Emil W. Ciurczak, Contributing Editor02.06.23
When I was part of the US-FDS PAT sub-committee in 2002, Dr. Ajaz Hussain (FDA) brought in people from industries that were using in-process controls (e.g., the automotive industry) to run workshops, using their experience to address “stress-points” in “normal” GMP manufacturing. It was quite informative and helped form the basis of the FDA Guidances and ICH (International Conference on Harmonization) Q8, 9, 10, 11, etc. I was invited largely because of my work using NIRS for incoming raw material qualification, implemented at Sandoz in 1984.
My only point of dispute with Dr. Hussain was his suggestion to shut down an entire process line, upgrade each step with validated monitors, then validate the whole system on a product-by-product basis. Now, when you buy your first home, unless you were born into money, you start small and, as the family and your income(s) grow, you can move upward in size and cost. My suggestion was to pick a “low-hanging fruit” project, such as blend uniformity—first performed by Pfizer in late 1980s—or drying, both already using NIR monitors at some early-initiator companies. So, when you show management that your first project 1) works and 2) can be validated, and 3) it speeds up production and minimizes waste, and 4) and most important to the “bean-counters,” it saves money, allowing for a larger profit.
Then, and only then, you can ask for funds to add other monitors along the process stream, slowly building to a full PAT/QbD process line. In other words, you don’t have to buy/build all the monitors before you begin gathering PAT experience, it can be done gradually. Likewise, a company does not (normally) spend millions of dollars/euros for a full-sized, fully automated, Continuous Manufacturing (CM) suite. Just as we have numerous and efficient contract research organizations (CROs), there are a number of locations in the U.S. and Europe where they perform contract PAT/QbD R&D and a number that can generated a continuous manufacturing protocol for you. Figure 1 below shows a small unit at the U. of Eastern Finland, used to perform contract PAT/QbD development.
1. With difficulties in obtaining timely shipments from traditional raw material vendors (i.e., API, excipients, and packaging materials), validating new sources is far speedier if the process is allowed to vary—QbD sets product specs not rigid manufacturing steps. That means that, instead of a lengthy validation process to integrate a new vendor’s product into an existing MMF (set in stone), the process may be varied to use the new material to generate a product, already defined as “good.” That is, the final product meets established performance criteria.
2. Even if the original source becomes easily available again, any alternate source might be a) less expensive, b) closer, therefore, faster to deliver, and c) could help fulfill requests to “buy locally” instead of importing materials. Note: the FDA can more easily inspect a facility in-country than overseas, assuring GMP compliance at those facilities.
3. The costs involved with GMP production versus QbD production are strikingly different: OOS lots, alone, can be quite costly due to failure and destruction in-house or the need to have a recall, mandated by the Agency. For a generic product, a recall can be doubly expensive: loss of the income for that lot and potential loss of business, due to bad publicity.
1. Amounts of API(s) needed during Process Development studies are smaller and require less active. Generating a Design Space (DS), using a Design of Experiments (DoE) study requires “production-sized” batches. Under GMP, generating 20-30 lots of, at least, 500,000 tablets would use up large amounts of (potentially expensive) API and take weeks, involving much equipment, personnel, and waiting for cleaning validation results. A continuously manufactured “batch” is essentially the same for 1,000 or 1,000,000 units. This is important because, for a new product (or updating a legacy product) there is often not an abundance of API in existence.
2. Scale-up studies are not required. This is a twin benefit: a) the six or so months that are normally required for scale-up is obviated , since the R&D/clinical sized batches are merely “up-sized” by running the CM equipment as long as needed to generate the required number of units, and b) the ideal final specs already exist from the DoE process, so further development is not needed before commercialization.
3. CM improves both the quality and consistency of products. There is, by definition, a steady state of control because the automated process, continuously monitored, assures instant response to any variation. This results in a consistent, high-quality product, obviating OOS (out of specification) products, virtually eliminating any recalled product.
4. CM-based products can be expanded or slowed, responding rapidly to changes in market demand. This flexibility allows for a quick rapid response to an increase in product demand, reducing shortage risks or prevents over-stocking when sales slow.
5. You can reduce your total cost of supplies. Allowing for the process (under QbD) to use any GMP-generated raw materials, ensures that supply chain shortages do not occur, meaning you needn’t carry an oversized inventory of raw materials, fearing an outage of available product on the market.
6. In a similar manner, you can achieve a lowering of COGS (cost of goods sold). This COGS reduction results in increased speed to clinic or market and lowers the cost of delivering a higher quality product.
