Chemical Authentication of Botanical Ingredients: A Review of Commercial Herbal Products

Chemical methods are the most important and widely used traditional plant identification techniques recommended by national and international pharmacopoeias. We have reviewed the successful use of different chemical methods for the botanical authentication of 2,386 commercial herbal products, sold in 37 countries spread over six continents. The majority of the analyzed products were reported to be authentic (73%) but more than a quarter proved to be adulterated (27%). At a national level, the number of products and the adulteration proportions varied very widely. Yet, the adulteration reported for the four countries, from which more than 100 commercial products were purchased and their botanical ingredients chemically authenticated, was 37% (United Kingdom), 31% (Italy), 27% (United States), and 21% (China). Simple or hyphenated chemical analytical techniques have identified the total absence of labeled botanical ingredients, substitution with closely related or unrelated species, the use of biological filler material, and the hidden presence of regulated, forbidden or allergenic species. Additionally, affecting the safety and efficacy of the commercial herbal products, other low quality aspects were reported: considerable variability of the labeled metabolic profile and/or phytochemical content, significant product-to-product variation of botanical ingredients or even between batches by the same manufacturer, and misleading quality and quantity label claims. Choosing an appropriate chemical technique can be the only possibility for assessing the botanical authenticity of samples which have lost their diagnostic microscopic characteristics or were processed so that DNA cannot be adequately recovered.


INTRODUCTION
Herbal products are being sold under many and diverse commercial descriptions in the international marketplace, including herbal drugs, botanical drugs, botanicals, phytomedicines, traditional medicines (TMs), herbal medicines (HMs), traditional herbal medicines products (THMPs), natural health products (NHPs), dietary supplements (DSs), plant food supplements (PFSs), nutraceuticals (NCs) and food supplements (FSs) (Ichim, 2019), the differences being mainly due to the prevailing national legislation under which they are marketed (Simmler et al., 2018). Herbal products are commercialized as medicines or foods, according to their officially declared intended final use by their manufacturers operating under various regulatory frameworks, and they are purchased, and subsequently used and consumed, for their medicinal claims (herbal medicines) or their expected health benefits (food supplements) (Thakkar et al., 2020). In the United Kingdom, for example, plant products are regulated under two main criteria, the first being what is claimed, i.e. if a manufacturer claims a medicinal effect, the product will automatically fall under medicines legislation; the second consideration being the activity of the plant in vivo, if it has shown to have a strong medicinal or pharmacological action then it is deemed a medicine regardless of the claims, the most notable plant in this category being Hypericum perforatum L. (St John's Wort). Whereas in the United States most plant products are regulated as food supplements (botanicals) and in Germany the majority are considered medicines. Unfortunately, these marketing differences, due to significant differences between the regulatory approaches across jurisdictions (Low et al., 2017), are further contributing to their poor regulation on the international market.
Accidental contamination or the deliberate use of filler or substitute species (Shanmughanandhan et al., 2016) leads inherently to non-authentic, adulterated products (Simmler et al., 2018). The adulteration of commercial herbal products is an internationally widespread problem, as it has been reported for many countries from all inhabited continents (Ichim, 2019;Ichim et al., 2020). Moreover, large percentages of adulterated products have been reviewed, irrespective of the formal category of herbal products, being affected food and dietary supplements and medicines altogether (Ichim and de Boer, 2021), including products used in centuries or even millennia-old Ayurveda (Revathy et al., 2012;Seethapathy et al., 2019) and Asian traditional medicine systems (Masada, 2016;Xu et al., 2019). The substantial proportion of adulterated commercial herbal products described appears to be independent of the methods used for their analysis, traditional pharmacopoeial methods being employed, such as macroscopic inspection (van der Valk et al., 2017), microscopy (Ichim et al., 2020), chemical techniques (Li et al., 2008;Upton et al., 2020), or even the more recently developed DNA-based ones, such as the rapidly technologically evolving DNA barcoding and metabarcoding (Ichim, 2019;Grazina et al., 2020).
On the global market, herbal products are sold in an extremely diverse variety of forms, from single ingredient, unprocessed, raw, whole plants to multi-species, highly processed extracts. Therefore, the successful authentication of commercial herbal products reported by peer reviewed studies are a valuable and useful source of information which provide the necessary practicalities, including their strengths and the limitations, of employing the right methods for a specific type of product along the length of its value chain (Booker et al., 2012). Such analyses of peer-reviewed authentication reports focused exclusively on commercial herbal products have concluded that, microscopy, a traditional pharmacopoeial identification method, is costefficient and can cope with mixtures and impurities but it has limited applicability for highly processed commercial samples e.g. extracts (Ichim et al., 2020). On the other hand, DNA-based identification, only recently adopted by the first two national Pharmacopoeias (Pharmacopoeia Committee of P. R. China, 2015;British Pharmacopoeia Commission, 2018), facilitate simultaneous multi-taxa identification by using the DNA of different origins extracted from complex mixtures and matrices but false-negatives can be expected if the DNA has been degraded or lost during post-harvest processing or manufacturing (Raclariu et al., 2018a;Ichim, 2019;Grazina et al., 2020). In this respect, our review adds the much needed peer-reviewed, systematically searched information, about the successful use of chemical identification for the authentication of commercial herbal products. While doing so, our review also provides some missing pieces of the commercial herbal products' authenticity puzzle.

