Recent Articles

Recent Comments

« | Main | »

THC in Hemp Foods and Cosmetics: The Appropriate Risk Assessment

By Hempology | January 27, 2002

[Ed: Dr. Geiwitz was part of the Ad Hoc Committee on Hemp Risks, an impartial scientific organization commisioned by Health Canada to assess the risks of hemp-based food and cosmetic products. This paper is a cornucopia of knowledge just waiting to be soaked up by your brains about the effects of THC, other cannabanoids, and cannabis in general on the physical and mental health of both humans and animals. Dr. Geiwitz is appearing as an expert witness in Ted's Court Case. Many thanks to Dr.
Geiwitz for providing Hempology with this incredible literature and giving us permission to publish it on our website! - Crackerjack]

THC in Hemp Foods and Cosmetics: The Appropriate Risk Assessment
by James Geiwitz, Ph.D., and the Ad Hoc Committee on Hemp Risks

January 15, 2001


In 1998, industrial hemp became a legal crop in Canada, promising
environmentally-sound farming and processing of fibre for paper, textiles, and
building products. In addition, hemp seed is among the world’s most nutritious
foods, and its oil is an exceptional bodycare emollient.

In 1999, Health Canada issued a draft report entitled Industrial Hemp Risk
Assessment. The report dealt only with hemp foods and cosmetics (bodycare
products) and focused on tetrahydrocannabinol (THC), the psychoactive ingredient
in cannabis hemp. By law, hemp foods and cosmetics must contain less than 10
parts per million THC. Health Canada concluded that, even with THC content
limited to 10 ppm, “inadequate margins of safety exist between potential exposure
and adverse effect levels for cannabinoids in cosmetics, food, and nutraceutical
products made from industrial hemp.” Health Canada, therefore, is considering a
ban on hemp foods and cosmetics.

The purpose of the Ad Hoc Committee on Hemp Risks is to respond scientifically to

the Health Canada risk assessment. We focused on four allegations by Health
Canada: acute neurological effects and toxic effects on brain development,
reproductive system, and immune system. In contrast to the Health Canada
conclusions, we found absolutely no health risks from the extremely low doses of
THC present in hemp foods and bodycare products. In fact, the best research
indicated some health benefits of THC, most notably in the strengthening of the
human immune system.

How can it be that scientists at Health Canada review the research literature on the
effects of THC and conclude that hemp foods and cosmetics are unsafe, while
another group, our Ad Hoc Committee, reviews the same research and concludes
the exact opposite? In our analysis of the science of THC risk assessment, we
identified the major problems with the research referenced by Health Canada,
including extreme dosing, inappropriate extrapolation to humans from animal
studies, and political pressures on scientific disinterest. Also, contrary to
assumptions made by Health Canada, children are at less risk from THC than adults;
hemp THC must be heated to be biologically active (which means the THC in cold-
pressed hemp oil is inactive); and only two or three cannabinoids are candidates for
investigation in hemp foods, not 66.

We next calculated, from our data, the appropriate standards for THC in hemp foods
and cosmetics. Although there are no health risks from THC, we set the standard at
the threshold for psychoactive effects, with a safety factor of 10. Scientifically
determined, the maximum THC in hemp oil (the most efficient carrier) should be
set at 20 parts per million  a conservative estimate, with other studies
recommending limits as high as 50 ppm. Current Canadian regulations,which set
the standard at 10 ppm, represent a difficult but achievable practical limit for bulk
hemp manufacturers.

Finally, to complete a cost/benefit analysis, the health benefits of hemp foods and
cosmetics were explored. In foods, the essential fatty acids (EFAs) and the high level
of protein make hemp nuts an exceptionally nutritious food; a healthy heart is
perhaps the most well-documented benefit. The EFAs and the protein are basic
building blocks of the body, involved in health at the cellular level. The same EFAs
are the primary ingredient in hemp bodycare products, which heal and nurture the
skin and prevent infections.

We conclude that a ban on hemp foods and cosmetics would be ill-advised policy
based on a flawed review of the research literature. Rather than protecting the
health of Canadians, such a ban would be damaging. We propose instead that
current THC regulations be retained, and the health benefits of hemp foods and
cosmetics be the topic of Health Canada reports on industrial hemp.

Contact: James Geiwitz, Ph.D., <>, 250-598-4075

1.0 INTRODUCTIONIndustrial hemp (Cannabis sativa L.) became legal in Canada on March 13, 1998, and
the first crops were planted in the summer of 1998.

On March 3, 1999, Health Canada (HC) issued a draft report entitled Industrial Hemp
Risk Assessment. The report was prepared by Joan Orr, M.Sc., and Mary Starodub,
M.Sc., for Hugh Davis, Head of the Microbiology and Cosmetics Division of the
Product Safety Bureau of Health Canada. On July 27, 1999, the Toronto Globe & Mail
published a story on the draft risk assessment. The report was revised several times,
the latest dated September 9, 1999.

The HC report dealt only with hemp foods and cosmetics, not the fibre products
such as paper and textiles. The focus of the report was tetrahydrocannabinol (THC),
the psychoactive ingredient in cannabis, responsible for the “high” in marijuana,
the high-THC cousin of industrial hemp. Other cannabinoid ingredients were
occasionally considered, although research on the other cannabinoids is sparse.

By law, industrial hemp must contain less than 0.3% THC, 10-100 times less than the
amount of THC in marijuana. The industrial hemp seed (botanically, a nut), from
which hemp foods and cosmetics are made, contains only trace amounts of THC, but
it cannot be cleaned and processed without some microscopic contamination from
the bracts and leaves of the hemp plant, the primary sites of THC production by the
plant. By law, industrial hemp foods and cosmetics must contain less than 10 parts
per million (ppm) THC.

The question posed by the HC report was this: Do hemp foods and cosmetics
containing 10 ppm THC or less present health risks for Canadian consumers?

