- Email: [email protected]
Cannabis use disorder and age at onset of psychosis — A study in first-episode patients
Schizophrenia Research 129 (2011) 52–56
Contents lists available at ScienceDirect
Schizophrenia Research j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s c h r e s
Cannabis use disorder and age at onset of psychosis — A study in first-episode patients Benno G. Schimmelmann a,⁎, Philippe Conus b,c, Sue M. Cotton c, Stephan Kupferschmid a, Anne Karow d, Frauke Schultze-Lutter a, Patrick D. McGorry c, Martin Lambert c,d a
University Hospital of Child and Adolescent Psychiatry, Bern, Switzerland Treatment and early Intervention in Psychosis Program (TIPP), Department of Psychiatry, CHUV, Lausanne, Switzerland Orygen Youth Health and Research Centre, Centre for Youth Mental Health, Melbourne, Australia d Psychosis Early Detection and Intervention Centre (PEDIC), Department of Psychiatry and Psychotherapy, University Medical Center Hamburg-Eppendorf, Germany b c
a r t i c l e
i n f o
Article history: Received 17 December 2010 Received in revised form 21 March 2011 Accepted 25 March 2011 Available online 17 April 2011 Keywords: First-episode psychosis Schizophrenia Substance use Adolescence
a b s t r a c t Introduction: Age at onset of psychosis (AAO) may be younger in patients with cannabis use disorders (CUD) compared to those without CUD (NCUD). Previous studies included CUD co-morbid with other substance use disorders (SUD), and many did not control for confounders. Methods: Controlling for relevant confounders, differences in AAO between patients with and without CUD excluding those with any other SUD were analyzed in a large representative file audit of 625 first-episode psychosis (FEP) patients (age 14 to 29 years) admitted to the Early Psychosis Prevention and Intervention Centre in Melbourne, Australia. Results: Three quarters of the 625 FEP patients had a CUD. Cannabis use started before psychosis onset in 87.6% of patients. AAO was not significantly different between CUD (without other SUD, n = 201) and NCUD (n = 157). However, AAO was younger in those with early CUD (starting age 14 or younger) compared to NCUD (F(1) = 5.2; p = 0.024; partial η2 = 0.026). Earlier age at onset of cannabis use predicted earlier age at onset of psychosis (β = − 0.49, R2-change = 0.25, p b 0.001). Conclusion: Only CUD starting age 14 or younger was associated with an earlier AAO at a small effect size. These findings suggest that CUD may exert an indirect effect on brain maturation resulting in earlier AAO potentially only in cannabis sensitive subjects. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Previous studies support the role of cannabis as a risk factor for psychosis. Arseneault et al. (2004) systematically reviewed all prospective epidemiological studies. The authors concluded that cannabis use confers an overall twofold increase in the relative risk for later schizophrenia. Notably, the effect of cannabis exposure on the presence of psychotic symptoms was reported to be significantly stronger when exposure occurs in early adolescence (Arseneault et al., 2002; Konings et al., 2008) and when cannabis use was persistent (Kuepper et al., 2011). Konings et al. (2008) used age 14 as cut-off for early cannabis exposure. Interestingly, the literature also suggests that age at onset of psychosis (AAO) is younger in patients abusing cannabis compared to those who do not, on the basis of data stemming from retrospective studies conducted in adult-onset multiple- or first-episode psychosis (Large et al., 2011). ⁎ Corresponding author at: University Hospital of Child and Adolescent Psychiatry, Bolligenstr. 110, 3000 Bern 60, Switzerland. Tel.: +41 31 932 8564; fax: +41 31 932 8569. E-mail address: [email protected] (B.G. Schimmelmann). 0920-9964/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.schres.2011.03.023
Several explanations have been suggested for the relationship between cannabis use and psychosis: First, there is evidence that the endocannabinoid system plays an important role in brain developmental processes in those brain regions (prefrontal cortex, limbic structures and hippocampus) and neurotransmitter systems (GABAergic and dopaminergic transmission), which are pathophysiologically relevant to schizophrenia (Malone et al., 2010). Thus, a causal link between cannabis use and the emergence of psychosis and potentially also between early cannabis use and a younger AAO is intuitively likely, particularly if early cannabis use or abuse interferes with brain developmental processes in adolescence. A second hypothesis proposes that cannabis use causes psychosis by means of gene environment interaction (Caspi et al., 2005). For example, the study by Caspi et al. (2005) found that a functional polymorphism in the catechol-O-methyltransferase (COMT) gene moderated the influence of adolescent cannabis use on developing adult psychosis. A third hypothesis postulates that the earlier AAO in cannabis users may be the result of factors unrelated to cannabis such as gender or pre-morbid adjustment. The respective studies, however, report inconsistent results (Barnes et al., 2006; Gonzalez-Pinto et al., 2008; Sevy et al., 2010).
B.G. Schimmelmann et al. / Schizophrenia Research 129 (2011) 52–56
Limitations of many previous studies on the link between cannabis use and AAO included (Zammit et al., 2008): (i) small sample sizes; (ii) failure to report effect sizes; (iii) failure to control for relevant confounders, and (iv) lack of representativeness of samples due to an informed consent procedure, which patients with severe and/or severe psychoses are more likely to refuse (Friis et al., 2003; Wade et al., 2006). To date, no study explored AAO differences between those with cannabis use disorders (CUD) only, excluding co-morbidity with other substance use disorders (SUD), and those without CUD. Therefore, based on a large cohort of 625 adolescent and adult first episode psychosis patients (age 14–29 years), we examined differences in AAO between patients with lifetime CUD only (n = 201) and patients without CUD (n = 158), taking into account the potential confounding effect of gender and/or premorbid functioning. Based on the literature, we expected that (i) AAO would be younger in patients with CUD only compared to those without CUD (NCUD) (Large et al., 2011) and (ii) the effect of CUD on younger AAO would be even more pronounced if early CUD (cannabis use starting at age 14 or younger) were compared to NCUD. 2. Methods 2.1. Context and sample This is a retrospective file-audit study. Data were sourced from the First Episode Psychosis Outcome Study (Conus et al., 2007). The initial sample comprised a population-based cohort of 786 first-episode psychosis (FEP) patients, consecutively admitted to the Early Psychosis Prevention and Intervention Centre (EPPIC) in Melbourne, Australia, between 1998 and 2000. EPPIC covered a catchment area of approximately 880,000 people, and had a mandate to offer treatment to all FEP patients aged 14–29. As such, this clinical sample is approaching epidemiological representativeness for urban settings. The EPPIC program comprises a comprehensive early intervention treatment program, with a usual episode of care of 18 months, which encompasses extensive assessments. Each patient is assigned to a core team, who treats the patient across settings and makes entries into the chart (McGorry et al., 1990a,b). Of the 786 patients admitted to EPPIC, 82 files (10%) had been sent to other services. The excluded patients did not differ in available demographic characteristics (age and gender) and diagnostic distribution at baseline. 57 (8.1%) of the remaining 704 patients were excluded from the study due to a final diagnosis of non-psychotic disorder, substance-induced psychosis or psychosis due to general medical conditions. Further, 11 (1.6%) patients were excluded due to missing data regarding CUD and 8 (1.1%) due to missing data regarding AAO. Further, three outliers regarding AAO (age 8.21, 8.36 and 9.97) were detected in preliminary analyses and excluded. Out of the remaining 625 patients, 267 (42.7%) were excluded due to CUD co-morbid with other SUD. Data from 358 FEP patients with CUD only (n = 201) or NCUD (n = 157) was analyzed.
