Intervention Review

You have free access to this content

Haloperidol versus placebo for schizophrenia

  1. Clive E Adams1,*,
  2. Hanna Bergman2,
  3. Claire B Irving1,
  4. Stephen Lawrie3

Editorial Group: Cochrane Schizophrenia Group

Published Online: 15 NOV 2013

Assessed as up-to-date: 15 MAY 2012

DOI: 10.1002/14651858.CD003082.pub3


How to Cite

Adams CE, Bergman H, Irving CB, Lawrie S. Haloperidol versus placebo for schizophrenia. Cochrane Database of Systematic Reviews 2013, Issue 11. Art. No.: CD003082. DOI: 10.1002/14651858.CD003082.pub3.

Author Information

  1. 1

    The University of Nottingham, Cochrane Schizophrenia Group, Nottingham, UK

  2. 2

    Enhance Reviews Ltd, Wantage, UK

  3. 3

    University of Edinburgh, Department of Psychiatry, Edinburgh, Scotland, UK

*Clive E Adams, Cochrane Schizophrenia Group, The University of Nottingham, Institute of Mental Health, University of Nottingham Innovation Park, Triumph Road,, Nottingham, NG7 2TU, UK. clive.adams@nottingham.ac.uk.

Publication History

  1. Publication Status: New search for studies and content updated (no change to conclusions)
  2. Published Online: 15 NOV 2013

SEARCH

 
Characteristics of included studies [ordered by study ID]
Arvanitis 1997

MethodsAllocation: random assignment.
Blindness: double-blind.
Duration: 6 weeks (preceded by a 7-day single blind placebo washout).
Location: multicentre.
Design: parallel.
Setting: inpatients.
Country: USA and Canada.

Consent: written.


ParticipantsDiagnosis: (DSM-III-R) schizophrenia.
N = 361.
History: sub/chronic hospitalised, currently experiencing acute exacerbation of psychotic symptoms.
Sex: 76% M, 24% F.
Age: 18 - 64 years (mean ˜ 37 years).
Exclusions: BPRS score < 27, CGI score < 4, history of seizures, other significant medical condition, participation in other drug trial within 30 days, use of depot antipsychotics within 1 dosing interval, pregnancy, placebo responders, non completion of dose escalation.


Interventions1. Haloperidol: fixed dose (FD) 12 mg/day, increased day 1-14. N = 52
2. Placebo. N = 51.
3. Quetiapine: (FD) 75 mg/day, increased day 1-14. N = 53.
4. Quetiapine: (FD) 150 mg/day, increased day 1-14. N = 48.
5. Quetiapine: (FD) 300 mg/day, increased day 1-14. N = 52.
6. Quetiapine: (FD) 600 mg/day, increased day 1-14. N = 51.
7. Quetiapine: (FD) 750 mg/day, increased day 1-14. N = 54.
Chloral hydrate, lorazepam, benztropine mesylate as required.


OutcomesLeaving the study early.

Unable to use -
Global effect: improved/not improved, CGI (> 50% loss).
Dose response (> 50% loss).
Time to response (> 50% loss).
Mental state: BPRS, SANS (> 50% loss).
Adverse events: AIMS, SAS, other various observed effects (> 50% loss).


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"Randomised" no further details reported.

Allocation concealment (selection bias)Unclear riskNo information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk"Double-blind" no further details reported.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk"Double-blind"; "[Blood] samples were shipped to the sponsor and analysed". No further details reported.

Incomplete outcome data (attrition bias)
All outcomes
High risk"Of all patients evaluated, 149 (41%) completed 6 weeks of treatment. Lack of efficacy was the primary reason for withdrawal and was seen most often in the placebo group". No further details reported.

Selective reporting (reporting bias)Low riskAll expected outcomes reported.

Other biasHigh riskSupported by a grant from Zeneca Pharmaceuticals.

Beasley 1996

MethodsAllocation: random assignment.
Blindness: double-blind.
Duration: 6 weeks (preceded by a 4-7 day single blind placebo washout).
Location: multicentre.
Design: parallel.
Setting: inpatients and outpatients.
Country: USA.

Consent: written.


ParticipantsDiagnosis: (DSM-III-R) schizophrenia.
N = 335.
History: currently suffering from acute exacerbation of symptoms.
Sex: 88% M, 12% F.
Age: 18 - 65 years.
Exclusions: other serious physical or neurological disorder/condition, substance abuse within 3 months of study, abnormal lab results, placebo responders.


Interventions1. Haloperidol: dose 10, 15 or 20 mg/day; initial dose 15 mg/day, adjusted accordingly thereafter. N = 69.
2. Placebo. N = 68.
3. Olanzapine: dose 2.5, 5 or 7.5 mg/day; initial dose 5 mg/day, adjusted accordingly thereafter. N = 65.
4. Olanzapine: dose 7.5, 10 or 12.5 mg/day; initial dose 10 mg/day, adjusted accordingly thereafter. N = 64.
5. Olanzapine: dose 12.5, 15 or 17.5 mg/day; initial dose 15 mg/day, adjusted accordingly thereafter. N = 69.
Lorazepam, benztropine mesylate as required.


OutcomesLeaving the study early.

Unable to use -
Global effect: CGI, PGI (> 50% loss).
Hospitalisation: admission/discharge, days in hospital (> 50% loss).
Dose response (> 50% loss).
Mental state: BPRS, SANS (> 50% loss).
Adverse events: SAS, AIMS, BAS, other observed effects (> 50% loss).
Compliance (> 50% loss).


NotesData only taken from the initial 'acute phase' trial.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"Randomised" no further details reported.

Allocation concealment (selection bias)Unclear riskNo information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk"Double-blind" no further details reported.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo information reported.

Incomplete outcome data (attrition bias)
All outcomes
High risk Losses to follow-up ranged from 51% - 68% in the five treatment arms.

Selective reporting (reporting bias)High riskPatient Global Impression (PGI) assessed but not reported.

Other biasHigh riskSource of funding from industry "From the Psychopharmacology Division, Lilly Research Laboratories, Eli Lilly and Company"

Bechelli 1983

MethodsAllocation: random assignment.
Blindness: double-blind.
Duration: 21 days (preceded by 3 days stabilisation with chlorpromazine and haloperidol followed by 2 day washout).
Location: hospital.
Design: parallel.
Setting: inpatients.
Country: Brazil.

Consent: not stated.


ParticipantsDiagnosis: (ICD-9) schizophrenia.
N = 90.
History: recently admitted to hospital, currently acute.
Sex: male.
Age: mean ˜ 29 years.
Exclusions: substance abuse, other clinically significant medical or neurological pathologies.


Interventions1. Haloperidol: dose 5 - 20 mg/day. N = 30.
2. Placebo: N = 31.
3. Pipotiazine: dose 10 - 40 mg/day. N = 29.
Biperiden and trihexyphenidyl as required on day 1 - 3 only.


OutcomesAdverse event: various observed effects.
Global effect: improved/not improved.
Leaving the study early.
Mental state: BPRS.


NotesIf, after entering the trial, participants showed improvement they could be discharged from hospital. If, however, they were readmitted due to relapse they could then re-enter the study.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"The patients were assigned to 3 groups of 30 each in a random and probabilistic manner, after stratification", no further details reported.

Allocation concealment (selection bias)Unclear riskNo information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk"Double-blind" no further details reported.

Blinding of outcome assessment (detection bias)
All outcomes
Low risk"The assessments were always performed by the same investigator in a double-blind trial. Only at the end of the study after the data were analysed were the patient group assignments identified".

Incomplete outcome data (attrition bias)
All outcomes
Low risk"Five patients ran away between the 6th and 27th day of the study. Three belonged to the pipotiazine group, 1 belonged to the haloperidol group, and 1 to the placebo group". "These patients were excluded from the analysis".

Selective reporting (reporting bias)High riskNot all expected outcomes reported.

Other biasUnclear riskSource of funding not reported.

Borison 1989

MethodsAllocation: random assignment.
Blindness: double-blind.
Duration: 6 weeks (preceded by a 7 day single blind placebo wash out).
Location: not stated.
Design: parallel.
Setting: not reported.
Country: not reported.

Consent: written.


ParticipantsDiagnosis: (DSM-III) schizophrenia.
N = 32.
History: not stated.
Sex: not stated.
Age: 18 - 60 years.
Exclusions: unstable physical health.


Interventions1. Haloperidol: dose 15 - 75 mg/day. N = 8.
2. Placebo. N = 8.
3. Tiospirone: dose 45 - 225 mg/day. N = 8.
4. Thioridazine: dose 150 - 750 mg/day. N = 8.
Chloral hydrate as required.


OutcomesAdverse events: various observed effects, use of antiparkinson medication.
Leaving the study early.

Unable to use -
Mental state: BPRS (no SD).


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"...subjects were assigned on a randomized blind schedule to treatment" no further details reported.

Allocation concealment (selection bias)Unclear risk"...subjects were assigned on a randomized blind schedule to treatment" no further details reported.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk"Double- blind" no further details reported.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo information reported.

Incomplete outcome data (attrition bias)
All outcomes
Low riskLosses to follow-up or missing data were balanced across intervention groups, with similar reasons for missing data. "...three placebo-treated patients were terminated prior to study completion due to lack of efficacy. Two patients receiving haloperidol terminated early due to positive response and the desire to leave the hospital, and one patient in the haloperidol treatment group was terminated for administrative reasons. The only patient who left the study prematurely due to an apparent adverse reaction was one receiving placebo, who developed chest pain and electrocardiographic changes"

Selective reporting (reporting bias)Unclear riskStudy does not state in methods which outcomes will be measured, and/or no protocol available.

Other biasHigh riskResearchers currently imprisoned for research fraud.

Borison 1992a

MethodsAllocation: random assignment.
Blindness: double-blind.
Duration: 7 weeks (preceded by a 7-day single blind placebo washout).
Location: multicentre, part of larger trial.
Design: parallel.
Setting: inpatients.
Country: not reported.

Consent: written.


