Examining the profile of high-potency cannabis and its association with severity of cannabis dependence

Background Cannabis use is decreasing in England and Wales, while demand for cannabis treatment in addiction services continues to rise. This could be partly due to an increased availability of high-potency cannabis. Method Adults residing in the UK were questioned about their drug use, including three types of cannabis (high potency: skunk; low potency: other grass, resin). Cannabis types were profiled and examined for possible associations between frequency of use and (i) cannabis dependence, (ii) cannabis-related concerns. Results Frequent use of high-potency cannabis predicted a greater severity of dependence [days of skunk use per month: b = 0.254, 95% confidence interval (CI) 0.161–0.357, p < 0.001] and this effect became stronger as age decreased (b = −0.006, 95% CI −0.010 to −0.002, p = 0.004). By contrast, use of low-potency cannabis was not associated with dependence (days of other grass use per month: b = 0.020, 95% CI −0.029 to 0.070, p = 0.436; days of resin use per month: b = 0.025, 95% CI −0.019 to 0.067, p = 0.245). Frequency of cannabis use (all types) did not predict severity of cannabis-related concerns. High-potency cannabis was clearly distinct from low-potency varieties by its marked effects on memory and paranoia. It also produced the best high, was preferred, and most available. Conclusions High-potency cannabis use is associated with an increased severity of dependence, especially in young people. Its profile is strongly defined by negative effects (memory, paranoia), but also positive characteristics (best high, preferred type), which may be important when considering clinical or public health interventions focusing on cannabis potency.


Introduction
There is huge variation in the types of cannabis (marijuana) available worldwide (UNODC, 2014). This is evident in illicit markets and also legal ones. For example, an unprecedented number of cannabis products and preparations are now available in Colorado (Coombes, 2014). By contrast, sales in Uruguay may be restricted to five strains only, with an upper limit on potency (Coombes, 2014).
Cannabis potency is typically judged according to concentrations of delta-9-tetrahydrocannabinol (THC), the primary psychoactive constituent in cannabis. However, the cannabis plant contains many other cannabinoids, most notably cannabidiol (CBD). These other cannabinoids (and possibly other plant chemicals known as terpenoids; Russo, 2011) contribute to potency by moderating the effects of THC. For example, CBD can block or dampen the effects of THC across a range of domains (Zuardi et al. 1982;Morgan & Curran, 2008;Morgan et al. 2010aMorgan et al. ,b, 2012Englund et al. 2012;Hindocha et al. 2015). These findings concur with users' ratings of cannabis potency, which are positively correlated with THC and negatively with CBD (Freeman et al. 2014).
Natural cannabinoid synthesis (and therefore cannabis potency) is influenced by a range of factors including genetics, growing conditions (especially light), harvest time, the part of the plant used, drying, storing and processing (Potter, 2014). Most products can be classified into three broad types: (1) high potencyindoorgrown floral material of unfertilized plants, whereby energy is diverted from seed production to cannabinoid synthesis ('skunk', 'sinsemilla'; meaning 'without seeds'); (2) low potencyoutdoor-grown imported floral material ('herbal', 'grass', 'weed'); and (3) compressed blocks of plant matter ('resin', 'hashish'). Skunk is characterized by the highest THC content (∼15%), followed by imported herbal/grass (∼9%) and then resin (∼5%), although there is considerable variation within these categories (Hardwick & King, 2008). Concentrations of CBD are typically low or completely absent in skunk and other herbal/grass preparations. By contrast, resin/hashish (and presumably landrace populations of cannabis plants) typically contain comparable quantities of THC and CBD (Potter et al. 2008). Thus, indoor-grown floral cannabis (skunk) is the clearly most potent type of cannabis (followed by imported herbal/grass, and then resin/hashish), and might be expected to be most strongly associated with any adverse effects of cannabis use. This is currently an under-researched area, although preliminary evidence suggests that regular use of high-potency (skunk) cannabis is predictive of first-episode psychosis (Di Forti et al. 2009) and an earlier onset of psychosis, particularly among daily users (Di Forti et al. 2013). By contrast, resin/hashish is not linked to an increased risk of psychosis, even among daily users (Di Forti et al. 2015).
It is estimated that 3.8% of the world's population used cannabis in the last year (UNODC, 2014) and this figure has remained relatively stable in the last decade. In England and Wales, however, prevalence of last year use dropped from 10.7% to 6. 6% between 2002/2003(Home Office, 2014. Despite this overall reduction in use, demand for cannabis in addictiontreatment services has continued to rise across the same time period: between 2005/2006 and 2013/2014 new admissions for cannabis rose from 7579 to 11 821 in adults (NDTMS, 2014) and from 9043 to 13 659 among under 18s (NDTMS, 2015). There are now more first-time clients for cannabis treatment in Europe than any other illicit drug (EMCDDA, 2014).
One possible explanation for these trends is an increase in cannabis potency. Data from cannabis seizures have documented rising THC concentrations in the UK. This is predominantly due to an increase in the availability of high-potency, indoor-grown (skunk) cannabis which made up 15% of police seizures in 1999-2002(King et al. 2004), 55% in 2004-2005(Potter et al. 2008) and 81% in 2007(Hardwick & King, 2008. These trends are matched by seizure data across Europe (EMCDDA, 2014). Cannabis users titrate (use less) as THC rises but only partially (Freeman et al. 2014;van der Pol et al. 2014) and not in response to CBD (Freeman et al. 2014). It is therefore possible that repeated exposure to high THC concentrations, and little if any CBD, may have increased users' dependence on cannabis.
In this study, we recorded detailed information on use of and experiences with different types of cannabis through an online drug survey. This approach made it possible to recruit a large sample who had used all of three different cannabis types (skunk, other grass, resin) within the last 12 months, permitting withinsubject comparison of cannabis types. We aimed to test the hypothesis that severity of dependence, and concerns about cannabis use, are more strongly associated with use of high-potency than low-potency cannabis. Additionally, we explored users' experiences of each type of cannabis in relation to effects on memory, paranoia, quality of high, preference, value for money and availability.

