Phytocannabinoids are characteristic metabolites of Cannabis sativa L. Approximately 200 cannabinoids have been discovered from this species, each with unique therapeutic and/or abuse potential. These bioactive terpenoids have recently been discovered in Rhododendron and Radula species, as well as in certain legumes and fungi (Gülck and Møller, 2020; Govindarajan et al., 2023; Hesami et al., 2023; Sainz Martinez et al., 2023). Cannabinoids are insoluble in water but soluble in alcohols, non-polar organic solvents (e.g., ether and hexane), and lipids. That is why the first products encountered in 2015 were mostly oil-based (Messina et al., 2015; Thomas and ElSohly, 2016). The revival of plant-based products and the steady shift in attitude surrounding Cannabis sativa L. consumption created an opportunity for cannabidiol (CBD) products, which inundated offline and online health markets (Hazekamp, 2018). In 2018, the “Farm Bill,” known as the Agricultural Improvement Act, permitted hemp-derived products to be treated as agricultural products, with a restriction of maximum 0.3% (w/w) tetrahydrocannabinol (THC) content by dry weight, and removed these products from the list of controlled substances under the supervision of the Drug Enforcement Administration (DEA) in the United States (Drug Enforcement Administration, 2018; Dabrowska et al., 2020). The opioid epidemic, which raised demand for non-opioid analgesic alternatives, also contributed to the emergence of the Cannabis and CBD product markets (Olfson et al., 2018).

The rise of CBD products began in 2014–2015, with the introduction of numerous food supplements not only in the United States but also in Europe. Since its distribution as food is incompatible with the novel food regulation (European Parliament, 2015), it can no longer be sold as a food supplement or conventional food ingredient. As a result, manufacturers resorted to other non-medical health product categories, such as medical devices or cosmetics. This legal ambiguity, along with the increase in demand over the past decade, has spawned a multibillion-dollar global market (Sarma et al., 2020; Bhamra et al., 2021; Miller et al., 2022; Lachenmeier et al., 2023).

Given that these plant-based products were initially marketed as food supplements, it is not surprising that health claims for CBD products have appeared. Regulatory agencies identified several regulatory violations, including the use of medical claims, and such incidents are still frequent among these products (Amann et al., 2022; Spindle et al., 2022). The potential long-term adverse health effects, including liver toxicity, possible harm to the male reproductive system, and the increased risk of drug interactions, also call for stricter control of CBD-containing products and emphasize the dangers of their uncontrolled use as food supplements (Huestis et al., 2019).

Globally, there are five cannabinoid-containing products authorized by regulatory agencies as medicine: cannabidiol (Epidiolex^® or Epidyolex^®), dronabinol/delta-tetrahydro-cannabinol (Marinol^®, Syndros^®, Reduvo^®, and Adversa^®), nabilone (Cesamet^® and Canemes^®), and nabiximols (CBD:THC = 1:1; Sativex^®). The Summary of Product Characteristics (SmPC) of these products may help identify the potential adverse health consequences of unregulated CBD oil and specific patient populations that will not benefit from their use (Kalant and Porath-Waller, 2016; Bajtel et al., 2021; Schlag et al., 2022).

National regulatory agencies are responding differently to the unresolved issue of CBD products because the current regulatory frameworks are incapable of mitigating these risks, and new approaches are required. The Hungarian National Institute of Pharmacy and Nutrition—similarly to other national authorities—has chosen a restrictive approach (Food and Drug Administration, 2019; Hughes et al., 2022; Hungarian National Institute of Pharmacy and Nutrition, 2022; Woodcock, 2023). In the European Union, foods containing CBD extracts and CBD-enriched foods are considered novel foods, and their distribution is, therefore, prohibited, regardless of their THC content, as cannabinoids are not permitted as novel foods. Only hemp seed and products derived from hemp seed processing (e.g., hemp seed oil) may be used in food supplements, provided that the CBD level does not exceed the impurity limit. Similarly, THC content, regarded as an impurity, needs to be evaluated case by case, taking into account the overall characteristics of the product.

