| Study | Reason for exclusion |
|---|---|
| Allen 1997 | Allocation: controlled clinical trial, not randomised. |
| Bai 2003 | Allocation: randomised. Participants: not stabilised on antipsychotic drugs. |
| Bechdolf 2016 | Allocation: randomised. Participants: including only those at clinical high risk for psychosis (CHR). |
| Bo 2017 | Allocation: randomised. Participants: clinically stable for at least 4 weeks, treated with risperidone monotherapy at an optimal dose. Intervention: risperidone (baseline dose), risperidone (gradual dose reduction by 50%), no placebo arm. |
| Bourin 2008 | Allocation: randomised. Participants: not stabilised on antipsychotic drugs. |
| Branchey 1981 | Allocation: not randomised, matched groups. |
| Breier 1987 | Allocation: not randomised. |
| Brown 2018 | Allocation: randomised. Participants: mild‐to‐moderate schizophrenia, unclear clinical stability. Intervention: BI 409306 (inhibitor of phosphodiesterase 9A), not currently approved for schizophrenia, versus placebo. |
| Cather 2018 | Allocation: cluster randomised. Participants: first episode of non‐affective psychosis, Intervention: community care compared to NAVIGATE program, which included individual resilience therapy, family education, supported employment and personalised medication management. The specific effect of maintenance therapy with antipsychotic treatment cannot be assumed from this design. |
| Cheng 2019 | Allocation: randomised. Participants: first episode schizophrenia, stabilised on antipsychotic medication. Intervention: risperidone versus olanzapine versus aripiprazole; participants failing the initially‐assigned antipsychotic were switched to one of the other two. No real placebo or discontinuation arm. |
| Chopra 2019 | Allocation: randomised,no hint for real maintenance study design. Participants: first episode schizophrenia; inclusion criteria allow acute patients. |
| Chouinard 1980 | Allocation: not randomised. |
| Chouinard 1993 | Allocation: randomised. Participants: not clinically stable. |
| Claghorn 1974 | Allocation: randomised. Participants: schizophrenia. Intervention: thiothixene alone versus thiothixene plus group therapy versus chlorpromazine alone versus chlorpromazine plus group therapy. |
| Clark 1967 | Allocation: randomised. Participants: not stabilised on antipsychotic drugs (discontinued medication for at least 6 months before study entry). |
| Collins 1967 | Allocation: not randomised. |
| Condray 1995 | Allocation: not randomised. |
| Curson 1985 | Allocation: not randomised. |
| Degkwitz 1970 | Allocation: not randomised. |
| Diamond 1960 | Allocation: not randomised. |
| Double 1993 | Allocation: randomised. Participants: schizophrenia. Intervention: all participants were on neuroleptics and antiparkinson medication at baseline. They were then randomised to neuroleptics plus continuation of antiparkinson medication versus neuroleptics alone. |
| Durgam 2016 | Allocation: randomised. Participants: acute exacerbation of schizophrenia. |
| Engelhardt 1967 | Allocation: randomised. Participants: outpatients with chronic schizophrenia not truly stabilised on antipsychotic drugs. |
| Fleischhacker 2014 | Allocation: randomised. Participants: schizophrenia, receiving maintenance treatment, stabilised on study drug for at least 8 weeks. Intervention: aripiprazole LAI (400 mg/4 weeks), aripiprazole oral (10 mg/day to 30 mg/day), aripiprazole LAI suboptimal dose (50 mg/4 weeks), no real placebo arm. |
| Francey 2018 | Allocation: randomised. Participants: first episode of psychosis, not stable on antipsychotic medication. |
| Freedman 1982 | Allocation: randomised. Participants: discontinued antipsychotic medication for some weeks, then kept in the study only if showing signs of psychotic exacerbation. |
| Gallant 1964 | Allocation: randomised. Participants: unclear baseline clinical status and unclear whether they were stabilised on antipsychotic medication. Intervention: I. butaperazine, trifluoperazine, inert placebo; II. trifluperidol, trifluoperazine, phenobarbital. Outcome: no predefined outcome of interest. |
| Gitlin 1988 | Allocation: randomised (no further details). Participants: schizophrenia or schizoaffective disorder, stabilised on the same depot medication for 1 year. Intervention: fluphenazine decanoate, placebo (cross‐over design). Outcome: no usable data for relevant outcomes (data up to the point of first cross‐over are not available). |
| Gitlin 2001 | Allocation: randomised. Participants: schizophrenia or schizoaffective disorder, clinically stable and with stabilised maintenance antipsychotic therapy. Intervention: fluphenazine decanoate, placebo; cross‐over design. Outcome: no usable data (data up to the point of first cross‐over are not available). |