Design and participants
An online cross-sectional drugs survey, The Global Drug Survey, was conducted in November 2009 as reported elsewhere (Winstock et al. 2011). All participants confirmed that they were aged 518 years, and consented for the information they gave to be analysed. Ethical approval was received from the joint South London and Maudsley and Institute of Psychiatry National Health Service (NHS) Research Ethics Committee.

Assessments
The survey collected demographic data and detailed information on use of and experience with a number of substances. The data presented and analysed in this report is the UK data only; cases living in England, Scotland, Wales, and Northern Ireland form the UK dataset. Rather than recording information on cannabis use generally (e.g. age when cannabis was first used, days of cannabis use per month), separate questions were provided for (1) resin, (2) skunk grass (hereafter 'skunk'), (3) grass other (hereafter 'other grass'). This enabled each type to be rated as a separate drug. The following information was collected for each type of cannabis:

Comparing use of cannabis types
Used in the last 12 months? (yes/no). Days used in the last month. How long does ⅛th last you (in days). † How many joints from ⅛th?

Profiling cannabis types
Respondents who had used all three cannabis types in the last 12 months were asked to choose one type for each of the following questions: Which gives the best high? Which is the best value for money? Which is most likely to get you paranoid?
Which is most likely to affect your memory? Which is your preferred type? Which is most available?

Route of administration
Respondents were asked whether they had ever used cannabis using the following methods: smoked in joint without tobacco, smoked in joint with tobacco, smoked in bong/water pipe without tobacco, smoked in bong/water pipe with tobacco, eaten/cooked, used in a vaporizer.

Severity of dependence and cannabis-related concerns
These questions were assessed with reference to cannabis use generally. Cannabis dependence was assessed using the Severity of Dependence Scale (SDS; Gossop et al. 1995), which was adapted for the survey with abbreviated response options as shown below. Scores can range from 0 to 15, and scores 53 on the original scale indicate dependence on cannabis (Swift et al. 1998).
(2) Does the prospect of missing a smoke make you very anxious or worried? [never (0); sometimes (1); often (2); always (3) Additionally, the following questions were asked (yes/no): Have you ever discussed your cannabis use with a healthcare professional? Have you ever thought you might need treatment for your cannabis use? Have you ever sought treatment for cannabis use? Have you ever tried to stop smoking cannabis?
Participants were also asked about a range of concerns relating to their cannabis use: We are interested in what worries you about smoking cannabis. Please rate the following possible health-related consequences of smoking cannabis on a scale of 1-10, where 1 = no concern for you and 10 = big concern for you: cancer, chronic lung disease, effect on memory, effect on mental health, legal issues, effect on relationships, effect on work or study, lack of motivation.

