The health and psychological consequences of cannabis use - chapter 7

THIS DOCUMENT HAS BEEN RESCINDED: Chapter 7.5 Chronic cannabis use and brain damage.

Page last updated: 1994

Please note the information in this publication may no longer be current but is retained on our website for historical or research purposes.

<< Previous Chapter | Contents | Next Chapter>>

7. The psychological effects of chronic cannabis use

7.5 Chronic cannabis use and brain damage

A major concern about the recreational use of cannabis has been whether it may lead to functional or structural neurotoxicity, or "brain damage" in ordinary language. Fehr and Kalant (1983) defined neurotoxicity as "functional aberrations qualitatively distinct from the characteristic usual pattern of reversible acute and chronic effects, and that may be caused by identified or identifiable neuronal damage" (p27). On this definition, an enduring impairment of cognitive functioning may be interpreted as a manifestation of neurotoxicity. Since such impairments are discussed in the previous chapter on chronic effects on cognitive functioning, this chapter will concentrate on direct investigations of neurological function and structural brain damage arising from exposure to cannabinoids. The review begins with an examination of the evidence for behavioural neurotoxicity from animal studies. Neurochemical, electrophysiological and brain substrate investigations of functionality follow, and the chapter concludes with the findings of more invasive examinations of brain structure and morphology in animals, and of less invasive techniques for imaging the human brain.

Top of page

7.5.1 Behavioural neurotoxicity in animals

Animal research provides the ultimate degree of control over extraneous variables; it is possible to eliminate factors known to influence research findings in humans, e.g. nutritional status, age, sex, previous drug history, and concurrent drug use. The results, however, are often difficult to extrapolate to humans because of between-species differences in brain and behaviour and in drug dose, patterns of use, routes of administration and methods of assessment.

Animal research on the effects of cannabis on brain function has typically administered known quantities of cannabinoids to animals for an extended period of time and then examined performance on various tasks assessing brain function, before using histological and morphometric methods to study the brains of the exposed animals. In general, the results of studies with primates produce results that most closely resemble the likely effects in humans; the monkey is physiologically similar to humans, while rats, for example, metabolise drugs in a different way; and monkeys are able to perform complex behavioural tasks. Nevertheless, every animal species examined to date has been found to have cannabinoid receptors in the brain. In animal models, non-targeted staring into space following administration of cannabinoids is suggestive of psychoactivity comparable to that in humans. The most characteristic responses to cannabinoids in animals are mild behavioural aberrations following small doses, and signs of gross neurotoxicity manifested by tremors and convulsions following excessively large doses. Where small doses are given for a prolonged period of time, evidence of behavioural neurotoxicity has emerged (see Rosenkrantz, 1983). Chronic exposure produces lethargy, sedation and depression in many species, and/or aggressive irritability in monkeys.

A clear manifestation of neurotoxicity in rats is what has been called the "popcorn reaction" (Luthra, Rosenkrantz and Braude, 1976), characterised by a pattern of sudden vertical jumping in rats exposed to cannabinoids for five weeks or longer. It is also seen in young animals exposed to cannabinoids in utero and then given a small dose challenge at 30 days of age. Several studies of prenatal exposure indicate that the offspring of cannabis treated animals show small delays in various stages of post-natal development, such as eye opening, various reflexes and open field exploration, although after several weeks or months their development is indistinguishable from normal (e.g. Fried and Charlebois, 1979). This means that either the developmental delay was not chronic, the remaining damage is too subtle to be detected by available measures, or the "plasticity of nervous system organisation in the newborn permitted adequate compensation for the loss of function of any damaged cells" (Fehr and Kalant, 1983, p29).

Top of page

Behavioural tests in rodents have included conventional and radial arm maze learning, operant behaviour involving time discriminations, open field exploration and two-way shuttle box avoidance learning. Correct performance on these tests is dependent on spatial orientation or on response inhibition, both of which are believed to depend heavily on intact hippocampal functioning. Some studies have found decreased learning ability on such tasks several months after long-term treatment with cannabinoids (see Fehr and Kalant, 1983). For example, Stiglick and Kalant (1982a, 1982b) reported altered learning behaviour in rats one to six months after a three-month oral dosing regimen of marijuana extract or THC. They claimed that the deficits were reminiscent of behavioural changes seen after damage to the hippocampus. Long lasting impairment of learning ability and hippocampal dysfunction suggests that long-lasting damage may result from exposure to cannabis. However, some studies have been carried out too soon after last drug administration to exclude the possibility that the observed effects are residual effects, that is, due to the continued action of accumulated cannabinoids.

Memory function in monkeys has often been assessed by delayed matching-to-sample tasks. In a recent study (Slikker et al, 1992), rhesus monkeys were trained for one year to perform five operant tasks before one year of chronic administration of cannabis commenced. One group was exposed daily to the smoke of one standard joint, another on weekends only, and control groups received sham smoke exposure (N=15 or 16 per group). Performance on the tasks indicated the induction of an "amotivational syndrome" during chronic exposure to cannabis, as manifested in a decrease in motivation to respond, regardless of whether the monkeys were exposed daily or only on weekends. This led the authors to suggest that motivational problems can occur at relatively low or recreational levels of use (in fact the effect was maximal with intermittent exposure). Task performance was grossly impaired for more than a week following last exposure, although performance returned to baseline levels two to three months after cessation of use. Thus, the effects of chronic exposure were slowly reversible with no long-term behavioural effects, and the authors concluded that persistent exposure to compounds that are very slowly cleared from the brain could account for their results. This hypothesis is consistent with the long half life of THC in the body (see Section 4.7 on metabolism).

