What Change are in the Future of Pharma?

Emil W. Ciurczak

Introduction: Some years ago, the “Dean of Science Fiction writers” (Isaac Asimov) was complaining that many of his newer “fiction” stories were being placed on the shelves of libraries’ “science” reference sections. The “march of science” had become a full sprint and was running ahead of his whimsy. In a similar mode, many of what I may feel are “future” improvements are, most likely, already in place in any number of locations. Nonetheless, they remain a secret to most practitioners or, at best, a rumor.

Many “Future” improvements are, most likely, already in place in any number of locations

Since the USFDA began encouraging Pharma to come into the 21^st Century (part of the title of a Guidance), a number of forward- thinking groups within larger companies have taken the FDA at its word. However, the inertia in the pharmaceutical industry is greater than most industries. This is natural, considering that the USFDA came into existence for the sole purpose of halting abuses and excesses of the drug producing industry; in other words, a policeman. Over the years, thousands of complicated and arcane rules have become “the law.” It did not help that the most common answer from Agencies to “how should this section (of an NDA or ANDA) be written?” was “Just give us something and we’ll tell you if it’s OK.”

To be safe, Pharma companies simply took the most conservative pathways, namely:

1. Make each submission just like the last 200 submissions. [“Heck, it’s worked for us for 60 years, why trifle with success?”]
2. Use the most advanced technologies [NMR, Raman, LCLC, LC-MS, etc.} to assure that the API is safe and active, then simply use USP or BP “wet” analysis methods for the final product, assuring speedy acceptance by the Agency (and retaining as much patent-protection as possible) since there will be less validation needed.
3. Since any “new” concept or production technology/monitoring needed to be arduously (read: time consuming and expensive) validated, manufacturers continued using 1950’s approaches to development and production.
4. Since 21^st century production is simply larger versions of mid-20^th century techniques, minimum testing was considered “safer” than a higher number of tests.

The fourth point is, in essence, the main reason for resistance to the idea of PAT, namely that, if we test too much, we will find problems with every batch. This is a fair point as every batch has percent active, hardness, weight, etc. distributed as best can be described by a Gaussain curve. So, unless a process actually is a six-sigma process, there are a number of outliers that, for a million-dose run, could run into the hundreds or even thousands. This is, of course, based on the traditional “make-it-and-then-check-it” paradigm. If PAT is used to merely assay the old-fashioned method of production, its main outcome will be discovering all the weak points and outliers. But can/should PAT (and QbD) be used to speed up AND improve the quality of the product?

Actual PAT to QbD

While the idea of PAT goes back 15 or so years, it is still not a bad idea to adopt a PAT program. If we assume that a company is heavily invested in traditional batch-type manufacturing, then PAT can help in several ways:

1. The 1^st part of PAT/QbD, ofttimes overlooked, is qualitative examination of incoming raw materials (both excipients and APIs) I also recommend 100% examination, to avoid potentially mislabeled or tainted conteners (“√n + 1” testing, commonly used, would imply only examining 11 out of every hundred containers, allowing potential errors in 89 of them).
2. The measured parameters of the raw materials allow the operators to estimate the mixing times, for example: smaller mean particle sizes equal to less time, on average, than larger ones.
3. Even if lots of incoming raw materials are found to be acceptable, the time saved in real-time qualification (e.g., an hour for 200 bags of lactose vs. a week in the QC lab) more than makes up for the time and expense of validating the program. [This doesn’t even take into account labeling incoming materials as “in testing,” quarantining them, relabeling them approved or destroy, and transferring them to a different part of the warehouse; they may be either used immediately or sent back on the truck that brought them.]
4. The physical information gleaned from incoming qualification may then be given to R&D/Pilot Plant/Production to allow them to plan their blending, etc.
5. The next easiest step in PAT, putting an in-line, real-time blend uniformity unit allows for optimal mixing time. All too often, the blend is mixed longer than needed, just to “be sure.” Over the course of a year, optimal blending can shave several minutes from each lot, effectively adding equipment without cost. [Blending generates heat which, in turn, could cause degradation, change in polymorphism, or simply grinding the particles to smaller sizes.]
6. In addition, active monitors at each production step fulfills the GMP requirement to “have meaningful in process tests.” Not only will these data be meaningful, but obtained in a timely fashion, potentially avoiding lot failure.
7. Another oft-ignored GMP requirement is “a statistically significant number of dosage forms need to be tested.” Clearly, 10, 20, or even 30 tablets from 3-5,000,000-unit run does not constitute a “significant number” in any statistics course taught on Earth. Using current HPLC testing is neither timely nor is it conducive to many samples per lot (potentially thousands?). Time and costs are reasons for not doing large number of HPLC samples, so we “wink-wink nod nod” and make believe we are properly sampling production runs. An inline spectrometer would give both control and a picture of the “goodness” of a batch from beginning to end.
8. One last point: any contract manufacturer, having this ability, makes itself more appealing to clients since they can produce better products, faster.

