Emil W. Ciurczak

In the 1970s, the analytical device industry introduced laboratory instruments that performed titrations, better thermal analyzers to show crystallinity, and all types of chromatography (e.g., gas-liquid chromatography [GLC], capillary gas chromatography [GC], high-performance liquid chromatography [HPLC], and ultra high-pressure liquid chromatography [UHPLC]), which took off like a rocket and became the backbone of the R&D, quality control (QC), and stability departments. In the 1980s, near-infrared and Raman instruments became accurate and operator-friendly and the first commercial Chemometric software packages were introduced. Thus, the industry expanded the infrared (IR) range from “merely” midrange-IR to near-IR (NIR) and later (c. 2005) to a viable, information-rich far-IR, now dubbed “terahertz.” Despite the speed, accuracy, and throughput of modern analytical instrumentation, it has remained safely ensconced in the laboratory, year after year.
	
	Production samples were obtained by operators from the process line, labeled, and hand-carried to the QC lab. Samples were logged and assigned to analysts for assay while the lots were held in quarantine in the warehouse until analytical results were delivered, often a week or so later. The slow turn-around meant stocking more raw materials than immediately needed and more product than orders for those products. Campaigns--where many process lines were incorporated for the same product for weeks on end-were run in anticipation of need, storing the product until packaged and shipped. All in all, there was a massive overhead involved, which was folded into the cost of goods sold (COGS).

	For more than 50 years, pharmaceutical product manufacturers were content to hum along, doing the same thing, year after year. They received the raw materials; quarantined them until QC managed to perform some arcane, compendial assays and either passed or failed them; blended the API and excipients and stored the mix; granulated, dried and stored the blend; punched tablets or filled capsules, and stored them; finally, coated the tablets and stored them until QC approved them to be bottled and sold.

	There was little pressure to change because most proprietary brands had little to no competition from other branded products or the few generic houses in existence at the time. Why would they spend the time and money to make any adjustments in the manner the products were manufactured?
	And, for analytical purity and “goodness,” taking only 20 units for assay was pretty good, especially when the batch just numbered in the thousands. As far as testing the raw materials, quick compendial checks were good enough, because the supply chain (a word not yet being bandied about) was domestic and most large companies synthesized their own APIs.
	
	


	With the trend to outsource every facet of the business (e.g., manufacturing, API production, packaging, clinical trials, and even basic research), the term “supply chain” came to the forefront. Many pieces of the pharma puzzle became opaque. APIs previously taken for granted because they were made in house now came into question because of their external and possibly offshore origin. In addition, simple 

) tests proved inadequate (e.g., heparin). The 10-20 tablets taken for “goodness check” and the mere six dissolution samples for “performance testing” suddenly seemed a trifle thin in the face of multi-million unit lots, possibly made by a surrogate (branch or contract manufacturing organization [CMO]) in a developing country.
	
	For the companies still making their own products, the cost of in-house production under the 50-year-old production paradigm kept rising. Also, because of the increasing number of questionable products being imported and the exploding number of generics--with the concomitant exploding number of GMP violations cited by the European Medicines Agency (EMA) and FDA--the agencies increased their scrutiny of their domestic, formerly above reproach, large branded products. Unfortunately, the agencies were finding that “after the process” tests were not as good a measure of product goodness for large lots as the old “mother’s kitchen”-sized lots of the past.

	Several things happened simultaneously. Pfizer wanted a way to speed the production and release of Viagra, and the US Congress, alarmed at loss of jobs and production to overseas pharma companies, suggested that FDA help domestic companies to compete. Part of the initiative came with a number of guidance documents and initiatives (1).

	The FDA draft guidance for process analytical technologies (PAT) was issued in 2002 and generated so much debate that the comments section of the FDA webpage almost crashed the system with the number (and tone) of submissions, and it took more than two years to get the final guidance released (2). It seems that the quality assurance (QA) departments went wild at the thought of using “good scientific judgement” in lieu of hard and fast procedures. The idea of flexible parameters, such as “blend until well-blended” or “dry until dried” or “vary hardness to give proper dissolution” were anathemas to QA.

	A number of companies, however, learned that assuring a product is well-made during the process is preferable to failing “X” number of lots (or having major recalls) every year. And, to help smooth the way for PAT, quite independently, Pfizer was developing tools for PAT to become feasible.


	The first major tool was a stand-alone, wireless, battery-operated near-infrared analyzer, designed to be mounted on a blender. Working at its Sandwich, UK facility with technicians from Zeiss (Switzerland), Pfizer designed, tested, and built a unit for determining end-points for blend “uniformity.” The author points out that powders cannot be “homogeneous;” only solutions may be described as such. A more accurate term is simply, “well-blended.” When a series of measurements (moving average) came to a constant standard deviation, the blending was halted. There were estimated times, but, depending on the size of the lot and the variable parameters of the excipients and APIs, the end-time would vary.

