Penfluridol for schizophrenia

Abstract

Background

Penfluridol, available since 1970, is an unusual long acting oral antipsychotic agent for the treatment of schizophrenia. It may be considered a depot medication as it is administered once a week.

Objectives

To review the effects of penfluridol for treatment of those with schizophrenia and schizophrenia‐like illnesses in comparison to placebo, other antipsychotic medication or no intervention.

Search methods

We undertook electronic searches of the Cochrane Schizophrenia Group's Register (2005), the Cochrane Central Register of Controlled Trials (2003‐5) and LILACS (1982‐2005). We hand searched references of all identified studies and sought citations of these studies in the Science Citation Index. We contacted the authors of trials and the manufacturer of penfluridol.

We updated this search September 2012 and added eight new trials to the awaiting classification section.

Selection criteria

We reliably selected all randomised clinical trials comparing penfluridol to placebo or typical or atypical antipsychotic drugs for schizophrenia or serious mental illness.

Data collection and analysis

We independently extracted and analysed data on an intention‐to‐treat basis. We calculated the relative risk (RR) and 95% confidence intervals (CI) of homogeneous dichotomous data using a random effects model, and where possible calculated the number needed to treat. We calculated weighted mean differences (WMD) for continuous data.

Main results

We included twenty‐five studies with a total of 1024 participants. Most of these studies were undertaken in the 1970s when penfluridol was launched. Ten studies, with 365 patients, compared penfluridol to placebo. In the meta‐analysis of medium‐term lasting studies, penfluridol was superior to placebo in the main efficacy measures: 'improvement in global state' (n=159, 4 RCTs, RR 0.69 CI 0.6 to 0.8, NNT 3 CI 2 to 10) and 'needing additional antipsychotic' (n=138, 5 RCTs, RR 0.43 CI 0.2 to 0.8, NNT 3 CI 1.8 to 20).

A total of 449 patients from eleven studies were randomised to penfluridol or oral typical antipsychotics. There were no particular differences between penfluridol versus chlorpromazine, fluphenazine, trifluoperazine, thioridazine, or thiothixene for the main outcome measures in medium‐term trials: 'improvement on global state' (N=2 studies), 'leaving the study early' (N=6), 'needing additional antipsychotic' (N=3), needing antiparkinsonian medication (N=2), and side‐effects.

Six studies, with 274 patients, compared penfluridol to depot typical antipsychotics. In general, for the efficacy and safety measures, no differences were established, but penfluridol was superior in keeping the patients in treatment; 'leaving the study early' (n=218, 5RCTs, RR 0.55 CI 0.3 to 0.97, NNT 6 CI 3.4 to 50).

Authors' conclusions

Although there are shortcomings and gaps in the data, there appears to be enough overall consistency for different outcomes. The efficacy and adverse effects profile of penfluridol are similar to other typical antipsychotics; both oral and depot. Furthermore, penfluridol is shown to be an adequate treatment option for people with schizophrenia, especially those who do not respond to oral medication on a daily basis and do not adapt well to depot drugs. One of the results favouring penfluridol was a lower drop out rate in medium term when compared to depot medications. It is also an option for chronic sufferers of schizophrenia with residual psychotic symptoms who nevertheless need continuous use of antipsychotic medication. An additional benefit of penfluridol is that it is a low‐cost intervention.

Note: the eight citations in the awaiting classification section of the review may alter the conclusions of the review once assessed.

Plain language summary

Penfluridol for schizophrenia

Synopsis pending.

Background

Schizophrenia is a serious, disabling, enduring and relapsing mental illness, the onset of which frequently occurs relatively early in life. It can result in a person experiencing problems with the ability to think, feel and perceive things clearly (Carpenter 1994). The worldwide lifetime prevalence of the disorder is about 1% (Torrey 1987, Almeida‐Filho 1997).

One cause of relapse and readmission to hospital is poor compliance with antipsychotic medication often due to its adverse effects. Schizophrenia may also affect a person's insight, interfering with their ability to appreciate the benefit of taking medication long term. The relapse rate is significantly higher in those who have discontinued antipsychotic medication (Gallant 1974). Penfluridol is an oral antipsychotic which is reputed to give antipsychotic protection for one week. This could prove a favourable option for treating those who do not wish to take daily or depot medication (Gallant 1974).

Technical background 
 It has been said that Penfluridol is the first truly long‐acting oral antipsychotic agent for the treatment of acute psychosis (Van Praag 1971), chronic schizophrenia (Ionescu 1983) and Tourette's syndrome (Shapiro 1983). First synthesised in 1968, it is a white micro‐crystalline tertiary amine, only slightly soluble in water and belongs to the diphenylbutylpiperidines; pimozide also belongs to this group (Janssen 1970). Penfluridol is absorbed from the gastrointestinal tract and deposited in fatty tissue from which it is slowly released, resulting in a prolonged duration of action with a half‐life of 70 hours. It is primarily excreted unchanged in the faeces; however, some hepatic metabolism and enterohepatic recycling occur ( Janssen 1972, Gerlach 1975). Penfluridol is characterised pharmacologically by the traditional neuroleptic properties: cataleptogenic effect, antagonism against apomorphine and amphetamine induced stereotypes and inhibition of controlled behaviour. It also has a relatively specific antidopaminergic effect via the blocking of the dopaminergic receptor membrane (Janssen 1970, Janssen 1972, Gerlach 1975).

