Why Fitting a Square Plug in a Round hole is Profitable for Pharma and Most Likely Will Stay?

In 2005 I had raised a question about Batch or a Continuous Process: A Choice. At that time it seemed like a logical question and still is. However, I left part of the question unanswered. Missing was the discussion of components of pharma manufacturing, API manufacture and their formulations. Generally most of the discussion about pharma manufacturing focuses on formulations. API manufacturing is treated as an orphan and is not discussed.

In this discussion I have used Synthroid, # 1 prescribed and # 59 by revenue drug from top 100 US drugs from 2013 to review both segments of manufacturing and related options. Information could be used to create a better business model that will incorporate better manufacturing technologies and move away from ‘regulation centricity” to “process centricity”. End result would be improved profits and expanded global healthcare and coverage with lower costs. If properly done global pharmaceutical landscape will change.  

API manufacturing and their formulations need to be dealt separately as the technologies involved (unit processes and operations) are different. However, they influence the total business process. First has reactive processes along with purification. Resulting products could be solid or liquid. Formulation in the simplest form is basically mixing of excipients and creating dose that delivers the expected performance and can be easily dispensed and consumed. Packaging is part of the formulation.

Why Continuous Manufacturing of API’s would be a challenge:

In the last few years many on increasing frequency have chimed in for the continuous processes for API. However, it seems like that the rational principles of chemical engineering have not been applied in coming to that conclusion. Continuous usually means operating 24 hours per day, seven days per week with infrequent maintenance shutdowns, such as semi-annual or annual. Generally 15% downtime is acceptable. Anything short of this definition is not a continuous process. Continuous process also means starting with raw materials and producing finished salable product.

The following dictate the rationality of what type of process would produce lowest cost and highest quality products.

• Product volume per year
• Process
• Equipment

In the development of a commercial process chemist/chemical engineer have to know and understand these. They can have the best process but equipment and product volume dictate the course of action. Generally the first thought is to use the existing equipment if the volume does not justify a continuous process. We all know and understand that a continuous process most of the times means capital investment.

For batch or continuous process complete command of the operating conditions and methods is necessary to produce repeatable quality product. Anything short impacts product quality and business process.

Batch cycle time exceeding e.g. 48 hours necessitates a thorough review and effort even going back to the lab bench to reduce the cycle time. Long batch processes impact asset utilization and the whole business process. Every effort needs to be made to minimize the batch cycle time. In pharma extended API manufacturing batch cycle times are normal as there is no “process centricity”. “Regulation centricity” rules and is an impediment to innovation.

Table 1 is self-explanatory and presents very interesting numbers. There are some extrapolating assumptions. It is assumed that 5% of the global population uses synthroid. This most likely is a high number. Thus the actual demand for the active ingredient would be less than illustrated.

Table 1 illustrates that at 112 microgram dose at 100% formulation efficiency about 15,000 kilograms of the active would be needed to satisfy the global demand. A continuous API plant operating [24X7X350x0.85 =7,140 hrs.] would produce at about five pounds per hour, an extremely low production rate for a continuous process. Use of currently available equipment would pose many challenges and be really trying to fit square plug in a round hole. If a continuous process plant were to be built, it will require special equipment and process controls that might not be available.

All said and done there is no justification to have a continuous plant for the manufacture of active Levothyroxine (synthroid). I have not looked into it but I am sure that today multiple plants are producing the active ingredient. Each possibly has low process yields, are inefficient, have variable site to site and batch to batch product quality. Significant and expensive manpower would be needed to have consistent quality product. In reality a single plant using a batch process would satisfy total global demand.

Synthroid (Levothyroxine) sales number and prescriptions [Table 2] presents another interesting hypothesis. We can reverse calculate the price of the active ingredient. It would be based on certain assumptions and would give us a picture of profitability at different levels. US sales per prescription per month are about $3.05. These compare to sale price of $4.00 for thirty or $10.00 for ninety day supply at Walmart and Target etc.