53
approval for this ‘non-informed consent’ file audit study allowing for the unselected inclusion of patients. 2.3. Assessment of diagnoses Clinical diagnoses (psychoses and SUD including CUD) according to DSM-IV criteria (APA, 2000) were based on the consensus of trained clinicians of the specialized assessment and crisis intervention team following an intensive diagnostic process within the first 6 weeks after entry into service. The main investigators (ML & PC) re-assessed all information available in charts with respect to baseline diagnoses. Validity of the file audit diagnoses was established by the following procedure: Between 1998 and 2000, 230 of the 786 patients treated at EPPIC had been included in prospective trials. Their main and co-morbid diagnoses were defined within 6 weeks of admission using the Structured Clinical Interview for DSM-IV (SCID-I/P; Ventura et al., 1998). The SCID and file audit diagnoses of 115 patients randomly selected within this sample of 230 were compared. The calculated kappa values revealed a very good concordance for both psychosis diagnoses (kappa = 0.80) and SUD diagnoses (kappa = 0.74). 2.4. Assessment of cannabis use disorders The prevalence of lifetime CUD, i.e., cannabis abuse or dependency according to DSM-IV criteria, as well as the age at onset of cannabis use was assessed with the Drug and Alcohol Assessment Schedule (DAAS) (McGorry et al., 1990a,b) during the first 6 weeks after service entry. Any cannabis use pattern, which did not fulfill the DSM-IV criteria for cannabis use disorders, was not assessed. The principal investigators (ML and PC) extracted this data from the files. As for some patients, the existence of lifetime CUD became only known during treatment, the CUD classification of the study was not based on DAAS results alone but on all available information in the files collected throughout the 18-month treatment period at EPPIC (Schimmelmann et al., in press). The calculated kappa value revealed a good concordance between prospectively assessed SUD diagnoses and those from the file-audit used in this study (kappa = 0.74). 2.5. Assessment of duration of untreated psychosis (DUP) and age at onset DUP was defined as age at entry into EPPIC subtracted by age when first sustained positive psychotic symptoms started (Schimmelmann et al., 2007, 2008). Clinicians at EPPIC were trained to determine DUP with patients and their relatives. In several cases, this date was modified in the file over the course of treatment on the basis of new information being gathered: For example, as patients became clinically more stable and therefore more able to reflect on symptoms' onset, they provided further details on the evolution of symptoms allowing for a more accurate estimate of DUP. The main investigators made a final decision based on all available data in the file. 2.6. Assessment of pre-treatment variables
2.2. Procedure All information on pre-treatment, baseline (at the time of admission to EPPIC), treatment and outcome variables for each patient treated at EPPIC is systematically documented in one standardized file including information over the treatment period as assessed with the Royal Park Multi-diagnostic Instrument for Psychosis (McGorry et al., 1990a,b). Using a systematic comprehensive approach (Early Psychosis File Questionnaire (EPFQ)) (Conus et al., 2007), all charts were rated by two experienced psychiatrists well acquainted with the EPPIC clinical service and treatment of young patients with FEP (ML & PC). A local ethics committee granted
Pre-treatment characteristics extracted from the files included the highest premorbid functioning according to the Global Assessment of Functioning Scale (GAF; APA, 1994), as the commonly used Premorbid Adjustment Scale was too complex to reliably extract data from files. Inter-rater reliability (ML and PC) was established for premorbid GAFscores (kappa = 0.88). 2.7. Data analysis and statistical tests Descriptive statistics were applied to compare CUD and NCUD in terms of gender and pre-treatment characteristics using Mann–
54
B.G. Schimmelmann et al. / Schizophrenia Research 129 (2011) 52–56
Table 1 Sample characteristics in patients with and without CUD (N = 358). Clinical characteristics
Pre-treatment characteristics Gender, n (%) Male Female Premorbid GAF score, mean (SD) DUP in months, median (quartiles) Diagnosis, n (%) Schizophrenia Schizophreniform disorder Schizoaffective disorder Bipolar disorder I Other psychoses
Cannabis use disordersa
Totala
Statistics
NCUD (n = 157; 43.9%)
CUD (n = 201; 56.1%)
(N = 358)
p-valueb
Effect sizec
68 89 71.5 2.0
151 50 71.8 2.0
(75.1) (24.9) (9.9) (0.3–6.1)
219 (61.2) 139 (38.8) 71.7 (10.2) 2.0 (1.0–7.0)
b0.001
0.32
0.853 0.424 0.380
NAd NAd NAd
(19.4) (44.8) (6.5) (18.4) (10.9)
74 (20.7) 155 (43.3) 17 (4.7) 68 (19.0) 44 (12.3)
35 65 4 31 22
(43.3) (56.7) (10.6) (1.0–7.1) (22.3) (41.4) (2.5) (19.7) (14.0)
39 90 13 37 22
Abbreviations: GAF = Global Assessment of Functioning Scale, DUP = duration of untreated psychosis, CUD = cannabis use disorder, NCUD = no cannabis use disorder, and SD = standard deviation. a Ns vary due to missing data. b Descriptive statistics: p-values χ2 for categorical variables and Mann–Whitney non-parametric tests for continuous variables. c Effect size = Cramer-V for χ2 tests. d NA (not applicable) is used for non-significant findings, for whom no effect size is provided.
