To investigate CBD degradation into ∆ ⁹-THC under acidic conditions, differently concentrated CBD in methanolic solutions was used in a range corresponding to typical amounts consumed with supplements based on commercial CBD (Supelco Cerilliant, certified reference material, #C-045, 1.0 mg/mL in methanol) supplied by Merck (Darmstadt, Germany). These solutions were exposed to an artificial gastric juice as well as different incubation times and stress factors such as storage under light and heat (see Table 1 for full experimental design). The solutions were stored either in standard freezer (-18°C) or refrigerator (8°C) or at room temperature (20°C). Increased temperatures were achieved using a thermostatically controlled laboratory drying oven type “UT6120” (Heraeus, Langenselbold, Germany) set to either 37°C or 60°C. The daylight condition was achieved by storage at a window (south side). For ultraviolet light exposure, six 25 W ultraviolet (UV) fluorescent tubes type “excellent E” (99.1% UVA) built into a facial tanner type “NT 446 U” (Dr. Kern GmbH, Mademühlen, Germany) were placed 15 cm from the surface of the solutions (open sample vials). In deviation of an experimental protocol of Merrick et al. ¹⁷ , a gastric juice without addition of surfactants was used, which was strictly produced according to the European pharmacopoeia ¹⁸ (0.020 g NaCl + 0.032 g pepsin + 0.8 mL HCl (1 mol/L), filled up to 10 mL with water). As pure CBD was available only in methanolic solution, the final experimental setups contained 0.08 mol/L HCl and 1% methanol due to dilution (methanol residues in this order of magnitude are not interfering with the analysis).

To ensure the utmost analytical validity, all samples were independently measured on two different instruments, using a triple quadrupole mass spectrometer (TSQ Vantage, Thermo Fisher Scientific, San Jose, CA, USA) coupled with an LC system (1100 series, Agilent, Waldbronn, Germany) and also using a quadrupole time-of-flight (QTOF) mass spectrometer (X500, Sciex, Darmstadt, Germany) coupled with an UPLC system (1290 series, Agilent, Waldbronn, Germany). Both systems used the same type of separation column (Luna Omega Polar C18, 150 × 2.1 mm, 1.6 μm, 100 Å, Phenomenex, Aschaffenburg, Germany). The separation was isocratic with 25 % water (0.1 % formic acid) and 75 % acetonitrile (0.1 % formic acid) and a flow of 0.3 mL/min. In case of QTOF with 35 % water (0.1 % formic acid) and 65 % acetonitrile (0.1 % formic acid) and a flow of 0.45 mL/min. The evaluation took place after fragmentation of the mother ion into three mass traces for each compound. As quantifier for ∆ ⁹-THC and CBD, the mass transition m/z 315 to 193 was used. In case of QTOF, quantification was conducted over accurate mass and control of fragmentation pattern. CBD eluted as one of the first cannabinoids, a few minutes before ∆ ⁹-THC. As internal standards ∆ ⁹-THC-d ₃ (Supelco Cerilliant #T-011, 1.0 mg/mL in methanol) was used for the quantification of ∆ ⁹-THC (Supelco Cerilliant #T-005, 1.0 mg/mL in methanol), and cannabidiol-d ₃ (Supelco Cerilliant #C-084, 100 μg/mL in methanol) for quantification of CBD (Supelco Cerilliant #C-045, 1.0 mg/mL in methanol). The certified reference materials were obtained as solutions in ampoules of 1 mL, all supplied by Merck (Darmstadt, Germany). A limit of detection (LOD) of 5 ng/mL was determined. For both procedures, relative standard deviations better than 5% were achieved. Both methods are able to chromatographically separate ∆ ⁹-THC and CBD from their acids. Specificity was ensured using a certified reference material as a reference standard of THCA (Supelco Cerilliant #T-093, 1.0 mg/mL in acetonitrile). Baseline separation was achieved between ∆ ⁹-THC, ∆ ⁸-THC and THCA. Therefore, the reported values in this study are specific for ∆ ⁹-THC and CBD. In contrast to some previous studies based on gas chromatography, we do not report “total THC” or “total CBD”, which would be a sum of the free form and its acid.