Here is their answer: “Overall, the data considered for this assessment support the
conclusion that inadequate margins of safety exist between potential exposure and
adverse effect levels for cannabinoids in cosmetics, food, and nutraceutical products
made from industrial hemp.”

Dr. Hugh Davis, for whom the risk assessment was performed, believes that hemp
foods and cosmetics should be prohibited unless the industry can develop
techniques for the total elimination of THC (and 65 other cannabinoids) from the
hemp seeds. As the hemp nutmeat itself may contain faint traces of THC (Ross et
al., 2000), such absolute elimination may not be practical.

A further problem is that methods for detecting THC levels in hemp products are
constantly evolving, allowing for the detection of more and more minute
quantities. Thus, the definition of “zero tolerance” is constantly shifting, and an
industry “in compliance” today might be violating the standards tomorrow. The
concept of “zero tolerance” developed from the U.S. War on Drugs, not from health
research practices. Not even deadly food contaminants (e.g., pesticides) are held to
such impossible restrictions, and popular foods containing trace amounts of other
natural drugs (e.g., morphine, cocaine) are not a matter of official concern.

Dr. Davis suggested that the industry explore genetic engineering of the hemp plant
to eliminate the THC. We believe that genetic engineering might pose significantly
more and greater risks than THC, both to the environment and to the health of
Canadians, although longterm studies have yet to be done. Elimination of
cannabinoids may also affect crop susceptibility to pests (Pate, 1999b). In any case,
genetic engineering would violate official definitions of “organic” food and thus
significantly damage the marketability of hemp products.

1.1 The Ad Hoc Committee

To evaluate the claims in the risk assessment and to provide an industry response
to the report, Dr. James Geiwitz and Dr. Chuck Schom formed the Ad Hoc
Committee on Hemp Risks and invited hemp scientists from around the world to
join. Here is a list of current members, briefly identified:

In addition to the members of the committee, we have 14 associates who follow our
research and exchange information. These include scientists from The Body Shop
(England), which markets hemp cosmetics. The Body Shop has produced its own
response to the HC risk assessment; we incorporate their findings into our response.

The Committee’s work was accomplished entirely by e-mail, from July 27, 1999, to
the present. The Committee relied entirely on volunteer labour, which accounts for
the brevity of this response. A rebuttal of Health Canada’s risk assessment, over 400
pages long, deserves a point-by-point, fully referenced response of comparable
length, which, unfunded, we were unable to provide. For instance, we were unable
to reference the original data sources for many of the statements in this document,
although, for critical issues, the primary research is cited. In other cases and
especially for research summaries, our secondary sources were book-length reviews
of THC research by two committee members, Dr. Franjo Grotenhermen
(Grotenhermen et al., 1998) and Dr. John P. Morgan (Zimmer & Morgan, 1997). In
these documents, the reader will also find references to primary sources. (It is
worthy of note that the Grotenhermen and the Zimmer/Morgan reviews agree in
every significant conclusion.)

1.2 Outline of the Committee’s Response

An outline of Health Canada’s Risk Assessment report is provided in the Appendix.
Our response comprises three major sections. The first (Section 2) focuses on
research on the toxicology of THC, specifically four toxic effects alleged by HC:

Section 3 will examine some of the inappropriate methods and conclusions of the
Health Canada risk assessment: what went wrong with this badly flawed report, and

Section 4 of this response will focus on the determination of safe exposure levels for
THC in hemp foods and cosmetics, that is, the appropriate risk assessment. Topics
covered include:

Section 5 provides a brief description of the health benefits of hemp foods and
cosmetics, benefits that would be lost to Canadians if hemp foods and bodycare
products were banned.
2.0 THE TOXICOLOGY OF THCIt is our position that THC is one of the least toxic chemicals that humans ingest. At
normal doses, there is no evidence of genetic damage due to THC exposure or effects
on fertility, pregnancy, or offspring. Similarly, there is no evidence of damage to the
hormonal or immune systems. These statements apply to humans who ingest large
quantities of marijuana daily, and much more so to humans who ingest trace
amounts of THC through hemp foods. Ingestion through hemp bodycare products
is completely undocumented and highly unlikely.

Research that finds damaging effects of THC generally falls into one of two

  1. studies that are not replicated by later research using more appropriate
    experimental designs; and
  2. studies that use massive quantities of THC, far beyond
    the doses employed by heavy marijuana users.
  3. 2.1 Genetic Effects

In doses typical for consumers of marijuana, THC is not genotoxic, mutagenic, or
carcinogenic, and it has no effect on cell metabolism. THC does not result in
chromosomal breaks.

At extremely high doses applied directly to cells, THC reduces the synthesis of DNA,
RNA, and proteins. These effects are nonspecific, that is, unrelated to the typical
receptor activation in the human body.

In regard to genotoxic effects, the trace amounts of THC in hemp foods and
cosmetics are obviously safe for consumers.

2.2 Pregnancy and Offspring

Animal studies of the effects of THC on pregnancy are inconsistent, even with doses
of 10-20 mg/kg, a hundred times higher than the LOEL for psychotropic effects. A
few studies purported to show impairment of cerebral development in children of
chronic cannabis consumers, but these studies were never replicated and are now
discredited. The NOEL for pregnancy variables (parturition, duration of pregnancy,
infantile abnormalities, birth weight) is above the range of human consumption by
chronic marijuana consumers, which is much higher than THC levels from hemp
foods and cosmetics.

There is no realistically demonstrated danger to pregnant women or their offspring
from consumption of hemp foods, and clearly none at all from use of hemp
bodycare products.

2.2.1 Pregnancy

Greenland et al., 1982, found more meconium staining and longer duration of
labour in marijuana users, but this study has never been replicated, even by
Greenland’s lab. For centuries, cannabis has been used for pain relief during birth.
The general conclusions permitted by the research are that no birth complications
can be observed in mothers who ingest marijuana levels of THC over a long period
of time and that the trace levels of THC in hemp foods and cosmetics are obviously

Gibson et al., 1983, found more premature births in marijuana users, but this study
has never been replicated. Most studies find no marijuana-induced change in the
duration of gestation.

