Study	Reason for exclusion
Albus 1985	Allocation: not randomised
Ananth 1977	Allocation: not randomised
Andersson 1988	Allocation: not randomised
Andia 1998	Allocation: randomised Participants: schizophrenia (DSM‐III‐R), N = 26, 14 with TD Intervention: haloperidol vs clozapine Outcomes: no data available for AIMS in clozapine group, the study also reported on plasma homovanillic acid levels, an outcome not relevant for this review We contacted the study authors but no information was received. This study is over 15years old and was excluded.
Asnis 1979	Allocation: not randomised
Auberger 1985	Allocation: not randomised
Barnes 2002	Allocation: random not mentioned in this short trial registration Participants: elderly patients, many started treatment with antipsychotics at start of study (no baseline TD)
Bateman 1979	Allocation: random Participants: psychiatric patients with TD Interventions: metoclopramide (10 mg, 20 mg or 40 mg) vs haloperidol (5 mg or 10 mg) Outcomes: no outcome data was provided for the first period before cross‐over. We were unable to find contact details for the study authors; study is over 35 years old and was excluded
Bitter 2000	Allocation: random Participants: people with schizophrenia, no TD symptoms at baseline Interventions: olanzapine vs clozapine
Blaha 1980	Allocation: not randomised
Borison 1987	Allocation: randomised Participants: schizophrenia (DSM‐III criteria); TD (Schooler and Kane criteria) Interventions: molindone versus haloperidol Outcomes: no usable efficacy data; only P values were reported. We were unable to find contact details for the study authors; study is over 25 years old and was excluded
Branchey 1981	Allocation: not randomised
Brecher 1999	Allocation: randomised Participants: people with dementia, not schizophrenia, not TD at baseline
Buchanan 1992	Allocation: not randomised
Burner 1989	Allocation: 'randomly assigned' Participants: people with schizophrenia, no TD symptoms at baseline Interventions: progabide vs placebo
Buruma 1982	Allocation: randomised, cross‐over Participants: "patients with tardive dyskinesia" ‐ no further details Interventions: tiapride vs placebo Outcomes: doppler ratings, none before cross‐over
Cai 1988	Allocation: randomisation not mentioned Participants: patients with antipsychotic‐induced TD Intervention: 1‐stepholidine (herbal product that has shown antipsychotic properties in animals) versus placebo Assessed and data extracted by Sai Zhao
Caine 1979	Allocation: "allocated by toss of a coin" Participants: Gilles de la Tourette's, Huntington's disease and drug‐induced atypical dyskinesia, no TD symptoms at baseline Interventions: clozapine vs placebo
Calne 1974	Allocation: not randomised
Campbell 1988	Allocation: not randomised
Carpenter 1980	Allocation: not randomised
Casey 1977	Allocation: not randomised
Casey 1979	Allocation: not randomised
Casey 1981	Allocation: not randomised
Casey 1983	Allocation: not randomised
Cassady 1992	Allocation: not randomised
Chouinard 1978	Allocation: random Participants: people with schizophrenia, no TD symptoms at baseline Interventions: fluphenazine ethanoate vs pipothiazine palmitate
Chouinard 1979	Allocation: random Participants: people with schizophrenia, and parkinsonism, no TD symptoms at baseline Interventions: ethopropazine vs benztropine
Chouinard 1989	Allocation: random Participants: people with schizophrenia, no TD symptoms at baseline Interventions: haloperidol decanoate vs fluphenazine decanoate
Chouinard 1994	Allocation: random Participants: people with schizophrenia, no TD symptoms at baseline Interventions: clozapine vs risperidone
Claveria 1975	Allocation: not randomised
Cookson 1991	Allocation: random Participants: people with schizophrenia, no TD symptoms at baseline Interventions: haloperidol decanoate vs fluphenazine decanoate
Cortese 2008	Allocation: random Participants: people with schizophrenia, no TD symptoms at baseline Interventions: quetiapine vs continuation of usual antipsychotic
Cowen 1997	Allocation: not randomised
Crane 1968	Allocation: not randomised, review article
Crane 1969	Allocation: not randomised
