Methods

Allocation: randomised.
Blindness: double.
Duration: 1 week before cross‐over.
Design: cross‐over.
Country: Denmark.

Funding: Not mentioned.

Participants

Diagnosis: paranoid schizophrenia.
History: not clear in article but states duration of illness 2 months ‐ 2 years.
N = 9.
Age: mean 58 years, range 39‐81.
Sex: all male.
Inclusion criteria: all patients who had been previously treated with cis(Z)/trans(E)‐clopenthixol (from 2 months to several years).

Exclusion: not mentioned.
Setting: not reported.

Interventions

1. Cis(Z)‐clopenthixol hydrochloride: dose range 5 mg to 50 mg/day. Taken twice daily. N = 4.
2. Clopenthixol hydrochloride (cis(Z) and trans(E) forms): dose 10 mg to 100 mg/day. Taken twice daily. N = 5.

Outcomes

5.2 Leaving the study early for General Reasons (zero people left the study early).

Unable to use: Physiological ‐ serum concentrations of Cis(Z)‐clopenthixol and clopenthixol. Non‐clinical data all reported post‐cross‐over.

Notes

Cross‐over trial. Patient 4 excluded during statistical analysis (also received multiple drugs).

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Done "in a randomised way". Authors do not state how.

Attempted to contact authors but Dikemark Hospital no longer exists.

Allocation concealment (selection bias)

Unclear risk

Authors do not comment.

Attempted to contact authors but Dikemark Hospital no longer exists.

Blinding

Low risk

Authors state "Double blindly".

Incomplete outcome data (attrition bias)
None

Low risk

None detected.

Selective reporting (reporting bias)

Low risk

All data appear to have been reported that was sought by the study's aims.

Other bias

High risk

Selection bias (n = 9) ‐ diagnostic purity bias.

Methods

Allocation: randomised

Blindness: none ‐ open‐label

Duration: 1 year

Funding: Theodore and Vada Stanley Foundation Grant

Setting: 3 inpatient units

Participants

Diagnosis: schizophrenia

N = 46

Age: mean ˜34 years, range 24‐44

Sex: 38 male, 8 female

Inclusion:

DSM IV criteria for schizophrenia, a violent episode in the previous year with a score of greater than or equal to 3 on the physical aggression sub‐scale or the MOAS. A family member living with the patient and willing to collaborate with the researchers. Informed consent signed by the patient and the informant.

Exclusion:

Any other axis I disorders, including alcohol and/or drug abuse/dependence, mental retardation.

Interventions

1. Zuclopenthixol (oral) 35 mg/day n = 20

2. Zuclopenthixol (depot) 233 mg every 14 days (˜16.6 mg/day) n = 26

Additional medication: All received biperiden, other psychotropics were permitted.

Outcomes

Useable:

4. Mental State ‐ PANSS positive only

5. Leaving the study Early

7. Behaviour ‐ number of violent episodes

8. Adverse Effects ‐ additional medication

Unable to use:

7. Behaviour ‐ MOAS, baseline data only

8. Adverse effects ‐ UKU ‐ no data provided by group

Notes

Madrid

Inclusion criteria do not match included data and significant selective reporting.

Authors will need to be contacted at next update to obtain clarification.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

A researcher at one centre did the randomisation.

Allocation concealment (selection bias)

Unclear risk

Not stated.

Blinding

High risk

Open‐label.

Incomplete outcome data (attrition bias)
None

High risk

Missing UKU, missing PANSS.

Selective reporting (reporting bias)

High risk

MOAS not reported other than at baseline.

Other bias

High risk

Selection bias ‐ Berkson

Methods

Allocation: randomised.
Blindness: double.
Duration: 10 weeks.
Design: outpatient and Inpatient
Country: UK.

Participants

Diagnosis: acute functional psychosis (schizophrenia n = 68, schizoaffective n = 7, hypomania n = 4, unknown n = 4 and unreported n = 11.
History: acutely ill / exacerbation of chronic illness, 0‐20 episodes.
N = 94.
Age: mean ˜ 36.9 years, range 18‐65. Age of one patient not recorded.
Sex: 48 male, 34 female, 13 unspecified (12 patients not analysed by authors and sex not reported in 1 patient). Lundbeck UK contacted to request information.
Inclusion criteria: BPRS 15 or more.
Exclusion criteria: previous intolerance to neuroleptics, additional serious psychiatric or neurological disorder, serious physical illness, pregnant or lactating women and those of child‐bearing potential who intend to become pregnant, addiction to drugs or alcohol, depot neuroleptic at an unstable dose or dosing frequency.
Setting: hospital and community.

Interventions

1. Zuclopenthixol: dose range 25 mg to 150 mg/day. N = 50.
2. Chlorpromazine: dose range 100 mg to 600 mg/day. N = 44.

Additional medication: amitriptyline, temazepam, procyclidine.

