factor between perfusion and batch/fed-batch processes. Many publications and researchers have noted in their studies that perfusion processing yields much higher quantities of product as compared to a comparable fed-batch process (making the same amount over a period of time – but with the perfusion process taking longer). While volumetric productivities obtained from a typical fed-batch process may equal to 0.3 g/L/d (gram per Liter of medium per day), a steady-state perfusion process can result in 1 – 3 g/L/day of productivity. This observation was agreed with by more than 50% of interviewees.

Nearly all the respondents agreed that at very high volumes of operation, the perfusion process may not be the most economical, cheapest option. Extremely large volumes of media replacement significantly increase the materials costs and consumables costs; and at some point, as scales go up, larger stainless processing becomes more cost-beneficial, as fully expected. Batch stainless steel processing becomes more cost-effective, with lower COGs, at these larger scales.

Since the culture medium is a major cost-determining factor for any perfusion process, using a high quality, optimized culture medium that can provide optimal cell densities and improve product yields at low CSPR is highly recommended to reduce overall perfusion costs (in addition to contributing to improved product quality). That is, optimized perfusion media can presumably save even more than the model calculates.


  1. Langer, E.S., et al.,17th Annual Report and Survey of Biopharmaceutical Capacity and Manufacturing, BioPlan Associates, April 2020, 527 pages (see www.bioplanassociates. com/17th).
  2. “Continuous Bioprocessing and Perfusion: Wider Adoption Coming as Bioprocessing Matures,” Langer Rader, BioProcessing J., Spring 2014, Vol 13 (1) pg. 43-49.

Author Biographies

Ronald A. Rader is the Senior Director, Technical Research, BioPlan Associates. He has 35+ years’ experience as a biotechnology and pharmaceutical professional, particularly as a biopharmaceutical information specialist, analyst and publisher, and has been responsible for the Antiviral Agents Bulletin periodical; the Federal Bio-Technology Transfer Directory; BIOPHARMA: Biopharmaceutical Products in the U.S. and European Markets; and the Biosimilars/Biobetters Pipeline Directory. [email protected], +1 301-921-5979,

Eric S. Langer is the President and Managing Partner at BioPlan Associates, Inc., a biotechnology and life sciences marketing research and publishing firm established in Rockville, MD in 1989. He is editor of numerous studies, including “Biopharmaceutical Technology in China,” “Advances in Largescale Biopharmaceutical Manufacturing,” and many other industry reports. [email protected], +1 301-921-5979,

Cell Culture in Perfusion Processes
  • <<
  • >>