Truth Seeker
New Member
Abstract
An increasing number of novel therapeutic agents are targeted at cannabinoid receptors. Drug development programmes of new cannabinoid drugs may be facilitated by the identification of useful biomarkers. This systemic literature review aims to assess the usefulness of direct biomarkers for the effects of cannabis and tetrahydrocannabinol (THC) in healthy volunteers. One hundred and sixty-five useful articles were found that investigated the acute effects of cannabis or THC on the central nervous system (CNS) and heart rate in healthy volunteers. Three hundred and eighteen tests (or test variants) were grouped in test clusters and functional domains, to allow their evaluation as a useful biomarker and to study their dose—response effects. Cannabis/THC affected a wide range of CNS domains. In addition to heart rate, subjective effects were the most reliable biomarkers, showing significant responses to cannabis in almost all studies. Some CNS domains showed indications of depression at lower and stimulation at higher doses. Subjective effects and heart rate are currently the most reliable biomarkers to study the effect of cannabis. Cannabis affects most CNS domains, but too many different CNS tests are used to quantify the drug—response relationships reliably. Test standardization, particularly in motor and memory domains, may reveal additional biomarkers.
Introduction
The discovery of cannabinoid receptors and endocannabinoids has pointed to the physiological and possibly pathophysiological relevance of cannabinoids in humans. So far, two cannabis receptors (CB1 and CB2) have been identified with certainty. The CB1 receptors are predominantly situated in the brain and the CB2 receptors are predominantly present in the spleen and in haematopoietic cells. CB2 receptors seem also to be widely distributed in the brain, but their function is still not clear. The discovery of the endocannabinoid system has stimulated the development of synthetic cannabinoids, which have been used in preclinical research to investigate further the role of the endocannabinoid system in health and disease. However, the clinical development of cannabinoids as medicines is only just beginning. At present, most research in humans has been performed with tetrahydrocannabinol (THC), a CB1/CB2 agonist and the main psychoactive ingredient of cannabis. THC is a highly lipophilic compound that is rapidly absorbed and distributed to highly vascularized tissues, including the brain, where it causes its pleasurable effects. Smoking is the preferred route of cannabis use, with high bioavailability of the THC content that is not lost by combustion or vaporization. In humans, plasma THC concentration profiles are similar after smoking or intravenous administration, with prompt onset and steady decline. In contrast, slow absorption and limited and variable bioavailability are observed after oral administration.
Although a large number of studies have been performed with cannabis and THC in healthy volunteers, it is not clear which biomarkers are useful in early cannabinoid drug development, and how cannabis affects different central nervous system (CNS) functions. The effects of THC/cannabis can provide important tools during the early development of cannabinoid agonists and antagonists, if the effects can be qualified as valid biomarkers. A biomarker is a characteristic that is measured and evaluated as an indicator of normal biological processes, pathogenic processes, or pharmacological responses to a therapeutic intervention [1]. A validated biomarker in early Phase I studies that provides useful information on the potential therapeutic effects of an investigational drug could support the drug development programme of the new compound. In general, a useful biomarker for activity of a drug class should meet the following criteria: (i) a clear, consistent response across studies (from different research groups) and drugs from the same class; (ii) a clear response of the biomarker to therapeutic doses; (iii) a dose (concentration)—response relationship; and (iv) a plausible relationship between the biomarker, the pharmacology of the drug class and/or the pathogenesis of the therapeutic area. Previously, these criteria have been used to evaluate the literature for the usefulness of biomarkers for the effects in healthy volunteers of antipsychotic drugs [2], benzodiazepines [3], selective serotonin reuptake inhibitors [4] and 3,4-methylene-dioxy-methamphetamine (ecstasy) [5]. In the current review, the effects of cannabis and THC in healthy volunteers were systematically evaluated using the same methodology.
Methods
Structured literature evaluation
A literature search was performed up to 15 November 2007 using MedLine, Web of Science and Embase. The following keywords were used: marijuana, marihuana, cannabis, THC, tetrahydrocannabinol and delta(9)-tetrahydrocannabinol. The searches were limited to healthy adults and papers in English. The resulting studies were subject to several selection criteria.
This review aimed to assess the usefulness of direct CNS biomarkers and heart rate for studies of cannabinoids in healthy volunteers. Reviews, studies in experimental animals or patients, and studies of interactions of cannabis use with personality features, behavioural characteristics, metabolic variations, other drugs, pain models or environmental factors (including secondary or subgroup analyses) were excluded from this review.
Studies with <10 subjects were not included. Study participants were divided into non-users and users. No distinction was made according to the levels of previous or current usage, which ranged from occasional to chronic frequent use and was rarely documented in detail. Frequent and infrequent users were grouped as users. The review was restricted to the effects of acute cannabis exposure. Hence, abstinence effects, 'morning after effects' (including sleep effects after dosing on the preceding day), long-term effects in chronic users or effects of repeated dosing were not incorporated in this review.
The study characteristics and each individual test result of all articles that complied with the criteria were put into a database (Microsoft Excel) (Appendix S1). The following items were recorded: number of subjects, sex (male; female), age, past cannabis use (users; non-users; unknown), abstinence period (yes; no; unknown), blinding (double blind; single blind; open; unknown), design (crossover; partial crossover; parallel; unknown), drug name (cannabis, including hashish and marijuana); THC(/dronabinol)), dose, route of administration (oral; intrapulmonary; intravenous; unknown), THC equivalence (<7 mg; 7—18 mg; >18 mg), test name, test effect, test cluster and functional domain. Most studies used different tests on different doses of cannabis, which were all regarded as independent measures of the cannabis effect. Thus, the total number of evaluated tests (cases) was a product of the number of articles, drugs, doses and tests (including secondary outcomes).
Individual test results
Based on previous reviews, it was anticipated that in most cases no consistent quantitative results could be recorded for individual tests, because of the large diversity of methods, parameters and treatments. Therefore, the ability of a test to show a statistically significant difference from placebo or baseline was scored as + (improvement/increase), = (no significant effect) or — (impairment/decrease). Subjective assessments with a desirable effect (e.g. increase of a high scale) were scored as an improvement/increase, and unwanted effects (e.g. increase of sedation) as an impairment/decrease.
Different parameters of a single test were always grouped together if they provided information on the same cluster. Many single tasks provided different outcome parameters, which sometimes showed apparently opposite responses. If these opposite responses were part of the same cluster, two items were scored for the same test: e.g. one (+) and one (−). More frequently, one of the parameters that improved was from a different functional cluster than the one that deteriorated. In these cases, both items were scored separately on different clusters. In the table an asterisk was added to the item that was considered a secondary parameter from a test of a different primary function.
