European fingerprint study on omeprazole drug substances using a multi analytical approach and chemometrics as a tool for the discrimination of manufacturing sources.

Like drug products, Active Pharmaceutical Ingredients (APIs) are subject to substandard and falsification issues, which represent a threat to patient health. In order to monitor the quality of drug substances and prevent the use of non-compliant APIs, Official Medicine Control Laboratories work together in a European network developing coordinated strategies and programmes. The API working group proposed a market surveillance study on omeprazole and omeprazole magnesium with the objectives of controlling the pharmaceutical quality of samples, checking compliance with the monographs of the European Pharmacopoeia, and collecting analytical fingerprints that could be further used to differentiate manufacturing sources for future authenticity investigations. The study described in this article reports the analysis carried out by 7 European laboratories on 28 samples from 11 manufacturers with 5 analytical techniques (related substances with HPLC, residual solvents with GC-MS, near infrared spectroscopy, proton nuclear magnetic resonance spectroscopy and X-ray powder diffractometry). The large amount of resulting analytical data were centralized and treated with two chemometric methods: Principal Component Analysis and Hierarchical Clustering Analysis. Data were analyzed separately and in combination (data fusion), allowing us to conclude that NMR and XRPD were suitable to differentiate samples originating from 9 out of 11 manufacturers. Analytical fingerprints associated with chemometrics were demonstrated to be a valuable methodology to discriminate manufacturers of omeprazole and omeprazole magnesium APIs and detect future substandard and falsified APIs.

GEONs API fingerprint project: Selection of analytical techniques for clustering of sildenafil citrate API samples.

Through its Active Pharmaceutical Ingredient Working Group (API-WG) the General European Official Medicines Control Laboratory (OMCL) Network (GEON), co-ordinated by the European Directorate for the Quality of Medicines & HealthCare (EDQM), regularly organises market surveillance studies for specific APIs for conformity to their monograph in the European Pharmacopoeia. During the past years some studies were combined with a fingerprint study of the APIs. The idea is to obtain a fingerprint for each manufacturer of the API under investigation, allowing the OMCL network to identify future samples as well as to detect substandard and falsified APIs. This paper reports the results of the latest fingerprint study, organised on sildenafil citrate API samples. Seventy-nine samples from 14 different manufacturers were collected throughout the Network. Fingerprint data was collected through Mid-Infrared spectroscopy, Raman spectroscopy, liquid chromatography for related substances, gas chromatography for residual solvents, X-ray diffraction and Nuclear Magnetic Resonance (NMR) spectroscopy. Chemometrics applied to the collected data showed that all manufacturers could be discriminated based on the data of only three of these tests, i.e. gas chromatography for residual solvents, X-ray diffraction and proton NMR. Suspicious API samples for sildenafil citrate will therefore be analysed in the future with the selected techniques in order to link the sample to a manufacturer or demonstrate the absence of such link. If the sample cannot be attributed to one of the manufacturers, further analysis and research on provenance and identity will be required. Of course, if the suspected sample claims to originate from one of the manufacturers included in the study, analysis can be limited to the test distinguishing this manufacturer.

A strategy based on fingerprinting and chemometrics for the detection of regulated plants in plant food supplements from the Belgian market: Two case studies.

The sale and consumption of plant food supplements is increasing, especially in the western world. A lot of these supplements can be bought through internet, where a lot of illegal trade is going on. Every year seized dietary supplements are send to laboratories in order to screen for the presence of chemical adulterants or illegally added active pharmaceutical ingredients, though also herbal adulteration occurs and is given less attention. In this paper a two-step approach is presented based on fingerprints recorded by both infrared spectroscopy as liquid chromatography with UV-detection for the screening of five regulated plants used in respectively dietary supplements for slimming and potency enhancement. Both types of fingerprints are combined with chemometric techniques in order to obtain classification models. A first classification model is calculated based on the infrared data and gives a first idea about the plant suspected to be present. This suspicion is then confirmed based on binary classification models calculated with the chromatographic data obtained for the suspected plant. In general, good classification models were obtained for each of the targeted plants. The approach was applied in a small market study comprising 35 dietary supplements for slimming and 34 for male potency enhancement. In total 21 samples were found to contain one of the five targeted plants.

Combining attenuated total reflectance- infrared spectroscopy and chemometrics for the identification and the dosage estimation of MDMA tablets.