ParticipantsDiagnosis: (DSM-III-R) schizophrenia.
N = 36.
History: chronic.
Sex: 97% M, 3% F.
Age: ˜ 31-52 years.
Exclusions: substance abuse, other clinically significant medical or neurological pathologies, women of child bearing capacity.


Interventions1. Haloperidol: dose 4-20 mg/day, dose adjusted as required days 1-18. N = 12.
2. Placebo. N = 12.
3. Risperidone: dose 2-10 mg/day, dose adjusted as required days 1-18. N = 12.
Lorazepam, sodium amytal, chloral hydrate, benztropine or trihexyphenidyl as required.


OutcomesAdverse events: various observed effects.
Leaving study early.
Mental state: no clinical improvement (< 20% reduction in BPRS score).

Unable to use -
Adverse events: AIMS (no data), ESRS (no SD).
Global effect: CGI (no SD).
Mental state: BPRS, SANS (no SD).


NotesParticipants already part of larger multicentre trial.
May be part of North American Trial.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"Randomised" no further details reported.

Allocation concealment (selection bias)Unclear riskNo information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk"Double-blind" no further details reported.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk"Double-blind" no further details reported.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo details reported.

Selective reporting (reporting bias)High riskSeveral outcomes not fully reported (BPRS, CGI, SANS, or ESRS).

Other biasHigh riskSource of funding not reported.

Richard Borison, MD, former psychiatry chief at the Augusta Veterans Affairs medical center and Medical College of Georgia, was sentenced to 15 years in prison for a $10 million clinical trial fraud.

Chouinard 1993

MethodsAllocation: random assignment.
Blindness: double-blind.
Duration: 8 weeks (preceded by a 7-day single blind placebo washout).
Location: multicentre.
Design: parallel.
Setting: inpatients.
Country: Canada.

Consent: written.


ParticipantsDiagnosis: (DSM-III-R) schizophrenia.
N = 135.
History: chronic, hospitalised.
Sex: 71% M, 29% F.
Age: mean ˜ 37 years.
Exclusions: other clinically significant neurological or psychical disorder, substance abuse, pregnancy, placebo responders.


Interventions1. Haloperidol: dose 20 mg/day, initial dose 2 mg/day increased in fixed increments day 2-7. N=21.
2. Placebo. N = 22.
3. Risperidone: dose 2 mg/day
4. Risperidone: dose 6 mg/day, initial dose 2 mg/day increased in fixed increments day 2-4.
5. Risperidone: dose 10 mg/day, initial dose 2 mg/day, increased in fixed increments day 2-5.
6. Risperidone: dose 16 mg/day, initial dose 2 mg/day, increased in fixed increments day 2-7.
Chloral hydrate, benzodiazepine, biperiden or procyclidine as required.


OutcomesLeaving the study early.

Unable to use -
Global effect: CGI (> 50% loss).
Level of medication required (> 50% loss).
Mental state: BPRS, GPS, PANSS (> 50% loss).
Adverse events: ESRS, various observed effects (UKU), use of antiparkinson medication (> 50% loss).
Cost of treatment (> 50% loss).


NotesPart of North American Trial.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"Randomly assigned" no further details reported.

Allocation concealment (selection bias)Unclear riskNo information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low risk"Double-blind", "Study medication was administered under double-blind conditions as identical tablets of risperidone, haloperidol and placebo".

Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk"Double-blind", "Assessment of symptoms was based on clinical interviews conducted by a psychiatrist". No further details reported.

Incomplete outcome data (attrition bias)
All outcomes
High riskA total of 65 participants (48.1%) left the study early: 16 (72.7%) from the placebo group and 13 (61.9%) from the haloperidol group.

"Statistical analyses of efficacy and safety parameters were conducted according to the intent-to-treat analysis principle".

Selective reporting (reporting bias)Low riskAll expected outcomes reported.

Other biasUnclear riskSource of funding not reported.

Durost 1964

MethodsAllocation: random assignment.
Duration: 10 days - 3 months (mean 3 weeks).
Location: hospital.
Design: parallel.
Setting: inpatients.
Country: not reported.

Consent: unknown.


ParticipantsDiagnosis: schizophrenia (40%), neurosis (60%).*
N = 84 (schizophrenia 34, neurosis 50).
History: unknown.
Sex: 60% M, 40% F.
Age: mean ˜ 39 years.
Exclusions: unknown.


Interventions1. Haloperidol: dose 2-25 mg/day, mean 6 mg/day. N = 19.
2. Placebo. N = 15.


OutcomesGlobal effect: improved/not improved.
Leaving the study early.

Unable to use -
Adverse events: various observed effects (no data).


Notes* use only data for those suffering from schizophrenia


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"...all drugs were given at random" no further details reported.

Allocation concealment (selection bias)Unclear riskNo details reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low risk"...'pharmacotherapeutically blind unit,' that is, in a hospital service where all drugs were given at random and used without the service team (made up of one intern, two assistant residents, one resident, one psychologist, two to four nurses, one social worker and one occupational therapist) knowing what drugs were being investigated."

Blinding of outcome assessment (detection bias)
All outcomes
Low risk"The assessment of the patients (and the drugs) was a result of the pooling of the opinions of the different members of the team."

Incomplete outcome data (attrition bias)
All outcomes
Low riskThere were no drop-outs from the study.

Selective reporting (reporting bias)High riskDetails on side effects were not fully reported although "Side effects were relatively numerous and disturbing to the patients."

Other biasUnclear risk"Haloperidol was generously supplied by G. D. Searle & Co." 

Garcia 2009

MethodsAllocation: randomised.
Blinding: double-blind.
Duration: 6 weeks.
Design: parallel.
Setting: inpatient.
Country: USA, Bulgaria, the Czech Republic and Russia.

Consent: written.


ParticipantsDiagnosis: schizophrenia DSM-4.
N = 307.
Age: 18-65 (Mean 38.1, SD 11.1).
Sex: M 183, F 124.
History: acute exacerbation of illness, hospitalised for < 2 weeks at screening due to the exacerbation. PANSS score of at least 70.
Exclusions: patients improving > 20% in the PANSS total score or 1 point in the CGI-S scale at baseline compared with screening; resistant to antipsychotic treatment; DSM-IV-TR-defined substance abuse/dependency within the preceding 3 months (or a positive urine drug test); treated with depot antipsychotics, unless the last injection was administered within greater than one treatment cycle before study entry;clinically significant or currently relevant illness or those judged by the investigator to be at serious suicidal risk.


Interventions1. Blonanserin: dose 2.5 mg/day. N = 61.

2. Blonanserin: dose 5 mg/day. N = 58.

3. Blonanserin: dose 10 mg/day. N = 64.

4. Haloperidol: dose 10 mg/day. N = 60.

5. Placebo. N = 64.


OutcomesLeaving the study early

Global state: No overall improvement (< 20% reduction in PANSS-total score)

Mental state: PANSS-total score, mean change from baseline at 6 weeks

Mental state: PANSS-positive score, mean change from baseline at 6 weeks

Mental state: PANSS-negative score, mean change from baseline at 6 weeks

Global state: CGI-S, mean change from baseline at 6 weeks

Adverse effects: Extrapyramidal symptoms (akathisia, dyskinesia, dystonia, parkinsonism, rigidity, tremor)

Adverse effects: Cardiovascular (bradycardia)

Adverse effects: Other (insomnia, drooling, headache, weight loss, agitation, anxiety, sleepiness)

Unusable data (no measurement of variance reported) -

Adverse effects: Extrapyramidal symptoms: SAS, BAS and AIMS scales


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low risk"Randomization of medication was performed using a computer-generated schedule".

Allocation concealment (selection bias)Unclear risk No information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low risk“Double-blind”, "...blinding was ensured by over-encapsulating all capsules to ensure the same appearance".

Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk“Double-blind”, no further details reported.

Incomplete outcome data (attrition bias)
All outcomes
Low riskLosses to follow-up or missing data were balanced across intervention groups, with similar reasons for missing data. Missing data have been imputed using appropriate methods (intention-to-treat analysis).

Selective reporting (reporting bias)High risk Not all of the study's pre-specified primary outcomes have been reported (BPRS, UKU assessments).

Other biasHigh risk"Laboratorios Almirall SA provided financial support for performing the study and Dainippon Sumitomo Pharma Co., Ltd funded the preparation of this manuscript. Drs Garcia, Robert and Peris are employees of Laboratorios Aimirall SA. H. Nakamura, Dr Sato and Y. Terazawa are employees of Dainippon Sumitomo Pharma Co., Ltd. Drs Garcia and Peris have received stock options from Laboratorios Almirall SA." 

Garry 1962

MethodsAllocation: random assignment.
Blindness: double-blind.
Duration: 12 weeks (preceded by a 2-week medication free period).
Location: not stated.
Design: parallel.
Setting: inpatients.
Country: not stated.

Consent: not stated.


ParticipantsDiagnosis: schizophrenia.
N = 52.
History: chronic, hospitalised.
Sex: 62% M, 38% F.
Age: mean ˜ 46 years.
Exclusions: not stated.


Interventions1. Haloperidol: dose 0.75 - 4.5 mg/day, increased day 1 - 42. N = 26.
2. Placebo. N=26.


OutcomesAdverse events: various observed effects*
Global effect: improved/not improved.
Leaving the study early.


Notes*Adverse effects reported for haloperidol group only, reviewers assumed that placebo group had no adverse effects.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskParticipants "were divided into two groups on the basis of a list of random numbers supplied by the drug company"

Allocation concealment (selection bias)Low riskCentral allocation "Our pharmacist allotted the patients from an alphabetical list supplied  by us"

Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk"Double-blind" no further details reported.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk"Double-blind" no further details reported.

Incomplete outcome data (attrition bias)
All outcomes
Low risk"Two patients were withdrawn from the trial, one had a recurrence of a skin rash (he was eventually found to be on the placebo) and the other was found to have early pulmonary tuberculosis on routine chest X-ray. We were finally left with 50 patients (25 on drug and 25 on controls) who completed the trial".

Selective reporting (reporting bias)Low riskAll expected outcomes reported.

Other biasUnclear riskMessrs. G. D. Searle and Co. Ltd. supplied the haloperidol and control tablets.