| Gleeson 2004 | Allocation: randomised. Participants: first‐episode psychosis. Intervention: treatment as usual (including antipsychotics) versus multimodal relapse prevention therapy (including antipsychotics and cognitive behavioral therapy/family intervention). |
| Goldberg 1967 | Allocation: not randomised. |
| Good 1958 | Allocation: randomised. Participants: schizophrenia. Interventions: chlorpromazine versus placebo. Outcomes: no usable outcomes. |
| Greenberg 1966 | Allocation: randomised. Participants: patients with chronic schizophrenia. Intervention: abrupt versus gradual withdrawal of chlorpromazine, but chlorpromazine was withdrawn from both groups. Thus not appropriate control group. |
| Hine 1958 | Allocation: not randomised. |
| Hirsch 1989 | Allocation: randomised. Participants: schizophrenia, clinically stable for at least 6 months, no florid psychotic symptoms. Intervention: fluphenazine decanoate, placebo. Outcome: no usable data for relevant outcomes. |
| Hirsch 1996 | Allocation: randomised (no further details). Participants: schizophrenia (DSM‐III‐R), clinically stable and receiving maintenance treatment. Intervention: fluphenazine depot, placebo. Outcome: no usable data for relevant outcomes (not presented for the randomised subset). |
| Hunt 1967 | Allocation: not randomised. |
| Ionescu 1983 | Allocation: not randomised. |
| Janecek 1963 | Allocation: randomised. Participants: 50% not diagnosed as with schizophrenia. |
| Johnstone 1988 | Allocation: not randomised. |
| Keefe 2018 | Allocation: randomised. Participants: schizophrenia with relevant negative symptoms, unclear clinical stability. Intervention: MIN‐101 (roluperidone), not currently approved, versus placebo. |
| Kellam 1971 | Allocation: not randomised. |
| Lauriello 2005 | Allocation: randomised. Participants: participants were acutely ill, not stable. |
| Lecrubier 1997 | Allocation: randomised. Participants: not stable, not all on antipsychotics before the study. |
| Liu 2018 | Allocation: randomised. Participants: schizophrenia‐related psychotic disorders, under remitted states. Intervention: maintenance therapy with antipsychotics versus guided dose reduction; no real discontinuation or placebo arm. |
| Loo 1997 | Allocation: randomised. Participants: participants were not stable, most not on antipsychotics before the study. |
| Mahal 1975 | Allocation: randomised. Participants: schizophrenia, on maintenance phenotiazine medication Intervention: pimozide, placebo; cross‐over design. Outcome: no usable data for relevant outcomes. |
| Marder 1994 | Allocation: randomised. Participants: not clinically stable. |
| Mathur 1981 | Allocation: randomised. Participants: chronic schizophrenia, stabilised on antipsychotic treatment for at least 6 months before study entry. Intervention: chlorpromazine, placebo; cross‐over design. Outcome: no usable data for relevant outcomes (data up to the point of first cross‐over are not available). |
| Meehan 2019 | Allocation: randomised. Participants: schizophrenia, not adequately stable. |
| Mefferd 1958 | Allocation: randomised. Participants: men with schizophrenia. Intervention: chlorpromazine versus placebo. Outcome: no usable outcome. |
| Mosher 1975 | Allocation: not randomised. |
| Müller 1982 | Allocation: some of the participants were matched, not randomised. |
| NCT03559426 | Allocation: randomised. Participants: schizophrenia spectrum disorders, currently on antipsychotic medication. Intervention: maintenance treatment with antipsychotics versus dose reduction programme; no real discontinuation or placebo arm. |
| Nishikawa 1989 | Allocation: randomised. Participants: outpatients diagnosed with schizophrenia, in remission at baseline. Intervention: timiperone, sulpiride, placebo. The placebo arm was only retrospective, derived from data from previous studies. |
| Oosthuizen 2003 | Allocation: randomised. Participants: first episode of psychosis, not clinically stable. |
| Pasamanick 1967 | Allocation: randomised. Participants: 152 state hospital patients with schizophrenia (severely impaired at time of enrollment), 29 ambulatory schizophrenia patients (not acutely ill but recruited only if severe enough to warrant hospitalisation). |
| Paul 1972 | Allocation: not randomised. |
| Peet 1981 | Allocation: randomised. Participants: schizophrenia. Intervention: chlorpromazine versus chlorpromazine plus propranolol. |
| Pickar 1986 | Allocation: not randomised. |
| Pickar 2003 | Allocation: not randomised. |
| Pigache 1993 | Allocation: randomised. Participants: chronic schizophrenia. Intervention: chlorpromazine, placebo, orphenadrine. Outcome: no relevant outcome, only auditory attention task. |