Outcomes

2. Global state: CGI.

Table 1

5. Leaving the study early.

Zuclo n = 32 completed study, Chlor n = 21 completed the study.
8. Adverse effects.

Table 1

Unable to use:
Mental state: BPRS (> 50% loss to follow‐up in one group).
Adverse effects: UKU side effects rating scale (data not reported by group).

Notes

Zuclopenthixol group previous episodes 0‐17. Chlorpromazine group previous episodes 0‐20.

Most common zuclopenthixol dose 75 mg/day.

Most common chlorpromazine dose 600 mg/day

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Authors do not mention. Lundbeck UK contacted to request information regarding potential biases 5th January 2016.

Allocation concealment (selection bias)

Unclear risk

Authors do not mention.

Blinding

Low risk

Clearly stated in the paper as double‐blind design.

Incomplete outcome data (attrition bias)
None

High risk

11 patients were excluded and not analysed. Missing outcome data.

Selective reporting (reporting bias)

Low risk

All data appear to have been reported on.

Other bias

High risk

Attrition bias ‐ per protocol analysis.

Methods

Allocation: randomised.
Blindness: double.
Duration: 12 weeks.
Design: single‐centre.
Country: Canada.

Funding: Public health service research grant ‐ Pfizer

Participants

Diagnosis: schizophrenia.
N = 30.
Age: mean 37 years, range 18‐58.
Sex: 13 male, 17 female.
Inclusion: diagnosis of schizophrenia

Exclusion criteria: not reported.
Setting: hospital.

Interventions

1. Clopenthixol: dose range 50 mg to 200 mg/day. Average 150 mg/day. N = 10.
2. Thiothixene: dose range 10 mg to 40 mg/day. average 30 mg/day. N = 10.
3. Chlorprothixene: dose range 150 mg to 600 mg/day. Average 450 mg/day. N = 10.

Outcomes

Useable:

2. Global state: CGI.

5. Leaving the study early
8. Adverse effects: number of adverse effects in every group, partial TESS

Unable to use:
4. Mental state: BPRS (data not reported by group).
7. Behaviour: NOSIE (no data available).
8. Adverse effects: TESS (no data available).

Notes

15 chronic and 15 new admissions. Original Cochrane authors did not include the two‐week washout period in the duration of the study.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Suggested but not reported by authors. "10 week double blind."

Allocation concealment (selection bias)

Unclear risk

Not mentioned.

Blinding

Low risk

"10 week double blind."

Incomplete outcome data (attrition bias)
None

Low risk

None obvious and none mentioned by authors.

Selective reporting (reporting bias)

High risk

Multiple areas where data not reported.

Several papers produced with same data.

Other bias

High risk

Selection bias ‐ sampling ‐ Berkson

Methods

Allocation: inpatient

Blindness: randomised by specially assigned person

Duraton:: 6 weeks

Participants

Diagnosis: CCMD ‐ 3

N = 120

Age: mean ˜33 years, range 24‐42

Exclusion: severe physical diseases, alcohol and substance dependence

Interventions

1. Clopenthixol 40 mg to 70 mg/day. N = 60

2. Chlorpromazine 300 mg to 550 mg/day. N = 60

Other antipsychotics were not allowed during the treatment.

Benzhexol and diazepam were used if necessary.

Outcomes

Able to use:

4. Mental state: BPRS, PANSS

8. Adverse Effects: EPSEs and excessive sedation

Unable to use:

4. Mentals State: SANS ‐ no useable data

Notes

Paper kindly translated and data extracted by Jun Xia.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Not mentioned. No detailed information was described in the study. No Concerns raised by the data extractor.

Allocation concealment (selection bias)

Low risk

Not mentioned. No detailed information was described in the study. No Concerns raised by the data extractor.

Blinding

Low risk

Randomised. No detailed information was described in the study. No Concerns raised by the data extractor.

Incomplete outcome data (attrition bias)
None

Low risk

All outcomes reported.

Selective reporting (reporting bias)

Low risk

None found.

Other bias

Low risk

None obvious.

Methods

Allocation: randomised.
Blindness: double.
Duration: 13 weeks.
Design: single centre.
Country: USA.

Funding: Not reported.

Participants

Diagnosis: chronic schizophrenia.
History: hospitalised patients with schizophrenia with duration of illness 2 months‐to 3 years.
N = 50.
Age: mean ˜ 43 years, range 25‐59.
Sex: all male.
Inclusion criteria: patients under 60 years of age and free of significant physical illness such as liver disease or severe cardiovascular disease.

Exclusion criteria: not discussed.
Setting: not reported.

Interventions

1. Clopenthixol: dose 132.5 mg/day, range 25 mg to 250 mg/day. N = 25.
2. Perphenazine: dose 51.2 mg/day, range 8 mg to 80 mg/day. N = 25.