Howard 1974

MethodsAllocation: random assignment.
Blindness: double-blind.
Duration: max 12 weeks (preceded by a 14 day placebo washout).
Location: hospital.
Design: parallel and cross-over.
Setting: inpatients and outpatients.
Country: USA.

Consent: not stated.


ParticipantsDiagnosis: schizophrenia (80%).
N = 49.
History: treatment resistant, hospitalised.
Sex: female.
Age: 25 - 65 years.
Exclusions: other serious physical or neurological disorder, pregnancy, severe hyposensitivity to haloperidol or thiothixene, placebo responders.


Interventions1. Haloperidol: dose < 200 mg/day. N = 17.
2. Placebo. N = 16.
3. Thiothixene: dose < 200 mg/day. N = 16.


OutcomesGlobal effect: improved/not improved.
Hospital discharge.
Leaving the study early.
Adverse events: various observed effects.

Unable to use -
Behaviour: NOSIE (no SD).
Mental state: BPRS (no mean, no SD), MSC (data given only for those remaining in hospital).


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"the patients were randomly assigned" no further details reported.

Allocation concealment (selection bias)Unclear riskNo details reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low risk"The study medications were prepared in identical appearing capsules".

Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk No details reported.

Incomplete outcome data (attrition bias)
All outcomes
Low risk Losses to follow-up were balanced across intervention groups, with similar reasons for missing data.

Selective reporting (reporting bias)High riskAll outcomes were not fully reported (no SD s were reported for BPRS and NOSIE).

Other biasUnclear riskInsufficient information to assess whether an important risk of bias exists (does not report source of funding).

Jann 1997

MethodsAllocation: random assignment.
Blindness: unsure.
Duration: 6 weeks.
Location: not stated.
Design: parallel.
Setting: inpatients.
Country: not reported.

Consent: written.


ParticipantsDiagnosis: (DSM-III-R) schizophrenia.
N = 36.
History: not stated.
Sex: not stated.
Age: ˜ 25 - 43 years.
Exclusions: not stated.


Interventions1. Haloperidol: dose up to 75 mg/day, dose individually adjusted weekly. N = 18.
2. Placebo. N = 18.
Lorazepam or chloral hydrate as required.


OutcomesAdverse events: various observed effects.
Leaving the study early.
Mental state: BPRS.

Unable to use -
Adverse events: AIMS (no data).


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"Randomly assigned" no further details reported.

Allocation concealment (selection bias)Unclear riskNo information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low risk"Identical looking capsules" were used.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo information provided.

Incomplete outcome data (attrition bias)
All outcomes
High risk"Only 8 patients in the placebo group completed the 6-week study". "For the haloperidol group, 12 patients completed the study".

Selective reporting (reporting bias)High riskAll pre-stated outcomes were not reported (no data reported for the AIMS scale).

Other biasUnclear riskSource of funding not reported.

Kane 2002

MethodsAllocation: random.
Blindness: double-blind.
Duration: 4 weeks.
Location: hospital, multicentre.
Design: parallel.
Setting: inpatients.
Country: USA.

Consent: given.


ParticipantsDiagnosis: (DSM-IV) schizophrenia or schizoaffective disorder.
N = 414*.
History: acute relapse, hospitalised.
Age: 18 - 65 years.
Sex: 70% M, 30% F.
Exclusions: other psychiatric disorder, history of violence or self-harm.


Interventions1. Aripiprazole: dose 15 mg/day. N = 102.
2. Aripiprazole: dose 30 mg/day. N = 102.
3. Haloperidol: dose 10 mg/day. N = 104.
4. Placebo. N = 106.
lozepam for anxiety or insomnia


OutcomesLeaving the study early.
Adverse events: various observed effects.

Unable to use:
Global effect: CGI (no SD).
Mental state: BPRS, PANSS (no SD).


NotesData taken from haloperidol and placebo groups only.
Data provided for some outcomes have only 103 people in haloperidol group and 104 in placebo group.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"Randomly assigned" no further details reported.

Allocation concealment (selection bias)Unclear riskNo information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk"Double-blind" no further information reported.

Blinding of outcome assessment (detection bias)
All outcomes
Low risk"The same rater conducted the assessment throughout the study and was blinded to the patient's treatment".

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNumber of losses and reasons for leaving the study similar across groups. "Analysis of efficacy parameters was performed on an intention-to-treat basis using data obtained from each patient's last visit (i.e. last observation carried forward analysis at week 4".

"Of the 414 randomised patients, 248 competed the 4-week study period".

Selective reporting (reporting bias)High riskAll pre-stated outcomes were not fully reported (SDs were not reported for PANSS, BPRS, CGI, AIMS, SAS and BAS scales, weight and serum prolactin levels).

Other biasHigh riskSponsored by Otsuka Pharmaceutical Co. Ltd (Tokyo, Japan) and Bristol-Meyers Squibb Company (Princeton, NJ).

Kane 2010

MethodsAllocation: randomised.
Blinding: double-blind.
Duration: 6 weeks.
Design: parallel.
Setting: inpatients and outpatients*.
Country: 43 centres (United States, 17 sites; Russia, 11 sites; India, 7 sites; Romania, 7 sites; Canada, 1 site).

Consent: written.


ParticipantsDiagnosis: schizophrenia, DSM-IV.
N = 458.
Age: 37 to 40 years.
Sex: M 270, F 188.
History: acute exacerbation of psychotic symptoms.
Exclusions: a clinically significant medical condition or abnormal laboratory or physical examination findings; diagnosis of residual-type schizophrenia, schizoaffective disorder, or coexisting psychiatric disorder coded on Axis I; current or past substance abuse; 20% or higher decrease in PANSS total score from screening to baseline; known allergy or sensitivity to haloperidol; imminent risk of self-harm or harm to others; previous participation in an asenapine trial.


Interventions1. Asenapine: dose 5 mg/day. N = 114.

2. Asenapine: dose 10 mg/day. N = 106.

3. Placebo. N = 123.

4. Haloperidol: dose 4 mg/day. N = 115.


OutcomesLeaving the study early

Global state: no overall improvement (no CGI-I score of 1 [very much improved] or 2 [much improved]) (also measured as < 30% reduction in PANSS total score, not used)

Mental state: PANSS subscale scores (positive, negative), mean change from baseline at 6 weeks

Global state: CGI-S and CGI-I, mean change from baseline at 6 weeks

Mental state: Calgary Depression Scale for Schizophrenia (CDSS), mean change from baseline at 6 weeks

Adverse effects: Extrapyramidal symptoms (SAS, BAS and AIMS scales), mean change from baseline at 6 weeks

Adverse effects: Extrapyramidal symptoms (parkinsonism, akathisia, dystonia, rigidity)

Adverse effects: Other (insomnia, oral hypoaesthesia, sleepiness, agitation, headache, anxiety, weight loss, weight gain)

Adverse effects: Autonomic (sedation)

Use of anti-Parkinson medication

Unable to use (measurements of variance not reported) -

PANSS total score mean change from baseline to day 42 (primary outcome)

Modified International Suicide Prevention Trial (InterSePT) Scale for Suicidal Thinking

Readiness to Discharge Questionnaire (RDQ)


Notes*Patients were hospitalised for at least 2 weeks.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk“Randomized”, no further details provided.

Allocation concealment (selection bias)Unclear risk“Randomized”, no further details provided.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk"Double-blind" no further details reported.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk"Double-blind", no further details reported.

Incomplete outcome data (attrition bias)
All outcomes
Low risk"Safety assessments were made using data from the treated population (all patients who received >1 dose of study medication); efficacy analyses were based on data from the intent-to-treat (ITT) population (treated patients who had >1 postbaseline PANSS assessment)"

Selective reporting (reporting bias)High riskNot all pre-specified outcomes were reported fully.

Other biasHigh risk"funded by Schering-Plough Corporation, now Merck & Co, Inc" 

Klieser 1989

MethodsAllocation: random assignment.
Blindness: double-blind.
Duration: 3 weeks.
Location: hospital.
Design: parallel.
Setting: inpatients.
Country: not reported.

Consent: not stated.


ParticipantsDiagnosis: schizophrenia (63%), major depressive disorder (37%).
N = 120.
History: chronic, hospitalised.
Sex: 41% M, 59% F.
Age: mean ˜ 43 years.
Exclusions: not stated.


Interventions1. Haloperidol: dose 20 mg/day. N = 20*.
2. Placebo. N = 16.
3. Trazodone: dose 400 mg/day. N = 17.
4. Amitriptyline: dose 150 mg/day. N = 22.
Biperiden as required.


OutcomesLeaving the study early
Mental state: BPRS.


Notes*number of people with schizophrenia.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"Randomised" no further details reported.

Allocation concealment (selection bias)Unclear riskNo information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk"Double-blind" no further details reported.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk"Double-blind" no further details reported.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNumber of losses from each treatment group and reasons for loss to follow-up not reported. "Fourteen patients left the study on day 3, 19 patients left on day 7, and 6 patients left on day 14".

Selective reporting (reporting bias)High riskNot all expected outcomes were reported.

Other biasUnclear riskSource of funding not reported.

Marder 1994

MethodsAllocation: random assignment.
Blindness: double-blind.
Duration: 8 weeks (preceded by 1 week placebo wash out).
Location: multicentre.
Design: parallel.
Setting: inpatients.
Country: USA.

Consent: given.


ParticipantsDiagnosis: (DSM-III-R) schizophrenia.
N = 388.
History: hospitalised, chronic.
Sex: 89% M, 11% F.
Age: mean ˜37 years.
Exclusions: physically unhealthy, schizoaffective disorder.


Interventions1. Haloperidol: dose 20 mg/day. N = 66.
2. Placebo: N = 66.
3. Risperidone: dose 2 mg/day. N = 63.
4. Risperidone: dose 6 mg/day. N = 64.
5. Risperidone: dose 10 mg/day. N = 65.
6. Risperidone: dose 16 mg/day. N = 64.
Lorazepam or chloral hydrate as required.


