THC-limits for food: A scientific study

Franjo Grotenhermen, Michael Karus and Daike Lohmeyer

nova-Institute, Hürth, Germany

     By now, a large number of food items from hemp, primarily based on hemp seed, is on the market in Europe, the USA, Canada, and Australia. They include hemp oil, hemp breads and pastries, hemp granola bars, hemp ice cream with hulled hemp seeds and cider blended with hemp flavored soda water, to mention just a few.
        This recent increase in the use of hemp seed products for nutrition, and the fact that small quantities of THC may enter into the food during processing, calls for the expeditious establishment of a legal framework, specifically a standard for the permissible residual content of THC (Δ9-tetrahydrocannabinol), the most relevant psychoactive ingredient of hemp.
        The "meat" of hemp seed itself does not contain any THC. It is primarily found in the flowers and perigonal bracts surrounding the seeds and, to a lesser extent, in the leaves. During seed processing, residuals of the flowers may enter the hemp food, particularly the oil.
        Generally, the THC content of hemp food is so low that psychotropic or even pharmacological effects can be excluded with certainty, even if larger quantities of food are consumed. Biologically relevant concentrations of THC were found in hemp oil only in a few isolated cases. The oil in question was of Swiss origin. It had been pressed from the improperly cleaned seeds of high-THC varieties, which are legal to grow in Switzerland. Due to these occurrences, the Swiss government introduced, in 1996, legal limits for THC in oil and other foods made from hemp seeds.
        So far, Switzerland is the only country worldwide where THC limits for food have been adopted. Two main reasons call for such action in other countries: The consumer must be able to expect no, or only insignificant, quantities of THC when consuming food based on hemp seeds. Not only should psychotropic effects be excluded with certainty, but also undesired short-term and long-term physiological effects.
        The spectre of contamination by THC, combined with the current un-clear legal situation regarding THC limits, tends to obstruct the increased use of hemp in food. Only when this matter is settled will hemp seed and hemp oil achieve the widespread distribution they deserve based on their nutritional value and flavor. As THC limits in hemp food are being considered by the authorities in Europe, North America and Australia, it becomes apparent that a rational and scientifically founded basis for such limits is still lacking. A detailed study, summarized here, was prepared by the nova-Institute in order to create such a basis.

First, our derivation of THC limits for hemp food required the establishment of maximum single doses, as well as maximum daily doses, of THC innocuous to health. A safety margin, commonly used in toxicology, was then applied in order to guarantee sufficient protection against undesired effects on sensitive people.
        To this end, first the placebo limit for acute psychotropic and physical effects from THC ingestion had to be determined from the literature. Due its long elimination half-life, THC may accumulate in the human body. Thus, some researchers have expressed concern that chronic health effects may arise at THC concentrations in food below the limit for psychotropic effects. Above all, possible effects on the hormonal system and fertility, on the immune system, as well as on pregnancy were mentioned. For that reason, a detailed account of this query is provided. Once safely permissible oral THC doses were determined, the corresponding maximum permissible THC concentrations in food had to be derived using consumption patterns for relevant hemp foods. While typical consumption habits served as a base-line, the consumption of quantities higher than average had to be considered as well. Since hemp-based foods may be consumed in strongly varying quantities, several categories and their respective limits were established.

Table 1. Ranges of THC doses and resulting concentrations in blood plasma.
  Oral dose of THC
per kg. body weight
(order of magnitude)
Maximum plasma
  0.014 mg/kg
0.1-0.2 mg/kg
1-2 mg/kg
10-20 mg/kg
100-200 mg/kg
Recommended maximum daily dose
Placebo limit, psychotropic threshold
Heavy Cannabis consumption
Medium dose in animal studies
High dose in animal studies
< 1 ng/ml
3-5 ng/ml
(15-30 ng/ml ?)
100-150 ng/ml
(500-1000 ng/ml ?)
  - High dose in cellular studies = 10,000 ng/ml (10 mg/ml).