All these benefits are largely to flexible batch sizes and speed to market. With government mandated price limits (mainly for Medicare patients) in the U.S. and Europe, lowering the COGS must be the mode of increasing profits. Clearly, merely raising prices will not be the future for fiscal success in the future.
I may have touched on other cost savings in the past, but the COVID-19 shut-downs, followed by supply chain disasters and a shrinking labor pool (accompanied by increases in salaries) and utility costs (oil prices skyrocketing) have made “normal” GMP production more expensive and prone to outages. Moving to QbD and CM allows for some major cost savings in addition to those outlined above:
1. Smaller warehouse space needed. Using PAT/QbD from the arrival of raw materials to the departure of finished products, the needed warehouse space diminishes for two main reasons (both are based on the GMP “waiting for Godot, er, uh, waiting for QC results” paradigm): a) RM supplies are approved or failed, on-the-spot, so, they can proceed to the active portion of the warehouse or sent back to the supplier. This greatly diminishes needed warehouse space and, b) finished products are, in essence, either passed by PAT monitors (up to 100% of the units) in real time, so they do not need to be stored in a quarantine area, awaiting QC analysis results.
2. A CM unit has a much smaller footprint than discreet, individual pieces of equipment needed for GMP production. Far more equipment is needed for GMP production, if only because so much stands idle, waiting for cleaning or cleaning validation, CM units may be cleaned as assembled, so fewer units are required.
3. Operational/personnel reductions are huge. Switching to CM allows for smaller and fewer buildings which drastically lowers a) electricity costs, b) heating/HVAC costs, c) lower personnel requirements (we already have a smaller labor pool, so this relieves that problem) since moving materials, equipment, and cleaning and validating equipment is lowered, needing fewer operators. Add to that, a smaller overall footprint could lower property taxes and maintenance costs, too.
In future columns, I will discuss biopharma applications of PAT/QbD. Clearly, there is a greater requirement for sterility and the “API” production is often a large and complex process. The packaging/final dosage form, unlike a solid, small molecule, is less complex and the biochemical process becomes the major focus for the PAT/QbD R&D. We will address the heterogeneity of the process and what water does to the most used technology in CM: Near-Infrared Spectroscopy.
Emil W. Ciurczak has worked in the pharmaceutical industry since 1970 for companies that include Ciba-Geigy, Sandoz, Berlex, Merck, and Purdue Pharma, where he specialized in performing method development on most types of analytical equipment.
My only point of dispute with Dr. Hussain was his suggestion to shut down an entire process line, upgrade each step with validated monitors, then validate the whole system on a product-by-product basis. Now, when you buy your first home, unless you were born into money, you start small and, as the family and your income(s) grow, you can move upward in size and cost. My suggestion was to pick a “low-hanging fruit” project, such as blend uniformity—first performed by Pfizer in late 1980s—or drying, both already using NIR monitors at some early-initiator companies. So, when you show management that your first project 1) works and 2) can be validated, and 3) it speeds up production and minimizes waste, and 4) and most important to the “bean-counters,” it saves money, allowing for a larger profit.
Then, and only then, you can ask for funds to add other monitors along the process stream, slowly building to a full PAT/QbD process line. In other words, you don’t have to buy/build all the monitors before you begin gathering PAT experience, it can be done gradually. Likewise, a company does not (normally) spend millions of dollars/euros for a full-sized, fully automated, Continuous Manufacturing (CM) suite. Just as we have numerous and efficient contract research organizations (CROs), there are a number of locations in the U.S. and Europe where they perform contract PAT/QbD R&D and a number that can generated a continuous manufacturing protocol for you. Figure 1 below shows a small unit at the U. of Eastern Finland, used to perform contract PAT/QbD development.
Figure 1: A small unit at the U. of Eastern Finland used to perform contract PAT/QbD development.
Who Can/Should Use/Switch to PAT/QbD and, eventually, CM?
It is understood that a straight-forward contract producer cannot arbitrarily change the MMF of the client/patent-holder. However, considering the pandemic-inspired supply chain kerfuffle and recent legislation aimed at reducing drug costs to consumers, it is in everyone’s interest to gradually make the shift, for new and legacy products, from GMP to QbD. Some reasons are:1. With difficulties in obtaining timely shipments from traditional raw material vendors (i.e., API, excipients, and packaging materials), validating new sources is far speedier if the process is allowed to vary—QbD sets product specs not rigid manufacturing steps. That means that, instead of a lengthy validation process to integrate a new vendor’s product into an existing MMF (set in stone), the process may be varied to use the new material to generate a product, already defined as “good.” That is, the final product meets established performance criteria.
