Which biological and self-report measures of cannabis use predict cannabis dependency and acute psychotic-like effects ?, H. Valerie Curran et a., 2018

Which biological and self-report measures of cannabis use predict cannabis dependency and acute psychotic-like effects ?

H. Valerie Curran, Chandni Hindocha, Celia J. A. Morgan, Natacha Shaban, Ravi K. Das and Tom P. Freeman

Psychological Medicine, 2018, 1-7.



Background : Changes in cannabis regulation globally make it increasingly important to determine what predicts an individual’s risk of experiencing adverse drug effects. Relevant studies have used diverse self-report measures of cannabis use, and few include multiple biological measures. Here we aimed to determine which biological and self-report measures of cannabis use predict cannabis dependency and acute psychotic-like symptoms.

Method : In a naturalistic study, 410 young cannabis users were assessed once when intoxicated with their own cannabis and once when drug-free in counterbalanced order. Biological measures of cannabinoids [(Δ9-tetrahydrocannabinol (THC), cannabidiol (CBD), cannabinol (CBN) and their metabolites)] were derived from three samples: each participant’s own cannabis (THC, CBD), a sample of their hair (THC, THC-OH, THC-COOH, CBN, CBD) and their urine (THC-COOH/creatinine). Comprehensive self-report measures were also obtained. Self-reported and clinician-rated assessments were taken for cannabis dependency [Severity of Dependence Scale (SDS), DSM-IV-TR] and acute psychotic-like symptoms [Psychotomimetic State Inventory (PSI) and Brief Psychiatric Rating Scale (BPRS)].

Results : Cannabis dependency was positively associated with days per month of cannabis use on both measures, and with urinary THC-COOH/creatinine for the SDS. Acute psychotic-like symptoms were positively associated with age of first cannabis use and negatively with urinary THC-COOH/creatinine; no predictors emerged for BPRS.

Conclusions : Levels of THC exposure are positively associated with both cannabis dependency and tolerance to the acute psychotic-like effects of cannabis. Combining urinary and self-report assessments (use frequency; age first used) enhances the measurement of cannabis use and its association with adverse outcomes.


Changes in the regulation of cannabis for recreational as well as medical use are currently continuing apace in many parts of the world. How patterns of use will subsequently change is not known, but even a small percentage increase in the current 183 million users worldwide will mean a considerable surge in absolute numbers. Quantifying the relative adverse and beneficial effects of cannabis and its constituent cannabinoids is, therefore, an important research priority (Curran et al., 2016). Cannabis use is associated with a 2-fold increased risk of developing a psychotic disorder (Marconi et al., 2016). Less attention has been paid to the much more common problem of cannabis addiction. It is estimated that 31% of past year cannabis users in the USA meet DSM-IV criteria for abuse or dependence (Hasin et al., 2015). However, the majority will be resilient and use cannabis without incurring serious mental health harms.

What predicts an individual’s vulnerability or resilience to experiencing the harmful effects of cannabis? Several factors are currently thought to be important including early adolescent initiation of use (Coffey et al., 2003; Mokrysz et al., 2016), genetic factors (Di Forti et al., 2012; Morgan et al., 2016), concurrent tobacco use (Hindocha et al., 2015) and frequent (especially daily) cannabis use (Coffey et al., 2003; Chen et al., 2005). Other factors that may be important include the level of Δ9 tetrahydro-cannabinol (THC) and other cannabinoids – especially cannabidiol (CBD) – in the strains that individuals use (Morgan and Curran, 2008; Morgan et al., 2010; Di Forti et al., 2015; Curran et al., 2016; Freeman et al., 2018).

One impediment to drawing conclusions about these risk factors is the varying measures of cannabis use that different studies employ. Despite growing international interest in this issue, there are currently no agreed standardised measures for assessing cannabis use in research (Yücel et al., 2016; Hindocha et al., 2017; Kögel et al., 2017). Although the majority of studies employ self-report measures (e.g. frequency of use; years used) few include questions on potency, dose and strain of cannabis (van der Pol et al., 2014). A minority employ biological measures, and when these are used there is much diversity in both the types of samples taken (e.g. hair, saliva, plasma, urine, actual cannabis used) and analyses subsequently carried out. Most estimate levels of Δ9−tetrahydrocannabinol (THC) and/or its metabolites, sometimes also cannabidiol (CBD) and less often other cannabinoids (Morgan and Curran, 2008; Demirakca et al., 2011; Freeman et al., 2014; Yücel et al., 2016).

If we could predict which variations in measures of cannabis use are and are not associated with adverse effects, then this would inform harm-reduction strategies which in turn would benefit those using cannabis for either recreational or medicinal purposes. Further, there are pragmatic reasons to explore which measures may be more or less associated with adverse outcomes because biological measures can be seen as personally intrusive and samples can be expensive to analyse.

We, therefore, set out to explore associations between multiple measures of cannabis use and two types of outcomes: the main harm we focussed on was dependence on cannabis; we also investigated acute psychotic-like effects that individuals experienced after ingesting the drug. Each of these outcomes was assessed by both self-report and by clinician-ratings. Three types of biological measures were used: analyses of cannabinoids (THC, CBD) and related metabolites in (i) participants’ hair, (ii) their urine and (iii) samples of cannabis each had used acutely. Self-report measures of use were: age of onset, years used, amount, frequency, time to smoke 3.5 g, the amount spent per week on buying cannabis, time since last use and preference or not for high potency cannabis strains. Our aim was to determine which measure or combination of measures best predicted the two outcomes.

We hypothesised firstly that using cannabis more frequently and using high (as opposed to low) potency varieties would be associated with increased rates of cannabis dependency (Morgan et al., 2010; Freeman and Winstock, 2015; Freeman et al., 2018). Secondly, we hypothesised that the use of high potency strains would lead to more acute psychotic-like experiences than the use of low potency strains (Di Forti et al., 2015). Thirdly, we hypothesised that CBD might mitigate the harmful effects of THC on both cannabis dependence and acute psychosislike symptoms (Morgan and Curran, 2008; Bhattacharyya et al., 2010; Morgan et al., 2010; Englund et al., 2013).