Methodological Basis for Determination of THC Limits
     Concepts, such as the "lowest observed adverse effect level" (LOAEL) and the "no observed adverse effect level" (NOAEL) are commonly applied when developing standards intended to prevent negative health effects from chemicals which are known, or suspected, to cause such effects. In order to provide a wide margin of protection, safety factors of about 10 are usually applied.
        The majority of toxicological data for THC results from animal studies in which high doses were applied, as well as from cell experimental studies (Table 1). For many of the potential health effects of THC, the NOAEL was found to range at such high concentrations. Occurrence of other ad-verse effects at lower dose ranges have also been discussed. They may be representative of the situation of a chronic heavy Cannabis consumer (50 to 400 mg of THC per day). Even these doses exceed the threshold under question, i.e. that for psychotropic effects, by an order of magnitude. Thus, the available data provide a sound basis for the exclusion for most of these effects with a high margin of safety.
        THC differs from non-specifically acting harmful chemicals in food in that it acts on compound-specific binding sites (cannabinoid receptors) on the surface of body cells. However, at high concentrations, which are not encountered in the case of hemp foods, non-specific harmful effects, such as direct effects on the cell membrane, are also observed. The receptor mode of action provides an additional mar-gin of safety for two reasons:
        • As a rule, for most harmful chemicals, the toxicity increases and the NOAEL correspondingly decreases, with the duration of exposure. In the case of THC, the opposite applies since its effect decreases with in-creasing exposure. This is due to the development of a tolerance to THC by the receptors.
        • Children are generally considered particularly sensitive to various harmful chemicals. Consequently, higher safety factors are chosen to provide sufficient protection. However, children have a significantly lower density of cannabinoid receptor sites. Thus, compared to adults, psychotropic effects occur only at higher THC doses. There are some non-specific actions of cannabinoids, e.g. some anti-oxidative activity in doses possibly relevant for the situation of a Cannabis consumer, but so far there seem to be no relevant harmful non-specific actions in this dose range.

Biological Basis for Determination of THC Limits
     THC is a highly active pharmacological substance which shows, as a function of dose, effects on a multitude of organic systems and body functions. The physical toxicity is low, e.g., tests to establish a lethal THC dose for monkeys have been unsuccessful to date because a maximum administered dose of 9,000 mg/kg body weight did not result in their death.
        Acute psychotropic effects from the consumption of products from Cannabis of the drug type, such as marijuana and hashish, are changes in mood ("high") typical for marijuana, and changes in the sensory perception, the feeling for elapsed time, etc. Acute physical effects are, for example, the acceleration of heartbeat and a dry mouth.
        The limit for psychotropic effects is about 0.2 to 0.3 mg THC (single dose) per kilogram body weight when administered orally in a lipophilic (oily) matrix. Depending on body weight, this corresponds to a dose of 10 to 20 mg of THC for an adult person.
        An oral single dose of 5 mg THC in a lipophilic medium is rated as a placebo dose with respect to acute perceptible psychotropic and physical effects, i.e., it cannot be distinguished from a placebo. Due to the duration of THC’s effectiveness in therapeutic doses (4 to 12 hours) 5 mg may be taken twice per day, resulting in a NOAEL for acute perceptible effects of 10 mg THC per day.
        In connection with a chronic consumption of Cannabis for intoxication purposes, the perils of a possible impairment of the lungs, caused by inhalation of marijuana or hashish smoke, and the possible chronic effects on the psyche have been discussed. These effects are of no importance when THC is taken orally in sub-psychotropic doses, such as with food. While the maximum daily quantity of THC ingested with food must be safely below the psychotropic threshold, the question of whether non-perceptible and chronic physical health impairment may occur below the psychotropic threshold, must also be addressed.
        By virtue of the effects observed in studies relating to animal tests, four areas are of potential interest. They include effects on the genetic material, on pregnancy, on the hormonal system and fertility, as well as on the immune system. The available literature suggests the following evidence with respect to these effects.
        Genetic material: If taken in doses typical for consumers of marijuana, THC is not mutagenic, not carcinogenic, and has no effect on cell metabolism. The NOAEL is above the concentrations relevant to this type of human consumption.
        Pregnancy: There are only weak references to an influence on pregnancy caused by the consumption of marijuana. Animal studies showed only inconsistent effects, even after the administration of doses of 10-20 mg THC per kilogram of body weight and more, i.e., hundred times higher than those causing psychotropic effects. There have been reported indications of a slight impairment of the cerebral development of children of chronic Cannabis consumers, though this could not be confirmed by other authors. The NOAEL for most parameters related to pregnancy, e.g., parturition, duration of pregnancy, infantile abnormalities, weight at birth, and cerebral development is above the psychotropic threshold and, in most cases, above the range of chronic marijuana consumers.
        Hormonal system and fertility: There are no consistent findings of THC influence on male and female sex hormones, or on fertility, caused by the THC intake of regular Cannabis consumers. The strongest indications of such effects relate to hormonal malfunctions during puberty and to a temporary influence on the concentration of prolactin and LH in women during a particular phase of the menstrual cycle. However, these were only isolated observations. The NOAEL for THC concerning the influence on sex hormones, other hormonal effects, or other effects relevant to reproduction, are usually above the range of human marijuana consumption. Again, they are above the threshold for psychotropic effects.
        Immune system: Animal and cellular studies indicate that THC produces suppressive effects on cellular and humoral immunity. However, they can be attributed mainly to non-specific toxic effects and require extremely high doses. In fact, at low doses, various effects stimulating the immune system or no effects at all were found. The NOAEL for many relevant parameters relating to the immune system is clearly above levels relevant to human Cannabis consumption. For some effects, conflicting observations could be made in the examination of humans and of cells of marijuana consumers. When such effects were described, they were, even in the case of a heavy Cannabis consumption, very weak and with a doubtful relevance to health. The NOAEL for this parameter is, therefore, above the psychotropic threshold.

