Emil W. Ciurczak, DoraMaxx Consulting09.08.16
For some time, I have been extolling the benefits of dragging labs and production into the 21st century for contract research organizations (CROs) and contract manufacturing organizations (CMOs). And, while I have alluded to some techniques, I may have been lacking some detail. I would like to address some of the applications of near-infrared spectroscopy (NIRS) in pharmaceutical applications, which has for some time now, been moving from the lab and into production. Let’s look at the process from start to finish.
Raw materials
Essentially, raw materials represent the first application of NIRS in pharma, which occurred in 1984. We initiated the idea at Sandoz in New Jersey because the European Medicines Agency (EMEA)—now EMA—suggested that 100% of all incoming active pharmaceutical ingredients (APIs) and raw materials be qualified. To do this, for example, for more than 200 bags of lactose alone would take weeks using U.S. Pharmacopeial (USP) tests. With all the other materials, we would have needed to more than double the quality control (QC) staff. We chose to use NIRS instead.
At that time the instruments were lab-based and samples needed to be collected in the warehouse, brought to the lab, placed in cups, and scanned. Despite all that it only took a half-day to collect, label, and test 220 bags of lactose.
One other benefit was finding differences not seen with a combo sample. Two of the samples from that first batch passed by the QC lab but failed by NIRS. When the individual samples were tested one failed pH and the other failed for moisture level. These were hidden in the combined sample.
By 1990 several companies built their same-sized instruments so that they could be wheeled into the warehouse and, using their fiber optic probe, could be used to examine each container individually without removing samples or even opening the plastic liners. The software, which became available in the 1990s, allowed the analyst to determine in seconds the type of lactose—classifying it as spray-dried, freeze-dried, fast-flowing, anhydrous, or anhydrous. If the company wished, other parameters could be determined such as mean particle size, moisture, degree of crystallinity, and polymorphic form.
Sadly, 30 years after its introduction, NIRS is still nowhere near universally used for raw materials and APIs. The inertia so prevalent in regulated industries seems to be keeping so many companies both large and small from applying this “new” technique.
If and when a company decides to employ PAT/QbD, the operators will need to know the physical parameters of the raw materials in order to make allowed adjustments to keep the final product within specifications not currently allowed under cGMP.
Blending
Blend uniformity testing is required for every batch for generic manufacturers. Taking discreet samples and sending them to the QC lab and waiting for assays is time and labor intensive. Many companies proceed at risk, assuming the assays will be fine. Far too many over mix to avoid unmixed batches. This is extremely wasteful. Even if all samples turn out to be fine the manpower, lab time, and chemicals are a waste. Also, if powders are mixed a little longer to assure blending, several things happen: Equipment is tied up needlessly; crystallinity and polymorphs change with extra mixing; APIs can change from amorphous to crystalline—critical flaw in poorly soluble drugs; and the heat generated may cause needless degradation of API.
A simple removable and easily transferred wireless NIR may be attached to almost any type blender. The materials are then only blended to an optimal level and all the listed negatives disappear. The paybacks are faster turn-around times for equipment, less breakdown of particles, less exposure of the API to heat generated by the blending process, and no need for keeping materials in quarantine awaiting analysis or risking making a product with under-mixed blends.
Granulating/drying
Currently, samples are taken manually, sent to a lab for, perhaps a Karl Fischer or a nearby station for LOD, and the moisture measured. A sieve analysis may also be run for granule size measurement. The current choices are to keep the drier/granulator running and risk overdoing the process, or stop the process and allow moisture to re-adsorb/re-distribute while waiting for a number from the test(s).
We have two critical parameters in this step: moisture and agglomerate size. There are several commercial NIR units that can measure both in real time, keeping the process from over-drying or causing the granules to grind down to powder, once again, destroying the advantages of granulation. These use both visible ranges for particle characterization and NIR for moisture levels. With multiple fiber probes multiple levels may be monitored simultaneously.
Compressing
And capsule filling, but since most solid dosage forms are tablets, I will address that type. With a potential for drift due to materials segregation over the course of a long batch compression, taking 20 tablets post production is almost worthless. At best it can detect an OOS batch while at worst it shows nothing about the goodness of the lot and where and why it went awry, if it is OOS.
A NIR placed immediately after compression will at least tell the operator that the tablets are uniform or are drifting high or low in real time with enough time to stop the process or make any adjustments.