Some studies explicitly reported the use of several different tests in the methods section, without presentation of the results for any apparent reason. In these cases, it was assumed that these tests had not shown any significant effects. In some studies with different drug doses, overall significances were reported for drug effects, without (post hoc) quantifications of the statistical significance levels for each individual dose. In these cases, efforts were made to estimate the individual dose effects from graphs or tables provided in the article. If this was impossible, only the effect of the highest dose was assumed to be significant (in case of overall statistical significance) and lower doses were considered nonsignificant.
Grouping of individual test results
Because of apparent lack of standardization between the studies even for the same tests, a structured procedure was adopted as described previously [2—5] in order to obtain an overview. This approach allowed the preservation of individual study data in early stages, followed by a progressive condensation of results into logical test clusters and functional domains. For the subjective assessments, most subjective scales can, for example, be grouped under scores of feeling high, craving, alertness, general drug effect, etc. A compendium of neuropsychological tests from Strauss et al. [6] was primarily consulted to group functional tests into clusters of related tests or test variants. If necessary, the compendium of Lezak was consulted [7]. Sometimes, these compendia did not mention the specific test. In these cases, the author's classification was followed or, if necessary, the test was looked up in other literature and classified by consensus. A single, more complicated test can sometimes measure several aspects (e.g. of memory, executive function, subjective effects, etc.) and can therefore provide information on different clusters. Examples are Babcock Story Recall Test, Buschke Selective Reminding Test, Digit Recall Test, Ratings of Narrative Quality.
Tests and clusters were grouped further into domains that represent higher aggregates of integration of subjective, neuropsychological, neuroendocrine, neurophysiological or autonomic functions. For each test (cluster), the compendia and other literature were used to determine which function was principally assessed by the test. Neuropsychological domains consisted of executive functions, memory, attention, motor functions, language and perception. Some tests provided different parameters with information on more than one functional domain. The results of the effects of a single test on different domains were scored separately, and the secondary effects were marked.
Results from tests that were used only occasionally or tests used only by a single research group could not be generalized. Therefore, these were not analysed individually, but grouped with other comparable tests. This step started with the grouping of tests that could be regarded as variants from a basic form (e.g. individual scores that are also part of more comprehensive tools such as Profiles of Mood States, Addiction Research Center Inventory (ARCI) or Bond & Lader Visual Analogue Scales (VAS) [8]). Subscales of such inventories were grouped if they fell in the same cluster. Within such clusters, all scales showing a significant effect were grouped, whereas all scales showing no effect were grouped separately. In this way, scales within the same cluster that showed mixed results were scored equivocally. Comprehensive scoring instruments like Waskow's Drug Effect Questionnaire can be subdivided into different subjective clusters (e.g. drug effect, high effect, etc.), but these subscales were not always reported separately. In these cases, the results were presented as part of the overall scale drug effect cluster. In a few articles, a couple of composite scores of different CNS functions were presented that could not be grouped according to the clusters or domains used in this review. These tests were not included in the analysis.
All subdivisions of the tests and effect scores were initially performed by two of the authors (E.M. and A.E.I.) and subsequently checked and discussed by the other authors (L.Z., A.E.I. and J.M.A.v.G.).
Dose—effect relationships
The chance that a test will detect a difference from placebo is expected to increase with dose. For each test that was used ≥10 times and for all clusters, potential dose—response relationships were determined. Dose-related increases or decreases of the average percentages of tests or clusters were reported without formal statistical analyses. Since the review yielded no immediately quantitative test effects, dose relationships were represented by the proportions of statistically significant results for a given test or cluster. Similarly, since THC doses were not reported uniformly, cannabis/THC dosages were pooled into 'lower', 'medium' and 'higher' dosages. The 'lower' dose was chosen to be a dose <7 mg (roughly corresponding to half a cigarette), the 'medium' dose lay between 7 and 18 mg (approximately corresponding to one to one-and-a-half cigarette), and the 'higher' doses were all dosages >18 mg (comparable with one-and-a-half cigarettes or more) [9—11].
Cigarette smoking was the predominant form of administration. In many articles the exact THC content of a cigarette was mentioned. However, some articles mentioned the THC contents in percentage without the weight of the cigarette. In these cases a cigarette weight of 700 mg was assumed since most cigarettes weigh 500—900 mg. In other articles the number of puffs taken was documented. In these instances the dose was calculated as eight puffs corresponding to one marijuana cigarette [11]. Some studies provided weight-adjusted doses, without specifying the (average) body weight. In these cases, the 70 kg adult general population body weight was used to calculate the average administered dose.
To be able to compare the test results obtained for oral and intravenous administration with the results obtained for smoking, all doses were normalized to smoking. After smoking, roughly 50% of the THC contents of a cigarette is delivered into the smoke [12] and another 50% of the inhaled smoke is exhaled again [13]. In practice, smoking a cannabis cigarette of 10 mg causes 50% loss due to heating, which leaves 5 mg. Next, half the inhaled 5 mg is exhaled again. Ultimately, 2.5 mg or 25% of the 10-mg cigarette is delivered to the systemic circulation. Bioavailability after oral administration was assumed to be around 10% [14, 15]. Consequently, 25 mg THC would have to be ingested to get the same 2.5 mg systemic exposure as after smoking a 10-mg cigarette. This means that oral doses were divided by 2.5 to calculate the equivalent intrapulmonary THC doses. The THC plasma concentrations after smoking a 19-mg marijuana cigarette are equal to intravenous administration of 5 mg THC [16]. Therefore, all intravenous dosages were multiplied by four for dose normalization. In this way, all routes of administration could be compared.
Results
Study design
The literature search yielded 165 different studies on cannabis and THC that met all criteria, published between 1966 and 15 November 2007. The numbers of participants ranged from 10 to 161, where 115 studies (70%) included 10—20 subjects and six studies included >75 subjects (9%). Ages ranged from 18 to 59 years, but the vast majority were young adults aged 18—35 years. In 57% of studies only healthy men were included, and 2% of studies included only women. Thirty-three percent of studies included men and women, whereas the sex of the subjects was not mentioned in 8%.
Most studies (80%) included subjects that were familiar with the effects of cannabis. In contrast, non-users were included in only 3%. Eleven percent of studies reported inclusion of both cannabis users and non-users. Previous cannabis use was not mentioned in 6% of studies. A small majority of studies (53%) described an abstinence period or the use of a THC drug screen. Four percent of studies reported the lack of an abstinence period, whereas 44% did not mention this topic.