While the idea of PAT goes back 15 or so years, it is still not a bad idea to adopt a PAT program

Continuous Manufacturing

The logical conclusion to an effective PAT/QbD program is to take the batch-wise (moving from step one to step two to…)production, now “controlled” by PAT and simply link the steps together. [Not necessarily simple, but the ROI is so great, it is worth the trouble.] There are several larger companies engaging in CM currently (Vertex and Pfizer, to name two) and several more seriously working on the technology needed.

Since a well-designed PAT program already would assure that each step (mixing, granulating/ribbon, lubricating, granulating, coating is within proper parameters, it almost seems silly to drag several containers of powder and/or tablets from room to room, charging and discharging apparatuses, for days at a time. This “classic” approach requires more rooms, more cleaning/cleaning validation, more personnel, more lighting/heating/venting, and a larger footprint than simply “downsizing” each unit and placing them all in one room.

Since they’re all in the same room, why not attach them and make a single entity? Since we are now skilled at ascertaining blend uniformity (in real time), let’s just keep adding API and excipient, in the correct proportions, and continuously blend them (a screw blender seems the most logical choice). Granulating can be performed, but extruding a ribbon has been seen to be more amenable to CM and easily attached to the blender. The ribbon can be chipped and lubricated continuously and the lubricated mix easily tableted (or encapsulated) continuously. [The easiest approach in CM is seen to be direct compression, where possible. Since the mixture is small and continuously monitored, stratification, seen in large granulation feeders, is less common in CM.]

There are several excellent advantage in using CM in lieu of batch production:

1. One not immediately apparent advantage is when your company commits to true QbD, based on a well-run design of experiments. To perform the DoE correctly, full-sized batches should be run to account for physical interactions (mixing time, polymorph change, etc.). In a classic batch production, this would entail a lot of time, equipment, excipients and API. A small-ish DoE could take more than a month to complete (not including all the lab work needed). Using a CM unit, the ratios of API/excipients mixing time/speed, and all such combination could realistically be performed in a couple of days with all the analytical data available at the end of the experiments.
2. Batch size is only dependent on the time of a run. That is, a company can easily set up for a 500,000-unit run as a 10,000,000 run. Of course, it goes without saying that there is no waiting for lab results, as the entire batch has been monitored and, for all intents and purposes, analyzed in real-time.
3. Cleaning is much simpler, too. After conferring with several companies who use CM, it is seen that “appropriate” powder, run through the intact system, can efficiently clean the components as well, if not better than a water/alcohol/detergent approach, often the “go-to” cleaning approach today. This cuts even more time between products and increases throughput and eliminates almost all of the down-time associated with conventional cleaning.
4. The most obvious positive attribute is the footprint needed. Since there are no massive tableting machines, the massively thick floor (sometimes a meter thick) is unnecessary. The entire suite may be housed in a two story room, not much bigger than a small auditorium. This means less land, HVAC, electricity, personnel, warehousing, and all associated expenses.

CM is much like the automobile: horses and blacksmiths were nice and needed for many years, but very few people, now exposed to cars, would willingly ride a horse to work or shopping. As a co-instructor of mine once stated, “There is no law stating you must stay in business.”

