All opinions are my own.

Monday, November 4, 2019

Pharmaceutical Quality: Concepts, Misconceptions, Realities and Remedies

Pharmaceutical quality, especially for generic drugs, has been and is going through its cyclical ups and downs. Discussion heats up and then goes dormant until the next major quality issue appears. One would expect that every pharma manufacturer (API and their formulator) will be on their toes and prevent regulators from issuing 483 or equivalent citations and news reporters from writing about out of compliance issues. Unfortunately, oversights at companies keep occurring and are the press are forced to report this. 

There are many cases, but a few have been highlighted from Heparin (1, 2, 3), Ranbaxy 2005 (4, 5) and most recently Valsartan (6). We all know there are multiple other incidents and negativity rears its ugly head. Delays and lethargy of regulators compromises public health (7, 8). Pharma companies, it seems, are immune to every controversy. This leads sometimes to the feeling that their cheer is higher revenue and higher profits rather than higher and more consistent quality. Any negativity is considered par for the course and will come to pass with time. Their thinking “less than quality happens” and “patients die eventually so why worry” should not be a normal demeanor. Quality is good for the companies but if their product/s deviate out of desired specification range, unless caught, are seldom admitted, admitted after the fact, or with reluctance.

Quality of a pharma product generally follow Sine curve (9) with specification amplitude being between the upper and lower control limits. Developed countries set the limits for their drugs. It is necessary that companies adhere to these limits. Since these limits are tighter than the developing countries, many question tougher standards. Manufacturing would be greatly simplified there was one standard across the board. 

Figure 1: Product Quality Range

Product quality to most of us reflects a company’s integrity, intelligence, knowledge and ability to manufacture products that are the best in their class. The irony is cost to achieve first time quality is very low or nothing if done right the first time (10). Properly designed and executed manufacturing processes are supposed to deliver quality products. Recurrence of quality issues (deviation from established specifications, lack of data integrity and cGMP practices) especially in pharma have a common theme. They are reflection of shortfall in company’s integrity, management, knowledge and manufacturing practices. Every oversight can result in an issue. 

It is critical that we understand pharma’s manufacturing landscape as it effects product quality. Every company knows what is needed but the ultimate question is “do they produce repeat quality product that meets established specs using cGMP practices?” 

In drug manufacturing there are two components (API or formulations) and their sequential execution is necessary to produce a dose. Before we discuss manufacturing process preferences, it is helpful if we understand what is a batch and/or a continuous process. This review could be considered unnecessary by some but they have quality implications. They are elaborated. It is always good to re-visit the established definitions for different processes. There is no financial relationship with any company. 

Financial Model: 

Companies have to follow criterion that gives them acceptable financial return. It is also applicable to Pharma’s older cousin fine/specialty chemicals. However, Pharma’s profitability criterion, my conjecture, is based on how to maximize profits irrespective of the processes used. I believe that their criterion of maximizing revenue and profits is based on exploiting the emotional need of humans to extend life. Lack of affordable drugs for masses have forced companies to increase revenue and profits through year over year price increases. Pharma has not explored increased profitability through continued process improvements and my conjecture is that the regulators are the obstacle in this process. Cipla (11) did dent the established model some. In the long run it was a blip. Companies from the developed countries have made sure such perturbation does not happen again (12).

Again, pharma’s profit model is based on tradition of capitalizing on a financial opportunity rather than based on manufacturing excellence, my conjecture. The current practice may have been derived from initial limited and dire need for the brand drug. Once the drug is widely sold, tradition of multiple manufacturing sites (inefficient processes) similar to when the drug had limited distribution has continued. It seems value of alternate or efficient processes has not been explored. As explained later too many API producers and formulators could also be part of the current quality variations and problems. 

