Cannabis withdrawal severity and short-term course among cannabis-dependent adolescent and young adult inpatients

Drug and Alcohol Dependence 106 (2010) 133–141

Contents lists available at ScienceDirect

Drug and Alcohol Dependence journal homepage: www.elsevier.com/locate/drugalcdep

Cannabis withdrawal severity and short-term course among cannabis-dependent adolescent and young adult inpatients U.W. Preuss a,c,∗,1 , A.B. Watzke c,1 , J. Zimmermann a , J.W.M. Wong d , C.O. Schmidt b a

Martin-Luther University, Halle-Wittenberg (MLU), Department of Psychiatry, Psychotherapie and Psychosomatics, Julius-Kühn-Straße 7, 06097 Halle/Saale, Germany Institute for Community Medicine, University of Greifswald, Walther Rathenau Strasse 48, 17475 Greifswald, Germany Fachklinik Gristower Wiek, Riemser Weg 19, 17498 Gristow, Germany d AWO Psychiatry Centre Halle, Zscherbener Strasse 11, 06114 Halle/Saale, Germany b c

a r t i c l e

i n f o

Article history: Received 22 December 2008 Received in revised form 5 August 2009 Accepted 7 August 2009 Available online 23 September 2009 Keywords: Cannabis withdrawal Cannabis dependence Inpatients Latent class models Personality

a b s t r a c t Objective: While previous studies questioned the existence of a cannabis withdrawal syndrome (CWS), recent research provided increasing evidence of a number of clinical symptoms after cessation of frequent cannabis consumption. The aim of this study is to prospectively assess the course of CWS in a sample of cannabis-dependent inpatients and to provide an estimate of the proportion of subjects experiencing CWS. Methods: 118 subjects, aged 16–36 years, diagnosed with a cannabis dependence (DSM-IV, assessed by SCID I) were enrolled in the study. CWS was assessed prospectively over 10 days using a modified version of the Marijuana Withdrawal Checklist. Personality dimensions were assessed with the NEO-FFI. Results: 73 subjects (61.3%) completed all assessments over the observation period. Most symptoms peaked on day 1. Model-based analyses revealed a high and low intensity CWS group. Less than half of the patients belonged to the high intensity craving, psychological, or physical withdrawal symptoms group. Symptom intensity decreased almost linearly over time. Indicators of cannabis consumption intensity as well as personality dimensions, but not recalled withdrawal were related to membership in the high intensity CWS group. Discussion: A clinically relevant CWS may only be expected in a subgroup of cannabis-dependent patients. Most subjects with an elevated CWS experience physical and psychological symptoms. The small to negligible associations between recalled and prospectively assessed symptoms raise questions about the validity of the former approach. © 2009 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Cannabis products are the most frequently consumed illicit substances in Western countries. It is estimated that about 65 million European adults have tried this substance at least once (EMCDDA, 2006). Around 5–15% of all lifetime consumers subsequently develop a substance use disorder (Perkonigg et al., 1999; Wittchen et al., 2007), 1–7% suffer from cannabis dependence (Chen et al., 2005; Perkonigg et al., 2008). One important criterion for substance dependence according to ICD10 and DSM-IV is a withdrawal state when substance use is reduced or ceased. This is evidenced by the characteristic withdrawal syndrome for the substance, or by use of the same (or closely

∗ Corresponding author at: Klinik für Psychiatrie, Psychotherapie und Psychosomatik der, Martin-Luther-Universität Halle, Julius-Kühn-Straße 7, 06097 Halle/Saale, Germany. Tel.: +49 345 557 4595; fax: +49 345 557 3500. E-mail address: [email protected] (U.W. Preuss). 1 Equally contributed. 0376-8716/$ – see front matter © 2009 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.drugalcdep.2009.08.008

related) substance with the intention of relieving or avoiding withdrawal symptoms. While the withdrawal syndromes of other legal and illegal substances, including cocaine, opiates, alcohol, or even nicotine are well documented, a cannabis withdrawal syndrome (CWS) is mentioned but poorly defined. Correspondingly, diagnostic criteria are not well established (Budney et al., 2004). However, findings from experimental and observational research with animals and human subjects support a CWS comprising psychological and physiological symptoms (Budney et al., 2004; Budney and Hughes, 2006). Discovery of the cannabinoid CB1 receptor and the development of a cannabinoid antagonist (SR141617A, rimonabant) resulted in studies which precipitated CWS in mice, rats, and dogs (Justinova et al., 2005). Commonly observed symptoms in humans include anger, aggression, anxiety, decreased appetite, weight loss, irritability, restlessness, and sleep difficulty (e.g. Budney et al., 2004). Depressed mood, stomach pain, physical discomfort, tremor, and sweating have also been reported, but occur less frequently. Most symptoms begin within 24 h of abstinence, peak before day 3 after cessation of cannabis consumption, and last approximately 1–2

