A Guide to QbD for APIs
All of us talk about using common sense for everything we do. However, in certain instances we do things that are quite contrary because we are in our comfort zone. Many times the world tells us that we did not use our “common sense” or pay attention to the feedback or there is a time lapse. When this happens, the world wants to rationalize our intentions and motives.
Even infants use their “common sense” to figure out the virtues of moving from crawling to walking to running in about one year. They understand the exhilaration and benefits of their newfound freedom. Similar euphoria is experienced when we have technology innovations or exceed financial goals.
In the manufacturing world common sense tells us that producing quality product the first time without interruptions (QbD) is more satisfying compared to the process where we have to stop and examine the progress at intermediate steps (QbA). It seems that the pharmaceutical manufacturing industry has considerable hesitation (1) in it’s ability to move from Qb (Quality by) A (Analysis) to D (Design). This observation is based on my review of discussions on different websites, in print and at conferences.
After more than five years of discussions we are still talking about how to get to “D”. This suggests that there are some reasons for not getting there. Pharma either is not able to overcome “B (bureaucracy) and “C (consternation) in its drive from “A” to “D” or has defiantly decided to live with analysis of raw materials, intermediates and final products to ensure we produce consistent quality products. Regulatory bodies have jumped into the fray by proposing methods of “how to” of QbD (2,3). They might be useful for new processes and products.
My question to all readers is how we are going to have a QbD process if we do not understand and apply fundamentals of chemistry and chemical engineering to our processes. Regulatory guidance is not going to deliver QbD. There are just too many guidances and they could complicate matters.
Had pharma moved from “A to D” we would be discussing its virtues, values and the amounts of monies saved and the reduction in compliance paper work. In defense of the pharmaceutical industry it has not lost “common sense” in its move from QbA to QbD. There are opportunities that have been missed and it is up to the technocrats to show their value. I believe that the industry is also victim of its own success.
We do need to understand and explore some of the hurdles that are in the way to move from “A” to “D”. These could be that we do not understand value of improved profits (most unlikely) or do not have financial justification (likely) or are reluctant to move out of our comfort zone to a simpler lifestyle of less work and regulation (likely). Actually the causes are a combination of all of the above plus more. They need to be discussed.
Much of the “how to” of the QbD discussion is geared towards formulations. APIs are excluded from the discussion. It is suggested that modeling, six sigma, design of experiments and other statistical tools along with Process Analytical Technologies (PAT) will facilitate “A” to “D” move. I believe that we are putting the cart before the horse or spinning wheels. These methods have value and will help “only” after we have a complete understanding and command of the interaction of chemistry, chemicals, process variables and unit operations.
Modeling without understanding of process fundamentals is time consuming, expensive and is not going to get us to “D”. Since significant data is needed for mathematical analysis, without process understanding we are creating a difficult situation for ourselves. If the modelers believe that volumes of data and its analysis, without understanding and control of the process variables, will deliver an economic and sustainable process that will produce quality product, which would not require repeated analysis of intermediates and final product and we will able to move the process from “A” to “D”, then they know something some of us do not know or understand.
In addition, there is a perception that various analytical instruments and tools will automatically deliver quality product. These tools tell us “the state of the process” but do not fix the process unless correct process control schemes are incorporated. Only our understanding along with proper use of these tools will create an optimum process. I think our misbeliefs have created a difficult situation. Simplicity and economic methods are needed to deliver quality products.
Some additional reasons for our reluctance to convert existing processes to QbD processes can be our fear of the unknown such as regulatory re-approval hurdles and associated costs or the profits are so high that everyone is comfortable with their current levels or, as said earlier,(1) we cannot justify innovation. Another reason for lack of QbD is a lack of economies of scale in our processes.
QbD is an opportunity. We need to take simple steps and have to have small wins to bring common sense back into our daily fold of process development, scale up and commercialization. Without wins we will be stuck in “analysis paralysis” mode and “A” to “D” movement will become the fad of the day and like any other fashion will disappear. Small wins are an inspiration for big wins.
The manufacture of a drug dose is a two-step process. The first step is the manufacture of the active pharmaceutical ingredient and the second step is its formulation to a dispensable dose. Since APIs are left out of pharmaceutical manufacturing conversations, discussion in this paper is focused on their manufacturing. The considerations are applicable to formulations also. Without a quality API and correct dose the drug would not deliver the desired effect. Under or over dosing would have detrimental effects.
Pharmaceutical current state
Business and manufacturing aspects - At branded (ethical) and generic pharmaceutical companies, we are always in a hurry to get the product to the market. Depending on the dosage, drug efficacy and the sale price API, their annual volume can vary from less than thousand pounds to few thousand pounds. Volume at a branded company during the patent life belongs to one company but it could be produced at multiple sites i.e. lower economies of scale.