Process Selection:

It is necessary to understand the manufacturing process selection methodology. Reaction chemistry, formulation recipe, product demand and overall economics dictate process (batch or continuous) selection. This is not a new revelation. Selection processes in the chemical and petrochemical industry have been discussed and taught in chemical engineering curriculums as early as 1926 (too many references to cite). In the current pharma business model batch or continuous manufacturing should influence revenues and profits but they do not because the process selection criterion is not based on economics or technology. It is based on exploitation of patient emotions and need to please shareholders. 

Batch Process:

There is an established definition for batch process (13). Figure 2 is a simple schematic. It is noteworthy that in such processes it is critical and necessary to have storage space/tanks/vessels where the intermediate products are held before the next processing step. 

Figure 2: Schematic of a Batch Process :- INTERMEDIATES HELD FOR FURTHER PROCESSING

Batch processes are stop and go and many different processes and products can be fitted in the same equipment (14). They provide flexibility. Invariably intermediate products are tested to assure process is working as planned. If a processing step does not produce the desired quality product such intermediate holding tanks or other tanks can be used to rework the intermediate to produce specification product. One way or the other this practice increases product cost either through intermediate inventory related costs or re-work or disposal. Cash flow is impacted. All added costs are passed on to the customers. 

Continuous Process:

Continuous processes have an established definition (15). Figure 3 is a generic schematic of a continuous process. There are two distinct differences between a batch and continuous process. In continuous processes flow of materials does not stop during the operating year [24x7x50=8,400 hours] except for the necessary downtime for maintenance or planned shutdown etc. which generally is as short as possible for economic gain. Each selected process is based on the reaction or formulation chemistry/method and financial justification. In addition, each process is designed to produce a single product or an exactly similar product. There is no intermediate product hold tank/space any place in any continuous process. This is the most critical aspect of the process. Since the process is producing product every operating second, its quality cannot deviate outside the upper and lower quality limits at any instant. Feedback control loops that are well established and practiced maintain quality regimentation. Since the process is time independent, in-line testing gives instant image of the process. Deviation outside the control limits (Figure 2) means product is not meeting specs and will result in significant financial losses. Batch processes are a stop and go operation. 
Figure 3: Schematic of a Continuous Process :- NO INTERMEDIATE PRODUCT HOLD

Figure 4 is a schematic of a real continuous process. It operated about 7,500 hours per year producing same fine/specialty chemical. Again, we have to recognize that chemistries of API are similar to their older cousins fine/specialty chemicals. Only differentiation is API have disease curing value whereas fine/specialty chemicals don’t. 

In Figure 4 process chemicals A, B and C are introduced at predefined rate based on demand in a pipe flow reactor, reacted and continuously pumped to Reactors 1 & 2 where chemical D is introduced and reacted to produce the product. Chemicals A, B, C, D, Solvent and the catalyst are fed in to the system 24x7x50 hours per year. Stoichiometry was precisely controlled using existing process controllers. Product was withdrawn from the reaction system continuously. This process is one example of many continuous processes that have been commercial since early seventies. Thus, continuous processing/manufacturing is not a new technology but is being touted by many as NEW (FDA and others who are equipment suppliers rather than actual practitioners). 

Figure 4: A Continuous Flow Process Schematic

Batch Vs. Continuous Process:

It is necessary to understand why batch processes are the preferred processes for the manufacture of APIs and their formulations. Pharma, it seems, due to its own volition has never considered alternate manufacturing technologies. It is sad, but it seems it is due to its sustained profitability even in quality by analysis regimentation. 

Selected process, their method of execution and equipment do influence product quality unless an effort is made to have an exact replica or dedicated equipment. Even then, there can be a batch to batch variation at the same company. Understanding of equipment, raw materials and execution method influence quality. Judicious review of product demand is necessary. 