Whitney U-tests for premorbid functioning and DUP and χ2-tests, when the dependent variable was categorical. Group differences in AAO were examined by analyses of covariance (ANCOVA) controlling for relevant confounders. Further, ANCOVA was also used to compare AAO in early CUD (cannabis use starting age 14 or younger) and NCUD. Within the CUD group, we further explored the predictive validity of age, when cannabis use started (henceforth cannabis onset) on AAO controlling for gender by means of linear regression analysis and reported change of R2, i.e., the variance of AAO explained by cannabis onset independent from gender. Careful inspection of residuals and multivariate outliers was performed in all analyses. To estimate effect sizes, Cramer's-V was used for χ2 analyses and partial η2 in ANCOVAS. Analyses were performed using the Statistical Package for Social Sciences Version 17 (SPSS 17.0; Chicago, IL). 3. Results 3.1. Sample characteristics and prevalence of CUD The sample is characterized in Table 1. Two hundred and one patients had co-morbid CUD and no other SUD; this number is equivalent to 32.2% of the original representative sample of 625 first episode patients; 158 patients had no CUD equivalent to 25.3% of the original sample. Compared to NCUD patients, CUD patients were more likely to be male (χ2 = 37.6; df = 1; p b 0.001, Cramer's V = 0.32). The groups did not differ with respect to premorbid functioning and DUP (Table 1). Accordingly, AAO comparisons between CUD and NCUD were controlled for gender. The mean AAO was 21.5 years (Md = 21.4; SD = 3.6; range 12 to 29), and the mean age of cannabis onset was 17.1 (Md = 17.0; SD = 3.1; range 10 to 27). Nineteen percent of those with CUD (n = 37) started cannabis at age 14 or younger. Notably, the majority of patients with CUD (87.6%; n = 176) started cannabis use before onset of sustained positive symptoms. Table 2 displays the timing of
psychosis and cannabis onset in males and females. All differences between males and females were non-significant. 3.2. Psychosis onset and CUD CUD and NCUD did not significantly differ with respect to AAO, when gender was controlled for (F(1) = 3.4; p = 0.067; partial n2 = 0.01). The mean age at onset of psychosis (adjusted for gender) in CUD was 21.8 (95% CI = 21.3–22.3) and 21.1 (95% CI = 20.5–21.7) in NCUD. However, patients whose cannabis onset was age 14 or younger (n = 37) had a significantly younger AAO compared to NCUD when gender was controlled for (F(1) = 5.2; p = 0.024; partial η2 = 0.026). The mean age at onset of psychosis (adjusted for gender) in early CUD was 19.4 (95% CI = 18.2–20.7) and 21.1 (95% CI = 20.5– 21.7) in NCUD. Notably, within the CUD group (n = 201) and controlling for gender, age at cannabis onset predicted AAO (linear regression analysis; β = − 0.49, R2-change = 0.25, p b 0.001). In other words, age at cannabis onset explained 25% variance of AAO. 4. Discussion This is the first study comparing age at onset of psychosis in those who only had CUD (and no other SUD) to those without CUD. Using a file audit design, common biases of less representative informed consent studies were avoided, such as the exclusion of patients with more severe psychopathology and substance use disorders (Friis et al., 2003; Menezes et al., 2006). 4.1. Sample In the original sample of 625 FEP patients, almost three quarters had lifetime CUD, and about one third had CUD only and no other SUD. This rate is higher than reported in other samples (for review see Sevy et al., 2010) and may be due to the high rate of cannabis use in the
Table 2 Timing of cannabis use and psychosis onset in male and female patients (N = 358). Male
Age at onset of psychosis Age at onset of cannabis use Start in pre-psychotic phase, n (%) Start in DUP, n (%)
Female
CUD (n = 151; 68.9%)
NCUD (n = 68; 31.1%)
CUD (n = 50; 36.0%)
NCUD (n = 89; 64.0%)
21.9 (3.3) 17.1 (3.1) 135 (91.9%) 12 (8.2%)
21.0 (3.7) NA NA NA
21.7 17.3 41 7
21.1 (4.1) NA NA NA
Abbreviations: CUD = cannabis use disorder, NCUD = no cannabis use disorder; and DUP = duration of untreated psychosis.
(3.5) (3.1) (85.5%) (14.6%)
B.G. Schimmelmann et al. / Schizophrenia Research 129 (2011) 52–56
catchment area of EPPIC in Melbourne (Wade et al., 2006) or due to the non-informed consent design avoiding selection biases towards patients without SUD. Two hundred and one FEP patients had CUD only and were compared to 158 patients without CUD. Previous studies have reported that onset of cannabis use precedes onset of psychosis in most patients. This was clearly confirmed by our study in that almost 90% of FEP patients with CUD only had started cannabis use before psychosis onset. Patients with CUD only were more likely male compared to NCUD. Thus, gender was controlled for, when AAO differences were considered. 4.2. Key findings Contrary to findings of a general effect of cannabis use on lowering AAO in psychotic patients (Large et al., 2011), we did not detect an AAO difference between those with CUD only and NCUD. Only early cannabis use was associated with an earlier AAO in adults when compared to NCUD. This effect was relatively small and explained merely 2.6% of the variance in AAO. Recently, Goldberger et al. (2010) also failed to find a general difference in AAO between lifetime cannabis users and non-users in a large sample of male schizophrenia patients. Yet, a markedly earlier AAO (2.6 years) in so-called cannabissensitive users in comparison to non-cannabis-sensitive users was detected. Thereby, cannabis-sensitive patients were characterized by an onset of psychotic symptoms shortly after initiation or a marked increase of cannabis use. Interestingly, cannabis-sensitivity was independently related to both a positive family history of psychosis and younger age, when cannabis use started. While the design of our study, unfortunately, did not allow for a similar measure of cannabissensitivity, our finding, that early start of cannabis use was associated with earlier AAO may be a proxy of a negative influence of cannabis on brain maturation in a sensitive (early) developmental period. The small effect size of this association is also in line with the cannabissensitivity hypothesis and might be explained by gene environment interactions, i.e., cannabis may exert its effect on an earlier onset of psychosis only in genetically vulnerable subjects (Henquet et al., 2008; Genetic Risk and Outcome in Psychosis (GROUP) Investigators, 2011). Another possible explanation for the weak effect size might lie in our cut-off for cannabis use. The CUD group only contained patients with manifest cannabis abuse or dependency according to DSM-IV, while other occasional cannabis use was not assessed and included in the NCUD group. Since the threshold for cannabis use causing earlier AAO might in fact be lower, stratifying the sample according to the quantity of cannabis used would potentially allow a better differentiation of the influence of cannabis use on AAO. 4.3. Limitations Limitations of file audit studies are usually linked to the following elements: poor quality of file entries and absence of strategies to ensure inter-rater reliability and validity of the data. In the present study, huge efforts were made to reduce these limitations in order to minimize the price to pay for getting representative data as described in detail previously (Conus et al., 2007). Kappa-values revealed a good concordance for substance use disorder diagnoses (kappa = 0.74). Also, psychoses diagnoses were retrospectively established by reviewing all information available in charts, and good convergent validity was confirmed in 114 patients of the sample (kappa = 0.80). Another limitation is the absence of more standardized measures for the assessment of premorbid functioning, e.g., the Premorbid Adjustment Scale. Further, it is important to critically consider the validity of data regarding the chronology of onset of cannabis use and of psychosis. Although the exactness of retrospective dating might be limited by recall bias, retrospective chronology assessments of cannabis use are the only way to gather this data in already psychotic patients. Alternatively, large and expensive prospective epidemiolog-
55
ical studies, continuously assessing both the emergence of cannabis use and of psychosis, are required (Kuepper et al., 2011). With a median reported time between onset of cannabis use and onset of psychosis of 4.6 (quartiles 2.5–7.1) years, however, our data seem to justify the statement that the vast majority of patients with CUD started cannabis use before onset of psychosis. Generally, a team of researchers with established inter-rater reliability applying one standardized interview on DUP and CUD onset per patient will most likely produce more reliable results. However, these data may not be more valid than those extracted from files in this study, because longitudinal and external information is commonly disregarded. The strengths of this study are (i) the large sample size allowing for the comparison of patients with CUD only – excluding other SUD – and those without CUD or any other SUD and (ii) that a well-trained and experienced clinical team at EPPIC routinely rated the chronology of cannabis and psychosis onset based on all information in the file collected over an 18-month treatment period from patients and their families (McGorry et al., 1990a,b). 5. Conclusions Lifetime CUD was not associated with an earlier AAO. However, in line with previous epidemiological studies showing an effect of early cannabis use on the emergence of psychosis, cannabis use starting at age 14 or younger was associated with an earlier AAO when compared to NCUD at a small effect size. Also, within the patient group with CUD, early cannabis onset was predictive of an earlier AAO. These results are in line with the hypothesis that it is not a general psychotogenetic effect of cannabis that may lower AAO in adults, but an indirect effect on brain maturation in a sensitive developmental period and only in vulnerable subjects (Goldberger et al., 2010; Genetic Risk and Outcome in Psychosis (GROUP) Investigators, 2011). Future studies should model the interactive effects of age at onset and amount/ severity of cannabis use on AAO and further explore the concept of cannabis-sensitivity. Role of funding source Eli Lilly Australia funded the assessment of files on a subgroup of this cohort treated with olanzapine and risperidone. Eli Lilly did not participate in or influence the data collection, data analyses or write-up of the manuscript. Contributors Drs. Lambert, Conus and McGorry designed this large First Episode Outcome Study. Dr. Lambert and Conus collected the data. Dr. Schimmelmann analyzed and interpreted the data. Drs. Schimmelmann and Lambert wrote the first draft of this manuscript. All authors contributed to and have approved the final manuscript. Conflicts of interest The acquisition of data was in part supported by the Colonial Foundation, National Health & Medical Research Council and Eli Lilly Australia. Drs. Schimmelmann, Conus, McGorry, and Lambert have received research funding from and/or served as paid speakers for Eli Lilly. Acknowledgments The study was in part supported by Eli Lilly Australia. Philippe Conus was supported by a grant from the Leenaards Foundation, Switzerland. The authors express their gratitude towards Pietro Ballinari who provided additional statistical support.