134

U.W. Preuss et al. / Drug and Alcohol Dependence 106 (2010) 133–141

weeks (Budney et al., 2003, 2004; Kouri and Pope, 2000). Regarding the frequency of symptoms, an early inpatient study of habitual cannabis users reported that 55–89% experienced irritability, restlessness, insomnia, or anorexia after discontinuation of oral THC (Tetrahydrocannabinol, Jones et al., 1976). Based on the Marijuana Withdrawal Checklist, MWC (Budney et al., 1999), a more recent study of adolescent frequent consumers reported a clinically relevant CWS with four or more severe symptoms in only 10% of the subjects whereas 58% reported at least four symptoms with mild intensity (Vandrey et al., 2005). In an adult outpatient sample comprising cannabis-dependent subjects, 85% reported four or more symptoms of any intensity and 33% reported four or more severe symptoms (Budney et al., 1999). Based on other assessment instruments lower proportions were reported from epidemiologic studies (Agrawal et al., 2008; Hasin et al., 2008). Comparing subjects within and across these studies suggests that withdrawal severity is greater among more frequent cannabis users. Consumption frequency as well as CWS intensity are positively associated with comorbid psychiatric symptoms and personality traits in adults as well as in adolescents (Budney et al., 1999; Barnes et al., 2005; Chabrol et al., 2004; Cornelius et al., 2008; Hasin et al., 2008). While the existence of a CWS is supported in the literature, several shortcomings need to be addressed: Observational research is mostly based on retrospective assessments of CWS which is subject to recall bias and fails to provide valid information on the symptom course (for example Budney et al., 1999; Cornelius et al., 2008; Hasin et al., 2008; Vandrey et al., 2005; Wiesbeck et al., 1996). Data on the latter is provided from experimental research but these samples rarely comprised dependent consumers according to DSM-IV criteria (Budney et al., 2004). Comparatively little observational research is based on samples with a diagnosed cannabis dependence in all study participants (for example Budney et al., 1999; Arendt et al., 2007). Less evidence is available from adolescents or young adult samples. However, these results are essentially similar (Cornelius et al., 2008; Nocon et al., 2006; Vandrey et al., 2005). Results from a prospective CWS assessment in a small sample of cannabis-dependent adolescent outpatients have recently been made available (Milin et al., 2008). In their study, a 1-week interval between each of their four assessments may be too long to appropriately assess symptom course after cessation of consumption. Retrospective assessments of withdrawal symptoms suggest the importance of subgrouping subjects with a high and low or no CWS. However, there is currently no evidence available which is based on a prospective assessment of physical and psychological symptoms during cannabis withdrawal. Therefore, this study prospectively assesses the course of CWS among cannabisdependent subjects seeking detoxification, and aims to classify subjects according to their reported symptoms. It aims to provide an estimate of the proportion of dependent subjects experiencing a CWS without the threat of bias typically associated with long-term recall periods. Specifically, this paper has four aims: First, to assess the characteristics, severity and course of CWS over a period of 10 days among in cannabis-dependent inpatients. We hypothesized an initial increase of symptom intensity during the first days of the stay followed by a subsequent decrease. Second, we analysed the necessity to distinguish between subgroups with a characteristically different CWS course. We expected at least two subgroups, the first of which should comprise subjects with moderate-to-strong symptoms and the second subjects with no or only mild symptoms. Third, we assessed potential predictors of CWS. As such, we focused on variables related to drug consumption and personality. We assumed that subjects with more intense cannabis consumption and more severe cannabis dependence should experience a stronger CWS. Concerning the hypothesis on symptom course, we assumed that subjects who had their last dose almost immediately prior to admission are in an earlier stage of CWS and should experi-

ence milder symptoms at the beginning of their stay. However, this is not expected to influence the overall intensity of CWS. Fourth, we studied the association between retrospective reports of a CWS and the prospectively assessed CWS. We hypothesized a positive association between these assessments. 2. Methods 2.1. Sample, in- and exclusion criteria Inpatients meeting the inclusion criteria as described below were consecutively enrolled into the study at the beginning of their detoxification in a closed addiction treatment ward at the Johanna-Odebrecht-Foundation psychiatric hospital, Greifswald, Germany. All patients were seeking planned detoxification and were admitted voluntarily to the ward for detoxification treatment for a time period of 5–10 days. None of the participants was volunteering for an inpatient stay as part of a non-routine hospitalization or was admitted for an emergency or unplanned hospitalization. Inpatient treatment for the purpose of drug detoxification is commonly offered by German psychiatric hospitals; but cannabis dependence alone is not an indication for admission to a detoxification treatment in all regions and institutions. Greifswald is a small town in the least populated state of Germany with comparatively few individuals addicted to illicit drugs such as opioids. For this reason adolescents or young adults with a shorter history of abuse without concomitant DSM-IV Axis I disorder were admitted. Inclusion criteria were a diagnosed cannabis dependence according to DSM-IV, without concomitant alcohol- or other (multiple) substance dependence. Subjects were excluded from the study if they tested positive for any other drug other than cannabis on the day of admission to avoid confounding by multiple substance abuse. Furthermore, subjects with other current DSM-IV Axis I diagnosis, or severe somatic or neurological disorders were transferred to another psychiatric ward and excluded from the study. Patients agreed to non-psychopharmacological treatment during the study period. They had free and unlimited access to coffee and cigarettes during the hospitalization. In total 118 adolescents and young adults met the inclusion criteria. The sample comprised 101 males and 17 females with an average age of 19.6 ± 2.9 years (range 16–36 years, two subjects being older than 25 years). During the 10-day observation period drop-out occurred (Table 1). All drop-outs (except for one case) were attributed to discharge when no more cannabinoids were detectable in the urine drug screening from day 4 onwards. Patients were discharged because healthcare providers will cover detoxification treatment only as long as the substance is detectable. One person did not provide information on symptom intensity and was excluded from the analyses. To assess the sensitivity of our results to attrition we computed the models as described in Section 2.5 on the first 6 days with more than 80% complete cases. These results revealed no relevant differences compared to an analysis based on all measurement points. Sensitivity analysis with the number of valid measurement points per person as predictor was not associated with inclusion in the high versus low withdrawal symptoms group as described in results (physical: odds ratio 1.0 (95% CI 0.8–1.3); psychological: odds ratio 0.9 (95% CI 0.7–1.2); craving: odds ratio 1.1 (95% CI 0.9–1.5)).

2.2. Instruments A modified German version of the Marijuana Withdrawal Checklist (MWC, Budney et al., 1999) was used to measure withdrawal symptoms, and is further described in the following section. History of cannabis consumption, criteria for cannabis dependence and patient’s psychosocial characteristics were assessed using the German Version of the SCID I (Wittchen et al., 1997). As part of the SCID I interview, cannabis withdrawal was assessed retrospectively. This comprised the general previous experience of a CWS, and a subsequent open-ended question asking for the recall of specific single withdrawal symptoms. Psychosocial characteristics concerning legal history, family background, and treatment history were assessed with a semi-structured interview as recommended by the German Society for Addiction Research and Addiction Treatment (DGS, 2001). Study participants were interviewed on their history of cannabis use. This encompassed an open-ended question concerning the last amount of cannabis consumed, and a second open-ended question concerning the time of last consumption before admission. On the day of admission, inpatient’s urine samples were taken to assess the baseline THC level and to exclude other recent or current substance use, including opioids, cocaine, benzodiazepines, ecstasy, and amphetamines. Moreover, breath alcohol level was measured. Based on the THC urine sample at admission, subjects were subgrouped into “low” and “high” THC level group, using the median as cut-off point. All subjects in the study consumed cannabis daily or almost daily. To control for concomitant nicotine consumption, cigarette smoking was assessed over the course of the study. A validated German translation of the NeoFive Factor Inventory (Costa and McCrae, 1992) was used to assess personality dimensions (neuroticism, extraversion, openness, agreeableness and conscientiousness). All instruments that were employed repeatedly are presented in Table 1. The remaining instruments were administered during the first 2 days of the stay.