Globally, hemp and CBD products may be subject to varying regulations, which may lead to confusion. According to current European legislation, the generally accepted cannabinoid profile of hemp is 1.5%–3% CBD and less than 0.2% THC. For native and produced hemp seeds, the maximum THC content is 0.3%, while the maximum THC content for hemp seed oil is 0.75% (European Commission, 2023). CBD concentrations can surpass 10% in the United States of America, but THC concentrations cannot exceed 0.3% on a dry-weight basis. In Hungary, products containing less than 10 mg/kg THC are permitted; CBD supplements and novel foods made from cannabis seeds must contain less than 0.2 mg/kg of THC per kilogram of dry weight, and the maximum quantity of CBD is 25 mg/kg (EMCDDA, 2018; Kirilov et al., 2020).

There is public pressure on pharmaceutical authorities to establish clear and concise regulatory frameworks for CBD products, as information regarding product content is frequently restricted or unknown, and lot-to-lot variation and long-term stability are debatable. International organizations involved in the regulation of controlled substances, such as the International Narcotics Control Board (INCB), likewise struggle with the questions and problems associated with these products. Although the INCB has clarified that CBD is not under international control in 2020, the European regulation of Cannabis sp. products between Member States is still a matter of debate (Evans, 2020; Adelstone, 2022; Johnson et al., 2022b).

Due to the ongoing issues and debates around CBD oils, it is of public health significance to call the attention of consumers and policymakers to prevent misperceptions about CBD products and reduce the potential harms and health risks associated with their use.

Our aim was to analyze the CBD content and label accuracy of CBD oil products available on the internet. Furthermore, we aimed to evaluate the health claims presented on the vendor’s website and on the test-purchased products and discuss the strength of the evidence behind the claimed health benefits.

Our study consisted of three sections. First, we conducted an online market analysis, focusing on the content analysis of web stores selling CBD oils for consumers in an effort to identify potentially misleading health claims. Second, selected CBD oils were test-purchased online, and the packaging and labeling of products were evaluated to assess the reliability of the information provided on the delivered products. Finally, product quality was assessed using a quantitative analysis of the CBD content.

In 2015 and 2016, the FDA conducted the first research assessing the content and labeling accuracy of CBD products and issued warning letters (warning of products with negligible or less than 1% CBD content) (Food and Drug Administration, 2021). In 2016, Bonn-Miller et al. ordered 84 CBD products and analyzed their content and labeling. Only 31% of the products were within ±10% of the stated CBD content (concentration ranged between 0.10 mg/mL and 655.27 mg/mL), and less than 50% (n = 18) were appropriately labeled. In general, the concentration of unlabeled cannabinoids was modest; nonetheless, THC was detected in 21.43% (18/84) of the products purchased (up to 6.43 mg/mL) (Bonn-Miller et al., 2017).

The officers of the Mississippi Bureau of Narcotics in 2020 purchased CBD-containing products from offline retailers in the state of Mississippi, according to a separate study. Of the 25 oils and electronic cigarette vaping products analyzed, 88% (n = 22) contained measurable quantities of CBD, and only three products (12%) were within 20% of the label claim. Three products contained THC, and four products (three vaping liquids) were adulterated with synthetic cannabinoids (4-fluoro MDMB-BUTINACA, 5-fluoro MDMB-PICA, and 5-fluoro MDMB-PINACA) (Gurley et al., 2020). Topical CBD products were purchased in local shops and online in 2020 in Baltimore, Maryland, and of the 105 products, 8% were over-labeled, 58% were under-labeled, and 24% were accurately labeled for CBD (applying the 10% rule like the other publications). THC content was detected in 35% of the products (less than 0.3%). The authors highlighted the misleading therapeutic or cosmetic claims (Spindle et al., 2022).