2. Even if the original source becomes easily available again, any alternate source might be a) less expensive, b) closer, therefore, faster to deliver, and c) could help fulfill requests to “buy locally” instead of importing materials. Note: the FDA can more easily inspect a facility in-country than overseas, assuring GMP compliance at those facilities.
3. The costs involved with GMP production versus QbD production are strikingly different: OOS lots, alone, can be quite costly due to failure and destruction in-house or the need to have a recall, mandated by the Agency. For a generic product, a recall can be doubly expensive: loss of the income for that lot and potential loss of business, due to bad publicity.
Operational Cost Savings and Profitability
A few less obvious savings of CM may be overlooked. Obviously, the major selling point was control over the quality of the product. A few other time and cost savings are also other “selling points” that should be addressed:1. Amounts of API(s) needed during Process Development studies are smaller and require less active. Generating a Design Space (DS), using a Design of Experiments (DoE) study requires “production-sized” batches. Under GMP, generating 20-30 lots of, at least, 500,000 tablets would use up large amounts of (potentially expensive) API and take weeks, involving much equipment, personnel, and waiting for cleaning validation results. A continuously manufactured “batch” is essentially the same for 1,000 or 1,000,000 units. This is important because, for a new product (or updating a legacy product) there is often not an abundance of API in existence.
2. Scale-up studies are not required. This is a twin benefit: a) the six or so months that are normally required for scale-up is obviated , since the R&D/clinical sized batches are merely “up-sized” by running the CM equipment as long as needed to generate the required number of units, and b) the ideal final specs already exist from the DoE process, so further development is not needed before commercialization.
3. CM improves both the quality and consistency of products. There is, by definition, a steady state of control because the automated process, continuously monitored, assures instant response to any variation. This results in a consistent, high-quality product, obviating OOS (out of specification) products, virtually eliminating any recalled product.
4. CM-based products can be expanded or slowed, responding rapidly to changes in market demand. This flexibility allows for a quick rapid response to an increase in product demand, reducing shortage risks or prevents over-stocking when sales slow.
5. You can reduce your total cost of supplies. Allowing for the process (under QbD) to use any GMP-generated raw materials, ensures that supply chain shortages do not occur, meaning you needn’t carry an oversized inventory of raw materials, fearing an outage of available product on the market.
6. In a similar manner, you can achieve a lowering of COGS (cost of goods sold). This COGS reduction results in increased speed to clinic or market and lowers the cost of delivering a higher quality product.
All these benefits are largely to flexible batch sizes and speed to market. With government mandated price limits (mainly for Medicare patients) in the U.S. and Europe, lowering the COGS must be the mode of increasing profits. Clearly, merely raising prices will not be the future for fiscal success in the future.
I may have touched on other cost savings in the past, but the COVID-19 shut-downs, followed by supply chain disasters and a shrinking labor pool (accompanied by increases in salaries) and utility costs (oil prices skyrocketing) have made “normal” GMP production more expensive and prone to outages. Moving to QbD and CM allows for some major cost savings in addition to those outlined above:
1. Smaller warehouse space needed. Using PAT/QbD from the arrival of raw materials to the departure of finished products, the needed warehouse space diminishes for two main reasons (both are based on the GMP “waiting for Godot, er, uh, waiting for QC results” paradigm): a) RM supplies are approved or failed, on-the-spot, so, they can proceed to the active portion of the warehouse or sent back to the supplier. This greatly diminishes needed warehouse space and, b) finished products are, in essence, either passed by PAT monitors (up to 100% of the units) in real time, so they do not need to be stored in a quarantine area, awaiting QC analysis results.
2. A CM unit has a much smaller footprint than discreet, individual pieces of equipment needed for GMP production. Far more equipment is needed for GMP production, if only because so much stands idle, waiting for cleaning or cleaning validation, CM units may be cleaned as assembled, so fewer units are required.
3. Operational/personnel reductions are huge. Switching to CM allows for smaller and fewer buildings which drastically lowers a) electricity costs, b) heating/HVAC costs, c) lower personnel requirements (we already have a smaller labor pool, so this relieves that problem) since moving materials, equipment, and cleaning and validating equipment is lowered, needing fewer operators. Add to that, a smaller overall footprint could lower property taxes and maintenance costs, too.
In future columns, I will discuss biopharma applications of PAT/QbD. Clearly, there is a greater requirement for sterility and the “API” production is often a large and complex process. The packaging/final dosage form, unlike a solid, small molecule, is less complex and the biochemical process becomes the major focus for the PAT/QbD R&D. We will address the heterogeneity of the process and what water does to the most used technology in CM: Near-Infrared Spectroscopy.