To study the possible influence of natively contained ∆ ⁹-THC in hemp products as a cause for side effects, a sampling of all available CBD products registered as food supplement in the German State Baden-Württemberg, other available hemp extract products in retail, as well as all products available at the warehouse of a large internet retailer were sampled between December 2018 and December 2019. A total of 67 samples (see Table 2 for product designations) were analysed using the above described liquid chromatographic method with tandem mass spectrometry (LC-MS/MS) for ∆ ⁹-THC content. For toxicological evaluation of the results, the lowest observed adverse effect level (LOAEL) of 2.5 mg ∆ ⁹-THC per day published by the European food safety authority (EFSA) based on human data (central nervous system effects and pulse increase) was used ²⁰ . Taking safety factors (factor 3 for extrapolation from LOAEL to no observed adverse effect level (NOAEL) and factor 10 for interindividual differences, total factor 30) into account, an acute reference dose (ARfD) of 1 μg ∆ ⁹-THC per kg body weight was derived ²⁰ . In their assessment, the Panel on Contaminants in the Food Chain of EFSA also considered interaction between ∆ ⁹-THC and CBD, but found the information controversial and not consistently antagonistic ²⁰ . This is consistent with more recent research of Solowij et al. ²¹ that the effects of ∆ ⁹-THC may even be enhanced by low-dose CBD (e.g., as found in food supplements) and may be particular prominent in infrequent cannabis users. However, the current scientific evidence does not allow for considering cumulative effects. The applicability of the acute reference dose (ARfD) of 1 μg ∆ ⁹-THC per kg body weight was re-confirmed by EFSA in 2020 ²² . For further details on interpretation of results and toxicity assessment, see Lachenmeier et al. ² .

There is not much evidence to assume that chemically pure CBD may exhibit ∆ ⁹-THC-like side-effects. The World Health Organization (WHO) judged the compound as being well tolerated with a good safety profile ³ . CBD doses in the food supplements on the market are typically much lower than the ones tested in clinical studies. Additionally, there is a 90-day experiment in rats with a hemp extract (consisting of 26% cannabinoids, out of which 96% were CBD and less than 1% ∆ ⁹-THC) from which a NOAEL of 100 mg/kg bw/day could be derived ²³ . Based on 100 mg/kg bw/day × 26% × 96%, this would be about 25 mg/kg bw/day for CBD (or 1750 mg/day for a person with a body weight of 70 kg). This NOAEL would not typically be reached by the CBD dosages in food supplements.

Some, partly older, in vitro studies put up hypotheses about the conversion of CBD to ∆ ⁹-THC under acidic conditions such as in artificial gastric juice ^{17, 24– 26} . If these proposals could be confirmed with in vivo data, consumers taking CBD orally could be exposed to such high ∆ ⁹-THC levels that the threshold for pharmacological action could be exceeded ²⁷ . However, taking a closer look at these in vitro studies raises some doubts. If CBD was to be converted to ∆ ⁹-THC in the stomach, typical ∆ ⁹-THC metabolites should be detectable in blood and urine, but this has not been observed in oral CBD studies ^{28, 29} . Due to the contradicting results, a replication of the in vitro study of Merrick et al. ¹⁷ was conducted using an extended experimental design. A more selective LC-MS/MS method and also an ultra-high pressure liquid chromatographic method with quadrupole time-of-flight mass spectrometry (UPLC-QTOF) were used to investigate the CBD degradation.