Search Strategy
Four databases were systematically searched for peer reviewed records following the PRISMA guidelines (Moher et al., 2009) using combinations of relevant keywords, Boolean operators and wildcards: [("herbal product" OR "herbal medicine" OR "traditional medicine" OR "food supplement" OR "dietary supplement" OR "herbal supplement" OR nutraceutical) AND (authentic* OR contaminat* OR substitut*)] for Web of Science, PubMed, Scopus, and [("herbal product" OR "herbal medicine" OR "food supplement" OR "dietary supplement" OR "herbal supplement" OR nutraceutical) AND (authentication OR contamination OR substitution)] for ScienceDirect. The option "search alert" was activated for all four databases, to receive weekly updates after the literature search was performed. Furthermore, we used cross-referencing to identify additional peer-reviewed publications.

Selection Process and Criteria
Identification: 10,497 records were identified through database searching (WoS 1,317,PubMed 3,253,Scopus 5,446,and ScienceDirect 481), and 196 additional records from crossreferencing and the weekly updates from the four databases. Screening: after the duplicates had been removed, 2,326 records were collected and their abstracts screened. After screening, 1,745 records were excluded for not reporting data relevant for the chemical authentication of herbal products. Eligibility: 581 fulltext articles were assessed and screened based on the following eligibility criteria: 1) The reported products had to be "herbal products"; the full wide range of commercial names was searched for and accepted for being included in our analysis. 2) The analyzed products had to be "commercial"; keywords such as "purchased", "bought", were accepted. Our analysis excluded samples which were obtained "cost-free", a "gift" or "donated" by a person, institution or company. 3) The products had to be clearly allocated to a "country" or "territory" (e.g., European Union). 4) The conclusion "authentic"/"adulterated" had to be drawn by the authors of the analyzed studies. 5) The products had to be analyzed with a "chemical" method or techniques.                    The set of retrieved full-text articles was further reduced by 446 that did not meet all eligibility criteria. Included: 135 records.

RESULTS
Different chemical methods have been successfully employed for the botanical authentication of 2,386 commercial herbal products, sold in 37 countries spread on six continents. The majority of the analyzed products were reported to be authentic (73%) but more than a quarter proved to be adulterated (27%), when the botanical identity of their content was compared with the label stated ingredients ( Table 1).
The herbal products were purchased from 37 countries scattered over six continents: Europe (n 20), Asia (n 9), North America (n 3), Australia (n 2), South America (n 2), and Africa (n 1) (Supplementary Table S1). The numbers of reported samples were geographically heterogeneous, at continental level the highest number of commercial herbal products was reported for Asia (n 877), North America (n 767), Europe (n 573), followed distantly by South America (n 86), Australia (n 25) and Africa (n 5). The proportion of adulterated products varies significantly among continents, being highest in Africa (60%), South America (57%), Australia (44%), and lower in Europe (28%), North America (27%), and Asia (25%). The adulteration percentage of the last three continents enumerated is close to the global one (27%) which can be influenced also by the significantly higher number of commercial products analyzed and reported, compared with the samples analyzed from the other three continents.
In twelve countries, out of the total of thirty-seven, all the analyzed commercial herbal products (100%) were reported as authentic, albeit, for eight of them, less than 10 samples were reported. Notably, the botanical identity of the samples purchased from South Korea (n 96) and Denmark (n 12) matched the labeled information. The adulterated proportion in the remaining twenty-five countries varied widely, from 8% up to as much as 80%. From the countries where more than 10 samples from their marketplace have been chemically authenticated and nonauthenticated products have been reported, the majority of the commercial products was adulterated, being the highest in India (69%), followed closely by Australia (58%), Brazil (58%), Turkey (57%) and Malaysia (51%). Noticeably, the adulteration percentage of the four countries with more than 100 commercial products reported is 37% (United Kingdom), 31% (Italy), 27% (United States) and the lowest is reported for China (21%).