U.W. Preuss et al. / Drug and Alcohol Dependence 106 (2010) 133–141

135

Table 1 Study flow-chart of repeated measures. Instruments

Day 1, admission

Day 2

Day 3

Day 4

Day 5

Day 6

Marijuana withdrawal checklist (MWC) Cigarette smoking Drug urine test Number of subjects in study

× × × 118

× ×

× ×

× ×

118

118

× × × 115

× × × 100

2.3. Assessment of cannabis withdrawal The MWC lists common as well as less frequently observed withdrawal symptoms. Participants rated each item on a 5-point scale (0 = not at all, 1 = mild, 2 = moderate, 3 = strong, and 4 = very strong). Differing from the original scale, the additional “very strong” category was added for two reasons: Firstly, the literal German translation of “severe” is a highly uncommon endpoint in German rating scales. Secondly, the questionnaire was aimed at an inpatient subgroup comprising only cannabis-dependent consumers. We aimed to be able to distinguish less affected from severely affected subjects. The questionnaire was administered eight times (Table 1). Based on previous research (Budney et al., 1999, 2004) selected symptoms were subgrouped into two composite measures of physical and psychological withdrawal. Sweating, hot flashes, sleeping difficulties (including difficulties falling asleep and early awakening), appetite changes, muscle pain, dizziness (see Table 4) were conceived as physical symptoms. Nervousness, depressed mood, restlessness, irritability, aggressive behavior and tension (see Table 4) were subsumed under psychological symptoms. While craving can be conceived as part of a psychological symptom scale, we analysed this item separately because it cannot be expected to follow the same course as the other psychological symptoms (Haney, 2005; Budney et al., 2003). The scalability of the MWC item classification was assessed by means of two uni-dimensional item response models (Masters partial credit model), using data from the first measurement point. The first day was chosen because most symptoms peaked on day 1. Due to sparse data in the underlying contingency table, common 2 -statistics are unreliable, and should not be used for the assessment of model fit. Instead, we used a bootstrapped Pearson 2 -value based on 1000 bootstrap samples. The higher the p-value, the better the model fit. The bootstrap estimates for goodness of fit were p = 0.31 for the psychological symptoms scale, and p = 0.24 for the physical symptoms scale. These results support the use of both symptom lists as unidimensional measures for the intensity of physical and psychological withdrawal symptoms, respectively.

2.4. Ethical standards A signed informed consent was obtained from all study participants after complete and extensive description of the study. In case of the eight enrolled minors their parents signed written consent was obtained as well. The Ethics Committee of the Ernst–Moritz – Arndt–University of Greifswald approved the study.

2.5. Statistical analysis Cannabis withdrawal symptoms were first studied using common descriptive statistics. To analyse the course of cannabis (CB) withdrawal over time, we used a mixture of latent growth/latent class models. The applied models take into account the correlated nature of measurements in individuals at different time points as well as measurement error. To assess the hypothesized nonlinear time course of withdrawal symptoms we tested quadratic time models against linear models. A quadratic model is able to describe the hypothesized time course of withdrawal symptoms with an initial increase, and a subsequent decrease. Dependent variables were the sum scores of the physical and psychological withdrawal syndrome scales. Based on previous results, a separate model was computed for craving (Haney, 2005; Budney et al., 2003). Missing values were treated under the missing-at-random assumption. To distinguish patients with a characteristically different CWS course we compared the statistical models between solutions with a single group, two subgroups, or three subgroups. Should a model with more than one subgroup perform better compared to the single group model this indicates that the population is not homogeneous with regard to the symptom course. We used the Bayes information criterion (BIC) to compare models with a different number of subgroups. If the BIC fails to decrease for a more complex model, the improvement in explanatory power is small in relation to the added complexity of the model. Logistic regression analysis was used for the prediction of CWS group membership by consumption-related indicators as well as personality dimensions. Associations between retrospectively and prospectively measured withdrawal symptoms were assessed based on the product–moment correlation coefficient. A linear regression was used to assess the association between time elapsed since last cannabis intake and symptom intensity on the first day. For the latter, standardized beta coefficients (ˇ) are reported.

107

Day 7

Day 8

Day 9

× × × 78

Day 10 × × × 73

95% CI and p-values were estimated for two-sided tests. Associations are labelled as being “significant” if p < 0.05. STATA 9.2 (Stata Corporation, College Station, TX), and MPLUS 4.2 (Muthen & Muthen, CA) were used for the statistical analyses.

3. Results 3.1. Sample characteristics The mean age of the sample was just under 20 years, the majority of subjects were male (Table 2). Since the sample comprised adolescents and young adults, some patients were still in their educational training or lived together with their parents. The educational level was mostly low, almost two-thirds of the patients were unemployed, and more than half reported having been in detention. Almost one-third of patients reported prior detention related to illegal drugs. Compared to the German NEO-FFI norm sample (Borkenau and Ostendorf, 1993), the patients scored higher on neuroticism, but lower an all the remaining personality dimensions. 3.2. Characteristics of cannabis consumption and other illicit substances All patients met the DSM-IV criteria of cannabis dependence (Table 3). More than 70% endorsed at least five of the seven dependence criteria. More than two-thirds of the sample consumed THC within the last 12 h before admission. Stimulants and hallucinogens were the most preferred illicit substances other than cannabis. Patients smoked on average about 18–19 cigarettes per day (SD approx. 10) with few changes during their stay (MANOVA F-value: 1.46; df: 7; p: 0.20). 3.3. Characteristics of cannabis withdrawal—descriptive results The intensity of most self-reported symptoms peaked on day 1 and decreased subsequently. Most symptoms ranged on average between low to moderate intensity. The most frequently mentioned physical symptoms of strong or very strong intensity on the first day were sleeping problems (20.7%), sweating (28.2%), hot flashes (20.7%) and decreased appetite (15.4%, Table 4). The most frequent “strong” or “very strong” ratings were given for craving (37.9%). Other often highly rated psychological symptoms included restlessness (19.8%), nervousness (19.7%), and sadness (19.2%). Proportions of the “very strong” response category did not exceed 13.8% on any physical item, 12.2% on any psychological item, or 14.7% for craving. Four withdrawal symptoms of at least moderate intensity were reported by the majority of subjects (69.8%) on the first day. 3.4. Characteristics of cannabis withdrawal—model-based results Multiple group models performed statistically substantially better compared to a single subgroup model: Using a linear time model, BIC dropped from 1318 in the single subgroup solution to 1202 in a two-subgroup solution for physical symptoms and from 1479 to 1293 for psychological symptoms. Further improvements by adding a third subgroup were substantially smaller (BIC = 1187 for physical, and 1256 for psychological symptoms).