2013 Sales, $	2013 # prescriptions	Dosage
858,725,708	23,452,848 [One prescription per person]	There are eleven different doses between (25-200 microgram) available. To facilitate calculations an average dose of 112 micrograms has been used.
Since synthroid has to be taken every day of the year we can calculate the total micrograms needed assuming 100% formulation yield. One prescription = one patient
Total API needed, micrograms per yr.		=23,452,848x112x365= 958,752,426,240
One kilogram = 1,000,000,000 micrograms
Total API needed to serve US population, Kilograms per yr.		958.75
US population taking Synthroid		(23,452,848x100)/320,000,000* = 7% [*US population]
Extrapolating number to project global Synthroid API demand per year
Global population seven billion. Assumption 5% takes synthroid = 350,000,000
Total Synthroid global API need Kg. per year			=(350,000,000x958.75/23,452,848) = 14,308

Table 1: Levothyroxine active ingredient needed for Global and US population

Sales, $	# Prescriptions
858,725,708	23,452,848
Avg. US patient cost $ per month = 858,725,708/(23,452,848x12) =3.05
Avg. dose, micrograms =112	API, Kg needed to fill US need = 958.75 Per Table 1
Profit assumption at wholesale level	100%
Sale price at wholesale level, $	= 0.5x858725708 = 429,362,854
Formulation profit 40%
Formulation level factory cost	=0.6x 429362854 = 257,617,712
Excipient and conversion cost	70% of Factory cost
Total API Purchase price, $	= 0.3x257617712= 77,285,313
API Selling price, $ per kilo	= 77,285,313/958.75= 80,610

Table 2: Reverse calculation of Levothyroxine selling price

Reverse calculation using US sales numbers suggest that active ingredient Levothyroxine cost to the formulator should be about $80,610 per kilo. Current selling price of levothyroxine on the world market is less than $3000 per kilo. This suggests even after generous profit margins being factored in Table 2 everyone in the supply chain has significant profits. These margins also indicate that all of the inefficiency costs can be passed on and there is no incentive to improve manufacturing practices. Average sale price of $3.05 per month would considered low by US standards but it might be considered expensive in the developing countries even when it is sold at $1.00 per month’s supply.  

Formulation Processes

Since dosages are in micrograms or milligrams, one kilogram can go a long way. One kilogram can make one million of one milligram and one billion of one microgram tablets at 100% yield. It tells us that a large volume of high value product can be produced from a small quantity. Using 10,000 kilograms Table 3 illustrates different production rates. At 100% yield we can produce ten billion of ten milligram or 10,000 billion tables containing one microgram active ingredient.

Tablet dose	10 milligram	100 microgram
10,000 Kilogram	10,000,000,000 Milligram	10,000,000,000,000 Microgram
# Tablets	1,000,000,000	100,000,000,000
Hours per year using 85% operating time	=350*24*0.85= 7140
Minutes per year using 85% operating time	=7140*60= 428,400
Tablets production rate per minute	=1,000,000,000 /428,400 =2,334	=100,000,000,000 /428,000 =233,426
Production rate Synthroid, 112 microgram tablets per minute		=233,426*1.43/1.12 =298,202

Table 3: Formulation production rates

Review suggests that a single plant could produce the necessary active ingredient to fill global synthroid demand. However, it would take multiple tableting lines/sites to convert the active in salable product. Since multiple doses are needed it would be worth to have parallel tableting lines. Multiple lines would give flexibility to meet the customer demand. 

Very precise and high degree of process control is needed at every step in each line. Various technologies to produce and package tablets exist. Blending of excipients and actives to deliver a uniform product requires precise controls. 

Analysis:

Point of this exercise is not to suggest that continuous processing is dead. Each manufacturing component has to be looked at separately to see what is the best for business and patient base. Going continuous for formulations is much easier than for API. Basis is ONE kilogram active ingredient can produce ONE million one milligram tablets that will serve about 27,000 patients per year at one milligram dose.

Many companies produce same API. Processes for these APIs are generally inefficient. Square plug in a round hole scenario for such molecules will exist unless a concerted business effort is made to alter the landscape through consolidation. If we look around, out of thousands of small molecule drugs may be less than ten APIs are produced using continuous processes.

Going from “B”(batch) to “C” (continuous) is not like going to the next alphabet. It will require a significant change in business thought process. Omeprazole, metformin hydrochloride and HCTZ (hydrochlorothiazide) are some APIs that could be produced using continuous process. For that matter any API that has a global requirement of more than 350,000 pounds per year could be produced using continuous process. Breakthrough chemistry and brilliant execution would be needed. Use of any API that is produced at these volumes (350,000 pounds per year or more) suggests that their consumption could be increased if costs come down. Pharmaceutical is the largest business segment where cost reduction can increase consumption by 20-30% i.e. by billions.

I am not giving up hope for continuous processes. I hope you don't either. It will take effort.

Girish Malhotra, PE

President

EPCOT International