Cannabis Use and Symptomatic Relapse in First Episode Schizophrenia: Trigger or Consequence? Data From the OPTIMISE Study

Abstract

Background and Hypothesis

This analysis examined the relationship between cannabis use, compliance with antipsychotics and risk for relapse in patients in remission following a first episode of schizophrenia, schizophreniform, or schizoaffective disorder.

Study Design

Analyses were performed on data from a large European study on first episode of schizophrenia, schizophreniform, or schizoaffective disorder (OPTiMiSE). After 10 weeks of antipsychotic treatment, 282/446 patients (63%) met criteria for symptomatic remission; of whom 134/282 (47.5%) then completed a 1-year follow-up. Cross-lagged models and mediation models investigated the temporal relationships between cannabis use, compliance with antipsychotics, social functioning, and symptomatic worsening/relapse.

Study Results

Compared to nonusers, cannabis use increased risk for relapse, adjusted hazard ratio (HR) = 3.03 (SE = 0.32), P < .001, even in patients who were compliant with antipsychotic medication, adjusted HR = 2.89, (SE = 0.32), P < .001. Cannabis use preceded symptomatic worsening and was followed by worsening of Positive and Negative Syndrome Scale total score at the 1-year end-point (standardized β = 0.62, SE = 0.19, P = .001) and by worsening of social functioning (coef = −0.66, P ≤ .001).

Conclusions

In patients in remission from their first episode of schizophrenia, schizophreniform, or schizoaffective disorder, cannabis use increases the rate of relapse in both compliant and noncompliant individuals. Importantly, the temporal relationship between cannabis and relapse was that cannabis use preceded later relapse, noncompliance, and decrease in social functioning, and not that patients began to relapse, then used cannabis. Further research with a precision psychiatry approach might identify those patients in particular danger of relapse when using cannabis.

relapse, cannabis, first episode psychosis

Introduction

Between 60% and 80% of patients treated with antipsychotics for their first psychotic episode experience symptomatic improvement, and some achieve remission.^1,2 However, the first relapse responds less well to treatment, and is a harbinger of a deteriorating course of illness.^3–6 Hence, identifying preventable risks for relapse in patients in remission after the first episode of schizophrenia, schizophreniform, or schizoaffective disorder is of paramount importance. The relationship between cannabis use and relapse has been tested in recent studies⁷ and presented in meta-analyses on cannabis use in both early-stage and chronic multi-episode schizophrenia patients, showing that continued cannabis use is associated with more relapses, longer hospital admissions, and more positive symptoms at follow-up.^8,9 The association between cannabis use, poor compliance, and relapse has also been studied in cohort studies of both first episode^10–12 and chronic schizophrenia patients.^13–17 However, the temporal relationship between cannabis use and poor compliance in relapse are far from elucidated.

This study is unique in that it addresses questions on the effect of cannabis on the pathway to relapse in patients in remission from their first episode of schizophrenia, schizophreniform, or schizoaffective disorder. Do patients in remission experience symptomatic worsening, become noncompliant with antipsychotic medication and as part of the relapse use cannabis; or do they use cannabis, become noncompliant with antipsychotic medication, which in turn leads to relapse? Can patients in remission after their first episode of schizophrenia, schizophreniform, or schizoaffective disorder, who are compliant with antipsychotics, safely use cannabis? What is the effect of cannabis use after remission on symptomatology and social functioning?