136

U.W. Preuss et al. / Drug and Alcohol Dependence 106 (2010) 133–141

Table 2 Sample characteristics.

Subjects (N) Age (Mean, SD)

Total

Females

Males

118 19.7 (2.9)

17 19.6 (2.6)

101 19.7 (2.9)

Total, %

Females, %

Males, %

Marital status Living alone Living with a partner Other

40.6 12.9 46.5

21.4 42.9 35.7

43.7 8.0 48.3

Housing Alone With parents With partner With partner and children With children Other

32.0 37.9 10.7 2.9 1.0 15.5

28.6 35.7 7.1 7.1 7.1 14.3

32.6 38.2 11.2 2.2 0 15.7

Educational achievements School drop-out Completed 9 years of schooling Completed 10+ years of schooling Currently in school Other

20.8 40.6 20.8 5.0 12.9

21.4 50.0 7.1 7.1 14.3

20.7 39.1 23.0 4.6 12.6

Unemployeda

60.9

75.0

58.8

Detention history

50.5

21.4

55.4

Detention history related to illegal drugs

30.2

7.1

34.1

Treatment history Previous CWS (in- or outpatient) Previous rehabilitations

27.7 16.0

21.4 7.1

28.7 17.4

Personality dimensions

Total, M (SD)

Females, M (SD)

Males, M (SD)

Neuroticism Extraversion Openness Agreeableness Conscientiousness

2.24 (0.53) 2.19 (0.52) 2.25 (0.53) 2.21 (0.37) 2.33 (0.56)

2.20 (0.52) 2.20 (0.53) 2.24 (0.51) 2.17 (0.36) 2.36 (0.52)

2.56 (0.57) 2.10 (0.66) 2.30 (0.64) 2.42 (0.41) 2.19 (0.77)

a Excluding patients still in school. All percentages refer to the total study sample with the exception of detention history, and personality dimensions. Data from 15 cases were not recoverable due to a technical failure of the electronic input mask.

The three-subgroup analysis revealed that inpatients with elevated symptoms consisted of two fractions: one with repeated elevated symptoms and one with elevated symptoms of a mostly low intensity on a single day. Only the first group was conceived as a high CWS group in the subsequent analyses. A similar observation was made for craving comparing the one- and two-subgroup solutions: BIC was 2726 in the linear one-subgroup model as compared to 2334 in the two-subgroup model. Statistical fit was not improved by a three-subgroup model (BIC = 2358). The probability of belonging to the assigned group always exceeded 90%. The interpretation of the two encountered groups corresponds to a high versus low intensity withdrawal group as depicted in Fig. 1. The models showed less than 50% of the sample with repeated elevated reporting of physical or psychological symptoms (Fig. 1). Had we taken the original two-subgroup solution, the proportion would have been about 10% points higher, respectively. About 41% reported an elevated level of craving. In all longitudinal analysis, no significant or clinically meaningful quadratic time effect with an initial increase as hypothesized emerged for any of the observed subgroups. The intensity of physical and psychological symptoms as well as craving decreased linearly or almost linearly over time (Fig. 1). The overlap between psychological and physical symptoms is large: 79.6% with psychological symptoms reported elevated physical symptoms. Vice versa, the proportion was 75.4%. In comparison,

42.4% reported neither increased physical nor psychological symptoms. The overlap with craving was large as well: 70.8% in the high craving subgroup reported elevated psychological or physical symptoms. 3.5. Predictors of withdrawal Age, sex, as well as THC urine levels, and the number of dependence criteria were not significantly related to CWS group membership (p > 0.1, Table 5). However, inclusion in the elevated psychological symptoms subgroup or in the elevated physical symptoms subgroup was predicted by the last amount of cannabis and the time elapsed since last consumption when controlling for other cannabis-related variables in the logistic regression model. Using the same variables to predict symptom intensity on the first day rendered a positive association between the time elapsed since last consumption and physical (ˇ = 0.27, p < 0.01) as well as psychological (ˇ = 0.22, p = 0.03) withdrawal symptom intensity. Having consumed other illicit drugs consistently increased the chance of inclusion in the high psychological symptoms subgroup. However these effects were not significant. Cigarette consumption was significantly associated with craving. Scoring high on neuroticism and low on agreeableness was significantly associated with more psychological withdrawal symptoms. Smaller and insignificant effects were detected for the influence of NEO-FFI personality dimensions on physical symptoms and craving with the exception

U.W. Preuss et al. / Drug and Alcohol Dependence 106 (2010) 133–141

137

Table 3 Consumption characteristics of cannabis and other illicit substances. Total N

Females

118

Males

17

101

Continuous variables

Total, mean (SD)

Females, mean (SD)

Males, mean (SD)

THC urine concentration at admission (ng/ml) Last amount of cannabis consumed before admission (g)

682.9 (462.5) 1.5 (1.1)

817.5 (519.9) 1.2 (0.7)

661.8 (452.0) 1.5 (1.1)

Categorical variables

Total, %

Females, %

Males, %

Daily cigarette consumption before admission <10 per day >10 per day

15.9 84.1

7.1 92.9

17.2 82.8

Last THC consumption before admission <6 h 6–12 h 12–24 h 24–48 h >48 h

29.1 32.5 7.7 16.3 14.4

50.0 18.8 0 18.8 12.4

25.7 34.7 8.9 15.9 14.9

Illicit drug consumption Consumption by partner Consumption by parents

29.5 56.3

76.9 61.5

21.9 55.4

DSM-IV cannabis dependence criteria (1) Tolerance (2) Withdrawal (3) Taken in larger amounts or over a longer period (4) Persistent desire/unsuccessful efforts to cut down substance use (5) Much time spent in activities to obtain substance (6) Neglect of important activities (7) Consumption continued despite problems

