Reflect

European Neuropsychopharmacology (2017) 27, 732–743

http://dx.doi.org/1 0924-977X/& 2017 E

nCorrespondence Hawthorn, VIC 3122

E-mail address: a

www.elsevier.com/locate/euroneuro

DSM-5 cannabis use disorder, substance use and DSM-5 specific substance-use disorders: Evaluating comorbidity in a population- based sample

Amie C. Hayleya,n, Con Stougha, Luke A. Downeya,b

aCentre for Human Psychopharmacology, Swinburne University of Technology, Hawthorn, Australia bInstitute for Breathing and Sleep, Austin Hospital, Melbourne, Australia

Received 16 January 2017; received in revised form 23 May 2017; accepted 10 June 2017

KEYWORDS DSM-5; Cannabis use disorder; Substance use disorder; Illicit; Prescription; NESARC

0.1016/j.euroneur lsevier B.V. and E

to: Centre for Hu , Australia. hayley@swin.edu

Abstract Cannabis use disorder (CUD) is frequently associated with concurrent substance use and/or comorbid substance use disorders (SUDs); however there is little specificity with regard to commonly abused individual drug types/classes. This study therefore aimed to provide insight into the degree of these co-occurring relationships across several specific newer and older generation illicit and prescription drugs. 36,309 adults aged 18+ from wave 3 of the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC-III) were assessed. Weighted cross-tabulations and multivariable logistic regression analyses were used to evaluate comorbid- ity between current DSM-5 CUD, substance use and DSM-5 SUD. Current DSM-5 CUD is associated with greater lifetime use of all examined drug classes, and previous 12-month use of several newer-class illicit and prescription stimulant-based substances (all po 0.05). Current DSM-5 CUD was similarly associated with increased incidence of a range of DSM-5 SUDs and was independently associated with concurrently reporting current DSM-5; sedative (Adjusted OR= 5.1, 95%CI 2.9–9.0), cocaine (AOR= 9.3, 95%CI 5.6–15.5), stimulant (AOR= 4.3, 95%CI 2.3–7.9), club drug (AOR= 16.1, 95%CI 6.3–40.8), opioid (AOR= 4.6, 95%CI 3.0–6.8) and alcohol-use disorder (AOR= 3.0, 95%CI 2.5–3.7); but not heroin or ‘other’ drug use disorder (both p40.05). High comorbidity exists between DSM-5 CUD and many specific DSM-5 SUDs. Newer-class illicit and prescription stimulant-based drug use disorders are overrepresented among those with DSM-5 CUD. These findings underscore the need for tailored treatment programs for those presenting with DSM-5 CUD, and for greater treatment specification where poly-drug use is evident. & 2017 Elsevier B.V. and ECNP. All rights reserved.

o.2017.06.004 CNP. All rights reserved.

man Psychopharmacology, Faculty of Health, Arts and Design, Swinburne University of Technology,

.au (A.C. Hayley).

733Comorbid cannabis use disorder and substance use disorders

1. Introduction increased rates of drug-related deaths and nonfatal drug

Cannabis (marijuana) is the most frequently consumed drug after alcohol and tobacco, and is the most commonly cultivated, trafficked and abused illicit substance worldwide (Degenhardt and Hall, 2012; UNODC, 2016). Rates of cannabis use and abuse are on the rise (Compton et al., 2004), and this is considered by some to be at least partially driven by greater availably (Freisthler and Gruenewald, 2014), recent decriminalisation in select regions (Miech et al., 2015), updated medicinal status (Freisthler and Gruenewald, 2014) and subsequent shifts in the public perception of the drug which typically align with public policy and legalisation status (Schuermeyer et al., 2014); however the exact role of marijuana liberalisation remains somewhat inconclusive (Pacula et al., 2017). Approximately one in ten of those who have ever consumed cannabis will also develop defined instance of cannabis use disorder (CUD) (Wagner and Anthony, 2002); and the progression from cannabis use to dependence is often quite rapid (Ridenour et al., 2003). Recent estimates suggest that between 2.5 (Hasin et al., 2016) and 19% (Farmer et al., 2015) of the adult population, or as many as 13.1 million individuals’ globally, meet criteria for CUD disorder. Notable peaks in prevalence rates of CUD are observed among those who have previously used any other illicit substance (Chen et al., 2005), in young adults (20–24 years), among males, and those who live in higher income countries (Degenhardt et al., 2013); however some of these characteristic differences appear to be diminishing among more recent cohorts assessments (Degenhardt et al., 2008). Both epidemiological (Fergusson and Horwood, 2000) and preclinical data (Agrawal et al., 2004) indicates the potential gateway liability of cannabis as a preceding factor in later substance use and/or abuse. Indeed, a dose-response relationship exists between the frequency of previous can- nabis use and rates of alcohol use (Stinson et al., 2006) and consumption of select illicit drugs such as amphetamine and non-medical opiates (Degenhardt et al., 2013); however the exact role of cannabis as a preceding factor in the develop- ment of other drug use disorders is still regarded as contentious.

Use or abuse of any psychoactive substances is asso- ciated with a marked increase in the probability of similarly abusing other categories of drugs (Tsuang et al., 1998). Longitudinal epidemiological studies con- firm the predictive value of previous cannabis use in the later development of other defined substance use dis- orders (SUD) (Brook et al., 2002). This has also been shown to extend to cannabis dependence, as individuals with CUD have a similarly significantly elevated risk of other comorbid illicit drug use and defined instances of SUDs (Grant and Pickering, 1998). Indeed, emergent epidemiological survey data indicate that among those with DSM-4 and IDSM-4 diagnosed CUD (APA, 1994), 20% and 28%, respectively, reported concurrent use of other illicit drugs, and 14% and 23%, respectively, also met criteria for another illicit drug use disorder in the previous year (Grant and Pickering, 1998). Poly- substance abuse is increasingly reported as a contributing factor associated with acute hospital admissions (Mowbray et al., 1997) and has also been linked to

overdoses (Kerr et al., 2007). Despite this, previous evaluations of comorbidity between CUD and other drugs of abuse lack specificity with regard to types of comorbid drug classes examined, as research often utilises aggre- gate measures of substance abuse only (Hasin et al., 2016), and does not provide individual assessment of each drug type/class. Moreover, few studies encompass a wide range of drug types and/or classes; particularly newer generation drugs of abuse among younger people, such as prescription amphetamines (Low and Gendaszek, 2002). This is problematic, as collapsing abuse and dependence categories into global measures of abuse and only evalu- ating a narrow range of substances neglects the more subtle archetypal aspects of poly-drug use, and provides little updated or relevant information for clinicians or those directly involved in both acute and longer-term treatment programs.

Despite the rapidly rising rate of those seeking treat- ment for CUD disorder (Dennis et al., 2002), systematic development of cannabis-specific clinical treatment interventions are currently lacking. Extant models of therapy for CUD often employ cognitive behavioural techniques derived from alcohol interventions which are modified to meet the needs of cannabis dependent patients (Copeland et al., 2001b); however the efficacy of long-term treatment programs is unclear, and the impact of complex diagnoses on these outcomes is largely unknown. Comorbidity of drug use significantly com- pounds the efficacy of a single-approach substance abuse treatment programs, and successful amelioration of co- occurring drug use often predicts favourable long-term treatment outcomes (such as a reduction in drug use and occurrence of non-fatal overdoses) (Stewart et al., 2002). Among cannabis users seeking treatment, co-occurring substance use and/or abuse has been linked to an elevated risk on measures of cannabis dependence, co- morbid psychopathology, dysfunctional cannabis cogni- tions (Connor et al., 2013) and poorer treatment out- comes (Belendiuk et al. 2015). Multifaceted interventions are often recommended for those with poly-substance use (Stephens et al., 1993), and treatment avenues are often informed by the types of drugs used. Examination of specific characteristic poly-drug use among those with CUD may therefore assist in the development of tailored interventions for the growing number of individuals seek- ing treatment for this disorder, and thus help inform this urgent area of need within the clinical community.