Statistical analysis
Repeated-measures ANOVA models were used to compare each cannabis type for indices of use. Post-hoc t tests were corrected locally using the Bonferroni method. χ 2 tests were used for comparing the profile of cannabis types. Current age was split into quartiles (<21, 21-22, 23-27 and >27 years) for analysis of first use, and profile of effects. Pearson correlational analyses were used to establish associations between cannabis use variables and SDS scores. Multiple regression was used to predict severity of cannabis dependence and cannabis-related concerns from indices of cannabis use. Analysis of gender was coded as female = 1, male = 2, and age was entered as a continuous variable. For all regression models, bias-corrected accelerated 95% confidence intervals (CIs) were estimated using 10 000 boostrapping samples.

Demographics
Data were available for 2514 respondents. In the last year, prevalence of use was 72.5% for skunk, 68.6% for other grass, and 58.7% for resin cannabis preparations. Thirty-seven per cent (929 respondents) had used all three cannabis preparations in the last year; all further analyses were conducted in this sample. These participants had a mean age of 24.25 (S.D.=6.86) years and 70.2% were male. Routes of administration (ever used/most common use) were as follows: smoked in joint with tobacco (98.9%/85.2%), smoked in joint without tobacco (75.5%/5.5%), smoked in bong/water pipe with tobacco (69.6%/3.3%), smoked in bong/ water pipe without tobacco (85.2%/4.5%), eaten/ cooked (80.1%/1.1%), used in a vaporizer (36.8%/0.4%).
Comparing use of cannabis types (Table 1) Participants reported differences in the number of days they had used each type of cannabis in the last month (F 2,761 = 38.332, p < 0.001, ŋ p 2 = 0.087). Skunk was used for more days than other grass (p < 0.001) and resin (p < 0.001), while other grass was used for more days than resin (p < 0.001). No differences were found for the number of days to smoke ⅛th (F 2,895 = 1.655, p = 0.197, ŋ p 2 = 0.003). Differences emerged for the number of joints made out of ⅛th (F 2,893 = 62.710, p < 0.001, ŋ p 2 = 0.108), reflecting a larger number of joints made from resin compared to skunk (p < 0.001) or other grass (p < 0.001); a similar number of joints were made for skunk and other grass (p = 0.078). There were also differences in age of first use (F 2,1432 = 41.059, p < 0.001, ŋ p 2 = 0.043); resin was used earlier than skunk (p < 0.001) and other grass (p = 0.002); while other grass was used earlier than skunk (p < 0.001).
Post-hoc tests showed that in young people (under 21's and 21-to 22-year-olds), all three types of cannabis were first tried at similar ages (all p's > 0.06). By contrast, 23to 27-year-olds had tried resin earlier than other grass (p < 0.001) and skunk (p = 0.010), which were both tried at an equivalent age (p = 1.000). In the over 27's, there was a marked delay in first trying skunk relative to other grass (2.00 years, p < 0.001) and resin (2.42 years, p < 0.001), and resin was again tried earliest (p = 0.002 compared to other grass). These data are consistent with a shift in the relative availability of resin and skunk over time, alongside a tendency for younger people to try cannabis at an earlier age.