One of the problems with these studies is that animals are often only exposed for a relatively short period of time, for example, one year or less. Slikker and colleagues acknowledge that it remains to be determined whether longer or greater exposures would cause more severe or additional behavioural effects. It may be that chronic dysfunction becomes manifest only after many years of exposure. Nevertheless, it is of concern that behavioural impairments have been shown to last for several months after exposure, but reassuring that they have generally resolved over time.

A further difficulty with animal studies is a consequence of differences between animals and humans in route of cannabinoid administration. In humans, the most common route of exposure to THC is via the inhalation of marijuana smoke, whereas most animals studies have relied upon the oral administration or injection of THC because of the difficulty in efficiently delivering smoke to animals and the concern about the complications introduced by carbon monoxide toxicity. While it may well be impossible to evaluate the pharmacological and toxicological consequences of exposure to the hundreds of compounds in cannabis simultaneously, it is arguably inappropriate to assess the long-term consequences of human cannabis smoking by administering THC alone. Hundreds of additional compounds are produced by pyrolysis when marijuana is smoked, which may contribute either to acute effects or to long-term toxicity. Future studies need to address these issues for comparability to human usage. Appropriate controls, including those which mimic the carbon monoxide exposure experienced during the smoking of marijuana may be necessary.

Top of page

7.5.2 Neurochemistry

The discovery of the cannabinoid receptor and its endogenous ligand anandamide revolutionised previous conceptions of the mode of action of the cannabinoids. However, much further research is required before the interactions between ingested cannabis, anandamide and the cannabinoid receptor are fully understood. Nor should the anandamide pathways be seen as responsible for all of the central effects of the psychoactive cannabinoids. There is good evidence that cannabinoids affect the concentration, turnover, or release of endogenous substances (see Pertwee, 1988). Much research has been devoted to examining the interactions between cannabinoids and several neurotransmitter receptor systems (e.g. norepinephrine, dopamine, 5-hydroxytryptamine, acetylcholine, gamma-aminobutyric acid (GABA), histamine, opioid peptides, and prostaglandins). The results suggest that all these substances have some role in the neuropharmacology of cannabinoids, although little is known about the precise nature of this involvement. Cannabinoids may alter the activities of neurochemical systems in the central nervous system by altering the synaptic concentrations of these mediators through an effect on their synthesis, release, or metabolism, and/or by modulating mediator-receptor interactions. There have been numerous reports of neurotransmitter perturbations in vitro and after short-term administration (see Martin, 1986; Pertwee, 1988).

Domino (1981) demonstrated in cats that large doses of THC elevate brain acetylcholine and reduce its turnover by producing a decrease in acetylcholine release from the neocortex. At large doses, THC may depress the brain stem activating system. With lower doses, brain acetylcholine utilisation was reduced primarily only in the hippocampus. Some of the potential undesirable side effects of cannabinoids may be related to a decrease in acetylcholine release and turnover. Domino also reported that THC decreased EEG activation and induced slow wave activity: high voltage slow waves in neocortical EEG were produced in frontal regions and tended to be exaggerated by small doses but reduced by larger doses. These findings support the general observation made in a variety of studies, that low doses of THC stimulate, while high doses depress the noradrenergic and dopaminergic system.

Bloom (1984) reported that cannabinoids increase the synthesis and turnover of dopamine and norepinephrine in rat and mouse brain, while producing little or no change in endogenous levels of catecholamines. However, THC and other cannabinoids were reported to alter functional aspects of catecholaminergic neurotransmission. THC was found to increase the utilisation of the catecholamine neurotransmitters, to alter the active uptake of biogenic amine neurotransmitters and their precursors into synaptosomes, and to alter transmitter release from synaptosomes. Further, THC was reported to alter the activity of enzymes involved in the synthesis and degradation of the catecholamines. THC and other cannabinoids can selectively alter the binding of ligands to several different membrane-bound neurotransmitter receptors.

Relatively few studies have examined whether long-term exposure to cannabinoids results in lasting changes in brain neurotransmitter and neuromodulator levels. An early study examined cerebral and cerebellar neurochemical changes accompanying behavioural manifestations of neurotoxicity (involuntary vertical jumping) in rats exposed to marijuana smoke for up to 87 days (Luthra, Rosenkrantz and Braude, 1976). Sex differences emerged in the neurochemical consequences of chronic exposure: in females, AChE showed a cyclic increase and cerebellar enzyme activity declined. For both sexes cerebellar RNA increased, but at different times for each sex, and at 87 days remained elevated only in females. Some of these neurochemical changes persisted during a 20-day recovery period, but the authors predicted the return to normality after a much longer recovery period. Cannabinoids administered prenatally not only impaired developmental processes in rats, but produced significant decrements in RNA, DNA and protein concentrations and reductions in amine concentrations (dopamine, norepinephrine) in mice, which could be important in the role of protein and nucleic acids in learning and memory (see Fehr and Kalant, 1983). Bloom (1984) reported that chronic exposure to cannabinoids has been shown to lead to increased activity of tyrosine in rat brain.