The principal objective of the Belgian Early Warning System on drugs (BEWSD) of the Reitox national focal point is to collect and disseminate information on the appearance of drugs circulating on the local drug market. To this end, one of the actions of BEWSD is to annually collect substances that are circulating on the Belgian summer festivals. Every year a high number of seized samples is collected and submitted to qualitative and quantitative analysis. The goal is twofold: next to the monitoring of the drug market, a surveillance on the quality of drugs is established with main focus on the evolution of MDMA dosages in ecstasy as well as the emergence of new psychoactive substances (NPS). The presented paper focusses on the characterization of ecstasy tablets, seized at summer festivals in the seasons 2016 and 2017. Samples were classically analysed with GC-MS and UV spectroscopy, representing a certain cost and time for analysis. In this paper, an approach is presented based on infrared spectroscopy to first differentiate ecstasy from other party drugs and further quantify MDMA in the tablets. To this end, a comparison was done between NIR and Mid-IR spectroscopy in combination with partial least squares-discriminant analysis (PLS-DA) and -regression (PLS). For the differentiation between MDMA positive and negative tablets, the best results were obtained by NIR and PLS-DA and resulted in a correct classification rate of 96% for an external test set. In addition, for the quantification of MDMA, the best results were also obtained with a PLS model based on NIR spectra. A RMSEP for the external test set of 3.86 was obtained and an R2 value between real and predicted values of 0.88. This means that for the dosage MDMA per tablet, when taking into account the range of masses of the tablets in our sample set, the error varies between 8 mg for low mass tablets and 27.8 mg for high mass tablets. These are acceptable values for a first estimation and give a first indication of the risk the tablet may represent. The presented approach will be of use for on-site analysis. Moreover, when applied in a laboratory environment, it will reduce the workload for the analysis of MDMA tablets and thus liberating resources.

Detection and identification of multiple adulterants in plant food supplements using attenuated total reflectance-Infrared spectroscopy.

Due to the rising popularity of dietary supplements, especially plant food supplements, and alternative herbal medicines, a whole market developed and these products became freely available through internet. Though several searches revealed that at least a part of these products, especially the ones obtained from websites disclosing their physical identity, are aldulterated with pharmaceutical compounds. This causes a threat for public health, since these compounds are not declared and therefore adverse effects will not immediately be related to the product. The more the adulterants can interfere with other medicinal treatments. Since the present active pharmaceutical ingredients are not declared on the package and the products are sold as 100% natural or herbal in nature, it is very difficult for custom personnel to discriminate between products to be confiscated or not. Therefore easy to apply analytical approaches to discriminate between adulterated and non-adulterated products are necessary. This paper presents an approach based on infrared spectroscopy combined with attenuated total reflectance (ATR) and partial least squares- discriminant analysis (PLS-DA) to easily differentiate between adulterated and non- adulterated plant food supplements and to get a first idea of the nature of the adulterant present. The performance of PLS-DA models based on Mid-IR and NIR data were compared as well as models based on the combined data. Further three preprocessing strategies were compared. The best performance was obtained for a PLS-DA model using Mid-IR data with the second derivative as preprocessing method. This model showed a correct classification rate of 98.3% for an external test set. Also eight real samples were screened using the model and for seven of these samples a correct classification was obtained. Generally it could be concluded that the obtained model and the presented approach could be used at customs to discriminate between adulterated and non-adulterated herbal food supplements and even get a first idea of the nature of the adulterant present. The more the presented approach hardly needs sample preparation.

Chemometrics and chromatographic fingerprints to classify plant food supplements according to the content of regulated plants.

Plant food supplements are gaining popularity, resulting in a broader spectrum of available products and an increased consumption. Next to the problem of adulteration of these products with synthetic drugs the presence of regulated or toxic plants is an important issue, especially when the products are purchased from irregular sources. This paper focusses on this problem by using specific chromatographic fingerprints for five targeted plants and chemometric classification techniques in order to extract the important information from the fingerprints and determine the presence of the targeted plants in plant food supplements in an objective way. Two approaches were followed: (1) a multiclass model, (2) 2-class model for each of the targeted plants separately. For both approaches good classification models were obtained, especially when using SIMCA and PLS-DA. For each model, misclassification rates for the external test set of maximum one sample could be obtained. The models were applied to five real samples resulting in the identification of the correct plants, confirmed by mass spectrometry. Therefore chromatographic fingerprinting combined with chemometric modelling can be considered interesting to make a more objective decision on whether a regulated plant is present in a plant food supplement or not, especially when no mass spectrometry equipment is available. The results suggest also that the use of a battery of 2-class models to screen for several plants is the approach to be preferred.