Sampling Heterogeneity and Unavoidable Bias
The authentication raw data were all retrieved from peerreviewed articles, the vast majority of them after they were indexed in the four major international databases which were systematically searched for while some other few articles were identified after cross-referencing. Although no limiting criteria (e.g. publication year, or language) was used, the authentication data reported in journals with limited-impact and international visibility might be underrepresented in the retrieved data. Moreover, the access of researchers from the economically depressed economies to high-impact journals, and especially to the OA journals, is a further limiting factor for publicly communicating the authentication results relevant for a certain country. On the other hand, as it was previously mentioned as possible bias, also the countries with a functional consumer safety system might be underrepresented as the authentication results of the commercial samples screened by the respective institutions will be published in internal bulletins or protocols, rather than in peer-reviewed journals (Ichim et al., 2020).

DISCUSSION
The chemical identification methods have confirmed that a substantial proportion (27%) of herbal products from the international market place is adulterated: on average, more than one in each four products sold in the 37 countries included in our analysis was proved to be non-authentic regarding their botanical identity. This adulteration percentage, revealed by employing many and very diverse chemical analytical methods, almost matches the figure obtained after the use of DNA-based techniques were assessed for their use for the authentication of commercial herbal products in a comparable number of countries: 27% (Ichim, 2019). Indeed, this percentage was obtained after almost a triple number of commercial herbal products (n 5,957) were analyzed and their results reviewed recently. Notably, the microscopic authentication of commercial herbal products have reported a much higher adulteration rate (41%) but the number of analyzed samples was considerably much smaller (n 508) which can be a possible bias of this finding (Ichim et al., 2020).
As it was previously reported by many peer-reviewed reports (Hoban et al., 2018;Seethapathy et al., 2019;Amritha et al., 2020;Anthoons et al., 2021;Palhares et al., 2021), irrespective of the authentication method, adulterated commercial HPs are geographically present across all continents (Supplementary Table S1). Moreover, this highly relevant category of commercial products was found to not comply with the labeled botanical ingredients in proportions almost identical (26 ± 2%), irrespective if they are traditionally used as herbal medicines, as commonly found in Asia, or overwhelmingly consumed as food supplements as in Europe or North America. These two main categories of herbal products commercialized in the global marketplace have many types of value chains (Booker et al., 2012), with some different stakeholders and entities along their shorter or more complex trade chains. Nevertheless, the end-users of both systems seem to be equally affected by non-authentic, accidental contamination or fraudulent substitution of labeled botanical ingredients and even the addition of compounds in an attempt to fool quality control testing e.g. as in adding food dyes to H. perforatum in order to achieve higher UV spectroscopy readings (Booker et al., 2018). Indeed, although monographs for herbal raw materials (e.g., Ph. Eur, USP) allow a minor presence of foreign organic matter (Parveen et al., 2016), the adulteration patters documented by employing different chemical methods, are very diverse and most of them are made possible only by the intentional, economically motivated and fraudulent actions of onerous producers or traders.
The total absence of labeled botanical ingredients and/or their extracts from the commercial herbal products tested was detected by using chemical methods. Commercial samples devoid of labeled botanical ingredient species (Carlson and Thompson, 1998;Ardila et al., 2015;Geng et al., 2019;Zhu et al., 2019) or not even substituted with their related species (Wan et al., 2016). An easy way to increase the profit margin of the products was the use of cheaper plant material as it was the use of other plant parts than the ones recommended, labeled and expected by the product's users, senna (Senna alexandrina Mill.) stems substituted with leaves and midribs (Kojima et al., 2000), Panax ginseng C.A.Mey roots with other plant parts (leaf or stem) (Govindaraghavan, 2017), or Panax notoginseng Burkill F.H.Chen roots with flowers (Liu et al., 2015). Another similar deceptive adulteration strategy was the reported use of extracts obtained from plant parts other than the recommended ones, such as the decoction of the stem bark to substitute the genuine "jatoba" sap products (Hymenaea stigonocarpa Hayne, Hymenaea martiana Hayne) and the adulteration of Aquilariae Lignum Resinatum (Aquilaria sinensis (Lour.) Spreng) products with cheap resin (e.g. rosin) (Qu et al., 2017).