Due to the high global prevalence of cannabis use and increasing prevalence of CUD, coupled with the incurred burden and comorbidity potential with other drug use and specific drug use disorders, explicit and systematic evalua- tions of poly-substance abuse have high clinical and public health relevance. Many of the current estimates and clinical interpretations examining CUD and other substance use are limited to aggregate or cursory evaluations only and lack specificity; and thus are of little clinical relevance. Com- prehensive assessments as to the extent and strength of these associations may assist in the formulation of tailored and specific treatment modalities for individuals who fall within this diagnostic stratum.

A.C. Hayley et al.734

2. Experimental procedures

2.1. NESARC-III survey

Data were drawn from the 2012–2013 National Epidemiologic Survey on Alcohol and Related Conditions (NESARC-III) (Grant et al., 2014a). The NESARC-III comprises a nationally representative sample of 36,309 civilian, non-institutionalised American adults aged 18 years and over (response rate = 61.1%) who reside in the continental United States, Alaska, and Hawaii. The sample included persons living in households and select group quarters, such as boarding houses, rooming houses, military personnel living off base, shelters, college quarters, and group homes. Veterans of the United States Armed Forces were included in sampling procedures; however those on current active military duty were excluded as they are not offered anonymity under Certificates of Confidentiality (Grant et al., 2014b). The NESARC-III employs a complex multistage probability sampling methodology incorporating primary sampling units (counties/groups of contiguous counties), secondary sampling units (groups of census- defined blocks), and tertiary sampling units (households within secondary sampling units from which respondents were selected, with blacks, Asians, and Hispanics oversampled) (Grant et al., 2014a). Data were adequately adjusted to compensate for partici- pant nonresponse and were weighted to be representative of the U. S. adult population based on 2012 American Community figures (Census, 2012). More detailed descriptions of sampling procedures for the NESARC-III can be found elsewhere (Grant et al., 2014a). Face-to-face computerised semi-structured interviews were con- ducted in respondents’ homes, and information pertaining to alcohol use, psychiatric health and physical disabilities was obtained. Ethical approval for the NESARC-III was obtained from the Westat Institu- tional Review Board and the Combined Neuroscience Institutional Review Board of the National Institutes of Health prior to data collection, and respondents were informed in writing about the nature of the survey.

2.2. Measures

2.2.1. Demographics Sociodemographic factors examined in this study included ethnicity (American Indian or Alaska native, Asian, Black or African American, Native Hawaiian or other Pacific Islander, White), gender, age, educational status (collapsed), marital status, and urbanicity (rural or urban dwelling).

2.2.1.1. Substance use. Substance use is assessed by the medica- tion module of the Alcohol Use Disorder and Associated Disabilities Interview Schedule-IV (AUDADIS-V).

2.2.2. Other substance consumption and use Drug use is assessed by the medication module of the AUDADIS-V. Participants were asked if they had ever used, had used in the past 12 month period, and what approximate age they had first used any of the following drugs; sedative and/or tranquilisers (e.g. Barbitu- rates), painkillers (e.g. Paracetamol), marijuana (e.g. cannabis, hashish, hemp), cocaine (e.g. crack, ‘blow’), legal stimulants, club drugs (e.g. ketamine, MDMA, methamphetamine), hallucinogens (e.g. LSD, mescalin), inhalant or solvents (e.g. nitrous oxide, glue), heroin (e.g. ‘smack’, black tar heroin) or ‘other’ drug/medication use (e.g. steroids). Alcohol use was recorded, and average daily ethanol use for the past 12 months was derived from summing beverage specific volumes across the four beverage types (‘cool- ers’, ‘beers’, ‘wine’ and ‘liquor’).

2.2.3. Tobacco use Smoking status (current, ex-smoker and lifetime non-user) was indicated by a positive response to whether participants had ever

(in entire life) used a tobacco product (100 cigarettes, at least 50 cigars, smoked pipe at least 50 times, used snuff at least 20 times, used E-cigarettes or E-liquid), in addition to a positive response to whether they had used a tobacco product (any) in the previous 12 month period (NIAAA, 2014) [if they had smoked at least one (cigarette/cigar/bowl of tobacco)/use at least (one pinch, dip, rub, plug, wad, or chew/E-cigarette cartridge/4 drops of E-Liquid)]. Ex- smokers were identified as those who indicated using a tobacco product in their lifetime, however had not used in the previous 12- month period. Lifetime non-smokers did not indicate any substan- tial tobacco use (e.g., less than 100 cigarettes).

2.3. DSM-5 substance use disorders

2.3.1. DSM-5 Cannabis use disorder (CUD) This study included survey respondents who met current DSM-5 criteria for CUD (APA, 2013). Diagnosis of current DSM-5 cannabis use disorder requires at least two of 11 diagnostic criteria within the 12-month period. Psychometric assessment of reliability of current diagnoses of drug use disorders (including CUD) have been shown to be good to excellent when evaluated in a general population samples (Grant et al., 2015).

2.3.2. DSM-5 substance use disorder Other DSM-5 SUD to be included as outcomes for this study were derived in a manner similar to that for CUD and comprised current (previous 12 month) diagnoses. Diagnoses evaluated in the current study included DSM-5; sedative and/or tranquiliser use disorder, cocaine use disorder, stimulant use disorder, club drug use disorder, hallucinogen use disorder, inhalant or solvent use disorder, heroin use disorder, opioid use disorder, alcohol use disorder and ‘other’ drug/medication use disorder.

2.4. Mental health factors

Psychiatric disorders: The DSM-5 criteria exclude substance-induced medical disorders and mood and/or anxiety disorders in order to reduce cross-diagnoses. DSM-5 mood disorders to be included for covariate analysis were: Current (previous 12 months) diagnosis of; DSM-5 primary Major depressive disorder (MDD), DSM-5 Bipolar I disorder and DSM-5 Post-traumatic stress disorder (PTSD). DSM-5 anxiety disorders examined included Panic disorder. These psychia- tric conditions were examined as they have a demonstrated close relationship with diagnoses of DSM-5 CUD in population-based samples (Hasin et al., 2016).

2. Statistical methods

Analyses were computed using SPSS version 21 Complex Samples design package (SPSS, 2012). This system imple- ments a Taylor series linearization to adjust standard errors of estimates for complex survey sampling design effects including clustered data. Frequency statistics were used to evaluate the prevalence of diagnoses of DSM-5 CUD and DSM-5 specific SUDs. Cross-tabulations were used to calculate the co-occurrence of diagnoses of current (pre- vious 12 month) and lifetime DSM-5 CUD and several demographic, health and lifestyle factors, and co- occurrence with both 12-month and lifetime diagnoses of DSM-5 specific SUDs. Odds ratios (OR) derived from logistic regression analyses were used to study associations between diagnoses of current DSM-5 CUD and current diagnoses of DSM-5 SUDs. Demographic and health vari- ables associated with diagnoses of DSM-5 CUD at po .05

735Comorbid cannabis use disorder and substance use disorders

were included in multivariate analysis. Separate multi- variable binary logistic regressions were performed to examine the association between current DSM-5 CUD and current DSM-5 SUDs (each for; DSM-5 sedative use disorder, cocaine use disorder, club drug use disorder, hallucinogen use disorder, inhalant/solvent use disorder, heroin use disorder, opioid use disorder, alcohol use disorder and ‘other’ drug/medication use disorder). The presence of each separate SUD (yes/no) was expressed as a binary outcome. Age, gender, educational status (collapsed), marital status, urbanicity (rural or urban dwelling), ethni- city, alcohol use, smoking status and psychiatric diagnoses were tested sequentially, and potential confounders and effect modifiers were checked in all statistical models. Significance was set at conventional p o .05 (two-tailed) for both univariate and multivariable analyses.