Top of page

However, recent evidence suggests that there are few, if any, irreversible effects of THC on known brain chemistry. Ali and colleagues (1989) administered various doses of THC to rats for 90 days and then assessed several brain neurotransmitter systems 24 hours or two months after the last drug dose. Examination of dopamine, serotonin, acetylcholine, GABA, benzodiazepine and opioid neurotransmitter systems revealed that no significant changes occurred. A larger study with both rats and monkeys examined receptor binding of the above neurotransmitters and the tissue levels of monoamines and their metabolites (Ali et al, 1991). No significant irreversible changes were demonstrated in the rats chronically treated with THC. Monkeys exposed to a chronic treatment of marijuana smoke for one year and then sacrificed after a seven-month recovery period were found to have no changes in neurotransmitter concentration in caudate, frontal cortex, hypothalamus, or brainstem regions. The authors concluded that there are no significant irreversible alterations in major neuromodulator pathways in the rat and monkey brain following long-term exposure to the active compounds in marijuana.

Slikker et al (1992), reporting on the same series of studies, noted that there were virtually no differences between placebo, low dose or high dose groups of monkeys in blood chemistry values. The general health of the monkeys was unaffected, but the exposure served as a chronic physiological stressor, evidenced by increases in urinary cortisol levels which were not subject to tolerance (although plasma cortisol levels did not differ). Urinary cortisol elevation has not been demonstrated in other studies with monkeys. Slikker et al reported a 50 per cent reduction in circulating testosterone levels in the high dosed group, with a dramatic (albeit non-significant) rebound one to four weeks after cessation of treatment. It is worth noting that these monkeys were three years of age at the commencement of the study and would have experienced hormonal changes over the course of entering adolescence during the study.

A recent pilot study compared monoamine levels in cerebrospinal fluid (CSF) in a small sample of human cannabis users and age and sex-matched normal controls (Musselman et al, 1993). The justification for the study was that THC administered to animals has been shown to produce increases in serotonin and decreases in dopamine activity. No differences were found between the user and non-user groups in the CSF concentration of HVA, 5HIAA, MHPG, ACTH and CRH. The authors proposed a number of explanations for these results: (1) cannabis use has no chronic effect on levels of brain monoamines; (2) those who use cannabis have abnormal levels of brain monoamines which are normalised over long periods of time by cannabis use; or (3) those who use cannabis have normal levels of brain monoamines which are transiently altered with cannabis use and then return to normal. However, there is insufficient data from this study to permit a choice between these hypotheses to be made. The frequency and duration of cannabis use, and the time since last use in the user group could not be determined. All users had denied using cannabis, having been drawn from a larger normative sample and identified as cannabis users by the detection of substantial levels of cannabinoids in urine screens. Furthermore, the "normal" controls were assumed to be non-users on the basis of their drug free urines, a far from adequate source of evidence for or against cannabis use. Thus, the small sample size and faulty methodology preclude any conclusions to be drawn from this study about possible alterations in monoamine levels in cannabis users.

Top of page

7.5.3 Electrophysiological effects

Cannabis is clearly capable of causing marked changes in brain electrophysiology as determined by electroencephalographic (EEG) recordings. Long-term residual abnormalities in EEG tracings from cortex and hippocampus have been shown in cats (Barratt & Adams, 1972; Domino, 1981; Hockman et al, 1971), rats (see Fehr and Kalant, 1983) and monkeys (Heath et al, 1980) exposed to cannabinoids. Withdrawal effects are also apparent in the EEG (see Fehr and Kalant, 1983), with epileptiform and spike-like activity most often seen.

Shannon and Fried (1972) related EEG changes in rat to the distribution of bound and unbound radioactive THC. Disposition of the tracer was primarily in the extra-pyramidal motor system and some limbic structures, and 0.8 per cent of the total injected drug which was weakly bound in the brain accounted for the EEG changes. In monkeys, serious subcortical EEG anomalies were observed in those exposed to marijuana smoke for six months (Heath et al, 1980). The septal region, hippocampus and amygdala were most profoundly affected, showing bursts of high amplitude spindles and slow wave activity. Such early studies often lacked critical quantitative analysis. The definition of abnormal spike-like waveforms in EEG were not made to rigorous criteria,and EEG frequency was not assessed quantitatively.

More recent studies have examined the effects of THC on extracellular action potentials recorded from the dentate gyrus of the rat hippocampus (Campbell et al, 1986a; 1986b). THC produced a suppression of cell firing patterns and a decrease in the amplitude of sensory-evoked potentials, also impairing performance on a tone discrimination task. The evoked-potential changes recovered rapidly (within four hours), but the spontaneous and tone-evoked cellular activity remained significantly depressed, indicating an abnormal state of hippocampal/limbic system operation. The authors proposed that such changes accounted for decreased learning, memory function and general cognitive performance following exposure to cannabis. The long-lasting effects of prolonged cannabis administration on animal electrophysiology has not been investigated to any degree of specificity.