85.0 68.8 69.0 86.7 69.9 75.2 82.3

82.4 81.2 82.4 88.2 69.8 70.6 76.5

85.4 66.7 66.7 86.5 70.6 76.0 83.3

Number of endorsed DSM-IV cannabis dependence criteria Three Four Five Six Seven

16.8 11.8 22.7 24.4 24.4

11.8 11.8 23.5 17.6 35.3

17.6 11.8 22.5 25.5 22.5

Consumption history of other illicit drugs Sedatives Stimulants Opioids Cocaine Hallucinogens Others

10.6 49.6 3.5 36.3 51.3 18.6

23.5 41.2 5.9 41.2 41.2 11.8

8.3 51.0 3.1 35.4 53.1 19.8

of the latter being significantly predicted by agreeableness. Using the original two-subgroup solution would have yielded a similar association pattern. 3.6. Prospective and retrospective withdrawal symptoms As presented in Table 3, almost 70% of the patients reported having experienced withdrawal symptoms previously. The five most common symptoms were sweating (28.6%), sleeping problems (22.7%), restlessness (18.5%), aggressiveness (13.5%), and depressed mood (13.5%). We conducted bivariate correlations between all these retrospectively assessed variables and (a) the CWS physical, psychological, and craving group membership, (b) the sum score of the physical, psychological, and craving symptoms scale on days 1–3, (c) the corresponding single symptoms on days 1–3. While almost all associations were non-significant and of a negligible effect size, a few statistically significant correlations were around 0.2 for the association between having reported a previous CWS and membership in the high intensity psychological CWS group, as well as for the association between present and past sweating and aggressiveness during cannabis withdrawal. 4. Discussion Our findings contribute to existing knowledge on the CWS in important aspects because the present CWS assessment was con-

ducted prospectively among adolescent and young adult patients during an inpatient stay in a highly controlled environment. Self-reported symptoms resembled those of previous studies and comprised psychological as well as physical symptoms. However, only a subgroup experienced a CWS of clinical significance despite the fact that all patients had a diagnosis of cannabis dependence according to DSM-IV criteria. Our results support subgrouping cannabis-dependent subjects in those with no or only a very mild CWS and those with a moderate-to-strong CWS. Recent cannabis intake and the last amount of cannabis consumed prior to hospitalization predicted inclusion in the low versus high symptom intensity subgroup but not the number of fulfilled DSM-IV dependence criteria. Significant associations of personality characteristics with psychological withdrawal symptoms suggest that at least some of the elevated symptoms are related to factors other than cannabis consumption patterns. Methodologically, it is noteworthy that retrospective and prospective symptoms were virtually not associated. This raises questions about previous findings based on retrospective assessments with regard to a CWS. As to the first research question, key self-reported withdrawal symptoms in the current study resembled those of previous studies. Psychological symptoms comprised aggression, depressive mood, anxiety, irritability, nervousness, and restlessness, physical symptoms comprised appetite changes, sleeping difficulties, sweating, dizziness, and hot flashes (Budney et al., 2003, 2004; Haney et al.,

138

U.W. Preuss et al. / Drug and Alcohol Dependence 106 (2010) 133–141

Table 4 Intensity of physical and psychological symptoms at day 1. Intensity Not at all, %

Mild, %

Moderate, %

Strong, %

Very strong, %

Physical symptoms Problems sleeping through the night Problems falling asleep Waking up early Increased appetite Decreased appetite Nausea Muscle pain Sweating Feverish Diarrhea, abdominal discomfort Headache Hot flashes Dizziness Stuffy nose Hiccups

39.7 28.4 45.6 62.4 59.8 84.6 77.8 25.6 98.3 75.2 67.2 40.5 69.0 51.7 91.4

18.1 14.7 19.3 13.7 12.8 8.5 14.5 24.8 0.9 13.7 13.8 25.9 16.4 19.0 7.8

21.6 21.6 16.7 17.9 12.0 5.1 5.1 21.4 – 7.7 12.9 12.9 10.3 21.6 0.9

14.7 21.6 12.3 4.3 9.4 5.1 0.9 18.8 0.9 1.7 4.3 14.7 3.4 2.6 –

6.0 13.8 6.1 1.7 6.0 1.7 1.7 9.4 – 1.7 1.7 6.0 0.9 5.2 –

Psychological symptoms Nervousness Depressed mood Restlessness Irritability Anxious Tension Sadness Anger Nightmares Aggressiveness

23.9 37.1 31.9 43.6 74.4 28.4 36.5 64.3 77.9 75.9

28.2 22.4 24.1 22.2 15.4 34.5 28.7 21.7 9.7 12.1

28.2 25.9 24.1 17.9 4.3 20.7 15.7 7.8 5.3 6.0

14.5 11.2 16.4 11.1 3.4 14.7 7.0 3.5 3.5 4.3

5.1 3.4 3.4 5.1 2.6 1.7 12.2 2.6 3.5 1.7

Craving

19.8

25.9

16.4

23.3

14.7

1999; Kouri and Pope, 2000). A recent study in adolescent outpatients aged 16.2 ± 1.1 years reported craving, depressed mood, irritability, restlessness, sleep difficulty, anger and aggression as the most commonly reported CWS symptoms (Vandrey et al., 2005). Subjects in our study reported withdrawal symptoms more often compared to this adolescent sample (Vandrey et al., 2005), but less often than adults in a previous study (Budney et al., 1999). The results of the former investigation may be related to the more

lenient inclusion criteria employed. However, comparisons are complicated for two reasons: Firstly, contrary to our study the MWC was applied to previous episodes of cannabis abstinence. Secondly, we applied a translated version of the scale with a partially different coding of categories. In line with previous findings from other authors, there is a significant overlap of psychological and physical symptoms (Arendt et al., 2007; Budney et al., 2004; Vandrey et al., 2005). This casts doubt

Table 5 Associations between drug consumption and withdrawal symptoms. Physical OR (95% CI)

Psychological OR (95% CI)

Craving OR (95% CI)

Age Gender

1.0 (0.91–2.0) 0.8 (0.3–2.3)

0.9 (0.8–1.1) 0.9 (0.3–2.3)

0.9 (0.8–1.1) 0.6 (0.2–1.9)

Cannabis consumption characteristicsa # of DSM-IV dependence criteria endorsed THC urine level >500 ng/ml Last consumption less than 6h before admission Last amount of cannabis consumed >1 g