OutcomesLeaving the study early

Unable to use -
Global state: CGI (> 50% loss).
Mental state: PANSS (> 50% loss).
Adverse events: EPS, UKU scale (> 50% loss).


NotesPart of 'North America 1997'.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"...randomization was in blocks of 12" no further details reported.

Allocation concealment (selection bias)Unclear riskNo information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk"Double-blind" no further details reported.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo information reported.

Incomplete outcome data (attrition bias)
All outcomes
High risk"early termination in 53% of the patients...62 % of the placebo patients; and 38% of patients receiving haloperidol...Both observed case and last observation carried forward (or end point) analyses were performed".

Selective reporting (reporting bias)Low riskAll expected outcomes are reported.

Other biasUnclear riskSource of funding from a pharmaceutical company "Supported by a grant from the Janssen Research Foundation".

Meltzer 2004

MethodsAllocation: randomised.
Blinding: double-blind.
Duration: 6 weeks.
Design: parallel.
Setting: inpatients up to day 15 after randomisation.
Country: USA.

Consent: written.


ParticipantsDiagnosis: schizophrenia and schizoaffective disorder DSM-IV.
N = 481.
Age: 18-64 ( range 35.4-37.5).
Sex: M 355, F 126.
History: hospitalised at baseline; a total score on the PANSS greater than 65 at screening and baseline, a minimum severity of illness score of 4 (moderately ill) on the CGI at screening and baseline.
Exclusions: patients with other axis I DSM-IV diagnoses; patients considered by the investigator to have been non-responsive to treatment with at least two different classes of antipsychotic medications; patients with any clinically significant medical illnesses; patients with clinical laboratory or ECG abnormalities; patients with evidence of current substance abuse or dependence; patients who were a danger to themselves or others.


Interventions1. Haloperidol: dose 10 mg/day. N = 98.

2. Placebo. N = 98.

3. 5-HT2A/2C antagonist: dose 5 mg/day. N = 70.

4. NK3 antagonist: dose 200 mg/day. N = 67.

5. CB1 antagonist: dose 20 mg/day. N = 69.

6. NTS1 antagonist: dose 180 mg/day. N = 63.


OutcomesLeaving the study early*

Unable to use (losses to follow-up > 50%) -

Mental state: PANSS (total, positive, negative, general), mean change from baseline at 6 weeks

BPRS (total), mean change from baseline at 6 weeks

Global state: CGI-I, mean endpoint score at 6 weeks

Global state: CGI-S, mean change from baseline at 6 weeks

Mental state: Calgary Depression Scale (CDS), mean change from baseline at 6 weeks

Adverse effects: Extrapyramidal symptoms: various scales (SAS, BAS, AIMS), mean change from baseline at 6 weeks

Adverse effects: Other (headache, insomnia, psychosis, agitation, abdominal pain, dyspepsia, vomiting, extrapyramidal symptoms, hyperkinesia)


Notes*This study had > 50% losses to follow-up, only the outcome Leaving the study early could be used.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk“Randomized”, no further details reported.

Allocation concealment (selection bias)Unclear risk No information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk“Double-blind”, no further details reported.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk“Double-blind”, no further details reported.

Incomplete outcome data (attrition bias)
All outcomes
High riskLosses to follow-up greater than 50%.

Selective reporting (reporting bias)Low risk All expected outcomes are reported.

Other biasHigh riskFunding sources are pharmaceutical companies, "Dr. Meltzer has received grant support from and is a consultant to Acadia, AstraZeneca, Eli Lilly, GlaxoSmithKline, Janssen Pharmaceutica, Lundbeck, Novartis, Pfizer, Sanofi-Synthelabo, and Solvay. He is a consultant to Psychiatric Genomics, Precision Med, Pharmacia, and Roche. Drs. Arvanitis and Rein and Ms. Bauer are employees of Sanofi-Synthelabo."

NCT00044044 2002

MethodsAllocation: randomised.
Blinding: double-blind.
Duration: 6 weeks.
Design: parallel.
Setting: inpatients.
Country: USA.

Consent: not stated.


ParticipantsDiagnosis: schizophrenia.
N = 330.
Age: mean (SD): 41.2 (10.0) years.
Sex: M 262, F 91.
History: hospitalised with acute or relapsing schizophrenia within 3 weeks of screening, a duration of illness of at least one year.
Exclusions: psychiatric hospitalisations other than current hospitalisations within 1 month prior to screening; treatment resistant; substance abuse; prolactin level of > 200 ng/mL at baseline; pregnancy.


Interventions1. Lurasidone: dose 20mg/day. N = 71.

2. Lurasidone: dose 40mg/day. N = 69.

3. Lurasidone: dose 80mg/day. N = 71.

4. Haloperidol: dose 10mg/day. N = 73.

5. Placebo. N = 72.


OutcomesLeaving the study early*

Unable to use (losses to follow-up > 50%) -

Adverse effects: Extrapyramidal symptoms (akathisia, tremor, dystonia, EPS)

Adverse effects: Autonomic (sedation)

Adverse effects: Cardiovascular and gastric (dizziness, diarrhoea, constipation, abdominal discomfort)

Adverse effects: Other (nausea, vomiting, headache, sleepiness, agitation, anxiety, insomnia)

Mental state: BPRS total, mean change from baseline at 6 weeks

Mental state: PANSS, mean change from baseline at 6 weeks

Global state: CGI-S, mean change from baseline at 6 weeks

Mental state: MADRS, mean change from baseline at 6 weeks


NotesClinicalTrials.gov Identifier: NCT00044044.

*This study had > 50% losses to follow-up, only the outcome Leaving the study early could be used.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk“Randomized”, no further details reported.

Allocation concealment (selection bias)Unclear risk“Randomized”, no further details reported.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskThe placebo was an "Oral Capsule matching treatment" indicating blinding of participants. Further, the study was described as “Double blind”, but there were no further details on blinding of personnel.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk“Double blind”, no further details reported.

Incomplete outcome data (attrition bias)
All outcomes
High riskThere was a high rate of drop-outs (> 50%).

Selective reporting (reporting bias)Low riskAll pre-stated outcomes were reported.

Other biasHigh risk"Sponsored by: Sumitomo Pharmaceuticals America" 

Nishikawa 1982

MethodsAllocation: random assignment.
Blindness: double-blind.
Duration: 3 years*.
Location: not stated.
Design: cross-over.
Setting: outpatients.
Country: Japan.

Consent: not stated.


ParticipantsDiagnosis: schizophrenia
N = 55.
History: outpatients, currently in remission, but have a history of several relapse episodes.
Sex: 67% M, 33% F.
Age: ˜ 25 - 41 years.
Exclusions: history of taking medication irregularly.


Interventions1. Haloperidol: dose 3 mg/day. N = 10.
2. Placebo. N = 10.
3. Chlorpromazine: dose 75 mg/day. N = 10.
4. Diazepam: dose 15 mg/day. N = 13.
5. Imipramine: dose 50 mg/day. N = 12.
Nitrazepam or biperiden as required.


OutcomesRelapse: number remaining in remission < 1 year.

Unable to use -
Leaving the study early (unclear when losses occurred, no individual group data given).


Notes*Initial 'cross-over' design of trial was disregarded after participant/clinician reluctance to switch neuroleptics. Data taken from first arm only.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"Randomly assigned" no further information reported.

Allocation concealment (selection bias)Unclear riskNo information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low risk"Double-blind", "Drug appearance, with respect to powder colour, taste and volumes, was made identical by adding a kind of stomachics, SMP (Sankyo, Japan)."

Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk"Double-blind" no further details reported.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskStudy does not report which treatment groups the losses to follow-up were from "Nine patients were dropped from the study for various reasons. The reasons included: failure to report to the hospital for scheduled appointment (N=3); admissions to other hospitals (N=2); strong requests from the patient not to change the previous drug (N=3); and a suicide after admission to the hospital (N=1)".

Selective reporting (reporting bias)High riskNot all expected outcomes were reported.

Other biasUnclear riskSource of funding was not reported.

Nishikawa 1984

MethodsAllocation: random assignment.
Blindness: double-blind.
Duration: 1 year.
Location: not stated.
Design: parallel.
Setting: outpatients.
Country: Japan.

Consent: not stated.


ParticipantsDiagnosis: (DSM III) schizophrenia.
N = 87.
History: in recovery stage of remission.
Sex: 61% M, 39% F.
Age: ˜ 28 - 54 years.
Exclusions: not stated.


Interventions1. Haloperidol: dose 1 mg/day. N = 13.
2. Haloperidol: dose 3 mg/day. N = 12.
3. Haloperidol: dose 6 mg/day. N = 12.
4. Placebo. N = 13.
5. Propericiazine: dose 10 mg/day. N = 13.
6. Propericiazine: dose 30 mg/day. N = 13.
6. Propericiazine: dose 60 mg/day. N = 13.
Each drug combined with nitrazepam and biperiden.


OutcomesRelapse: number remaining in remission < 1 year.
Leaving the study early.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"Randomly assigned" no further details reported.

Allocation concealment (selection bias)Unclear riskNo information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low risk"Double-blind", "Drug appearance, with respect to powder colour, taste and volume, was made identical by gastric aid, SMP (Sankyo, Japan)".

Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk"Double-blind" no further details reported.

Incomplete outcome data (attrition bias)
All outcomes
High risk"Among the entire group of 87 patients, 19 patients continued to receive the assigned drugs since they were in remission during the entire year of the trial, while other patients discontinued use of the assigned drugs due to overdose (N=16), relapse (N=48) and drop-out (N=4)".

Selective reporting (reporting bias)High riskResults not reported for placebo group for number of symptom free days and serum prolactin.

Other biasUnclear riskSource of funding not reported.

Potkin 2008

MethodsAllocation: randomised.
Blinding: double-blind.
Duration: 6 weeks.
Design: parallel.
Setting: not stated.
Country: not stated.

Consent: written.


ParticipantsDiagnosis: schizophrenia, DSM-IV.
N = 621.
Age: 18-65, (range 37-40.1).
Sex: M 443, F 178.
History: acute or subacute exacerbation of schizophrenia and PANSS total score of at least 60 at screening and at baseline.
Exclusions: not stated.