Outcomes

Useable:

2. Global state: CGI.

5. Leaving the study early.
8. Adverse effects: requiring additional medication ‐ benztropine methanesulphate; other medications.

Unable to use:
4. Mental state: BPRS (no SD), PIP (no data by group).

8. Adverse Effects ‐ table 3 ‐ week 12 missing

Notes

All but two patients were taking 1 or more psychoactive drugs (mostly phenothiazines) before starting the study.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Not reported, suggested but not clear.

Allocation concealment (selection bias)

Unclear risk

Not reported, suggested but not clear.

Blinding

Low risk

Reasonable methods employed for blinding.

Incomplete outcome data (attrition bias)
None

Low risk

Reasons given by authors.

Selective reporting (reporting bias)

Low risk

Not detected.

Other bias

High risk

Per‐protocol analysis.

Methods

Allocation: Random

Blindness: open‐label

Duration: 13 weeks

Funding: Danish Medical Counscil, H:S Research council, University of Copenhagen Faculty of Humanities

Location: Copenhagen

Participants

Diagnosis: ICD10 F20 Schizophrenia

N = 31

Age: 27.3 years (+/‐5.9), range 19‐37

Sex: not discussed

History: untreated psychosis 4‐78 months

Inclusion: anti‐psychotic naive admitted for treatment and written informed consent

Exclusion: known retardation, need acute medication, compulsorily hospitalised

Interventions

1. Zuclopenthixol 6 mg to 26 mg average 9.6 mg. N = 10

2. Risperidone 2 mg to 7 mg average 3.6 mg. N = 15

3: Healthy controls N = 25

Additional Medication: Benzodiazepiens and anticholinergics but not on the day of examination

Outcomes

Usable:

4. Mental State: PANSS ‐ table 2

5. Leaving the study Early

Not Useable:

6. General Functioning: Cognitive Functions (CANTAB, WCST, verbal fluency, Figural fluency, trail making tests A & B, DART) ‐ no useable data

8. Adverse Effects: ESRS ‐ no data

Notes

Not all data published by authors. When review next updated the authors should be contacted to obtain this information.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Stated but not elaborated on.

Allocation concealment (selection bias)

High risk

Open‐label.

Blinding

High risk

Open‐label.

Incomplete outcome data (attrition bias)
None

Unclear risk

Authors state they will publish other data elsewhere though not found in searches to date.

Selective reporting (reporting bias)

Unclear risk

Authors state they will publish other data elsewhere though not found in searches to date.

Other bias

High risk

Selection bias, Berkson

Methods

Allocation: randomised not clear.
Blindness: double.
Duration: 7 weeks.
Design: multicentre.
Country: Switzerland.

Funding: not stated

Participants

Diagnosis: acute paranoid schizophrenia.
History: moderate to severe hospitalised acute patients with schizophrenia.
N = 723.
Age: not reported.
Sex: not reported.
Inclusion criteria: moderate to severe acute patients with schizophrenia.
Setting: hospital.

Interventions

1. Clopenthixol: dose 100 mg/day. N = 36.
2. Clozapine: dose 200 mg to 300 mg/day. N = 371.
3. Chlorpromazine: dose 200 mg to 350 mg/day. N = 212.
4. Haloperidol: dose 4 mg to 8 mg/day. N = 78.
5. Trifluoperazine: dose 20 mg to 30 mg/day. N = 36.

Outcomes

Useable:

8. Adverse effects: drowsiness, stimulation, confusion, gastrointestinal side‐effects, anticholinergic side‐effects, dizziness, orthostatic reaction, headache, hypersalivation, hypokinesia, hyperkinesia, dyskinesia, rigor, tremor, akathisia.

Unable to use:
2. Global state: number of patients completing the study not provided.
4. Mental state: BPRS (no clear data).

Notes

All of the higher dose ranges are medians.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Not reported. Sandoz UK contacted by online form to clarify bases and seek additional data 5th January 2016.

Allocation concealment (selection bias)

Unclear risk

Not reported

Blinding

Unclear risk

Authors state double‐blind but do not say how.

Incomplete outcome data (attrition bias)
None

Unclear risk

Total figures are not provided.

Selective reporting (reporting bias)

High risk

Sandoz side‐effect check list not reported.

Other bias

High risk

Selection bias, performance bias, unclear if intention‐to‐treat analysis was used.

Methods

Allocation: random

Blindness: not reported

Duration: 12 weeks

Funding: committee on medicine and science ‐ personal grant and an undisclosed non‐restricted grant

Participants

Diagnosis: F20 schizophrenia

n = 45

Age: ˜21‐33 years

Sex: 25 male, 13 female, 7 unknown

Included: drug naive first episode patients with schizophrenia, F20 ICD10 criteria for schizophrenia

Excluded: mental retardation, severe somatic diseases, severe head trauma, pregnancy, presence of MRI contraindications, acute need of medication, compulsory hospitalisation in patients, presence of psychotic illness in controls or their first‐degree relatives.