However, these findings are not country-specific and reflect a worldwide trend since other European studies have shown the same discrepancies in labeling accuracy as the International Cannabis and Cannabinoids Institute reported in 2017 (Cannabis and Cannabinoids Institute, 2017). In a Dutch study, 46 cannabis oil products (29 home-made and 17 web-purchased) were collected directly from consumers, and it was found that only 46% of the products had label information (CBD/THC content), seven products did not contain any measurable phytocannabinoids, and 26 samples (57%) had a THC content >1%, with one sample reaching 57.5% (575 mg/mL), whereas the CBD content ranged from 0.1% to 27% (1 mg/mL and 270 mg/mL) (Hazekamp and Epifanova, 2017). In 2018, an Italian study assessing the overall quality of CBD oil products purchased online found that only five of 14 products contained at least 10% of the CBD amount stated on the label. The CBD content varied between 0.24 and 4.96 w/w%. Additional investigation indicated that 12 of 14 contained THC (Pavlovic et al., 2018). In a 2019 study from the UK, Liebling et al. analyzed 29 products purchased online and offline from 27 different suppliers to test cannabinoid content, heavy metals, and residual solvents. Regarding the CBD content, only 11/29 (38%) products were within 10% of the labeled content. Ten percent of the overall cannabinoid content was comprised of non-CBD phytocannabinoids, including THC (0%–0.22%), THCA, THCV, CBN (0%–0.12%), CBG, CBGA, and CBC, when the other cannabinoid components were evaluated, including THC (0%–0.22%), THCA, THCV, CBN (0%–0.12%), CBG, CBGA, and CBC. Some products contained residual solvents like n-pentane, ethanol, ethyl acetate, isopropanol, heptane, and cyclohexane; in addition, small quantities of lead (0.01–0.24 ppm) and arsenic (0.01–0.06 ppm) were found in some products. All values were below the ICH guidelines for medicinal products but above safety levels of the food limit (Liebling et al., 2022). In 2021, a Belgian study investigated 18 CBD oils that were seized by inspectors and concluded that 45% of the analyzed items met the acceptance limits for the CBD content. The provenance of the products ranged from the Netherlands to Switzerland to Barcelona. Δ⁹-THC was not detected in the products analyzed (Duchateau et al., 2021).

Another study in the United States investigated hemp-derived products from both online and local retailers in Kentucky (n = 80) from 2 April to 9 May 2021. Unique to the experiment were the use of authorized CBD medicinal products as a positive control (Epidiolex) and the inclusion of only oil products. Of the 80 products, 31% were under-labeled, 15% were over-labeled, and 54% were accurately labeled. The CBD concentration varied between 9.3 and 60.5 mg/mL (24.1 ± 15.3 mg/mL). The THC levels were presented in another publication, where the authors reported that the Δ⁹-THC concentration of the 51 analyzed products (64%) ranged from 0.008 mg/mL to 2.071 mg/mL (Johnson et al., 2022a; Johnson et al., 2022b). In a study led by a pharmacy student in the southwest Wisconsin area in 2021, 39 cannabidiol products were purchased from local retail shops, and of the 11 oils, only 36.36% (n = 4) were appropriately labeled. THC was detected in 54.55% (n = 6) of the oils with a maximum concentration of 0.2% w/v (Miller et al., 2022). As part of an international study conducted in 2020–2021, 24 samples were purchased online from Austria, France, Germany, Great Britain, Luxembourg, the Netherlands, and Spain. Seven samples were within 10% of the reported CBD values (CBD and CBDA), sixteen samples were over-labeled, and one sample had a concentration that was more than five times lower than the declared concentration. One sample had a THC concentration above 0.2% w/w (the legal limit in most European countries) (Schneider, 2021).

During our literature search, we also discovered a review of label accuracy studies in which, based on the results of five studies, it was concluded that the proportion of correctly labeled products ranged from 17% to 86% and that further well-designed research is required in this field due to the paucity and heterogeneity of the available studies. In this product category, reporting the potential for adverse health effects and practicing pharmacovigilance are crucial (Bonn-Miller et al., 2017; Stevenson, 2018; Blebea et al., 2019; Herbst and Musgrave, 2020; Oldfield et al., 2021; Vandrey et al., 2022).