Detection of regulated herbs and plants in plant food supplements and traditional medicines using infrared spectroscopy.

The identification of a specific toxic or regulated plant in herbal preparations or plant food supplements is a real challenge, since they are often powdered, mixed with other herbal or synthetic powders and compressed into tablets or capsules. The classical identification approaches based on micro- and macroscopy are therefore not possible anymore. In this paper infrared spectroscopy, combined with attenuated total reflectance was evaluated for the screening of plant based preparations for nine specific plants (five regulated and four common plants for herbal supplements). IR and NIR spectra were recorded for a series of self-made triturations of the targeted plants. After pretreatment of the spectral data chemometric classification techniques were applied to both data sets (IR and NIR) separately and the combination of both. The results show that the screening of herbal preparations or plant food supplements for specific plants, using infrared spectroscopy, is feasible. The best model was obtained with the Mid-IR data, using SIMCA as modelling technique. During validation of the model, using an external test set, 21 of 25 were correctly classified and six of the nine targeted plants showed no misclassifications for the selected test set. For the other three a success rate of 50% was obtained. Mid-IR combined with SIMCA can therefore be applied as a first step in the screening of unknown samples, before applying more sophisticated fingerprint approaches or identification tests described in several national and international pharmacopoeia. As a proof of concept five real suspicious samples were successfully screened for the targeted regulated plants.

The use of Stationary Phase Optimized Selectivity Liquid Chromatography for the development of herbal fingerprints to detect targeted plants in plant food supplements.

The consumption of plant food supplements is increasing steadily and more and more, these products are bought through internet. Often the products sold through internet are not registered or declared with a national authority, meaning that no or minimal quality control is performed and that they could contain herbs or plants that are regulated. Stationary Phase Optimized Selectivity Liquid Chromatography (SOS-LC) was evaluated for the development of specific fingerprints, to be used for the detection of targeted plants in plant food supplements. Three commonly used plants in plant food supplements and two regulated plants were used to develop fingerprints with SOS-LC. It was shown that for all plants specific fingerprints could be obtained, allowing the detection of these targeted plants in triturations with different herbal matrices as well as in real samples of suspicious supplements seized by the authorities. For three of the five plants a more specific fingerprint was obtained, compared to the ones developed on traditional columns described in literature. It could therefore be concluded that the combination of segments of different types of stationary phases, as used in SOS-LC, has the potential of becoming a valuable tool in the quality control and the identification of crude herbal or plant material and in the detection of regulated plants in plant food supplements or other herbal preparations.

Chromatographic fingerprinting as a strategy to identify regulated plants in illegal herbal supplements.

Erectile dysfunction (ED) is a sexual disorder characterized by the inability to achieve or maintain a sufficiently rigid erection. Despite the availability of non-invasive oral treatment options, many patients turn to herbal alternatives. Furthermore, herbal supplements are increasingly gaining popularity in industrialized countries and, as a consequence, quality control is a highly important issue. Unfortunately, this is not a simple task since plants are often crushed and mixed with other plants, which complicates their identification by usage of classical approaches such as microscopy. The aim of this study was to explore the potential use of chromatographic fingerprinting to identify plants present in herbal preparations intended for the treatment of ED. To achieve this goal, a HPLC-PDA and a HPLC-MS method were developed, using a full factorial experimental design in order to acquire characteristic fingerprints of three plants which are potentially beneficial for treating ED: Epimedium spp., Pausinystalia yohimbe and Tribulus terrestris. The full factorial design demonstrated that for all three plant references a C8 column (250mm×4.6mm; 5µm particle size) is best suited; methanol and an ammonium formate buffer (pH 3) were found to be the best constituents for the mobile phase. The suitability of this strategy was demonstrated by analysing several self-made triturations in three different botanical matrices, which mimic the influential effects that could be expected when analysing herbal supplements. To conclude, this study demonstrates that chromatographic fingerprinting could provide a useful means to identify plants in a complex herbal mixture.

Clustering and diagnostic modelling of slimming aids based on chromatographic and mass spectrometric fingerprints.