Interventions

1. Zuclopenthixol (n = 8) 10.3 mg/day average (3.8 mg to 16.8 mg)

2. Risperidone (n = 11) 3.4 mg/day average (1.9 mg to 4.9 mg)

3. Control (n = 20)

Of the original n = 45 it is unclear as to which groups the remaining n = 6 were attributed.

Other medication: Anticholinergic n = 1‐, antidepressant n = 1, benzodiazepines n = 11

Outcomes

Usable:

4. Mental State: PANSS ‐ table 2 (PANSS general, mean and SD baseline and follow‐up).

5. Leaving the study early

8: Adverse effects ‐ ESRS ‐ table 2

Notes

3 patients minimally medicated prior to the start of the study.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Not reported. DRCMR research group contacted 5th January 2016 to clarify these biases and seek any missing information.

Allocation concealment (selection bias)

High risk

Not reported.

Blinding

High risk

Not reported.

Incomplete outcome data (attrition bias)
None

High risk

Conflicting reports from several papers using same population cohort.

Selective reporting (reporting bias)

High risk

Supplementary publications highlight dropout and inconsistencies in the randomisation.

Other bias

High risk

Berkson bias

Methods

Allocation: unclear. "In the investigation patients ... were allocated." Implied when table 1 reviewed.
Blindness: double.
Duration: 4 weeks
Design: RCT
Country: Norway

Participants

Diagnosis: acute and chronic psychoses, mainly schizophrenia
N = 20
Age: 20‐66 years (mean 36)
Sex: 13 male, 7 female
Inclusion criteria: acute psychoses and chronic exacerbations.

Exclusion: below 15 years age, serious somatic disease, pathological laboratory findings, pregnant patients.
Setting: Inpatient

Interventions

1. Cis(Z)‐clopenthixol 48mg/day range 10 mg to 150 mg/day. N = 10.
2. Cis(Z)/trans(E) ‐ clopenthixol 88 mg/day range 10 mg to 200 mg/day. N = 10.

Outcomes

2. 6 Average Endpoint in global state scores

Table 7, CGI

5. Leaving the study early n = 2, (one for refusal due to side‐effects, oculogyric crisis and one for non‐compliance)

Unable to use:
Physiological: serum concentrations of Cis(Z)‐clopenthixol and clopenthixol (non‐clinical data all reported post‐cross‐over).

4. 4 Average endpoint general mental state score

Table 3 and 4 ‐ CGI time 0, 2 weeks and 4 weeks no SD

4. 4 Average change in general mental state scores

Table 5 ‐ BPRS ‐ total score time 0, 2 weeks and 4 weeks. no SD

4. 8 Average change in specific symptom scores

Thinking disturbance: Table 5 ‐ BPRS ‐ time 0, 2 weeks and 4 weeks. no SD

4. 8 Average change in specific symptom scores

Withdrawal ‐ retardation: Table 5 ‐ BPRS ‐ time 0, 2 weeks and 4 weeks. noS.D

4. 8 Average change in specific symptom scores

Hostile‐suspiciousness: Table 5 ‐ BPRS ‐ time 0, 2 weeks and 4 weeks. no SD

4. 8 Average change in specific symptom scores

Anxious‐depression: Table 5 ‐ BPRS ‐ time 0, 2 weeks and 4 weeks. no SD

8. 1 Clinically important general adverse effects

Table 8, individual side‐effects and totals. Frequencies not patient numbers.

Notes

none

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Not reported but implied; "A double blind clinical investigation".

Allocation concealment (selection bias)

Unclear risk

patients were "allocated"

Attempted to contact authors but Dikemark Hospital no longer exists. Google search did not reveal any contact details of the authors.

Blinding

Low risk

Authors state double‐blinding but the details are not clear in the paper.

Incomplete outcome data (attrition bias)
None

Low risk

The two excluded patients had reasons given by the authors.

Selective reporting (reporting bias)

Low risk

Reported all outcomes stated in the study aims.

Other bias

Unclear risk

Berkson and diagnostic purity bias.

Methods

Allocation: patients are allocated to each group and it is unclear if a process of randomisation occurred. A review of table one suggests an element of randomisation ‐ implied randomisation.
Blindness: double.
Duration: 8 weeks.
Design: cross‐over.
Country: Denmark.

Funding: not declared.

Participants

Diagnosis: chronic schizophrenia.
History: illness duration 5‐20 years
N = 57
Age: mean ˜50 years, range 24‐79
Sex: 53 male and 4 female
Inclusion criteria: chronic patients already in neuroleptic treatment, preferably clopenthixol

Exclusion: not mentioned
Setting: inpatient

Interventions

1. Cis(Z)‐clopenthixol hydrochloride, n = 29 average dose 47.6 mg/day
2. Clopenthixol hydrochloride (cis(Z) and trans(E) forms), n = 28 average dose 150 mg/day

Table 2 dosing reported by authors is unclear.