3. Results

3.1. Demographic and health/psychiatric factors

Demographic factors for the whole group and for those with current DSM-5 CUD are presented in Table 1. There was near-equal distribution of men and women, and proportion of individuals across age-ranges was comparable. A con- siderable proportion of the sample comprised White, non- Hispanic ethnicity, and reported living in ‘rural’ areas. A large proportion of the overall sample reported that they were currently married, and many identified as being a ‘former smoker’. Psychiatric diagnoses among the whole sample indicated that one in ten individuals reported current diagnoses off MDD, and one in twenty people reported current diagnoses of PTSD. Lifetime drug use was high among the sample, with one third of participants reporting having ever used marijuana (32.2%), and about one in ten reporting having ever used painkillers (11.3%) or cocaine (10%). Lifetime usage rates of other illicit drug use were similarly high, with 9.3% reporting the use of halluci- nogens, and as many as 8.3% reported the use of prescrip- tion stimulants (such as Ritalin). Current drug use (previous 12 month period) was highest for ‘other’ medication/drugs (such as steroids or Haloperidol) (12.4%), club drugs (5.4%) and painkillers (4.9%). Past year marijuana use was reported by only 1.1% of the total sample.

Among those diagnosed with current DSM-5 CUD, a significantly greater proportion were male, aged 18–24 years old and identified as White, non-Hispanic ethnicity com- pared to those without DSM-5 CUD diagnoses (all po0.001). A large proportion reported having had never married, resided in a rural area, and were significantly less likely to be a college graduate. A greater proportion of those with DSM-5 CUD were current smokers, and indicated signifi- cantly elevated rates of DSM-5 MDD, PTSD, Panic Disorder and Bipolar I disorder. Current DSM-5 CUD diagnoses was similarly associated with significantly increased rates of concordant DSM-5 MDD, PTSD, Bipolar I Disorder and Panic Disorder (all po0.001).

3.2. Associations between current DSM-5 cannabis use disorder and concordant substance use and substance use disorders

Individuals with diagnosed cases of current DSM-5 CUD reported significantly higher rates of lifetime use of any of; sedative/tranquiliser, painkiller, marijuana, cocaine, sti- mulants, club-drug, hallucinogen, inhalant/solvents, her- oin and other medication/drug use than those without DSM-5 CUD (all po0.05) (see Table 1). For current (past 12- month) drug use, those with confirmed cases of current DSM-5 CUD reported significantly greater use of; mari- juana, cocaine, stimulants, club-drugs and hallucinogens, as well as lower rates of painkiller use (all po 0.05) compared to those without the diagnosis. No association was observed for sedatives/tranquilisers, inhalant/sol- vents, heroin or ‘other’ medication/drug use. Those with current DSM-5 CUD also reported significantly greater alcohol consumption (g/day) compared to those without the diagnosis (p= o0.001). When evaluating the associa- tion between current DSM-5 CUD and specific DSM-5 SUD, individuals with current DSM-5 CUD diagnoses reported significantly increased rates of concurrent diagnosis of DSM-5; sedative/painkiller, cocaine, stimulant, club-drug, hallucinogen, inhalant/solvent, heroin, opioid and alcohol use disorder (all po0.05). No association was observed for ‘other’ mediation/drug use disorder.

3.3. Associations between current DSM-5 cannabis use disorder and age of drug use

Age of first reported drug use for the whole sample and for those with current diagnosis of DSM-5 CUD are presented in Table 2. For all drug types (expect heroin), reported use commenced prior to age 25 years (heroin use was reported, on average, to start at age 32 years). Use of cannabis was reported at an earlier age than any other substance (17 years), followed by the use of hallucinogens (LSD) and club drugs (MDMA, methamphetamine), with the ages of first use reported as 18 and 19 years, respectively. Those with current diagnoses of DSM-5 CUD reported, on average, a two year earlier commencement of use of any other substance; with as much as a 5-year earlier commence- ment of use noted for the use of sedatives/tranquilisers and painkillers. Age of first cannabis use among those with DSM-5 CUD was 15.7 years, compared to 17.7 years for those without the disorder.

3.4. Multivariable associations between current DSM-5 CUD and specific SUDs:

The results of each of the unadjusted and fully adjusted multivariable regression models evaluating the relationship between DSM-5 CUD and specific DSM-5 SUDs are presented in Table 3. Unadjusted odds ratios for the association between DSM-5 CUD and all SUDs ranged from 2.2 (other drug use/medication use disorder) to 117.4 (Club drug use disorder) (all p o 0.001). No association was observed between current DSM-5 CUD and cases of current other

Table 1 Demographic, health and drug use characteristics for whole sample and by current DSM-5 Cannabis use disorder, aged 18 years and above (weighteda).

Characteristics Overall sample Current DSM-5 CUDb

Total N = 36,309 No Yes n = 35,337 n = 972

% 95%CI Design effect % 95%CI Design effect % 95%CI Design effect p

Gender o0.001 Male 48.1 47.5–48.7 1.27 47.6 47.0–48.2 1.31 66.1 61.9–70.1 1.77 Female 51.9 51.3–52.5 1.27 52.4 51.8–53.0 1.31 33.9 29.9–38.1 1.77 Age group (yr) o0.001 18–24 13.1 12.5–13.6 2.71 12.3 11.7–12.8 2.60 43.7 39.5–47.9 1.70 25–34 17.2 16.7–17.8 2.06 17.0 16.4–17.5 1.99 26.9 23.2–30.8 1.72 35–44 17.1 16.6–17.7 1.99 17.2 16.7–17.8 1.97 13.5 11.0–16.6 1.58 45–54 18.6 18.0–19.2 2.02 18.8 18.3–19.5 2.15 8.8 6.8–11.3 1.49 55–64 16.4 15.9–17.0 1.79 16.7 16.2–17.2 1.78 5.7 4.2–7.6 1.26 65+ 17.6 16.9–18.3 3.38 18.0 17.3–18.7 3.36 1.5 0.7–3.2 2.05 Ethnicity o0.001 White, non-Hispanic 66.2 64.6–67.7 9.60 66.4 64.4–67.7 9.54 58.2 54.4–62.0 1.40 Black, non-Hispanic 11.8 10.6–13.2 15.00 11.6 10.6–13.2 14.70 20.8 17.3–24.7 1.95 American Indian/Alaska Native 1.6 1.3–1.8 3.44 1.5 1.3–1.8 2.98 3.3 1.8–5.9 2.90 Asian/Native Hawaiian/Other Pacific Islander 5.7 4.9–6.7 15.09 5.8 4.9–6.7 14.73 2.8 1.7–4.6 1.70 Hispanic, any race 14.8 13.5–16.1 13.10 14.7 13.5–16.1 13.12 14.9 12.7–17.5 1.05 Urbanicity o0.001 Urban 1.3 1.2–1.5 1.61 1.3 1.2–1.5 1.57 2.6 1.7–3.9 1.05 Rural 32.1 31.2–33.0 3.15 31.7 30.8–32.6 3.38 45.6 42.2–49.2 1.16 Marital Status o0.001 Married 51.2 50.1–52.2 4.11 51.2 51.0–53.1 4.03 16.9 14.2–20.0 1.5 Defacto (living with someone) 6.6 6.3–7.0 1.90 6.5 6.1–6.9 1.94 12.3 10.1–14.9 1.2 Widowed 5.8 5.5–6.1 1.81 5.9 5.6–6.3 1.86 1.3 0.6–2.5 0.4 Divorced 10.9 10.5–11.3 1.52 11.0 10.6–11.4 1.43 9.1 6.9–11.9 1.3 Separated 2.9 2.7–3.2 1.85 2.9 2.7–3.2 1.95 4.0 2.6–6.3 0.9 Never married 22.5 21.6–23.4 4.12 21.6 20.8–22.5 3.82 56.3 52.2–60.4 2.1 Education o0.001 Less than 9th grade 4.5 4.1–5.0 4.48 4.6 4.2–5.1 4.50 2.3 1.6–3.3 0.80 9th–11th grade 8.5 8.0–9.0 3.27 8.3 7.8–8.9 3.39 13.9 11.6–16.5 1.20 High School/GED equivalent 25.8 24.8–26.8 5.02 25.6 24.6–26.7 5.22 30.7 27.6–34.0 1.16 Some college (no degree) 21.6 20.9–22.3 2.54 21.3 20.6–22.0 2.43 33.0 29.3–36.9 1.51 College graduate 39.6 38.1–41.1 8.75 40.1 38.6–41.6 8.74 20.1 17.4–23.2 1.26 Smoking status (tobacco products) o0.001 Current smoker 27.2 26.3–28.1 3.60 26.0 25.2–26.9 3.6 71.1 67.3–74.7 1.57 Former smoker 18.7 17.9–19.4 3.33 19.0 18.2–19.7 3.3 7.1 5.5–9.2 1.18 Never smoked 54.2 53.0–55.3 5.00 55.0 53.8–56.2 5.0 21.7 18.9–24.9 1.21 DSM-5 Psychiatric Disordersc

