1 Introduction

Medicinal and recreational products containing cannabidiol (CBD) are increasingly popular. Projections for the CBD market in Europe suggest potential revenue of over EUR 3.4 bn by 2026 (Statista 2022). One of the most prominent products is CBD oil, typically consisting of an industrial hemp extract mixed with oils such as hemp seed, olive, or MCT (Miller et al. 2022; Tonoyan et al. 2022). CBD oils appear as foods, food supplements, aroma oils, or cosmetics (for oral, dental, or skin care). CBD shows a wide range of biological effects and is approved as an anticonvulsant (Epidyolex®) in the EU and the USA (EFSA NDA Panel 2022; EMA 2021; FDA 2018). However, reliable clinical evidence is lacking for many other effects (Peng et al. 2022). The products in this study, commonly referred to as ‘CBD oils,’ are not medicines and may not have positive pharmacological effects, despite often being advertised with health claims (Fernández et al. 2023). In Germany, consumers use CBD products mainly for pain relief, curiosity, stress relief, relaxation, and improved sleep (Geppert et al. 2023).

CBD and its acidic form CBDA are the most abundant cannabinoids in industrial hemp varieties, while THC and THCA are most abundant in varieties used for medicinal and illicit purposes (Andre et al. 2016). Unlike the psychoactive Δ9-THC, CBD has no psychoactive effects (Bhattacharyya et al. 2010; Golombek et al. 2020; Zuardi et al. 2012). The European Food Safety Authority (EFSA) has paused the evaluation of pending applications for CBD under the novel food regulation due to data gaps and uncertainties regarding possible adverse health effects. Applicants must provide the necessary data (EFSA NDA Panel 2022; European Parliament 2015). CBD oils marketed as foods, including supplements, violate EU food law as they are considered non-authorized novel foods (Lachenmeier et al. 2023). Likewise, the German Federal Office of Consumer Protection and Food Safety (BVL) is stating no legal marketability of CBD in foods and supplements in the EU.Footnote 1 Several German administrative court decisions have upheld that CBD oils cannot be marketed without approval as novel foods or are classified as medicinal products (VGH BW 2019; 2022a; 2022b; 2023; VGH München 2023).

Δ9-THC is the main psychoactive substance responsible for the ‘high’ from marijuana and related drugs, causing euphoria and relaxation but also side effects like abdominal pain, dizziness, nausea, or paranoid reactions (CONTAM 2015). Δ9-THC is the only cannabinoid regulated as a food contaminant in the EU, with maximum levels of 3 mg/kg for hemp seeds and derived products and 7.5 mg/kg for hemp seed oils. CBD oils are not part of this regulation. Maximum levels apply to Δ9-THC equivalents, the sum of Δ9-THC and its acidic counterpart THCA, excluding split off CO2 (European Commission 2023). These levels were implemented because the acute reference dose (ARfD) was exceeded in some exposure estimates (European Commission 2022).

To obtain reliable data for risk assessment, an ultra-high performance liquid chromatography method coupled to tandem mass spectrometry (UHPLC-MS/MS) is used to investigate the cannabinoid profile of CBD, CBG (cannabigerol), and CBN (cannabinol) oils, with a focus on psychoactive Δ9-THC. The cannabinoid profiles of the CBD oils are compared to CBN and CBG oil profiles to investigate differences in cannabinoid concentrations, especially Δ9-THC. The manufacturers’ declared cannabinoid content is compared with the concentrations measured in this study. It was hypothesized that the Δ9-THC concentrations in the oils (CBD, CBG, and CBN) would be high enough to exceed the ARfD of 1 µg of Δ9-THC per kg of body weight for recommended use. To verify this, we calculated the number of drops of CBD, CBG, and CBN oils that would result in an exceedance of the ARfD for Δ9-THC.

2 Materials and methods

2.1 Samples

A total of 26 different products were purchased from hemp shops, drugstores, and online shops in Germany. These included ‘CBD oils’ (n = 21), ‘CBN oils’ (n = 2), ‘CBG oils’ (n = 2), and ‘CBD tinctures’ (n = 1). More detailed product information (e.g., origin, cannabinoid declaration, carrier oil) is provided in Table S1 (Supplementary Material).