Outcomes

2.6 Global State ‐ Average change in global state scores (Table 1)

4.3 Average endpoint general mental state score

5.2 Leaving the study early (for general reasons) ‐ zero people left the study

6.3 General Functioning ‐ average endpoint general functioning scores (Table 4)

Unable to use: Physiological ‐ serum concentrations of Cis(Z)‐clopenthixol and clopenthixol (non‐clinical data all reported post‐cross‐over).

Notes

none.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Not reported but implied; "A double blind clinical investigation".

Allocation concealment (selection bias)

Unclear risk

Not reported.

Blinding

Low risk

Stated in paper. "The drugs were given double blindly in a randomised way."

Incomplete outcome data (attrition bias)
None

Low risk

No incomplete outcome data.

Selective reporting (reporting bias)

Low risk

None noted.

Other bias

Unclear risk

Berkson and diagnostic purity bias.

Methods

Allocation: "The patients were allocated..." Implied randomisation.
Blindness: double.
Duration: 8 weeks
Design: RCT
Country: Finland

Funding: not stated

Participants

Diagnosis: 40 patients with schizophrenia and 14 patients with oligophrenia.
History: vast majority ill or hospitalised more than 5 years.
N = 54
Age: Mean ˜45 years
Sex: 28 male, 26 female
Inclusion/exclusion: not reported
Setting: not reported.

Interventions

1. Cis(Z)‐clopenthixol 73 mg/day (10 mg to 150 mg/day). N = 26
2. Cis(Z)/Trans(E)‐clopenthixol 114 mg/day (20 mg to 300 mg/day). N = 28

Outcomes

2. 5 Average endpoint global state score

Table 3 and 5: Severity of illness, scale unclear

5. Leaving the study early

Table 3 and 5: n = 5 left, reasons not stated in the article.

8. 3 Average endpoint general adverse effect score

Table 7: Single side‐effects week 8 total

Unable to use:

8. 4 Average change in general adverse effect scores

Table 6: Side‐effects interfering with patient's functioning ‐ CGI. No SD.

8. 8 Average change in specific adverse effects

Table 7: Single side‐effects. No SD.
Physiological: serum concentrations of Cis(Z)‐clopenthixol and clopenthixol (non‐clinical data all reported post‐cross‐over).

Notes

none

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Not reported ‐ Finland Central Hospital contacted by email 5th January 2016.

Allocation concealment (selection bias)

Unclear risk

Not reported ‐ Finland Central Hospital contacted by email 5th January 2016.

Blinding

Unclear risk

Reported but not detailed

Incomplete outcome data (attrition bias)
None

High risk

At week 8, five patients were not included in the outcome data and not accounted for by the authors.

Selective reporting (reporting bias)

Unclear risk

Not enough information to make a clear decision.

Other bias

Low risk

Per protocol analysis, attrition bias

Methods

Allocation: not stated
Blindness: double, implied randomisation
Duration: 12 weeks.
Design: multicentre.
Country: Finland.

Funding: Not stated

Participants

Diagnosis: chronic schizophrenia (n = 58) or other psychotic (n = 5, paranoic state, depressive/PD)
History: n = 40 duration of illness > 10 years and n = 11 for > five years.
N = 63.
Age: mean ˜ 43 years.
Sex: 41 male, 22 female.
Inclusion criteria: chronic schizophrenic inpatients or other psychotic inpatients who might benefit from either of these drugs or already receiving such treatment.
Exclusion criteria: patients under 15 years of age, with serious concomitant somatic illness or pathological laboratory findings, and pregnant patients.
Setting: hospitalised.

Interventions

1. Cis(Z)‐zuclopenthixol: dose 40 mg/day. N = 30.
2. Haloperidol: dose 10 mg/day. N = 33.

Outcomes

Useable:

2. Global state: CGI.

5. Leaving the study early [n = 27 left (13 Zuclo and 14 Halo) ‐ only 13 accounted for: n = 1 discharge, n = 9 insufficient effect and side‐effects, n = 3 reasons unrelated to treatment]
8. Adverse effects: use of other therapeutic drugs. CGI, frequency only.

Unable to use ‐
2. Global State: NOSIE 30 (unable to use data). CGI ‐ percentages only
4. Mental state: BPRS (no SD).

Notes

Previous neuroleptic medication: 14 different neuroleptics.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Not stated. Hospital District of South West Finland contacted to seek further information on biases 5th January 2016.

Allocation concealment (selection bias)

Unclear risk

Not stated.

Blinding

Low risk

Double‐blind.

