Near-infrared chemical imaging could hold key data in fight against counterfeiters

Counterfeit pharmaceuticals are complex products that can vary from their legitimate counterparts both chemically and physically. The poorer counterfeits can be identified readily from the appearance of the packaging material, or gross shape and colour variations from the original. However, the more sophisticated counterfeit products typically look like the original and are packaged in reasonable facsimiles of the original packaging. In these cases, a chemical analysis can be performed to measure the presence of known ingredients; in most cases, the presence and concentration of API is measured with traditional techniques, such as HPLC or near infrared spectroscopy. The correct amount of API ensures public safety in the sense that a patient will at least receive the right drug and in the right amount. However, it does not by any means ensure that the product is not a counterfeit or that the bioavailability of the API is correct. Much effort is spent in formulation; careful selection of excipients and formulation parameters result in a tablet with a desired structure, performance and stability. This is where near-infrared chemical imaging (NIRCI) was initially deployed and has provided invaluable information to the industry in both formulation development and root-cause analysis of performance failure.

Janies Dubois. Product Manager Americas, Analytical Imaging at Malvern Instruments.

The power of NIRCI

The NIRCI information that enables an understanding of performance also offers key advantages over traditional analytical techniques for counterfeit detection and sourcing. NIRCI can measure the spatial distribution of ingredients, including API and most excipients; in other words, it measures the structure and the chemical composition at once. Simply put, tablets made with the same ingredients in the same factory and following the same process will have very similar chemical compositions and structures. Tablets made with the same ingredients, but using different process machinery will probably show the same chemical composition and consequently pass HPLC and NIR spectroscopy tests, but will most likely have a different structure (Figure 1). This means that even if a counterfeiter were able to copy the list of ingredients, the structural fingerprint of the tablets would still enable its identification as a rogue product. Only spectroscopic imaging can provide this level of information, and only when it is performed in the NIR can it be done quickly and easily.

Figure 1: Array of tablets with the same chemical composition showing a similar spatial distribution.

The accuracy of NIR imaging is directly linked to the original product’s formulation consistency. The Quality by Design initiative is actually helping in this regard because the manufacturers of original products are taking steps to understand the formulation required to obtain a consistent product, which makes it easy to detect even small differences in the counterfeits.

Data is not being fully exploited

In spite of its benefits, the information obtained by NIRCI has not yet been fully exploited for counterfeit analysis. For example, it has been demonstrated that it is possible to not only determine that a group of products are counterfeit, but also to evaluate how many counterfeiting factories were involved in their production! This is possible because NIRCI can measure both API and excipients; counterfeiters may not all use the same excipients, even if they use the right API.

Even more importantly, since it is expected that geographically-close counterfeiters probably obtain their supplies from a small number of sources, there is a good chance that their products will contain similar excipients from the same or different sources. It is also likely that they purchased their manufacturing equipment from surplus and other inexpensive sources, which almost guarantees they don’t all have the same equipment. The “signature” of the equipment will be seen in the structure of the tablets. For example, blender geometry and conditions of operation impact the patterns of ingredient agglomeration. Meanwhile, the order of loading the blender, the presence of a pre‑granulation step, the humidity level in the processing plant or even the transfer of blends in drums prior to tabletting all impact the physical structure of the tablets by causing segregations, agglomerations, preferential agglomerations of subsets of ingredients and the way they fit together into the pressed product. It is the consequences of all theses variables that NIRCI measures, and it is this tremendous reach into the details of the differences in formulation (voluntary or not) that make it such a powerful tool.

Installing and operating the technology

NIR chemical imaging comes in many flavours and it is important to consider the objectives of the work in the selection of the right technology. The factors that vary between them include:

• data acquisition modality (i.e., mapping versus imaging) which affects speed
• need or not for sample preparation
• instrument stability and maintenance requirements over long periods of time
• ruggedness and the possibility to deploy in mobile or even non-laboratory environment
• availability of automated data processing
• operability by a non-specialist user

Installation and ease of operation depend on the implementation selected; for example, focal plane array camera-based NIRCI (true imaging systems) are usually easier to deploy and operate than a mapping system, but the cooling system of the camera influences the stability, ease of deployment and ruggedness and thus should be considered. A camera cooled with liquid nitrogen, for example, will require a supply of this inconvenient material to be available at all times and the operator to be trained for handling the hazmat. Mechanically-cooled cameras alleviate the need for liquid refrigerant and, as such, are typically used in instruments that may be brought outside a laboratory. Another consideration for field use is the ruggedness of the system in terms of resistance to shock, which may be expected in transport from one location to the next. Can the system be operated in a mobile unit and transported assembled? The technology exists to satisfy a very broad range of needs and it is the objective and conditions of deployment that should dictate the selection of a particular implementation.