2.2 Sample preparation and analysis

Sample preparation followed the analytical method for quantifying cannabinoids in edible oils proposed for standardization under § 64 of the German Food and Feed Code (LFGB) by the BVL (Amtliche Sammlung von Untersuchungsverfahren nach § 64 LFGB 2024), as detailed by Lindekamp et al. (2024). Modifications included reducing the sample weight to 100 mg oil in a 10 mL volumetric flask and adding higher dilution levels. For each sample, 2 dilutions in 2 orders of magnitude (4 in total) were prepared to cover the broad concentration spectrum of the different cannabinoids. To quantify low-concentration analytes, the dissolved oil was diluted by a factor of 250 or 1000, while for highly concentrated analytes, factors of 25,000 and 50,000 were used. Recovery and precision adhered to the guidance document on identifying mycotoxins and plant toxins in food and feed (EURL/NRL Mycotoxin/Plant Toxin Network 2023), even at high dilution factors (Table S2b, Supplementary Material).

Cannabinoids were quantified using the UHPLC-MS/MS method described in Lindekamp et al. (2024). In addition to the 16 cannabinoids listed therein, Δ8-THC and hexahydrocannabinol ((R)-HHC and (S)-HHC) were also determined, for a total of 19 cannabinoids. Quality control parameters (Table S2), MS parameters (Table S3), analyte structures (Fig. S1), and an exemplary chromatogram (Fig. S2) are provided in the supplementary material.

2.3 Conversion of cannabinoid concentration per mass to per volume

Cannabinoid concentrations in µg/g were converted to mg/mL assuming a density of 0.927 g/mL for hemp seed oil (Anwar et al. 2006). For simplified estimation, this density was applied to all oils. A drop of oil was estimated to be approximately 0.04 mL, based on a product declaration stating: “10 mL of product contains approximately 250 drops.” The conversion used Eq. 1.

Equation 1: Conversion of in µg per g to in µg per drop.

$${c}_{\text{Cannabinoid}} \left[\frac{{\upmu }\text{g}}{\text{drop}}\right]={c}_{\text{Cannabinoid}}\left[\frac{{\upmu }\text{g}}{\text{g}}\right]\cdot 0.927 \frac{\text{g}}{\text{m}\text{L}}\cdot 0.04 \frac{\text{m}\text{L}}{\text{d}\text{r}\text{o}\text{p}}$$

3 Results and discussion

3.1 Quantification of CBD and deviations from its declared concentrations

Nineteen cannabinoids were analyzed in 26 CBD, CBG, and CBN oil samples, with total cannabinoid concentrations ranging from 39,121 to 256,212 mg/kg (Table 1). CBD was the dominant cannabinoid in all CBD oils, comprising an average of 92.3% of the cannabinoid profile, increasing to 96.0% when CBDA was included.

Table 1 Concentrations of the 19 cannabinoids quantified in the 26 CBD oil samples. Each sample was worked up once and measured in 2 high and 2 low dilution levels (4 in total) to quantify major and minor cannabinoids
Full size table

All CBD products prominently declared the CBD concentration on the packaging, often with higher concentrations offered at higher prices. Declared CBD concentrations ranged from 2.5 to 20%, while measured CBD concentrations ranged from 3.1 to 24.2%. Actual concentrations ranged from 81 to 226% (average of 121%) of the declared values. Most samples (n = 18) had manufacturers providing certificates of analysis regarding cannabinoid concentrations on their websites. The minimum and maximum deviations from the declarations were 73% and 137%, respectively, in these certificates. Based on the present analysis, 23 samples had CBD concentrations within the ranges provided by manufacturers’ certificates (73% < c < 137%), while one sample drastically exceeded the declaration (sample 21 with 226% of the declared value). Deviations from labeled claims have also been reported in previous studies (Gurley et al. 2020; Liebling et al. 2022; Pavlovic et al. 2018). An overview of the deviations is shown in Fig. 1; the exact values are listed in Table S4 (Supplementary Material).