2.2.2 Birth defects and brain development

Birth defects associated with THC have been found only in animal studies in which
the THC was injected, in very high doses, directly into the abdomen. In humans,
there is no evidence whatsoever for a link between marijuana use and fetal
malformations or Minor Physical Anomalies (MPAs).

Studies that show a decreased birth weight in rat pups after THC ingestion have
been clearly discredited. The decrease, when it occurs (at high doses), is due to
reduced food and water intake of exposed dams; there is no difference between these
animals and pair-fed controls.

Evidence is accumulating that the cannabinoid-anandamide receptor system might
play a role in cerebral development in fetuses and neonates. Daily administration of
5 mg/kg THC to pregnant rats doubles the activity of the enzyme tyrosine
hydroxilase (TH) in specific brain cells of their fetuses (Hernandez et al., 1997). TH is
believed to be a key factor in the development of neurons. In contrast, one animal
study has established a disturbance of mesolimbic dopaminergic neurons among
perinatally THC-exposed males which persists in adult animals (Garcia-Gil et al.,
1997). However, the significance of these data for humans using hemp foods and
cosmetics is very probably nil.

Animal studies have generally found behavioural problems only at high doses. For
example, no behavioural effects in offspring were observed after dosing the
pregnant rats with 50 mg/kg/day. Hutchings et al., 1987, found nipple attachment
problems in rat offspring exposed to 50 mg/kg/day, but the problems were clearly
related to decreased food and water intake in the dams; the offspring of pair-fed
controls were indistinguishable from the offspring of experimental animals.

In humans, the offspring of chronic users show no differences from normal in
sleeping, eating, mental tests, and psychomotor tests. One researcher (Dreher, 1994,
1997) found the offspring of chronic users to be more lively and less irritable, with
fewer tremors; these babies were more easily quieted, yet more responsive to novel
stimuli. These results have not been replicated, but they show the extreme
inconsistency of marijuana studies. The more common finding is, simply, no

Studies that have attempted to find brain damage from THC have been
unsuccessful. Marijuana levels of THC do not kill brain cells. In one study,
monkeys were forced to inhale five marijuana cigarettes a day for a year; there was
no evidence of brain damage (Zimmer & Morgan, 1997). In humans, with brain
damage assessed by CAT scans, no damage was observed in spite of the high dose:
nine marijuana cigarettes a day.

2.3 Hormonal Systems and Reproductive Capabilities

Some high-dosage animal studies suggest that THC may act on the hypothalamus-
pituitary-adrenal axis and adversely affect the sex steroid hormones. However,
there is no reliable finding of adverse effects in animals (male or female) within the
range of human consumption of marijuana. The slight effects that sometimes
appear, disappear with repeated doses (tolerance). In humans, no effects were
discovered regarding the function or concentration of sexual hormones or other
parameters relevant for reproduction such as sperm quantity and quality.

In one representative study, men were dosed with up to 20 marijuana cigarettes a
day for a month (Hembree et al., 1979). The researchers found some decrease in
sperm concentrations and motility. The decreased factors were not outside of
“normal” range, and by the end of the month, the sperm factors had returned to
normal, despite continued dosing.

In men, a few studies found effects of chronic marijuana use on luteinizing
hormone (LH), which is related to testosterone production, although the effect
disappears with time, even if THC doses remain constant. Other studies found no
such LH effect. There is no effect of THC on testosterone, follicle stimulating
hormone (FSH), or prolactin. There are no effects on puberty. A representative
study (Mendelson et al., 1978) found no effect of marijuana smoking on testosterone
level, in spite of the high doses: 120 marijuana cigarettes in 21 days.

In women, the conclusions are the same: There are no reliable effects of THC on the
menstrual cycle, estrogen levels, progesterone, prolactin, LH, or FSH. The few
studies of positive effects involved high-dosage inhalation, effects that quickly
disappeared as tolerance developed.

In some animal studies, THC reduced the level of adrenocorticotropin (ACTH),
which is secreted by the adenohypophysis and stimulates the production of
glucocorticoids (cortisol) in the suprarenal cortex. This result could not be replicated
in human chronic consumers of marijuana. THC has no effect in humans on the
thyroid hormones or on glucose metabolism.

2.4 Immune System

“Cell experiments and animal studies demonstrate that THC has
suppressive effects on the humoral and cell-mediated immunity.
However, the majority of those can be attributed to toxic unspecific
effects. Many analysed parameters required extremely high doses to
exhibit any significant effect and the effects were dose-dependent with
the threshold concentration being precisely determinable. When
applying lower doses, one often observed differentially
immunostimulating effects or no effects at all. For many immune
parameters the NOEL is … irrelevant to the human consumption
situation. In studies of man or of cells of marijuana users the effects
observed were often contradictory. If such effects were found at all,
they were weak even in case of heavy cannabis use and of questionable
relevance to health. The World Health Organisation summarised in
its most recent cannabis report: ‘Many of their effects appear to be
relatively small, totally reversible after removal of the cannabinoids,
and produced only at concentrations or doses higher than those
required for psychoactivity (WHO, 1997, p. 27)’” (Grotenhermen et al,
1998, p. 53)

2.4.1 Suppression versus enhancement

THC and the immune system is the most thoroughly researched topic in the area of
subliminal biological effects. Much of the early research, which demonstrated
immune-system suppression, has been discredited. For example, Nahas et al. (1974)
found that THC decreases the number of T-lymphocytes � which control cell-
mediated, acquired immunity. Later studies found no such decrease. Dax et al.
(1989), for example, found no change in T- or B-lymphocytes (humoural immunity)
or in T-cell subtypes before, during, or subsequent to administration of THC to
chronic users. Wallace et al. (1988) reported similar findings, with a twist: an
increase in helper T-cells (CD4). These findings should be interpreted as
immunoenhancement, because helper T-cells stimulate the proliferation and
activation of other immune cells.