References APA, 1994. Diagnostic Criteria from DSM-IV. American Psychiatric Publishing, Washington, DC. APA, 2000. Diagnostic & statistical manual of mental disorders, Text Revision (DSM-IVTR). Fourth Edition. American Psychiatric Publishing, Washington, DC. 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. Arseneault, L., Cannon, M., Witton, J., Murray, R.M., 2004. Causal association between cannabis and psychosis: examination of the evidence. Br. J. Psychiatry 184, 110–117.
56
B.G. Schimmelmann et al. / Schizophrenia Research 129 (2011) 52–56
Barnes, T.R., Mutsatsa, S.H., Hutton, S.B., Watt, H.C., Joyce, E.M., 2006. Co-morbid substance use and age at onset of schizophrenia. Br. J. Psychiatry 188, 237–242. Caspi, A., Moffitt, T.E., Cannon, M., et al., 2005. Moderation of the effect of adolescentonset cannabis use on adult psychosis by a functional polymorphism in the catechol-O-methyltransferase gene: longitudinal evidence of a gene × environment interaction. Biol. Psychiatry 57 (10), 1117–1127. Conus, P., Cotton, S., Schimmelmann, B.G., McGorry, P.D., Lambert, M., 2007. The FirstEpisode Psychosis Outcome Study: premorbid and baseline characteristics of an epidemiological cohort of 661 first-episode psychosis patients. Early Interv. Psychiatry 1 (2), 191–200. Friis, S., Larsen, T.K., Melle, I., Opjordsmoen, S., Johannessen, J.O., Haahr, U., 2003. Methodological pitfalls in early detection studies — the NAPE Lecture 2002. Nordic Association for Psychiatric Epidemiology. Acta Psychiatry Scand. 107 (1), 3–9. Genetic Risk and Outcome in Psychosis (GROUP) Investigators, 2011. Evidence that familial liability for psychosis is expressed as differential sensitivity to cannabis: an analysis of patient–sibling and sibling–control pairs. Arch. Gen. Psychiatry 68, 138–147. Goldberger, C., Dervaux, A., Gourion, D., Bourdel, M.C., Loo, H., Laqueille, X., et al., 2010. Variable individual sensitivity to cannabis in patients with schizophrenia. Int. J. Neuropsychopharmacol. 13 (9), 1145–1154. Gonzalez-Pinto, A., Vega, P., Ibanez, B., Mosquera, F., Barbeito, S., Gutierrez, M., 2008. Impact of cannabis and other drugs on AAO. J. Clin. Psychiatry 69 (8), 1210–1216. Henquet, C., Di Forti, M., Morrison, P., Kuepper, R., Murray, R.M., 2008. Gene– environment interplay between cannabis and psychosis. Schizophr. Bull. 34, 1111–1121. Konings, M., Henquet, C., Maharajh, H.D., Hutchinson, G., Van Os, J., 2008. Early exposure to cannabis and risk for psychosis in young adolescents in Trinidad. Acta Psychiatry Scand. 118 (3), 209–213. Kuepper, R., van Os, J., Lieb, R., Wittchen, H.U., Höfler, M., Henquet, C., 2011. Continued cannabis use and risk of incidence and persistence of psychotic symptoms: 10 year follow-up cohort study. BMJ. doi:10.1136/bmj.d738. Large, M., Sharma, S., Compton, M.T., Slade, T., Nielssen, O., 2011. Cannabis Use and Earlier Onset of Psychosis: A Systematic Meta-analysis. Arch. Gen. Psychiatry. doi:10.1001/archgenpsychiatry.2011.5.
Malone, D.T., Hill, M.N., Rubino, T., 2010. Adolescent cannabis use and psychosis: epidemiology and neurodevelopmental models. Br. J. Pharmacol. 160 (3), 511–522. McGorry, P.D., Copolov, D.L., Singh, B.S., 1990a. Royal Park multidiagnostic instrument for psychosis: part I. Rationale and review. Schizophr. Bull. 16 (3), 501–515. McGorry, P.D., Singh, B.S., Copolov, D.L., Kaplan, I., Dossetor, C.R., van Riel, R.J., 1990b. Royal Park multidiagnostic instrument for psychosis: part II. Development, reliability, and validity. Schizophr. Bull. 16 (3), 517–536. Menezes, N.M., Arenovich, T., Zipursky, R.B., 2006. A systematic review of longitudinal outcome studies of first-episode psychosis. Psychol. Med. 36 (10), 1349–1362. Schimmelmann, B.G., Conus, P., Cotton, S.M., Kupferschmid, S., McGorry, P.D., Lambert, M. (in press) Prevalence and impact of cannabis use disorders in adolescents with early onset first episode psychosis. Eur. Psychiatry. Schimmelmann, B.G., Conus, P., Cotton, S., McGorry, P.D., Lambert, M., 2007. Pretreatment, baseline, and outcome differences between early-onset and adult-onset psychosis in an epidemiological cohort of 636 first-episode patients. Schizophr. Res. 95, 1–8. Schimmelmann, B.G., Huber, C.G., Lambert, M., Cotton, S., McGorry, P.D., Conus, P., 2008. Impact of duration of untreated psychosis on pre-treatment, baseline, and outcome characteristics in an epidemiological first-episode psychosis cohort. J. Psychiatry Res. 42, 982–990. Sevy, S., Robinson, D.G., Napolitano, B., Patel, R.C., Gunduz-Bruce, H., Miller, R., 2010. Are cannabis use disorders associated with an earlier AAO? A study in first episode schizophrenia. Schizophr. Res. 120 (1–3), 101–107. Ventura, J., Liberman, R.P., Green, M.F., Shaner, A., Mintz, J., 1998. Training and quality assurance with the Structured Clinical Interview for DSM-IV (SCID-I/P). Psychiatry Res. 79 (2), 163–173. Wade, D., Harrigan, S., Edwards, J., Burgess, P.M., Whelan, G., McGorry, P.D., 2006. Substance misuse in first-episode psychosis: 15-month prospective follow-up study. Br. J. Psychiatry 189, 229–234. Zammit, S., Moore, T.H., Lingford-Hughes, A., Barnes, T.R., Jones, P.B., Burke, M., 2008. Effects of cannabis use on outcomes of psychotic disorders: systematic review. Br. J. Psychiatry 193 (5), 357–363.
