I was reading an extended article about the reasons for the baby formula fiasco of recent months. It would appear that a series of errors, missed documents, and laissez-faire attitudes came together in a “perfect storm.” The plant was isolated and was a major part of the income for the town in which it was located. The FDA and plant management kept overlooking often obvious problems, which resulted in recalls, consent decrees, and eventually, a total shut-down of the plant.

There were reports of standing water, powder found in cracks of mixing vessels, voluntary recalls because of bacteria found in spot inspections of product, and still there was no action by the government, based solely on assurances of the company’s managers that all problems would be addressed and repaired. When the number of illnesses affecting young children became significant, the FDA acted.

A disturbing sidebar shows that, for a number of reasons including exclusive patents and licenses, infant formula was made in very, very few locations. A single pharma factory making, for example, analgesics, would hardly make a dent on pharmacy shelves were it to close. With a specialty item like baby formula, however, the shutdown was catastrophic. Shelves were empty for months and emergency supplies were flown in from around the world.

Likewise, several years ago, a major pharma firm needed to recall several different products because of some wrong product mixed with the proper meds in some bottles. But these two examples aren’t necessarily the points of this column.

The main point is simply, the pursuit of excellence (quality product) and a healthy profit margin are not exclusionary. Making a product safer, more consistent, and properly packaged is not an unnecessary expense, it is an investment. There is no such thing as a non-important component. For example, it was a simple, inexpensive O-ring, not tested for cold, as it was in Florida that day, that brought the multi-million-dollar Challenger space shuttle down and killed all the crew! Sometimes, it pays NOT to get the part from the lowest bidder!

There are two ways to improve quality and safety: 1) institute vast amounts of testing for GMP batch-type manufacturing or 2) move (over time) to QbD/Continuous Manufacturing. Continuous processes may be applied to API synthesis, solid dosage form production, or biologicals, but I will emphasize solid dosage forms. Let us address the first approach.

The steps in GMP batch-type manufacturing that need to be checked/controlled are:

1. Raw Materials. The most common method is to bring a large number of containers of materials into the warehouse, label then “quarantined,” sample a pre-determined number (e.g., N1/2 + 1, for example, where N = the number of containers sampled) of containers and send samples to the QC lab.

Not only is this a time-consuming process—some larger companies can take months to pass/fail the samples—but, if we have 100 bags or containers, only 11 will be sampled, with 89 possibly contaminated or be the improper grade/particle size. If the sampled materials are passed, they need to be relabeled and moved to another section of the warehouse, another time-burner. If they fail, they need to be moved to be discarded or returned to the supplier. This takes time and storage space, plus the materials need to be re-ordered, taking even more time. I have encountered companies who have to wait months for raw materials to be cleared; priorities are release, stability, in-process, R&D samples, then stability samples!

Control Mechanism(s): At the loading dock, containers may be opened (in a clean area, of course) and the contents of every container qualified by NIRS or Raman instantly. Depending on the model used, either just the ID or a series of parameters may be measured. NIR has been used since the mid-1980s to perform ID (clearly), mean particle size, moisture level, and polymorphic form.

Using this approach, the approved materials may be immediately labeled “passed” and any outliers returned to the vendor. The savings in time (lab and shuffling about the warehouse) and space (no need to store more than needed and only approved materials are kept) are substantial. This is one step closer to “just in time” manufacturing.

2. Blending. Larger companies perform testing during development for optimum blending conditions (i.e., RPM, blending time) then fix the time for process batches. Of course, this does not take into account variations in physical parameters of raw materials. That fact is why all generic batches are supposed to be sampled to assure optimum distribution of all ingredients. In reality, many generic firms who make batch after batch of a particular product and are comfortable with their track record, may proceed “at risk,” where the blended materials proceed to the next step while the samples are analyzed. The assumption is that there are so few failures, it is less expensive to destroy one lot than delay all their products.

Control Mechanism: an on-board Near-Infrared, Light Induced, or Raman spectrometer, measuring the degree of blending and stopping the blend at the maximum point.

