Randomised blinded placebo-placebo controlled studies or case-control studies with a duration of at least four weeks that assessed antipsychotic treatment in patients with substance use disorders were eligible for inclusion. Only randomised studies were considered for patients without psychosis (defined as schizophrenia, schizoaffective disorder and bipolar disorder). Two or more antipsychotic treatments had to be compared in studies of patients with psychosis. Case reports, open-label studies and cross-sectional designs were excluded. Eligible outcomes had to measure craving, alcohol-drug use and/or relapse.
Within included trials, length of follow-up ranged from six to 144 weeks. Within comorbid psychosis studies, almost all focused on alcohol, stimulant and tobacco disorders. Most patients had schizophrenia. Comorbid psychosis regimens included clozapine, olanzapine, risperidone, long-acting risperidone, quetiapine, zuclopenthixol depot and haloperidol. Within non-comorbid psychosis studies, all studies were either of stimulant or alcohol substance use disorder. Non-comorbid psychosis regimens included flupenthixol decanoate, tiapride, amisulpride, olanzapine, quetiapine, aripiprazole, risperidone, long-acting risperidone and flupenthixol decanoate (dosages were reported for most trials).
The number of reviewers who performed the study selection was not reported.