Fifty-seven percent of the reviewed studies had a double-blinded design; 26% were single-blinded; 7% had an open design and for 10% the blinding was unknown. In addition, a small majority of the studies had a crossover design (60%), 3% had a partial crossover design, 33% had a parallel design and 4% of studies did not mention the study design.
Study drug and dosing
Cannabis is also known as marijuana, and dronabinol is an analogue of THC, the predominant psychoactive component of cannabis. Cannabis was used in 63% of studies and THC in 34%. Intrapulmonary administration was the preferred route of administration in 71% of studies. Oral administration of the drug was mentioned in 25% of studies and intravenous administration was used in only 3%. Three percent of studies did not describe which form of cannabis was used and 1% did not mention the route of administration. In these cases it could be inferred from the doses and the design that cannabis was smoked.
Tests, clusters and domains
In total, 318 different tests were used. Only a minority of tests were used frequently enough to allow individual analysis. The majority of tests (196 tests, 61.6%) were used only once, and only heart rate (0.3%) was used >50 times (in 92 articles) (Table 1). VAS scale high/stoned was studied in 30 articles, whereas the subjective effect rating scale high/stoned/euphoria was assessed in 28 articles (Table 1). Taken together, the subjective high phenomenon was measured in >50 (35.2%) articles as well (Table 1). The Digit Symbol Substitution Test (DSST) or variants such as the Symbol Digit Substitution Test were the most frequently used neuropsychological tests (22 times) (Table 1). The ARCI was used in 18 articles (Table 1).
Although many different tests and test variants were used to evaluate the effects of cannabis, most actually measured a limited number of core features. Therefore, tests were grouped further into clusters and subsequently in domains. Table 2a—d is a progressive condensation of all reported tests, from test to cluster to domain, and includes the overall calculated significant drug effects on each cluster (impairment/decrease, no change or improvement/increase).
Table 2a—d shows that most drug-sensitive clusters demonstrate consistent functional impairment, and some an enhancement (heart rate, scale high). A few clusters show both impairments and improvements (e.g. time estimation, EEG alpha and evoked potential measurements, and scales for calmness, craving, mood and performance). Only a few frequently-used (>10 times) test clusters showed significant responses to THC/cannabis in >80% of studies, notably heart rate (n = 85/92), scale high (n = 67/70) and scale psychotomimetic (n = 14/18). Most other clusters reported significant drug effects in only about 30—50% of studies (Table 2a—d). All tests that were used five times or more showed a significant THC effect in at least one case, except for EEG delta, which never responded in any study.
Dose—response relationships
Tests and clusters that were used in >10 articles were inspected for potential dose—response relationships (Table 3). Heart rate showed a statistically significant increase in 78% of measurements in the THC equivalence dose group <7 mg, which increased to 99 and 98% after the use of 7—18 and >18 mg THC, respectively. The subjective high feeling included many different scoring methods, varying from observer rating scales to individual VAS scores, either in isolation or as part of multidimensional inventories (Table 2d). Despite this variability, the cluster scale high showed very consistent effects for all dose groups. The lowest dose group of <7 mg THC already showed a response of 94%, and the middle (7—18 mg) and highest dose group (>18 mg) scored close to 100%. The related subjective cluster scale psychotomimetic also showed consistent deterioration (i.e. an increase in psychotomimetic scores) with THC/cannabis of 76—83% without a clear dose—response relationship. A small increase with dose (from 56 to 78%) was observed for the cluster scale drug effect.
The relationship between memory and doses of THC/cannabis was more complex. Impairment increased with dose for auditory/verbal delayed recall (from 23% with the lowest doses to 78% with the highest dose range), but the effects were less clear for immediate recall (Table 3). Auditory/verbal delayed recognition also deteriorated with dose (from 17 to 50%), but this was assessed in only 11 studies. Working memory impairment, on the other hand, seemed to decrease with dose, from 52% impairments in the lowest dose group to 9% in the highest (Table 3). Other clusters that also appeared to show an inverse dose—response association were the DSST-like cluster, focused selective attention and tests of motor and visuomotor control (Table 3). The proportion of significant effects of THC/cannabis within the cluster scale aggression increased slightly with dose (from 20 to 40%). No clear dose—response relationships were observed for inhibition, reasoning/association or reaction time, or for most subjective scales (Table 3). For studies with different doses, we scored significance for the highest dose only, if significance was merely reported for the overall group effect, without allowing an estimate of the individual dose effects from graphs or tables. Although in such cases we could have artificially induced a dose—response relationship, this was observed in only 3% of all test scores.
Discussion
This review aimed to evaluate systematically the usefulness of tests for the effects of cannabis and THC in healthy volunteers. It should be noted that cannabis cigarettes contain a mixture of psychoactive compounds, which in combination may contribute differently to the psychological and physical effects of cannabis compared with single THC administration. However, since THC is the main psychoactive ingredient of cannabis, in this review studies with cannabis and THC were used. The results were comparable to those of similar reviews of biomarkers of different CNS-active drugs in healthy volunteers [2—5]. A striking number of 318 different tests or test variants were described, of which 61.6% were used only once. Grouping of tests in clusters and domains was required to evaluate the general usefulness of functional measurements, but this inevitably led to a loss of information. Even clustering tests with the same name and/or description could have bypassed differences among research groups or tests variants. In addition, this review investigated biomarkers for the effects of cannabis and THC in healthy volunteers, i.e. often with relatively small subject numbers; 70% of the studies had ≤20 participants. It is possible that some tests will be useful biomarkers in patient studies or studies with large numbers of subjects, or if their value is demonstrated in more studies. The observed variability in test results may have been enhanced by differences in prior cannabis use (non-users, occasional and frequent users). In this review these differences were not taken into consideration, although most participants were occasional cannabis users and only 3% had not used cannabis before. Chronic and occasional cannabis users show similar drug effects, although chronic users generally require higher doses and thus seem to be less sensitive [17]. A small majority of articles mentioned an abstinence period, but it is likely that this was also included in many other studies, without being reported. The neglect of prior use intensity or abstinence duration may have confounded the detection of dose—response relationships, which was only roughly possible in any case because of the many different doses and administration forms.