Factors reducing the biological efficacy of THC
     Since they may strongly affect the biological efficacy of THC, two aspects have to be considered in connection with oral THC intake with food:
        • THC must be consumed in its phenolic form in order to be bio-logically effective. However, in unprocessed hemp plants, THC occurs in the form of its largely ineffective carboxylic acid (THCA). It is decarboxylated, i.e., converted into its active form, primarily by heat during baking and other forms of food processing, and when smoked. Thus, largely unprocessed foods, such as cold-pressed oils, may be permitted to contain larger fractions of pharmacologically non-efficacious THC carboxylic acids.
        • The degree of THC absorption by the human intestines also depends on the physical and chemical properties of the carrier. Generally, lipophilic carriers, such as oils, promote absorption. In our study, this most un-favorable case, (i.e., maximum absorption of THC), was assumed for the determination of the maximum daily dose of 10 mg. If present in less fatty matrices, such as breads, pastries or drinks (hydrophilic environment), the bioavailability of THC is typically reduced by 50%.

Summary and conclusion on biological effects
     The experimental findings summarized above, suggest a safe maxi-mum THC single dose of 5 mg for a healthy person of 70 kg, corresponding to 70 micrograms per kg body weight (BW). The safe maximum THC daily dose is 10 mg, or 140 micrograms per kg BW. These doses do not result in any acute perceptible psychotropic or physical effects nor in any chronic adverse effects. These doses assume that THC is present in its active, phenolic form and is taken in a lipophilic basis. If these conditions are not met, the tolerable dose is higher due to the lower resorption by the intestine and due to a part of the THC being in the pharmacologically non-available carboxylic form.
        The application of a safety factor of 10 results in a tolerable daily dose of 14 micrograms THC/kg body weight or 1 milligram THC for a person of 70 kg. As already mentioned, the assumed safety factor of 10 represents a conservative choice for two reasons, development of tolerance despite accumulation of THC and a substantially lower density of cannabinoid receptors in children compared to adults. Both reduce the potential for specific psychic and physical effects conveyed by the receptors.

Table 2. Foods and their respective annual and daily rounded average consumption.
  Food Consumption per year Consumption per day
Finished products, breads and
pastries, non-perishable foods,
and pasta:
Alcoholic drinks:
Nonalcoholic drinks
(including herbal teas):
12 kg
105 kg
170 liters
251 liters
33 g
288 g
466 ml
688 ml