This is also a good time to perform content uniformity and assays on the cores, since most modern coating materials contain titanium dioxide and iron salts, making them nearly opaque to NIR radiation. For measuring coating levels, this becomes a good thing, as the NIR peaks from the core disappear, while the peaks of the coating material and its solvents predominate the spectrum.
Coating
Initially, coatings were placed on tablets for several reasons, including protection from moisture and oxygen and identification of the dosage form, especially where there are several levels of API available. Since the coating was often cosmetic, it was sufficient to simply weigh the amount of coating and visually ascertain that the tablets appeared uniformly coated.
Today’s coatings are more precisely applied and, in addition to protecting the API, they may also be used to affect the dissolution and, in some cases, contain another API. Clearly, accuracy is needed in the application of the solution.
NIRS has been used for years to control the coating process. The most common approach is to place a fiber optic probe into the coating pan and measure and control the spraying and drying cycles. In addition to controlling the spray/dry cycles, NIR may even be used to assure that the coating mix remains constant. As was mentioned in the previous point, the core peaks disappear and the coating peaks appear. A calibration equation is generated and used to control the process.
Packaging
Seemingly trivial, packaging is the last, best protection from environmental contamination and mix-ups. There are several areas where NIRS can be instrumental in making the packaging run smoothly and safely.
With plastic bottles, this entails confirming the proper polymer type, density, and crystallinity is done in the lab when they are not merely accepted from a certificate of analysis from the vendor. NIRS can do these quickly at the loading dock and, for safety, just before the products are filled by the same instrument used to assure the proper product has been sent from production for filling.
Another potential disaster is with blister packs. The bi-layer polymer portion comes in rolls where we assume the roller was loaded properly. That is, the adhesive is down and mere polymer is up. If the sheet was rolled in the wrong order, it will melt onto the packaging equipment causing a large and expensive mess. Merely checking with a hand-held NIR will assure that the proper side meets the aluminum backing, avoiding a costly error and missed shipping dates. Since it is neither possible nor prudent to open every pack another method may be used. Yes, you guessed it: NIR. Numerous studies have been published—one by yours truly—where 100% of the tablets/capsules/vials may be tested in less time than doing the traditional sacrificial testing in the QC lab. Even a single fiber-optic probe on a NIR spectrometer, operated by a single analyst, can scan 100% of a batch prior to shipping in less time than the HPLC methods now employed.
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.
Raw materials
Essentially, raw materials represent the first application of NIRS in pharma, which occurred in 1984. We initiated the idea at Sandoz in New Jersey because the European Medicines Agency (EMEA)—now EMA—suggested that 100% of all incoming active pharmaceutical ingredients (APIs) and raw materials be qualified. To do this, for example, for more than 200 bags of lactose alone would take weeks using U.S. Pharmacopeial (USP) tests. With all the other materials, we would have needed to more than double the quality control (QC) staff. We chose to use NIRS instead.
At that time the instruments were lab-based and samples needed to be collected in the warehouse, brought to the lab, placed in cups, and scanned. Despite all that it only took a half-day to collect, label, and test 220 bags of lactose.
One other benefit was finding differences not seen with a combo sample. Two of the samples from that first batch passed by the QC lab but failed by NIRS. When the individual samples were tested one failed pH and the other failed for moisture level. These were hidden in the combined sample.
By 1990 several companies built their same-sized instruments so that they could be wheeled into the warehouse and, using their fiber optic probe, could be used to examine each container individually without removing samples or even opening the plastic liners. The software, which became available in the 1990s, allowed the analyst to determine in seconds the type of lactose—classifying it as spray-dried, freeze-dried, fast-flowing, anhydrous, or anhydrous. If the company wished, other parameters could be determined such as mean particle size, moisture, degree of crystallinity, and polymorphic form.
Sadly, 30 years after its introduction, NIRS is still nowhere near universally used for raw materials and APIs. The inertia so prevalent in regulated industries seems to be keeping so many companies both large and small from applying this “new” technique.
If and when a company decides to employ PAT/QbD, the operators will need to know the physical parameters of the raw materials in order to make allowed adjustments to keep the final product within specifications not currently allowed under cGMP.