Useful cannabinoid biomarkers
The effects of cannabis were observed on all clusters and all domains and in almost all individual tests, which might be due to the wide distribution of cannabinoid receptors in the brain [18]. An increase in heart rate was the most consistent result (Tables 1 and and2a),2a), and almost all studies with heart rate measurements showed statistically significant effects. This was expected, since heart rate shows a sharp increase and rapid decline after intrapulmonary THC administration that is clearly concentration related and already considerable at low THC levels [9, 19]. Feeling high has previously also been shown to be closely related to THC plasma concentrations [19]. The high phenomenon was measured in many different ways, but despite this variability almost all studies showed statistically significant subjective drug effects. The predicted and highly consistent effects of THC/cannabis on the most clearly concentration-related effects (heart rate and feeling high [9, 19]) in this review also support the methodological approach that was adopted, to integrate the widely variable study designs, drug forms and doses, and tests reported in the literature. Feeling high seems to be the most sensitive CNS biomarker for the effects of cannabis, irrespective of how it is measured. The scales psychotomimetic and drug effect are not quite as sensitive, but they address subjective changes that are less specific for THC/cannabis. This is clearly illustrated by the only negative scores on the drug effect cluster, which are all due to the reductions on the benzedrine scale (BG scale) of the ARCI, which is considered as a measure of subjective stimulation. Most other clusters show low to medium sensitivity for the effects of THC/cannabis, with significant drug effects in roughly 30—60% of cases (Table 2a—d). These findings are similar to other drug classes, which show very comparable sensitivities of neurophysiological, neuropsychological and subjective tests of 30—60% with benzodiazepines [3] and neuroleptics [2]. In these reviews, saccadic peak velocity was highly sensitive to benzodiazepines in 100% [3], and prolactin release to neuroleptics in 96% [2]. These parameters were not particularly responsive to THC/cannabis in the current review, where heart rate and subjective high feeling scored 92—96%. This illustrates the differential effect profiles of different pharmacological groups, even among drug classes that are generally considered to be 'CNS depressant'. Such variability should be considered when methods are selected to study the CNS effects of neuropsychiatric agents. Furthermore, tests that showed a medium chance on an effect (approximately 60%) should also be critically considered.
Dose—response relationships
A useful biomarker should show a dose—response relationship starting at a low therapeutic dose. Consideration of dose—response relationships is also essential to compare the effects of THC (at different doses and administration routes) across studies. In this review, doses could be grouped only roughly, and effects could be scored only as either statistically significant or not. Moreover, hardly any test was measured frequently and quantified consistently enough for a meaningful analysis of dose—response associations. Perhaps due to these limitations, dose—response relationships were found for only a few clusters (Table 3). THC doses were categorized in a low (<7 mg, half a cannabis cigarette), medium (7—18 mg, one to one-and-a-half cigarette) and high (>18 mg, one-and-a-half cigarettes or more) dose. This pragmatic division was not based on well-established relations between doses, plasma concentrations and CNS effects. Nonetheless, it led to roughly similar numbers of tests at the three different dose levels (623—852 in each dose group), and thus reflects practical dose selection in the literature. However, this practice could be based on the habit of subjects to smoke enough cannabis to elicit a desirable subjective state that does not cause unpleasant effects. It is not illogical to assume that this is reflected in the dose of one cigarette, and that a 'standard dose' is near the maximum tolerated dose for most subjects (still pleasant and devoid of intolerable adverse effects). In this review, lower doses (<7 mg) were used in only about 30% of cases, and even this dose range caused subjective high feeling in 94% of cases. In a recent pharmacokinetic—pharmacodynamic (PK—PD) study, heart rate, VAS high and alertness, and postural stability were already sensitive to levels as low as 2 mg of THC administration by inhalation, and PK—PD effect relationships showed that near-maximum effects are reached with THC doses corresponding to roughly 10 mg of cannabis [9, 19], which corresponds to one or two cannabis cigarettes.
The memory effects of cannabis showed some dose—response relationships, but this differed for the various types of memory tests. Impairments increased with dose for auditory/verbal delayed recall and to a lesser extent for immediate recall and auditory/verbal delayed recognition (Table 3). Working memory (which included immediate recognition) on the other hand seemed to improve (i.e. normalize) with dose, with 52% impairments in the lowest dose group to 9% in the highest (Table 3). The clusters of focused selective attention and of motor and visuomotor control also appeared to show an inverse dose—response association (Table 3). All these functions are highly influenced by attention and concentration [6]. Decreases in subjective alertness were noted in 43% with the lowest doses and 35% with the highest. This may have been accompanied by some agitation. At the same time, dose-related increases in (subjective) aggression (which increased with dose from 20 to 40%) and anxiety (from 11 to 33%) were observed. All this suggests that lower doses of THC/cannabis generally cause pleasant effects of relaxation and reduced attention, whereas with high doses CNS depression is partly overcome by more stimulatory effects.
These results suggest that most CNS measurements are sensitive to the effects of THC/cannabis. Some parameters were very sensitive to THC/cannabis. For such biomarkers, most doses studied in the literature may have been too high to show clear dose—response relationships. The proportion of significant effects did not increase much at the highest dose range compared with the medium range, although this does not exclude a dose-related increase in the size of the effects per se. Except for some subjective effects (feeling high, psychotomimetic feelings, drug effects) and heart rate, most tests did not show consistent effects even at quite high THC/cannabis doses. It could be argued that less sensitive CNS tests would have shown effects at even higher doses, well above 18 mg. For most subjects, particularly for inexperienced cannabis users, two or more cannabis cigarettes would cause an overdose, with questionable pharmacological and functional selectivity and an unacceptable adverse event profile. Hence, sensitive biomarkers at low to medium doses are needed to characterize the concentration—effect relationships of CB1/CB2 agonists. At present, the literature indicates that the choice is limited to heart rate and subjective effects.
Summary
Biomarkers are useful tools to study drug effects, since they can provide information on the potential pharmacological effects of the investigational drug in early-phase drug development. However, the number of tests and test variants that is used in studies of THC and cannabis seems excessively large. This abundance thwarts good assessment of the physiological, neuropsychological and subjective effects of this drug class, and there is dire need for test standardization in these areas. In general, the doses studied in the literature reflect the patterns of recreational use, and are often too high to determine accurately pharmacological dose—response relationships. Cannabis/THC has an effect on a wide range of CNS domains. At lower doses, THC/cannabis seems to be relaxant and to reduce attention, which is accompanied by impaired performance on other CNS tests that require motivation and active participation. At high doses, the drug seems to be more stimulatory. Subjective effects are the most reliable biomarkers to study the effects of cannabis, in addition to heart rate increases that reflect peripheral cannabinoid activation. This review indicates that these parameters are useful biomarkers that can be used in future studies to investigate the effects of THC/cannabis and other cannabinoid agonists on the CNS.