| Ran 2002 | Allocation: randomised. Participants: chronic schizophrenia, unclear clinical status, 30% uncovered Intervention: antipsychotic therapy + family psychoeducational intervention, antipsychotic treatment alone, control group (in which quote: "medication was not encouraged nor discouraged"). |
| Rassidakis 1970 | Allocation: not randomised. |
| Ravaris 1965 | Allocation: randomised. Participants: chronic schizophrenia. Intervention: fluphenazine elixir plus placebo injection versus fluphenazine enanthate injection plus oral placebo. |
| Ruiz 1975 | Allocation: randomised. Participants: chronic schizophrenia, same antipsychotic treatment for at least one month before study entry. Intervention: pimozide, placebo. Outcome: no usable data for relevant outcomes (data for the double‐blind phase are not avaiable). |
| Ruiz Veguilla 2013 | Allocation: randomised. Participants: diagnosed with non‐affective psychosis (first episode), receiving antipsychotic treatment for 12 months since clinical stabilisation, at the same dose for at least 4 months. Intervention: continual antipsychotic treatment, treatment discontinuation. Outcome: study not performed (stopped after recruitment of 16 patients), no data available. |
| Schlossberg 1978 | Allocation: randomised. Participants: not stable. |
| Singer 1971 | Allocation: randomised. Participants: unclear if clinically stable, probabily not taking antipsychotics before study entry. Intervention: thiopropazate, placebo; cross‐over design. Outcome: no usable data for relevant outcomes (data up to the point of first cross‐over are not available). |
| Singh 1990 | Allocation: not randomised. |
| Smelson 2006 | Allocation: not randomised. |
| Soni 1990 | Allocation: randomised. Participants: schizophrenia, not stabilised on antipsychotic drugs, because all had been withdrawn from antipsychotic drugs for 8 to 20 months before study start. |
| Stuerup 2017 | Allocation: randomised. Participants: participants with newly diagnosed schizophrenia spectrum disorder, from the outpatient early intervention program (OPUS), meeting remission criteria for at least 3 months before study entry. Intervention: maintenance treatment with antipsychotics versus tapering/discontinuation; the control arm does not necessarily imply complete discontinuation of antipsychotic medication in all cases. |
| Sumitomo 2008 | Allocation: randomised. Participants: schizophrenia, not described as clinically stable in inclusion criteria. |
| Vaddadi 1986 | Allocation: randomised. Participants: schizophrenia. Intervention: depot antipsychotics (fluphenazine depot, flupenthixol depot or clopenthixol depot) plus oral dihomo gammalinolenic acid (DHLA) versus oral DHLA plus placebo injections versus DHLA placebo capsules and placebo injections. What is lacking is a depot antipsychotic only group. |
| Van Kammen 1982 | Allocation: not randomised. |
| Van Praag 1973 | Allocation: randomised. Participants: psychotic participants. Intervention: fluphenazine enanthate versus fluphenazine decanoate. |
| Vanover 2018 | Allocation: randomised. Participants: acute exacerbation of schizophrenia. |
| Weller 2018 | Allocation: randomised. Participants: young people with a first episode of affective/non‐affective psychosis (unclear proportion), meeting remission criteria for at least 3 months. Intervention: maintenance treatment with antipsychotic drugs versus dose reduction strategy; no real discontinuation or placebo arm. |
| Wiedemann 2001 | Allocation: randomised. Participants: schizophrenia. Intervention: continuation of current antipsychotic versus gradual withdrawal. However, antipsychotic was given again when early warning signs appeared, i.e. intermittent treatment, a design that was excluded a prior by our protocol. |
| Wright 1964 | Allocation: not randomised. |
| Wunderink 2006 | Allocation: randomised. Participants: schizophrenia and related psychotic disorder. Intervention: continuation of current antipsychotic versus gradual withdrawal. However, antipsychotic was given again when early warning signs appeared, i.e. intermittent treatment, a design that was excluded by the protocol. Approximately 50% of participants were never withdrawn. |
| Zeller 1956 | Allocation: all participants were in hospital. 95 were allocated to placebo (not randomly). Then 81 participants were quote: "selected at random to match” the intervention group. We feel that this is not an appropriate method of randomisation. |
| Zou 2018 | Allocation: randomised. Participants: people with schizophrenia, experiencing an acute episode. |
| Zwanikken 1973 | Allocation: randomised. Participants: more than 50% had mental retardation, not schizophrenia. |
DSM‐III‐R: Diagnostic and Statistical Manual of Mental Disorders, Third Edition‐Revised LAI: LAI: long‐acting injectable