Fig. 1
figure 1

Deviation of the measured primary cannabinoid (CBD, CBG, or CBN) concentrations compared to the product declaration and, if available, the concentration declared in the certificate of analysis provided by the manufacturer

Full size image

3.2 Minor cannabinoids and acid-to-neutral-ratios

In this study, no clear distinction between ‘full’ and ‘broad’ spectrum oils was possible based on their cannabinoid profiles. Both terms were used in product descriptions, but neither is accurately defined. For different manufacturers, ‘broad spectrum’ could imply the removal of specific cannabinoids or entire groups of phytochemicals (e.g., terpenes) from the extract (e.g., by distillation).

CBG was quantified in 24 samples, with concentrations ranging from 6 to 3503 mg/kg (excluding CBG oil samples 23 and 24). CBGA is the precursor of CBG and the main product of cannabinoid biosynthesis in C. sativa. CBN was quantified in 18 samples, with concentrations ranging from 8 to 1528 mg/kg (excluding CBN oil samples 25 and 26). CBNA and CBN are formed by the oxidation of THCA and THC due to storage, heat, and light (Russo et al. 2008). Thus, CBN concentrations in CBD oils may be related to the storage and treatment conditions of the production materials.

Additionally, cannabichromene (CBC) was found in 22 samples, with concentrations ranging from 4 to 2008 mg/kg. CBCA, the precursor to CBC, is a product of CBGA biosynthesis, along with CBDA and THCA, and is therefore common in hemp and hemp products (Andre et al. 2016). Other cannabinoids are derived from these primary and secondary cannabinoids and are usually found in lower concentrations (Lange and Zager 2021; Taura et al. 2007). Cannabidivarin (CBDV) concentrations ranged from 4 to 9718 mg/kg across all 26 samples. Similar to THCVA (tetrahydrocannabidivarin), CBDVA (the precursor of CBDV) is a minor product of CBDA biosynthesis. Consequently, high concentrations of CBDA in a cannabis plant correlate with relatively high concentrations of CBDVA (Hanuš et al. 2016; Taura et al. 2007; Yang et al. 2016). In this study, CBDV is the only other cannabinoid besides CBDA and CBD quantified in all 26 samples.

As cannabinoid acids are the direct products of biosynthesis, acidic forms dominate neutral forms in cannabis plant material. Processing and storage conditions lead to the decarboxylation of acids to their corresponding neutral forms, resulting in lower acid-to-neutral ratios (A/N) in hemp-derived products (Citti et al. 2018). In this study, the A/N ratio for all cannabinoids was less than 0.03 for 23 samples and ranged from 0.2 to 0.5 in 3 samples. This is consistent with manufacturers’ claims on their product websites, which often highlight the need to thermally process ingredients to maximize neutral cannabinoids as a unique selling point of their products.

3.3 CBG and CBN oils

Cannabinoid oils containing CBG or CBN as dominant cannabinoids have received limited attention in cannabinoid research. CBN has higher binding affinities to CB receptors compared to other minor cannabinoids, but its affinity is 10 times lower than Δ9-THC’s (Munro et al. 1993; Walsh et al. 2021). Both products are relatively new and less common than CBD oils. Two samples of each were obtained and analyzed in this study. In the CBG and CBN oils, CBD was a major component in 3 oils, comprising approximately 53%, 52%, and 34% of the cannabinoid profiles, while the fourth oil contained only 4%.

It is well-established that CBG-rich hemp varieties contain similar concentrations of CBD and CBG (de Meijer and Hammond 2005). However, the existence of CBN-rich hemp varieties is unlikely, as CBNA is not a direct product of cannabinoid biosynthesis but rather an oxidation product of THCA. Producing CBN oils by extraction and enrichment from plant material is economically infeasible due to relatively low concentrations. Thus, semi-synthesis, possibly from CBD, seems a more reasonable approach (Maioli et al. 2022).

In the 2 CBN oil samples, (R)-HHC was quantified, but not its isomer (S)-HHC, with values of 5 and 110 mg/kg. HHC is a semi-synthetic cannabinoid that can be added to cannabis products (Russo et al. 2023). Data on its toxicity and effects on humans is still insufficient, but available data suggests that (R)-HHC has psychoactive potential, though it is likely less potent than Δ9-THC (BfR 2023). It is noteworthy that HHC was found only in the CBN oils, possibly originating as an intentional or unintentional by-product of the suspected CBN semi-synthesis (Russo et al. 2023).