Herbal medicines and food supplements intended as slimming aids are increasingly gaining popularity worldwide, especially for treating obesity. In this study, an ultra-performance liquid chromatography coupled to photodiode array detection (UPLC-PDA) and an ultra-performance liquid chromatography mass spectrometry (UPLC-MS) method were developed to analyze 92 slimming aids (confiscated by customs), aimed at acquiring highly informative fingerprints. Three types of fingerprints were acquired (PDA, Total Ion Chromatograms (TIC), and MS fingerprints) which were used in the chemometric data analysis. Both unsupervised (i.e., Hierarchical Cluster Analysis (HCA)) and supervised techniques (i.e., Classification and Regression Tree (CART) and Partial Least Squares - Discriminant Analysis (PLS-DA)) were applied. The aim was to perform an in-depth study of the samples, thereby exploring potential patterns present in the data. HCA was able to generate a clustering which was mainly defined by chemical compounds detected in the samples, i.e., sibutramine, phenolphthalein and amfepramone. PLS-DA generated the best diagnostic models for both PDA and TIC fingerprints, characterized by correct classification rates of external validation of 85% and 80%, respectively. For the MS fingerprints, the best model was obtained by CART (65% correct classification rate of external validation). Despite a lower correct classification rate, exploration of the concerned misclassifications revealed that the MS fingerprints proved to be superior since even very low concentrations of sibutramine could be detected. This study shows that reliable chemometric models can be obtained, based on the presence of prohibited chemical substances, which allow high-throughput data analysis of such samples. Moreover, they generate a prime notion of potential threat to a patient's health posed by these kinds of slimming aids. Copyright © 2016 John Wiley & Sons, Ltd.

Overview of skin whitening agents with an insight into the illegal cosmetic market in Europe.

Lightening skin tone is an ancient and well-documented practice, and remains common practice among many cultures. Whitening agents such as corticosteroids, tretinoin and hydroquinone are medically applied to effectively lighten the skin tone of hyperpigmented lesions. However, when these agents are used cosmetically, they are associated with a variety of side-effect. Alternative agents, such as arbutin and its derivatives kojic acid and nicotinamide have been subsequently developed for cosmetic purposes. Unfortunately, some cosmetics contain whitening agents that are banned for use in cosmetic products. This article provides an overview of the mode of action and potential side-effects of cosmetic legal and illegal whitening agents, and the pattern of use of these types of products. Finally, an EU analysis of the health problems due to the presence of illegal products on the market is summarized.

Discriminating nicotine and non-nicotine containing e-liquids using infrared spectroscopy.

In a few countries, including Belgium, nicotine-containing e-cigarettes and e-liquids are considered medicines, and therefore cannot freely be sold, but should be distributed in a pharmacy. The fact that in the neighbouring countries these products are freely available, poses a problem for custom personnel, the more the nicotine content of the products is not always labelled, especially when they are bought through internet. Therefore there is a need for easy-to-use equipment and methods to perform a first on site screening of intercepted samples, both for border control as to check label compliance of the sample. The use of attenuated total reflectance-infrared spectroscopy (ATR-IR) and near infrared spectroscopy (NIR), combined with chemometrics was evaluated for the discrimination between nicotine containing and non-nicotine containing samples. It could be concluded that both ATR-IR and NIR could be used for the discrimination when combined with the appropriate chemometric techniques. The presented techniques do not need sample preparation and result in models with a minimum of false negative samples. If a large enough training set can be established the interpretation can be fully automated, making the presented approach suitable for on-site screening of e-liquid samples.

Chromatographic impurity fingerprinting of genuine and counterfeit Cialis® as a means to compare the discriminating ability of PDA and MS detection.