Methods

Study Population

Data were obtained from Optimization of Treatment and Management of Schizophrenia in Europe (OPTIMISE), a large EU-funded study that included 446 patients diagnosed with a first episode of schizophrenia, schizophreniform, or schizoaffective disorder, from 15 sites throughout Europe and Israel. Inclusion criteria required that patients were aged 18–40 and were in their first episode of schizophrenia, schizophreniform, or schizoaffective disorder, as defined by The Diagnostic and Statistical Manual of Mental Disorders 4th edition (DSM-IV) on the basis of the Mini International Neuropsychiatric Interview Plus (MINI Plus).¹⁸ Patients were excluded from the study if the time elapsed between onset of psychotic symptoms and entry into the study exceeded 2 years, and/or if they used antipsychotic medications for longer than 2 weeks in the previous year prior to recruitment, and/or 6 weeks lifetime. Of the 446 patients recruited who received open-label amisulpride for 4 weeks, 250/446 patients remitted. These rates of remission are in line with similar studies of open-label antipsychotics administered to patients in their first episode of schizophrenia, schizophreniform, or schizoaffective disorder.^19,20 Those who did not respond to open-label amisulpride were randomized to 6 weeks of continuing amisulpride, or switching to olanzapine, 32 additional patients remitted, giving a total of 282 (63%) patients in remission according to the Andreasen criteria after 10 weeks of treatment, who are the focus of this study. The trial was registered on www.clinicaltrials.gov (NCT01248195) and received approval from the respective national regulatory authorities.²¹

Study Design

All 446 patients included in the study first received open-label amisulpride 200–800 mg/day for 4 weeks (Phase I). Symptomatic remission was defined according to the stringent criteria of Andreasen et al. in which the following 8 PANSS items are scored 3 (mild) or below: Delusions, conceptual disorganization, hallucinatory behavior, blunted effect, passive/apathetic social withdrawal, lack of spontaneity, mannerisms and posturing, and unusual thought content.²² Those patients who were not in symptomatic remission based on the Andreasen criteria at the end of these 4 weeks entered (N = 93) a 6-week, randomized, double-blind phase of either continuation of amisulpride 200–800 mg/day or switching to olanzapine 5–20 mg/day (Phase II), 32 additional patients entered remission. A small number of patients who did not respond to either amisulpride or olanzapine received clozapine (Phase III). The main results of the study were published in 2018,²¹ finding that: 1. 56% responded to open-label amisulpride; 2. After 4 weeks of open-label amisulpride, there was no benefit in switching nonresponders to olanzapine, as opposed to continuing on amisulpride; 3. Amisulpride/olanzapine nonresponders received open-label clozapine, and had a decrease of 18.4 points in Positive and Negative Syndrome Scale (PANSS) total. Subsequent to this intervention period, patients were invited to participate in a 1-year follow-up period, during which patients were assessed 6, 12, 18, and 52 weeks from remission (see introduction to web supplementary material for details of measures assessed at each follow-up visit). The 282 patients who were in remission after receiving either 3 or 10 weeks of antipsychotic treatment are the focus of this study.

Patients were assessed using the Clinical Global Impression scale²³ and the PANSS,²⁴ using the original positive, negative, and general subscales, and also using the 5 factors identified by Davis et al..²⁵ These factors included: Withdrawal (N4, passive/apathetic social withdrawal; N2, emotional withdrawal; N1, blunted affect; N3, poor rapport; N6, lack of spontaneity), Delusions (P1, delusions; G9, unusual thought content; P6, suspiciousness/persecution; P3, hallucinatory behavior; P5, grandiosity), Anxiety and depression (G2, anxiety; G4, tension; G6, depression), Disorganization (G5, mannerisms and posturing; G13, disturbance of volition; G11, poor attention; G15: preoccupation), and Poor control (P7, hostility; G14, poor impulse control; P4, excitement; G8, uncooperativeness). Social functioning was measured using the Social and Occupational Functioning Assessment Scale (SOFAS).²⁶ Cannabis use was self-reported (yes/no) at all visits throughout the study. We defined “continued cannabis use” as self-report of cannabis use at remission and/or at any of the follow-up visits. The number of visits in which patients reported use of cannabis were added up and used as a measure of level of cannabis use. Compliance was assessed using the Sellwood Compliance Scale,²⁷ a 4-item scale, and the analyses on compliance examined it both using an approximation of a continuous variable and as a dichotomized variable. The primary outcome variable was “relapse,” defined as a specific question: “did relapse occur?” that the investigator answered at the end of the 1-year follow-up period, and/or psychiatric hospitalization during the follow-up period. The relationship between cannabis use and PANSS scores was also assessed. Patients were asked at each follow-up visit about any hospitalizations since the previous visits. If they reported being hospitalized, medical files from the hospital were also checked.