Incomplete outcome data (attrition bias)
None

High risk

27 patients dropped out and only 12 accounted for by the authors.

Selective reporting (reporting bias)

Unclear risk

Not able to ascertain from paper.

Other bias

High risk

Berkson bias, diagnostic purity bias

Methods

Allocation: randomised.
Blindness: double.
Duration: 6 weeks.
Design: multicentre.
Country: Finland.

Funding: not stated

Participants

Diagnosis: chronic schizophrenia or schizophreniform disorder with an acute exacerbation (DSM‐III‐R).
History: community patients with acute exacerbation of chronic schizophrenia.

N = 98.
Age: mean ˜ 47 years, range 18‐65.
Sex: 47 male, 51 female.

Inclusion criteria: acute exacerbations. 18‐65y, acute psychotic symptoms necessitating antipsychotic treatment. Diagnosis of chronic or sub‐chronic schizophrenia or schizophreniform disorder according to DSM III‐R

Exclusion criteria: patients with clinically significant organic or neurological disorder, serious psychotic disorder other than schizophrenia or schizophreniform disorder, clinically relevant abnormalities in laboratory tests, patients likely to be noncompliant, as well as pregnant or lactating women and those of reproductive age without adequate contraception.

Setting: community.

Interventions

1. Zuclopenthixol: dose 38 mg/day, range 10 mg to 100 mg/day. N = 50.
2. Risperidone: dose 8 mg/day, range 2 mg to 20 mg/day. N = 48.

Authors aimed to avoid anti‐parkinsonism medication and benzodiazepines.

Outcomes

Useable:

5. Leaving the study early (n = 98 to n = 40)

8. Adverse effects: use of antiparkinsonian medication.

8. Adverse effects: UKU side effect rating scale

Unable to use

1. Death: n = 1 patient on risperidone. died 1 month after the trial. This patient was only on the drug for one week and the pathology diagnosis was viral myocarditis. The authors deemed this to not be related to risperidone.

4. Mental state: PANSS score, BPRS (originally included but upon review data is unusable)
Vital signs, body weight, laboratory screening (no data by group).

8. Adverse effects: extrapyramidal symptom rating scale (data not useable)

Notes

Discontinuation of anti‐parkinsonism and other psychotropic medication.

Author contacted to seek clarification of some data: no reply at the time of writing. Information from multiple papers.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Randomised but not stated how. E‐mailed M.O Huttunen Summer 2015 with no response to date.

Allocation concealment (selection bias)

Unclear risk

Randomised but not stated how. E‐mailed M.O Huttunen Summer 2015 with no response to date.

Blinding

Unclear risk

Mentioned but not clear how. E‐mailed M.O Huttunen Summer 2015 with no response to date.

Incomplete outcome data (attrition bias)
None

Low risk

Intention‐to‐treat analysis

Selective reporting (reporting bias)

High risk

Report mean for non‐normal data instead of median. Only reported CGI data on comparison to previous neuroleptic treatment.

Other bias

High risk

Diagnostic purity and attrition biases.

Methods

Allocation: randomised.
Blindness: double.
Duration: 3 weeks.

Funding: Ayerst, McKenna and Harrison LTD

Participants

Diagnosis: schizophrenic reaction (different subtypes) plus one extreme incapacitating anxiety in a schizoid personality.

N = 41

Age: 18‐65 years, mean ˜31 years

Sex: 16 male, 25 female

History:

Duration of illness 0.4 ‐ 5.3 years

Included: acute psychotic symptomatology

Excluded: not mentioned.

Interventions

1. Clopenthixol (N = 20) 122 mg/day (75 mg to 600 mg), n = 5 additional medication

2. Chlorpromazine (N = 21) 435 mg/day (150 mg to 1800 mg), n = 2 additional medication

All patients had previous treatment.

Outcomes

Used:

2. Global state ‐ Global assessment ‐ table 2

4. Mental state ‐ BPRS ‐ table 1

5. Leaving the study early n = 4 (Chlorpromazine, 1 exhausted tablets and 3 due to side‐effects). n = 1 (zuclopenthixol, side‐effects).

8. Adverse effects ‐ table 3 and additional medication n = 7 (authors do not mention which additional medication)

Unable to use:

5. Leaving the study early

"Several patients ... diagnosis changed ... to a non‐psychotic disorder or to a brain syndrome were dropped from the study." No data.

Notes

none.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Implied randomisation.

Response: Patients assigned a number belonging to a bottle of medications.

Research unit at the Royal Victoria Hospital in Quebec contacted for more information on biases and missing patient information 5th January 2016.

Allocation concealment (selection bias)

Low risk

Contents of bottles not known to anyone in the study until the end.

Blinding

Low risk

Study states "double blind".

Incomplete outcome data (attrition bias)
None

High risk

Several patients not included in the study because of a change in diagnosis. The numbers were not given and not included in the analysis.