The ratio of accurately labeled products in our study sample was higher (n = 6; 50%) than in the previously published literature from the United States (31% by Bonn-Miller et al. in 2016, 36.36% by Miller et al. in 2021, and 24% by Spindle et al. in 2022) or Europe (23.5% by Hazekamp and Epifanova in 2017 in the Netherlands, 35.7% by Pavlovic et al. in 2018 in Italy, 38% by Liebling et al. in 2019 in the United Kingdom, 29.2% by Schneider in 2021, and 48% by Duchateau et al. in 2021 in Belgium). Similar results were found and published by Johnson et al. with 54% labeling accuracy in 2022 in the United States. (Bonn-Miller et al., 2017; Hazekamp and Epifanova, 2017; Pavlovic et al., 2018; Duchateau et al., 2021; Schneider, 2021; Johnson et al., 2022a; Johnson et al., 2022b; Liebling et al., 2022; Miller et al., 2022; Spindle et al., 2022).

Chronic pain and spasticity, multiple sclerosis, treatment-resistant epilepsy, nausea and vomiting due to chemotherapy, weight gain in HIV infection, sleep disorders, Tourette syndrome, anxiety, post-traumatic stress disorder, and schizophrenia are among the approved indications or applications of authorized CBD medicinal products. (Whiting et al., 2015; EMCDDA, 2018; Freeman et al., 2019; VanDolah et al., 2019; Sholler et al., 2020).

Although some cannabinoids have beneficial effects, CBD products that are not medications are prohibited from using therapeutic indications and health claims in their marketing. Similar to our findings, other studies have indicated that regulation violations occur often on the food supplement markets in the United States and Europe (Evans, 2020; Zenone et al., 2021; Amann et al., 2022; Bilbao and Spanagel, 2022; Spindle et al., 2022). A good and comparable example is anxiety, which, in our study, emerged as a leading “indication” for CBD oils. Similarly, a study by Soleymanpour et al. (2021) analyzed Twitter medical claims related to CBD-containing products and found pain, anxiety disorders, sleep disorders, and stress to be the four main “therapeutic” applications claimed on the social media platform (Soleymanpour et al., 2021).

CBD oils usually have a lower concentration of CBD than medicinal products in clinical trials; therefore, claims on efficacy should be handled with precautions. When evaluating the potency of CBD oils and their potential for health risk, a dose regimen and content analysis comparable to ours, as well as a comparison with approved medications containing the same substance, might be helpful (Table 4).

Admittedly, we could not identify toxic levels of the CBD content or recommended dose, and the preclinical and clinical data from medicine trials cannot be generally applied to other health products. However, the uncontrolled product quality in this market and unsupervised application or combination with other cannabis or addictive substances (e.g., alcohol and illegal substances) pose definite health risks. Uncontrolled consumption (especially when the product lacks application and dosage information) of a new ingredient regulated as a drug for which there is no EFSA novel food recommendation should not be promoted and regarded as a safe phenomenon. From a legal perspective, these products are regarded as counterfeit medicines: food supplements containing active pharmaceutical ingredients (Freeman et al., 2019; Bansal et al., 2020; Nasrin et al., 2021; Johnson et al., 2022b).

The EFSA NDA panel recently defined a 4.3 mg/kg bw/day lowest observed adverse effect level (LOAEL) value for cannabidiol in humans, and the authors suggest that CBD products available in the EU market as food supplements or other non-medical health products should be considered unsafe. Risk assessment should be implemented with the guidance of two values: a health-based guidance value (HBGV) of 10 mg/day, and products exceeding this should be titled “unfit for consumption,” while products exceeding human LOAEL should be considered “injurious to health.” When applying these limits (see Table 4) to our study samples, all of our ordered products can be considered as “unfit for consumption” as their recommended dosing exceeds the 10 mg/day limit (Lachenmeier et al., 2023).