Statistical Analysis

Cox regression and Kaplan–Meier analyses were used to examine the effect of cannabis use on the primary outcome measure (clinician’s diagnosis of relapse and/or hospitalization), see web supplementary material for further presentations of statistical methods and results (web supplementary material sections (1) survival, (2) mixed-effect model on PANSS, (3) random intercept cross-lagged panel model, and (4) Mediation models). A sensitivity analysis was performed, using only hospitalization as a relapse indicator.

A random intercept cross-lagged panel model^28,29 was used to examine the directionality of effects between cannabis use and the outcomes of interest—PANSS total, 5-factor PANSS scores, SOFAS, and compliance. To examine the effect of time-varying within-individual effects, time-invariant between-individual effects (stable trait-like individual differences) were controlled for in the model. All variables were treated as latent variables. The primary research question was whether patients in remission from their first episode of schizophrenia, schizophreniform, or schizoaffective disorder used cannabis, consequently stopped their medication and relapsed, or if patients who stop medication begin to relapse, and as part of the relapse used cannabis. Cannabis use at a previous time point was examined to predict psychotic symptoms and noncompliance at the next time point, and conversely, the presence of symptoms and noncompliance at earlier time points was examined to predict later cannabis use (see web supplementary material section 3 p. 42). Based on the directionality obtained above, mediation analysis was conducted to test whether cannabis use causes relapse through an increase in either continuous (or dichotomized) noncompliance to antipsychotics, worsening in the delusions, or poor control factors. The mediator and outcome were modeled using linear or logistic regression models depending on whether the dependent variables were continuous or binary. Statistical significance was calculated through bootstrapping with 1000 replicates (see web supplementary material section 4 p. 50). The mediation model is based on the regression model. In both mediation and outcome models, the baseline mediator was controlled for. This was to examine the role of changes in the mediator on the outcome. For example, cannabis use at remission was the predictor, noncompliance at the end of the study (controlling for noncompliance at remission) was the mediator, and relapse at the end of the study was the outcome (see web supplementary material, section 4.1).

Mixed-Effect Model with Repeated Measures with time interaction and with baseline cannabis use as a covariate examined the effect of cannabis use on each individual PANSS item, their subscales and factors, as well as the SOFAS over time, in order to identify which symptoms worsened with cannabis use. To account for the within-participant correlation, participants were treated as a random effect. The fixed effects comprised of continued cannabis use (Yes = 1, No = 0), time (in weeks), and the interaction of cannabis use and time.

An alpha level of 0.05 was used throughout all analyses; P-values between .05 and .10 were interpreted as trends.

R version 3.8.0 was used. The statistical analysis code is available upon request.

Results

Of the 282 patients in remission after receiving either 4 or 10 weeks of antipsychotic treatment, based on the M.I.N.I., 21/282 (7.4%) were diagnosed with schizoaffective disorder, 133/282 (47.2%) were diagnosed with schizophrenia, and 128/282 (45.4%) with schizophreniform disorder. Baseline demographics and clinical characteristics of the 282 patients who met remission criteria are described in table 1 and are the analytic sample. Of these, 134/282 (47.5%) patients remained in the study throughout the follow-up period of 1 year (see CONSORT diagram with details on number of patients and dropouts for each timepoint during follow-up; figure 1). Of these 282 patients, 50 (17.7%) reported using cannabis at remission and/or during the follow-up period, 232 (82.3%) did not report cannabis use.

Table 1.