Source, Graphs and Figures: Biomarkers for the effects of cannabis and THC in healthy volunteers
An increasing number of novel therapeutic agents are targeted at cannabinoid receptors. Drug development programmes of new cannabinoid drugs may be facilitated by the identification of useful biomarkers. This systemic literature review aims to assess the usefulness of direct biomarkers for the effects of cannabis and tetrahydrocannabinol (THC) in healthy volunteers. One hundred and sixty-five useful articles were found that investigated the acute effects of cannabis or THC on the central nervous system (CNS) and heart rate in healthy volunteers. Three hundred and eighteen tests (or test variants) were grouped in test clusters and functional domains, to allow their evaluation as a useful biomarker and to study their dose—response effects. Cannabis/THC affected a wide range of CNS domains. In addition to heart rate, subjective effects were the most reliable biomarkers, showing significant responses to cannabis in almost all studies. Some CNS domains showed indications of depression at lower and stimulation at higher doses. Subjective effects and heart rate are currently the most reliable biomarkers to study the effect of cannabis. Cannabis affects most CNS domains, but too many different CNS tests are used to quantify the drug—response relationships reliably. Test standardization, particularly in motor and memory domains, may reveal additional biomarkers.
Introduction
The discovery of cannabinoid receptors and endocannabinoids has pointed to the physiological and possibly pathophysiological relevance of cannabinoids in humans. So far, two cannabis receptors (CB1 and CB2) have been identified with certainty. The CB1 receptors are predominantly situated in the brain and the CB2 receptors are predominantly present in the spleen and in haematopoietic cells. CB2 receptors seem also to be widely distributed in the brain, but their function is still not clear. The discovery of the endocannabinoid system has stimulated the development of synthetic cannabinoids, which have been used in preclinical research to investigate further the role of the endocannabinoid system in health and disease. However, the clinical development of cannabinoids as medicines is only just beginning. At present, most research in humans has been performed with tetrahydrocannabinol (THC), a CB1/CB2 agonist and the main psychoactive ingredient of cannabis. THC is a highly lipophilic compound that is rapidly absorbed and distributed to highly vascularized tissues, including the brain, where it causes its pleasurable effects. Smoking is the preferred route of cannabis use, with high bioavailability of the THC content that is not lost by combustion or vaporization. In humans, plasma THC concentration profiles are similar after smoking or intravenous administration, with prompt onset and steady decline. In contrast, slow absorption and limited and variable bioavailability are observed after oral administration.
Although a large number of studies have been performed with cannabis and THC in healthy volunteers, it is not clear which biomarkers are useful in early cannabinoid drug development, and how cannabis affects different central nervous system (CNS) functions. The effects of THC/cannabis can provide important tools during the early development of cannabinoid agonists and antagonists, if the effects can be qualified as valid biomarkers. A biomarker is a characteristic that is measured and evaluated as an indicator of normal biological processes, pathogenic processes, or pharmacological responses to a therapeutic intervention [1]. A validated biomarker in early Phase I studies that provides useful information on the potential therapeutic effects of an investigational drug could support the drug development programme of the new compound. In general, a useful biomarker for activity of a drug class should meet the following criteria: (i) a clear, consistent response across studies (from different research groups) and drugs from the same class; (ii) a clear response of the biomarker to therapeutic doses; (iii) a dose (concentration)—response relationship; and (iv) a plausible relationship between the biomarker, the pharmacology of the drug class and/or the pathogenesis of the therapeutic area. Previously, these criteria have been used to evaluate the literature for the usefulness of biomarkers for the effects in healthy volunteers of antipsychotic drugs [2], benzodiazepines [3], selective serotonin reuptake inhibitors [4] and 3,4-methylene-dioxy-methamphetamine (ecstasy) [5]. In the current review, the effects of cannabis and THC in healthy volunteers were systematically evaluated using the same methodology.
Methods
Structured literature evaluation
A literature search was performed up to 15 November 2007 using MedLine, Web of Science and Embase. The following keywords were used: marijuana, marihuana, cannabis, THC, tetrahydrocannabinol and delta(9)-tetrahydrocannabinol. The searches were limited to healthy adults and papers in English. The resulting studies were subject to several selection criteria.
This review aimed to assess the usefulness of direct CNS biomarkers and heart rate for studies of cannabinoids in healthy volunteers. Reviews, studies in experimental animals or patients, and studies of interactions of cannabis use with personality features, behavioural characteristics, metabolic variations, other drugs, pain models or environmental factors (including secondary or subgroup analyses) were excluded from this review.
Studies with <10 subjects were not included. Study participants were divided into non-users and users. No distinction was made according to the levels of previous or current usage, which ranged from occasional to chronic frequent use and was rarely documented in detail. Frequent and infrequent users were grouped as users. The review was restricted to the effects of acute cannabis exposure. Hence, abstinence effects, 'morning after effects' (including sleep effects after dosing on the preceding day), long-term effects in chronic users or effects of repeated dosing were not incorporated in this review.
The study characteristics and each individual test result of all articles that complied with the criteria were put into a database (Microsoft Excel) (Appendix S1). The following items were recorded: number of subjects, sex (male; female), age, past cannabis use (users; non-users; unknown), abstinence period (yes; no; unknown), blinding (double blind; single blind; open; unknown), design (crossover; partial crossover; parallel; unknown), drug name (cannabis, including hashish and marijuana); THC(/dronabinol)), dose, route of administration (oral; intrapulmonary; intravenous; unknown), THC equivalence (<7 mg; 7—18 mg; >18 mg), test name, test effect, test cluster and functional domain. Most studies used different tests on different doses of cannabis, which were all regarded as independent measures of the cannabis effect. Thus, the total number of evaluated tests (cases) was a product of the number of articles, drugs, doses and tests (including secondary outcomes).
Individual test results
Based on previous reviews, it was anticipated that in most cases no consistent quantitative results could be recorded for individual tests, because of the large diversity of methods, parameters and treatments. Therefore, the ability of a test to show a statistically significant difference from placebo or baseline was scored as + (improvement/increase), = (no significant effect) or — (impairment/decrease). Subjective assessments with a desirable effect (e.g. increase of a high scale) were scored as an improvement/increase, and unwanted effects (e.g. increase of sedation) as an impairment/decrease.
Different parameters of a single test were always grouped together if they provided information on the same cluster. Many single tasks provided different outcome parameters, which sometimes showed apparently opposite responses. If these opposite responses were part of the same cluster, two items were scored for the same test: e.g. one (+) and one (−). More frequently, one of the parameters that improved was from a different functional cluster than the one that deteriorated. In these cases, both items were scored separately on different clusters. In the table an asterisk was added to the item that was considered a secondary parameter from a test of a different primary function.
