Everyone wants their business to be more profitable. For drug manufacturers, there are two (legal) ways in which to increase profits:

1. Increase product prices. This sometimes works, if you have a superior product or are underpriced at the moment. However, whether you are selling generic or branded drugs, the competition is terrific. In addition, a number of countries limit the cost of products to what they think is a reasonable level (results-based reimbursement in Germany, for example). In the U.S., HMOs and even Medicare are becoming cost conscious. So price rises will take you just so far before you hurt your sales.
2. Reduce your COGS (Cost of Goods Sold). Since you are limited by the raw materials that you can use (see the word “legal” in the first sentence), production costs come to mind. These are broken down into several categories:

a. Labor costs. The more people you have working (normal hours and overtime), the higher your labor costs.
b. Equipment needed. If you are performing your production just as it has been done for the last century, you have large amounts equipment sitting unused at any time, either dirty, awaiting cleaning, clean, awaiting cleaning validation, or sitting with an intermediate, awaiting assay results.
c. Utility costs. When you have a large number of production units (including blenders, granulators, reaction vessels and tabletting machines), there are costs in electricity, both to run the machines as well as light and air condition/heat the buildings.
d. Materials used in the manufacture of the drugs. Obviously, these are the chemicals used to synthesize the API, bottle the drugs, the API, itself, and all the excipients used in the dosage forms.
e. Taxes. While not needed to perform the work, these have to be thrown in the mix when considering “overhead.”
f. Time to market and recalls. Time to market can mean either how fast an ANDA is approved or keeping the shelves stocked. Recalls are (unfortunately) self-explanatory.

OK, having established that raising prices is not the best route to profitability, we can examine the pieces of the second approach: lowering costs. Some of the points are intertwined and are not easily seen one-on-one. For example, the longer a batch takes to manufacture, the more warehouse space will be needed to store intermediates. This warehouse space needs to be lit and heated, of course. On top of that, more land needs to be developed, increasing property taxes.

Add to those points the length of time required to deliver lots of product to retailers. If your product is competing with one or more other products (branded or generic) and your product is out of stock, a doctor well may prescribe a competing product. Seldom do patients switch back at a later date, unless there is some adverse effect, of course.

Since so many factors dovetail, let’s examine the various steps in the production of a solid dosage form. And, before you start complaining that, as a contract manufacturer, you must do things your clients’ way, allow me to explain how you can modernize and stick to your contracts. But before I start, let me simply state, “The FDA has never complained because you tested too much or too often.”

First, we must be very, very sure of our raw materials. With our supply chains stretching around the globe, accepting a certificate of analysis or a simple “trust me” is unacceptable today. Since the first approved method for “container-wise” qualitative testing of incoming raw materials (via near-infrared) was 30 years ago (1984), the Agencies (FDA, EMA) are quite receptive to such testing. Why switch?

Some years ago, I made a presentation (sponsored by a NIR company) to a major pharma company that was having raw materials bottlenecks in-house. The company’s QC department was taking from 30-90 days to clear raw materials, since all the other tasks (such as release testing, stability samples, pre-formulation and pilot plant samples) were considered more important than raw materials. As a consequence, when ordering simple excipients such as lactose, the company ordered two or three times more than needed, so they would have enough on hand, “just in case.”

There are two issues here:
1) speed of assay
2) number of samples “pulled.”

A typical suite of USP tests for lactose takes several days. The tests include IR and color tests for ID, sieve analysis for particle size, pH and moisture analyses, and so on. If we assume a lot of 200 bags of Anhydrous Lactose, USP, it could easily take a week to run all the tests on all 200 samples. Even an NIR or Raman test involving samples brought back to the lab could complete the task before lunch of “Day 1.” The values, especially using NIR, could also include being able to determine mean particle size and moisture levels.

Many companies don’t even open, much less test, all raw material containers. They run a composite, often based on a (√n +1) paradigm. That is, for 100 containers, they open and sample 11 and, often, combine them into a single sample.

My own experiences, while developing the first NIR method (all those years ago), show the folly in both using compendial methods and combined samples. As for composite samples, my first full-blown test on Lactose was meaningful. The QC lab passed the composite sample, while I found that two bags were out of specification (OOS), using the first multivariate analysis (MVA) algorithm. Upon re-test of those two samples, one was found to fail the pH test and the other contained too much moisture. Would these have killed a patient? No, but they signaled that there could be other latent problems. [For the record, all 220 samples took one morning to sample, deliver to the lab, scan, and evaluate.]

Worst case scenario? Another “event” occurred, which led me to measure for particle sizes (spoiler alert). The lab accepted a lot (several containers) of a barbiturate salt. When I scanned the samples they “failed” by a large margin. By that time, I had noticed physical differences in the spectra (diffuse reflection is quite sensitive to physical effects). I quickly sent the samples to the photo-microscopy lab and, sure enough, the samples were not 120 mesh, but were micronized! They were to be used for suppositories and there was no chance that the material would have dissolved in cocoa butter. The paperwork for destroying a controlled substance is ugly, by the way.

The USP method stated, for its particle size requirement, that “no more than 1% is retained on a 120-mesh screen.” Obviously, anything smaller passes that test. [I didn’t even mention the “test” for lactose, where a sample is boiled in a copper solution turning it red, proving that the sample is a reducing sugar. These simple, non-specific tests were designed to help a pharmacist, in his store, while compounding a batch, assuring him that he is probably using the correct materials and they do not, for example, contain heavy metals. As seen with the heparin scandal, compendial tests are absolutely no help against deliberate substitution of shoddy or OOS raw materials.]

So, without violating any contracts with clients, a contract manufacturer can easily switch to NIR for incoming raw materials, assuring that
1) they are, indeed, proper, safe, and within specs
2) time is saved, allowing materials to be delivered directly to the production floor, avoiding quarantine space in yet another warehouse, and
3) there are cost savings on personnel, chemicals, instruments, legal disposal of chemicals and solvents, lab space, etc., etc., etc. I won’t even mention the cost of a failed batch or recall due to poorly tested raw materials that turn out to be OOS.

Another time waster, yet critical step is blending. Since all batches of generic product, by law, need to show blend uniformity, plant personnel need to either “guess” at the correct time or stop the blender at several time points and take samples from the body of the mix. This leads to over or under-mixing of the blend and wastes time sampling, labeling, delivery, assay, reporting, etc. Many companies go with “historical” times, sample, then proceed “at risk,” assuming the lab results will be fine.

So much effort and time (and badly mixed batches) could be saved with a wireless NIR or LIF (light-induced fluorescence, best for <1% API content) riding the mixer. The “best” end point can be achieved for each lot, often in far less time than formerly assumed. That alone increases throughput and add equipment to the process stream. You also avoid the sampling, labeling, assaying, reporting, etc. involved with the procedure. As an added bonus, all the ingredients are “seen” by NIR, assuring that more than just the API is well-blended. Because of different physical differences, other excipients, dyes, lubricants, etc. all mix at slightly different rates. Spectroscopy assures a well-blended mix.

Another plus is that over-blending is avoided. Why is that crucial? When you are rotating (assume a V-blender) several tons of powder, over and over, you generate quite a bit of heat. That heat can change the polymorphic form and degree of crystallinity of both the API and excipients and cause degradation of the API.

I have merely touched on the obvious and immediate steps that can be taken to both assure safety and notch up the profit per batch, while not making any changes to the production of the product, assuring compliance with contracts. In later columns, I will show how to update other parts of the process to both assure quality and increase the profitability of the product line.

---

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.