Baseline Demographics and Clinical Characteristics For All Patients Meeting Remission

Characteristic	All Patients in Remission (N = 282)	According to cannabis use (reported at any point during follow-up)
		Nonusers (N = 232)	Users (N = 50)	P-value
Age, mean (SD)	26.0 (6.2)	26.3 (6.4)	24.7 (5.3)	.093
Female (%)	89 (31.6)	77 (33)	12 (24)	.205
Years of education (SD)	12.5 (3.1)	12.7 (3.2)	11.5 (1.9)	.001
Living Alone (%)	49 (17.4)	40 (17.2)	9 (18)	.898
Psychiatric Diagnosis
Schizophrenia (%)	133 (47.1)	114 (49.1)	19 (38)	.558
Schizoaffective (%)	21 (7.4)	17 (7.3)	4 (8)
Schizophreniform disorder (%)	128 (45.4)	101 (43.5)	27 (54)
Baseline PANSS score, mean (SD)
Total PANSS score	75.3 (19.1)	76.4 (19.7)	70.2 (15.3)	.015
Positive scale	19.6 (5.7)	19.6 (5.7)	19.2 (5.5)	.632
Negative scale	18.5 (7.1)	18.9 (7.2)	16.2 (6.1)	.012
General Psychopathology scale	37.2 (9.8)	37.8 (9.9)	34.7 (8.1)	.045
Baseline CGI Severity Scale, mean (SD)	5.4 (0.9)	5.4 (0.9)	5.3 (0.9)	.347

Fig. 1.

Consort diagram.

Of the 148 patients who dropped out, 16 (10.81%) reported using cannabis at remission. this is the same proportion of cannabis use reported at remission (14/134 = 10.44%) by the patients who remained in the study throughout the follow-up period (see the bottom of figure 1).

Cannabis Use and Relapse

The Cox regression without covariates showed a significant relationship between cannabis use and risk for relapse hazard ratio (HR = 3.03, P < .001; figure 2A). A sensitivity analysis using only hospitalization as an indicator of relapse also showed a significantly increased risk for relapse among cannabis users (HR = 3.51, P < .001, web supplementary material 1.1.1). When controlling for compliance with antipsychotic medication at remission and cannabis use before the amisulpride phase (figure 1A and web supplementary table 1.1.2), the Cox regression showed that continued cannabis use was associated with a 4-fold increased risk for relapse (HR = 4.04, SE (log HR) = 0.39, P < .001; figure 2A). A Kaplan–Meier survival analysis (chi-square test statistic = 13.1, d.f. 1, P = .0003; web supplementary figure 1.1.3) confirmed the association between continued cannabis use and increase in the risk of relapse. There was evidence of a dose–response effect: The more frequently cannabis use was reported, the higher the PANSS total score at the end of the study. (web supplementary table 1.2.1; figure 1.2.1).

Fig. 2.

Survival and mediation analysis. Note. (A) Cox regression with continued cannabis use as a predictor where the time to failure is defined as the earlier of either hospitalization or the clinician’s diagnosis of relapse; (B) Mediation analysis model with cannabis use as the predictor, nonadherence as a mediator, and relapse (the earlier of hospitalization or the clinician’s diagnosis of relapse) as the outcome; (C) Cox regression estimate of the failure probabilities according to cannabis use, the dichotomized adherence score at remission, and their interaction, where the failure time is defined as the earlier of hospitalization or diagnosis of relapse; (D) The proportion of relapse by cannabis use and week 52 noncompliance. The number in parentheses is the number of relapses over the number of total instances. E.g., 11 relapsed among the 80 cannabis nonusers with noncompliance of 1 by week 52.