Derivation of THC limits
     The establishment of a concentration-based THC limit in food requires that two questions be answered:
        • How much THC can be ingested safely with the diet? This question has been discussed above.
        • At what rates may hemp-based foods be ingested?
        Table 2 averages consumption habits by Germans for market foods which may contain hemp products. The figures are taken from the data-base of the Federal Statistical Office (Krueger 1998, Weber 1998).
        So far, legal THC limits for food have only been adopted in Switzerland, and only after THC-containing food resulted in side-effects for some consumers (Table 3).
        The Swiss THC limits have proven their value in practice. On the one hand, producers have been able to comply with them through proper manufacturing practice. On the other hand, no incidents of side effects due to the consumption of hemp based food have be-come known following their adoption.
        In 1998, Canada was the second country to pass THC limits. However, the limit does not refer directly to food, but to hempen raw and semi-finished products, such as hemp fibers and seeds. Products containing hemp are exempt from further regulation if they contain less than 10 milligrams THC per kilogram.
        According to the discussion above, the maximum tolerable daily dose is 1 mg THC (for a person of 70 kg). This dose includes a safety factor of 10. Table 2 furnishes the data for the average consumption of food containing THC.
        Our recommended THC limits for each food category can be derived from this data. They are presented in the fourth column of Table 4. In order to account for potentially higher consumption rates, the derived limit is multiplied by an additional safety factor which varies for the different groups of food (see column 3). This additional safety factor also takes into consideration the potential situation of a person ingesting hemp food with the maximum permissible THC quantity of all listed groups of food. In the case of extreme nutritional habits, the safety factor of 10 may be considered reduced to as little as 5, which is still sufficient due to the low toxicity of THC. Furthermore, this margin is higher than that for some other natural drug substances in food, e.g., the quinine, used as an additive in sodas or alcohol used for embellishment of meat dishes.
        A comparison of the nova-recommended THC limits for food to the Swiss limits suggests the following:
        Edible oil: The nova-recommended limit of 20 mg THC/kg is below the Swiss value of 50 mg THC/kg. The main reason is the difference in assumed daily consumption of oil (nova-study: 33 g/day, Switzerland: 10 g/day). This is due to the fact that the nova value includes edible oils employed by the food industry in processed foods. The additional safety factor of 1.5 for higher-than-normal consumption is lower than for our other food categories. This reflects the fact that the food industry rarely employs oils for food processing which are high in polyunsaturated fatty acids, such as hemp oil. High-grade oils with intense taste, such as walnut, olive and hemp oil are mainly used alone, for example as salad oil.
        Finished products, breads and pas-tries, and pasta: The nova-recommended limit of 1.5 mg THC/kg is somewhat lower than the Swiss limits (2-5 mg THC/kg). Nova took into account more extreme consumption habits here as well.
        Alcoholic drinks: The nova-re-commended limit is 0.7 mg THC/kg versus the Swiss limit of 0.2 mg THC/kg. We consider the value of 0.7 mg THC/kg sufficiently low, all the more since the effect of alcohol, by far, pharmacologically outweighs the effect of THC. In order to exceed the placebo limit for THC (5 mg), more than 7.0 liters of alcoholic drinks would have to be consumed.
        Nonalcoholic drinks: The nova-recommended limit of 0.3 mg THC/ kg is similar to the Swiss limit of 0.2 mg THC/kg. The additional safety factor is higher because particularly large quantities of nonalcoholic drinks may be consumed without - as in the case of alcohol - other pharmacological effects dominating.
        Considering the conservative nature of the assumptions underlying nova’s recommended limits, the latter provide the consumer with a high degree of safety and protection from any acute perceptible psychotropic and physiological effects, as well as ad-verse chronic health effects produced by THC. At the same time, the experience of hemp foods manufacturers in the last few years has shown that these limits can be achieved through good production practice, without posing an undue burden.
        This article is a summary of an approximately 65 page study by the nova-Institute, providing a detailed pharmacological and toxicological basis for THC limits for food and extensive references. Copies are avail-able in German from nova-Institute, Hürth, Germany, and in English from Hemptech, Sebastopol, California, USA. The English version is on the Website of the North American Indus-trial Hemp Council (http://www.naihc.org).
        The nova-Institute study was made possible by the kind support from the following individuals and entities (in alphabetical order). The authors would like to extend their sincerest thanks to them:
        Australian Hemp Resource And Manufacture, Australia; Dave Pate, International Hemp Association, The Netherlands; Davert-Mühle Rainer Welke GmbH, Germany; dupetit Natural Products, Germany; Euro-American Marketing & Promotion Service, The Netherlands; Genossenschaft Gärtnerei Enetbrugg, Switzerland; HanfDampf Großhandel & Produktentwicklung, Germany; Hempola, Ontario, Canada; Jace Callaway, University of Kuopio, Finland; MBR Agrar Service Taunus-Westerwald GmbH, Germany; Mörk Naturprodukte, Germany; North American Industrial Hemp Council, Wisconsin, USA; Richard Rose, The Hemp Corporation, California, USA; The Ohio Hempery Inc., Ohio, USA; Treuhanf AG, Germany.

Table 3. THC limits for food in Switzerland.
  Food Limit mg/kg Notes
  Hemp oil
Hemp seeds
Breads and pastries
Vegetable food
Nonalcoholic drinks
Alcoholic drinks
Herb and fruit teas
On a dry weight basis
On a dry weight basis
On a dry weight basis
mg per liter of pure alcohol
mg per liter of finished products
Excluding spirits
mg per liter of finished product. Assumes
15 grams of plant parts per kg of boiling water,
steeped for 30 minutes above 85ºC.

Table 4. THC limits for food recommended by the nova-Institute.
  Food Average consumption
per day and per capita
Additional safety
THC limit
  Edible oil:
Finished products, breads
and pastries, pasta:
Alcoholic drinks:
Nonalcoholic drinks:
33 g
288 g
466 ml
688 ml