The waking or sleep EEG is increasingly recognised as a particularly sensitive tool for evaluating the effects of drugs, especially drugs that affect the CNS, since EEG signals are sensitive to variables affecting the brain's neurophysiological substrate. The recording of the EEG is one of the few reasonably direct, non-intrusive methods of monitoring CNS activity in man. However, alterations in EEG activity are difficult to interpret in a functional sense. Struve and Straumanis (1990) provide a review of the human research dating from 1945 on the EEG and evoked potential studies of acute and chronic effects of cannabis use. While the data have often been contradictory, the most typical human alterations in EEG patterns include an increase in alpha activity and a slowing of alpha waves with decreased peak frequency of the alpha rhythm, and a decrease in beta activity (Fink, 1976; Fink et al, 1976). In general, this is consistent with a state of drowsiness. Desynchronisation, variable changes in theta activity, abnormal sleep EEG profiles and abnormal evoked responses have also been reported (see Fehr and Kalant, 1983). Clinical reports have associated cannabis with triggering seizures in epileptics (Feeney, 1979) and experimental studies have shown THC to trigger abnormal spike waveforms in the hippocampus, whereas cannabidiol has an opposite effect. Yet there is suggestive evidence that cannabis may be useful in the treatment of convulsions. Feeney (1979) discusses these paradoxical effects.

Top of page

A number of studies have investigated EEG in chronic cannabis users. The early cross-cultural studies were flawed in many respects (see Section 7.4 on cognitive functioning) and also failed to used quantitative techniques in analysing EEG spectra. No EEG abnormalities were found in the resting EEG of chronic users from Greek, Jamaican or Costa Rican populations compared to controls (Karacan et al, 1976; Rubin and Comitas, 1975; Stefanis, 1976). In all of these studies, only subjects who were in good health and who were functioning adequately in the community, were selected, thereby systematically eliminating subjects who may have been adversely affected by cannabis use and who may therefore have shown residual EEG changes. The evidence from many studies has been contradictory: users have been found to show either higher or lower percentages of alpha-components than non-users, and to have higher or lower visual evoked response amplitudes (Richmon et al, 1974). Subjects in a 94-day cannabis administration study (Cohen, 1976) showed lasting EEG changes. The abnormalities were more marked in subjects who had taken heavier doses, but it was observed that even in abstinence, cannabis users had more EEG irregularities than non-using controls. It was not determined for how long after cessation of use the EEG changes persisted. It has also been reported that chronic users develop tolerance to some of the acute EEG changes caused by cannabis (Feinberg et al, 1976). The question as to why chronic cannabis users can continue to display changes in EEG when tolerance is known to develop to such alterations remains unanswered.

In a series of well controlled ongoing studies, Struve and colleagues (1991, 1993) have been using quantitative techniques to investigate persistent EEG changes in long-term cannabis users, characterised by a "hyperfrontality of alpha". Significant increases in absolute power, relative power and interhemispheric coherence of EEG alpha activity over the bilateral frontal-central cortex in daily THC users compared to non-users were demonstrated and replicated several times. The quantitative EEGs of subjects with excessively long cumulative THC exposures (>15 years) appear to be characterised by increases in frontal-central theta activity in addition to the hyperfrontality of alpha found in THC users in general (or those with much shorter durations of use). Ultra-long-term users have shown significant elevations of theta absolute power over frontal-central cortex compared to short-term users and controls, and significant elevations in relative power of frontal-central theta in comparison to short-term users. Over most cortical regions, ultra-long-term users had significantly higher levels of theta interhemispheric coherence than short-term users or controls. Thus, excessively long duration of THC exposure (15-30 years) appears to be associated with additional topographic quantitative EEG features not seen with subjects using THC for short to moderately long time periods.

These findings have led to the suspicion that there may be a gradient of quantitative EEG change associated with progressive increases in the total cumulative exposure (duration in years) of daily THC use. Infrequent, sporadic or occasional THC use does not seem to be associated with persistent quantitative EEG change. As daily THC use begins and continues, the topographic quantitative EEG becomes characterised by the hyperfrontality of alpha. While it is not known at what point during cumulative exposure it occurs, at some stage substantial durations of daily THC use become associated with a downward shift in maximal EEG spectral power from the mid alpha range to the upper theta/low alpha range. Excessively long duration cumulative exposure of 15-30 years may be associated with increases of absolute power, relative power and coherence of theta activity over frontal-central cortex. One conjecture is that the EEG shift toward theta frequencies, if confirmed, may suggest organic change. These data are supplemented by neuropsychological test performance features separating long-term users from moderate users and non-users, however the relationship between neuropsychological test performance and EEG changes has not been investigated.

While the EEG provides little interpretable information about brain function, brain event-related potential measures are direct electrophysiological markers of cognitive processes. Studies by Herning et l (1979) demonstrated that THC administered orally to volunteers alters event-related potentials according to dose, duration of administration, and complexity of the task. Event-related potential studies of chronic cannabis users in the unintoxicated state have provided evidence for long-lasting functional brain impairment and subtle cognitive deficits (see Section 7.4 on cognitive functioning).

Top of page

7.5.4 Cerebral blood flow studies

Brain cerebral blood flow (CBF) is closely related to brain function. The use of CBF may help to identify brain regions responsible for the behavioural changes associated with drug intoxication. Since, however, psychoactive drugs may induce CBF changes through mechanisms other than alteration in brain function (e.g. by increasing carbon monoxide levels, changing blood gases or vasoactive properties, affecting blood viscosity, autonomic activation or inhibition of intraparenchymal innervation, acting on vasoactive neuropeptides), any conclusions drawn from drug-induced CBF changes must be treated with caution.