3.4 Quantification of Δ9-THC and related compounds

Δ9-THC and structurally related compounds (Δ8-THC, THCA, THCV, THCVA) were quantified in 21 of the 26 samples, with Δ9-THC being the most abundant in most samples. It was found in all 21 samples, with concentrations ranging from 5 to 1576 mg/kg. Similar THC concentration ranges in CBD oils have been previously reported (Dubrow et al. 2021; Johnson et al. 2022; Pavlovic et al. 2018). In contrast, Δ8-THC and THCVA were found at low levels (2 to 4 mg/kg) in only 2 samples each. THCV, the C3 side-chain equivalent of the C5 side-chain Δ9-THC, was the most abundant THC-related cannabinoid in 3 samples, with concentrations ranging from 751 to 2837 mg/kg, compared to Δ9-THC concentrations of 57 to 182 mg/kg. Notably, THCVA, a minor product of THCA biosynthesis in cannabis plants, typically occurs at much lower levels relative to THCA (Walsh et al. 2021). Consequently, high THC(A) concentrations are associated with elevated THCV(A) levels, typically found in medicinal hemp varieties. The abundant THCV(A) levels suggest the removal of excessive THC(A) during production.

In this study, Δ8-THC was quantified at concentrations of 4 and 3 mg/kg in samples 2 and 3, respectively. For comparison, the Δ9-THC concentrations in these samples were < LOQ (0.003 mg/kg) and 5 mg/kg, respectively. This is notable because the Δ8-THC concentrations in these two samples are higher than the average Δ9-THC concentrations in hemp seed edible oils (approx. 2 mg/kg) (Lindekamp et al. 2024). The occurrence of Δ8-THC in CBD oils has been previously described in a report by the UK’s Food Standards Agency (FSA), with concentrations ranging from 3 to 102 mg/kg (MacDonald et al. 2022). It is known that Δ9-THC can transform into its thermodynamically more stable isomer Δ8-THC under oxidative or mildly acidic conditions (Staginnus et al. 2014). Δ8-THC shows psychoactive effects similar to Δ9-THC (Akpunonu et al. 2021; Walsh et al. 2021), but it is currently not included in the regulated maximum Δ9-THC equivalent levels for hemp seeds and hemp seed food products (European Commission 2023).

3.5 Safety implications of the findings with regard to CBD and Δ9-THC

3.5.1 Measured CBD concentrations in CBD oils may pose a health concern

CBD oil manufacturers usually include daily consumption recommendations on their product declarations. In this study, these recommendations ranged from 1 to 20 drops per day. Consumers generally base their consumption on these recommendations, but printed values might often be lower than those advertised elsewhere by manufacturers. Some manufacturers or hemp lifestyle websites offer ‘CBD oil calculators’ that recommend dosages based on age, gender, body weight, and perceived severity of pain or sensitivity to CBD. Although these recommendations often fall within the range of printed guidelines, some calculators suggest daily doses of 100 drops or more. Depending on the parameters submitted, the same 100 drops may be suggested for both low (5%) and high concentration (40%) CBD oils.

Furthermore, the quantification of CBD in the products showed that the concentration was on average 21% higher than declared, which complicates the estimation of CBD intake and may result in a higher intake than expected. CBD oils with concentrations up to 40% are available, meaning a 10 mL bottle could contain 4 g of CBD. EFSA has not yet derived health-based guidance values due to gaps in knowledge about CBD’s safety and has requested more toxicological data from manufacturers (EFSA NDA Panel 2022). In the UK, the Advisory Committee on Novel Foods and Processes (ACNFP) and the Committee on Toxicity (COT) have established a provisional acceptable daily intake (ADI) for pure form (≥ 98%) CBD in foods in April 2024, based on new data from novel food applications. The provisional ADI was set at 0.15 mg/kg bw or 10 mg for a 70 kg person (ACNFP 2024).