Interventions1. Iloperidone: dose 4 mg/day. N = 121.

2. Iloperidone: dose 8 mg/day. N = 125.

3. Iloperidone: dose 12 mg/day. N = 124.

4. Haloperidol: dose 15 mg/day. N = 124.

5. Placebo. N = 127.


OutcomesLeaving the study early*

Unable to use (losses to follow-up > 50%, variance not reported) -

Mental state: PANSS total, PANSS positive, PANSS negative, PANSS general psychopathology, BPRS, mean change from baseline at 6 weeks

Adverse effects


Notes*This study had > 50% losses to follow-up, only the outcome Leaving the study early could be used.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk“Randomized”, no further details reported.

Allocation concealment (selection bias)Unclear risk No information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk“Double-blind”, no further details reported.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk “Double-blind”, no further details reported.

Incomplete outcome data (attrition bias)
All outcomes
High riskLosses to follow-up greater than 50%.

Selective reporting (reporting bias)High riskSDs not reported for PANSS or BPRS.

Other biasHigh riskSource of funding from pharmaceutical companies. "Dr Potkin has received grant funding from Astra- Zeneca, Bioline, Bristol-Myers Squibb, Dainippon-Sumitomo, Elan, Forest Laboratories, Fujisawa Healthcare, Janssen Pharmaceutica, Merck, Novartis, Ono, Organon, Otsuka, Pfizer Inc, Solvay Pharmaceuticals, Roche"

Selman 1976

MethodsAllocation: random assignment.
Blindness: double-blind.
Duration: 12 weeks (preceded by a 2 week medication free period).
Location: not stated.
Design: parallel.
Setting: inpatients.
Country: USA.

Consent: not stated.


ParticipantsDiagnosis: schizophrenia.
N = 87.
History: acute, hospitalised.
Sex: 80% M, 20% F.
Age: mean ˜ 33 years.
Exclusions: other significant physical or neurological disorder, < 16 or > 60 years, females of child bearing age, epileptics, substance abuse.


Interventions1. Haloperidol: dose 4-12 mg/day. N = 29.
2. Loxapine: dose 50-150 mg/day. N = 29.
3. Placebo. N = 29.
Chloral hydrate or paraldehyde as required.


OutcomesAdverse effects: various observed effects.
Global effect: CGI improved/not improved.
Leaving the study early.

Unable to use -
Mental state: BPRS (no SD).
Behaviour: NOSIE (no SD).


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"Randomly assigned" no further details reported.

Allocation concealment (selection bias)Unclear riskNo information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low risk"A double-blind process was used in which neither patient nor investigator knew what medication was received until after the study was completed", "All medication was administered in identically appearing capsules".

Blinding of outcome assessment (detection bias)
All outcomes
Low risk"A double-blind process was used in which neither patient nor investigator knew what medication was received until after the study was completed".

Incomplete outcome data (attrition bias)
All outcomes
Unclear risk"Eight patients were excluded (1 loxapine, 3 haloperidol and 4 placebo", "All eight were dropped for administrative reasons such as unauthorised departure from the hospital or family objections to patients' participation in the study".

Last observation carried forward method used "Patients who continued beyond the fourth week but did not complete the full 12 weeks were included in the analysis through their final rating period, at either 4 or 8 weeks. These included 29 patients".

Selective reporting (reporting bias)High riskAll pre-stated outcomes were not fully reported (SDs were not reported for the BPRS and NOSIE scales).

Other biasHigh riskLoxitane (Loxapine succinate) supplied by Lederle Laboratories, Division of America Cyanamid Co, Pearle River, New York, who also supported the study.

Serafetinides 1972

MethodsAllocation: random assignment.
Blindness: double-blind.
Duration 3 months.
Location: not stated.
Design: parallel.
Setting: inpatients.
Country: not stated.

Consent: not stated


ParticipantsDiagnosis: schizophrenia.
N = 57.
History: chronic.
Sex: 42% M, 48% F .
Age: mean ˜ 42 years.
Exclusions: other physical or neurological disorder.


Interventions1. Haloperidol: dose up to 15 mg/day. N = 14.
2. Placebo. N = 14.
3. Clopenthixol: dose up to 250 mg/day. N = 15.
4. Chlorpromazine: dose up to 1000 mg/day. N = 14.


OutcomesAdverse events: various observed effects.
Global effect: CGI improved/not improved.
Leaving the study early.

Unable to use -
Mental state: BPRS (no SD).
Behaviour: NOSIE, OBRS (no SD).
Cognitive response (no data given).


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"Randomly assigned" no further details reported.

Allocation concealment (selection bias)Unclear riskNo information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low risk"Double-blind", "All medications were prepared in identically appearing capsules".

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo information reported.

Incomplete outcome data (attrition bias)
All outcomes
Low risk"Four of the 57 subjects, three on CPZ, and one on PL, failed to complete the 12 weeks of study. The PL subject and one CPZ subject were terminated because of behavioural deterioration after 4 and 8 weeks respectively. The other two CPZ subjects developed intestinal obstruction and were terminated in the 7th week of study".

Selective reporting (reporting bias)High riskAll pre-stated outcomes were not reported (no data reported for cognitive response, no SDs reported for the BPRS, NOSIE and OBRS scales).

Other biasLow riskStudy supported in part by US Public Health Service Grant MH 11666 and by National Institute of Mental Health Research Scientist Development Award K135278.

Simpson 1967

MethodsAllocation: random assignment.
Blindness: double-blind.
Duration: 14 weeks.
Location: not stated.
Design: parallel.
Setting: inpatients.
Country: not stated.

Consent: not stated.


ParticipantsDiagnosis: schizophrenia.
N = 24.
History: chronic, hospitalised.
Sex: male.
Age: ˜ 37 years.
Exclusions: other significant physical or neurological disorder.


Interventions1. Haloperidol: dose 6 mg/day. N = 8.
2. Haloperidol: dose 30 mg/day. N = 8.
3. Placebo. N = 8.
Benztropine mesylate as required.


OutcomesAdverse events: use of antiparkinson medication.
Global effect: improved/not improved.
Leaving the study early.

Unable to use -
Mental state: IMS, PRP (no SD).


NotesGroup numbers not stated in text, reviewers have assumed that they were divided into three sets of 8 (see table 1).


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"Randomly assigned" no further details reported.

Allocation concealment (selection bias)Unclear riskNo information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk"Double-blind"; "the staff correctly guessed which patients were on active medication in a high percentage of cases".

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskThe code was broken at the end of the study.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo details reported.

Selective reporting (reporting bias)High riskAll pre-stated outcomes were not fully reported (SDs were not reported for IMPS and PRP).

Other biasUnclear risk"The haloperidol (HALDOL®) used in this study was supplied by McNeil Laboratories. Inc.," 

Spencer 1992

MethodsAllocation: random assignment.
Blindness: double-blind, cross-over.
Duration: 10 weeks (preceded by a 2-week single blind placebo washout).
Location: not stated.
Design: cross-over.
Setting: inpatients.
Country: USA.

Consent: not stated.


ParticipantsDiagnosis: (DSM-III-R) schizophrenia.
N = 12.
History: hospitalised.
Sex: 75% M, 25% F.
Age: ˜ 6 - 12 years.
Exclusions: other significant physical or neurological disorder, receipt of psychoactive medication within 4 weeks of study.


Interventions1. Haloperidol: dose 4 weeks 0.5-10 mg/day followed by 4 weeks placebo. N = 12.
2. Placebo: 4 weeks followed by 4 weeks 0.5-10 mg/day haloperidol. N = 12.


OutcomesGlobal effect: improved/not improved.
Leaving the study early.

Unable to use -
Global effect: CGI (no SD).
Mental state: CPRS, BPRS-C (no SD).
Adverse events: various observed effects (no data for placebo group).
Cognitive response: WISC-R, WPPSI, DICA-R (no data).


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"Random assignment" no further details reported.

Allocation concealment (selection bias)Unclear riskNo information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk"Double-blind" no further details reported.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk"Double-blind" For CPRS (one of the outcomes scales), the raters were "all blind to treatment status", no further details reported.

Incomplete outcome data (attrition bias)
All outcomes
Low riskThere were no drop-outs.

Selective reporting (reporting bias)High riskNo SDs reported for CGI, CPRS, BPRS-C. No data reported for adverse events for placebo group. No data reported for cognitive assessments (WISC-R, WPPSI, DICA-R).

Other biasUnclear riskStudy supported by NIMH Child and Adolescent Mental Health Academic Award MH-00763 and NIMH Institutional Training Grant MH-18915.

Haloperidol and placebo tablets provided by McNeil Pharmaceutical.

Vichaiya 1971

MethodsAllocation: random assignment.
Blindness: double-blind, cross-over.
Duration: 12 weeks (preceded by 4 weeks drug free period).
Location: hospital.
Design: cross-over.
Setting: inpatients.
Country: Thailand.

Consent: unknown.


ParticipantsDiagnosis: schizophrenia.
N = 30.
History: hospitalised, chronic.
Sex: female.
Age: mean ˜ 40 years.
Exclusions: unknown.


Interventions1. Haloperidol: dose 6 weeks 4.5 mg/day followed by 6 weeks placebo. N = 30.
2. Placebo: 6 weeks placebo followed by 6 weeks 4.5 mg/day haloperidol. N = 30.


OutcomesGlobal effect: FFS, improved/not improved.

Unable to use -
Adverse events: various observed effects (no group data).
Leaving the study early (no group data).


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"Randomized" no further details reported.

Allocation concealment (selection bias)Unclear riskNo information reported.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk"Double-blind"; "The code was worked out by the head of Female In-patient section, which kept it secret until the investigation had ended."; "...it soon became clear that the trial was not, in fact blind. The patients on haloperidol showed extrapyramidal symptoms".

Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk"Double-blind"; "The code was worked out by the head of Female In-patient section, which kept it secret until the investigation had ended."

Incomplete outcome data (attrition bias)
All outcomes
Low riskOnly one drop-out reported.

Selective reporting (reporting bias)Unclear riskSide-effects in placebo group not mentioned.