Public health is threatened worldwide by counterfeit medicines. Their quality, safety and efficacy cannot be guaranteed since no quality control is performed during and/or after the manufacturing process. Characterization of these products is a very important topic. During this study a High Performance Liquid Chromatography-Photodiode Array (HPLC-PDA) and a High Performance Liquid Chromatography - Mass Spectrometry (HPLC-MS) method were developed to analyse both genuine and counterfeit samples of Cialis®. The obtained PDA and MS fingerprints were explored and modelled using unsupervised Principal Component Analysis (PCA) and supervised Partial Least Squares and its discriminant variant (PLS, PLS-DA) as well the classification methods including Soft Independent Modelling of Class Analogy (SIMCA) and the k Nearest Neighbour classifier (kNN). Both MS1 and MS2 data and data measured at 254 nm and 270 nm were used with the aim to test the potential complementarity of PDA and MS detection. First, it was checked if both groups of fingerprints can support differentiation between genuine and counterfeit medicines. Then, it was verified if the obtained multivariate models could be improved by combining information present in MS and PDA fingerprints. Survey of the models obtained for the 254 nm data, 270 nm data and 254_270 nm data combination showed that a tendency of discrimination could be observed with PLS. For the 270 nm data and 254_270 nm data combination a perfect discrimination between genuine and counterfeit medicines is obtained with PLS-DA and SIMCA. This shows that 270 nm alone performs equally well compared to 254_270 nm. For the MS1 and MS1_MS2 data perfect models were obtained using PLS-DA and kNN, indicating that the MS2 data do not provide any extra useful information to acquire the aimed distinction. When combining MS1 and 270 nm perfect models were gained by PLS-DA and SIMCA, which is very similar to the results obtained for PDA alone. These results show that both detectors have a potential to reveal chemical differences between genuine and counterfeit medicines and thus enable the construction of diagnostic models with excellent recognition. However, if a larger sample set, including more possible sources of variation, is analysed more sophisticated techniques such as MS might be necessary.

Comparison of three development approaches for Stationary Phase Optimised Selectivity Liquid Chromatography based screening methods Part II: A group of structural analogues (PDE-5 inhibitors in food supplements).

Three approaches for the development of a screening method to detect adulterated dietary supplements, based on Stationary Phase Optimised Selectivity Liquid Chromatography were compared for their easiness/speed of development and the performance of the optimal method obtained. This comparison was performed for a heterogeneous group of molecules, i.e. slimming agents (Part I) and a group of structural analogues, i.e. PDE-5 inhibitors (Part II). The first approach makes use of primary runs at one isocratic level, the second of primary runs in gradient mode and the third of primary runs at three isocratic levels to calculate the optimal combination of segments of stationary phases. In each approach the selection of the stationary phase was followed by a gradient optimisation. For the PDE-5 inhibitors, the group of structural analogues, only the method obtained with the third approach was able to differentiate between all the molecules in the development set. Although not all molecules are baseline separated, the method allows the identification of the selected adulterants in dietary supplements using only diode array detection. Though, due to the mobile phases used, the method could also be coupled to mass spectrometry. The method was validated for its selectivity following the guidelines as described for the screening of pesticide residues and residues of veterinary medicines in food.

Comparison of three development approaches for Stationary Phase Optimised Selectivity Liquid Chromatography based screening methods Part I: A heterogeneous group of molecules (slimming agents in food supplements).

Three approaches for the development of a screening method to detect adulterated dietary supplement, based on Stationary Phase Optimised Selectivity Liquid Chromatography were compared for their easiness/speed of development and the performance of the optimal method obtained. This comparison was performed for a heterogeneous group of molecules, i.e. slimming agents (Part I) and a group of structural analogues, i.e. PDE-5 inhibitors (Part II). The first approach makes use of primary runs at one isocratic level, the second of primary runs in gradient mode and the third of primary runs at three isocratic levels to calculate the optimal combination of segments of stationary phases. In each approach the selection of the stationary phase was followed by a gradient optimisation. For the slimming agents, the heterogeneous group of molecules, the method obtained with the first approach was selected as optimal, based on the speed of development and the performance of the method. The method shows a good separation of the compounds, allowing the screening to be performed with diode array detection, and is fully compatible with mass spectrometry. The method was validated for its selectivity following the guidelines as described for the screening of pesticide residues and residues of veterinary medicines in food.

Development of a Stationary Phase Optimised Selectivity Liquid Chromatography based screening method for adulterations of food supplements for the treatment of pain.

Illegally adulterated dietary supplements are an increasing problem worldwide. One of the important groups of often adulterated products are the dietary supplements, sold for the treatment of pain. These often contain analgesics, a heterogeneous group of molecules, containing both hydrophilic and hydrophobic compounds. The development of a screening method for these components, especially when mass spectrometric detection is not available, necessitates chromatographic separation, difficult to achieve with traditional chromatographic columns. In this paper Stationary Phase Optimised Selectivity Liquid Chromatography was used for the development of a screening method for nine analgesics, codeine and caffeine, often present in this type of dietary supplements. The method shows a good separation of all the compounds, allowing the screening to be performed with diode array detection and is fully compatible with mass spectrometry. The method was validated for its selectivity following the guidelines as described for the screening of pesticide residues and residues of veterinary medicines in food.