Improved quality is being touted from continuous process vs. batch processes. That is true if there is such a process for the manufacture of APIs and their formulations which would produce the same product about 7,500 hours per year. However, volume is needed to have a steady run from the same equipment. If the process is stop and go, it is no different from any batch process. In addition, significant understanding chemistry, component interaction and execution control is needed. This can present formidable and different challenges for API manufacturing and their formulations. If all was easy continuous manufacturing especially for formulations would have been adopted in pharma manufacturing 60+ years ago.
Brand drugs, due to their high prices, have limited demand for the API and its formulation. As the demand increases, additional API and formulation sites are used to meet the increasing demand (7). Once a drug becomes generic, many companies enter the landscape (Table 2). Each uses a batch process even if there is a high demand. This is discussed later. Tradition prevails. e.g. Johnson & Johnson’s McNeil lab when it re-did its Tylenol formulation plant, it chose batch production when it, due to its high product demand, could have easily used continuous process (16)

Drug dose and product demand determine API and formulation needs (17). Table 1 illustrates API and tablets needed to satisfy the needs of 50 million patients per year. If principles of economics and chemical engineering were applied, most likely needed API (dose 1 milligram) could be produced in a single plant. Formulation of this API could also be done at a single plant requiring multiple parallel formulation lines operating year-round. For 50 milligram dose a single plant using a continuous process would suffice the global API need. However, in reality multiple plants are used to produce the API and their formulations. This is an extremely important fact as quality from each (API and their formulations) facility can vary even if every plant was an exact replica of each other. This is due to myriad factors (people, raw materials, equipment and even execution). 

API, Kilo/year
Table 1: Theoretical API and Formulation needs 

Table 2 is an illustration of reality for some of the widely used selected drugs. Process economics should dictate process selection but that is not the reality. Fundamentals of engineering and economics are not applied for manufacturing selection guidance. It is interesting to note that Pfizer produced 200 tons of Atorvastatin API at three sites (8) but now it is being produced at 44 (17) sites and is being formulated at over 800 sites (different companies). Every product will not be exactly the same. We can all draw our own conclusions about batch to batch and site to site quality variations. Another interesting fact we have to recognize that FDA or another regulatory body does not have adequate manpower to do even risk-based inspections just for these six drugs, my conjecture. 

Based on number of sites (17) (Table 2) that are being used to produce the API and their formulations any experienced engineer/entrepreneur could conjecture that accepted principles of process selection are not being used API production and their formulations. Only explanation for large number of sites clearly suggests that the involved companies producing the API and their formulations are quite profitable even when they do not have the most economic processes. They will also have quality variability. One way to assure profits is taking short cuts whatever they might be. Marginal quality might be easier and cheaper to achieve than to comply with USFDA quality requirements. As long as quality is close enough, the product can be shipped to many countries. It is my conjecture that such practices can overflow to the developed country shipments. 

Number of API Sites
Number of FDF Sites
Atorvastatin Calcium
Metformin HCl
Table 2: Number of sites for APIs and Formulations 

Again, rationale for so many API and formulation producers is “PROFITS”. Reverse calculations illustrate the profit math. Ciprofloxacin, an excellent in-demand antibiotic is used as an illustration. Table 3 is quick back of the envelope analysis. Series of articles(18) further illustrate the point. It is interesting to note that Ciprofloxacin API used to sell the drug in India indicates API price of about $16.40 per kilo. Export price ranges between $26.00--$46.00 per kilo (19), a significant incentive and margin at API seller level which prompts many to produce this and other APIs. As indicated earlier these API producers most likely do not have economic processes and quality is tested in rather than built in with high probability of significant quality deviations. No one should be surprised if there are multiple quality levels of the same product at the same site. 

Ciprofloxacin, 500 mg tablet
$/tablet, exchange rate Rs. 69/$

Selling price in India

Selling price in India is used to reverse calculate API cost
API reverse calculated cost
85% is used as profit margin, seller margin, formulation profit margin and costs
0.0082 [=$16.4/Kg.]