Mathew and Wilson (1992) report several studies of cannabis effects on cerebral blood flow. Acutely, cannabis administration to inexperienced users produced global CBF decreases, while acute intoxication increased CBF in both hemispheres, at frontal and the left temporal regions, in experienced users. There was an inverse relationship between anxiety and CBF. The authors attributed the decrease in CBF in naive subjects to their increased anxiety after cannabis administration, while the increased CBF in experienced users was attributed to the behavioural effects of cannabis. A further study showed that the largest increases in CBF occurred 30 minutes after smoking. The authors concluded that cannabis causes a dose related increase in global CBF, but also appears to have regional effects, with a greater increase in the frontal region and in particular in the right hemisphere. CBF increases were correlated with the "high", plasma THC levels and pulse rate, loss of time sense, depersonalisation, anxiety and somatisation scores (positively with frontal flow and inversely with parietal flow).

The authors claimed their results suggested that altered brain function was mainly, if not exclusively, responsible for the CBF changes. Carbon monoxide increased after both cannabis and placebo but did not correlate with CBF. Cannabis induced "red eye" lasted for several hours, but the CBF increases declined significantly within two hours of smoking. Nevertheless, the possibility remains that the CBF changes reflected drug-induced vascular (cerebral) change. Continued reduction in cerebral blood velocity was demonstrated following cannabis administration, and reports of dizziness but with normal blood pressure suggested that cannabis may impair cerebral autoregulation.

The time course of CBF changes resembled that of mood changes more closely than plasma THC levels. Global CBF was closely related to levels of arousal mediated by the reticular activating system. High arousal states generally show CBF increases while low arousal states show CBF decreases. Of all cortical regions, the frontal lobe has the most intimate connections with the thalamus which mediates arousal, and CBF increases after cannabis use were most pronounced in frontal lobe regions. The right hemisphere is known to mediate emotions, and the most marked changes after cannabis were seen there. Time sense and depersonalisation, which are associated with the temporal lobe, were severely affected, but there were no significant correlations between these scores and temporal flow. Perhaps CBF techniques are not sensitive enough in terms of spatial resolution to detect such effects. The parietal lobes are associated with perception and cognition. Cannabis reduces perceptual acuity, but during intoxication subjects report increased awareness of tactile, visual and auditory stimuli. Maybe their altered time sense and depersonalisation are related to such altered awareness.

There have been a few investigations of chronic effects of cannabis on CBF. Tunving et al, (1986) demonstrated globally reduced resting levels of CBF in chronic heavy users of 10 years compared to non-user controls, but no regional flow differences were observed. CBF increased by 12 per cent between nine and 60 days later, indicating reduced CBF in heavy users immediately after cessation of cannabis use, with a return to normal levels with abstinence. This study was flawed in that some subjects were given benzodiazepines (which are known to lower CBF) prior to the first measurement. Mathew and colleagues (1986) assessed chronic users of at least six months (mean 83 months) after two weeks of abstinence. No differences in CBF levels were found between users and non-user controls. The number of studies available on the effects of cannabis on CBF are relatively small. Use of techniques with better spatial resolution and the ability to quantify subcortical flow, such as positron emission tomography (PET), would yield more useful findings.

Top of page

7.5.5 Positron emission tomography (PET) studies

Positron emission tomography (PET) is a nuclear imaging technique which allows the concentration of a positron-labelled tracer to be imaged in the human brain. PET can measure the time course and regional distribution of positron-labelled compounds in the living human brain. Most PET studies have utilised an analog of glucose to measure regional brain glucose metabolism, since nervous tissue uses glucose as its main source of energy. Measurement of glucose metabolism reflects brain function, since activation of a given brain area is indicated by an increase in glucose consumption. PET may be used to assess the effects of acute drug administration by using regional brain glucose metabolism to determine the areas of the brain which are activated by a given drug. Assessment of brain glucose metabolism has been useful in identifying patterns of brain dysfunction in patients with psychiatric and neurological diseases. It is a direct and sensitive technique for identifying brain pathology, since it can detect abnormalities in the functioning of brain regions in the absence of structural changes, such as is likely to occur with the neurotoxic effects of chronic drug use. It is accordingly more sensitive than either computer-assisted tomography (CAT) scans or magnetic resonance imaging (MRI) in detecting early pathological changes in the brain.

Only one study to date has used the PET technique to investigate the effects of cannabis use. Volkow et al (1991) reported preliminary data from an investigation comparing the acute effects of cannabis in three normal subjects (who had used cannabis no more than once or twice per year) and in three chronic users (who had used at least twice a week for at least 10 years). The regions of interest were the prefrontal cortex, the left and right dorsolateral, temporal, and somatosensory parietal cortices, the occipital cortex, basal ganglia, thalamus and cerebellum. A measure of global brain metabolism was obtained using the average for the five central brain slices, and relative measures for each region were obtained using the ratios of region/global brain metabolism. Due to the small number of subjects, descriptive rather than inferential statistical procedures were used for comparison. The relation between changes in metabolism due to cannabis and the subjective sense of intoxication was tested with a regression analysis.

In the normal subjects, administration of cannabis led to an increase in metabolic activity in the prefrontal cortex and cerebellum; the largest relative increase was in the cerebellum and the largest relative decrease was in the occipital cortex. The degree of increase in metabolism in the cerebellar cortex was highly correlated with the subjective sense of intoxication. The cannabis users reported less subjective effects than the normal controls and showed less changes in regional brain metabolism, reflecting tolerance to the actions of cannabis. However, the authors did not report comparisons of baseline levels of activity in the users and controls, perhaps due to the limitations of the small sample size. Such a comparison would have enabled some evaluation of the consequences of long-term cannabis use on resting levels of glucose metabolism. The increases in regional metabolism are in accord with the increases in cerebral blood flow reported by Mathew and Wilson (1992). The regional pattern of response to cannabis in this study is consistent with the localisation of cannabinoid receptors in brain. A further application of PET would be to label cannabinoids themselves: labelling of cannabis with a positron emitter has been achieved and preliminary biodistribution studies have been carried out in mice and in the baboon (Charalambous et al, 1991; Marciniak et al, 1991). The use of PET in future human studies is promising.