1.2 (0.9–1.7) 1.0 (0.4–2.2) 3.9 (1.3–11.0)* 4.1 (1.7–10.5)*

1.2 (0.9–1.1) 0.8 (0.3–1.9) 5.2 (1.8–15.6)* 4.6 (1.8–11.7)*

1.1 (0.8–1.5) 1.0 (0.4–2.3) 1.4 (0.5–3.6) 2.1 (0.9–4.9)

Other drug and nicotine consumptionb Stimulants Hallucinogens Cocaine Other Cigarettesc

0.9 (0.9–2.0) 0.7 (0.3–1.6) 1.1 (0.5–2.4) 0.5 (0.2–1.4) 1.4 (0.6–3.5)

2.0 (0.9–4.6) 1.5 (0.7–3.5) 1.6 (0.7–3.7) 0.9 (0.3–2.4) 2.8 (1.0–5.2)

1.6 (0.7–3.3) 1.1 (0.5–2.4) 1.4 (0.6–3.2) 1.2 (0.5–3.1) 4.3 (1.5–12.4)*

Personality dimensionsd Neuroticism Extraversion Openness Agreeableness Conscientiousness

1.3 (0.8–1.9) 1.0 (0.7–1.5) 1.2 (0.8–1.8) 0.7 (0.5–1.1) 1.3 (0.8–1.9)

1.6 (1.1–2.5)* 0.9 (0.6–1.3) 1.4 (0.9–2.1) 0.5 (0.3–0.8)* 1.1 (0.7–1.6)

1.4 (0.9–2.2) 1.0 (0.7–1.6) 1.1 (0.7–1.6) 0.6 (0.4–1.0)* 1.0 (0.7–1.5)

a b c d *

Age, gender, and all cannabis consumption characteristics in the model. Controlling for age, gender, last consumption, and time since last consumption. Unit is packages per day. One package contains 19 cigarettes. Controlling for age and gender, scales z-standardized. p < 0.05.

U.W. Preuss et al. / Drug and Alcohol Dependence 106 (2010) 133–141

Fig. 1. Course of withdrawal symptoms.

on views that somatic symptoms predominantly occur in a small group of extreme cases only (Smith, 2002). Reporting bias based on established expectations could have contributed to the considerable overlap between psychological and physical symptoms in our study. However, the small associations between recalled and prospectively assessed withdrawal symptoms suggest minor effects of established expectations on symptom reporting. Physical symptoms such as sweating, sleeping and eating disorders may be related to a dysfunction of the autonomic nervous system occurring during cannabis withdrawal, since similar symptoms in alcohol withdrawal are related to a dysregulation of autonomic function (e.g. Björkqvist, 1975). A linear or nearly linear decrease of symptoms over the observed time period best described the symptom course in our sample. Thus, the hypothesis of an initial increase and subsequent decrease of CWS symptom intensity needs to be rejected. This contrasts with other outpatient studies which had an experimental design or at least an established symptom baseline (Budney et al., 2001, 2003;

139

Kouri and Pope, 2000). In the latter investigations, cessation of cannabis consumption was initiated per protocol at a fixed time point. There is more uncertainty about cannabis cessation in the present study despite an available self-report on the abstinence period before hospitalization. This problem might have been aggravated because most symptoms may need only a short time to reach peak intensity, as suggested by other studies. Furthermore, moving to a closed ward might have influenced symptom reporting. In line with the second hypothesis, our results underline the necessity to distinguish between subgroups with a characteristically different course of withdrawal symptoms. The model-based analysis showed that the observed variability in CWS intensity across subjects is unlikely to come from a single distribution. At least one subgroup with none-to-mild symptoms must be distinguished from the other subgroup with increased level of symptoms lasting several days. Our results confirm findings from retrospective assessments (Agrawal et al., 2008; Budney et al., 1999; Hasin et al., 2008; Vandrey et al., 2005). Moderate-to-strong psychological or physical symptoms may be expected only in a subgroup, even in strongly cannabis-dependent subjects seeking inpatient treatment. Our findings are of importance in the clinical context of the present study because they question the necessity for an inpatient detoxification treatment for the sake of cannabis withdrawal only. However, a longitudinal study on CWS over a longer time period would be useful to further validate this conclusion by comparing effects of inpatient and outpatient withdrawal on future drug consumption. Furthermore, a larger sample would be advisable to replicate and expand our findings on CWS subgroups. More subgroups might be identifiable in a larger sample. Regarding the third research question and in line with our hypothesis, subjects who consumed more cannabis before hospitalization, were more likely to report a CWS. This corresponds to findings from other studies (Budney et al., 1999; Vandrey et al., 2005). Yet, contrary to our assumptions, symptoms were stronger and a CWS was reported with a higher likelihood among those who consumed cannabis just before admission. Subjects who consumed cannabis just before their hospitalization had no time delay in their reported withdrawal symptom peak. This may be related to inconsistencies with regard to self-reported consumption but also to the fast onset of symptoms. Time of last cannabis intake may likely serve as an indirect measure for consumption intensity. This explanation would correspond better to the observed results. A baseline measure of physical and psychological symptoms before withdrawal would be necessary to further explain these results. The negligible association between CWS and THC urine levels may be attributable to factors such as differences in individual pharmacokinetics, presence of metabolites, patient variability (e.g. body mass), or urine pH (Moeller et al., 2008). Another finding that did not meet our expectations was the weak association between endorsed DSM-IV dependence criteria and CWS. Hence, the number of criteria may be an inappropriate measure for cannabis dependence severity. The comparatively small effects may also be due to a selective sample of highly dependent subjects who consumed cannabis almost every day. This reduced variance at the predictor side is likely to impair substantial associations. While subjects with other substance dependences were not enrolled into the study, the vast majority of the subjects reported prior use of other substances, in particular stimulants and hallucinogens. However, this non-dependent illicit substance use was only moderately related to psychological withdrawal symptoms. Marginal associations concerned physical withdrawal. The former corresponds to findings from Wiesbeck et al. (1996) who noted significant relationships between CWS and sedative/hypnotic and alcohol dependence in the COGA (Collaborative Study on Genetics in Alcoholism) sample. However, due to our small sample size,