US selling prices (20) vary from $0.28 to $4.77
Table 3: Reverse Calculation Illustration

Given the current profits landscape many companies will enter to fill the need. As stated earlier to assure profits, it is very likely that shortcuts will be taken and quality could be compromised. Many companies believe that about 80% of the population does not need drugs of USFDA standards. Their thinking is 98% is good enough then why 100% (USFDA standard) is needed. Such thinking is reflection of company’s integrity and ethics. Such thinking prevails in companies who do not have total command of their operations and will sooner than later be cited by regulators. 

Pharma companies have not explored increased profitability through continued process improvements and my conjecture is that the developed country regulations are the obstacle in this process. Too many filings are needed to comply with current regulations. Companies might not have the time to improve processes also. Thus, continuous process improvement effort is minimal at best. 

Possible Solution:

Ethical practices, integrity and quality products are expected from companies. Every company that produces health related products e.g. drugs, has to make sure that the products meet established quality specifications and follow cGMP practices (21). Data integrity must be paramount as well. Companies have to have complete command of their operations. Their ethics and integrity of companies are at stake when they produce any drug. If they cannot comply with such expectations, my conjecture is that they should need not be in this business. They have to ask themselves the question “would they or their next of kin consume the products they produce?” It is imperative that for quality they live by “Do or Do Not, There is no Try---Yoda”. It is incomprehensible when companies operate outside the specification limits they had committed and agreed to in the first place. 

Companies very well know that non-compliance with USFDA regulations will have negative consequences besides bad publicity. They will have to spend monies for remediation. Doing things right the first time has minimal financial costs but somehow, they miss this important fact. 

With USFDA being short staffed, my speculation is that even the “risk-based inspections” might not be sufficient to catch less than quality/cGMP producers. Some companies have and will figure out the system and take advantage thereby jeopardizing safety and the health of many. If that is the case, it may be necessary to change the prevailing regulatory landscape. 

Under the current regulations producing companies are given an opportunity to correct their non-compliance. Companies could be abusing this privilege. Repeat offence companies (7, 22) are still in business. Ranbaxy stayed in business for many years (8). Pharma lobby’s behavior, siding with less than quality manufacturers, should be considered shameful and unethical. 

FDA needs to change the rules of the game. It seems FDA/regulators are afraid of ensuing shortages that could result. Repeated non-compliance to FDA’s requirements and guidelines should be a cause to forbid shipments to the United States. It is my conjecture that API manufacturers and their formulators in China and India have taken advantage of lack of sufficient producers in the developed countries and some might have “take it or leave it” posturing. Developed country regulators have to take a tougher stance (8)With many API producers (17)of the same product, it is very likely that that each has an inefficient and the cheapest process with harmful impurities. Formulators and PBMs in order to maximize their profits are most likely purchasing the lowest cost API and formulations and might not be totally aware of the impurities and their consequences (22)

Combination of polite and drastic pathways need to be used by the regulators to convince manufacturers to abide by the current regulations and quality standards. Companies could be given single site exemption with the stipulation that they deposit $200,000 (refundable) after the first deviation from FDA’s expectations. Second offence on the same site should be the cause for forfeiture of the deposit and the company should be barred from exporting their product to the United States. I call this “one strike and you are out as in the game of cricket” a simple and clear strategy. 

FDA has taken bold and drastic steps to withdraw ANDA approvals in the past but they have not been publicized much (23,24,25,26). Recently FDA withdrew ANDAs approvals for Apotex (27), essentially stopping their exports to the US market. USFDA should withdraw ANDAs from the companies if they have not produced the product at the declared site for more than one year (28). They has no clue about if approved ANDAs are being produced. Basically, FDA has had a big stick it has not used properly. Maybe it is time to use it as often as possible. There could be some shortages, but they could become an opportunity for others. 

Companies, who have considered FDA to be lax, need to wake up and produce quality products. Their business future could be in jeopardy. Thoughts discussed above could be part of the going forward strategy. If promulgated, pharma companies that comply with USFDA standards will benefit from higher profits that will come from better technologies that will be the result of consolidation, economies of scale and competition. Regulators should facilitate “continuous improvements”. It is imperative that companies don’t abuse the trust and privilege. 