Blending
Blend uniformity testing is required for every batch for generic manufacturers. Taking discreet samples and sending them to the QC lab and waiting for assays is time and labor intensive. Many companies proceed at risk, assuming the assays will be fine. Far too many over mix to avoid unmixed batches. This is extremely wasteful. Even if all samples turn out to be fine the manpower, lab time, and chemicals are a waste. Also, if powders are mixed a little longer to assure blending, several things happen: Equipment is tied up needlessly; crystallinity and polymorphs change with extra mixing; APIs can change from amorphous to crystalline—critical flaw in poorly soluble drugs; and the heat generated may cause needless degradation of API.
A simple removable and easily transferred wireless NIR may be attached to almost any type blender. The materials are then only blended to an optimal level and all the listed negatives disappear. The paybacks are faster turn-around times for equipment, less breakdown of particles, less exposure of the API to heat generated by the blending process, and no need for keeping materials in quarantine awaiting analysis or risking making a product with under-mixed blends.
Granulating/drying
Currently, samples are taken manually, sent to a lab for, perhaps a Karl Fischer or a nearby station for LOD, and the moisture measured. A sieve analysis may also be run for granule size measurement. The current choices are to keep the drier/granulator running and risk overdoing the process, or stop the process and allow moisture to re-adsorb/re-distribute while waiting for a number from the test(s).
We have two critical parameters in this step: moisture and agglomerate size. There are several commercial NIR units that can measure both in real time, keeping the process from over-drying or causing the granules to grind down to powder, once again, destroying the advantages of granulation. These use both visible ranges for particle characterization and NIR for moisture levels. With multiple fiber probes multiple levels may be monitored simultaneously.
Compressing
And capsule filling, but since most solid dosage forms are tablets, I will address that type. With a potential for drift due to materials segregation over the course of a long batch compression, taking 20 tablets post production is almost worthless. At best it can detect an OOS batch while at worst it shows nothing about the goodness of the lot and where and why it went awry, if it is OOS.
A NIR placed immediately after compression will at least tell the operator that the tablets are uniform or are drifting high or low in real time with enough time to stop the process or make any adjustments.
This is also a good time to perform content uniformity and assays on the cores, since most modern coating materials contain titanium dioxide and iron salts, making them nearly opaque to NIR radiation. For measuring coating levels, this becomes a good thing, as the NIR peaks from the core disappear, while the peaks of the coating material and its solvents predominate the spectrum.
Coating
Initially, coatings were placed on tablets for several reasons, including protection from moisture and oxygen and identification of the dosage form, especially where there are several levels of API available. Since the coating was often cosmetic, it was sufficient to simply weigh the amount of coating and visually ascertain that the tablets appeared uniformly coated.
Today’s coatings are more precisely applied and, in addition to protecting the API, they may also be used to affect the dissolution and, in some cases, contain another API. Clearly, accuracy is needed in the application of the solution.
NIRS has been used for years to control the coating process. The most common approach is to place a fiber optic probe into the coating pan and measure and control the spraying and drying cycles. In addition to controlling the spray/dry cycles, NIR may even be used to assure that the coating mix remains constant. As was mentioned in the previous point, the core peaks disappear and the coating peaks appear. A calibration equation is generated and used to control the process.
Packaging
Seemingly trivial, packaging is the last, best protection from environmental contamination and mix-ups. There are several areas where NIRS can be instrumental in making the packaging run smoothly and safely.
With plastic bottles, this entails confirming the proper polymer type, density, and crystallinity is done in the lab when they are not merely accepted from a certificate of analysis from the vendor. NIRS can do these quickly at the loading dock and, for safety, just before the products are filled by the same instrument used to assure the proper product has been sent from production for filling.
Another potential disaster is with blister packs. The bi-layer polymer portion comes in rolls where we assume the roller was loaded properly. That is, the adhesive is down and mere polymer is up. If the sheet was rolled in the wrong order, it will melt onto the packaging equipment causing a large and expensive mess. Merely checking with a hand-held NIR will assure that the proper side meets the aluminum backing, avoiding a costly error and missed shipping dates. Since it is neither possible nor prudent to open every pack another method may be used. Yes, you guessed it: NIR. Numerous studies have been published—one by yours truly—where 100% of the tablets/capsules/vials may be tested in less time than doing the traditional sacrificial testing in the QC lab. Even a single fiber-optic probe on a NIR spectrometer, operated by a single analyst, can scan 100% of a batch prior to shipping in less time than the HPLC methods now employed.
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.