Since the APIs are fine/specialty chemicals, it is natural for any chemist or chemical engineer to produce them in the same equipment where other similar chemicals are produced. The practice of producing APIs, due to similarity of chemistries and equipment needs, began at fine/specialty chemical plants more than hundred years ago. From a financial perspective this practice makes sense as it avoids investment for every product. Since process inefficiencies are absorbed through product pricing, businesses do not have to be concerned about low process yields or lack of having sustainable processes. Thus, the need for an optimum process has never been a pharma business concern i.e. having a QbD process has never been an issue.
Regulatory Aspects - Since the drugs are for human consumption, their quality and consistency is paramount. This need has created different regulatory bureaucracies. As explained earlier, different products due to their low volume are produced in the same equipment where other products are produced. Regulatory bodies in order to prevent cross contamination established “good manufacturing practices”. These guidelines are minimal at best.
All of the above necessitate that record keeping of every step including raw materials, intermediates, equipment cleaning and testing be done so that no mistakes happen and if they happen, they can be tracked and corrected. This has become a gargantuan task and more time is spent on cleaning the equipment and paperwork than having the correct and optimum process. QbA has become a way of life.
Due to lack of competition the pharmaceutical industry has not been proactive in developing innovative manufacturing technologies because the current practices have delivered the desired profits. By not having the optimum process and equipment that should be used for a designated API manufacturing, companies have relied on in-process analysis to replicate laboratory milestones. However, with efficient and optimized processes the industry can exceed these guidelines. It is very possible that better technologies can reduce the paperwork but this option has never been shared, if explored. Reduced paperwork will improve profits.
As patents expire, additional companies enter the business due to the potential for high profits. Entry of “me-too” companies (definition: those who produce many of the same or similar products as their competitors) will lower production volumes per site from the “pre-patent expire” levels. Most of these companies will be from the developing countries. Since these new entrants attain their expected profits, the need for efficient processes and manufacturing technology innovation is not a priority.
The financial investment for process innovation and improvements is difficult to justify at low production levels. This exacerbates the “A” to “D” move further. Thus the pharmaceutical industry, branded and generic, will lumber along with the QbA methodology unless the business philosophy changes or is threatened. This is discussed later.
To recap, the pharmaceutical industry, unless it makes changes in its business model, is not ready and equipped to implement QbD for existing and new products - especially the API manufacturing step. This is also true for the formulated products.
Suggested regulatory guidances/guidelines are not a gospel but an indication of the kind of data that needs to be generated by the companies to create a quality product. From web and print chatter I can see that without understanding and command of the process, such an effort will add costs and confusion. Since no one, including pharmaceutical companies and regulatory bodies know the costs and related benefits for existing products, I am not sure how many would venture to move from “A” to “D”. Simplicity is needed.
If we want QbD practices to become the pharmaceutical manufacturing norm, industry will have to change (1). If we do not change voluntarily, business circumstances might force the change. We have seen this in many industries and need to pay attention. Creative destruction (4) is needed for pharmaceutical companies as well as regulatory bodies. Having QbD processes is not difficult and/or expensive. It requires a different mindset and a revamp of our thinking.
QbD processes can be arrived at two different ways. For lack of better names for these methods I have labeled them “Me-Too” and “Consolidated” model. They are discussed below.
“Me-Too” companies have to rely on “square plug in a round hole (5)” as their manufacturing philosophy. The process is fitted in the existing equipment and most likely is a copy of their lab process that has been submitted for regulatory approval. These processes are generally inefficient, unsustainable and require monitoring of everything as described above.
If the companies want to stay the course of producing every product that is approved, there are two paths to QbD. First is to audit the processes, operating methods and production planning to see how the processes can be simplified to minimize QbA practices. The second path for QbD is to have equipment that is “right sized” for the process. Of course chemical engineering practices have to be applied in each case.
“Right sizing” requires investing in new equipment and commercialization of the optimum process. Manpower and financial investment along with re-approval of the products might be needed. It is very likely few companies will want to go this route, as they might not be able to justify such an investment for the existing products.
Modular equipment might help to reduce the investment but would require careful equipment selection, sizing and production planning. If not done right, companies would have even worse asset utilization compared to current conditions. Many companies may consider these alternates too challenging and complex and might stay with the current QbA methodology.
In a recent article (6) Figure 2 compares the costs of API manufacture in the West, India and China. A review suggests opportunities. Companies in all three regions purchase their raw materials from the same source. Thus their material costs are about the same. Companies in each region have opted for different margins. These margins can be significantly improved through QbD practices.