Crane 1970	Allocation: random Participants: people with schizophrenia, only 2%‐3% with TD at baseline Interventions: trifluoperazine high dose vs trifluoperazine low dose vs placebo
Curran 1973	Allocation: not randomised
Curson 1985	Allocation: random Participants: people with schizophrenia, no TD symptoms at baseline Interventions: fluphenazine decanoate vs placebo
Davidson 2000	Allocation: not randomised
de Jesus Mari 2004	Allocation: randomised Participants: diagnosis of schizophrenia or related disorders (DSM‐IV criteria). < 50% of participants had TD at baseline Interventions: olanzapine vs "conventional antipsychotic drugs" Outcomes: author contacted for data regarding people with TD ‐ data no longer available
Delwaide 1979	Allocation: randomised Participants: hospitalised patients with TD on a psychogeriatric ward Intervention: thioproperazine vs tiapride vs placebo Outcome: all data unusable, unable to extract data from first arm of cross‐over The study is over 35 years old and we were unable to identify contact details for the author
Diamond 1986	Allocation: not randomised
Dixon 1993	Allocation: not randomised
Fahn 1983	Allocation: not randomised
Fahn 1985	Allocation: not randomised
Freeman 1980	Allocation: not randomised
Gardos 1984	Allocation: not randomised
Gerlach 1975	Allocation: random Participants: schizophrenia, no established, stable TD diagnosis at baseline Interventions: clozapine vs haloperidol
Gerlach 1978	Allocation: the randomisation was just in one arm of the study “Haloperidol + biperiden for 4 weeks (phase 2 and phase 3 in randomized sequence)”. All other arms thioridazine for 3 months, haloperidol for 4 weeks; thioridazine for 4 weeks, clozapine for 4 weeks were not Participants: elderly people with psychiatric history and neuroleptic‐induced TD Interventions: biperiden vs no treatment as an adjunct to haloperidol
Gerlach 1984a	Allocation: not randomised, cohort study
Gerlach 1984b	Allocation: not randomised
Gibson 1980	Allocation: not randomised
Glazer 1984	Allocation: not randomised
Glazer 1989	Allocation: not randomised
Goldberg 1981	Allocation: randomised Participants: people with schizophrenia (no history of TD) Interventions: withdrawal of fluphenazine decanoate vs continuation
Greil 1984	Allocation: "randomly assigned" Participants: people with schizophrenia Interventions: biperiden vs placebo
Haggstrom 1980	Allocation: not randomised
Heresco‐Levy 1993	Allocation: not randomised
Hershon 1972	Allocation: randomised Participants: people with schizophrenia (no history of TD) Interventions: trifluoperazine withdrawal vs trifluoperazine continuation
Herz 1991	Allocation: randomised Participants: people with schizophrenia Interventions: neuroleptic reduction (intermittent treatment) vs maintenance neuroleptic Outcomes: no usable data Dr Herz kindly replied to our request for more information. Unfortunately, individual baseline and endpoint AIMS score are no longer available
Hogarty 1976	Allocation: not randomised
Hogarty 1988	Allocation: quasi‐randomised
Inada 2003	Allocation: not randomised
Inderbitzin 1994	Allocation: Not randomised ("by alternate allocation")
Jean‐Noel 1999	Allocation: random Participants: people with schizophrenia, no TD symptoms at baseline Interventions: clozapine vs olanzapine
Jeste 1977	Allocation: not randomised Participants: chronic schizophrenia with TD, N = 2 Interventions: chlorpromazine schedule A vs chlorpromazine schedule B. Treatments in the 2 groups were the same except for timing of the doses (frequency and withdrawal)
Jeste 1979	Allocation: not randomised
Johnson 1983	Allocation: not randomised
Johnson 1987	Allocation: randomised Participants: people with schizophrenia Interventions: neuroleptic dose reduction vs maintenance dose (both arms used flupenthixol decanoate) Outcomes: no usable data Dr Johnson kindly replied to our letter. No further data available from the first author
Jolley 1990	Allocation: randomised Participants: people with schizophrenia (no history of TD) Interventions: brief intermittent antipsychotic treatment vs fluphenazine decanoate
Jus 1979	Allocation: not randomised