Other biasUnclear riskDoes not report source of funding.

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Akhondzadeh 2005Allocation: unclear.
Participants: people with schizophrenia.
Interventions: haloperidol versus haloperidol plus allopurinol.

Allison 2007Allocation: post-hoc analysis of two RCTs (no references provided, conference abstract).

Alpert 1995Allocation: unclear.

Alphs 1993Allocation: unclear.
Participants: people with schizophrenia.
Interventions: haloperidol versus remoxipride versus placebo.
Outcomes: all data unusable - conference proceedings.

Andrezina 2006Allocation: random.
Participants: people with schizophrenia.
Interventions: intramuscular injections of haloperidol versus aripiprazole versus placebo

Arvanitis 2002Allocation: random.
Participants: people with schizophrenia.
Interventions: SR compound, haloperidol or placebo.
Outcomes: all data unusable, conference proceedings.

AstraZeneca 2001Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol versus quetiapine.

Augustin 1996Allocation: random.
Participants: people with schizophrenia.
Interventions: haloperidol versus placebo.
Outcomes: all data unusable - conference proceedings.

Azima 1960Allocation: unclear.
Participants: only 34/84 were people with schizophrenia.

Ban 1969Allocation: unclear.
Participants: people with schizophrenia.
Interventions: haloperidol + phenothiazine medication versus trifluperidol + phenothiazine medication versus placebo, not haloperidol alone.

Barbee 1992Allocation: random.
Participants: people with schizophrenia.
Interventions: haloperidol versus haloperidol plus alprazolam.

Bateman 1979Allocation: random.
Participants: unclear if people with schizophrenia.
Interventions: supplementation of original neuroleptic with haloperidol or placebo.

Baymiller 2002Allocation: random.
Participants: people with schizophrenia.
Interventions: clozapine versus haloperidol.

Ben-dor 1998Allocation: random.
Participants: people hospitalised with chronic schizophrenia.
Interventions: haloperidol + vitamin E versus haloperidol + placebo, no placebo only group.

Bersudsky 2006Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol plus phenytoin versus haloperidol.

Beuzen 1996Allocation: random.
Participants: healthy elderly, not people with schizophrenia.

Blum 1969Allocation: unclear.

Bogeum 2008Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol plus aripiprazole versus haloperidol plus placebo.

Brandrup 1961Allocation: random, cross-over study .
Participants: people with chronic schizophrenia.
Intervention: haloperidol 8 mg/day versus placebo.
Outcomes: no usable data from period before first cross-over.

Browne 1988Allocation: random.
Participants: people with chronic schizophrenia.
Intervention: haloperidol 10-160 mg/day versus placebo.
Outcomes: all data unusable, > 50% loss during double-blind phase, people relapsing could re-enter study under single-blind conditions.

Buchsbaum 1992Allocation: unclear, no recording of random allocation.

Cai 2009Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol versus haloperidol plus Bezoar xiexin tang (Chinese herbs).

Cao 2006Allocation: random.
Participants: people with schizophrenia.
Intervention: ziprasidon versus haloperidol.

Caroli 1975Allocation: not random, ABA design.

Cho-Boon 1989Allocation: random.
Participants: people with schizophrenia.
Interventions: haloperidol 12-18, 30-45, 60-90 mg/day versus placebo.
Outcomes: all data unusable - conference proceedings.

Contreas 1988Allocation: not random, ABA design.

Craft 1965Allocation: unclear.

Crow 1986Allocation: random.
Participants: 120 people with schizophrenia.
Interventions: haloperidol withdrawal versus continuation, not instigation of haloperidol treatment.

Czobor 1992Allocation: not an RCT, only one treatment group.

Deberdt 1971Allocation: not random, ABA design.

Diamond 1991Allocation: random.
Participants: men with schizophrenia.
Interventions: haloperidol 5-75 mg/day versus tiospirone 75-375 mg/day versus placebo.
Outcomes: all data unusable - conference proceedings.

Eklund 1990Allocation: random.
Participants: people with schizophrenia.
Interventions: intramuscular depot administration of haloperidol versus placebo

Eli Lilly 2005Allocation: random.
Participants: people with schizophrenia.
Interventions: intramuscular injections of haloperidol versus olanzapine versus placebo

Gelders 1986Allocation: random.
Participants: people with chronic schizophrenia.
Interventions: haloperidol 10 mg/day versus ritanserin 20/mg/day versus placebo.
Outcomes: all data unusable, no group numbers given for CGI, no data given for BPRS or adverse effects.

George 2000Allocation: random.
Participants: people with schizophrenia.
Intervention: intramuscular injection of olanzapine versus intramuscular injection of haloperidol versus intramuscular injection of placebo.

GlaxoSmithKline 2005Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol plus placebo versus haloperidol plus alosetron.

Glovinsky 1992Allocation: not random, ABA design.

Herrera 2005Allocation: random.
Participants: people with schizophrenia.
Intervention: risperidone plus placebo of haloperidol versus haloperidol plus placebo of risperidone.

Huygens 1973Allocation: random.
Participants: 40 women with psychosis.
Interventions: dexetimide versus placebo, adjunct to haloperidol.

IRCT138809201457N6Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol verus haloperidol plus celecoxib.

IRCT138809201457N7Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol versus haloperidol plus ascorbic acid.

Itil 1981Allocation: not random, ABA design.

Jolley 1990Allocation: random.
Participants: people in remission from schizophrenia.
Interventions: haloperidol withdrawal versus placebo, no instigation of haloperidol.

Jung 2007Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol versus aripiprazole plus haloperidol.

Kapur 2004Allocation: not an RCT, narrative review.

Kasper 1996Allocation: random.
Participants: people with schizophrenia.
Intervention: 4, 8 and 16 mg/day haloperidol versus 12, 20 and 24 mg/day sertindole or placebo.
Outcomes: all data unusable, conference proceedings.

Kim 2005Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol plus placebo versus haloperidol plus donepezil.

Kinon 2012Allocation: retrospective study of 3 already excluded or included RCTs.

Ko 1989Allocation: not random, ABA design.

Kostic 2005Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol versus aripiprazole.

Kramer 1989Allocation: random.
Participants: 56 people with schizophrenia.
Interventions: amitriptyline versus desmethylimipramine versus placebo in addition to haloperidol and benztropine, haloperidol not randomised.

Kurland 1981Allocation: random.
Participants: 28 people with schizophrenia and scoring >17 on HAM-D.
Interventions: adjunctive viloxazine versus placebo, antipsychotics, such as haloperidol, continued as normal.

Labarca 1993Allocation: not random, ABA design.

Lee 1968Allocation: unclear.

Lee 2007Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol plus placebo versus haloperidol plus donepezil.

Lehmann 1967Allocation: random.
Participants: 30 people with chronic schizophrenia.
Interventions: continued on current phenothiazine medication, then randomised to adjunctive haloperidol (1.5 mg/day) versus trifluperidol (0.75 mg/day) versus placebo, not haloperidol alone.

Lemmer 1993Allocation: random.
Participants: people with acute schizophrenia.
Interventions: pramipexole versus haloperidol or placebo.
Outcomes: all data unusable, study stopped early.

Li 2007Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol plus placebo versus clonazepam plus placebo versus haloperidol plus clonazepam.

Liang 1987Allocation: random.
Participants: people with schizophrenia.
Interventions: haloperidol versus insulin shock therapy.

Lindborg 2003Allocation: not random, meta-analysis.

Magelund 1979Allocation: random, cross-over.
Participants: 12 people hospitalised with schizophrenia.
Interventions: AMPT versus haloperidol, placebo given after each active treatment period, not randomised.

Malaspina 1997Allocation: random.
Participants: people with schizophrenia.
Interventions: haloperidol decanoate 0.3 mg/kg versus placebo.
Outcomes: all data unusable - conference proceedings.

Maoz 2000Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol versus haloperidol plus propanolol.

Meltzer 2008Allocation: random.
Participants: people with schizophrenia.
Intervention: risperidone plus placebo versus risperidone plus pimavanserin versus haloperidol plus placebo versus haloperidol plus pimavanserin.

Mossaheb 2006Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol plus clozapine versus clozapine plus placebo.

Nagaraja 1977Allocation: not random.

NCT00018850Allocation: random.
Participants: people with schizophrenia.
Intervention: ondasteron versus nicotine versus haloperidol

NCT00156104Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol versus asenapine, the first phase of the trial includes a placebo group but no results are reported for this phase.

NCT00189995Allocation: random.
Participants: people with schizophrenia.
Intervention:clozapine versus haloperidol

NCT00947375Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol plus placebo versus haloperidol plus lamictal.

NCT01161277Allocation: random.
Participants: not schizophrenia, psychiatrically healthy participants.

Necomer 1992Allocation: random.
Participants: 24 males with schizophrenia.
Interventions: haloperidol withdrawal versus placebo, not instigation of haloperidol.

Nguyen 1984Allocation: unclear.
Participants: people with chronic schizophrenia.
Interventions: haloperidol versus placebo.
Outcomes: all data unusable - conference proceedings.

North America 1997Allocation: random.
Participants: 523 people with chronic schizophrenia.
Interventions: haloperidol 20 mg/day versus placebo versus risperidone 2, 6, 10 or 16 mg/day.
Outcomes: combines data from several centres but all unique data from these reports of combined analyses is unusable due to > 50% loss; data from publications of specific centres can be included (Chouinard 1993, Marder 1994).

Octavio 2004Allocation: random.
Participants: people with schizophrenia.
Interventions: aripiprazole versus haloperidol.

Okasha 1964Allocation: random.
Participants: 80 people hospitalised with chronic psychosis.
Interventions: haloperidol 4.5 mg/day versus placebo.
Outcomes: no usable data, leaving study early (no group data), behaviour (no total scale score).

Ortega-Soto 1994Allocation: random.
Participants: drug free people with acute schizophrenia.
Interventions: threshold dose of haloperidol + 20 mg haloperidol versus threshold dose of haloperidol + placebo, not haloperidol versus placebo.