The reduction of conversion cost from 50 to 35% in the West through process simplification, which is not out the realm of possibilities, can significantly improve profit margins (7,8). If similar improvements are adopted in processes in India and China, their margins will also improve, making them more aggressive for further expansion in API and formulated products. Simplification will result in cycle time reduction, for example, production capacity will increase with no or minimal investment. I do not believe anyone would mind a “free” side benefit.
“Me-Too” companies can adopt QbD for their new products. They will have to think through these practices at the laboratory level and carry them to commercialization. The lessons of benefits can be extended to their existing products.
Companies adopting this or a similar model will be producing a limited number of products instead of many in the “Me-Too” model. Since the number of products is limited they will supply a higher volume percentage of the global needs. Manufacturing technologies for these products will be the most innovative and sustainable. These technologies will be simple to execute and will be QbD based. Some of the products in their basket could be produced continuously.
Companies in this model would have some of the following attributes for their API manufacturing and their formulation practices.
1. Have a technology that is the best, revolutionary and simple.
2. Have complete command of the process and require no or minimal in-process quality analysis.
3. Produce better than 50% or a very large percentage of the global needs for the API and possibly the formulated product.
4. Process equipment is designed for the product.
5. Dedicated product lines at a site to produce multiple products that are independent of each other.
Companies in this model will be creating procedures and manufacturing methods that will far exceed what is outlined in regulatory guidelines. They would be able to show case what is possible with excellent processes.
Due to economies of scale they will be purchasing their raw materials in significantly large quantities, and would be able to get price discounts to achieve higher profit margins. QbD processes will reduce their paperwork, as there will be significantly reduced changeovers. Reduced paperwork is extra profit. Regulatory compliance will be easier.
Compared to the “Me-Too” model the “Consolidated” model will result in better business practices resulting in minimal supply shortages, better cash flow, lower inventories, lower investment in total equipment and infrastructure and definitely less regulatory paper work. Due to better and sustainable technology this model will have a better conversion yield and a lower cost.
Irrespective of what path companies choose to attain QbD, companies will have to incorporate and exploit chemicals, their chemistry and chemical engineering principles to get there.(9) Companies will have to select options. They can apply QbD to new products (least expensive) or existing products (more expensive compared to new products). QbD in the long run will be a win-win.
Expenses to achieve QbD processes will be offset through savings that will come from efficient processes and reduced paperwork. If companies want to stay with QbA methods it will be an option of their choosing but then they also will have to live with the threat of being driven out of business at some point.
QbA --> QbD is an imagination and creativity challenge (10), which can be met if the companies want to. There are two unknowns in the “A” to “D” move. They are: 1) what would be the combined cost of each re-approval (pharmaceutical companies and regulatory) and would the savings offset the expense and 2) do the regulatory bodies have sufficient and knowledgable manpower to handle the paperwork and inspections. My conjecture is that no one has the answer and any improvement or change will have to be justified on a case by case basis (11).
1. Malhotra, Girish: Who and What Killed QbD, Pharmaceutical Processing, Volume 26, No. 4, May 2011, pg 10-14
2. Quality by Design for ANDAs: An Example for Modified Release Dosage Forms Accessed February 13, 2012
3. Process Validation: General Principles and Practices Accessed February 13, 2012
4. Malhotra, Girish: Pharmaceutical Manufacturing: Is it antithesis of creative destruction? http://www.pharmamanufacturing.com/articles/2008/091.html October 2008
5. Malhotra Girish: Square Plug In A Round Hole: Does This Scenario Exist in Pharmaceuticals? Chemical Week, August 25, 2010
6. Pollack, P., Bardot, A, and Dach, R., API Manufacturing: Facts and Fiction, Contract Pharma, January/February 2012 Vol. 14, No. 1 Pgs. 56-66
7. Malhotra, Girish: Chemical Process Simplification: Improving Productivity and Sustainability, John Wiley & Sons,ISBN: 978-0-470-48754-9, January 2011
8. Malhotra, Girish: Can the Combination of Creative Destruction and "Steve Jobs Traits" Lead to Pharma QbD Spring? Profitability through Simplicity (http://pharmachemicalscoatings.blogspot.com/2012/04/can-combination-of-creative-destruction.html), April 15, 2012
9. Malhotra, Girish: Focus on Physical Properties To Improve Processes, Chemical Engineering, Vol 119, No. 4, April 2012, Pg 63-66
10. Malhotra, Girish: QbA -->QbD: An imagination and creativity challenge, Chemica OGGI/ Chemistry TODAY March/April 2011, Vol. 29 Number 2, Pgs. 42-44
11. Malhotra, Girish: Neglected Tropical Disease (Infectious Diseases) Drugs: What are they telling us about Innovations! March 7, 2012