140

U.W. Preuss et al. / Drug and Alcohol Dependence 106 (2010) 133–141

findings were statistically not significant, and a larger sample is required to corroborate these results. To assess influences of psychological factors on symptom reporting we measured personality dimensions. As observed for nicotine or benzodiazepines, neuroticism predicted psychological withdrawal symptoms (Madden et al., 1997; Reuter and Netter, 2001; Schweizer et al., 1998). To our knowledge, comparable associations for low agreeableness, the tendency to be pleasant and accommodating in social situations, have not been reported previously for CWS. The observed associations suggest that elevated symptom reporting may not only be attributable to cannabis withdrawal. Yet, when predicting CWS based on a joint model of consumption-related indicators and personality dimensions, the former remained more important. The proportion of dependent subjects reporting a previous CWS matched those of a previous study among cannabis-dependent subjects (Crowley et al., 1998). However, contrary to our fourth research question and the corresponding hypothesis, we found low to negligible associations between recalled and prospectively reported symptoms. As histories of symptoms are prone to recall bias, we consider prospectively assessed symptoms to be more reliable. Our findings indicate that results from previous observational studies (for example Budney et al., 1999; Cornelius et al., 2008; Hasin et al., 2008; Vandrey et al., 2005; Wiesbeck et al., 1996) may lack validity despite a reasonable reliability (Mennes et al., 2009). However, we have employed different assessment instruments to obtain symptoms retro- and prospectively. This may limit substantial conclusions regarding the weak associations between DSM-IV criteria endorsed and CWS intensity. The low associations may also be caused by potential shortcomings of the dependence syndrome assessment using the SCID. Further studies should therefore base corresponding analyses on comparable instruments. While the current study is, to our knowledge, the first to prospectively investigate cannabis withdrawal symptoms in an inpatient sample of cannabis-dependent subjects, there are several limitations. First, a control group of inpatient nicotine-dependent subjects would have been desirable to compare the short-term characteristics and course of cannabis with nicotine withdrawal. However, such a sample is virtually impossible to obtain, since nicotine dependence is not an indication for inpatient treatment in Germany. Treatment on a voluntary basis together with other drug addicts would have been ethically questionable. For several reasons it is reasonable to assume that the observed symptoms are – to a large extent – attributable to cannabis because subjects with any other severe psychopathology or somatic disorder have been excluded. Our study provided a better control than most other studies. Uncontrolled consumption of any drug could be excluded with high certainty in a closed ward. Subjects were able to smoke and drink coffee without limitations. Furthermore, the symptom patterns reported by our patients match those obtained in other studies. Second, attrition might have biased our findings as 46 of the 119 subjects dropped out of the study. However, statistical results barely changed, when the analyses were conducted without the last time points or when the number of available time points was used as a predictor for CWS group membership. Although about onethird of the sample dropped out over the course of their stay, the fraction of unavailable data was much smaller: 819 out of 944 data points were available (=87%). Because prior data on the symptom course was available from all patients and symptom intensity at the beginning of the stay predicted later symptom intensity, the actual amount of information lost is even smaller. In addition, the highest symptom intensity was reported during the first days of stay while the available sample was almost complete. It is therefore unlikely that attrition severely biased our results.

Third, the observation period of the study was shorter compared to some previous studies (Budney et al., 2001, 2003; Milin et al., 2008). For instance, complete degradation of cannabinoids from fatty body tissues might last several weeks. A number of symptoms, including nightmares and irritability may therefore last significantly longer than the 10-day period of the current study (Budney et al., 2004). Fourth, cross-cultural translation always poses a threat to meaningful comparisons of results. The scaling of the symptom checklist comprised some changes to the original scale to better match typical German rating scale formats. However, assessing scalability rendered similar results to previous research (Budney et al., 1999). The introduction of an additional “very strong” category for all MWC scale turned out to be of limited use even in this sample of dependent inpatients. In line with previous findings this underlines that most subjects do not experience strong or severe symptoms related to cannabis withdrawal (Agrawal et al., 2008; Budney et al., 1999; Hasin et al., 2008; Vandrey et al., 2005). Fifth, the comparability of our inpatient psychiatric sample to other study populations is an issue. However, subjects with other current DSM-IV Axis I diagnosis, or severe somatic or neurological disorders were transferred to other psychiatric wards and excluded from the study. A similar procedure was applied for instance by Budney et al. (1999). Therefore, the sample should not be substantially different from other dependent samples. Sixth, a more extensive assessment of recent cannabis and general drug use, including a detailed quantification of cannabis intake would have improved our analyses regarding their relationship to CWS course and intensity. Finally, only a small fraction of the subjects included into the study were female. A larger sample with more female subjects is desirable to assess potential gender differences in CWS. Role of funding Funding for this study was provided by the Johanna-OdebrechtFoundation. The Foundation had no further role in study design, in the collection, analysis and interpretation of data, in the writing of the report, or in the decision to submit the paper for publication. Conflicts of interest U.W.P. and J.Z. have financial interests/arrangements or affiliations with one or more organizations that could be perceived as a conflict of interest in the context of the subject of this manuscript. He received either research support, consultancy, or lecture fees from Pfizer, Astra-Zeneca, Bristol-Meyers-Squibbs, EliLilly, Janssen-Cilag, Novartis and Wyeth in the past 3 years. C.O.S. has received consultancy and lecture fees from Pfizer, and Grünenthal on subjects unrelated to this manuscript. ABW, COS and JWMW have no conflicts of interest to declare. References Agrawal, A., Pergadia, M.L., Lynskey, M.T., 2008. Is there evidence for symptoms of cannabis withdrawal in the national epidemiologic survey of alcohol and related conditions? Am. J. Addict. 17, 199–208. Arendt, M., Rosenberg, R., Foldager, L., Sher, L., Munk-Jorgensen, P., 2007. Withdrawal symptoms do not predict relapse among subjects treated for cannabis dependence. Am. J. Addict. 16, 461–467. Barnes, G.E., Barnes, M.D., Patton, D., 2005. Prevalence and predictors of “heavy” marijuana use in a Canadian youth sample. Subst. Use. Misuse. 40, 1849–1863. Björkqvist, S.E., 1975. Clonidine in alcohol withdrawal. Acta Psychiatr. Scand. 52, 256–263. Borkenau, P., Ostendorf, F., 1993. NEO-Fünf-Faktoren-Inventar (NEO-FFI) nach Costa und McCrae. Hogrefe, Göttingen. Budney, A.J., Hughes, J.R., 2006. The cannabis withdrawal syndrome. Curr. Opin. Psychiatry 19, 233–238.