For counterfeit analysis, it is intuitive to want the analysis to be performed either at customs control centres, or even in mobile units, so speed, automation, low maintenance and ruggedness are probably key factors in deciding on a technology. Systems that meet these requirements are currently deployed in the pharmaceutical industry, but the authorities have so far not deployed them for control.

Capturing the data

Data that enables the identification of a counterfeit product can be acquired in as little as a few seconds with NIRCI and up to 1–3 minutes for high‑resolution images containing the maximum amount of chemical information. Flexible NIRCI systems allow for the acquisition of either what is known as “a full spectral range” (i.e., maximum chemical information) or discrete wavelengths (fast access to target information). The full spectral range means that data is acquired at consecutive wavelengths within the NIR range (i.e., between 850 and 2500 nm). Instruments typically offer only one of the two ends of the range, 850–1700 nm or 1200–2500 nm, the latter being better suited for pharmaceutical products because of the increased chemical specificity of the absorption spectrum obtained in this range. Using targeted (discrete wavelengths) access enables the development of highly specific identification methods (Figure 2) for which data acquisition requires only a couple of seconds; when a set of tablets are presented to the instrument at once, the result can be obtained at a speed of tens of milliseconds per tablet.

Figure 2: Examples of the level of detail that can be investigated by NIRCI. A tablet of different chemical composition is highlighted in an array of tablets and zooming in reveals structural information.

With NIRCI, the chemical information is in the form of a NIR spectrum or intensity at selected wavelengths in the NIR range for each pixel of the image; the physical information lies in the relative spatial positioning of the pixels containing similar or differing spectra. Simply put, a chemical image is much like a photograph, but the shades of blue may represent the levels of API concentration at each pixel and red may represent lactose. As one can imagine, pictures are nice, but their interpretation could be biased or even quite arbitrary. This is why pictures are usually only reported for illustrative purposes; the images are normally converted into statistics of concentration distribution, agglomeration patterns and co-localisation for reporting. The statistics are calculated using automated data processing schemes, which include checks for data quality and can go as far as ending with flags about conformity to expected chemistries and structures.

How widespread is NIRCI use in the pharma industry?

There is a reasonable level of use of NIRCI for counterfeit analysis among the bigger players in the pharmaceutical industry and the literature contains some information about the measurements performed by these companies. The technology was generally acquired for purposes other than counterfeit detection: troubleshooting of dissolution failure, formulation development work or root-cause analysis of performance issues for example. It is the realisation that the same information, already available in their databases, could be used in the fight against counterfeits that has accelerated this application. The FDA also published compelling results almost a decade ago, which clearly demonstrated the power of the technique to identify counterfeits.¹ Furthermore, a paper published in Science in 2004² contained some of the first chemical images of pharmaceutical tablets acquired to demonstrate differences between medication purchased on the internet from various sources using the NIRCI. Today, the knowledge exists, and it is only a matter of will and resources that are required for it to be deployed on the ground.

How can this technology be realistically used?

It is not unreasonable to think that a database of chemical composition and physical structure could be built for original products, and compared with seized lots to determine their legality. A similar database of seized counterfeits could help build an understanding of the movement of these products across borders and markets. One important consideration for this type of deployment is that the complete fingerprint of a tablet or a set of tablets, both chemical and physical, is obtained in under 3 minutes by a non-specialised operator and all data processing is performed automatically. Therefore, data acquisition and processing is not a barrier to deployment; we need to gather and use all of the information that is already at our fingertips to fight counterfeiters.

The price tag for NIRCI technology is comparable to chromatography equipment, but the information content is unparalleled. While the industry has and will continue to benefit from advanced understanding of their products and the counterfeits found on the market, it is regulatory agencies and consequently the public that would benefit most from a broader deployment of NIRCI technology.

References

1. R.C. Lyon et al., at the Annual Meeting of American Pharmaceutical Scientists (Canada, November 2002).

2. M. A. Veronin and B-B. C. Youan, Science 305, 481 (July 2004).