Some studies explicitly reported the use of several different tests in the methods section, without presentation of the results for any apparent reason. In these cases, it was assumed that these tests had not shown any significant effects. In some studies with different drug doses, overall significances were reported for drug effects, without (post hoc) quantifications of the statistical significance levels for each individual dose. In these cases, efforts were made to estimate the individual dose effects from graphs or tables provided in the article. If this was impossible, only the effect of the highest dose was assumed to be significant (in case of overall statistical significance) and lower doses were considered nonsignificant.
Grouping of individual test results
Because of apparent lack of standardization between the studies even for the same tests, a structured procedure was adopted as described previously [2—5] in order to obtain an overview. This approach allowed the preservation of individual study data in early stages, followed by a progressive condensation of results into logical test clusters and functional domains. For the subjective assessments, most subjective scales can, for example, be grouped under scores of feeling high, craving, alertness, general drug effect, etc. A compendium of neuropsychological tests from Strauss et al. [6] was primarily consulted to group functional tests into clusters of related tests or test variants. If necessary, the compendium of Lezak was consulted [7]. Sometimes, these compendia did not mention the specific test. In these cases, the author's classification was followed or, if necessary, the test was looked up in other literature and classified by consensus. A single, more complicated test can sometimes measure several aspects (e.g. of memory, executive function, subjective effects, etc.) and can therefore provide information on different clusters. Examples are Babcock Story Recall Test, Buschke Selective Reminding Test, Digit Recall Test, Ratings of Narrative Quality.
Tests and clusters were grouped further into domains that represent higher aggregates of integration of subjective, neuropsychological, neuroendocrine, neurophysiological or autonomic functions. For each test (cluster), the compendia and other literature were used to determine which function was principally assessed by the test. Neuropsychological domains consisted of executive functions, memory, attention, motor functions, language and perception. Some tests provided different parameters with information on more than one functional domain. The results of the effects of a single test on different domains were scored separately, and the secondary effects were marked.
Results from tests that were used only occasionally or tests used only by a single research group could not be generalized. Therefore, these were not analysed individually, but grouped with other comparable tests. This step started with the grouping of tests that could be regarded as variants from a basic form (e.g. individual scores that are also part of more comprehensive tools such as Profiles of Mood States, Addiction Research Center Inventory (ARCI) or Bond & Lader Visual Analogue Scales (VAS) [8]). Subscales of such inventories were grouped if they fell in the same cluster. Within such clusters, all scales showing a significant effect were grouped, whereas all scales showing no effect were grouped separately. In this way, scales within the same cluster that showed mixed results were scored equivocally. Comprehensive scoring instruments like Waskow's Drug Effect Questionnaire can be subdivided into different subjective clusters (e.g. drug effect, high effect, etc.), but these subscales were not always reported separately. In these cases, the results were presented as part of the overall scale drug effect cluster. In a few articles, a couple of composite scores of different CNS functions were presented that could not be grouped according to the clusters or domains used in this review. These tests were not included in the analysis.
All subdivisions of the tests and effect scores were initially performed by two of the authors (E.M. and A.E.I.) and subsequently checked and discussed by the other authors (L.Z., A.E.I. and J.M.A.v.G.).
Dose—effect relationships
The chance that a test will detect a difference from placebo is expected to increase with dose. For each test that was used ≥10 times and for all clusters, potential dose—response relationships were determined. Dose-related increases or decreases of the average percentages of tests or clusters were reported without formal statistical analyses. Since the review yielded no immediately quantitative test effects, dose relationships were represented by the proportions of statistically significant results for a given test or cluster. Similarly, since THC doses were not reported uniformly, cannabis/THC dosages were pooled into 'lower', 'medium' and 'higher' dosages. The 'lower' dose was chosen to be a dose <7 mg (roughly corresponding to half a cigarette), the 'medium' dose lay between 7 and 18 mg (approximately corresponding to one to one-and-a-half cigarette), and the 'higher' doses were all dosages >18 mg (comparable with one-and-a-half cigarettes or more) [9—11].
Cigarette smoking was the predominant form of administration. In many articles the exact THC content of a cigarette was mentioned. However, some articles mentioned the THC contents in percentage without the weight of the cigarette. In these cases a cigarette weight of 700 mg was assumed since most cigarettes weigh 500—900 mg. In other articles the number of puffs taken was documented. In these instances the dose was calculated as eight puffs corresponding to one marijuana cigarette [11]. Some studies provided weight-adjusted doses, without specifying the (average) body weight. In these cases, the 70 kg adult general population body weight was used to calculate the average administered dose.
To be able to compare the test results obtained for oral and intravenous administration with the results obtained for smoking, all doses were normalized to smoking. After smoking, roughly 50% of the THC contents of a cigarette is delivered into the smoke [12] and another 50% of the inhaled smoke is exhaled again [13]. In practice, smoking a cannabis cigarette of 10 mg causes 50% loss due to heating, which leaves 5 mg. Next, half the inhaled 5 mg is exhaled again. Ultimately, 2.5 mg or 25% of the 10-mg cigarette is delivered to the systemic circulation. Bioavailability after oral administration was assumed to be around 10% [14, 15]. Consequently, 25 mg THC would have to be ingested to get the same 2.5 mg systemic exposure as after smoking a 10-mg cigarette. This means that oral doses were divided by 2.5 to calculate the equivalent intrapulmonary THC doses. The THC plasma concentrations after smoking a 19-mg marijuana cigarette are equal to intravenous administration of 5 mg THC [16]. Therefore, all intravenous dosages were multiplied by four for dose normalization. In this way, all routes of administration could be compared.
Results
Study design
The literature search yielded 165 different studies on cannabis and THC that met all criteria, published between 1966 and 15 November 2007. The numbers of participants ranged from 10 to 161, where 115 studies (70%) included 10—20 subjects and six studies included >75 subjects (9%). Ages ranged from 18 to 59 years, but the vast majority were young adults aged 18—35 years. In 57% of studies only healthy men were included, and 2% of studies included only women. Thirty-three percent of studies included men and women, whereas the sex of the subjects was not mentioned in 8%.
Most studies (80%) included subjects that were familiar with the effects of cannabis. In contrast, non-users were included in only 3%. Eleven percent of studies reported inclusion of both cannabis users and non-users. Previous cannabis use was not mentioned in 6% of studies. A small majority of studies (53%) described an abstinence period or the use of a THC drug screen. Four percent of studies reported the lack of an abstinence period, whereas 44% did not mention this topic.