Top of page

7.5.6 Brain morphology Animal studies

Early attempts to investigate the effects of chronic cannabinoid exposure on brain morphology in animals failed to demonstrate any effect on brain weight or histology under the light microscope. Electron microscopic examination, however, has revealed alterations in septal, hippocampal and amygdaloid morphology in monkeys after chronic treatment with THC or cannabis. A series of studies from the same laboratory (Harper et al, 1977; Myers and Heath, 1979; Heath et al, 1980 discussed below) reported widening of the synaptic cleft, clumping of synaptic vesicles in axon terminals, and an increase in intranuclear inclusions in the septum, hippocampus and amygdala. These findings incited a great deal of controversy, and the studies were criticised for possible technical flaws (Institute of Medicine, 1982), with claims that such alterations are not easily quantifiable.

Harper et al (1977) examined the brains of three rhesus monkeys six months after exposure to marijuana, THC or placebo, and two non-exposed control monkeys. In the treated group, one monkey was exposed to marijuana smoke three times each day, five days per week; another was injected with THC once each day and the third was exposed to placebo smoke conditions. The latter two had electrode implants for EEG recording and had shown persistent EEG abnormalities following their exposure to cannabis. Morphological differences were not observed by light microscopy, but electron microscopy revealed a widening of the synaptic cleft in the marijuana and THC treated animals, with no abnormalities detected in the placebo or control monkeys. Further, "clumping" of synaptic vesicles as observed in pre- and post-synaptic regions in the cannabinoid treated monkeys, and opaque granular material was present within the synaptic cleft. The authors concluded that chronic heavy use of cannabis alters the ultrastructure of the synapse, and proposed that the observed EEG abnormalities may have been related to these changes.

Myers and Heath (1979) examined the septal region of the same two cannabinoid treated monkeys, and found the volume density of the organised rough endoplasmic reticulum to be significantly lower than that of the controls, and fragmentation and disorganisation of the rough endoplasmic reticulum patterns, free ribosomal clusters in the cytoplasm, and swelling of the cisternal membranes was observed. The authors noted that similar lesions have been observed following administration of various toxins or after axonal damage, reflecting disruptions in protein synthesis.

Heath et al (1980) extended the above findings by examining a larger sample of rhesus monkeys (N=21) to determine the effects of marijuana on brain function and ultrastructure. Some animals were exposed to smoke of active marijuana, some were injected with THC and some were exposed to inactive marijuana smoke. After two to three months of exposure, those monkeys that were given moderate or heavy exposure to marijuana smoke developed chronic EEG changes at deep brain sites, which were most marked in the septal, hippocampal and amygdaloid regions. These changes persisted throughout the six to eight month exposure period, as well as the postexposure observation period of between one and eight months. Brain ultrastructural alterations were characterised by changes at the synapse, destruction of rough endoplasmic reticulum and development of nuclear inclusion bodies. The brains of the placebo and control monkeys showed no ultrastructural changes. The authors claimed that at the doses used, which were comparable to human usage, permanent alterations in brain function and ultrastructure were observed in these monkeys.

Top of page

Brain atrophy is a major non-specific organic alteration which must be preceded by more subtle cellular and molecular changes. Rumbaugh et al (1980) observed six human cases of cerebral atrophy in young male substance abusers of primarily alcohol and amphetamines. They then conducted an experimental study of six rhesus monkeys treated chronically with various doses of cannabis extracts orally for eight months and compared them to groups that were treated with barbiturates or amphetamines, or untreated. No signs of cerebral atrophy were demonstrated in the cannabis exposed group, and light microscopy revealed no histological abnormalities in four of the animals, but "equivocal" results for the other two. Brains were not examined under the electron microscope. The amphetamine treated group showed the greatest histological, cerebrovascular and atrophic changes.

More recently, McGahan et al (1984) used high resolution computerised tomography scans in three groups of four rhesus monkeys. One was a control group, a second was given 2.4mg/kg of oral THC per day for two to 10 months, and a third group received a similar daily dose over a five-year period. The dosage was considered the equivalent of smoking one joint a day. The groups receiving THC were studied one year after discontinuing the drug. There was a statistically significant enlargement of the frontal horns and the bicaudate distance in the brains of the five-year treated monkeys as compared to the control and short-term THC groups. This finding suggests that the head of the caudate nucleus and the frontal areas of the brain can atrophy after long-term administration of THC in doses relevant to human exposure.