Major Depressive Disorder (MDD) 10.4 9.9–10.9 2.48 10.0 9.5–10.5 2.23 25.5 22.0–29.4 1.66 o0.001 Post-Traumatic Stress Disorder (PTSD) 4.7 4.3–5.0 2.42 4.3 4.0–4.7 2.40 17.2 14.3–20.7 1.68 o0.001 Panic Disorder 3.1 2.9–3.3 2.42 2.9 2.7–3.1 1.50 10.0 7.7–12.8 1.67 o0.001 Bipolar I Disorder 1.5 1.4–1.7 1.76 1.4 1.2–1.5 1.70 8.8 7.0–11.1 1.18 o0.001 Alcohol use (g/day)d 0.64 0.61–0.67 2.5 0.60 0.57–0.63 2.4 1.74 1.58–1.92 0.84 o0.001 Drug Use: Evere

A .C . H ayley

et al.

736

Sedative/tranquilisers 7.5 7.0–8.0 3.12 7.0 6.5–7.4 2.87 27.8 23.9–32.1 1.96 o0.001 Painkillers 11.3 10.6–12.0 4.72 10.6 9.9–11.3 4.99 40.2 36.5–44.0 1.35 o0.001 Marijuana 32.2 31.1–33.2 4.82 30.4 29.4–31.4 4.65 100 0–100 - o0.001 Cocaine 10.0 9.5–10.5 2.48 9.4 9.0–9.9 2.16 30.3 26.2–34.7 1.98 o0.001 Stimulants 8.3 7.9–8.8 2.86 7.8 7.4–8.3 2.57 28.5 24.8–32.6 1.78 o0.001 Club drugs 4.4 4.1–4.7 2.01 3.9 3.6–4.2 1.91 24.7 21.1–28.6 1.77 o0.001 Hallucinogens 9.3 8.8–9.8 2.65 8.6 8.2–9.1 2.38 36.5 32.5–40.6 1.68 o0.001 Inhalants/solvents 3.1 2.9–3.4 2.06 2.9 2.6–3.1 1.90 11.9 9.5–14.7 1.47 o0.001 Heroin 1.6 1.5–1.8 1.49 1.5 1.4–1.7 1.51 5.5 3.7–8.0 2.01 o0.001 Other medication/drugs 0.8 0.7–1.0 1.90 0.8 0.7–0.9 1.87 1.8 0.9–3.4 1.78 0.014 Drug Use: Past 12 monthsc

Sedative/tranquilisers 4.6 3.8–5.7 1.27 4.3 3.4–5.3 1.21 8.3 4.4–15.4 2.11 0.05 Painkillers 4.9 4.1–5.9 1.57 5.1 4.2–6.0 1.48 3.7 1.9–7.0 1.51 0.33 Marijuana 1.1 31.1–33.2 4.8 1.0 0.8–1.3 1.53 2.0 1.2–3.4 1.34 o0.001 Cocaine 1.2 0.8–1.8 1.71 0.9 0.5–1.5 1.89 5.1 2.8–8.9 1.21 o0.001 Stimulants 3.3 2.5–4.4 1.90 2.9 2.1–4.0 1.92 6.9 3.4–13.5 2.13 0.03 Club drugs 5.2 3.8–7.2 2.26 3.9 2.5–6.1 2.72 12.9 7.8–20.5 1.99 0.001 Hallucinogens 2.0 1.4–2.8 1.75 0.8 0.5–1.4 1.74 12.4 8.4–17.9 1.60 o0.001 Inhalants/solvents 1.7 1.0–3.0 1.27 1.4 0.7–2.8 1.52 4.6 1.5–12.8 1.33 0.08 Heroin 3.2 1.6–6.0 2.18 2.7 1.3–5.5 1.86 8.7 1.6–35.5 3.15 0.17 Other medication/drugs 12.4 8.8–17.3 2.21 12.9 9.1–18.0 1.22 4.1 0.5–24.7 0.67 0.20 Substance Use Disorders: Currentc

Sedative/ tranquiliser use disorder 0.4 0.3–0.4 1.60 0.3 0.2–0.4 1.41 3.8 2.6–5.5 1.32 o0.001 Cocaine use disorder 0.3 0.3–0.4 1.27 0.2 0.2–0.3 1.32 5.0 3.7–6.7 1.13 o0.001 Stimulant use disorder 0.3 0.3–0.4 1.36 0.2 0.2–0.3 1.43 4.2 2.8–6.2 1.66 o0.001 Club drug use disorder 0.1 0.1–0.2 0.83 0.0 0.0–0.1 0.81 3.5 2.4–5.0 1.10 o0.001 Hallucinogen use disorder 0.0 0.0–0.1 2.60 0.0 0.0-0.0 0.43 1.7 1.0–3.0 2.27 o0.001 Inhalant or solvent use disorder 0.0 0.0–0.1 2.60 0.0 0.0–0.1 3.15 1.1 0.4–2.8 2.8 o0.001 Heroin Use Disorder 0.1 0.1–0.2 2.05 0.1 0.1–0.2 0.4 0.4 0.3–0.5 1.96 o0.001 Opioid use disorder 0.9 0.8–1.0 1.12 0.7 0.6–0.8 1.15 8.2 6.2–10.7 1.53 o0.001 Other medication/drug use disorder 0.0 0.0–0.1 1.18 0.0 0.0–0.1 1.22 0.1 0.0–0.5 0.68 0.45 Alcohol use disorder 13.9 13.3–14.5 3.00 12.7 12.1–13.3 2.81 59.4 56.1–62.7 1.08 o0.001

Values are given as mean (SD) and % estimate. Bold font represents statistically significant results (p o .05). aData are weighted by NESARC-calculated AUDADIS full-sample weight, clustering for county and stratification using the stratum variable, adjusting standard errors. In all cases complex

sampling procedures was used. bRefers to diagnosed cases of DSM-5 Cannabis use disorder as defined by the Alcohol Use Disorder and Associated Disabilities Interview Schedule-V (AUDADIS-V). cRefers to positive indication for the past 12-month period only dAverage daily ethanol intake over previous 12 months eRefers to positive indication of ‘ever used’ only.

737 C om

orbid cannabis

use disorder

and substance

use disorders

Table 2 Age of first reported drug use for whole sample and by current DSM-5 Cannabis use disorder, aged 18 years and above (weighteda).

Overall Sample Current DSM-5 CUD

No Yes†

Total N = 36,309 N = 35,337 N = 927

Drug use behaviours Mean SE 95%CI Mean SE 95%CI Mean SE 95%CI p

Age first used 17.6 .06 17.4–17.7 17.7 .06 17.6–17.8 15.7 .20 15.3–16.1 o0.001 Cannabis 24.6 .30 24.0–25.2 25.0 .32 24.4–25.7 20.4 .65 19.1–21.7 o0.001 Sedatives/tranquiliser 25.3 .30 24.7–25.9 25.8 .33 25.1–26.4 20.4 .54 19.3–21.5 o0.001 Painkiller 22.1 .13 21.9–22.4 22.2 .13 21.9–22.5 21.2 .50 20.3–22.2 o0.001 Cocaine 20.5 .17 20.2–20.9 20.7 .18 20.3–21.0 19.0 .29 18.4–19.6 o0.001 Stimulants 21.0 .18 20.7–21.4 21.1 .19 20.7–21.5 20.4 .39 19.6–21.1 o0.001 Club Drugs 19.0 .10 18.8–19.2 19.0 .10 18.8–19.2 18.8 .25 18.3–19.3 o0.001 Hallucinogens 18.2 .21 17.7–18.6 18.2 .23 17.8–18.7 17.4 .60 16.3–18.6 o0.001 Inhalants/solvents 22.8 .37 22.1–23.6 22.9 .40 22.1–23.7 21.5 1.0 19.6–23.5 o0.001 Heroin 32.1 .87 30.3–33.8 32.3 .86 30.5–34.0 29.1 5.2 18.6–39.6 o0.001 Other medication/drugs 33.3 .93 31.5–35.2 17.7 .06 17.6–17.8 15.7 .20 15.3–16.1 o0.001

1SE = standard error of the mean. 2The p-value is based on the t-test and tests for the association between the age of first reported drug use and DSM-5 CUD using CSGLM procedures and is done on the weighted data.

aData are weighted by NESARC-calculated AUDADIS full-sample weight, clustering for county and stratification using the stratum variable, adjusting standard errors. In all cases, complex sampling procedures were used.