3. Granulation. As a general description, the blended powders are placed in a vessel, some solvents or just water are added, the liquid(s) and powders are mixed, then dried to a pre-determined level of dryness. This is accomplished (normally) in a unit such as a fluid-bed drier, where warm, dry air is percolated through the mixed bed until “dryness” is reached.

The main problem with this method is that, with varying particle sizes, RH of the incoming air, and other variables, the time to “correct” dryness varies. Currently, most companies either stop the process (at a certain time) and do an analysis of the moisture or trust to a timed endpoint.

Control Mechanism: a NIR probe is inserted and the moisture levels are measured in real time. No guesswork involved and batch after batch dried to the same level.

4. Tableting/Encapsulation. Under “normal” conditions, the granulation is lubricated and fed into a tableting/encapsulating machine. The amount of granulation in a tablet or capsule will, obviously, determine the amount of API in each dose (assuming “homogeneity” of the blend). The weights are mitigated by some sort of continuous weighing mechanism (available since the 1970s), but they are averaged and OOS tablets or capsules are difficult to weed out, post-production.

Physical parameters, such as hardness and disintegration times are measured by pulling samples and testing off-line. The assay/content uniformity is not measured until after production is completed, so real-time changes can only be based on weight by adjusting the tablet press to accept more or less granulation or compress with more or less force. So, in short, there is no real control under typical GMP production.

Control Mechanism: In the early 1990s, Pfizer began introducing PAT into their production of Viagra—they couldn’t make it fast enough with the “tried and true” approaches. They developed the first “ride-along” wireless monitor (NIR, obviously) and built a QC booth between their tablet lines to do (nearly) real-time analyses. They took 10-20 tablets from each press, used a Carver press to flatten the tablets to disks, and ran then on NIR spectrometers, equipped with multiple sample holders. This meant that a typical lot could easily have hundreds of tablets assayed over the course of production, virtually assuring good uniformity.
Today, we have equipment that can measure the weight, hardness, and API content of every tablet and remove any on outside set parameters.

5. Coating (tablets). Tablets are coated for several reasons (i.e., protecting the API from O2 or moisture, controlling dissolution times, a possible second API, etc.), but in all cases, the coating needs to be consistent within and between lots.

The process consists of placing the tablet cores into a slowly rotating pan. The mixture of coating material and solvent is sprayed (over several periods) on the rotating bed of tablets. The tablets are allowed (assisted) to dry and more coating added. This continues until the optimal amount is added. The coating is “measured” by weight and seldom by examining more than a few tablets.

Control Mechanism: NIR or Raman are the most commonly applied tools. Figure 1 shows a Raman probe in a coating pan, monitoring and controlling the spraying and drying cycles. NIR or Raman can also be used to assure that the coating solution/suspension is in a proper ratio during the coating process.

6. Packaging. Without going into great detail, packaging is the company’s last chance to “mess up” the product. The actual filling of bottles and blister packs is pretty well established, so they need little changes. However, there is a gap in what might be called “chain of custody” between production and packaging. The recalls that I mention at the beginning of this column were simply due to not performing a simple, rapid check on what is being filled. Clearly, if we can do all the aforementioned analyses non-destructively, we can do a simple ID with a hand-held NIR or Raman instrument, avoiding mislabeled products.

The Ultimate Solution is simply continuous manufacturing (CM). One of the greatest time-burners is cleaning validation. Under GMP batch production, the majority of production equipment is either “dirty,” “being cleaned,” or “clean, but awaiting assay results.” This takes up space and ties up a fortune in hardware.

CM is, under most conditions, easily cleaned by running a common ingredient (e.g., lactose, talc) through the process train between lots of product. This process can easily be validated, then used continuously.

These changes have been made by some of the largest Pharma companies. They can and should be adopted by all companies to assure safe and consistent drugs. And, since they save money, profits will be assured.

---

Emil W. Ciurczak
DoraMaxx Consulting

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.