Kalachnik 1984	Allocation: not randomised, case‐control study Dr Kalachnick kindly provided additional information. After randomisation clinicians reviewed group allocations and re‐assigned selected individuals on clinical grounds
Kane 1993	Allocation: not randomised, 2 case series
Kinon 2004	Allocation: randomised Participants: schizophrenia and TD (Schooler and Kane criteria) Interventions: olanzapine (5 mg‐20 mg/d) with 1 set of intermittent dose‐reduction periods versus olanzapine (5 mg‐20 mg/d) with a different set of intermittent dose reduction periods
Kirch 1983	Allocation: not randomised
Kopala 2004	Allocation: random Participants: people with schizophrenia, no TD symptoms at baseline Interventions: haloperidol vs risperidone
Lal 1974	Allocation: randomised, cross‐over. Participants: people with schizophrenia. Interventions: thiopropazine vs trifluoperazine vs placebo. Outcomes: no usable data. Dr Lal kindly replied to inquiry. Unable to extract data from the first segment. Jadad score = 4/5
Leblanc 1994	Allocation: not randomised, cohort study
Leblhuber 1987	Allocation: not randomised
Levine 1980	Allocation: randomised Participants: people with schizophrenia (no history of TD) Interventions: fluphenazine withdrawal vs continuation
Lieberman 1988	Allocation: randomised Participants: TD according to the criteria of Schooler and Kane, schizophrenia, schizoaffective disorder, major affective disorder and attention deficit disorder Intervention: physostigmine vs bromocriptine vs benztropine vs haloperidol for 1 day, then crossed over. Outcomes: no outcome data provided for the first period before cross‐over. We contacted the study author but no information received. Study is over 25 years old years old and was excluded
Lieberman 1989	Allocation: not randomised, cohort study
Lin 2006	Allocation: not randomised: naturalistic observational study
Littrell 1993	Allocation: not randomised
MacKay 1980	Allocation: "patients were divided into pairs" Participants: people with schizophrenia Intervention: lithium vs placebo
Marder 1987	Allocation: randomised Participants: people with schizophrenia (no history of TD) Interventions: low vs conventional‐dose maintenance therapy with fluphenazine decanoate
McCreadie 1980	Allocation: random Participants: people with schizophrenia, no TD symptoms at baseline Interventions: intermittent pimozide vs fluphenazine decanoate
Meco 1989	Allocation: not randomised
Miller 1994	Allocation: not randomised
NDSG 1986	Allocation: randomised cross‐over Participants: psychiatric inpatients with TD Intervention: chlorprothixene vs haloperidol vs perphenazine vs haloperidol + biperiden Outcomes: no outcome data provided for the first period before cross‐over We contacted the study author but no reply. Study is 30 years old and was excluded.
Newcomer 1992	Allocation: randomised Participants: people with schizophrenia (no history of TD) Interventions: haloperidol dose reduction vs maintained dose
Newton 1989	Allocation: randomised Participants: people with schizophrenia (no history of TD) Interventions: haloperidol with 'drug holiday' versus haloperidol
Odejide 1982	Allocation: randomisation not mentioned Participants: people with schizophrenia (no history of TD) Interventions: fluphenazine decanoate vs vitamin B complex
Pai 2001	Allocation: not randomised Participants: people with schizophrenia and TD Interventions: risperidone vs placebo
Paulson 1975	Allocation: not randomised Dr Paulsen kindly provided additional information about this double‐blind study
Peacock 1996	Allocation: not randomised
Peluso 2012	Allocation: random Participants: people with schizophrenia, no TD symptoms at baseline Interventions: FGA vs second generation antipsychotic
Perry 1985	Allocation: randomised Participants: children with autism without history of TD Dr Campbell kindly provided all published and in‐press data. The authors found no difference in TD between the intermittent and continuous treatment groups but further details required for this review were not available
Pyke 1981	Allocation: not randomised
Quinn 1984	Allocation: randomised, double‐blind, cross‐over study Participants: people with schizophrenia Intervention: sulpiride (Dogmatil) 300 mg‐ 1200 mg/d Outcomes: no usable data. Drs Marsden and Quinn kindly replied to our letter, but no data suitable for this review could be provided