HS-GC-MS method for the analysis of fragrance allergens in complex cosmetic matrices.

Potential allergenic fragrances are part of the Cosmetic Regulation with labelling and concentration restrictions. This means that they have to be declared on the ingredients list, when their concentration exceeds the labelling limit of 10 ppm or 100 ppm for leave-on or rinse-off cosmetics, respectively. Labelling is important regarding consumer safety. In this way, sensitised people towards fragrances might select their products based on the ingredients list to prevent elicitation of an allergic reaction. It is therefore important to quantify potential allergenic ingredients in cosmetic products. An easy to perform liquid extraction was developed, combined with a new headspace GC-MS method. The latter was capable of analysing 24 volatile allergenic fragrances in complex cosmetic formulations, such as hydrophilic (O/W) and lipophilic (W/O) creams, lotions and gels. This method was successfully validated using the total error approach. The trueness deviations for all components were smaller than 8%, and the expectation tolerance limits did not exceed the acceptance limits of ± 20% at the labelling limit. The current methodology was used to analyse 18 cosmetic samples that were already identified as being illegal on the EU market for containing forbidden skin whitening substances. Our results showed that these cosmetic products also contained undeclared fragrances above the limit value for labelling, which imposes an additional health risk for the consumer.

ATR-FTIR spectroscopy and chemometrics: An interesting tool to discriminate and characterize counterfeit medicines.

Counterfeit medicines pose a huge threat to public health worldwide. High amounts of counterfeit pharmaceuticals enter the European market and therefore detection of these products is essential. Attenuated Total Reflection Fourier-Transform infrared spectroscopy (ATR-FTIR) might be useful for the screening of counterfeit medicines since it is easy to use and little sample preparation is required. Furthermore, this approach might be helpful to customs to obtain a first evaluation of suspected samples. This study proposes a combination of ATR-FTIR and chemometrics to discriminate and classify counterfeit medicines. A sample set, containing 209 samples in total, was analyzed using ATR-FTIR and the obtained spectra were used as fingerprints in the chemometric data-analysis which included Principal Component Analysis (PCA), k-Nearest Neighbours (k-NN), Classification and Regression Trees (CART) and Soft Independent Modelling of Class Analogy (SIMCA). First it was verified whether the mentioned techniques are capable to distinguish samples containing different active pharmaceutical ingredients (APIs). PCA showed a clear tendency of discrimination based on the API present; k-NN, CART and SIMCA were capable to create suitable prediction models based on the presence of different APIs. However k-NN performs the least while SIMCA performs the best. Secondly, it was tested whether these three models could be expanded to discriminate between genuine and counterfeit samples as well. k-NN was not able to make the desired discrimination and therefore it was not useful. CART performed better but also this model was less suited. SIMCA, on the other hand, resulted in a model with a 100% correct discrimination between genuine and counterfeit drugs. This study shows that chemometric analysis of ATR-FTIR fingerprints is a valuable tool to discriminate genuine from counterfeit samples and to classify counterfeit medicines.

Detection of sibutramine in adulterated dietary supplements using attenuated total reflectance-infrared spectroscopy.

Sibutramine is one of the most occurring adulterants encountered in dietary supplements with slimming as indication. These adulterated dietary supplements often contain a herbal matrix. When customs intercept these kind of supplements it is almost impossible to discriminate between the legal products and the adulterated ones, due to misleading packaging. Therefore in most cases these products are confiscated and send to laboratories for analysis. This results inherently in the confiscation of legal, non-adulterated products. Therefore there is a need for easy to use equipment and techniques to perform an initial screening of samples. Attenuated total reflectance-infrared (ATR-IR) spectroscopy was evaluated for the detection of sibutramine in adulterated dietary supplements. Data interpretation was performed using different basic chemometric techniques. It was found that the use of ATR-IR combined with the k-Nearest Neighbours (k-NN) was able to detect all adulterated dietary supplements in an external test set and this with a minimum of false positive results. This means that a small amount of legal products will still be confiscated and analyzed in a laboratory to be found negative, but no adulterated samples will pass the initial ATR-IR screening.