†Refers to cases of current (previous 12 month) diagnosis of DSM-5 Cannabis use disorder as defined by the Alcohol Use Disorder and Associated Disabilities Interview Schedule-V (AUDADIS-V).

Table 3 Odds Ratios (OR) and 95% Confidence Intervals (CI) for the association between DSM-5 Cannabis Use Disorder (past year) and specific DSM-5 Substance Use Disorders, unadjusted and fully adjusted models (weighteda).

Variable Unadjusted Fully Adjustedb

B SE OR 95%CI p B SE OR 95%CI p

Sedative/Tranquiliser Use Disorder 2.6 .25 13.8 8.4–22.6 o0.001 1.1 .29 5.1 2.9–9.0 o0.001 Cocaine Use Disorder 3.1 .20 23.1 15.4–34.6 o0.001 2.2 .26 9.3 5.6–15.5 o0.001 Stimulant Use Disorder 3.0 .26 19.8 11.9–33.1 o0.001 1.5 .31 4.3 2.3–7.9 o0.001c

Club Drug Use Disorder 4.8 .35 117.4 58.3–236.5 o0.001 2.8 .47 16.1 6.3–40.8 o0.001 Hallucinogen Use Disorderd – – – – – – – – – –

Inhalant or Solvent Use Disorderd – – – – – – – – – –

Heroin Use Disorder 2.7 .51 16.0 5.9–43.6 o0.001 0.5 .52 1.6 0.6–4.6 0.33 Opioid Use Disorder 2.5 .17 12.7 9.1–17.8 o0.001 1.5 .20 4.6 3.0–6.8 o0.001 Alcohol Use Disorder 2.3 .07 10.1 8.7–11.7 o0.001 1.1 .10 3.0 2.5–3.7 o0.001 Other Drug/Medication Use Disorder 0.8 1.1 2.2 0.3–17.8 40.05 – – – – –

Note: No diagnosis of DSM-5 CUD was used as the reference outcome category for all models. aData are weighted by NESARC-calculated AUDADIS full-sample weight, clustering for county and stratification using the stratum

variable, adjusting standard errors- in all cases complex sampling procedures was used. bAdjusted for age, gender, education, marital status, urbanicity, ethnicity, alcohol use, smoking status and DSM-5 primary MDD,

PTSD, Bipolar I and panic disorder (omitted from table). cIndependent of ethnicity. dDue to unexpected singularities in the Hessian matrix, this data set failed to allow us to investigate the associations between DSM-5

CUD and both (i) Hallucinogen use disorder or (ii) Inhalant or solvent use disorder.

A.C. Hayley et al.738

drug/medication use disorder (unadjusted OR = 2.14, 95% CI: 0.3–17.8, p = .46) in the unadjusted model, and so no multivariable model was developed for this variable.

Following adjustment for all relevant covariates (age, gender, education, marital status, ethnicity, alcohol use,

smoking status and DSM-5 primary MDD, PTSD, Bipolar I disorder and panic disorder), the association between DSM-5 CUD and independent DSM-5 SUDs were retained, albeit tempered, in all models except for DSM-5 Heroin use disorder (p4 0.05).

739Comorbid cannabis use disorder and substance use disorders

Due to unexpected singularities in the Hessian matrix, statistical evaluations between DSM-5 CUD and both (i) Hallucinogen use disorder or (ii) Inhalant or solvent use disorder were not performed (including univariate or multi- variable models).

4. Discussion

Results from this large, representative study of adults indicate a close association between current (past 12- month) DSM-5 CUD diagnoses and concordant lifetime and current use of several illicit and prescription substances. Specifically, diagnosis of current DSM-5 CUD was found to be strongly associated with a greater incidence of lifetime use of all examined illicit and prescription drugs, as well as current (previous 12 month) use of several types of newer- class, stimulant-based substances including club drugs (e.g. MDMA, methamphetamine), cocaine and prescription stimu- lants (e.g. Ritalin). DSM-5 CUD was similarly linked to greater incidence of current drug use disorders for all of the examined drug classes, except for ‘other’ medication/ drug use disorder. Multivariable modelling confirmed the existence of a strong and independent association between current DSM-5 CUD and a number of DSM-5 SUD; with an overrepresentation of newer class legal and illicit stimulant- based substances. These finding were independent of a number of salient lifestyle, health and psychiatric factors.

Cannabis is rapidly becoming one of the most frequently abused substances worldwide, and rates of individuals seeking treatment for cannabis-related abuse disorders has increased substantially in the last decade (SAMHSA). Increased drug accessibility (Freisthler and Gruenewald, 2014), changes in decriminalisation and/or legalisation status (Miech et al., 2015) and increasingly positive public attitudes coupled with shifts in the perceived risk profile of cannabis are potential contributory factors in this rise in overall consumption rates (Berg et al., 2015). Of note, salient and pervasive sociodemographic factors such as impoverished environments, systemic disadvantage and psychiatric pathology similarly contribute to instances of cannabis use and abuse (Stinson et al., 2006), and thus this association cannot be attributed to regional availability or current drug status alone. Although complex, it is possible that a combination of these factors likely also contribute to observed increases in rates of DSM-5 CUD, given that approximately 1 in 10 lifetime cannabis users later become dependent on the drug (Compton et al., 2004; Looby and Earleywine, 2007). Age is an important etiological factor in determining the longer-term progression from drug use to dependence (Tarter et al., 1999). Rates of CUD observed in the current study were greatest among young males aged 18–34 years, and DSM-5 CUD diagnosed individuals reported first cannabis use to occur during mid-adolescence (mean age of first use was less than 16 years of age). Cannabis use early in life often predicts later comorbid use of other illicit substances (Lessem et al., 2006), and has peripheral implications for longer-term adverse psychiatric outcomes (Arseneault et al., 2002). This risk seems to be dose dependent (Moore et al., 2004) and particularly elevated among adolescents who initiate cannabis use before the age of 17 years (Lynskey et al., 2003). Indeed, young initiates

demonstrate an eighteen-fold increase in subsequent can- nabis dependence later in life (Silins et al., 2014). Adoles- cence is a critical transitional period of growth and development; and is a developmental stage which is con- sidered particularly sensitive to the influence of drug use (Crews et al., 2007). Consequently, there is an acute need for comprehensive educative programs which aim to pro- mote adolescent development and adjustment through these critical periods in order to reduce the likelihood of concurrent drug use and thus the potential for poly-drug use and/or abuse later in life.