	Once industry process development people wrapped their heads around variables in master manufacturing formulas (MMFs), in-process testing started to sound reasonable. It was Pfizer, again, that started using in-process sampling and NIR analyses of Viagra. Using primitive, yet effective methods, they continuously sampled two lines and brought samples to a number of NIR instruments in a “lab” constructed between the two production lines. Thus, they analyzed up to 20 tablets per hour per production line. For a 12-hour run, a minimum of 240 samples (taken sequentially throughout the run) was analyzed. Compare this to scooping 20 tablets at the end of the run. Obviously, if any problems were seen, the tableting could be either suspended or corrected.

	From these humble beginnings, many instruments were designed specifically for the production line. Aside from NIR, Raman, terahertz, chemical imaging (e.g., NIR, Raman, and fluorescence) are just some of the tools available to control the process. Add to this the qualification of raw materials and one can monitor and, more importantly, control every step of the production stream.

	That is, one can characterize APIs, raw materials, and packaging materials; follow the blending, granulating, and drying of the mixture; direct the tableting and coating processes; even check the fill (mostly blister packs) and identity of the final package. This was PAT, which begat quality by design (QbD). FDA published several more guidance documents and the International Council on Harmonization (ICH) generated a series of guidances (ICH Q8, Q9, Q10, and Q11) to standardize how one finds design space (i.e., allowable variations within a process, determined by experimentation using design of experiment software and synthetic batches) and conducts a QbD program.

PAT as a gateway to continuous manufacturing


	The industry now has encouragement from FDA, ICH, and EMA to perform modern, real-time monitoring/adjustments and begin to consider “real-time” release (on-line, not after-the-fact in the QC lab). The use of process-hardened NIR, Raman, terahertz, and other tools have rendered after-the-fact analyses moot.

	The costs of traditional process lines places branded companies out of the generics marketplace-resulting in dumping or licensing a product at or near its patent expiry date-and making change a little less odious.

	Because the values of all the pieces of QbD (based on PAT) save time and money, improve quality of the product, and essentially eliminate recalls, it is time to take the next logical step: continuous manufacturing (CM).

	Now that more than one manufacturer of process equipment also supplies CM equipment, it is the next logical step in the march to the 21st Century.

	Vertex received FDA approval of their CM-based new drug application for their cystic fibrosis medicine and Janssen retro-validated an existing product to a continuous process (3,4).

	A CM unit consists of a series of serially connected operating pieces: weighing units for APIs and excipients, dispensing powders into a screw-like blender, optional in-line granulators, or ribbon blender (or direct compression), a drier when a granulator is used, compression, followed by coating (if needed) or filling into capsules.

	In most processes, within an hour of start-up, the first tablets begin coming off the line, ready for packaging. Why not send these tablets to QC for analysis? All the pieces of a real-time release procedure are in place: proper amounts of each component are weighed (and recorded at all times); the blending is continuously monitored and controlled; the ribbon compaction or granulation is monitored and controlled; the tablet pressing (or capsule filling) is controlled (weight, assay, and even dissolution prediction); and, if coated, the coating procedure is monitored and controlled. Therefore, if PAT leads to CM, what comes next? Does the phrase “3-D printing for orphan drugs” sound interesting?


	1. FDA, Pharmaceutical cGMPs for the 21st CenturyâA Risk-Based Approach (Rockville, MD, September 2004).
	2. FDA, 

Guidance for Industry, PAT-A Framework for Innovative Pharmaceutical Development, Manufacturing, and Quality Assurance

 (Rockville, MD, September 2004).
	3. L. Yu, “Continuous Manufacturing Has a Strong Impact on Drug Quality,” 

, April 12, 2016, http://blogs.fda.gov/fdavoice/index.php/2016/04/continuous-manufacturing-has-a-strong-impact-on-drug-quality/
	4. S. E. Kuehn, “Janssen Embraces Continuous Manufacturing for Prezista,” 

, Oct. 8, 2015, www.pharmamanufacturing.com/articles/2015/janssen-embraces-continuous-manufacturing-for-prezista/


	When referring to this article, please cite it as E. Ciurczak, "PAT: “Gateway Drug” to the 21st Century for the Pharma Industry," 

Articles

Process analytical technology paved the way for continuous manufacturing.

PAT: “Gateway Drug” to the 21st Century for the Pharma Industry

PAT has aided formulation, both pre-and post-filing, by reducing costs and time frames. Contrary to popular beliefs, it doesn’t require big investments.

PAT has aided formulation, both pre-and post-filing, by reducing costs and time frames.