Contents lists available at ScienceDirect
Schizophrenia Research j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s c h r e s
Cannabis use disorder and age at onset of psychosis — A study in first-episode patients Benno G. Schimmelmann a,⁎, Philippe Conus b,c, Sue M. Cotton c, Stephan Kupferschmid a, Anne Karow d, Frauke Schultze-Lutter a, Patrick D. McGorry c, Martin Lambert c,d a
University Hospital of Child and Adolescent Psychiatry, Bern, Switzerland Treatment and early Intervention in Psychosis Program (TIPP), Department of Psychiatry, CHUV, Lausanne, Switzerland Orygen Youth Health and Research Centre, Centre for Youth Mental Health, Melbourne, Australia d Psychosis Early Detection and Intervention Centre (PEDIC), Department of Psychiatry and Psychotherapy, University Medical Center Hamburg-Eppendorf, Germany b c
a r t i c l e
i n f o
Article history: Received 17 December 2010 Received in revised form 21 March 2011 Accepted 25 March 2011 Available online 17 April 2011 Keywords: First-episode psychosis Schizophrenia Substance use Adolescence
a b s t r a c t Introduction: Age at onset of psychosis (AAO) may be younger in patients with cannabis use disorders (CUD) compared to those without CUD (NCUD). Previous studies included CUD co-morbid with other substance use disorders (SUD), and many did not control for confounders. Methods: Controlling for relevant confounders, differences in AAO between patients with and without CUD excluding those with any other SUD were analyzed in a large representative file audit of 625 first-episode psychosis (FEP) patients (age 14 to 29 years) admitted to the Early Psychosis Prevention and Intervention Centre in Melbourne, Australia. Results: Three quarters of the 625 FEP patients had a CUD. Cannabis use started before psychosis onset in 87.6% of patients. AAO was not significantly different between CUD (without other SUD, n = 201) and NCUD (n = 157). However, AAO was younger in those with early CUD (starting age 14 or younger) compared to NCUD (F(1) = 5.2; p = 0.024; partial η2 = 0.026). Earlier age at onset of cannabis use predicted earlier age at onset of psychosis (β = − 0.49, R2-change = 0.25, p b 0.001). Conclusion: Only CUD starting age 14 or younger was associated with an earlier AAO at a small effect size. These findings suggest that CUD may exert an indirect effect on brain maturation resulting in earlier AAO potentially only in cannabis sensitive subjects. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Previous studies support the role of cannabis as a risk factor for psychosis. Arseneault et al. (2004) systematically reviewed all prospective epidemiological studies. The authors concluded that cannabis use confers an overall twofold increase in the relative risk for later schizophrenia. Notably, the effect of cannabis exposure on the presence of psychotic symptoms was reported to be significantly stronger when exposure occurs in early adolescence (Arseneault et al., 2002; Konings et al., 2008) and when cannabis use was persistent (Kuepper et al., 2011). Konings et al. (2008) used age 14 as cut-off for early cannabis exposure. Interestingly, the literature also suggests that age at onset of psychosis (AAO) is younger in patients abusing cannabis compared to those who do not, on the basis of data stemming from retrospective studies conducted in adult-onset multiple- or first-episode psychosis (Large et al., 2011). ⁎ Corresponding author at: University Hospital of Child and Adolescent Psychiatry, Bolligenstr. 110, 3000 Bern 60, Switzerland. Tel.: +41 31 932 8564; fax: +41 31 932 8569. E-mail address: [email protected] (B.G. Schimmelmann). 0920-9964/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.schres.2011.03.023
Several explanations have been suggested for the relationship between cannabis use and psychosis: First, there is evidence that the endocannabinoid system plays an important role in brain developmental processes in those brain regions (prefrontal cortex, limbic structures and hippocampus) and neurotransmitter systems (GABAergic and dopaminergic transmission), which are pathophysiologically relevant to schizophrenia (Malone et al., 2010). Thus, a causal link between cannabis use and the emergence of psychosis and potentially also between early cannabis use and a younger AAO is intuitively likely, particularly if early cannabis use or abuse interferes with brain developmental processes in adolescence. A second hypothesis proposes that cannabis use causes psychosis by means of gene environment interaction (Caspi et al., 2005). For example, the study by Caspi et al. (2005) found that a functional polymorphism in the catechol-O-methyltransferase (COMT) gene moderated the influence of adolescent cannabis use on developing adult psychosis. A third hypothesis postulates that the earlier AAO in cannabis users may be the result of factors unrelated to cannabis such as gender or pre-morbid adjustment. The respective studies, however, report inconsistent results (Barnes et al., 2006; Gonzalez-Pinto et al., 2008; Sevy et al., 2010).
B.G. Schimmelmann et al. / Schizophrenia Research 129 (2011) 52–56
Limitations of many previous studies on the link between cannabis use and AAO included (Zammit et al., 2008): (i) small sample sizes; (ii) failure to report effect sizes; (iii) failure to control for relevant confounders, and (iv) lack of representativeness of samples due to an informed consent procedure, which patients with severe and/or severe psychoses are more likely to refuse (Friis et al., 2003; Wade et al., 2006). To date, no study explored AAO differences between those with cannabis use disorders (CUD) only, excluding co-morbidity with other substance use disorders (SUD), and those without CUD. Therefore, based on a large cohort of 625 adolescent and adult first episode psychosis patients (age 14–29 years), we examined differences in AAO between patients with lifetime CUD only (n = 201) and patients without CUD (n = 158), taking into account the potential confounding effect of gender and/or premorbid functioning. Based on the literature, we expected that (i) AAO would be younger in patients with CUD only compared to those without CUD (NCUD) (Large et al., 2011) and (ii) the effect of CUD on younger AAO would be even more pronounced if early CUD (cannabis use starting at age 14 or younger) were compared to NCUD. 2. Methods 2.1. Context and sample This is a retrospective file-audit study. Data were sourced from the First Episode Psychosis Outcome Study (Conus et al., 2007). The initial sample comprised a population-based cohort of 786 first-episode psychosis (FEP) patients, consecutively admitted to the Early Psychosis Prevention and Intervention Centre (EPPIC) in Melbourne, Australia, between 1998 and 2000. EPPIC covered a catchment area of approximately 880,000 people, and had a mandate to offer treatment to all FEP patients aged 14–29. As such, this clinical sample is approaching epidemiological representativeness for urban settings. The EPPIC program comprises a comprehensive early intervention treatment program, with a usual episode of care of 18 months, which encompasses extensive assessments. Each patient is assigned to a core team, who treats the patient across settings and makes entries into the chart (McGorry et al., 1990a,b). Of the 786 patients admitted to EPPIC, 82 files (10%) had been sent to other services. The excluded patients did not differ in available demographic characteristics (age and gender) and diagnostic distribution at baseline. 57 (8.1%) of the remaining 704 patients were excluded from the study due to a final diagnosis of non-psychotic disorder, substance-induced psychosis or psychosis due to general medical conditions. Further, 11 (1.6%) patients were excluded due to missing data regarding CUD and 8 (1.1%) due to missing data regarding AAO. Further, three outliers regarding AAO (age 8.21, 8.36 and 9.97) were detected in preliminary analyses and excluded. Out of the remaining 625 patients, 267 (42.7%) were excluded due to CUD co-morbid with other SUD. Data from 358 FEP patients with CUD only (n = 201) or NCUD (n = 157) was analyzed.