Quitkin 1977	Allocation: not randomised
Rapoport 1997	Allocation: not described
Ringwald 1978	Allocation: not random
Rosenheck 2003	Allocation: random Participants: people with schizophrenia, no TD symptoms at baseline Interventions: haloperidol vs olanzapine
Roxburgh 1970	Allocation: not randomised
Schultz 1995	Allocation: not randomised
Schwartz 1990	Allocation: randomised Participants: psychiatric inpatients with TD Interventions: sulpiride vs placebo Outcomes: no outcome data provided for the first period before cross‐over. We were unable to find contact details for the study authors; this study is over 25 years old and was excluded
Seeman 1981	Allocation: not randomised
Simpson 1978	Allocation: not randomised, cohort study
Singer 1971	Allocation: randomised Participants: psychiatric inpatients with persistent TD Interventions: thiopropazate vs placebo Outcomes: no outcome data provided for the first period before cross‐over. We were unable to find contact details for the study authors; this study is 45 years old and was excluded
Singh 1990	Allocation: randomised Participants: people with schizophrenia (majority did not have TD) Intervention: abrupt antipsychotic withdrawal versus continuation of antipsychotic medication
Small 1987	Allocation: not randomised, cohort study
Smith 1979	Allocation: not randomised, cohort study
Soni 1984	Allocation: not randomised
Speller 1997	Allocation: randomised Participants: schizophrenia (DSM‐III‐R), majority with TD Intervention: amisulpride versus haloperidol Outcomes: schizophrenia symptom changes, especially negative symptoms, adverse events, and TD as adverse event. We excluded this reference because, although the majority had TD at baseline and the intervention drugs qualified, the drugs were not examined as a treatment for TD (as our inclusion criteria demand), but for negative symptoms of schizophrenia
Spivak 1997	Allocation: not randomised, cohort study
Spohn 1988	Allocation: randomised Participants: people with schizophrenia Interventions: abrupt neuroleptic cessation versus neuroleptic maintenance Outcomes: no usable data Dr Spohn kindly replied to our request for further information. Data on baseline and endpoint TD not available
Spohn 1993	Allocation: randomised Participants: people with schizophrenia Interventions: abrupt neuroleptic withdrawal versus maintenance Outcomes: no usable data Dr Spohn kindly replied to our letter, but no further data were available
Suh 2004	Allocation: randomised Participants: dementia and not TD
Thapa 1994	Allocation: randomised Participants: nursing home staff Interventions: education about neuroleptic prescribing vs no specific additional education
Tollefson 1997	Allocation: random Participants: no TD at baseline, investigates incidence of TD with long‐term treatment with olanzapine vs haloperidol
Tran 1997	Allocation: random Participants: people with schizophrenia, no TD symptoms at baseline Interventions: olanzapine versus haloperidol
Turek 1972	Allocation: not randomised ‐ allocated to treatment group in a "nonsystematic" fashion, but then participants were re‐allocated to alternate groups based on clinical judgement
Williamson 1995	Allocation: random Participants: schizophrenia, not TD Interventions: olanzapine 1 mg vs olanzapine 10 mg versus placebo
Wirshing 1999	Allocation: random Participants: people with treatment‐resistant schizophrenia, no TD symptoms at baseline Interventions: haloperidol vs risperidone
Wistedt 1983	Allocation: randomised Participants: people with schizophrenia (no history of TD) Interventions: fluphenazine/flupenthixol decanoate continuation vs withdrawal
Wolf 1991	Allocation: not randomised, cohort study
Wright 1998	Allocation: not randomised
Zander 1981	Allocation: not randomised
Zarebinski 1990	Allocation: not randomised, cohort study
Zeng 1994	Allocation: randomised Participants: patients with antipsychotic‐induced TD Intervention: flunarizine (calcium channel antagonist) vs placebo Assessed and data extracted by Sai Zhao

FGA: first‐generation antipsychotic
 IV = intravenous
 TD: tardive dyskinesia