Under these conditions in contrast to Merrick et al. ¹⁷ , no conversion of CBD to ∆ ⁹-THC was observed in any of the samples. Only in case of the positive control (2 week storage in 0.5 mol/L HCl and 50% methanol), a complete degradation of CBD into 27% ∆ ⁹-THC and other not identified products (with fragments similar to the ones found in cannabinol and ∆ ⁹-THC fragmentations but with other retention times) was observed (Table 1, underlying data ¹⁶ ). From an analytical viewpoint, the use of less selective and specific analytical methods, especially from the point of chromatographic separation, could result in a situation in which certain CBD degradation products might easily be confused with ∆ ⁹-THC due to structural similarities. Thus, similar fragmentation patterns and potentially overlapping peaks under certain chromatographic conditions might have led to false positive results in the previous studies. In conclusion of our degradation experiments, we agree with more recent literature ^{30, 31} that CBD would not likely react to ∆ ⁹-THC under in vivo conditions. The only detectable influence leading to degradation is strong acidity, which should be avoided in CBD formulations to ensure stability of products ³² .

Out of 67 samples, 17 samples (25% of the collective) have the potential to exceed the ∆ ⁹-THC LOAEL and were assessed as harmful to health. 29 samples (43% of the collective) were classified as unsuitable for human consumption due to exceeding the ARfD (see Table 2, underlying data ¹⁶ ). Furthermore, all samples (100%) have been classified as non-compliant to Regulation (EU) 2015/2283 of the European Parliament and of the Council of 25 November 2015 on novel foods ³³ and therefore being unauthorized novel foods ³⁴ . The labelling of all samples (100%) was also non-compliant to Regulation (EU) No 1169/2011 of the European Parliament and of the Council of 25 October 2011 on the provision of food information to consumers ³⁵ , e.g. due to lack of mandatory food information such as ingredients list or use of unapproved health claims in accordance to Regulation (EC) No 1924/2006 of the European Parliament and of the Council of 20 December 2006 on nutrition and health claims made on foods ³⁶ . In summary, none of the products in our survey was found as being fully compliant with European food regulations.

The ∆ ⁹-THC dose leading to intoxication is considered to be in the range of 10 to 20 mg (very high dose in heavy episodic cannabis users up to 60 mg) for cannabis smoking ³⁷ . The resorption of orally ingested ∆ ⁹-THC varies greatly inter-individually with respect to both total amount and resorption rate ³⁸ . This might be one of the reasons for the individually very different psychotropic effects. A single oral dose of 20 mg THC resulted in symptoms such as tachycardia, conjunctival irritation, “high sensation” or dysphoria in adults within one to four hours. In one in five adults, a single dose of 5 mg already showed corresponding symptoms ³⁹ .

Some of the CBD oil supplements contained ∆ ⁹-THC in doses up to 30 mg (in this case in the whole bottle of 10 ml), which can easily explain the adverse effects observed by some consumers. Most of the CBD oils with dosage of around 1 mg ∆ ⁹-THC per serving offer the possibility to achieve intoxicating and psychotropic effects due to this compound if the products are used off-label (i.e. increase of the labelled maximum recommended daily dose by factors of 3–5, which is probably not an unlikely scenario. Some manufacturers even suggest an increase of daily dosage over time). Generally, these products pose a risk to human health, especially in light of the German guideline value for total THC in these kind of products ^{19, 40} . The German guideline value of 150 μg total THC/kg for foods in general including food supplements is several orders of magnitude below the actual contents of ∆ ⁹-THC in the CBD products, even without consideration of THCA.

Hence our results provide compelling evidence that THC natively contained in CBD products by contamination may be a direct cause for side effects of these products. Obviously, there is an involuntary or deliberate lack of quality control of CBD products. Claims of “THC-free”, used by most manufacturers, even of the highly contaminated products – sometimes based on unsuitable analytical methodologies with limits of detection in the percentage range –, have to be treated as fraudulent or deceptive food information.

Open Science Framework: Dataset for “Are side effects of cannabidiol (CBD) products caused by delta9-tetrahydrocannabinol (THC) contamination?” (Version 2) https://doi.org/10.17605/OSF.IO/F7ZXY ¹⁶

This project contains the following underlying data:

Data are available under the terms of the Creative Commons Zero “No rights reserved” data waiver (CC0 1.0 Public domain dedication).