Misuse or abuse of any psychoactive substances is asso- ciated with a marked increase in the probability of similarly abusing other categories of drugs (Tsuang et al., 1998). As described here, a diagnosis of DSM-5 CUD was associated with significantly greater lifetime use of any other drug class as well as greater current (previous 12-month) use of several prescription and illicit substances. Notably, these comorbid associations for current (previous 12 month) drug use were strongest for newer-class, stimulant type drugs, such as club drugs (e.g. MDMA, methamphetamine), cocaine, legal stimulants and hallucinogens (e.g. LSD). Cannabis use is frequently linked to greater use of a select range of drugs of abuse, including sedatives (such as benzodiazepines) and stimulants such as cocaine (Degenhardt et al., 2001; Kendler et al., 2003). No studies have yet described this association among those with defined DSM-5 CUD (i.e. when the use extends to abuse). To our knowledge, only one other study has explicitly assessed associations between cannabis use and comorbid use of a comparably diverse range of other specific drug types as those examined here (Fergusson et al., 2006); however, this observation was confined to drug frequency only (i.e. did not extend to diagnoses of DSM-5 CUD). Indeed, many studies looking into poly-drug use among CUD patients utilise aggregated drug use indicators (Degenhardt et al., 2001) (even when DSM-5 criteria is utilised) (see; Grant et al. (2016)), examine a relatively limited range of specific drug types (Degenhardt et al., 2001), or are assessed in the context of alcohol use only (Stinson et al., 2005). Mechanistically, these close associa- tions reported here may reflect increases in regional avail- ability of these types of drugs among users (Bramness et al., 2015), greater abuse rates of these types of drugs in younger adults (i.e. reflective of the sample) (McCabe et al., 2005), greater availability of these types of drugs (Kaye and Darke, 2012), and/or be due to the desirability of opposing pharmacokinetic profiles to offset the stimulating effects of these types of drugs (such as using cannabis to help the ‘come-down’ following heavy amphetamine use). Of consideration, it is similarly probable that a complex system of underlying social, economic and personal factors contribute to co-abuse of these drugs. However, further research is urgently warranted to adequately describe these vulnerabilities, particularly where concurrent use of canna- bis and newer-class substances is evident. Irrespective of the etiological origin, these data highlights the high degree of wide-ranging drug use cross-over and poly-drug use among those with DSM-5 CUD.

Despite sharp increases in individuals seeking treatment for CUD in recent years (Compton et al., 2004) and greater prevalence of comorbid drug abuse behaviour patterns

A.C. Hayley et al.740

noted among inpatient facilities for those seeking treatment for DSM classified CUD (Copeland et al., 2001a), systematic evaluations of comorbid DSM-5 CUD and other specific DSM-defined SUD are markedly absent. We report a strong and independent association between DSM-5 CUD and nearly all examined specific DSM-5 SUD. As previously highlighted, existing literature has indicated an association between cannabis use and the comorbid use of sedating type drugs (such as benzodiazepines), and cocaine (Degenhardt et al., 2001; Kendler et al., 2003). However, no research has yet described these associations among the mutually classified diagnostic abuse category (i.e. specific SUD) for any drug class. Multivariable modelling revealed that those with DSM- 5 CUD reported a 16-fold increased risk of similarly record- ing DSM-5 Club drug use disorder (i.e. MDMA, methamphe- tamine). In terms of strength of associations, these were followed by cocaine use disorder, sedative use disorder, opioid use disorder and stimulant use disorder (Adjusted OR = 9.3, 5.1, 4.6 and 4. 3, respectively). Several national monitoring systems indicate an overall rise in the incidence of multidrug use of ‘club drugs’ (i.e. MDMA, methampheta- mine) among young adults (Gross et al., 2002; Wu et al., 2006), and a marked increase in young adults using pre- scription stimulants for recreational purposes (Arria and DuPont, 2010). Frequency and chronicity of prior drug use is associated with a relative increased risk of becoming dependent on that substance (Chen et al., 1997). Thus, it is possible that the observed association between CUD and both club-drug specific disorders and prescription stimulant- type disorders reflects a natural extension of these mala- daptive patterns of use as they transcend into more defined and entrenched abuse-type behaviours. This is of particular health concern, given the potential for greatly amplified psychiatric effects of these types of drugs, and the possi- bility of more rapid onset of when they are used in combination (Dalmau et al., 1999). High rates of concor- dance between DSM-5 CUD and both sedative/tranquiliser and opioid use disorder may similarly reflect an emergent trend of prescription-type medication abuse among young adults (Sung et al., 2005). Indeed, such observations some- what mirror previous observations of a close association between cannabis use and use of other sedative-type disorders; however provides an updated context for pre- ferred drug type (i.e. movement away from drugs such as heroin or traditional benzodiazepines for newer-class seda- tives such as Xanax). Use of these drugs in combination significantly increases the likelihood of morbidity (Subramaniam et al., 2010), and thus extra vigilance and urgency is required when evaluating comorbid use of these types of drugs among those seeking treatment.

The evident and potential health implications for CUD is at odds with its historical standing as a typically benign substance, and this has been significantly compounded by the current lack of effective and tailored treatment ave- nues (Dennis et al., 2002). No evidence-based pharma- cotherapies are currently available neither for the management of cannabis withdrawal and cravings, nor for complex presentations where comorbid drug abuse is evi- dent. Rather, the implementation of brief cognitive beha- vioural therapy and contingency management have indicated the strongest evidence for success (Copeland and Swift, 2009); however even these are shown to have

mixed efficacy and/or poorly defined treatment endpoints (Hser et al., 2001), and their effectiveness among complex cross-diagnoses is largely unclear. At present, the most widely utilised programs aim to provide an emphasis on traditional stress-reduction strategies to help with anxiety and emotional relief efficacy, educational strategies aimed at promoting opportunistic efficacy (to help identify the reason behind the poly-substance use), and acquiring skills to promote cannabis cessation and maintenance of absti- nence; yet these too are limited in their long-term effec- tiveness. Additional research is thus urgently needed to better define treatment models directly targeted at mana- ging cannabis use disorder, and how these might then extend to assist in treating complex presentations of the diagnoses where polydrug use is noted. Further assessment of the efficacy of early intervention protocols are similarly recommended, given the established link between age of use and progression into patterns of drug abuse behaviours.

Results from the current study must be considered in light of some methodological limitations. Firstly, the pre- sent work is a cross-sectional, not prospective study, and thus we are unable to comment on the direction or trajectories of these associations. However, given that DSM-5 CUD diagnosed individuals reported use of cannabis at an earlier age than other drug use, we tentatively suggest that this may further be reflected for diagnosed SUDs whereby cases of DSM-5 CUD precede the develop- ment of other SUDs. The nature of the survey design (i.e. many face to face interviews) may impact responses in relation to prior and current drug use (e.g., over/under- reporting), due to perceived issues of legality and reluc- tance to disclosing sensitive information. Despite this, many similar surveys have indicated good compliance with these types of questionnaire items in general population samples (Hser, 1997). Validity could be improved using biological screening measures (i.e. blood or urine), how- ever these assessments have a narrow window of detect- ability and are often cumbersome and expensive; which greatly reduces their usefulness for population-based research (Fendrich et al., 2004). Lastly, due to singularities in the dataset for two outcomes (hallucinogen-use disorder and inhalant/solvent use disorder), we were unable to compute multivariable models for these outcomes. Mitiga- tion of quasi or complete separation of data typically includes: (i) omission of the offending variable or (ii) imputation methods (Demirtas and Hedeker, 2008). Neither of these options are entirely desirable due to the risk of deliberate specification bias or increased risk of incor- rectly inflating missing binary variables (Enders, 2010). Thus, in order to retain integrity of the presented data, we omitted these from the main multivariable findings, and acknowledge that we are unable to comment on these associations in any great detail. Additional and targeted research is therefore urgently warranted in order to elucidate the more subtle aspects of this particular relationship among these specific drug types.

4.1. Conclusion

These findings provide the first evaluation as to the strength of associations between DSM-5 CUD, concurrent lifetime and

741Comorbid cannabis use disorder and substance use disorders

current (previous 12 month) substance use and other specific DSM-5 SUDs. Relationships were particularly salient for young people, and across several newer-class illicit and prescription stimulant-based drug types. Potential for greater abuse of other substances should thus be considered for people who present with problematic patterns of cannabis use and dependence, and tailored treatment programs are urgently required to address this growing problem.

Role of the funding source

No study sponsors were involved in the study design; collection, analysis and interpretation of data; the writing of the manuscript; or in the decision to submit the manu- script for publication.

Author contribution

ACH, LAD and CS were involved in the development and design of the study. ACH collated the data, and ACH, LAD and CS constructed variables for this paper and analysed the data. ACH interpreted the data and wrote the manuscript. ACH, LAD and CS were involved in drafting, editing and critical appraisal of the manuscript. All authors have approved the manuscript for submission.

Conflict of interest declaration

All authors declare that they have no conflicts of interest.