Compliance With Antipsychotic Medication, Cannabis, and Relapse

We next explored whether the association of cannabis with relapse might be explained by noncompliance, thinking that if cannabis users were noncompliant, that might be the mechanism accounting for relapse. Compliance was entered as an interaction in the Cox regression, either as a continuous variable or a dichotomous variable. Compliance with antipsychotics did not significantly change the effect of cannabis use on the risk of relapse, either when used as a continuous variable (HR = 1.64, SE (log HR) = 0.61, P = .41, web supplementary table 1.3.1; figure 1.3.1), or as a dichotomous variable (HR = 1.52, SE (log HR) = 0.83, P = .61; figure 2C; web supplementary material 1.3.2). When separately examining patients in the low and high compliance subgroups, even patients who are compliant with their antipsychotics, but use cannabis, are at higher risk for relapse (HR = 2.65, SE (log HR) = 0.37, P = .01; figure 2C; web supplementary table 1.3.3). The small number of patients (N = 13) who used cannabis and were noncompliant were at even higher risk for relapse, compared to patients who were compliant, at a trend level (HR = 4.04, SE (log HR) = 0.74, P = .06) (web supplementary table 1.3.3). This was confirmed by the mediation analysis, which showed that the increased risk of relapse in cannabis users was not mediated by compliance (figure 1B and 2B;web-supplementary figure 4.1.3; 4.1.4.) Moreover, when looking at the percent of relapse in subjects divided according to degree of adherence, and comparing cannabis users vs nonusers, the findings showed that most relapses occurred in cannabis users at all levels of compliance, whereas among the cannabis nonusers there is a tendency for fewer relapses at a high degree of noncompliance (see figure 2D).

Time Ordered Effects Between Relapse, PANSS Scores, Cannabis Use, and Compliance With Antipsychotic Medication

A cross-lagged model of PANSS total score and continued cannabis use (figure 3 and web supplementary table 3.2) found that cannabis use at week 12 of the follow-up period was associated with a later worsening of PANSS total score at the 1-year endpoint (standardized β = 0.62, SE = 0.19, P = .001). Furthermore, the effect of cannabis use on relapse was mediated by the PANSS delusion factor (web supplementary figure 4.1.1). In the reverse direction, the PANSS total score at the 12-week assessment was not associated with later cannabis use (standardized β = 0.10, SE = 0.23, P = .66, figure 3 and web supplementary table 3.2). Similarly, cannabis use was associated with later antipsychotic noncompliance (standardized β = 0.66, SE = 0.17, P < .001, figure 3 lower panel and web supplementary table 3.4). These results indicate that cannabis use may be associated with later noncompliance and worsening of symptoms, rather than patients first worsening clinically, for whatever reason, and/or stop medication and only then begin to use cannabis.

Fig. 3.

Random Intercept cross-lagged path model on cannabis use, Positive and Negative Syndrome Scale (PANSS) total, and nonadherence. Note. Numbers in the parentheses are 95% confidence intervals; variables in the square represent the observed variables and variables in the circle represent the latent variables; RI, random intercept; ε, error; (A) RI_P, random intercept of PANSS total; RI_C, random intercept of cannabis use; P1 – P4, observed PANSS total variables at the corresponding time point; C1 – C4, observed cannabis use variable at the corresponding time point; (B) RI_N, random intercept of nonadherence; RI_C, random intercept of cannabis use; N1 – N3, observed nonadherence variable at the corresponding time point; C1 – C3, observed cannabis use variable at the corresponding time point. * P < .05, ** P < .01, *** P < .001.

Cannabis and Social Functioning

The mixed-effect model with time interaction showed that social functioning worsened during follow-up in those who used cannabis (Coef = −0.19, SE = 0.06, P < .001, web supplementary material section 2.3). This is consistent with the cross-lagged model, which shows that patients who used cannabis at week 12 are more likely to later have poor social functioning at the end of follow-up (standardized β = −0.66, P < .001; figure 3 upper panel and web supplement table 3.3). This worsening in social functioning is mediated by an increase in poor control (indirect effect = −7.4, P = .001), see figure 4 lower panel and web supplementary material section 4.2. These results are consistent with the finding that cannabis use is associated with worsening in the PANSS active social avoidance item (G16), t = 2.2, P = .03, see web supplementary material 2.4.30). These analyses further elucidate the pathway to relapse, indicating that as patients use cannabis they may lose control, causing their social functioning to deteriorate.

