1. The role of substance abuse in prediction of long-term outcome of schizophrenia - systematic review and meta-analysis Jouko Miettunen jouko.miettunen@oulu.fi Department of Psychiatry, University of Oulu, Finland

2. Conflicts of interest: None

3. Alcohol and other substance use problems are common among individuals with schizophrenia. Many of the previous studies have focused on clinical comorbidity, not on longitudinal studies aiming to look for possible causal associations. Cannabis is a potential risk factor for schizophrenia We aimed to investigate if premorbid or early substance use predicts long-term clinical and social course in schizophrenia.

4. Koskinen J, Löhönen J, Koponen H, Isohanni M, Miettunen J. Rate of cannabis use disorders in clinical samples of patients with schizophrenia: a meta-analysis. Schizophr Bull 2010; 36: 1115-30. Approximately every fourth schizophrenia patient had a diagnosis of cannabis use disorders (35 studies). median current rate of cannabis use disorders was 16.0% and median lifetime rate was 27.1% higher in first-episode vs. long-term patients (current 28.6% vs. 22.0%) Approximately every fifth schizophrenia patient had a diagnosis of cannabis use disorders (60 studies). Koskinen J, Löhönen J, Koponen H, Isohanni M, Miettunen J. Prevalence of alcohol use disorders in schizophrenia - a systematic review and meta-analysis. Acta Psychiatr Scand 2009; 120: 85-96.

5. Prevalence of cannabis use disorder

6. A meta-analysis including nine mainly cross-sectional studies found that substance using patients have more positive, but less negative symptoms (Talamo et al. 2006). In depressive symptoms, Potvin et al. (2007) found that substance using schizophrenia patients have more depressive symptoms, although the difference was smaller than commonly assumed. Other outcomes, such as employment or social functioning, have not been studied in systematic reviews. substance abuse and outcome Potvin S, et al. Austr NZ J Psychiatry 2007; 41: 792-9; Talamo A, et al. Schizophr Res. 2006;86:251–255.

7. A systematic review to identify potentially relevant studies was conducted. Only studies with a follow-up period of at least two years were included. We studied following outcomes: negative, positive and total symptoms clinical remission hospitalizations social functioning employment global outcome Current data search until December 2012 substance abuse and outcome

8. The search identified 9343 unique potentially relevant articles of which 20 studies presenting results from 24 samples met our inclusion criteria. The meta-analysis included 5 to 13 studies in each outcome category. Identification of studies

9. ReferenceSample (M/F)Outcome(s) Addington and Addington 1998 (Canada) 80 (54/26) positive, negative and total symptoms (PANSS) Arndt et al. 1992 (USA)131 (93/38) symptoms (SANS/SAPS), hospitalization, global outcome (GAS), social and overall functioning Blow et al. 1998 (USA)682 (660/22) hospitalizations, symptoms (BPRS), functioning (GAF) Bodén et al. 2009 (Sweden)71 functioning (living independently, work/ studying 50%, and meeting friends) Bühler et al. 2002 (Germany)46 positive and negative symptoms, employment, hospitalization, social disability (DAS-M) Caspari 1999 (Germany)53 (41/12) global outcome (GAS), hospitalization, employment, positive and negative symptoms (BPRS, AMDP) Cleghorn et al. 1991 (Canada)27 (19/8)positive and negative symptoms Dubertret et al. 2006 (France)181 (~118/63)positive and negative symptoms Foti et al. 2010 (USA)162symptoms (SAPS, SANS), global (GAF)

10. ReferenceSample(M/F)Outcome(s) Haro et al. 2011 (worldwide: six areas) 11078clinical and functional (social) remission Hodgins et al. 2008 (UK)225 menreal life functioning Huguelet et al. 1996 (Switzerland)37 (18/19)hospitalization, relapse, functioning (GAF) Juola et al. 2013, Miettunen et al. 2007 (Finland) b 105 (63/43)employment Kozaric-Kovacic et al. 1995 (Croatia) 312 (150/162)social functioning, hospitalization, course Marwaha et al. 2009 (United Kingdom, France, Germany) 213 to 773Employment (losing or getting job) Orlandi and Bersani 2001 (Italy)125 negative, positive, total symptoms, hospitalizations Pencer et al. 2005 (Canada)49employment/ productivity Verdoux et al. 2010 (France)108 (72/36)employment Whitty et al. 2008 (United Kingdom) 97 (63/34) Strauss–Carpenter Scale outcome scale (total score) Yen et al. 2009 (Taiwan)118 (99/19)symptoms (PANSS)

11. In meta-analysis, substance use associated modestly with outcome, all the associations (Spearman’s r) were non- significant and between -0.05 and 0.10. The studied in outcome groups were mainly moderate to high heterogeneous. In meta-regression, studied substance or length of illness did not affect significantly the results (Note: small number of studies) results

12. summary of results

13. Odds Ratio < 1 indicates lower likelihood of achieving remission Haro JM, et al. Cross-national clinical and functional remission rates: Worldwide Schizophrenia Outpatient Health Outcomes (W-SOHO) study. Br J Psychiatry 2011; 199:194-201.

14. although comorbid substance use associates with poorer outcome in schizophrenia, the early substance use has only a modest effect as a predictor of long-term outcome this difference between these two designs may be explained for instance with poorer treatment adherence of dual diagnosed patients. effect of medication? changes in substance use? alcohol vs. cannabis ? covariates? milder disease?

15. Jouko Miettunen, PhD, Katja Kasurinen, BMed, Eetu Oinas, BMed, Juha Käkelä, BMed,Erika Jääskeläinen, MScD, Antti Alaräisänen, MScD Department of Psychiatry, University of Oulu, Oulu, Finland & Department of Psychiatry, Oulu University Hospital, Oulu, Finland Noora Hirvonen, MA Information studies, Faculty of Humanities, University of Oulu Email: jouko.miettunen@oulu.fi