The use of CBD products should be advised with caution for several patient or consumer groups, including young adults (18–25 years old) due to the unknown effects of CBD on developing brains; people with coexisting psychiatric conditions; elderly people taking multiple medications, as CBD may increase the risk of falls; people with decreased liver function or liver disease; women who are pregnant or nursing; and people who have allergies to cannabis or components in CBD products (Health Canada, 2022).

Given that CBD products are probably used under unsupervised circumstances, it is imperative to provide accurate product information about how to use and dose these items. Nevertheless, our study also demonstrated that consumers are more likely to take the wrong dosage, which is consistent with prior label accuracy studies from the US, Europe, or other nations. Therefore, these products can have serious negative effects on health when used by sensitive patient groups (such as those who have epilepsy) (Health Canada, 2022).

The Canadian government appointed a scientific committee to examine cannabis-containing goods, and the board made the following recommendations: for healthy adults, oral administration of CBD at doses ranging from 20 mg per day (mg/day) to a maximum of 200 mg/day is safe and tolerable for short-term use (a maximum of 30 days; e.g., enzyme-induction-related adverse effects require prolonged exposure, such as greater than 21 days), as long as they discuss the use of all other medications and substances with their pharmacist (Health Canada, 2022). It is important to take note of regional variations because the Australian standard from 2020 identified a low dose as 1 mg/kg/day or roughly 60 mg/day. It is also recommended that companies include warnings regarding special patient groups like pregnant and breastfeeding women, allergies, patients taking multiple prescribed medications, and the potential for drug interactions. Additionally, it is advised to include in the packaging of health products containing CBD clear dosing instructions and warnings of potential side effects, as they are dose-dependent. The committee recommends that since there are no conclusive studies that have validated its use for such indications, health products containing cannabis should bear a warning stating that they are not designed to help reduce consumption of opioids or alcohol. Along with the regulatory overview and change, the education of the public through awareness campaigns should also be carried out (Australian Government Department of Health and Aged Care Therapeutic Goods Administration, 2020). Labels on these products should encourage consumers to report adverse reactions on a defined platform or to the pharmacist. Finally, it is advisable to restrict the availability of these products; for example, health products containing CBD should only be available in pharmacies (National Mental Health and Substance Use Policy Laboratory, 2023). These recommendations might also be helpful to other countries that are having trouble regulating CBD products.

Since there was no significant history of consumption in the EU prior to May 1997, CBD and other cannabinoids were confirmed to fall under the novel food legislation in 2019. As a result, CBD oils must first undergo evaluation and authorization before they can be utilized as an ingredient in food or dietary supplements. The data currently available are insufficient to determine the definite no observed adverse effect level (NOAEL) and the LOAEL values, two crucial metrics for evaluating the toxicity of CBD. Furthermore, while evaluating CBD as a novel food, interactions should be taken into account due to the intricacy and importance of CBD receptors and pathways. Not surprisingly, the EFSA Panel could not conclude the safety of CBD as a novel food in 2022. In 2021, the Committee on Toxicity of Chemicals in Food, Consumer Products, and the Environment (COT) suggested setting a daily maximum for food exposure to CBD of 1 mg/kg body weight (The Committee on Toxicity of Chemicals in Food, 2021; EFSA Panel on Nutrition et al., 2022). Although there is a general increase in awareness and scrutiny from a regulatory perspective, distributors and companies have found other ways to sell the products, such as cosmetics. Our research group found a CBD product with a Cosmetic Products Notification Portal (CPNP) notification on their website (see United States MEDICAL CBD). Refer to the current legislation on hemp products and their CBD and THC content provided in Table 5.

In recent years, CBD products have gained global attention due to their inconsistent labeling and unregulated marketing, which pose a significant threat to consumer and patient safety since their diverse compositions might result in toxicities and drug interactions. The likelihood of serious adverse effects, such as liver failure, should highlight the dangers associated with the use of illegally marketed CBD oils. The improvement of pharmacovigilance, such as customers reporting adverse reactions to non-medicinal items (including dietary supplements), can be a valuable method for reducing risks. To build the regulatory framework for CBD containing products, additional research, including phytochemical investigations using validated methodologies, clinical and real world safety studies, is required.