Fig. 4.

Random intercept cross-lagged path model and mediation model on Social and Occupational Functioning Assessment Scale (SOFAS). Note. Numbers in the parentheses are 95% confidence intervals; variables in the square represent the observed variables and variables in the circle represent the latent variables; RI, random intercept; ε, error; (A) RI_S, random intercept of SOFAS; RI_C, random intercept of cannabis use; S1 – S3, observed SOFAS variables at the corresponding time point; C1 – C3, observed cannabis use variable at the corresponding time point; (B) Mediation analysis model with cannabis use as the predictor, poor control factor as the mediator, and SOFAS score as the outcome. * P < .05, ** P < .01, *** P < .001.

Effect of Cannabis on Symptoms

A mixed-effect model with time interaction examining which aspects of schizophrenia were affected by continued cannabis use over time showed a significant worsening of the PANSS total score (t = 5.03, P < .001), positive symptoms (t = 7.63, P < .001) and general psychopathology (t = 4.48, P < .001), but not negative symptoms (nonsignificant, see web supplementary material 2.1). The 5-factor PANSS analysis, found that continued cannabis use was associated with worsening delusions (t = 6.08, P < .001), poor control (t = 6.77, P < .001), and anxiety–depression (t = 2.20, P = .03), but not withdrawal or disorganization (both nonsignificant, see web supplementary material Section 2.2).

Discussion

This study focused on the sequence and potential causal relationship between cannabis use and symptomatic relapse in a well-defined group of first-episode psychosis patients who responded to antipsychotic treatment and met stringently defined criteria for remission. First, cannabis use is associated with later relapse and worsening of PANSS total at week 52, is mediated by worsening of delusions, and is associated with later medication noncompliance. The cross-lagged analyses show that cannabis use precedes the worsening of the PANSS total at week 52, hence the pathway is not that these patients begin to relapse, and then stop their medication and use cannabis. However, this finding should be considered in light of the fact that adherence was measured 12 weeks after remission, which might be too short of a time frame to accumulate many cannabis-related relapses. Second, cannabis use is associated with an increase in poor control, which proceeds worsening social functioning. These findings were shown with numerous analyses which support each other, thus strengthening the sensitivity of the results.

Finding that cannabis use is associated with relapse even in remitted first-episode patients who are compliant with their medication (cox regression HR = 2.7), is particularly relevant. These patients are mostly young, and are encouraged to interact socially with their peers, interactions which are often associated with cannabis use.³⁰ This puts patients in a bind, as it is hard to expect them to interact socially with their cannabis-using peers, while abstaining from using themselves.

These results are consistent with Schoeler et al. who performed a cross-lagged path analysis showing an effect of cannabis increasing risk of relapse in a time and dose-dependent manner.³¹ Our analysis is different in that we only studied remitted patients, who have a better prognosis. Foti et al. found a bidirectional association between cannabis use and psychotic symptoms.³²

These results are in line with the results of recent studies⁷ and meta-analyses on cannabis use in both early-stage and chronic multi-episode schizophrenia patients, which found that continued cannabis use was associated with more relapses, longer hospital admissions, and more positive symptoms at follow-up.⁸

Our observation that cannabis use worsens social functioning is important in that it identifies another domain of behavior in which cannabis is deleterious in schizophrenia, schizoaffective disorder, and schizophreniform disorder. Cannabis use precedes poor social functioning, and this effect is mediated by the effect of cannabis on losing control. In addition, previous studies have shown cannabis abuse to be related to violent behavior.^33–37 Our finding that cannabis use is associated with increases in the PANSS poor control factor (poor control, hostility, uncooperativeness, and excitement), which is related to violent behavior, supports this. These results indicate that patients with a history of poor control, particularly violent behavior should be particularly discouraged from using cannabis.