Going Forward:

Pharma landscape needs to be reviewed and changed to assure quality drugs from the get go. I believe regulators themselves are reluctant to adopt “continuous improvement” in their own operations which they expect companies to practice. They need to practice what they preach (28)FDA instead of preaching QbD (quality by design) needs to practice it. It relies of QbA (quality by analysis).

Current ANDA filing process is complex as most of the application involve multiple reviews and submissions taking as much as three years (29). FDA’s ANDA approval process should be such that any application filed by its own personnel should be approved by a fellow reviewer on the first review in 90 days (30, 31). If it cannot be, that suggests that the ANDA filing and approval process needs to be simplified and perfected. Complexity of the current process is confirmed by the recent Government Accountability Office (GAO) Report (29). Since certain drugs are given an expedited approval (32), it suggests possibilities of 90-day approval is possible, if an effort is made.

ANDA applications that can be reviewed and approved in 90 days will demand that companies filing such applications have a complete command of every facet of the manufacturing, product quality, labeling and whatever else is needed by USFDA. Many will say “this cannot be accomplished”. Such conjecture is unfathomable from successors of a generation that can send a human to the moon and bring him back. 

FY 2013
FY 2014
FY 2015
FY 2016
FY 2017
FY 2018
FY 2019 (May 2019)
    Table 4: ANDA Approvals and Withdrawn

Table 4 from US FDA (CDER and OGD) (33) illustrates ANDA approvals and withdrawals. A clarification on withdrawals (are these withdrawals by FDA only or companies only or a combination?) has been sought. My conjecture is that most of the withdrawals are by the company initiated. These numbers are not publicized. It is also interesting to note that number of ANDA approvals per fiscal year (Table 4) are significantly higher than the numbers published (31, 34). GAO (31) report suggests about 80 ANDAs are approved per year. Why not have 160 ANDA approvals per year? Such numbers are not out of the realm of reality if FDA can simplify the ANDA filing and approval process. It will take effort. There will be significant internal resistance. 

Approval of a product and its manufacturing process is equivalent to a binding contract between two parties (manufacturing company and FDA). If a company does not live up to its contract, FDA should exercise its power to withdraw the ANDAs. USFDA has a stealth weapon, withdrawing approved ANDAs, but has been reluctant to use it. Recently they have used it (27) and publicized it. It is an indication of wall handwriting to the companies to get their act together if they want to be player in the developed country markets. No one, including the legislators, should be surprised that ANDA withdrawal could become frequent to stop less than quality drugs being offered to the populous. EMA and other regulators have similar options. ANDA withdrawals could lead to temporary shortages but could be an opportunity also. 

US FDA should also refrain from telling/suggesting companies types of manufacturing processes (batch or continuous) companies need or practice. Regulators should focus on making sure companies have robust and repeatable manufacturing processes through continuous improvements. They, as suggested earlier, should facilitate the filing and approval process (31). Companies justify every investment and are responsible for their product and its quality. If they don’t know what is necessary to produce quality products and are reluctant to follow FDA guidelines that are necessary to export products to the regulated markets, they should not be in the business. Only best of the best should produce drugs. 

Possibility of companies losing their profitable markets will force them to stay on top of the product quality through economies of scale and better manufacturing technologies. If they succeed their revenues and profits will improve. In light of regulators withdrawing ANDAs or potentially imposing fines for repeat violations, companies have to re-look at their operations and technologies. Some might have to consider “does it pay them to be in the business”. 

Companies have choices to make but supplying drugs that do not meet established quality and regulatory standards should not be the choice. 

