Moving From QbA to QbD - An Example of Continuous Processing

Quality by Analysis (QBA) has been the MANTRA for manufacturing pharmaceuticals. This is due to the fact that, as we know, pharmaceuticals are manufactured using batch processes. Regulatory agencies would like the industry to move from QBA to QBD. QBD can result in a better batch process or a continuous process. Either of these process methodologies require a complete understanding of component chemistry and their mutual interaction.

Jumping from QB-"A" to QB-"D" might seem like a quantum leap for the pharmaceutical industry but it should not be. However, I can understand the challenge of moving from QBA --> QBD when "B" Bureaucracy and "C" Consternation stand in the way. Somehow we have to take a leap of faith based on our scientific prowess and education to stride from A-->D. If we can take this jump, we will be a step closer to a continuous process.

Pharmaceutical companies file and are granted patents for their small molecule process chemistries. Many of the patents are simple and excellent chemistries. Some present opportunities to produce their products continuously. However, even with these discoveries, the products are produced using a batch process. It almost seems as if there is an unknown force preventing companies from making the switch to a continuous process. One recent such patent is from AstraZeneca AB about the manufacture of "The Purple Pill".

Abstract of USP 7,227,024 (1) is " A process for the manufacture of omeprazole or esomeprazole from pyrmethyl alcohol via pyrmethyl chloride and pyrmetazole characterized in that the whole reaction sequence is carried out in a solvent without any isolation or purification of intermediates. Further, the reaction is carried out in a solvent system common for the whole reaction sequence and inert to the reactants formed during the process and used in the process and comprises a water immiscible organic solvent and a specified amount of water."

Based on the above abstract it is clear that for the production of "omeprazole or esomeprazole" an intermediate pyrmetazole is needed. AstraZeneca has other(2) patents related to the manufacture of this active pharmaceutical ingredient. There are other patents (3, 4) for the manufacture of the same API but are not part of the discussion.

To the right are the two steps needed to produce pyrmetazole. The patent '024 describes the two steps of the batch process.

Pyrmethyl alcohol, 8.82 g (52.7 mmol), was dissolved in toluene, saturated with water, 74 ml (water content 0.4 mg/ml according to Karl Fisher titration). To the stirred solution, at 10° C, thionyl chloride, 8.15 g (68.5 mmol), was added slowly over 60 minutes (flow rate 0.083 ml/min). A fast reaction was recorded, reaching 99% conversion after completed addition of thionyl chloride. Water (2.3 ml) was added to quench any excess of thionyl chloride.

Step 1: Pyrmethyl alcohol +Thionyl Chloride -> Pyrmethyl Chloride. Below, Step 2: Pyrmethyl Chloride + Metmercazole + Sulfur dioxide + NaOH-> Pyrmetazole + Water + NaCl+ NaHSO3
(Click image for larger version.)

An alkaline (13.5 g, 168.3 mmol; 50% w/w sodium hydroxide) aqueous (80ml) solution of metmercazole (9.8 g, 54.2 mmol) was added followed by additional sodium hydroxide (8.8 g, 110.5 mmol, 50% w/w sodium hydroxide) to reach pH>12. The temperature was allowed to increase to 45° C during the additions. The reaction mixture was left with vigorous stirring for approximately two hours at 45 °C. The agitating was interrupted and the phases were left to separate. The aqueous phase was discarded. The organic phase, comprising pyrmetazole, was washed with water and was analyzed for residues of pyrmethyl alcohol (less than 0.1% mol).

Based on the above description, anyone experienced in process commercialization would conclude that the reactions are fast i.e. zero order. Exact temperature and residence time need to be refined. They will depend on how a continuous process for Pyrmetazole will be executed.

In the following write up I have slightly modified the two steps of USP '024 to show how the two batch process steps could be converted to a continuous process. I have not defined the exact process conditions but any chemical engineer and chemist would understand my rationale and thinking. Some of the process details will be explained in an upcoming book (5). Conversion of Pyrmetazole to Omeprazole (API) can be extended via a continuous process.