One cannot do a randomized study to prove the negative effects of cannabis use. This study adds to the large literature on observational studies documenting the negative effects of cannabis on patients with schizophrenia.

Limitations

This study has substantial limitations, some reflecting that this is a secondary study, not a priori considered in the design of the mother study. This very likely has contributed to some degree of noise/error in the data and, in turn, to the degree of confidence one can have in the findings. For example, some of the analyses concerning cannabis users included very low numbers of patients. For that reason, we used different analytic approaches—survival analysis, cross-lagged analysis, and mixed effects analysis, yielding very similar results which are inherently highly correlated, and multiple testing correction methods that usually assume independence among tests would be unnecessarily conservative. Additionally, the mediators and outcome variables were measured at the same time point and any causal inferences based on the mediation analyses should be interpreted with caution.

The question regarding cannabis use was categorial (yes/no), and did not include quantitative or biological measures of cannabis use; in addition, the study did not include an objective confirmation for the self-reported Sellwood compliance score. Moreover, there were no objective criteria for measuring relapse, and clinicians asked a general question about symptomatic deterioration at any point during the follow-up period. In addition, hospitalization was also reported by patients and only then checked in hospital records. Therefore, if patients did not report being hospitalized this could have led to the risk of underreporting of hospitalizations. In addition, we do not know who are the patients who dropped out of the study relapsed or were doing well in the community. However, the dropouts had the same rates of cannabis use upon remission as the patients who remained in follow-up, indicating that cannabis was not the cause of dropout.

Conclusions

In conclusion, our study strongly supports that cannabis use significantly increases risk for relapse in remitted first-episode psychosis patients, even in those who are compliant with their medication. Further research with a precision psychiatry approach might identify those patients in particular danger of relapse when using cannabis.

Funding

This study is part of the OPTIMISE project, funded by the European Commission Seventh Framework Program (HEALTH-F2-2010-242114).

Conflicts of Interest

Linda Levi: Ms. Levi has no potential conflicts of interest to disclose. Mor Bar Haim: Ms. Bar Haim has no potential conflicts of interest to disclose. Inge Winter-van Rossum: Dr. Winter-van Rossum has no potential conflicts of interest to disclose. Michael Davidson: Dr. Davidson is an employee and owns stock options of Minerva Neurosciences a Biotech developing CNS drugs. Stefan Leucht: In the last three years Stefan Leucht has received honoraria as a consultant and/or advisor and/or for lectures from Alkermes, Angelini, Eisai, Gedeon Richter, Janssen, Johnson and Johnson, Lundbeck, Medichem, Merck Sharpp and Dome, Otsuka, Recordati, Rovi, Sandoz, Sanofi Aventis, Sunovion, TEVA, Medichem. Wolfgang W. Fleischhacker: Dr. Fleischhacker has received research grants from Boehringer Ingelheim, Janssen Cilag, Otsuka, and Lundbeck. He has received speaking fees and advisory board honoraria from AOP Orphan, Boehringer Ingelheim, Janssen, Richter, Roche, Lundbeck, Otsuka, Takeda, Amgen, Teva and Dainippon Sumitomo. Jinyoung Park: Ms. Park has no potential conflicts of interest to disclose.John M. Davis: Dr. Davis has no potential conflicts of interest to disclose. Renè S. Kahn: In the past three years, Dr. Kahn has served as a consultant for Alkermes, Otsuka, Janssen-Cilag, and Sunovion. Mark Weiser: Dr. Weiser has received advisory board/speaker’s fees/consultant fees/owns stock from Dexcel, Janssen, Lundbeck, Minerva, Pfizer, Acadia, Roche and Teva