Ota 1973Allocation: unclear .
Participants: 54 people with chronic schizophrenia.
Interventions: haloperidol withdrawal versus placebo, not instigation of haloperidol.

Pathiraja 1995Allocation: random.
Participants: men with schizophrenia.
Interventions: haloperidol 5-20 mg/day versus risperidone 2-16 mg/day versus MAR 327 50-400 mg/day versus placebo.
Outcomes: all data unusable - conference proceedings.

Paykel 2000Allocation: random.
Participants: women with peurperal psychosis .

Pool 1976Allocation: random but with non-random additions.
Participants: 75 people with schizophrenia.
Interventions: haloperidol versus loxapine versus placebo.
Outcomes: no usable data - those who withdrew during the study were replaced with new patients and data for those randomised not reported separately.

Potkin 1984Allocation: random.
Participants: people with schizophrenia.
Interventions: haloperidol + IC (insulin coma therapy) versus placebo + IC, not haloperidol alone.
Outcomes: all data unusable - conference proceedings.

Potkin 1995Allocation: random.
Participants: people with schizophrenia.
Interventions: haloperidol versus MAR 327 150-300 mg/day versus placebo.
Outcomes: all data unusable - conference proceedings.

Potkin 2000Allocation: random.
Participants: people with schizophrenia.
Interventions: haloperidol versus clozapine or placebo.
Outcomes: PET scan study, outcomes not relevant, data unusable.

Price 1985Allocation: not random, ABA design.

Price 1987Allocation: random.
Participants: people with schizophrenia.
Interventions: haloperidol 10-4- mg/day versus verapamil 80 mg/day versus placebo.
Outcomes: all data unusable.

Rees 1965Allocation: random, cross-over.
Participants: 14 people with schizophrenia.
Interventions: haloperidol versus placebo.
Outcomes: all data unusable, none available before cross-over.

Reschke 1974Please note: this study was included in the previous versions of this review; as it was clarified that the route of haloperidol administration was not oral, but by intramuscular injection, we decided to exclude this study for this update.

Allocation: randomised.
Participants: patients with psychotic symptoms.
Interventions: intramuscular injections of haloperidol versus placebo versus chlorpromazene.

Roitman 1989Allocation: random.
Participants: 16 people with schizoaffective disorder.
Interventions: adjunctive demeclocycline (DMC) versus placebo, all received haloperidol.

Romeo 2009Allocation: random.
Participants: not schizophrenia, participants had intellectual disabilities.

Ruskin 1991Allocation: random.
Participants: 35 people with schizophrenia.
Interventions: haloperidol withdrawal versus placebo, not instigation of haloperidol.

Samuels 1961Allocation: unclear.

Shim 2007Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol plus placebo versus haloperidol plus aripiprazole.

Singh 1972Allocation: not random, ABA design.

Soloff 1986Allocation: random.
Participants: people with "Borderline disorders", not schizophrenia.

Stankovska 2002Allocation: unclear if randomised.

Taverna 1972Allocation: quasi-randomised.

Teja 1975Allocation: random.
Participants: people with chronic schizophrenia.
Interventions: haloperidol 36 mg/week versus chlorpromazine > 1800 mg/week versus trifluoperazine > 90 mg/week versus thiothixene > 90 mg/week versus placebo.
Outcomes: all data unusable.

Tran-Johnson 2007Allocation: random.
Participants: people with schizophrenia.
Interventions: intramuscular injections of haloperidol versus aripiprazole versus placebo

Van Lommel 1974Allocation: random.
Participants: 19 psychotic males.
Interventions: haloperidol withdrawal versus placebo, not instigation of haloperidol.

Veser 2006Allocation: random.
Participants: people with psychoses mainly caused by substance abuse, and a minority with bipolar disorder and schizophrenia.

Volavka 1992Allocation: random.
Participants: 173 people with acutely exacerbated schizophrenia.
Interventions: dosage levels of haloperidol, no placebo group.

Wang 2009Allocation: random.
Participants: people with schizophrenia.
Intervention: aripiprazole plus haloperidol versus placebo plus haloperidol.

Wilson 1994Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol plus placebo versus haloperidol plus lithium.

Wright 2001Allocation: random.
Participants: people with schizophrenia.
Interventions: intramuscular injections of haloperidol versus olanzapine versus placebo

Zhan 2000Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol plus water versus haloperidol plus levamisole ointment versus levamisole ointment plus placebo.

Outcome: all data unusable (reported measures of immune index and P values of correlation of immune index and BPRS/SAPS scores).

Zhang 2001Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol plus extract of Ginkgo biloba versus haloperidol plus placebo.

Zhang 2006Allocation: random.
Participants: people with schizophrenia.
Intervention: haloperidol plus placebo versus haloperidol plus ondansetron.

Zimbroff 1997Allocation: random.
Participants: people with schizophrenia.
Interventions: haloperidol 4, 8, 16 mg/day versus sertindole 12, 20, 24 mg/day versus placebo.
Outcomes: all data unusable due to > 50% loss, leaving study early (not given as individual group data).

 
Comparison 1. HALOPERIDOL versus PLACEBO

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Global state: 1a. Overall improvement: No marked global improvement11Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    1.1 up to 6 weeks (clinician rated)
4472Risk Ratio (M-H, Fixed, 95% CI)0.67 [0.56, 0.80]

    1.2 > 6-24 weeks (clinician rated)
8307Risk Ratio (M-H, Fixed, 95% CI)0.67 [0.58, 0.78]

    1.3 > 6-24 weeks (nurse rated)
128Risk Ratio (M-H, Fixed, 95% CI)0.59 [0.37, 0.92]

 2 Global state: 1b. Overall improvement: Average change in CGI-S score (up to 6 weeks; high = poor)2353Mean Difference (IV, Fixed, 95% CI)-0.49 [-0.73, -0.25]

 3 Global state: 2. Hospital discharge: Not discharged from hospital (> 6-24 weeks)133Risk Ratio (M-H, Fixed, 95% CI)0.85 [0.47, 1.52]

 4 Global state: 3. Relapse (< 52 weeks)270Risk Ratio (M-H, Fixed, 95% CI)0.69 [0.55, 0.86]

 5 Global state: 4. Leaving the study early24Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    5.1 up to 6 weeks
161812Risk Ratio (M-H, Fixed, 95% CI)0.87 [0.80, 0.95]

    5.2 > 6-24 weeks
8304Risk Ratio (M-H, Fixed, 95% CI)0.54 [0.29, 1.00]

    5.3 < 52 weeks
150Risk Ratio (M-H, Fixed, 95% CI)2.58 [0.14, 46.83]

 6 Mental state: 1. No clinical improvement (< 20% reduction in BPRS score; up to 6 weeks)124Risk Ratio (M-H, Fixed, 95% CI)0.76 [0.54, 1.08]

 7 Mental state: 2a. General symptoms: Average BPRS total score (up to 6 weeks; high = poor)3108Mean Difference (IV, Fixed, 95% CI)-9.76 [-14.60, -4.93]

 8 Mental state: 2b. General symptoms: Average change in PANSS total score (up to 6 weeks; high = poor)1119Mean Difference (IV, Fixed, 95% CI)-15.58 [-23.92, -7.24]

 9 Mental state: 3. Positive symptoms: Average change in PANSS positive score (up to 6 weeks; high = poor)2353Mean Difference (IV, Fixed, 95% CI)-3.29 [-4.70, -1.89]

 10 Mental state: 4. Negative symptoms: Average change in PANSS negative score (up to 6 weeks; high = poor)2353Mean Difference (IV, Fixed, 95% CI)-1.18 [-2.32, -0.04]

 11 Mental state: 5. Mood: Average change in CDS score (up to 6 weeks; high = poor)1234Mean Difference (IV, Fixed, 95% CI)-0.30 [-1.20, 0.60]

 12 Adverse effects: 1a. Movement disorders: Extrapyramidal symptoms11Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    12.1 akathisia
6695Risk Ratio (M-H, Fixed, 95% CI)3.66 [2.24, 5.97]

    12.2 dystonia
5471Risk Ratio (M-H, Fixed, 95% CI)11.49 [3.23, 40.85]

    12.3 needing antiparkinson medication
4480Risk Ratio (M-H, Fixed, 95% CI)3.23 [2.20, 4.72]

    12.4 oculogyric crises
283Risk Ratio (M-H, Fixed, 95% CI)0.97 [0.14, 6.57]

    12.5 parkinsonism (including EPS)
5485Risk Ratio (M-H, Fixed, 95% CI)5.48 [2.68, 11.22]

    12.6 rigidity
5461Risk Ratio (M-H, Fixed, 95% CI)4.98 [2.74, 9.05]

    12.7 teeth grinding
133Risk Ratio (M-H, Fixed, 95% CI)2.53 [0.11, 57.83]

    12.8 'thick' speech
133Risk Ratio (M-H, Fixed, 95% CI)5.89 [0.33, 105.81]

    12.9 tremor
5447Risk Ratio (M-H, Fixed, 95% CI)3.93 [1.96, 7.91]

 13 Adverse effects: 1b. Movement disorders: Tardive dyskinesia2Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    13.1 dyskinesia and tardive dyskinesia
2157Risk Ratio (M-H, Fixed, 95% CI)1.00 [0.14, 7.13]

 14 Adverse effects: 1c. Movement disorders: Average changes scores (various scales; up to 6 weeks)1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    14.1 AIMS (high = poor)
1231Mean Difference (IV, Fixed, 95% CI)-0.29 [-0.71, 0.13]

    14.2 BAS (high = poor)
1231Mean Difference (IV, Fixed, 95% CI)0.31 [0.10, 0.52]

    14.3 SAS (high = poor)
1231Mean Difference (IV, Fixed, 95% CI)1.48 [0.76, 2.20]

 15 Adverse effects: 2. Other CNS3Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    15.1 blurred vision
2240Risk Ratio (M-H, Fixed, 95% CI)3.96 [1.21, 12.93]

    15.2 confusion
133Risk Ratio (M-H, Fixed, 95% CI)2.53 [0.11, 57.83]

    15.3 dry mouth
133Risk Ratio (M-H, Fixed, 95% CI)1.67 [0.62, 4.46]