Inventors at AstraZeneca improved the yield of its US Patent '955 by improving mass transfer. Many times we are not sure if certain reactions will progress because there is no literature reference of an unconventional reaction scheme. If there is no mention in literature that does not mean it cannot be done. Unless we try it, we would never know. USP '024 is one of those cases where the inventor has done what the literature suggests cannot be done.

Continuous Process description

Dissolve pyrmethyl alcohol in wet toluene that meets the moisture specification defined in the patent. Alcohol/toluene mix is fed to the reactor at a continuous rate in a glass-lined jacketed reactor. Temperature of the reactor is to be maintained as high as possible. Azeotrope can be used as an additional method to control the exotherm. Reaction vessel has to have sufficient residence time to ensure complete conversion of the alcohol to the corresponding hydrochloride. Thionyl chloride flow rate is linked to the feed rate of the alcohol.

An alkaline solution of metmercazole is prepared in 50% caustic. Its concentration is controlled based on solubility. This alkaline solution is fed into the second reactor and its feed rate is controlled based on the flow rate of the hydrochloride coming from the first reactor. pH of the reaction mass is controlled at about 12.5 to 13.5. Reaction temperature is controlled at 55-65oC. Reaction temperatures ranges can be further refined to improve the reaction time. Instead of jacketed glass-lined reactors, other reactor configurations can be used for the process.

pH of the reaction system would define the material of construction of the piping and the reactors. Temperature range in the first reaction has to be such that any sulfur dioxide generated in the first reaction stays dissolved in the liquid. If necessary the generated sulfur dioxide can be neutralized with an alkali
Pyrmetazole produced in step # 2 is oxidized to produce omeprazole, which is taken to water phase and further processed to produce the API. Organic phase is processed and toluene is recovered for re-use. In this particular process we are finessing the physical properties of the reactants to our advantage to create a continuous process. Since we can control every reaction step, we do not have to do any process sampling as we are controlling the reaction steps to produce the desired quality product. The final product is crystallized using commercially available crystallizers.

Figure 1: A process scheme showing two reactors. This set-up can be used for the two steps described
(Click image for larger version.)

Figure 1 depicts a process scheme showing two reactors that can be used for the two steps. The final step can be added to the process. Another way to have a continuous process is by using back mix pipe reactors. Based on the above description, a process that can operate 7140 hours/year (based on 24/7/350 days at 85% "on-stream time") can be designed. A 500 to 1000 pounds per hour plant can easily produce about 3.5 - 7.1 million pounds of the Nexium API. A single plant can be designed to serve the global needs.

A recent process development(6) - production of primary amines from alcohols and ammonia can be done continuously if the physical properties and kinetics are manipulated.

Similarly the active ingredient of Provigil (Modafinil) (7,8) 2-(benzhydrylsulfinyl) acetamide can be produced continuously by finessing the chemical addition, reaction sequence and crystallization. A single plant can easily fulfill the global needs for this API even after the expiration of the patents.

For a continuous process, a complete understanding of the chemistry and kinetics of an API is needed. In addition, if the chemistry and physical properties can be manipulated, conversion of a batch to a continuous process is facilitated. These are not difficult to put in practice.


1 USP 7,227,024 Method for the Synthesis of a Benzimidazole Compound *expires October 14, 2023)
2 USP 5,958,955 Method for the Synthesis of a Benzimidazole Compound (expires December 4, 2017)
3 USP 5,386,032 Method of Synthesis of 5-Methoxy-2- [(4-Methoxy-3, 5-Dimethyl-2-Pyridinyl)-Methyl] Sulfinyl-1H-Benzimidazole (Omeprazole) (expires May 24, 2013)
4 USP 6,303,788 Process for Preparing Omeprazole (expires November 30, 2018)
5 "Chemical Process Simplification: Improving Productivity and Sustainability" by Girish Malhotra to be published John Wiley & Sons Inc. 2010
6 C. Gunanathan and D. Milstein, Angew. Chem. 2008, 120, 8789-8792
7 USP 6,875,893 Preparation of a Sulfinyl Acetamide (expires May 21, 2023)
8 USP 7,057,069 Preparation of a Sulfinyl Acetamide (expires May 27, 2025)
About the author: Girish Malhotra is president of EPCOT International. email:; phone: 216-292-0626