Back in the 1960s, gas chromatography was being considered by the U.S. FDA for analysis. In the 1970s, HPLC was not only not being considered, but openly not allowed as a final analysis method in new drug applications (NDAs); the FDA had neither the expertise nor the equipment to examine HPLC analyses. With some hesitation, the Agency allowed GLC to be used. However, there were a few speed-bumps to ease of use. Not the least of these was that there were no commercial columns or standard methods for assuring any columns were performing equally. In order to use gas chromatography, the analyst needed to:

1. Invest in a lot of silica gel, preferably a larger amount, since there was no guarantee that the same material could be purchased in the future.
2. Purchase various lots of silicone oil from polymer companies, usually in 5-10 gallon lots, since finding that exact same material again was problematic.
3. Mix an amount of the silicone in a volatile solvent, pour it over a weighed amount of silica, and mix under vacuum until the solvent was absent (Roto-Vac).
4. Pack the material into the appropriate glass or stainless-steel column(s) of varying lengths.
5. Try to evaluate the column with various volatile materials, covering various polarities and functional groups to attempt to characterize it for future uses.
6. Send either a column or sample of the packing to the FDA for them to be able to duplicate your methodology.

So, if this much trouble was involved in GLC, imagine the difficulty in producing reproducible HPLC columns. It was no wonder the FDA wanted no part of HPLC assays in 1970. When the suppliers of GC packings adopted McReynolds polarity index, using standard solutes (i.e., Benzene, n-Butanol, Pentan-2-one, 1-Nitropropane, or Pyridine) to characterize their column coatings and settled on percent coverage, etc., the industry soared and became standard in every analytical lab. In fact, there are just a few major suppliers of these packings, which most instrument vendors simply re-sell to their customers, packed in their columns. The liquid chromatography column market has been a bit slower to standardize, but numerous suppliers can produce product that now mirrors the performance of competitors.

When I first began to use HPLC in earnest (c. 1980), LC units had evolved from a one-piece, clunky (actually, primitive) unit to discreet components: injection port/injection device, pump, column (and heater), and detector (and, of course, some sort of recording device). While this was good for companies with the proper talent—experienced chromatographers, instrument repair and maintenance personnel, etc.—most smaller labs couldn’t “make magic” and were hesitant to make the plunge or would perform poor chromatography. In less time than you would expect, manufacturers re-integrated all the parts back into a single unit. This proved successful and greatly expanded the applications and quality of HPLC analyses.

The pharmaceutical industry, on the other hand, went in the other direction: in the 1970s, each major manufacturer performed most functions at a single location. Other than purchased raw materials, almost always purchased in the country where the facility was operating. The API (active pharmaceutical ingredients = actives) was synthesized in-house, the tablets (capsules) were made in-house, the final product was packaged in-house, and, finally, the product was shipped directly from the company warehouse to the hospital or pharmacy. This paradigm worked quite well for decades, but several things happened:

1. Blockbuster products happened. This was the largest spur to expansion and growth (and, eventually to mergers and acquisitions). Multiple locations making similar products meant a need for replication of engineers, formulators, operators, and lab personnel. Quite frankly, the facilities expanded faster than experienced professionals became available and with attendant consequences—some loss of quality, recalls, etc., which happens when you make massive amounts of a product. This was the equivalent of the HPLC units becoming “de-constructed” to borrow a term from the Food Network. While a product might be considered by the company as exactly the same, there needed to be as many inspections as there were new locations, putting a new level of stress on the regulators.
2. Generics happened. This rapid growth of generic substitutes echoed the diaspora of the innovator companies. Some better selling drugs might have dozens of generic substitutes, so the burden on FDA and EMA inspectors also increased exponentially. And, quite frankly, since most generics had a shorter time frame of experience, there tended to be more procedural errors (warning letters, 483s, etc.) than for proprietary companies.
3. International outsourcing happened. About the same time generics multiplied, larger companies began using contract manufacturers, both domestic and foreign. Even though these contractors were purportedly using the same manufacturing formula as the parent company, they have their own equipment, formulators, engineers, and analysts. In fact, they also most likely have different sources of raw materials from the parent company.
4. Specialized companies along the supply chain happened. Now, we add many, many new, specialized companies, much like the automotive industry. In the automotive industry one supplier will supply a steering wheel, another the air bag, a third the cover and horn actuator, and so forth. So, no one point’s QC department is responsible to the steering wheel. Likewise, a pump-actuated basal spray might have the pump supplied by “company A,” the bottle by “company B,” the drug solution by “company C,” the final product assembled by “company D,” and the product shipped to destinations by “company E.” One would hope that the final assembly had QC oversight. This “deconstructed” product now depends on four or five suppliers along an extended supply chain.
5. Competition among sub-contractors happened. In parallel to generic drug products, who claim to all be the same (at least have the same API, if not the same excipients or manufacturing conditions) and may be freely exchanged in a prescription for any patient, multiple sub-contractors may supply the same parts for a complex delivery system (e.g., the pump mentioned above).

In all honesty, each supplier uses their own supplies of polymer, their own proprietary mix of mold release agents, antioxidants, etc., in their polymer formulation and their own molds (injection or blow molds). So, in reality, there are minor variations among suppliers and true interchangeability could be problematic.

On top of small differences, there is a barrier of secrecy, sown by competition, so information about each other’s parts needs to be gleaned by physical examining the part in question.

As with any complex system, the more moving parts, the greater a chance of a breakdown. John Glenn was quoted as saying he worried that he was sitting on a million pounds of liquid oxygen contained in a rocket of 1,000,000 parts, all supplied by the lowest bidder. A fine-tuned sports car works wonderfully, until it doesn’t. A salamander or starfish can regrow a missing limb, simply because of their simplicity. In the auto industry, “just-in-time” manufacturing (one of the popular models for many Pharma companies) was an economic boon, saving millions when it worked. A typical American car contains thousands of parts, supplied by hundreds of suppliers (for dozens of models), so warehousing all the parts is an enormous expense, which led to the just-in-time model.

However, since most car manufacturers only had three days’ worth of parts on hand, a fire, strike, or other natural disaster could sideline a popular model. Many parts are also supplied by alternate vendors, but not all. One example of depending on a single critical part is Apple’s smartphone. The screens are produced in China. Based on today’s newspapers, that alone should mean something. As I am writing this, the DOW Jones average has dropped 800 points (today alone) due to the coronavirus’ upset of supply chains from China to the U.S.; now S. Korea is hit and there are signs of disease in Northern Italy.

This natural disaster may be less permanent than a tsunami or volcano, but it will be hindering commerce for quite some time and, worse yet, make countries more suspicious of each other and erode confidence in transparency. China “delayed” announcing the existence of the problem for a month, then effectively shut down major cities to travel and commerce. I suspect that the freewheeling days of “outsource everything, everywhere” may never completely return.

The effects of this break-down could lead many of the smaller, specialized vendors to consolidate and/or communicate better. If for no other reason than it would be nice for a CMO or CRO to have more than one source for any part/API/excipient, there may be some standardization among suppliers, perhaps led by CROs and CMOs setting stricter specifications for parts and chemicals, making the disparate vendors (kicking and screaming) work towards a commonality. Then just using better pricing and delivery would be their competitive edge, not some secret way of producing a part, which makes it irreplaceable when missing, and shuts down a company when it cannot be delivered.

In short, perhaps this disruption will, despite the human tragedy, lead to some good in the industry, afterall.

---

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.