 3-D Printing

Not likely to rapidly replace products that are either easily produced or are made in large numbers, this methodology is the answer to a number of complicated questions. At several recent meeting (e.g., IFPAC), I was thrilled to see actual products being made with 3-DP. These were usually low-volume batches, typically specialty drugs for small populations of patients, but there are also a number of more esoteric applications.

3-D printing is the answer to a number of complicated questions…

Special delivery systems (osmotic pumps) and even long duration (controlled-release) dosage forms benefit from 3-DP. Getting a 12-hour delivery system to repeat its performance via standard compression technology, even when aided by CM, can be more art than science. 3-DP give a far more reproducible dosage form than any other system.

Several other benefits are realized with 3-DP:

1. Cleaning, again, is simpler. Since the components are “sprayed” onto a platform (stationary of moving), the reservoirs can be dedicated to various excipients and APIs. The proper ones simply put into place for the production of the various products and seldom need cleaning (covered by SOP’s, obviously).
2. Again, DoE is very inexpensive and rapid. The company can make even smaller experimental units via 3-DP, which can be important when the unit is complex and/or the API is rare or expensive.

Currently, 3-DP is mainly used for fabrication of medical devices, such as replacement knees, and work is being done on cells to form muscles and organs. The joints are a “slam-dunk” while organs are still down the road a bit. As the units fall in price and experience is gained, I would say this is the most interesting new technology to come along in decades.

Predictions

Of course, we are not saying that there will be a technological revolution over the next year or two. For decades after the automobile was introduced, many people complained that they were smelly, loud, there were no gasoline stations, no one knew how to repair them, and so forth. Horses were known, traditional, comfortably familiar, etc. How did that turn out?

Several students, over the years, have asked, “We covered that last year. Do we still need to know it?” My stock answer is, “If you are an English major, you do not need to remember Chaucer to study Shakespeare.” However, if you are majoring in science, what was covered last year and the year before are important. Likewise, the Pharma (or BioPharma) industry cannot jump to continuous manufacturing without "prior knowledge" of PAT principles (risk management, design or experiment, statistical process control, etc.), which lead to QbD concepts (including lifecycle concepts, supply chain control, LEAN, six sigma).

When these concepts are understood, embraced, and mastered; when the proper personnel are in place; when upper management embraces the program (acknowledging that the process has been enhanced, with regard to speed and quality) are we prepared for CM. CM is the peak of the pyramid and, like the roof of a building, cannot be built without the base being strong.

Yes, technology will be a strong impetus for change, but, as always, economics is the largest driving force for change. As price controls are instituted in country after country, the pressure to make less expensive drugs will only increase. On top of that, the Agencies of all these countries are becoming increasingly strict, obviating the temptation to cut quality to cut costs.

Slowly, larger companies will move to 1) their own QbD/CM processing or 2) outsource to CMO's that are QbD/CM capable. Smaller/generic companies may have difficulty engaging the personnel/technology needed to move to successful QbD/CM manufacturing, so they will have three options: 1) close (there is no law forcing them to stay in business), 2) merge with one or more small companies, so they can afford/support QbD/CM, or 3) (happening now) the larger Pharma companies engage and support one or more smaller companies to produce their off patent, older (but still in-demand) brands.

So, will there be a tsunami or slow geological-speed change in Pharma? Like a garden, some seeds sprout early (larger, innovative companies like Pfizer, GSK, and Merck), some come later, but the last seeds face an existential choice: sprout or be crowded out by the other plants. There will be an entirely different looking industry in a decade: there will be fewer companies and the only ones making a profit will be the ones who understand and properly implement QbD.

Hint: lf the US FDA and EMA were to announce that they would enforce the two parts of cGMPs I mentioned earlier by a set date (e.g., 2022),then line for purchasing PAT/QbD equipment, software, and experts would resemble the line at an Apple store when a new iPhone is introduced.

CM is the peak of the pyramid and, like the roof of a building, cannot be built without the base being strong.

Source: The Pharma Review