Fifty-seven percent of the reviewed studies had a double-blinded design; 26% were single-blinded; 7% had an open design and for 10% the blinding was unknown. In addition, a small majority of the studies had a crossover design (60%), 3% had a partial crossover design, 33% had a parallel design and 4% of studies did not mention the study design.
Study drug and dosing
Cannabis is also known as marijuana, and dronabinol is an analogue of THC, the predominant psychoactive component of cannabis. Cannabis was used in 63% of studies and THC in 34%. Intrapulmonary administration was the preferred route of administration in 71% of studies. Oral administration of the drug was mentioned in 25% of studies and intravenous administration was used in only 3%. Three percent of studies did not describe which form of cannabis was used and 1% did not mention the route of administration. In these cases it could be inferred from the doses and the design that cannabis was smoked.
Tests, clusters and domains
In total, 318 different tests were used. Only a minority of tests were used frequently enough to allow individual analysis. The majority of tests (196 tests, 61.6%) were used only once, and only heart rate (0.3%) was used >50 times (in 92 articles) (Table 1). VAS scale high/stoned was studied in 30 articles, whereas the subjective effect rating scale high/stoned/euphoria was assessed in 28 articles (Table 1). Taken together, the subjective high phenomenon was measured in >50 (35.2%) articles as well (Table 1). The Digit Symbol Substitution Test (DSST) or variants such as the Symbol Digit Substitution Test were the most frequently used neuropsychological tests (22 times) (Table 1). The ARCI was used in 18 articles (Table 1).
Although many different tests and test variants were used to evaluate the effects of cannabis, most actually measured a limited number of core features. Therefore, tests were grouped further into clusters and subsequently in domains. Table 2a—d is a progressive condensation of all reported tests, from test to cluster to domain, and includes the overall calculated significant drug effects on each cluster (impairment/decrease, no change or improvement/increase).
Table 2a—d shows that most drug-sensitive clusters demonstrate consistent functional impairment, and some an enhancement (heart rate, scale high). A few clusters show both impairments and improvements (e.g. time estimation, EEG alpha and evoked potential measurements, and scales for calmness, craving, mood and performance). Only a few frequently-used (>10 times) test clusters showed significant responses to THC/cannabis in >80% of studies, notably heart rate (n = 85/92), scale high (n = 67/70) and scale psychotomimetic (n = 14/18). Most other clusters reported significant drug effects in only about 30—50% of studies (Table 2a—d). All tests that were used five times or more showed a significant THC effect in at least one case, except for EEG delta, which never responded in any study.
Dose—response relationships
Tests and clusters that were used in >10 articles were inspected for potential dose—response relationships (Table 3). Heart rate showed a statistically significant increase in 78% of measurements in the THC equivalence dose group <7 mg, which increased to 99 and 98% after the use of 7—18 and >18 mg THC, respectively. The subjective high feeling included many different scoring methods, varying from observer rating scales to individual VAS scores, either in isolation or as part of multidimensional inventories (Table 2d). Despite this variability, the cluster scale high showed very consistent effects for all dose groups. The lowest dose group of <7 mg THC already showed a response of 94%, and the middle (7—18 mg) and highest dose group (>18 mg) scored close to 100%. The related subjective cluster scale psychotomimetic also showed consistent deterioration (i.e. an increase in psychotomimetic scores) with THC/cannabis of 76—83% without a clear dose—response relationship. A small increase with dose (from 56 to 78%) was observed for the cluster scale drug effect.
The relationship between memory and doses of THC/cannabis was more complex. Impairment increased with dose for auditory/verbal delayed recall (from 23% with the lowest doses to 78% with the highest dose range), but the effects were less clear for immediate recall (Table 3). Auditory/verbal delayed recognition also deteriorated with dose (from 17 to 50%), but this was assessed in only 11 studies. Working memory impairment, on the other hand, seemed to decrease with dose, from 52% impairments in the lowest dose group to 9% in the highest (Table 3). Other clusters that also appeared to show an inverse dose—response association were the DSST-like cluster, focused selective attention and tests of motor and visuomotor control (Table 3). The proportion of significant effects of THC/cannabis within the cluster scale aggression increased slightly with dose (from 20 to 40%). No clear dose—response relationships were observed for inhibition, reasoning/association or reaction time, or for most subjective scales (Table 3). For studies with different doses, we scored significance for the highest dose only, if significance was merely reported for the overall group effect, without allowing an estimate of the individual dose effects from graphs or tables. Although in such cases we could have artificially induced a dose—response relationship, this was observed in only 3% of all test scores.
Discussion
This review aimed to evaluate systematically the usefulness of tests for the effects of cannabis and THC in healthy volunteers. It should be noted that cannabis cigarettes contain a mixture of psychoactive compounds, which in combination may contribute differently to the psychological and physical effects of cannabis compared with single THC administration. However, since THC is the main psychoactive ingredient of cannabis, in this review studies with cannabis and THC were used. The results were comparable to those of similar reviews of biomarkers of different CNS-active drugs in healthy volunteers [2—5]. A striking number of 318 different tests or test variants were described, of which 61.6% were used only once. Grouping of tests in clusters and domains was required to evaluate the general usefulness of functional measurements, but this inevitably led to a loss of information. Even clustering tests with the same name and/or description could have bypassed differences among research groups or tests variants. In addition, this review investigated biomarkers for the effects of cannabis and THC in healthy volunteers, i.e. often with relatively small subject numbers; 70% of the studies had ≤20 participants. It is possible that some tests will be useful biomarkers in patient studies or studies with large numbers of subjects, or if their value is demonstrated in more studies. The observed variability in test results may have been enhanced by differences in prior cannabis use (non-users, occasional and frequent users). In this review these differences were not taken into consideration, although most participants were occasional cannabis users and only 3% had not used cannabis before. Chronic and occasional cannabis users show similar drug effects, although chronic users generally require higher doses and thus seem to be less sensitive [17]. A small majority of articles mentioned an abstinence period, but it is likely that this was also included in many other studies, without being reported. The neglect of prior use intensity or abstinence duration may have confounded the detection of dose—response relationships, which was only roughly possible in any case because of the many different doses and administration forms.