In a study cited in the Health Canada risk assessment, Nahas et al. (1977) found in
vitro suppression of T-cell proliferation in response to mitogens, which stimulate
cell division. Other researchers criticized Nahas’s method � applying THC in
massive doses to human cells in a petri dish � and called the results “meaningless.”
Better studies failed to replicate Nahas’s work and, instead, found immune system
stimulation at lower doses (Pross et al., 1993; Luo et al., 1992).

Let us be clear about these findings: What the research shows is immune system
suppression at very high doses, but immune system stimulation (enhancement) at
low doses. These effects have been demonstrated for both the T- and B-lymphocytes.
This means that the trace amounts of THC in hemp foods probably strengthen the
immune system of humans. High doses have nonspecific toxic effects, likely the
cause of any damage, whereas low doses act through specific receptor-based effects.
It’s a basic principle of pharmacology: low doses may be curative whereas high doses
are poisonous.

One last point: With an oral dose of THC of 0.1-0.2 mg/kg (the psychotropic
threshold), the blood plasma reaches a maximum concentration of 3-5 ng/ml. In
the cell studies, the concentration is 10 ug/ml, or 10,000 ng/ml � 2000 to 3000 times
the dose that produces the marijuana “high.”

2.4.2 Humans and disease

Marijuana smokers show an enhanced response to antigens (which trigger
antibodies) compared to cigarette smokers and cancer patients (Hollister, 1992),
which supports the conclusion of THC strengthening the immune system and casts
additional doubt on the high dosage cell studies. On a more general level.
absolutely no epidemiological evidence exists relating marijuana use and infectious
diseases (Hall et al., 1994). In cancer and AIDS patients, THC is used to reduce pain
and depression, stimulate appetite, and prevent nausea and vomiting. AIDS
patients, who suffer from a damaged immune system, are not harmed by THC (Di
Franco et al., 1996).

2.5 THC and Cancer

Immune-system stimulation by THC at low doses should be apparent in macro-
level health benefits. The stunning (but rarely reported) success of THC treatments
of cancer may be representative. One of the first studies had rats ingest a large dose
(50 mg/kg) of THC daily for two years. At the completion of the experiment, 70
percent of the dosed animals were still alive, but only 45 percent of the control
(undosed) animals survived. This sizeable difference was due almost entirely to a
reduced incidence of cancer in the animals given THC (Chan et al., 1996).

A more direct test of THC’s cancer-fighting properties was performed on rats with
brain tumours (Galve-Roperh et al., 2000). The tumours, called gliomas, are fatal in
humans. The researchers infused THC directly into the rats’ brains. The control
rats (no THC) died in two to three weeks. In a third of the THC-dosed rats, the
tumour was eliminated. Another third lived eight to nine weeks, instead of the two
to three weeks of the control (no THC) rats. A third of the THC-dosed rats gained no
benefit. The researchers claim that the THC works by stimulating the cancer cells to
“commit suicide” in a natural process called “apoptosis.” Normal cells were
unharmed. The THC in this experiment was very low dosage, and the cancers were
at a late stage, when untreated rats were already starting to die. The researchers
suggest that THC would work even better if given earlier.
3.0 THE SCIENCE OF THC RISK ASSESSMENTSection 4 will focus on the determination of safe exposure levels for THC. Before
we turn to the appropriate risk assessment for THC in hemp foods and cosmetics,
we will examine some of the inappropriate methods and conclusions of the Health
Canada risk assessment. We have come to precisely the opposite conclusions to
those of Health Canada regarding the risks of hemp products. Both of us claim
scientific data in support of our position. In this section, we will list some of the
ways in which Health Canada was mistaken, by inadvertently citing inadequate

The science of THC is not unlike other areas of science: Science does not prove
anything. It deals in probabilities, and its methods are designed to estimate the
degree of error in an estimate or in a probabilistic relationship. Most scientists view
their procedures as a search for error, whereas the general public perceives it as a
search for truth. In reality, it is a search for truth by way of estimating error.

The nature of science is such that one can always argue the opposite to a suggested
proposition, with some evidence in support. Global warming, for example, is
supported by the bulk of the evidence, but there are enough data leaning toward the
opposite conclusion that the National Post can claim that global warming is a hoax.
Similarly, scientists paid by the tobacco industry can mount a claim, with data
support, that smoking does not cause lung cancer.

When a scientific question has political ramifications (such as global warming or
smoking), the goals of science are often perverted, as different camps seek to
generate evidence for their position. The US War on Drugs is such a camp.
Beginning in the 1960s, the US government offered scientists millions of dollars to
“prove that marijuana is harmful.” The research cited by Health Canada includes
much of this “advocacy science,” which produced misleading conclusions about the
effects of THC.

The following section is, in effect, a manual on how to do advocacy science.

3.1 Extreme Dosing

The major deficiency with most reports of harm from THC is the massive doses
required to demonstrate such effects. In one study, monkeys were given the human
equivalent of 15 kg of marijuana in a single dose. Similarly, the petri-dish studies of
the effects of THC on body cells used concentrations 2000 to 3000 times the threshold
level for psychotropic effects.

The Body Shop, which markets hemp cosmetics, noted that the Health Canada
estimate of skin penetration of THC (33%) is wildly inaccurate because the oil used
to calculate the estimate had extremely high levels of THC (26mg/g). The high
concentration of THC outside the skin encourages penetration, which is a function
of the difference between outside and inside (where the concentration is essentially
zero). If hemp oil with 4 ug/g THC constituted 10 percent of a cosmetic, as is the case
with Body Shop lotions, then about 0.4 ug/g THC would be available for skin
absorption, that is, about 65,000 times less than the dose used by Health Canada
(Adams, 1999). In addition, attempts to deliver therapeutic THC via skin patch have
been unsuccessful (ElSohly, 1998), a further indication of the safety of hemp
bodycare products.

In a review of the effects of THC on the human immune system (which found
none), the reviewers note that some animals given large doses do show effects;
doses are forty to one thousand times the psychoactive doses for humans (Zimmer
& Morgan, 1997). Similarly, an attempt to find brain damage in monkeys failed to
do so, in spite of the dose: five marijuana cigarettes a day for a year.