Severity of dependence and cannabis-related concerns
Exploratory correlations were conducted between SDS scores and all 12 indices of use (for skunk, other grass and resin: age of first use, days used in the last month, days taken to smoke ⅛th, number of joints from ⅛th). For days used in the last month, Pearson's r values reflected a medium-large effect size for skunk (r = 0.432) and small-medium effect sizes for other grass (r = 0.290) and resin (r = 0.247). For all other indices of use, effect sizes were small (all r's 40.155). Scores for severity of dependence (and individual concerns about cannabis) were therefore regressed onto days of skunk, other grass, and resin use in the last month.
Four hundred and three respondents had used each of the three cannabis types at least once in the last month; the following analyses were conducted in those individuals. Within their lifetime, 23.0% had discussed their cannabis use with a healthcare professional, 17.8% thought they might need treatment for their cannabis use, 5.3% had sought treatment for cannabis use, and 47.4% had tried to stop smoking cannabis. On the SDS, they had a mean score of 2.82 (S.D. = 3.29). Scores ranged from 0 to 14 (out of a maximum of 15) and quartiles were 0.00, 2.00, and 4.50. When classified using the cut-off of 53 (Swift et al. 1998), 38% of the sample currently met criteria for cannabis dependence.
As shown in Table 2, frequency of cannabis use in the last month predicted severity of dependence, accounting for 14.4% of the variance in these scores. This was driven by skunk; no associations emerged for other grass or resin. We additionally investigated whether this effect was moderated by gender or age. Adding gender and age into the model (step 2) did not account for additional variance, but including them as moderators of skunk use (step 3) improved model fit. Removing redundant predictors (step 4) did not result in a loss of variance explained and accounted for a total of 15.5%. SDS scores increased (b = 0.254, 95% CI 0.161-0.357, p < 0.001) for each additional day of skunk per month. This relationship became stronger as age decreased (b = −0.006, 95% CI −0.010 to −0.002, p = 0.004). SDS scores also increased with age (b = 0.081, 95% CI 0.014-0.170, p = 0.039). Frequency of cannabis use (skunk, other grass, resin) did not predict scores for any of these concerns. Fig. 1. Current age and first use of cannabis. Young people in the sample (currently under 23) were exposed to all three types of cannabis at similar ages. Older people were exposed to resin earlier than other types of cannabis, and skunk use was markedly delayed in the over 27's. These results support a shift in the relative availability of resin and skunk over time. High potency cannabis and severity of dependence 3185