53
approval for this ‘non-informed consent’ file audit study allowing for the unselected inclusion of patients. 2.3. Assessment of diagnoses Clinical diagnoses (psychoses and SUD including CUD) according to DSM-IV criteria (APA, 2000) were based on the consensus of trained clinicians of the specialized assessment and crisis intervention team following an intensive diagnostic process within the first 6 weeks after entry into service. The main investigators (ML & PC) re-assessed all information available in charts with respect to baseline diagnoses. Validity of the file audit diagnoses was established by the following procedure: Between 1998 and 2000, 230 of the 786 patients treated at EPPIC had been included in prospective trials. Their main and co-morbid diagnoses were defined within 6 weeks of admission using the Structured Clinical Interview for DSM-IV (SCID-I/P; Ventura et al., 1998). The SCID and file audit diagnoses of 115 patients randomly selected within this sample of 230 were compared. The calculated kappa values revealed a very good concordance for both psychosis diagnoses (kappa = 0.80) and SUD diagnoses (kappa = 0.74). 2.4. Assessment of cannabis use disorders The prevalence of lifetime CUD, i.e., cannabis abuse or dependency according to DSM-IV criteria, as well as the age at onset of cannabis use was assessed with the Drug and Alcohol Assessment Schedule (DAAS) (McGorry et al., 1990a,b) during the first 6 weeks after service entry. Any cannabis use pattern, which did not fulfill the DSM-IV criteria for cannabis use disorders, was not assessed. The principal investigators (ML and PC) extracted this data from the files. As for some patients, the existence of lifetime CUD became only known during treatment, the CUD classification of the study was not based on DAAS results alone but on all available information in the files collected throughout the 18-month treatment period at EPPIC (Schimmelmann et al., in press). The calculated kappa value revealed a good concordance between prospectively assessed SUD diagnoses and those from the file-audit used in this study (kappa = 0.74). 2.5. Assessment of duration of untreated psychosis (DUP) and age at onset DUP was defined as age at entry into EPPIC subtracted by age when first sustained positive psychotic symptoms started (Schimmelmann et al., 2007, 2008). Clinicians at EPPIC were trained to determine DUP with patients and their relatives. In several cases, this date was modified in the file over the course of treatment on the basis of new information being gathered: For example, as patients became clinically more stable and therefore more able to reflect on symptoms' onset, they provided further details on the evolution of symptoms allowing for a more accurate estimate of DUP. The main investigators made a final decision based on all available data in the file. 2.6. Assessment of pre-treatment variables
2.2. Procedure All information on pre-treatment, baseline (at the time of admission to EPPIC), treatment and outcome variables for each patient treated at EPPIC is systematically documented in one standardized file including information over the treatment period as assessed with the Royal Park Multi-diagnostic Instrument for Psychosis (McGorry et al., 1990a,b). Using a systematic comprehensive approach (Early Psychosis File Questionnaire (EPFQ)) (Conus et al., 2007), all charts were rated by two experienced psychiatrists well acquainted with the EPPIC clinical service and treatment of young patients with FEP (ML & PC). A local ethics committee granted
Pre-treatment characteristics extracted from the files included the highest premorbid functioning according to the Global Assessment of Functioning Scale (GAF; APA, 1994), as the commonly used Premorbid Adjustment Scale was too complex to reliably extract data from files. Inter-rater reliability (ML and PC) was established for premorbid GAFscores (kappa = 0.88). 2.7. Data analysis and statistical tests Descriptive statistics were applied to compare CUD and NCUD in terms of gender and pre-treatment characteristics using Mann–
54
B.G. Schimmelmann et al. / Schizophrenia Research 129 (2011) 52–56
Table 1 Sample characteristics in patients with and without CUD (N = 358). Clinical characteristics
Pre-treatment characteristics Gender, n (%) Male Female Premorbid GAF score, mean (SD) DUP in months, median (quartiles) Diagnosis, n (%) Schizophrenia Schizophreniform disorder Schizoaffective disorder Bipolar disorder I Other psychoses
Cannabis use disordersa
Totala
Statistics
NCUD (n = 157; 43.9%)
CUD (n = 201; 56.1%)
(N = 358)
p-valueb
Effect sizec
68 89 71.5 2.0
151 50 71.8 2.0
(75.1) (24.9) (9.9) (0.3–6.1)
219 (61.2) 139 (38.8) 71.7 (10.2) 2.0 (1.0–7.0)
b0.001
0.32
0.853 0.424 0.380
NAd NAd NAd
(19.4) (44.8) (6.5) (18.4) (10.9)
74 (20.7) 155 (43.3) 17 (4.7) 68 (19.0) 44 (12.3)
35 65 4 31 22
(43.3) (56.7) (10.6) (1.0–7.1) (22.3) (41.4) (2.5) (19.7) (14.0)
39 90 13 37 22
Abbreviations: GAF = Global Assessment of Functioning Scale, DUP = duration of untreated psychosis, CUD = cannabis use disorder, NCUD = no cannabis use disorder, and SD = standard deviation. a Ns vary due to missing data. b Descriptive statistics: p-values χ2 for categorical variables and Mann–Whitney non-parametric tests for continuous variables. c Effect size = Cramer-V for χ2 tests. d NA (not applicable) is used for non-significant findings, for whom no effect size is provided.