These are extreme examples, but far from rare. Almost all of the studies that show
damage from THC use high to very high doses, even compared to marijuana levels.
When compared to the low doses from hemp foods and cosmetics, the high-dose
studies are irrelevant.

THC at reasonable levels such as those in marijuana and hemp foods acts on
compound-specific binding sites (cannabinoid receptors). Only at high
concentrations (which are not encountered in hemp foods and cosmetics) do
nonspecific, toxic effects occur. Most if not all chemicals will damage body cells and
systems at high concentrations  for example, numerous deaths have been recorded
in people who for psychiatric reasons drink excessive amounts of water. And
pharmaceuticals that are toxic at high concentrations are beneficial at low doses, as
seems to be the case with THC and the immune system.

3.2 Cannabinoid Receptors and Tolerance

The fact that THC at reasonable doses acts not nonspecifically but, rather, specifically
at receptor sites on neurons provides a further margin of safety for hemp foods and
cosmetics. For one reason, neurochemical receptors generally show tolerance  that
is, decreasing effect with repeated or sustained exposure. For most harmful
chemicals, the toxicity increases (and the NOEL decreases) with duration of
exposure. But, with THC, the opposite occurs, because of tolerance. For example,
high doses of THC in female monkeys resulted in hormonal changes and a
disruption of their menstrual cycle. After six months of high doses, the hormone
levels and the menstrual cycles returned to normal (Smith et al., 1983). Tolerance
can be observed in the cases of most THC effects.

Chronic exposure to THC does not irreversibly alter the cannabinoid receptors
(Westlake et al, 1991).

At the low doses of hemp foods and cosmetics, THC’s effects are almost entirely
receptor based, with little or no nonspecific toxicity. This means that even if a
troubling effect of low-dose THC were to be established, the risk would be

3.3 Cannabinoid Receptors in Children

According to the Health Canada risk assessment, infants experience greater exposure
to THC from hemp foods and cosmetics for four reasons:

We have some concern with the first three points. First, compared to adults, infants
have a higher percentage of body fat relative to lean mass, although the absolute
volume of fat is of course less. Second, although infants do have less lipoprotein,
the level reaches adult proportions by two to three years. And third, hepatic
metabolism may not be a desirable function, if THC metabolites are more
psychoactive than THC itself. In any case, we consider these facts irrelevant, since
low levels of THC present no risk.

Our research indicates that the fourth point is mistaken. Children have a
significantly lower density of cannabinoid receptor sites. They are therefore less, not
more, susceptible to the effects of THC (Grotenhermen et al., 1998). We recognize
that this issue is a controversial one, with research supporting both positions.
Because this is such an important point, we examine the research in some detail.

The preponderance of research data supports our position. One of the studies that
does not, one that is the basis for Health Canada’s claim, attempted to determine the
density of cannabinoid receptors in the fetus and neonate (Glass et al., 1997). These
researchers concluded that concentrations of receptors in these subjects were
“extremely high.” One fetal brain (33 weeks gestation) and two neonatal brains (3
and 6 months of age) were examined, a sample too small for valid conclusions. The
fetus died in utero of bowel obstruction, one neonate died of congenital heart
disease and the other of asphyxia; the first two subjects must be considered
“abnormal.” What this means we do not know. It’s possible that the fetus was
reacting to its bowel obstruction by producing high levels of endocannabinoids,
which could stimulate the production of cannabinoid receptors (Callaway, 2000).

There was a post mortem delay of up to 21 hours, a delay that may affect receptor
profiles. The autoradiograms of fetal and neonatal tissue were of poor to fair
quality, in contrast to those of adult tissue, which were extremely high quality. The
fetal and neonatal tissue was processed separately from that from adults, which
even the authors agree, requires considerable care in comparing results from
children and adults. And their conclusion of more receptors in young brains is
qualified: “Due to the small number of cases available for the study, it is not possible
to draw any definitive conclusions of the precise levels of cannabinoid receptors …
within the developing brain” (Glass et al., 1997).

This is hardly a “definitive study,” certainly not one on which to base public policy.

Research that supports our position includes a study of rats that discovered an
increase of cannabinoid binding from birth to day 30, which corresponds roughly to
human adolescence (Rodriquez de Fonseca et al., 1993). These data indicate a lower
density of receptors in younger subjects. Another rat study found a 470% increase
from birth to day 60, in all brain areas investigated (Belue et al., 1995). A third found
receptor binding increasing almost 50 percent with increasing age (McLaughlin et al.,

It is true that children generally respond more severely to chemical toxins and
require a greater margin of safety than adults. But in the case of THC, which
operates on specific receptors, children require a smaller margin of safety because
they have many fewer receptors. Children with cancer, for example, tolerate
considerably higher doses of THC than adults, with no symptoms of psychoactivity
(Abrahamov et al., 1995). This research group later studied, in mice, the response to
anandamide and THC; there was no response to anandamide for the first 23 days,
whereas a small response to THC began between days 15 and 20. The researchers felt
that their results were compatible with their human data showing that children
respond to the antiemetic effects of THC without psychotropic side effects (Fride &
Mechoulam, 1996). A similar study of children with cancer taking nabilone, a THC
analog, found that high doses were well tolerated: “Particularly for some adolescent
patients, it can turn a five day course of chemotherapy from a dreaded ordeal into
something accepted with a shrug of the shoulders” (Dalzell et al., 1986).

To summarize, most researchers find cannabinoid receptors in newborns, but
receptor populations in children are significantly smaller than in adults; also,
receptor binding in children is significantly less. Clinical studies of children with
cancer find that children tolerate much higher doses of THC than adults. While
more research needs to be done, the pattern of data is quite clear: Children can
tolerate much higher levels of THC in hemp foods and bodycare products than

3.4 Effective Forms of THC

In unprocessed hemp, THC occurs in the form of a monocarbon acid (THCA) that is
not absorbed well by the intestines. One cannot, for example, eat uncooked
marijuana and expect much of an effect. THC must be converted (decarboxylated) to
its phenolic form to be bio-effective, which is accomplished by the application of
heat. Smoking and baking are typical conversion methods. This means unheated
hemp foods, such as cold-pressed oils, contain mostly inactive forms of THC.