In this study, CBD concentrations ranged from 0.1 to 9.0 mg per drop (Table S5, Supplementary Material), meaning that 24 of the 26 samples would exceed this provisional ADI after consuming 2–7 drops (Fig. 2a). Epidyolex®, a prescription medicine approved in the EU for treating epileptic seizures, uses CBD as the active ingredient and is administered at doses of 2.5 to 12.5 mg/kg bw twice daily, equalling 350 to 1750 mg/day for a 70 kg person (EMA 2021). Assuming a lower-end recommendation of 4 drops per day with a 5% CBD oil at 2 mg per drop, daily CBD intake totals 8 mg. In a higher-end recommendation scenario, with 20 drops daily of a 20% CBD oil at 9 mg per drop, the daily CBD consumption reaches 180 mg, which is half the starting therapeutic dose – and a consumption of 100 drops would clearly result in a dose level used therapeutically.

3.5.2 Measured Δ9-THC concentrations in CBD oils may result in exceedance of the ARfD

To assess the health risk posed by CBD oils regarding the psychoactive cannabinoid Δ9-THC, the quantified Δ9-THC concentrations were used to determine how many drops may be consumed by a 70 kg person before exceeding the ARfD of 1 µg/kg bw derived by the EFSA (EFSA 2020).

Considering the highest daily dose recommended by one manufacturer from our sample list (20 drops), the ARfD is exceeded if the oil has a Δ9-THC concentration of at least 100 mg/kg. In this study, 13 samples (50%) contained Δ9-THC levels that would lead to an exceedance of the ARfD if 20 drops of the product were consumed. Assuming consumption of 100 drops, an additional 4 samples would lead to an exceedance. In 7 of these samples, consuming just 2 drops would be sufficient to exceed the ARfD. The ARfD specifies the estimated maximum quantity of a substance that can be consumed in one day, either during one meal or across several meals, without a detectable risk to health. Figure 2b shows the number of drops needed to exceed the ARfD, with the calculated values listed in Table S5 (Supplementary Material).

A study conducted by Lachenmeier et al. (2023) quantified Δ9-THC levels in various CBD products, including ‘oils’ and ‘oil supplements’. Among 30 samples across both categories, Δ9-THC concentrations were found to exceed the ARfD (Lachenmeier et al. 2023). Several other studies have also reported Δ9-THC concentrations in CBD oil products in the same order of magnitude as presented here (Dubrow et al. 2021; Johnson et al. 2022; Miller et al. 2022). In the present study, however, the determined Δ9-THC concentrations per drop were related to the ARfD for the first time.

Fig. 2
figure 2

Number of drops required to exceed the provisional ADI for CBD derived by the FSA and the ARfD of Δ9-THC for a 70 kg person

Full size image

On the y-axis, samples are scattered for clarity. Only samples with quantifiable amounts of CBD and Δ9-THC are shown. The daily limit of 20 drops represents the highest recommended daily dosage specified by one manufacturer in our selected sample list.

4 Conclusion

In this study, 19 cannabinoids were quantified in 22 CBD, 2 CBG, and 2 CBN oil samples marketed as food supplements, aroma oils, or cosmetics. The quantification of CBD concentrations showed that the concentration was on average 21% higher than the manufacturer’s declaration, resulting in a higher intake than expected by consumers. This is concerning given the uncertainties identified by EFSA regarding the safety of CBD. Furthermore, 24 of 26 samples (92%) would exceed the provisional ADI set by the UK FSA for pure form CBD (≥ 98%, used as a food ingredient) of 0.15 mg/kg bw with a consumption of only 2–7 drops per day.

Δ9-THC was quantified in 20 of 26 samples (77%), with high concentrations above 100 mg/kg found in 13 samples (50 %). The Δ9-THC concentration averaged 536 mg/kg (median: 232 mg/kg). Δ8-THC and (R)-HHC were both found in two samples each. Due to their psychoactive potential, it is crucial to monitor these compounds when analyzing CBD oils. Consumption of such CBD oils can easily result in exceeding the ARfD for Δ9-THC of 1 µg/kg bw. Remarkably, in 10 of the 26 samples, just 2 to 10 drops—common consumption recommendations by manufacturers for adults—would suffice to exceed the ARfD. Thus, consumption of these products may pose a risk to human health due to both CBD and THC content. Additionally, as these products are available in flavored forms like lemon or raspberry, there is a risk that children may consume them. Therefore, close investigation and monitoring of CBD oils, for both CBD concentration and psychoactive Δ9-THC, is crucial.