Selective reporting (reporting bias)

High risk

Missing data, areas not reported but alluded to in discussion.

Other bias

High risk

Berkson bias, attrition bias, per protocol analysis

Methods

Allocation: "Patients were subsequently allocated" ‐ randomisation implied
Blindness: single (patient)
Duration: 6 weeks.
Design: single centre.
Country: Greece.

Participants

Diagnosis: chronic schizophrenia.
History: duration of illness 6 years and over.
N = 54.
Age: mean ˜43 years, range 26‐64.
Sex: 29 male, 25 female.
Inclusion: established diagnosis of schizophrenia in hospital records

Exclusion criteria: not reported.
setting: hospital.

Interventions

1. Clopenthixol: dose 150 mg/day. N = 18.
2. Chlorpromazine: dose 300 mg/day. N = 18.
3. Placebo. N = 18.

Additional Medication: Artane and Phenergan.

Outcomes

Used:

5. Leaving the study early n = 0

8. Adverse effects: number of patients experiencing side‐effects.

(all on clopenthixol, n = 3 restarted menstruation, n = 2 parkinsonism, n = 1 oculogyric crisis

Not used:

7. Behaviour: behavioural scale, Aof V (table III) ‐ no usable data

Notes

none

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Not mentioned. Contacted Dromokaītion Hospital Athens to obtain further information on bias 5th January 2016.

Allocation concealment (selection bias)

Unclear risk

"...patients were subsequently allocated into three groups..." Unclear how.

Blinding

Unclear risk

"He was the only person knowing the identity of the tablets and to which group the patients belonged." Unclear if single‐ or double‐blind.

Incomplete outcome data (attrition bias)
None

Unclear risk

First rating deleted from analysis as was considered to be training by the researchers.

Selective reporting (reporting bias)

High risk

See above

Other bias

High risk

Hospitalised patients used ‐ Berkson bias.

Methods

Allocation: randomised.
Blindness: single.
Duration: 10 weeks.
Design: multicentre.
Country: UK.

Funding: not declared.

Participants

Diagnosis: acute schizophrenia (RDC).
N = 61.
Age: mean ˜ 39 years, range 18‐65.
Sex: 26 male, 35 female.

Inclusion criteria: aged 18‐65 years inclusive, were experiencing an acute schizophrenic episode according to RDC and minimum BPRS score of 15.

Exclusion criteria: concurrent serious physical illness, patients who had previously demonstrated intolerance to neuroleptics, received a depot neuroleptic in the previous 2 weeks, or were suffering from a serious additional psychiatric or neurological disorder, patients who were pregnant or lactating; dependence or addiction to drugs or alcohol.

Setting: hospital.

Interventions

1. Zuclopenthixol dihydrochloride: dose range 25 mg to 150 mg/day. N = 30.
2. Sulpiride: dose 200 mg to 1200 mg/day. N = 31.

Additional medication: temazepam (insomnia), procyclidine (EPSEs), amitriptyline (depression).

Outcomes

usable:

2. Global state: CGI.
4. Mental state: BPRS.

5. Leaving the study Early
8. Adverse effects: change in mean weight.

Notes

Multicentre

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Randomised but not elaborated on by authors. Bolton General Hospital contacted 5th January 2016 to seek clarification from the first author regarding biases.

Allocation concealment (selection bias)

Low risk

Randomly allocated but not elaborated on.

Blinding

Low risk

Clearly stated.

Incomplete outcome data (attrition bias)
None

Low risk

Authors discuss

Selective reporting (reporting bias)

High risk

Not all data was reported in the paper.

Other bias

High risk

Diagnostic purity bias, attrition bias, per‐protocol analysis

Methods

Randomised, double‐blind, 12 weeks

Participants

Acute psychosis

N = 54

Age: 20‐60 years (mean 38)

Sex: 25 female, 15 male, 14 unspecified

Inclusion: hospitalised in the preceding 7 days, acute psychosis, exacerbation of chronic psychosis, neuroleptic treatment was anticipated to be at least 3 weeks, informed consent.

Excluded: mania/depression, serious somatic disease, organic brain damage, pregnant, history of abuse, previous good response to neuroleptic

Interventions

1. Zuclopenthixol 37 mg/day (10 mg to 20 mg) n = 27

2. Perphenazine 30 mg/day (8 mg to 72 mg) n = 27

Additional medication: benzodiazepines, methotrimeprazine, anti‐parkinsonism medication

Outcomes

5. Leaving the study early

Unable to use:

2. Global state: CGI

4. Mental state: CPRS

8 Adverse effects

Notes

Multicentre

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Not stated. Glostrup hospital contacted 5th January 2016 to obtain further information to clarify biases and seek additional outcome data if available.

Allocation concealment (selection bias)

Unclear risk

Not stated.