Absorption of THC by human intestines depends on properties of the carrier.
Lipophilic carriers, such as hemp oil, promote absorption of decarbosylated THC. If
the carrier is less fatty, as in hemp breads or beverages, the bioavailability of THC is
reduced by 50 percent or more. Even Health Canada accepts that hemp beer and
wine presents no risks.

3.4.1 Cannabinoids other than THC

The Health Canada risk assessment repeatedly makes the point that there are 66
cannabinoids in industrial hemp, that THC is only the best known and most
frequently studied. This number is misleading. It represents the sum total of
cannabinoids found in detectable quantity in at least one cannabis variety in at least
one study in the history of cannabis research. Health Canada contends that, even if
the risks of hemp foods and cosmetics from ingestion of THC are shown to be
minimal, the hemp industry must also show that the risks from the other 65
cannabinoids are also so. Such research would require years, if not decades. This
open-ended contention is unprecedented for a natural food product or drug source
(e.g., coffee, alcohol, tobacco), which may contain scores of untested chemical

The only cannabinoids proven to be manufactured by the hemp plant are THC,
CBD, CBC, and (presumably) their common biogenetic precursor, CBG (Pate, 2000).
CBD predominates, with an accompanying fraction of THC. CBC is found in
significant quantities only in tropical marijuana. CBG is found only in very small
amounts. To this short list can be added minor quantities of the THC degradation
products, CBN and delta-8 THC. The remaining 60 cannabinoids exist in almost
undetectable amounts  in fact, usually none at all  in any given hemp sample.
Health Canada admits that CBD poses few risks. CBN, according to Health Canada,
is as dangerous as THC, but the research that “proves” this is the same research that
“proves” that THC is risky. We believe this research to be problematic, if not

Since we have been unable to discover any significant health risks from the far
more potent marijuana, it is unlikely that any ingredient of hemp foods and
bodycare products pose health risks to human consumers.

3.5 Extrapolation from Animal Studies

A major disagreement exists between the Health Canada report and our research
group regarding the value of animal studies for the determination of risks to
human consumers of hemp foods and cosmetics. Many of the risks reported by
Health Canada come from studies in which high doses were given to rats or mice.
That’s OK, says Health Canada, because of “similarities” between humans and
rodents in the pharmacokinetics and metabolism of THC and in the brain
distribution of cannabinoid receptors.

However, the application of rat data to human risk assessments is an uncertain and
often misleading extrapolation, with numerous pitfalls. For example, the
extrapolation of doses is problematic. Typically, a dose given to rats is reported in
milligrams of THC per kilogram of body weight. The dose for humans to produce
the same effects is then calculated using the body weight of humans. The average
human weighs about 70 kg. So an effect caused by a 2 mg dose to a rat weighing 0.2
kg translates to a 700 mg dose to humans (about 50 times the dose for a human

This kind of extrapolation may be meaningless, because many biological processes
(e.g., metabolic rate) are unrelated to body weight. For this reason, some researchers
use comparisons of body surface (mg/m2) instead of body weight. It has been found
that body-surface comparisons predict more accurately human tolerance for anti-
cancer drugs from animal data than do body-weight comparisons. But body surface
is also a poor basis for extrapolation for many drugs. Other bases include
pharmacokinetics (absorption, metabolism, excretion, etc.) and toxicological
estimates such as the “lethal dose” studies.

The lethal-dose studies are a lesson: In rats, the lethal dose is around 1300 mg/kg.
Extrapolated on the basis of body surface, the lethal dose in dogs should be about
350 mg/kg and in monkeys, about 650 mg/kg. But dogs lived after a dose of 3000
mg/kg, and monkeys survived 9000 mg/kg. The lethal dose in these animals could
not be established. The primates should have been 50 percent more sensitive to
THC than rats, but were at least five to ten times less sensitive. The extrapolation
from rats to higher mammals was wildly inaccurate.

There are significant differences between the reproductive and hormonal systems of
rats and mice and those of humans (Mendelson and Mello, 1984). Mice, for
example, are especially disposed to fetal malformations. In general, data on smaller
animals leads to highly inaccurate estimates of THC toxicity in larger animals.

Reliable data on the toxicity of THC in humans must be based on studies with
human subjects.

3.6 The Fallibility and Abuse of Science

Studies of the effects of THC on humans are inconsistent, for a number of reasons:
Many studies use small samples (that is, few subjects), and small-sample studies are
notoriously unreliable (that is, inconsistent). For scientific purposes, small-sample
studies are practically worthless. A young man who smokes pot fails to go through
puberty; the child of a pot smoker develops cancer: These are meaningless
anecdotes, although such studies are widely touted as proof of THC’s dangers.

Most of the marijuana studies on humans compare chronic users with “matched”
control subjects. This experimental design produces data that are often misleading,
because the researchers are comparing two groups that differ in many ways. True
matching is impossible, since one can never know all the factors that influence the
life of a test subject. For example, many chronic marijuana smokers use other drugs
as well, including cigarettes and alcohol. In addition, human subjects often lie about
their drug use, making assignment to groups difficult. Results from such studies are
often unreliable or difficult to interpret.

As we’ve mentioned, the US War on Drugs has distorted the scientific
infrastructure and produced a plethora of biased findings. A study that purports to
have found deleterious THC effects is quickly published, whereas a study that finds
THC safe is not. In the latter case, researchers may suppress the data or peer review
might disparage the experiments (Levy and Koren, 1990). Finally, if well-designed
experiments demonstrating the safety of THC are published, government
publications often ignore them, focusing instead on the studies that support the
official view. This pseudo-science we have termed “advocacy science.”