A number of rat studies have found similar results to those in rhesus monkeys described above. Investigators have reported that after high dose cannabinoid administration, there is a decrease in the mean volume of rat hippocampal neurons and their nuclei, and that after low dose administration, there is a shortening of hippocampal dendritic spines. Scallet and coworkers (1987) used quantitative neuropathological techniques to examine the brains of rats seven to eight months after 90-day oral administration of THC. The anatomical integrity of the CA3 area of rat hippocampus was examined using light and electron microscopy. High doses of THC resulted in striking ultrastructural alterations, with a significant reduction in hippocampal neuronal and cytoplasmic volume, detached axodendritic elements, disrupted membranes, increased extracellular space and a reduction in the number of synapses per unit volume (i.e. decreased synaptic density). These structural changes were present up to seven months following treatment. Lower doses of THC produced a reduction in the dendritic length of hippocampal pyramidal neurons two months after the last dose, and a reduction in GABA receptor binding in the hippocampus, although the ultrastructural appearance and synaptic density appeared normal. The authors suggested that such hippocampal changes may constitute a morphological basis for the persistent behavioural effects demonstrated following chronic exposure to THC in rats, effects which resemble those of hippocampal brain lesions. These findings are in accord with those of Heath et al (1980) with rhesus monkeys, and the doses administered correspond to daily use of approximately six joints in humans.

A study by Landfield et al (1988) showed that chronic exposure to THC reduced the number of nucleoli per unit length of the CA1 pyramidal cell somal layer in the rat hippocampus. The brains of rats treated five times per week for four or eight months with 4-10mg/kg injected subcutaneously were examined by light and electron microscopy. Significant THC-induced changes were found in hippocampal structure; pyramidal neuronal cell density decreased and there was an increase in glial reactivity, reflected by cytoplasmic inclusions similar to that seen during normal aging or following experimentally induced brain lesions. However, no effects were observed on ultrastructural variables such as synaptic density. Adrenal-pituitary activity increased, resulting in elevated ACTH and corticosterone elevations during acute stress. The authors claimed that the observed hippocampal morphometric changes produced by THC exposure were similar to glucocorticoid-dependent changes that develop in rat hippocampus during normal aging. They proposed that, given the chemical structural similarity between cannabinoids and steroids, chronic exposure to THC may alter hippocampal anatomical structure by interacting with adrenal steroid activity. More recently, Eldridge et al (1992) reported that delta-8-THC bound with the glucocorticoid receptors in the rat hippocampus, and was displaced by corticosterone or delta-9-THC. A glucocorticoid agonist action of delta-9-THC injections was demonstrated. Injection of corticosterone increased hippocampal cannabinoid receptor binding. These interactions suggest that cannabinoids may accelerate brain aging.

Top pf page

It should be noted that where THC has been administered to monkeys for six months, this represents only 2 per cent of their life span and may not have been long enough to detect the gradual effects that could arise from interactions with steroid systems (and affect the aging process). In contrast, eight months administration to rats represents approximately 30 per cent of their life span. The differences in the ultrastructural findings of Landfield's and Scallet's studies may be due to the largely different doses administered; the 8mg/kg of Landfield's study was not sufficient to produce any marked behavioural effects. Further, the two studies examined slightly different hippocampal areas (CA1 or CA3).

Most recently, Slikker and colleagues (1992) reported the results of their neurohistochemical and electronmicroscopic evaluation of the rhesus monkeys whose dosing regime, behavioural and histochemical data were reported above. They failed to replicate earlier findings: no effects of drug exposure were found on the total area of hippocampus, or any of its subfields; there were no differences in hippocampal volume, neuronal size, number, length or degree of branching of CA3 pyramidal cell dendrites. Nor were there effects on synaptic length or width, but there were trends toward increased synaptic density (the number of synapses per cubic mm), increased soma size, and decreased basilar dendrite number in the CA3 region with marijuana treatment. Slikker et al (1992) were able to demonstrate an effect of enriched environments upon neuroanatomy: daily performance of operant tasks increased the total area of hippocampus and particularly the CA3 stratum oriens, producing longer, more highly branched dendrites and less synaptic density, while the reverse occurred in the animals deprived of the daily operant tasks. The extent of drug interaction with these changes was not clear and may explain some of the inconsistencies between this study and those described above. Clearly, the question of whether prolonged exposure to cannabis results in structural brain damage has not been fully resolved.

The development of tolerance following chronic administration of psychoactive compounds is often mediated by a down-regulation of receptors. Thus, chronic exposure to THC could result in a decreased number of cannabinoid receptors in the brain. Such receptor down-regulation and reduced binding has recently been demonstrated in rats (Oviedo, Glowa and Herkenham, 1993). However, previously Westlake et al (1991) found that cannabinoid receptor properties were not irreversibly altered in rat brain 60 days following 90-day administration of THC, nor in monkey brain seven months after one year of exposure to marijuana smoke. It was argued that these recovery periods were sufficient to allow the full recovery of any receptors that would have been lost during treatment. Nevertheless, studies have not yet confirmed the parameters of any alterations in cannabinoid receptor number and function that may result from chronic exposure to cannabinoids, and the extent of reversibility following longer exposures has not been determined.

Top of of page Human studies

There is very little evidence from human studies of structural brain damage. In their controversial paper, Campbell et al (1971) were the first to present evidence suggesting that structural/morphological brain damage was associated with cannabis use. They used air encephalography to measure cerebral ventricular size, and claimed to have demonstrated evidence of cerebral atrophy in ten young males who had used cannabis for three to 11 years, and who complained of neurological symptoms, including headaches, memory dysfunction and other cognitive impairment. Compared to controls, the cannabis users showed significantly enlarged lateral and third ventricular areas. Although this study was widely publicised in the media because of its serious implications, it was heavily criticised on methodological grounds. Most subjects had also used significant quantities of LSD and amphetamines, and the measurement technique was claimed to be inaccurate, particularly since there were great difficulties in assessing ventricular size and volume to any degree of accuracy (e.g. Bull, 1971; Susser, 1972; Brewer, 1972). Moreover, the findings could not be replicated. Stefanis (1976) reported that echoencephalographic measurements of the third ventricle in 14 chronic hashish users and 21 non-users did not support Campbell et al's pneumoencephalographic findings of ventricular dilation.

