- All APIs were synthesized and purified, in-house;
- Pre-formulation studies were performed, in-house;
- Clinical supplies were made in-house and the trials were run by in-house staff;
- Formulation and scale-up work was done, in-house;
- Product manufacture, QC analyses, and packaging was all performed, in-house; and
- Finally, through the 1970s, product was stored and sent to hospitals and pharmacies from the company’s own warehouses.
Then the good news/bad news occurred: the markets became international, making it easier to produce a product in the country where it was sold than try to export it from “home base.” While it was, in some ways, simpler to deliver the goods, other details became difficult. The smallest difficulty was the inability of the Quality staff to lay eyes upon any piece of documentation. Yes, as the years went on, more and more of the information was available on computer files, but locating them was still primitive.
Later on, as “blockbuster” products became the norm, the sheer number of lots produced precluded a single point of production, even within a given country. Now ‘Frisbee 25mg’ was being produced at several sites within many different countries—either by wholly owned locations or contract manufacturing organizations (CMOs). This meant that the product of a single name was being produced at numerous sites, each potentially with its own sources of API and raw materials. Keeping track of cGMP adherence was becoming a nightmare. The Quality staff was now enhanced and functions such as “vendor validation” became full-time jobs. Travel between sites—much of it overseas to sites in developing countries, where concepts like GMP were, quite literally, foreign concepts, became more and more extensive, and expensive. FDA warning letters of cases where data was destroyed and product tested over and over, until passing results were obtained, became more common. In addition, it is economically impractical to station someone from QA at each CRO and CMO to assure there is compliance with the clients’ SOPs, MMFs, etc.
On top of that, authorities (FDA, EMA, etc.) had been required to receive advance permission from the Chinese government before entering the country for an inspection. Do we suspect that maybe, just maybe there could be some “sanitization” of the facility and records prior to the site visit? Of course, not all violations or deviations are from nefarious behavior. Many “evolving” or developing nations do not have a long history of scientific training and application. Many “lightly trained” technicians are literally learning on-the-job and will make “beginner” mistakes. Hiding these mistakes—some mechanical, some computational—is often based on embarrassment, not criminal intent. Thus, what is needed is a more immediate way to obtain an overview of contract facilities, both operational and paperwork.
The answer can be traced to the FDA’s PAT Guidance and personnel therein. When I was having a conversation with Moheb Nasr, formerly CDER, FDA, now principal of Nasr Pharma Regulatory Consulting, he described what HE would consider the best-case scenario of an FDA inspection, remembering that all comments are personal and do not reflect Agency policy. He said, in his view, the “ideal” oversight of a company, using an approved PAT/QbD-based process line, would be to call the QA manager, ask for the code to sign into the computer system from, say, 12:00PM until 3:00 PM and XX Month, 20XX, and observe the XYZ process.
Now, why would that be considered an “inspection?” In short, any PAT/QbD-based production has been heavily validated and proven to be effective. All the in-process data up to and including RTRT (a.k.a., real-time release testing), is subject to 21CFR part 11 (Electronic Records; Electronic Signatures). That is, all data must be stored in its raw form for up to twenty years, enabling any inspector, or appropriate QA person or manager, to recall the data and apply the algorithms that convert it into “results.” That is, these data must be shown to produce the same “answers” or results when recalled from storage and challenged well after the time they were generated. [E.g., if the API value for tablet #34,506 was 20.7 mg (103.5% of label for a 30 mg tablet) last July, it should also be 20.7 mg, today, when the algorithm is applied.]
As was mentioned above, when the QA office was within walking/driving distance of the location where these data were generated/stored, a simple “drop-in” inspection was child’s play. As production spread to other (same country) locations, it was harder, but still feasible to “drop in” on production and quality control for impromptu inspections. With the supply chain spread over several continents, surprise visits became a thing of the past. Clearly, multinational visits are becoming painfully inefficient as controls for quality for reasons beyond cost, time and personnel constraints. Of course, I seldom merely state problems without some type of suggestion for redress of the problem. I refer to a software solution.
When LIMS (lab information management systems) were first introduced in the 1970s, the computers were, kindly stated, primitive. The large data disks, which were the size of pizza pies, needed to be replaced daily and stored for further inspection. The lab computer that I was using for NIRS research, with attached 9 Mbyte drive, surpassed the one available to the entire QC department. Today, most smart-phones far exceed the capacity and ability of departmental LIMS units of decades ago. With the international capacity of the internet and “cloud computing,” namely a third-party server with large capacity, it became quite easy to keep control of all the aspects of a production process within a company at least. As the market for better software packages was established, a number were introduced, some better and more versatile than others. One in particular with which I am familiar stands as an example of what could be accomplished within a company AND with the client-CMO relationship.
A Lisbon-based company (4Tune Engineering) has a product named i-Risk that began its life as a solution to communications within a corporation. The purpose was to allow a QA person in, for example, Germany, to access the records in New York, allowing them to ascertain which Guidances were being followed, etc. Obviously, records of raw materials, production progress, and so on could also be examined at any time. Shortly after the introduction of the in-company version, they rolled out the version with the ability to communicate with other companies, including contract organizations such as CMOs and CROs.
The software allows the Quality representative to check on processes, incoming raw materials, finished product analyses, and any other items necessary. Clearly, actual qualification of incoming raw materials (NIRS, Raman) is superior to a mere certificate. But, if the R.M. suppliers have been certified and even occasionally visited, the CoA will suffice.
The process will work best if there are actual data available for each process step (PAT), but GMP measures, once validated, could be used instead. The software can immediately show the QA person the status of each step of each CMO. Some headers are Workflows, Folders and Entities. As seen in Figure 1, the board shows numerous CMOs as well as status, what happened/is happening, what was done for any circumstance, what effect the procedure had, and so forth. Any incident can be opened and details examined. While the QA personnel are still free to travel to any CMO site, this type of software allows them to follow contracted materials’ manufacture quality in real time. Welcome to the 21st Century. Maybe clouds in the sky do signal good weather?
Emil W. Ciurczak
Emil W. Ciurczak has worked in the pharmaceutical industry since 1970 for companies that include Ciba-Geigy, Sandoz, Berlex, Merck, and Purdue Pharma, where he specialized in performing method development on most types of analytical equipment. In 1983, he introduced NIR spectroscopy to pharmaceutical applications, and is generally credited as one of the first to use process analytical technologies (PAT) in drug manufacturing and development.