True science consists of a search for conclusions to explain previously established
facts, theories to explain observed data. Advocacy science consists of a search for
facts to support a previously established opinion.
4.0 DETERMINATION OF TRUE HEMP RISKSIn this section, we present our determination of a THC level in hemp foods and
cosmetics that is clearly safe for human consumption. This determination was
conducted for the German hemp industry by Dr. Franjo Grotenhermen, a member
of our committee (Grotenhermen et al., 1998).

4.1 LOELs and NOELs

Our first step is to determine the No Observed Effect Levels (NOELs) and Lowest
Observed Effect Levels (LOELs) for the psychoactive effects of THC. Our review of
the research clearly shows that if THC levels are below the NOEL for psychoactive
effects, there will be no other risks to health.

The LOEL for THC’s psychoactive effects is 0.2 to 0.3 mg/kg, about 10 – 20 mg THC in
a single dose to an average adult. The NOEL, the level of THC that cannot be
distinguished from placebo (no THC) effects, is .07 mg/kg, about 5 mg for an average
adult. At the effect duration of four hours, the NOEL is 5 mg twice a day, or 10

The application of a safety factor of 10 results in a tolerable daily dose of 14 ug/kg,
about 1 mg THC for a 70 kg adult. This dose will have no psychoactive effects and
no adverse health effects.

4.1.1 Appropriate safety margins

Health Canada applied a safety margin of 1000 to its flawed determination of THC
LOELs. We believe that this safety margin is ridiculously large, a hundred times the
industry standard of 10 (Kendal, 2000). Health Canada justified its safety margin as
follows: 10-fold for interspecies differences, 10-fold for intraspecies differences, and
10-fold for lack of data from chronic studies. As we have seen in the review of
research, the interspecies differences (especially comparing human risks to toxic
effects from high-dose rodent studies) are in the opposite direction from that
proposed by Health Canada, that is, humans are less at risk than would be assumed
from the rodent findings (even if extrapolation calculations were accurate, which
they are not). Similarly, the intraspecies differences are such that estimation of
NOELs with adults will protect children, with fewer receptors, even more.

Chronic consumption of THC will not increase risk, it will decrease it. THC
receptors typically develop tolerance, so that a continuous supply of THC does not
lead to an increase in possible health impairments (and a decrease in NOEL), as is
common with most toxic chemicals. And, although we can always use more
chronic studies of humans, there is no evidence that we will be misled by using the
many chronic studies we now have.

We believe that a safety factor of 10 is conservative.

4.2 Exposure and NOEL

The research on absorption of THC clearly shows that the greatest risk is with adults
ingesting hemp oil that contains THC in its active phenolic form. (We disregard the
fact that cold-pressed hemp oil is likely to contain more inactive THC acids than
active THC forms.) Lipophilic carriers such as oils promote absorption; THC
absorption in hemp breads, in contrast, is reduced by 50 percent; in hemp beverages
and cosmetics, absorption levels are even less, or nil.

The average daily consumption of hemp oil for Germans who consume this
product is 33g/day. The German figure is probably high for the Canadian situation,
since Germany has had a vital hemp industry for many years. But we will use it
anyway, and add a further safety factor, 1.5, to account for increased consumption as
the world recognizes the health benefits of hemp foods. So our exposure figure is

4.2.1 Maximum levels of THC, properly determined

The NOEL for THC, as determined in Section 4.1, is 1 mg for an average adult. With
consumption of hemp oil (in which absorption is greatest) of 50 g/day/capita, the
maximum THC content of the oil, if we don’t want to exceed the NOEL, should be
.020 mg per g of oil, or 20 mg/kg. Scientifically determined, the maximum THC in
hemp oil should be set at 20 parts per million (20 ppm).

4.2.2 Comparisons of THC limits

Our recommendation for THC limits in hemp foods and cosmetics is 20 ppm.
Current Canadian legislation sets Canadian limits at 10 ppm, and the Health Canada
report erroneously determined that even this amount was dangerous. The hemp
oil available in Canada before industrial hemp was legalized (from Don
Wirtshafter’s Ohio Hempery, for example) had 15 ppm and was rejected by Health

Switzerland is the only country other than Canada to set limits on the THC in hemp
foods. After careful scientific evaluation, the Swiss set limits of 50 ppm for hemp
oil, with less restrictive limits for other foods and bodycare products. In Europe, the
Swiss standard is thought to be liberal, the German standard (20 ppm) is considered
conservative. The Canadian standard of 10 ppm is considered severe, and the “zero
tolerance” recommended by the Health Canada risk assessment is considered


Health authorities typically do cost/benefit analyses of new drugs, trying to
determine whether the benefits outweigh the toxic costs associated with the drugs.
Health Canada has chosen not to pursue the “benefits” portion of such analyses,
asserting that their mandate is only to identify possible risks to the health of
Canadians. Of course, all drugs and all foods have health risks. However, the
notable risks of aspirin or red meat are deemed insufficient to warrant banning
these substances.

The risks of THC in hemp foods and cosmetics are practically nil, as we have shown.
However, even if minor risks could be demonstrated, we would argue that
depriving Canadians of the health benefits of hemp foods and bodycare products
would constitute more of a threat to their health than unregulated consumption.

The purpose of this paper is to refute the flawed conclusions of the Health Canada
risk assessment. For that reason, we will not dwell excessively on the virtues of
hemp foods and bodycare products, which are well-described elsewhere (e.g.,
Conrad, 1997). Instead, we will briefly list some of their abundant health benefits.

5.1 Benefits of Hemp Foods

Hemp foods are made from the hemp seed, which is botanically a nut. There are
two major components of the hemp nut, the oil and the nutmeat. Hemp oil is
made by cold-pressing the seed; what’s left is the hempseed “press cake,” which is
commonly converted to flour for hemp breads and similar foods. The hemp nut
can be eaten whole, as it’s very nutritious and quite tasty. Hemp foods have been a
dietary staple for millions of people in Europe for centuries and for tens of millions
in China and other parts of Asia for millennia.