The introduction of more accurate and non-invasive techniques, in the form of computerised tomographic (CT) scans, (also known as computer-assisted tomographic (CAT) scans), permitted better studies of possible cerebral atrophy in chronic cannabis users (Co et al, 1977; Kuehnle et al, 1977). Co et al (1977), for example, compared 12 cannabis users recruited from the general community, with 34 non-drug using controls, all within the ages of 20-30. The cannabis users had used cannabis for at least five years at the level of at least five joints per day, and most had also consumed significant quantities of a variety of other drugs, particularly LSD. Kuehnle et al's (1977) subjects were 19 heavy users aged 21-27 years, also recruited from the general community who had used on average between 25 and 62 joints per month in the preceding year, although their duration of use was not reported. CT scans were obtained presumably at the end of a 31-day study, which included 21 days of ad libitum smoking of marijuana (generally five joints per day), and were compared against a separate normative sample. No evidence for cerebral atrophy in terms of ventricular size and subarachnoid space was found in either study. Although these studies could also be criticised for their research design (e.g. inappropriate control groups, and the fact that cannabis users had used other drugs), these flaws would only have biased the studies in the direction of detecting significant differences between groups, yet none were found. The results were interpreted as a refutation of Campbell's findings, and supporting the absence of cortical atrophy demonstrated by Rumbaugh et al's (1980) CAT scans of monkeys. A further study (Hannerz and Hindmarsh, 1983) investigated 12 subjects who had smoked on average 1g of cannabis daily for between six and 20 years, by thorough clinical neurological examination and CT scans. As in the studies above, no cannabis related abnormalities were found on any assessment measure.

Top of page

7.5.7 Discussion

Surprisingly few studies of neurotoxicity have been published, and the results have been equivocal. There is convincing evidence that chronic administration of large doses of THC leads to residual changes in rodent behaviours which are believed to depend upon hippocampal function. There is evidence for long-term changes in hippocampal ultrastructure and morphology in rodents and monkeys. Animal neurobehavioural toxicity is characterised by residual impairment in learning, EEG and biochemical alterations, impaired motivation and impaired ability to exhibit appropriate adaptive behaviour. Although extrapolation to man is not possible, the results of these experimental studies have demonstrated cannabinoid toxicity at doses comparable to those consumed by humans using cannabis several times a day. There is sufficient evidence from human research to suggest that the cannabinoids act on the hippocampal region, producing behavioural changes similar to those caused by traumatic injury to that region.

The cognitive, behavioural and functional responses to long-term cannabis consumption in animals and man appear to be the most consistent manifestation of its potential neurotoxicity. The extent of damage appears to be more pronounced at two critical stages of central nervous system development: in neonates when exposed to cannabis during intrauterine life; and in adolescence, during puberty when neuroendocrine, cognitive and affective functions and structures of the brain are in the process of integration. As discussed in Section 7.4 on cognitive functioning, research needs to investigate the possibility that more severe consequences may occur in adolescents exposed to cannabinoids. Human research has defined a pattern of acute CNS changes following cannabis administration; there is convincing evidence for long-lasting changes in brain function after long-term heavy use; whether or not these changes are permanent has not been established.

Human studies of brain morphology have yielded generally negative results, failing to find gross signs of "brain damage" after chronic exposure to cannabis. Nevertheless, the results of many human studies are indicative of more subtle brain dysfunction. It may be that existing methods of brain imaging are not sensitive enough to establish subcellular alterations produced in the CNS. Many psychoactive substances exert their action through molecular biochemical mechanisms which do not distort gross cell architecture. The most convincing evidence on brain damage would come from postmortem studies, but this type of information has not been available.

In 1983, Fehr and Kalant concluded that "The state of the evidence at the present time does not permit one either to conclude that cannabis produces structural brain damage or to rule it out" (p602). Nahas (1984) wrote "The brain is the organ of the mind. Can one repetitively disturb the mental function without impairing brain mechanisms? The brain, like all other organs of the human body, has very large functional reserves which allow it to resist and adapt to stressful abnormal demands. It seems that chronic use of cannabis derivatives slowly erodes these reserves" (p299). In 1986, Wert and Raulin (1986) proposed, that on the available evidence "there are no gross structural or neurological deficits in marijuana-using subjects, although subtle neurological features may be present. However, the type of deficit most likely to occur would be a subtle, functional deficit which could be assessed more easily with either psychological or neuropsychological assessment techniques." (p624). In 1993, little further evidence has emerged to challenge or refute these earlier conclusions.

This conclusion was anticipated by the Parisian physician Moreau as early as 1845 when he observed:

...unquestionably there are modifications (I do not dare use the word "lesion") in the organ which is in charge of mental functions. But these modifications are not those one would generally expect. They will always escape the investigations of the researchers seeking alleged or imagined structural changes. One must not look for particular, abnormal changes in either the gross anatomical or the fine histological structure of the brain; but one must look for any alterations of its sensibility, that is to say, for an irregular, enhanced, diminished or distorted activity of the specific mechanisms upon which depends the performance of mental functions. (Moreau (de Tours), 1845).