Acknowledgements

This manuscript was prepared using a limited access dataset obtained from the National institute of Alcohol and Alcoholism (NIAAA) and does not reflect the opinions of the NIAAA or the U.S Government.

References

Agrawal, A., Neale, M.C., Prescott, C.A., Kendler, K., 2004. A twin study of early cannabis use and subsequent use and abuse/ dependence of other illicit drugs. Psychol. Med. 34, 1227–1237.

American Psychiatric Association (APA), 1994. Diagnostic and Sta- tistical Manual of Mental Disorders, Fourth Edition, Washington.

American Psychiatric Association (APA), 2013. Diagnostic and Sta- tistical Manual of Mental Disorders, Fifth edition, Washington.

Arria, A.M., DuPont, R.L., 2010. Nonmedical prescription stimulant use among college students: why we need to do something and what we need to do. J. Addict. Dis. 29, 417–426.

Arseneault, L., Cannon, M., Poulton, R., Murray, R., Caspi, A., Moffitt, T.E., 2002. Cannabis use in adolescence and risk for adult psychosis: longitudinal prospective study. BMJ 325, 1212–1213.

Belendiuk, K., Black, T., Thurstone, C., 2015. The effect of marijuana commercialization on adolescent polysubstance use and treatment. Drug Alcohol Depend. 146, e221.

Berg, C.J., Stratton, E., Schauer, G.L., Lewis, M., Wang, Y., Windle, M., Kegler, M., 2015. Perceived harm, addictiveness, and social acceptability of tobacco products and Marijuana among young

adults: Marijuana, Hookah, and Electronic Cigarettes Win. Subst. Use Misuse 50, 79–89.

Bramness, J.G., Reid, M.J., Solvik, K.F., Vindenes, V., 2015. Recent trends in the availability and use of amphetamine and metham- phetamine in Norway. Forensic Sci. Int. 246, 92–97.

Brook, D.W., Brook, J.S., Zhang, C., Cohen, P., Whiteman, M., 2002. DRug use and the risk of major depressive disorder, alcohol dependence, and substance use disorders. Arch. Gen. Psychiatry 59, 1039–1044.

Bureau of the Census, 2012. American Community Survey Data. Suitland, Maryland.

Chen, C.-Y., O’Brien, M.S., Anthony, J.C., 2005. Who becomes cannabis dependent soon after onset of use? Epidemiological evidence from the United States: 2000–2001. Drug. Alcohol Depend. 79, 11–22.

Chen, K., Kandel, D.B., Davies, M., 1997. Relationships between frequency and quantity of marijuana use and last year proxy dependence among adolescents and adults in the United States. Drug Alcohol Depend. 46, 53–67.

Compton, W.M., Grant, B.F., Colliver, J.D., Glantz, M.D., Stinson, F. S., 2004. Prevalence of marijuana use disorders in the united states: 1991–1992 and 2001–2002. JAMA 291, 2114–2121.

Connor, J., Gullo, M., Chan, G., Young, R., Hall, W., Feeney, G., 2013. Polysubstance Use in Cannabis users referred for treat- ment: drug use profiles, psychiatric comorbidity and Cannabis- related beliefs. Front. Psychiatry 4, 79.

Copeland, J., Swift, W., 2009. Cannabis use disorder: epidemiology and management. Int. Rev. Psychiatry 21, 96–103.

Copeland, J., Swift, W., Rees, V., 2001a. Clinical profile of participants in a brief intervention program for cannabis use disorder. J. Subst. Abuse Treat. 20, 45–52.

Copeland, J., Swift, W., Roffman, R., Stephens, R., 2001b. A randomized controlled trial of brief cognitive–behavioral inter- ventions for cannabis use disorder. J. Subst. Abuse Treat. 21, 55–64.

Crews, F., He, J., Hodge, C., 2007. Adolescent cortical develop- ment: a critical period of vulnerability for addiction. Pharmacol. Biochem. Behav. 86, 189–199.

Dalmau, A., Bergman, B., Brismar, B., 1999. Psychotic disorders among inpatients with abuse of cannabis, amphetamine and opiates. Do dopaminergic stimulants facilitate psychiatric ill- ness? Eur. Psychiatry 14, 366–371.

Degenhardt, L., Hall, W., 2012. Extent of illicit drug use and dependence, and their contribution to the global burden of disease. Lancet 379, 55–70.

Degenhardt, L., Hall, W., Lynskey, M., 2001. The relationship between cannabis use and other substance use in the general population. Drug Alcohol Depend. 64, 319–327.

Degenhardt, L., Coffey, C., Carlin, J.B., Swift, W., Patton, G.C., 2008. Are diagnostic orphans at risk of developing cannabis abuse or dependence?: four-year follow-up of young adult cannabis users not meeting diagnostic criteria. Drug Alcohol Depend. 92, 86–90.

Degenhardt, L., Ferrari, A.J., Calabria, B., Hall, W.D., Norman, R. E., McGrath, J., Flaxman, A.D., Engell, R.E., Freedman, G.D., Whiteford, H.A., Vos, T., 2013. The global epidemiology and contribution of Cannabis use and dependence to the global burden of disease: results from the GBD 2010 Study. PLoS One 8, e76635.

Demirtas, H., Hedeker, D., 2008. An imputation strategy for incomplete longitudinal ordinal data. Stat. Med. 27, 4086–4093.

Dennis, M., Babor, T.F., Roebuck, M.C., Donaldon, J., 2002. Chan- ging the focus: the case for recognizing and treating cannabis use disorders. Addict 97, 4–15.

Enders, C.K., 2010. Applied Missing Data Analysis. Guilford Press, New York, NY, US.

A.C. Hayley et al.742

Farmer, R.F., Kosty, D.B., Seeley, J.R., Duncan, S.C., Lynskey, M.T., Rohde, P., Klein, D.N., Lewinsohn, P.M., 2015. Natural course of Cannabis use disorders. Psychol. Med. 45, 63–72.

Fendrich, M., Johnson, T.P., Wislar, J.S., Hubbell, A., Spiehler, V., 2004. The utility of drug testing in epidemiological research: results from a general population survey. Addict 99, 197–208.

Fergusson, D.M., Horwood, L.J., 2000. Does cannabis use encourage other forms of illicit drug use? Addict 95, 505–520.

Fergusson, D.M., Boden, J.M., Horwood, L.J., 2006. Cannabis use and other illicit drug use: testing the cannabis gateway hypoth- esis. Addict 101, 556–569.

Freisthler, B., Gruenewald, P.J., 2014. Examining the relationship between the physical availability of medical marijuana and marijuana use across fifty California cities. Drug Alcohol Depend. 143, 244–250.

Grant, B.F., Pickering, R., 1998. The relationship between cannabis use and DSM-IV cannabis abuse and dependence: results from the national longitudinal alcohol epidemiologic survey. J. Subst. Abuse 10, 255–264.

Grant, B.F., Amsbary, M., Chu, A., Sigman, R., Kali, J., Sugawana, Y., Jiao, R., Goldstein, R.B., Jung, J., Zhang, H., Chou, P.S., 2014a. Source and Accuracy Statement: National Epidemiologic Survey on Alcohol and Related Conditions-III (NESARC-III). National Institute on Alcohol Abuse and Alcoholism, Rockville, MD.

Grant, B.F., Chu, A., Sigman, R., Amsbary, M., Kali, J., Sugawara, Y., Jiao, R., Ren, W., Goldstein, R., 2014b. National Institute on Alcohol Abuse and Alcoholism National Epidemiologic Survey on Alcohol and Related Conditions-III (NESARC- III) Source and Accuracy Statement, Rockville, MD.

Grant, B.F., Goldstein, R.B., Smith, S.M., Jung, J., Zhang, H., Chou, S.P., Pickering, R.P., Ruan, W.J., Huang, B., Saha, T.D., Aivad- yan, C., Greenstein, E., Hasin, D.S., 2015. The Alcohol Use Disorder and Associated Disabilities Interview Schedule-5 (AUDA- DIS-5): reliability of substance use and psychiatric disorder modules in a general population sample. Drug Alcohol Depend. 148, 27–33.