Whitney U-tests for premorbid functioning and DUP and χ2-tests, when the dependent variable was categorical. Group differences in AAO were examined by analyses of covariance (ANCOVA) controlling for relevant confounders. Further, ANCOVA was also used to compare AAO in early CUD (cannabis use starting age 14 or younger) and NCUD. Within the CUD group, we further explored the predictive validity of age, when cannabis use started (henceforth cannabis onset) on AAO controlling for gender by means of linear regression analysis and reported change of R2, i.e., the variance of AAO explained by cannabis onset independent from gender. Careful inspection of residuals and multivariate outliers was performed in all analyses. To estimate effect sizes, Cramer's-V was used for χ2 analyses and partial η2 in ANCOVAS. Analyses were performed using the Statistical Package for Social Sciences Version 17 (SPSS 17.0; Chicago, IL). 3. Results 3.1. Sample characteristics and prevalence of CUD The sample is characterized in Table 1. Two hundred and one patients had co-morbid CUD and no other SUD; this number is equivalent to 32.2% of the original representative sample of 625 first episode patients; 158 patients had no CUD equivalent to 25.3% of the original sample. Compared to NCUD patients, CUD patients were more likely to be male (χ2 = 37.6; df = 1; p b 0.001, Cramer's V = 0.32). The groups did not differ with respect to premorbid functioning and DUP (Table 1). Accordingly, AAO comparisons between CUD and NCUD were controlled for gender. The mean AAO was 21.5 years (Md = 21.4; SD = 3.6; range 12 to 29), and the mean age of cannabis onset was 17.1 (Md = 17.0; SD = 3.1; range 10 to 27). Nineteen percent of those with CUD (n = 37) started cannabis at age 14 or younger. Notably, the majority of patients with CUD (87.6%; n = 176) started cannabis use before onset of sustained positive symptoms. Table 2 displays the timing of
psychosis and cannabis onset in males and females. All differences between males and females were non-significant. 3.2. Psychosis onset and CUD CUD and NCUD did not significantly differ with respect to AAO, when gender was controlled for (F(1) = 3.4; p = 0.067; partial n2 = 0.01). The mean age at onset of psychosis (adjusted for gender) in CUD was 21.8 (95% CI = 21.3–22.3) and 21.1 (95% CI = 20.5–21.7) in NCUD. However, patients whose cannabis onset was age 14 or younger (n = 37) had a significantly younger AAO compared to NCUD when gender was controlled for (F(1) = 5.2; p = 0.024; partial η2 = 0.026). The mean age at onset of psychosis (adjusted for gender) in early CUD was 19.4 (95% CI = 18.2–20.7) and 21.1 (95% CI = 20.5– 21.7) in NCUD. Notably, within the CUD group (n = 201) and controlling for gender, age at cannabis onset predicted AAO (linear regression analysis; β = − 0.49, R2-change = 0.25, p b 0.001). In other words, age at cannabis onset explained 25% variance of AAO. 4. Discussion This is the first study comparing age at onset of psychosis in those who only had CUD (and no other SUD) to those without CUD. Using a file audit design, common biases of less representative informed consent studies were avoided, such as the exclusion of patients with more severe psychopathology and substance use disorders (Friis et al., 2003; Menezes et al., 2006). 4.1. Sample In the original sample of 625 FEP patients, almost three quarters had lifetime CUD, and about one third had CUD only and no other SUD. This rate is higher than reported in other samples (for review see Sevy et al., 2010) and may be due to the high rate of cannabis use in the
Table 2 Timing of cannabis use and psychosis onset in male and female patients (N = 358). Male
Age at onset of psychosis Age at onset of cannabis use Start in pre-psychotic phase, n (%) Start in DUP, n (%)
Female
CUD (n = 151; 68.9%)
NCUD (n = 68; 31.1%)
CUD (n = 50; 36.0%)
NCUD (n = 89; 64.0%)
21.9 (3.3) 17.1 (3.1) 135 (91.9%) 12 (8.2%)
21.0 (3.7) NA NA NA
21.7 17.3 41 7
21.1 (4.1) NA NA NA
Abbreviations: CUD = cannabis use disorder, NCUD = no cannabis use disorder; and DUP = duration of untreated psychosis.
(3.5) (3.1) (85.5%) (14.6%)
B.G. Schimmelmann et al. / Schizophrenia Research 129 (2011) 52–56
catchment area of EPPIC in Melbourne (Wade et al., 2006) or due to the non-informed consent design avoiding selection biases towards patients without SUD. Two hundred and one FEP patients had CUD only and were compared to 158 patients without CUD. Previous studies have reported that onset of cannabis use precedes onset of psychosis in most patients. This was clearly confirmed by our study in that almost 90% of FEP patients with CUD only had started cannabis use before psychosis onset. Patients with CUD only were more likely male compared to NCUD. Thus, gender was controlled for, when AAO differences were considered. 4.2. Key findings Contrary to findings of a general effect of cannabis use on lowering AAO in psychotic patients (Large et al., 2011), we did not detect an AAO difference between those with CUD only and NCUD. Only early cannabis use was associated with an earlier AAO in adults when compared to NCUD. This effect was relatively small and explained merely 2.6% of the variance in AAO. Recently, Goldberger et al. (2010) also failed to find a general difference in AAO between lifetime cannabis users and non-users in a large sample of male schizophrenia patients. Yet, a markedly earlier AAO (2.6 years) in so-called cannabissensitive users in comparison to non-cannabis-sensitive users was detected. Thereby, cannabis-sensitive patients were characterized by an onset of psychotic symptoms shortly after initiation or a marked increase of cannabis use. Interestingly, cannabis-sensitivity was independently related to both a positive family history of psychosis and younger age, when cannabis use started. While the design of our study, unfortunately, did not allow for a similar measure of cannabissensitivity, our finding, that early start of cannabis use was associated with earlier AAO may be a proxy of a negative influence of cannabis on brain maturation in a sensitive (early) developmental period. The small effect size of this association is also in line with the cannabissensitivity hypothesis and might be explained by gene environment interactions, i.e., cannabis may exert its effect on an earlier onset of psychosis only in genetically vulnerable subjects (Henquet et al., 2008; Genetic Risk and Outcome in Psychosis (GROUP) Investigators, 2011). Another possible explanation for the weak effect size might lie in our cut-off for cannabis use. The CUD group only contained patients with manifest cannabis abuse or dependency according to DSM-IV, while other occasional cannabis use was not assessed and included in the NCUD group. Since the threshold for cannabis use causing earlier AAO might in fact be lower, stratifying the sample according to the quantity of cannabis used would potentially allow a better differentiation of the influence of cannabis use on AAO. 4.3. Limitations Limitations of file audit studies are usually linked to the following elements: poor quality of file entries and absence of strategies to ensure inter-rater reliability and validity of the data. In the present study, huge efforts were made to reduce these limitations in order to minimize the price to pay for getting representative data as described in detail previously (Conus et al., 2007). Kappa-values revealed a good concordance for substance use disorder diagnoses (kappa = 0.74). Also, psychoses diagnoses were retrospectively established by reviewing all information available in charts, and good convergent validity was confirmed in 114 patients of the sample (kappa = 0.80). Another limitation is the absence of more standardized measures for the assessment of premorbid functioning, e.g., the Premorbid Adjustment Scale. Further, it is important to critically consider the validity of data regarding the chronology of onset of cannabis use and of psychosis. Although the exactness of retrospective dating might be limited by recall bias, retrospective chronology assessments of cannabis use are the only way to gather this data in already psychotic patients. Alternatively, large and expensive prospective epidemiolog-
55
ical studies, continuously assessing both the emergence of cannabis use and of psychosis, are required (Kuepper et al., 2011). With a median reported time between onset of cannabis use and onset of psychosis of 4.6 (quartiles 2.5–7.1) years, however, our data seem to justify the statement that the vast majority of patients with CUD started cannabis use before onset of psychosis. Generally, a team of researchers with established inter-rater reliability applying one standardized interview on DUP and CUD onset per patient will most likely produce more reliable results. However, these data may not be more valid than those extracted from files in this study, because longitudinal and external information is commonly disregarded. The strengths of this study are (i) the large sample size allowing for the comparison of patients with CUD only – excluding other SUD – and those without CUD or any other SUD and (ii) that a well-trained and experienced clinical team at EPPIC routinely rated the chronology of cannabis and psychosis onset based on all information in the file collected over an 18-month treatment period from patients and their families (McGorry et al., 1990a,b). 5. Conclusions Lifetime CUD was not associated with an earlier AAO. However, in line with previous epidemiological studies showing an effect of early cannabis use on the emergence of psychosis, cannabis use starting at age 14 or younger was associated with an earlier AAO when compared to NCUD at a small effect size. Also, within the patient group with CUD, early cannabis onset was predictive of an earlier AAO. These results are in line with the hypothesis that it is not a general psychotogenetic effect of cannabis that may lower AAO in adults, but an indirect effect on brain maturation in a sensitive developmental period and only in vulnerable subjects (Goldberger et al., 2010; Genetic Risk and Outcome in Psychosis (GROUP) Investigators, 2011). Future studies should model the interactive effects of age at onset and amount/ severity of cannabis use on AAO and further explore the concept of cannabis-sensitivity. Role of funding source Eli Lilly Australia funded the assessment of files on a subgroup of this cohort treated with olanzapine and risperidone. Eli Lilly did not participate in or influence the data collection, data analyses or write-up of the manuscript. Contributors Drs. Lambert, Conus and McGorry designed this large First Episode Outcome Study. Dr. Lambert and Conus collected the data. Dr. Schimmelmann analyzed and interpreted the data. Drs. Schimmelmann and Lambert wrote the first draft of this manuscript. All authors contributed to and have approved the final manuscript. Conflicts of interest The acquisition of data was in part supported by the Colonial Foundation, National Health & Medical Research Council and Eli Lilly Australia. Drs. Schimmelmann, Conus, McGorry, and Lambert have received research funding from and/or served as paid speakers for Eli Lilly. Acknowledgments The study was in part supported by Eli Lilly Australia. Philippe Conus was supported by a grant from the Leenaards Foundation, Switzerland. The authors express their gratitude towards Pietro Ballinari who provided additional statistical support.