Discussion
This study compared the profile of three cannabis types and their associations with cannabis dependence. Our findings clearly show that use of high-potency (skunk) but not low-potency (other grass, resin) cannabis is associated with an increased severity of dependence, especially in young people. Furthermore, the profile of high-potency (skunk) cannabis was marked in terms of negative effects (memory and paranoia) but also positive characteristics (preferred type and best high). It was also rated as the most available type, but was not the best value for money.
The past decade has seen a huge increase in the prevalence of high-potency (skunk-type) cannabis in England and Wales (King et al. 2004;Hardwick & King, 2008;Potter et al. 2008;Freeman et al. 2014) alongside rising demand for cannabis treatment in addiction services (NDTMS, 2014(NDTMS, , 2015. Our findings are consistent with these observations, and are the first to our knowledge reporting a link between cannabis potency and severity of drug dependence. Thus, clinically, it might be useful (and desirable) to encourage skunk users at risk of/experiencing dependence to move to less potent forms of cannabis if they are not motivated to quit.
Younger people were especially vulnerable, displaying a stronger relationship between extent of skunk use and severity of dependence. This is in agreement with observations that more under 18's seek treatment for cannabis than all adults combined, unlike any other drug (NDTMS, 2014(NDTMS, , 2015. Young people in our sample were also exposed to skunk from an earlier age; older adults tried resin first and had not tried skunk until an average of 2.42 years later. Given that these changes in the illicit market may have increased rates of cannabis dependence in the UK, it will be important to evaluate the impact of careful regulation of cannabis potency (e.g. as planned in Uruguay) and other legislative changes (e.g. in the US) on cannabis dependence.
One explanation for our findings is that greater THC exposure enhances the dependence-forming properties of cannabis. This interpretation is in keeping with preclinical research showing that THC is reinforcing in a dose-dependent manner (Tanda et al. 2000;Justinova et al. 2003). Titration by cannabis users appears to counteract higher THC concentrations, but only partially (Freeman et al. 2014;van der Pol et al. 2014). As a result, episodic use of high-potency cannabis will typically deliver larger doses of THC.
Interestingly, people in this study added less cannabis to their joints (based on the number of joints made out of 3.5 g) when using resin compared to other types. This may have further reduced their dose of THC, in addition to the low potency typical of resin. However, they took a similar number of days to smoke 3.5 g in totalperhaps suggesting that resin users smoke more joints with smaller amounts of cannabis (and possibly more tobacco), resulting in similar consumption of raw cannabis overall. The presence of CBD may also be relevant. Cannabis with a high ratio of CBD:THC (i.e. resin) reduced attentional bias to drug cues (a process implicated in addiction; Field & Cox, 2008) relative to low CBD:THC cannabis (i.e. skunk) (Morgan et al. 2010b). CBD was also found to reduce symptoms of  (Crippa et al. 2012).
Although our results support a relationship between cannabis potency and severity of dependence, they do not imply a causal relationship, and many other factors are likely to be involved. For example, a prospective study found no independent associations between indices of cannabis use (including preferred type and THC concentrations) and subsequent incidence of dependence (van der Pol et al. 2013). This study differed from ours in a number of respects, and included a number of additional predictors (e.g. socio-demographic, vulnerability and stress factors). Additionally, it used a betweensubjects comparison of preferred cannabis type and potency as opposed to our within-subject analysis.
Contrary to our expectations, degree of cannabis use did not predict level of concerns about cannabis (memory, mental health, work or study, relationships, motivation, chronic lung disease, cancer, legal issues). This might reflect the varying susceptibility to cannabis-related harms between individuals. Another possible explanation is that more frequent users hold the belief that their use is not problematic. These findings also suggest that all levels of use can be associated with modest health concerns. This may imply that infrequent users who are not currently using treatment services are nevertheless worried about the effects of cannabis, and might benefit from help at an individual or population-based level.
It is also noteworthy that memory emerged as the strongest concern about cannabis use in this study, as this was also the most defining feature of high-potency cannabis. This is consistent with the high THC, low CBD profile of skunk. When acutely administered, THC produces robust and dose-dependent impairments in verbal memory (Curran et al. 2002;D'Souza et al. 2004) and these impairments can be ameliorated by co-administration of CBD (Morgan et al. 2010a;Englund et al. 2012). Similarly, skunk was identified as the type of cannabis most strongly associated with paranoia. This is consistent with evidence that the paranoia-inducing effects of THC can also be inhibited by CBD (Englund et al. 2012), and that regular skunk use is associated with an increased risk (Di Forti et al. 2009) and earlier onset (Di Forti et al. 2013) of psychosis, while resin/hashish is not, even in daily users (Di Forti et al. 2015).
Although skunk was most clearly defined by these negative effects, it was also rated as having the 'best high' and was considered the 'preferred type'. Clinical and public health interventions related to high-potency cannabis, focusing in its negative effects, should be interpreted in the context of users' own preferences. Given that skunk was rated as the most available type of cannabis, consistent with previous findings (Hardwick & King, 2008;Potter et al. 2008;Freeman et al. 2014) people who do prefer it will probably find it easy to obtain in the illicit market. On the other hand, those who do not prefer skunkor find that its negative effects outweigh the desirable onesmay have little choice due to the current lack of available alternatives. Perhaps varieties of cannabis with weaker effects on memory and paranoia (e.g. other grass, resin) may be more desirable in this respect.
When comparing these lower potency varieties, it is somewhat surprising that resin was rated as having a better high and stronger effects on memory and paranoia, given that it generally contains lower THC and higher CBD than imported herbal cannabis (Hardwick & King, 2008). It may be relevant that variation in cannabinoid content is comparatively greater in resin (Potter et al. 2008) and some resin can be highly potent (e.g. 39.3% THC in the Netherlands; Pijlman et al. 2005). Such preparations are incredibly rare in the UK (Potter et al. 2008) but perhaps experience with especially potent forms of resin could have led some people to rate it as having the best high, and strongest effects on memory and paranoia.
This study has some limitations. First, it used a selfselecting (drug using) sample. This enabled a large number of cannabis users to be recruited, but it does limit the extent to which the findings can be attributed to the general population, or more problematic users, as dependence scores were modest on average. Second, because dependence was estimated using the SDS rather than a structured clinical interview, it was not possible to tease apart specific aspects of cannabis use disorder such as tolerance, withdrawal, craving, failing obligations, giving up other recreational interests, and persistent use in spite of problems. Additionally, the study was cross-sectional and causality cannot be established on the basis of these results. Indeed, it is quite plausible that reverse causation might explain our findings (e.g. as a result of dependence, people use more skunk). Third, skunk was used for more days per month than the other types, which might explain the reported association with dependence. However, variance in days per month of use was similar for each of the three types, suggesting that these data were equally appropriate to detect the existence of possible associations with dependence. Fourth, although we quantified use of three different cannabis preparations, we cannot be sure that the terms we used (e.g. skunk) were meaningful to the population tested (Potter & Chatwin, 2012), although similarly named types were predictive of actual THC and CBD concentrations elsewhere (Freeman et al. 2014).

Conclusion
Use of high-potency (skunk) cannabis is associated with an increased severity of dependence, especially in young people. Skunk is also rated as having stronger effects on memory impairment and paranoia than other types of cannabis, but at the same time it produces the best high and is users' preferred type.