5.1.1 Hemp oil

Hemp oil has many desirable ingredients (Pate, 1999a), the most important of which
are the essential fatty acids (EFAs), linoleic acid (Omega 6) and linolenic acid (Omega
3). These fatty acids are present in hemp oil in the ratio of 3:1, which is the “optimal
ratio” for health benefits (Erasmus, 1993).

“The membrane (coating) of every cell in our body is composed of oil.
It is this oil that acts as a superconductor allowing an unimpeded flow
of the bio-electric currents that govern nerve, muscle, heart and
membrane functions. The oil component of our diet normally comes
from fresh, well stored seeds, nuts, vegetables, and a few fruits. If the
oils in your diet are primarily of low quality, such as supermarket oils
and fats that stay solid at room temperature, then the oil coating on
your cells are going to have some of the insulating properties of tar and
be a less than ideal conductor. Conversely, if the oils in your diet are in
a pure, unadulterated form, the bio-electric current will flow much
smoother, and all our bodily functions will be easier to perform:
everything from the pancreas secreting insulin to keep our blood sugar
levels balanced, to keeping our hormone system in check, to
alleviating the buildup of the heavy LDL cholesterol that plugs our
arteries leading to heart disease and arterial sclerosis. Pioneers in the
fields of biochemistry and human nutrition now believe
cardiovascular disease (CVD) and most cancers are really diseases of
fatty degeneration caused by the continued over consumption of
saturated fats and refined vegetable oils that turn essential fatty acids
into carcinogenic killers. One out of two Americans will die from the
effects of CVD. One out of four Americans will die from cancer.
Researchers believe cancers erupt when immune system response is
weakened” (Thorpe, 1999, p. 32).

If, however, cell membranes are constructed from “fats that heal”  the best of
which is hemp oil (Erasmus, 1993)  the health benefits are considerable. Perhaps
the primary benefit is the effect of the EFAs on the heart. Hemp oil reduces the
level of bad cholesterol (LDL), reduces inflammation in blood vessels, thins the
blood (by reducing platelet stickiness), and reduces blood pressure. Thus,
hypertension is relieved and the risk of heart attacks and strokes is reduced. The
chances of heart disease in general are significantly reduced. In October, 2000, the
American Heart Association issued a recommendation that Americans consume
foods with high levels of Omega 3 (Gorman, 2000); the most balanced common
source of this EFA is hemp oil.

A second major benefit of hemp oil is a strengthening of the immune system. It
inhibits tumour growth, kills bacteria (including staph), and heals wounds
(Erasmus, 1993).

In summary, the EFAs in hemp oil are used to:

The EFAs in hemp oil will beneficially affect (partial list):

5.1.2 Hemp nuts, hemp flour, and hemp protein

When the oil is squeezed from hemp seed, the remaining press cake is made into
flour for hemp breads, pastries, and other products. This press cake contains two
high-quality proteins called edestin and albumin. These proteins contain all eight of
the essential amino acids in highly favourable proportions, and they are easier to
digest than the protein in soybeans and other foods. Like EFAs, proteins are the
basic building blocks of the human body. There are few bodily functions that are not
affected, in a positive way, by hemp protein.

The dehulled hemp seed (hempnut) is perhaps the best way to ingest hemp foods.
The delicious hempnut contains not only the proteins mentioned above, but also
the highly beneficial EFAs, better preserved in the nutmeat matrix.

An interesting report has turned up on the use of hemp protein to treat tuberculosis
in Czechoslovakia during and after World War II (Sirek, 1954). At an institution for
children with TB, doctors had no medicine and very little food. The doctors decided
to treat the children with hemp seeds, because of the protein (edestin) in the
nutmeat. Edestin containes not only the appropriate amino acids (including
arginine, essential for formation and growth of new tissue) but also a wealth of
healthy enzymes. A total of 26 children were treated with a diet of hemp seed, oats,
and cottage cheese. All 26 were cured or significantly improved, and all grew to be
healthy young adults.

5.2 Benefits of Hemp Bodycare Products

Hemp oils are used to make body lotions, soaps, and other products that heal the
skin, restoring natural health and beauty. The essential fatty acids (EFAs) that are
used by the body to build and maintain healthy body cells (especially the
membranes) work directly on epidermal cells, entering the lipid layers of dry skin
cells to replenish their oils (Ohio Hempery, undated).

The EFAs also repair skin damage, promoting healing in wounds and burns, and
they are antibiotic. Research has shown that EFAs are effective treatments for atopic
dermatitis, eczema, and psoriasis (Fitzpatrick, 2000).
6.0 CONCLUSIONSHemp foods and bodycare products are among the healthiest substances that
humans consume. In their essential fatty acids and proteins, hemp products
provide the basic building blocks that our bodies use to construct cells and tissue for
healthy and efficient functioning.

Health Canada, in a draft risk assessment, has raised the question of possible health
risks associated with 10 ppm levels of THC, the psychoactive ingredient in hemp.
We have reviewed the relevant research and concluded that there are no health
risks from low level doses of THC. None. There may indeed be health benefits:
several studies have shown strengthening of the immune system.

The Health Canada risk assessment is based on poorly-designed research. Most of
the research showing possible health risks with THC ingestion uses massive doses
of THC, far more than even those levels consumed by the heaviest marijuana
smokers. Every study showing health risks has been discredited or refuted; cannot
be replicated; or has been shown to be in error by a majority of studies on a given

The appropriate risk assessment for hemp foods and cosmetics would show that
there are no health risks, only benefits. We believe that the current Canadian
standard, requiring less than 10 ppm THC in hemp products, is too low; we have
calculated 20 ppm as sufficient to protect consumers from any possible psychoactive
reactions (and even these reactions are not a health risk). But with considerable
effort, the hemp industry has found it possible to prepare hemp products with less
than 10 ppm, and is willing to accept that standard.

Topics: Studies | Comments Off

Comments are closed.