References APA, 1994. Diagnostic Criteria from DSM-IV. American Psychiatric Publishing, Washington, DC. APA, 2000. Diagnostic & statistical manual of mental disorders, Text Revision (DSM-IVTR). Fourth Edition. American Psychiatric Publishing, Washington, DC. 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. Arseneault, L., Cannon, M., Witton, J., Murray, R.M., 2004. Causal association between cannabis and psychosis: examination of the evidence. Br. J. Psychiatry 184, 110–117.
56
B.G. Schimmelmann et al. / Schizophrenia Research 129 (2011) 52–56
Barnes, T.R., Mutsatsa, S.H., Hutton, S.B., Watt, H.C., Joyce, E.M., 2006. Co-morbid substance use and age at onset of schizophrenia. Br. J. Psychiatry 188, 237–242. Caspi, A., Moffitt, T.E., Cannon, M., et al., 2005. Moderation of the effect of adolescentonset cannabis use on adult psychosis by a functional polymorphism in the catechol-O-methyltransferase gene: longitudinal evidence of a gene × environment interaction. Biol. Psychiatry 57 (10), 1117–1127. Conus, P., Cotton, S., Schimmelmann, B.G., McGorry, P.D., Lambert, M., 2007. The FirstEpisode Psychosis Outcome Study: premorbid and baseline characteristics of an epidemiological cohort of 661 first-episode psychosis patients. Early Interv. Psychiatry 1 (2), 191–200. Friis, S., Larsen, T.K., Melle, I., Opjordsmoen, S., Johannessen, J.O., Haahr, U., 2003. Methodological pitfalls in early detection studies — the NAPE Lecture 2002. Nordic Association for Psychiatric Epidemiology. Acta Psychiatry Scand. 107 (1), 3–9. Genetic Risk and Outcome in Psychosis (GROUP) Investigators, 2011. Evidence that familial liability for psychosis is expressed as differential sensitivity to cannabis: an analysis of patient–sibling and sibling–control pairs. Arch. Gen. Psychiatry 68, 138–147. Goldberger, C., Dervaux, A., Gourion, D., Bourdel, M.C., Loo, H., Laqueille, X., et al., 2010. Variable individual sensitivity to cannabis in patients with schizophrenia. Int. J. Neuropsychopharmacol. 13 (9), 1145–1154. Gonzalez-Pinto, A., Vega, P., Ibanez, B., Mosquera, F., Barbeito, S., Gutierrez, M., 2008. Impact of cannabis and other drugs on AAO. J. Clin. Psychiatry 69 (8), 1210–1216. Henquet, C., Di Forti, M., Morrison, P., Kuepper, R., Murray, R.M., 2008. Gene– environment interplay between cannabis and psychosis. Schizophr. Bull. 34, 1111–1121. Konings, M., Henquet, C., Maharajh, H.D., Hutchinson, G., Van Os, J., 2008. Early exposure to cannabis and risk for psychosis in young adolescents in Trinidad. Acta Psychiatry Scand. 118 (3), 209–213. Kuepper, R., van Os, J., Lieb, R., Wittchen, H.U., Höfler, M., Henquet, C., 2011. Continued cannabis use and risk of incidence and persistence of psychotic symptoms: 10 year follow-up cohort study. BMJ. doi:10.1136/bmj.d738. Large, M., Sharma, S., Compton, M.T., Slade, T., Nielssen, O., 2011. Cannabis Use and Earlier Onset of Psychosis: A Systematic Meta-analysis. Arch. Gen. Psychiatry. doi:10.1001/archgenpsychiatry.2011.5.
Malone, D.T., Hill, M.N., Rubino, T., 2010. Adolescent cannabis use and psychosis: epidemiology and neurodevelopmental models. Br. J. Pharmacol. 160 (3), 511–522. McGorry, P.D., Copolov, D.L., Singh, B.S., 1990a. Royal Park multidiagnostic instrument for psychosis: part I. Rationale and review. Schizophr. Bull. 16 (3), 501–515. McGorry, P.D., Singh, B.S., Copolov, D.L., Kaplan, I., Dossetor, C.R., van Riel, R.J., 1990b. Royal Park multidiagnostic instrument for psychosis: part II. Development, reliability, and validity. Schizophr. Bull. 16 (3), 517–536. Menezes, N.M., Arenovich, T., Zipursky, R.B., 2006. A systematic review of longitudinal outcome studies of first-episode psychosis. Psychol. Med. 36 (10), 1349–1362. Schimmelmann, B.G., Conus, P., Cotton, S.M., Kupferschmid, S., McGorry, P.D., Lambert, M. (in press) Prevalence and impact of cannabis use disorders in adolescents with early onset first episode psychosis. Eur. Psychiatry. Schimmelmann, B.G., Conus, P., Cotton, S., McGorry, P.D., Lambert, M., 2007. Pretreatment, baseline, and outcome differences between early-onset and adult-onset psychosis in an epidemiological cohort of 636 first-episode patients. Schizophr. Res. 95, 1–8. Schimmelmann, B.G., Huber, C.G., Lambert, M., Cotton, S., McGorry, P.D., Conus, P., 2008. Impact of duration of untreated psychosis on pre-treatment, baseline, and outcome characteristics in an epidemiological first-episode psychosis cohort. J. Psychiatry Res. 42, 982–990. Sevy, S., Robinson, D.G., Napolitano, B., Patel, R.C., Gunduz-Bruce, H., Miller, R., 2010. Are cannabis use disorders associated with an earlier AAO? A study in first episode schizophrenia. Schizophr. Res. 120 (1–3), 101–107. Ventura, J., Liberman, R.P., Green, M.F., Shaner, A., Mintz, J., 1998. Training and quality assurance with the Structured Clinical Interview for DSM-IV (SCID-I/P). Psychiatry Res. 79 (2), 163–173. Wade, D., Harrigan, S., Edwards, J., Burgess, P.M., Whelan, G., McGorry, P.D., 2006. Substance misuse in first-episode psychosis: 15-month prospective follow-up study. Br. J. Psychiatry 189, 229–234. Zammit, S., Moore, T.H., Lingford-Hughes, A., Barnes, T.R., Jones, P.B., Burke, M., 2008. Effects of cannabis use on outcomes of psychotic disorders: systematic review. Br. J. Psychiatry 193 (5), 357–363.