    15.4 sedation
1238Risk Ratio (M-H, Fixed, 95% CI)0.86 [0.24, 3.11]

 16 Adverse effects: 3. Cardiovascular effects5Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    16.1 blood pressure - dizziness/low BP
3245Risk Ratio (M-H, Fixed, 95% CI)1.01 [0.36, 2.79]

    16.2 blood pressure - high BP
116Risk Ratio (M-H, Fixed, 95% CI)3.0 [0.14, 64.26]

    16.3 bradycardia
1124Risk Ratio (M-H, Fixed, 95% CI)4.27 [0.49, 37.10]

 17 Adverse effects: 4. Other adverse effects9Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    17.1 agitation
2362Risk Ratio (M-H, Fixed, 95% CI)1.07 [0.54, 2.12]

    17.2 anxiety
2362Risk Ratio (M-H, Fixed, 95% CI)0.84 [0.33, 2.16]

    17.3 drooling
3207Risk Ratio (M-H, Fixed, 95% CI)4.00 [0.88, 18.21]

    17.4 facial oedema
133Risk Ratio (M-H, Fixed, 95% CI)2.83 [0.12, 64.89]

    17.5 headache
4593Risk Ratio (M-H, Fixed, 95% CI)0.93 [0.62, 1.39]

    17.6 infection
124Risk Ratio (M-H, Fixed, 95% CI)7.0 [0.40, 122.44]

    17.7 insomnia
4629Risk Ratio (M-H, Fixed, 95% CI)1.11 [0.76, 1.63]

    17.8 nausea/vomiting
2231Risk Ratio (M-H, Fixed, 95% CI)0.90 [0.49, 1.65]

    17.9 oral hypoaesthesia
1238Risk Ratio (M-H, Fixed, 95% CI)0.15 [0.01, 2.92]

    17.10 perspiration
293Risk Ratio (M-H, Fixed, 95% CI)4.74 [0.58, 38.81]

    17.11 sleepiness
7686Risk Ratio (M-H, Fixed, 95% CI)3.09 [1.51, 6.31]

    17.12 weight gain
2441Risk Ratio (M-H, Fixed, 95% CI)4.89 [1.41, 16.95]

    17.13 weight loss
3385Risk Ratio (M-H, Fixed, 95% CI)0.77 [0.36, 1.64]

 18 SUBGROUP ANALYSIS: 1. MEN vs WOMEN: Global state: Overall improvement: No marked global improvement, > 6-24 weeks (clinician rated)3Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    18.1 only men
124Risk Ratio (M-H, Fixed, 95% CI)0.47 [0.27, 0.82]

    18.2 only women
288Risk Ratio (M-H, Fixed, 95% CI)0.69 [0.55, 0.87]

 19 SUBGROUP ANALYSIS: 2. 18-65 YEARS vs < 18 YEARS: Global state: Overall improvement: No marked global improvement, > 6-24 weeks (clinician rated)8Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    19.1 18-65 years
7284Risk Ratio (M-H, Fixed, 95% CI)0.70 [0.61, 0.80]

    19.2 < 18 years
124Risk Ratio (M-H, Fixed, 95% CI)0.27 [0.10, 0.74]

 20 SUBGROUP ANALYSIS: 3. ACUTE vs CHRONIC: Global state: Overall improvement: No marked global improvement, > 6-24 weeks (clinician rated)8Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    20.1 acute phase of illness
158Risk Ratio (M-H, Fixed, 95% CI)0.56 [0.30, 1.06]

    20.2 chronic phase of illness
7250Risk Ratio (M-H, Fixed, 95% CI)0.68 [0.59, 0.78]

 21 SUBGROUP ANALYSIS: 4. LOW DOSE vs MEDIUM TO HIGH DOSE11Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    21.1 low dose (≤ 5 mg/day) - Global state: Overall improvement: No marked global improvement, up to 6 weeks (clinician rated)
1234Risk Ratio (M-H, Fixed, 95% CI)0.85 [0.69, 1.04]

    21.2 medium to high dose (> 5 mg/day) - Global state: Overall improvement: No marked global improvement, up to 6 weeks (clinician rated)
3238Risk Ratio (M-H, Fixed, 95% CI)0.48 [0.35, 0.66]

    21.3 low dose (≤ 5 mg/day) - Global state: Overall improvement: No marked global improvement, > 6-24 weeks (clinician rated)
4149Risk Ratio (M-H, Fixed, 95% CI)0.67 [0.55, 0.81]

    21.4 medium to high dose (> 5 mg/day) - Global state: Overall improvement: No marked global improvement, > 6-24 weeks (clinician rated)
5165Risk Ratio (M-H, Fixed, 95% CI)0.67 [0.54, 0.83]

 22 SUBGROUP ANALYSIS: 5. DIAGNOSTIC CRITERIA vs NO DIAGNOSTIC CRITERIA13Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    22.1 operational criteria used for diagnosis - Global state: Overall improvement: No marked global improvement, up to 6 weeks (clinician rated)
3414Risk Ratio (M-H, Fixed, 95% CI)0.67 [0.56, 0.81]

    22.2 no operational criteria used for diagnosis - Global state: Overall improvement: No marked global improvement, up to 6 weeks (clinician rated)
158Risk Ratio (M-H, Fixed, 95% CI)0.64 [0.33, 1.24]

    22.3 operational criteria used for diagnosis - Global state: Overall improvement: No marked global improvement, > 6-24 weeks (clinician rated)
124Risk Ratio (M-H, Fixed, 95% CI)0.27 [0.10, 0.74]

    22.4 no operational criteria used for diagnosis - Global state: Overall improvement: No marked global improvement, > 6-24 weeks (clinician rated)
7284Risk Ratio (M-H, Fixed, 95% CI)0.70 [0.61, 0.80]

    22.5 operational criteria used for diagnosis - Global state: Relapse (< 52 weeks)
150Risk Ratio (M-H, Fixed, 95% CI)0.72 [0.57, 0.91]

    22.6 no operational criteria used for diagnosis - Global state: Relapse (< 52 weeks)
120Risk Ratio (M-H, Fixed, 95% CI)0.62 [0.37, 1.03]

 23 SENSITIVITY ANALYSIS: 1. ASSUMPTIONS FOR MISSING DATA vs NO ASSUMPTIONS FOR MISSING DATA: Global state: Overall improvement: No marked global improvement, up to 6 weeks (clinician rated)3Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    23.1 no assumptions for missing data
3353Risk Ratio (M-H, Fixed, 95% CI)0.72 [0.60, 0.87]

 
Summary of findings for the main comparison. HALOPERIDOL versus PLACEBO for schizophrenia

HALOPERIDOL versus PLACEBO for schizophrenia

Patient or population: patients with schizophrenia
Settings: hospital and community
Intervention: HALOPERIDOL versus PLACEBO

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

ControlHALOPERIDOL versus PLACEBO

Death - suicide and natural causesSee commentSee commentNot estimable0
(0)
See commentNo studies reported on this outcome.

Overall improvement: No marked global improvement
Rated by clinician
Follow-up: >6-24 weeks
841 per 1000564 per 1000
(488 to 656)
RR 0.67
(0.58 to 0.78)
307
(8 studies)
⊕⊕⊕⊝
moderate1
Another four trials reported on this outcome at up to six weeks follow-up, and one trial at > 6-24 weeks follow-up using a nurse-rated scale, both sub-analyses showed significant results in favour of haloperidol.

Not discharged from hospital
Follow-up: > 6-24 weeks
625 per 1000531 per 1000
(294 to 950)
RR 0.85
(0.47 to 1.52)
33
(1 study)
⊕⊝⊝⊝
very low2,3,4

Relapse
Follow-up: < 52 weeks
1000 per 1000690 per 1000
(550 to 860)
RR 0.69
(0.55 to 0.86)
70
(2 studies)
⊕⊝⊝⊝
very low3,5,6

Leaving the study early
Follow-up: > 6-24 weeks
134 per 100072 per 1000
(39 to 134)
RR 0.54
(0.29 to 1)
304
(8 studies)
⊕⊕⊕⊝
moderate1
Another 16 trials reported on this outcome at up to six weeks follow-up showing a significant result in favour of haloperidol. One trial at < 52 weeks follow-up showed no difference between haloperidol and placebo.

Satisfaction with treatmentSee commentSee commentNot estimable0
(0)
See commentNo studies reported on this outcome.

Adverse effects: Movement disorders - parkinsonism
Follow-up: 3 weeks to 3 months
28 per 1000154 per 1000
(75 to 315)
RR 5.48
(2.68 to 11.22)
485
(5 studies)
⊕⊕⊕⊝
moderate7
Several studies also reported on other, specific movement disorders: there was a significant result favouring placebo for akathisia, dystonia, needing anti-Parkinson medication, rigidity and tremor; there was no difference between haloperidol and placebo for tardive dyskinesia, oculogyric crises, teeth grinding and 'thick' speech.

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval; RR: Risk ratio;

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

 1 Seven out of the eight included studies had an unclear risk of bias for random sequence generation and for allocation concealment. Blinding of participants and personnel was unclear in four studies and blinding of assessors was unclear in six. Two studies had an unclear risk of bias for incomplete outcome data. One study had a high risk of other bias as they were funded by industry and three an unclear risk of bias as the drugs were provided by a pharmaceutical company.
2 The included study had an unclear risk of bias for random sequence generation, allocation concealment, and blinding of outcome assessors.
3 The total number of participants and events were very low.
4 Only one out of the 25 included studies reported on this outcome.
5 The two included studies had an unclear risk of bias for random sequence generation, allocation concealment, and for blinding of outcome assessors. One study had a high risk of bias for incomplete outcome data, the other an unclear risk of bias.
6 Only two out of the 25 included studies reported on this outcome.
7 Four out of the five included studies had an unclear risk of bias for random sequence generation, and all five studies had an unclear risk of bias for allocation concealment. Blinding of participants and personnel was unclear in two studies and blinding of assessors was unclear in four. One study had a high risk of bias for incomplete outcome data, and two for other bias as they were funded by industry or the drugs were provided by a pharmaceutical company.