Girish Malhotra, PE
EPCOT International

1.     Heparin Scare: Deaths from Tainted Blood-Thinner Spur Race for Safe Replacement, Scientific American, November 2008 Accessed June 24, 2019 
2.     Chinese Heparin Contamination Questions Return With New FDA Warning Letter, Regulatory Focus
3.     Baxter recalling sub-potent heparin, FiercePharma, June 23, 2015, Accessed June 3, 2019
4.     US FDA blocks imports from Ranbaxy for poor quality, The Economic Times, September 17, 2008 Accessed June 20, 2019
5.     USFDA pulls 27 ANDA for drugs made at banned Ranbaxy plants August 12, 2012, Accessed June 20, 2019
6.     Search List of Recalled Angiotensin II Receptor Blockers (ARBs) including Valsartan, Losartan and Irbesartan, USFDA, June 18, 2019 Accessed June 24, 2019
7.     Apotex Research Private Limited, Warning Letter 320-18-69, USFDA, Accessed June 3, 2019
8.     Bottle of Lies, Katherine Eban, Accessed May 31, 2019 
9.     Sine Curve, Merriam-Webster, Accessed June 24, 2019
10.  Cost of Quality, Industry Week, June 19, 1995 Accessed June 20, 2019 
11.  Indian Company Offers to Supply AIDS Drugs at Low Cost in Africa, The New York Times, February 7, 2001, Accessed June 2, 2019
12.  Gilead to license Sovaldi to Indian generics drug firms, Livemint, Accessed July 8, 2019 
13.  Batch Production, http://bit.ly/31dzpo3http://bit.ly/31UiV4k  Accessed June 10, 2019 
14.  Malhotra, Girish: Square Plug In A Round Hole: Does This Scenario Exist in Pharmaceuticals? Profitability through Simplicity,August 10, 2010
15.  Continuous Production, Accessed June 24, 2019
16.  'Shocking' conditions at Tylenol plant, CNN Money, May 14, 2010
17.  Malhotra, Girish: Opportunities for Generic Pharma to Clear the Quality Stigma, Profitability through Simplicity, May 23, 2019 
18.  Malhotra, Girish: Profitability through Simplicity Accessed June 26, 2019
20.  Malhotra, Girish: Opportunities to Lower Drug Prices and Improve Affordability: From Creation (Manufacturing) to Consumption (Patient), Profitability through Simplicity, March 9, 2018
21.  Data Integrity and Compliance With Drug cGMP, FDA, December 2018, 1Accessed July 10, 2019 
23.  Strides Pharma Receives FDA Warning Letter and Defers 10 ANDAs, in-PharmaTechnologists.com, July 2, 2019, Accessed July 9, 2019 
27.  Apotex, Inc.; Withdrawal of Approval of 31 Abbreviated New Drug Applications, govinfo.gov, July 10, 2019, Accessed July 10, 2019
28.  Malhotra, Girish: Impact of Regulations, Drug Manufacturing and Pharma Supply Chain (PBMs and allies) in Drug Shortages and Affordability Part 1, Profitability through Simplicity, March 8, 2019 Accessed July 10, 2019
29.  GENERIC DRUG APPLICATIONS, Government Accountability Office, August 10, 2019 
30.  Malhotra, Girish: Can the Review and Approval Process for ANDA at USFDA be Reduced from Ten Months to Three Months?,Profitability through SimplicityMarch 25, 2017
31.  Malhotra, Girish: Simplified Roadmap for ANDA/NDA Submission and Approval will change Pharma Landscape, Profitability through Simplicity, November 25, 2018
32.  Fast Track, Breakthrough Therapy, Accelerated Approval, Priority Review, FDA, Accessed August 29, 2019
33.  Correspondence from Ms. Mary Dempsey, Associate Director for Regulatory Affairs, Office of Generic Drugs, USFDA, July 29, 2109, [Activities Report of the Generic Drugs Program (FY 2019) Monthly Performance | FDA]
34.  First Generic (ANDA) Drug Approvals, FDA, August 16, 2019, Accessed August 16, 2019

No comments:

Post a Comment