U.W. Preuss et al. / Drug and Alcohol Dependence 106 (2010) 133–141 Budney, A.J., Hughes, J.R., Moore, B.A., Novy, P.L., 2001. Marijuana abstinence effects in marijuana smokers maintained in their home environment. Arch. Gen. Psychiatry 58, 917–924. Budney, A.J., Hughes, J.R., Moore, B.A., Vandrey, R., 2004. Review of the validity and significance of cannabis withdrawal syndrome. Am. J. Psychiatry 161, 1967–1977. Budney, A.J., Moore, B.A., Vandrey, R.G., Hughes, J.R., 2003. The time course and significance of cannabis withdrawal. J. Abnorm. Psychol. 112, 393–402. Budney, A.J., Novy, P., Hughes, J.R., 1999. Marijuana withdrawal among adults seeking treatment for marijuana dependence. Addiction 94, 1311–1322. Chabrol, H., Duconge, E., Roura, C., Casas, C., 2004. Relations between anxious, depressive and borderline symptomatology and frequency of cannabis use and dependence. Encephale 30, 141–146. 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. Costa, P.T., McCrae, R.R., 1992. Revised NEO Personality Inventory (NEO PI-R) and NEO Five Factor Inventory. Professional Manual. Psychological Assessment Resources, Odessa, FL. Cornelius, J.R., Chung, T., Martin, C., Wood, D.S., Clark, D.B., 2008. Cannabis withdrawal is common among treatment-seeking adolescents with cannabis dependence and major depression, and is associated with rapid relapse to dependence. Addict. Behav. 33, 1500–1505. Crowley, T.J., Macdonald, M.J., Whitmore, E.A., Mikulich, S.K., 1998. Cannabis dependence, withdrawal, and reinforcing effects among adolescents with conduct symptoms and substance use disorders. Drug Alcohol Depend. 50, 27–37. DGS, 2001. Documentation Standards III for the Evaluation of Addiction Treatment. Sucht 47, special issue no. 2. European Monitoring Centre for Drugs and Drug Addiction (EMCDDA), 2006. Annual Report 2006: The State of the Drugs Problem in Europe. European Monitoring Centre for Drugs and Drug Addiction, Lisbon. Haney, M., 2005. The marijuana withdrawal syndrome: diagnosis and treatment. Curr. Psychiatry Rep. 7, 360–366. Haney, M., Comer, S.D., Ward, A.S., 1999. Abstinence symptoms following oral THC administration to humans. Psychopharmacology (Berl.) 14, 385–394. Hasin, D.S., Keyes, K.M., Alderson, D., Wang, S., Aharonovich, E., Grant, B.F., 2008. Cannabis withdrawal in the United States: results from NESARC. J. Clin. Psychiatry 69, 1354–1363. Jones, R.T., Benowitz, N., Bachman, J., 1976. Clinical studies of cannabis tolerance and dependence. Ann. N.Y. Acad. Sci. 282, 221–239. Justinova, Z., Goldberg, S.R., Heishman, S.J., Tanda, G., 2005. Self-administration of cannabinoids by experimental animals and human marijuana smokers. Pharmacol. Biochem. Behav. 81, 285–299.

141

Kouri, E.M., Pope Jr., H.G., 2000. Abstinence symptoms during withdrawal from chronic marijuana use. Exp. Clin. Psychopharmacol. 8, 483–592. Madden, P.A., Bucholz, K.K., Dinwiddie, S.H., Slutske, W.S., Bierut, L.J., Statham, D.J., Dunne, M.P., Martin, N.G., Heath, A.C., 1997. Nicotine withdrawal in women. Addiction 92, 889–902. Moeller, K.E., Lee, K.C., Kissack, J.C., 2008. Urine drug screening: practical guide for clinicians. Mayo Clin. Proc. 83 (1), 66–76. Mennes, C.E., Ben, A.A., Cottler, L.B., 2009. The reliability of self-reported cannabis abuse, dependence and withdrawal symptoms: multisite study of differences between general population and treatment groups. Addict. Behav. 34, 223–226. Milin, R., Manion, I., Dare, G., Walker, S., 2008. Prospective assessment of cannabis withdrawal in adolescents with cannabis dependence: a pilot study. J. Am. Acad. Child Adolesc. Psychiatry 47, 174–178. Nocon, A., Wittchen, H.U., Pfister, H., Zimmermann, P., Lieb, R., 2006. Dependence symptoms in young cannabis users? A prospective epidemiological study. J. Psychiatr. Res. 40, 394–403. Perkonigg, A., Goodwin, R.D., Fiedler, A., Behrendt, S., Beesdo, K., Lieb, R., Wittchen, H.U., 2008. The natural course of cannabis use, abuse and dependence during the first decades of life. Addiction 103, 439–449. Perkonigg, A., Lieb, R., Höfler, M., Schuster, P., Sonntag, H., Wittchen, H.U., 1999. Patterns of cannabis use, abuse and dependence over time: incidence, progression and stability in a sample of 1228 adolescents. Addiction 94, 1663–1678. Reuter, M., Netter, P., 2001. The influence of personality on nicotine craving: a hierarchical multivariate statistical prediction model. Neuropsychobiology 44, 47–53. Schweizer, E., Rickels, K., De, M.N., Case, G., Garcia-Espana, F., 1998. The effect of personality on withdrawal severity and taper outcome in benzodiazepine dependent patients. Psychol. Med. 28, 713–720. Smith, N.T., 2002. Review of the published literature into cannabis withdrawal symptoms in human users. Addiction 97, 621–632. Vandrey, R., Budney, A.J., Kamon, J.L., Stanger, C., 2005. Cannabis withdrawal in adolescent treatment seekers. Drug Alcohol Depend. 78, 205–210. Wiesbeck, G.A., Schuckit, M.A., Kalmijn, J.A., Tipp, J.E., Bucholz, K.K., Smith, T.L., 1996. An evaluation of the history of a marijuana withdrawal syndrome in a large population. Addiction 91, 1469–1478. Wittchen, H.U., Zaudig, M., Fydrich, T., 1997. Strukturiertes klinisches Interview für DSM-IV (SKID) Achse I und II. Hogrefe, Göttingen. Wittchen, H.U., Fröhlich, C., Behrendt, S., Günther, A., Rehm, J., Zimmermann, P., Lieb, R., Perkonigg, A., 2007. Cannabis use and cannabis use disorders and their relationship to mental disorders: a 10-year prospective-longitudinal community study in adolescents. Drug Alcohol Depend. 88 (Suppl 1), S60–S70.