Last month, I ended with a thought or two about cooperative projects between initiator companies and (usually) smaller generic or contract manufacturing firms with, of course, the proviso that such “marriages” do not violate anti-trust laws. Before you say this cooperative is impossible, let me give you an example—albeit much smaller—of such an effort.
Back in the late1980s, early 1990s, companies were beginning to take an interest in measuring processes in real time. The control methods, at that time, for example, consisted of such methods as using a “sample thief” to extract a number of samples from a blending bin, send them to QC and, based on the results, declare the blend mixed, or not. I had published a paper using a fiber optic probe, pushed into the bed to determine uniformity by Near-Infrared Spectroscopy. Pfizer went quite a bit farther.
Understanding that any one of their blockbuster drugs (Lipitor, for example) had sales in excess of the combined sales of all instrument companies, worldwide, they realized they could not ask a much smaller instrument company to invest a significant portion of its time and fortune in developing a specialized instrument that a Pfizer might or might not buy, after all. So, they came up with a plan that has become a “new normal” in PAT/QbD circles.
The first (“PAT-type”) in-process instrument was a joint project between Pfizer (Sandwich, UK) and Zeiss optics (Zurich, CH). A lab-style near-infrared spectrometer was mounted on a v-blender with a window in one end cap. Since no R&D was done prior to the experiment, there were still wires coming from the instrument: one from a power source and another to the computer. The blender was rotated a single one-half spin in one direction, the one-half spin in the other direction to avoid breaking or binding the wires. When it appeared to be giving meaningful results, Zeiss developed the unit with an internal power source and wireless capacity. The device/technique was patented, but not “defended,” allowing all companies to use the procedure in their processes. This process monitoring device changed the entire paradigm of the powder blending step in a manufacturing process.
This big/small company cooperative model can work in many ways:
- An initiator company would develop the dosage form/process as usual and, as the successful product increases in sales, it can clone its process equipment and distribute them to subsidiary companies (either co-ops or wholly-owned), quickly validate them, and proceed to produce the drug at multiple sites.
- The initiator company can develop the process and sub-let it to a contract manufacturing company. This would be an exclusive license and all the products produced are sold under the initiator’s name. When the patent expires, the contract would end and a new paradigm would need to be negotiated.
- Then, we have the concept of a sponsor-company subsidizing an entire PAT/QbD process(es) at a contract location, with the CRO/CMO developing everything from the clinical supplies, through the final dosage form, through actual production and distribution of the commercial supplies.
The best part of this (these) situation(s) is that any or all of the contract locations would not be dependent on a single supply source for raw materials. Since the overall scope of QbD is to modify the process to affect the outcome, variances in raw materials are overcome through changing the settings/parameters of the Design Space—all previously validated, accepted changes to the process, which allows the intermediate and final product to meet specifications. The reason that we have specified excipients from specific vendors is that, under the cGMP (batch) manufacturing paradigm, processes (and amounts of materials, e.g., lubricants, such as magnesium stearate) cannot be legally varied even though some experienced operators make “corrections” to the batches to “assure” they meet specs.
This does not mean that all standards for raw materials will be discarded, but that two things are (painfully) apparent:
- Current specifications are based on purity, not performance. Since almost all the specs (as in USP, EP, etc.) were formulated decades ago when pharmacists made small batches of pills or capsules in the back rooms of their pharmacies, easy tests, such as tap density, color, heavy metals, and sieve size were all that were needed, since there was no processing as we know it today. This assumed that the supplies were all locally obtained (or from a small number – or one – nationally based vendor(s)), so only id and purity needed to be tested. Manually blended small batches didn’t need physical parameters specified.
- Since modern pharmaceutical manufacturing is so dependent on the physical structure of the ingredients, the two choices are to sit down with vendors and agree upon delivery of tightly controlled materials or, the companies can adopt a true QbD program, which allows physical variation (of course, still maintaining chemical purity), which allows manufacturers to use local vendors, saving inter-state and inter-country shipping of materials.
This ability to not be held captive by a supply chain is step one. Now, we can look at quality. Quality isn’t a vague concept; it is rather easily defined. Some of the hallmarks of a quality product (and its process) are:
- Obviously, a product that consistently meets product performance criteria in physical and chemical tests (assay, dissolution, etc.).
- Monitoring each step to avoid waste and shorten the time between steps (minimum PAT set-up). This leads to a faster through-put, making better use of manufacturing equipment.
- Quality, from a somewhat selfish point of view, means avoiding the hassle of recalls and or destroyed batches that do not meet standards.
- All these (1-3) points lead to a reduced COGS and assuring that their orders to distributers are met (especially since, if a generic product, there are many competitors waiting to step in).
Now that we have the ingredient difference problem addressed, we can look at the manufacturing sites. The traditional (a.k.a., “old-fashioned”) production process has been:
- Do the preliminary formulation work. (Including clinical batches)
- Do a scale-up (batch process) on the chosen formulation and write the SOPs for the process.
- Begin producing the final product, using traditional multi-step processes, at the site of origin.
- If the product is successful, attempt to mirror the process at your other sites or, as we are seeing in recent years, contract the (GMP) process to CMOs around the world.
At this point of a successful product, the traditional approaches as the patent life expires were:
- Drop prices during the patent’s last year, hoping to sell a percentage of the product, based on name recognition and reputation.
- Change the product design at the last minute, hoping to delay generic competition.
- Invent (excuse me), “discover” new uses for the older product to extend its patent life—think a lower dose Proscar (non-cancerous prostate tumors) being marketed as Propecia (male pattern baldness).
Now, in the time of continuous manufacturing, built on sound scientific principles (QbD, Design of Experiments, real-time process monitoring instruments), it is possible and economically sound to reproduce the same manufacturing conditions in multiple locations: including contract manufacturing sites.
This is where creative business decisions are needed. Getting back to the three scenarios, where CM units may be “cloned” and placed in subsidiaries to help keep up with demand, or placed on CMOs to either use them as contract producers with short-term contracts, ending or re-negotiated upon patent expiry date.
The truly imaginative companies will find ways to “partner” with contract manufacturers. These solutions will range from outright paying them to produce the product(s) on equipment owned by and installed by the proprietary company to loaning the CMO the equipment for the duration of the product’s patent life to working out a plan wherein the CMO takes a portion of its fees towards eventually owning the continuous manufacturing unit(s).
Clearly, there are positive aspects and speed-bumps in each case. The one clear positive that exists in all cases is that the speed and cost-savings and consistency of CM-made product will bring the ever-sought-after “quality” to the product. The scenario wherein the CM unit is merely “lent” to the CMO is the lowest cost to the contract facility, allowing smaller producers to bid successfully for the contract.
Where the CMO is allowed to purchase the hardware (and expertise and training), either on a lay-away plan or through some profit-sharing method, it sets them up for being able to easily accept more contracts for CM-produced drugs. Ultimately, these well-equipped facilities can be used as resources for DoE experiments (virtual scale-ups) and potentially, with assistance, actual formulation work. A great advantage to that scenario is that all the R&D would be performed with locally sourced raw materials, obviating the need to modify the work performed by a central R&D facility.
Since I am merely a humble chemist, I cannot even suggest methods for the financial structure of these proposals, nor can I imagine how the distribution of finished products could be integrated into the existing structure. But, all in all, I am sure that the financial wise people in both companies can work out something that will both lower costs and improve quality.
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.