Useful cannabinoid biomarkers
The effects of cannabis were observed on all clusters and all domains and in almost all individual tests, which might be due to the wide distribution of cannabinoid receptors in the brain [18]. An increase in heart rate was the most consistent result (Tables 1 and and2a),2a), and almost all studies with heart rate measurements showed statistically significant effects. This was expected, since heart rate shows a sharp increase and rapid decline after intrapulmonary THC administration that is clearly concentration related and already considerable at low THC levels [9, 19]. Feeling high has previously also been shown to be closely related to THC plasma concentrations [19]. The high phenomenon was measured in many different ways, but despite this variability almost all studies showed statistically significant subjective drug effects. The predicted and highly consistent effects of THC/cannabis on the most clearly concentration-related effects (heart rate and feeling high [9, 19]) in this review also support the methodological approach that was adopted, to integrate the widely variable study designs, drug forms and doses, and tests reported in the literature. Feeling high seems to be the most sensitive CNS biomarker for the effects of cannabis, irrespective of how it is measured. The scales psychotomimetic and drug effect are not quite as sensitive, but they address subjective changes that are less specific for THC/cannabis. This is clearly illustrated by the only negative scores on the drug effect cluster, which are all due to the reductions on the benzedrine scale (BG scale) of the ARCI, which is considered as a measure of subjective stimulation. Most other clusters show low to medium sensitivity for the effects of THC/cannabis, with significant drug effects in roughly 30—60% of cases (Table 2a—d). These findings are similar to other drug classes, which show very comparable sensitivities of neurophysiological, neuropsychological and subjective tests of 30—60% with benzodiazepines [3] and neuroleptics [2]. In these reviews, saccadic peak velocity was highly sensitive to benzodiazepines in 100% [3], and prolactin release to neuroleptics in 96% [2]. These parameters were not particularly responsive to THC/cannabis in the current review, where heart rate and subjective high feeling scored 92—96%. This illustrates the differential effect profiles of different pharmacological groups, even among drug classes that are generally considered to be 'CNS depressant'. Such variability should be considered when methods are selected to study the CNS effects of neuropsychiatric agents. Furthermore, tests that showed a medium chance on an effect (approximately 60%) should also be critically considered.
Dose—response relationships
A useful biomarker should show a dose—response relationship starting at a low therapeutic dose. Consideration of dose—response relationships is also essential to compare the effects of THC (at different doses and administration routes) across studies. In this review, doses could be grouped only roughly, and effects could be scored only as either statistically significant or not. Moreover, hardly any test was measured frequently and quantified consistently enough for a meaningful analysis of dose—response associations. Perhaps due to these limitations, dose—response relationships were found for only a few clusters (Table 3). THC doses were categorized in a low (<7 mg, half a cannabis cigarette), medium (7—18 mg, one to one-and-a-half cigarette) and high (>18 mg, one-and-a-half cigarettes or more) dose. This pragmatic division was not based on well-established relations between doses, plasma concentrations and CNS effects. Nonetheless, it led to roughly similar numbers of tests at the three different dose levels (623—852 in each dose group), and thus reflects practical dose selection in the literature. However, this practice could be based on the habit of subjects to smoke enough cannabis to elicit a desirable subjective state that does not cause unpleasant effects. It is not illogical to assume that this is reflected in the dose of one cigarette, and that a 'standard dose' is near the maximum tolerated dose for most subjects (still pleasant and devoid of intolerable adverse effects). In this review, lower doses (<7 mg) were used in only about 30% of cases, and even this dose range caused subjective high feeling in 94% of cases. In a recent pharmacokinetic—pharmacodynamic (PK—PD) study, heart rate, VAS high and alertness, and postural stability were already sensitive to levels as low as 2 mg of THC administration by inhalation, and PK—PD effect relationships showed that near-maximum effects are reached with THC doses corresponding to roughly 10 mg of cannabis [9, 19], which corresponds to one or two cannabis cigarettes.
The memory effects of cannabis showed some dose—response relationships, but this differed for the various types of memory tests. Impairments increased with dose for auditory/verbal delayed recall and to a lesser extent for immediate recall and auditory/verbal delayed recognition (Table 3). Working memory (which included immediate recognition) on the other hand seemed to improve (i.e. normalize) with dose, with 52% impairments in the lowest dose group to 9% in the highest (Table 3). The clusters of focused selective attention and of motor and visuomotor control also appeared to show an inverse dose—response association (Table 3). All these functions are highly influenced by attention and concentration [6]. Decreases in subjective alertness were noted in 43% with the lowest doses and 35% with the highest. This may have been accompanied by some agitation. At the same time, dose-related increases in (subjective) aggression (which increased with dose from 20 to 40%) and anxiety (from 11 to 33%) were observed. All this suggests that lower doses of THC/cannabis generally cause pleasant effects of relaxation and reduced attention, whereas with high doses CNS depression is partly overcome by more stimulatory effects.
These results suggest that most CNS measurements are sensitive to the effects of THC/cannabis. Some parameters were very sensitive to THC/cannabis. For such biomarkers, most doses studied in the literature may have been too high to show clear dose—response relationships. The proportion of significant effects did not increase much at the highest dose range compared with the medium range, although this does not exclude a dose-related increase in the size of the effects per se. Except for some subjective effects (feeling high, psychotomimetic feelings, drug effects) and heart rate, most tests did not show consistent effects even at quite high THC/cannabis doses. It could be argued that less sensitive CNS tests would have shown effects at even higher doses, well above 18 mg. For most subjects, particularly for inexperienced cannabis users, two or more cannabis cigarettes would cause an overdose, with questionable pharmacological and functional selectivity and an unacceptable adverse event profile. Hence, sensitive biomarkers at low to medium doses are needed to characterize the concentration—effect relationships of CB1/CB2 agonists. At present, the literature indicates that the choice is limited to heart rate and subjective effects.
Summary
Biomarkers are useful tools to study drug effects, since they can provide information on the potential pharmacological effects of the investigational drug in early-phase drug development. However, the number of tests and test variants that is used in studies of THC and cannabis seems excessively large. This abundance thwarts good assessment of the physiological, neuropsychological and subjective effects of this drug class, and there is dire need for test standardization in these areas. In general, the doses studied in the literature reflect the patterns of recreational use, and are often too high to determine accurately pharmacological dose—response relationships. Cannabis/THC has an effect on a wide range of CNS domains. At lower doses, THC/cannabis seems to be relaxant and to reduce attention, which is accompanied by impaired performance on other CNS tests that require motivation and active participation. At high doses, the drug seems to be more stimulatory. Subjective effects are the most reliable biomarkers to study the effects of cannabis, in addition to heart rate increases that reflect peripheral cannabinoid activation. This review indicates that these parameters are useful biomarkers that can be used in future studies to investigate the effects of THC/cannabis and other cannabinoid agonists on the CNS.
Source, Graphs and Figures: Biomarkers for the effects of cannabis and THC in healthy volunteers