Grant, B.F., Saha, T.D., Ruan, W., et al., 2016. Epidemiology of dsm- 5 drug use disorder: results from the national epidemiologic survey on alcohol and related conditions–III. JAMA Psychiatry 73, 39–47.

Gross, S.R., Barrett, S.P., Shestowsky, J.S., Pihl, R.O., 2002. Ecstasy and drug consumption patterns: a Canadian Rave population study. Can. J. Psychiatry 47, 546–551.

Hasin, D.S., Kerridge, B.T., Saha, T.D., Huang, B., Pickering, R., Smith, S.M., Jung, J., Zhang, H., Grant, B.F., 2016. Prevalence and correlates of DSM-5 Cannabis use disorder, 2012–2013: findings from the National Epidemiologic Survey on Alcohol and Related Conditions–III. Am. J. Psychiatry 173, 588–599.

Hser, Y.I., 1997. Self-reported drug use: results of selected empiri- cal investigations of validity. NIDA Res. Monogr. 167, 320–343.

Hser, Y.-I., Grella, C.E., Hubbard, R.L., Hsieh, S.-C., Fletcher, B.W., Brown, B.S., Anglin, M.D., 2001. An evaluation of drug treat- ments for adolescents in 4 US cities. Arch. Gen. Psychiatry 58, 689–695.

Kaye, S., Darke, S., 2012. The diversion and misuse of pharmaceu- tical stimulants: what do we know and why should we care? Addict 107, 467–477.

Kendler, K.S., Jacobson, K.C., Prescott, C.A., Neale, M.C., 2003. Specificity of genetic and environmental risk factors for use and abuse/dependence of Cannabis, Cocaine, Hallucinogens, Seda- tives, Stimulants, and Opiates in Male Twins. Am. J. Psychiatry 160, 687–695.

Kerr, T., Fairbairn, N., Tyndall, M., Marsh, D., Li, K., Montaner, J., Wood, E., 2007. Predictors of non-fatal overdose among a cohort of polysubstance-using injection drug users. Drug Alcohol Depend. 87, 39–45.

Lessem, J.M., Hopfer, C.J., Haberstick, B.C., Timberlake, D., Ehringer, M.A., Smolen, A., Hewitt, J.K., 2006. Relationship

between adolescent Marijuana use and young adult illicit drug use. Behav. Gen. 36, 498–506.

Looby, A., Earleywine, M., 2007. Negative consequences associated with dependence in daily cannabis users. Subst. Abuse Treat. Prev. Policy 2, 1–7.

Low, K.G., Gendaszek, A.E., 2002. Illicit use of psychostimulants among college students: a preliminary study. Psychol. Health Med. 7, 283–287.

Lynskey, M.T., Heath, A.C., Bucholz, K.K., et al., 2003. Escalation of drug use in early-onset cannabis users vs co-twin controls. JAMA 289, 427–433.

McCabe, S.E., Knight, J.R., Teter, C.J., Wechsler, H., 2005. Non- medical use of prescription stimulants among US college stu- dents: prevalence and correlates from a national survey. Addict 100, 96–106.

Miech, R.A., Johnston, L., O’Malley, P.M., Bachman, J.G., Schulen- berg, J., Patrick, M.E., 2015. Trends in use of marijuana and attitudes toward marijuana among youth before and after decriminalization: the case of California 2007–2013. Int. J. Drug Policy 26, 336–344.

Moore, B.A., Augustson, E.M., Moser, R.P., Budney, A.J., 2004. Respiratory effects of Marijuana and tobacco use in a U.S. Sample. J. Gen. Intern. Med., 20.

Mowbray, C.T., Ribisl, K.M., Solomon, M., Luke, D.A., Kewson, T.P., 1997. Characteristics of dual diagnosis patients admitted to an urban, public psychiatric hospital: an examination of individual, social, and community domains. Am. J. Drug Alcohol Abuse 23, 309–326.

National Institute of Alcohol Abuse and Alcoholism (NIAAA), 2014. NESARC-III Data Notes.

Pacula, R.L., Smart, R., 2017. Effects of changes in marijuana laws on marijuana use and disorders: medical marijuana and mar- ijuana legalization. Ann. Rev. Clin. Psychol. 13, 397–419.

Ridenour, T.A., Cottler, L.B., Compton, W.M., Spitznagel, E.L., Cunningham-Williams, R.M., 2003. Is there a progression from abuse disorders to dependence disorders? Addict 98, 635–644.

Schuermeyer, J., Salomonsen-Sautel, S., Price, R.K., Balan, S., Thurstone, C., Min, S.-J., Sakai, J.T., 2014. Temporal trends in marijuana attitudes, availability and use in Colorado compared to non-medical marijuana states: 2003-11. Drug Alcohol Depend. 140, 145–155.

Silins, E., Horwood, L.J., Patton, G.C., Fergusson, D.M., Olsson, C. A., Hutchinson, D.M., Spry, E., Toumbourou, J.W., Degenhardt, L., Swift, W., Coffey, C., Tait, R.J., Letcher, P., Copeland, J., Mattick, R.P., 2014. Young adult sequelae of adolescent cannabis use: an integrative analysis. Lancet Psychiatry 1, 286–293.

SPSS, 2012. IBM SPSS statistics version 21. International Business Machines Corp, Boston, MA.

Stephens, R.S., Roffman, R.A., Simpson, E.E., 1993. Adult mari- juana users seeking treatment. J. Consult. Clin. Psychol. 61, 1100–1104.

Stewart, D., Gossop, M., Marsden, J., 2002. Reductions in non-fatal overdose after drug misuse treatment: results from the National Treatment Outcome Research Study (NTORS). J. Subst. Abuse Treat. 22, 1–9.

Stinson, F.S., Grant, B.F., Dawson, D.A., Ruan, W.J., Huang, B., Saha, T., 2005. Comorbidity between DSM-IV alcohol and specific drug use disorders in the United States: results from the National Epidemiologic Survey on Alcohol and Related Condi- tions. Drug Alcohol Depend. 80, 105–116.

Stinson, F.S., Ruan, W.J., Pickering, R., Grant, B.F., 2006. Cannabis use disorders in the USA: prevalence, correlates and co- morbidity. Psychol. Med. 36, 1447–1460.

Subramaniam, G.A., Ives, M.L., Stitzer, M.L., Dennis, M.L., 2010. The added risk of opioid problem use among treatment-seeking youth with Marijuana and/or alcohol problem use. Addict 105, 686–698.

743Comorbid cannabis use disorder and substance use disorders

Substance Abuse and Mental health Services Admisntration (SAMHSA). Substance Abuse Facilities Data. ⟨https://www. samhsa.gov/data/substance-abuse-facilities-data-nssats⟩.

Sung, H.-E., Richter, L., Vaughan, R., Johnson, P.B., Thom, B., 2005. Nonmedical use of prescription opioids among teenagers in the United States: trends and correlates. J. Adolesc. Health 37, 44–51.

Tarter, R., Vanyukov, M., Giancola, P., Dawes, M., Blackson, T., Mezzich, A.D.A., Clark, D.B., 1999. Etiology of early age onset substance use disorder: a maturational perspective. Dev. Psy- chopathol. 11, 657–683.

Tsuang, M.T., Lyons, M.J., Meyer, J.M., et al., 1998. Co-occurrence of abuse ofy different drugs in men: the role of drug-specific and shared vulnerabilities. Arch. Gen. Psychiatry 55, 967–972.

United Nations Office of Drug Control and Crime Prevention

(UNODC), 2016. World Drug Report. Oxford University Press;

Oxford New York, 43-51. Available at: ⟨http://www.unodc.org/

doc/wdr2016/WORLD_DRUG_REPORT_2016_web.pdf⟩. Wagner, F.A., Anthony, J.C., 2002. From first drug use to drug

dependence: developmental periods of risk for dependence

upon marijuana, cocaine, and alcohol. Neuropsychopharmacol

26, 479–488. Wu, L.-T., Schlenger, W.E., Galvin, D.M., 2006. Concurrent Use of

Methamphetamine, MDMA, LSD, Ketamine, GHB, and Flunitra-

zepam among American Youths. Drug Alcohol Depend. 84,

102–113.