Frontiers in Pharmacology | www.frontiersin.org April 2021 | Volume 12 | Article 666850 The economically motivated adulteration includes also the use of unlabeled filler species as the DNA of species such as rice (Oryza sativa L.), soybean (Glycine max (L.) Merr.) and wheat (Triticum spp.) was previously identified in commercial herbal products (Newmaster et al., 2013;Ivanova et al., 2016). Yet, the TLC alone was able to detect the fraudulent use of soybean oil as filler in "copaiba" (Copaifera multijuga Hayne) oil-resin products (Barbosa et al., 2009). The detection of unlabeled species with allergenic potential and known or suspected toxicity was previously reported by the use of DNA-based authentication techniques (Newmaster et al., 2013;Speranskaya et al., 2018). The same potential was shown by the phytochemical analyses which have been able to unmask the presence of unwanted and hazardous botanic ingredients, such as species that should have been notified to authorities (e.g. Ilex paraguariensis A. St-Hil., Epimedium spp., Tribulus terrestris L.), or forbidden toxic plants (e.g. Aristolochia fangchi Y.C.Wu exL.D.Chow and S.M.Hwang) (Deconinck et al., 2019) or even health hazardous contaminations, with Digitalis lanata Ehrh. added to plantain (Plantago major L.) products (Slifman et al., 1998). Moreover, as peanut allergy is a major public health concern and can be severe or even life-threatening (Gray, 2020), chemical methods have proved able to detect adulteration with the peanut skin extract of grape seedcontaining herbal products (Vitis vinifera L.) from Australia (Govindaraghavan, 2019) and United States (Villani et al., 2015).
All the intentional adulteration practices documented and reported repeatedly till now (Li et al., 2008;Ichim, 2019;Xu et al., 2019;Ichim et al., 2020;Upton et al., 2020) can be evidenced by peer-reviewed reports referring to the top selling herbal products containing highly valued or widely used medicinal species across countries and cultures. The prices of ginseng herbal medicines and supplements vary widely based on the species, quality, and purity of the ginseng, and this provides a strong driver for intentional adulteration (Ichim and de Boer, 2021). Indeed, several chemical methods were able to identify ginseng products totally or partially devoid of the labeled P. ginseng plant material (Mihalov et al., 2000;Yang et al., 2016) and prove that, in most cases, labeled Panax species were substituted with other Panax species (Li et al., 2010;Yu et al., 2014;Dong et al., 2020), but also the substitution of ginseng root with leaves, stems or flowers (Liu et al., 2015;Govindaraghavan, 2017). Notably, chemical analysis was even able to detect the adulteration and substitution of wild with cultivated ginseng  as well as a white ginseng products (P. ginseng) not composed of 6 years old ginseng radix only (Li et al., 2010).
Studies carried out at UCL School of Pharmacy, London have consistently shown that product adulteration is commonplace, with 25-40% of products typically being found to be of poor quality or adulterated, and especially with products obtained via the internet. Although with products that have been registered as Traditional Herbal Medicines under the Traditional Herbal Medicinal Products Directive (THMPD), no adulteration has so far been found and these products have shown to be of acceptable quality (Booker et al., 2016a;Booker et al., 2016b;Booker et al., 2018). This does not necessarily mean that all non-registered products (e.g. food supplements) are of poor quality but the problem being that it is difficult for the general public to be able to reliably discern high quality products from inferior ones. Organic certification provides some assurances regarding traceability, including origin, cultivation methods and manufacturing practices and so until more formal regulations are introduced for these food supplement products, buying organic may be the best option.
The many cases of substituted or adulterated herbal products purchased from a very high number of national marketplaces, where the labeled botanical ingredients did not match the chemically identified ones are, unfortunately, accompanied by other low-quality issues which additionally affect the safety and potential efficacy of commercial herbal products. As many as forty-one peer reviewed research articles, which have reported a case of adulteration among analyzed commercial samples, have also reported other quality issues which further lower the overall quality expected by their users and consumers. Additionally, another nineteen studies reported quality issues of the tested products without identifying any proof for their botanical identity adulteration. For the majority of herbal products reported, considerable variability of their labeled metabolic profile and/ or content, such as the alkaloid content of "ma-huang" (Ephedra sinica Stapf) products (Gurley, 1998) or Menispermi Rhizoma (Menispermum dauricum DC) products (Liu et al., 2013b), selected triterpene glycosides and phenolic constituents in black cohosh (A. racemosa) products (Jiang et al., 2006) or the PAC content of cranberry products (Turbitt et al., 2020). Furthermore, aside of significant product-to-product variability, the marked differences of the content of individual flavonoids/flavonolignans in milk thistle (Silybum marianum (L.) Gaertn.) products have revealed quality difference also between different batches by the same manufacturers (Fenclova et al., 2019).
The peer-reviewed authentication results and the methods which were successfully employed to analyze commercial herbal products and significantly contribute to a better understanding of authenticity issues affecting the herbal industry and provides an as close-to-reality possible picture of the commercial herbal products' authenticity as well as examples of techniques to be efficiently and accurately used for their authentication.
It is clear that chemical analysis alone can only identify existing problems. In order to prevent these problems from arising in the first place, better governance needs to be implemented along all stages of the supply chain. Regulation can help with this process but resources are scarce and real progress on quality is more achievable through having closer and more focused co-operation between the regulators and the producers, manufacturers and retailers of herbal products.

AUTHOR CONTRIBUTIONS
MI performed the literature systematic search and analyzed the results. MI and AB wrote the manuscript together.