Blinding

Low risk

Double‐blind stated.

Incomplete outcome data (attrition bias)
None

Low risk

Not detected.

Selective reporting (reporting bias)

High risk

Multiple data not reported and often unclear.

Other bias

High risk

Berkson bias evident.

Methods

Allocation: randomised.
Blindness: double.
Duration: 24 weeks.
Design: single centre.
Country: USA.

Funding: United states Public health service grant

Participants

Diagnosis: chronic schizophrenia.
History: duration of illness 2 years or longer.
N = 57.
Age: mean ˜ 42 years, range 21‐61 years.
Sex: 25 male, 32 female.
Inclusion criteria: no evidence of organic disease, duration of illness 2 years or longer, chronic schizophrenia, permission from a family member

Exclusion: no complicating organic illness or known brain damage
Setting: hospital.

Interventions

1. Clopenthixol: dose 205 mg/day. N = 15.
2. Haloperidol: dose 12.3 mg/day. N = 14.
3. Chlorpromazine: dose 830 mg/day. N = 14.
4. Placebo. N = 13.

Additional medication: antiparkinsonian medication, sedative.

Outcomes

Useable:

5. Leaving the study early: n = 4 (behavioural deterioration n = 2, 1 Chlorpromazine and 1 placebo, Intestinal obstruction secondary faecal impaction n = 2, both chlorpromazine)
8. Adverse effects ‐ table 10 and 11

Unable to use ‐
2. Global state: CGI (no SD)

4. Mental state: BPRS (no SD).

Short‐term data not reported (first 12 weeks medication free)

Notes

Originally reported as a 12‐week study in previous review.

Combination of several papers (6 in total) all reporting on the same cohort.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

"In a double blind placebo controlled trial..."

Implied but not clear.

Allocation concealment (selection bias)

High risk

Not specifically discussed by the authors but "subjects were housed together on a special two‐wing ward, males in one wing and females in the other" so it is unlikely that allocation concealment was maintained.

Blinding

Low risk

Authors state double‐blind. Not discussed how.

Incomplete outcome data (attrition bias)
None

Unclear risk

Incomplete data described by the authors but not used in outcome data.

Selective reporting (reporting bias)

High risk

Multiple papers, same patient cohort.

Other bias

High risk

Selection bias (Berkson), Suspected data mining, likely per‐protocol analysis.

Methods

Allocation: Quote” Patients were randomly divided into two groups ”(p.4)

Blinding: Double‐blind

Duration: Quote”8 weeks”(p. 4)
Location: Quote ”Inpatients, WuHan, China” (p.4)

Participants

Schizophrenia, CCMD‐3;
Sex: male 31 female 29
Age：mean˜34.1 years, SD ˜12.5
Total N = 64

Inclusion and exclusion criteria:
Patients with schizophrenia fulfilled CCMD‐3 were included;
Patients with severe heart, liver, kidney diseases were excluded.

Interventions

1. Zuclopenthixol Group: (n = 33)
Management: newly hospitalised patients had to stop the previous drugs for three days and then were randomly divided into Zuclopenthixol Group or Chlorpromazine Group.
Chlorpromazine or Zuclopenthixol was enclosed in the same colour, style capsule by pharmacist.
In the first week of treatment, the dosage of Zuclopenthixol was 20 mg/day and from the second week, it was increased 10 mg/day every week.
No other antipsychotic drugs were allowed during the treatment.
Trihexyphenidyl or scopolamine was used when necessary.
MECT was used two times per week for 12 times.

2. Chlorpromazine Group: (n = 27)
management:
Newly hospitalised patients had to stop the previous drugs for three days and then were randomly divided into Zuclopenthixol Group or Chlorpromazine Group.
Chlorpromazine or Zuclopenthixol was enclosed in the same colour, style capsule by pharmacist.
In the first week of treatment, the dosage of Chlorpromazine was 200 mg/day and from the second week, it was increased 50 mg/day every week.

No other antipsychotic drugs were allowed during the treatment;

Trihexyphenidyl or scopolamine was used when necessary;

Outcomes

2. Global State ‐ Clinical response
4. Mental State ‐ BPRS score
8. Adverse effects ‐ Adverse events, TESS;

Unable to use: electrocardiogram; electroencephalogram; liver function; blood routine examination.

Notes

We have noticed that the N number does not always add up in this study, for example, the total sample size randomised was claimed to be 64, and 61 completed, but the number randomised to each group was reported as 33 and 27. Author to be contacted at next update.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Authors state patients randomised, not how

Allocation concealment (selection bias)

Unclear risk

Not reported after translation and data extraction

Blinding

Low risk

Double‐blind

Incomplete outcome data (attrition bias)
None

Unclear risk

N number unclear

Selective reporting (reporting bias)

Unclear risk

Not reported after translation and data extraction

Other bias

High risk

Berkson bias