Objectives

To examine the effects of penfluridol for schizophrenia and schizophrenia‐like illnesses in comparison to placebo, other antipsychotic medication or no intervention.

Methods

Criteria for considering studies for this review

Types of studies

We included all relevant randomised controlled trials. Where a trial was described as 'double‐blind', but it was implied that the study was randomised, we included the trial in a sensitivity analysis. If there was no substantive difference within primary outcomes (see 'types of outcome measures') when these 'implied randomisation' studies were added, then we included these in the final analysis. If there was a substantive difference, we only analysed clearly randomised trials and described the results of the sensitivity analysis in the text. We excluded quasi‐randomised studies, such as those allocating by using alternate days of the week.

Types of participants

We included people with schizophrenia and non‐affective serious/chronic mental illness, irrespective of criteria of diagnosis, sex, and length of illness.

Types of interventions

1. Penfluridol: any dose taken orally. 
 2. Placebo (or no intervention). 
 3. Typical antipsychotic drugs: any dose and mode or pattern of administration. Examples of such drugs are chlorpromazine and haloperidol. 
 4. Atypical antipsychotic drugs: any dose and mode or pattern of administration. Examples of such drugs are clozapine, olanzapine, quetiapine and risperidone.

Types of outcome measures

As schizophrenia is often a life‐long illness and penfluridol is used as an ongoing treatment, outcomes were grouped according to time periods: short‐term (less than three months), medium‐term (3‐12 months) and long‐term (more than one year). It is assumed that long‐term outcomes will be particularly relevant for this review given the properties of penfluridol.

Primary outcomes

1. Death

1.1 Suicide 
 1.2 Natural causes

2. Service utilisation outcomes 
 2.1 Hospital admission

3. Global outcomes 
 3.1 Clinically significant response in global state ‐ as defined by each of the studies

4. Mental state 
 4.1 Clinically significant response in mental state ‐ as defined by each of the studies

5. Behaviour 
 5.1 Leaving the study early

Secondary outcomes

1. Service utilisation outcomes 
 1.1 Days in hospital

2. Global outcomes 
 2.1 Average score/change in global state

3. Mental state 
 3.1 Average score/change in mental state 
 3.2 Clinically significant response on negative symptoms ‐ as defined by each of the studies 
 3.3 Average score/change in negative symptoms 
 3.4 Relapse as defined in the study

4. Behaviour 
 4.1 Clinically significant response in behaviour ‐ as defined by each of the studies 
 4.2 Average score/change in behaviour

5. Extrapyramidal side effects 
 5.1 Incidence of use of antiparkinson drugs 
 5.2 Clinically significant extrapyramidal side effects‐ as defined by each of the studies 
 5.3 Average score/change in extrapyramidal side effects

6. Other adverse effects, general and specific 
 6.1 Number of people dropping out due to adverse effects 
 6.2 Cardiac effects 
 6.3 Anticholinergic effects 
 6.4 Antihistamine effects 
 6.5 Prolactin related symptoms

7. Social functioning 
 7.1 Clinically significant response in social functioning ‐ as defined by each of the studies 
 7.2 Average score/change in social functioning

8. Economic outcomes

9. Quality of life/satisfaction with care for either recipients of care or carers 
 9.1 Significant change in quality of life/satisfaction ‐ as defined by each of the studies 
 9.2 Average score/change in quality of life/satisfaction 
 9.3 Employment status

10. Cognitive functioning

Search methods for identification of studies

Electronic searches

1. We identified relevant randomised trials by searching the following electronic databases:

1.1. We searched the Cochrane Schizophrenia Group Trials Register (November 2005 and September 2000) using the phrase: [penfluridol* or micefal* or semap* or flupidol* in REFERENCE and penfluridol* or micefal* or semap* or flupidol* in STUDY)]

This register is compiled by systematic searches of major databases, hand searches and conference proceedings (see Group Module).

1.2. We searched LILACS (January 1982 to October 2005) using the Cochrane Schizophrenia Group's phrase for randomised controlled trials and for schizophrenia (see Group search strategy) combined with: [AND (penfluridol or semap or penflu)]

1.3. We searched Cochrane Central Register of Controlled Trials (2003‐5) using the phrase: [penfluridol or micefal or semap or flupidol] associated to the Cochrane Schizophrenia Group's phrase for schizophrenia.

1.4 Cochrane Schizophrenia Group Trials Register (September 2012)

The Trials Search Co‐ordinator searched the Cochrane Schizophrenia Group’s Trials Register

1.1 Title, Abstract and indexing terms search

These fields will be searched using the phrase:

*penfluridol* or *micefal* or *semap* or *flupidol* or *cyperon*

1.2 Intervention search

The ‘health care condition’ field will be searched using the phrase:

*penfluridol* or *micefal* or *semap* or *flupidol* or *cyperon*

The Cochrane Schizophrenia Group’s Trials Register is compiled by systematic searches of major databases, handsearches of journals and conference proceedings (see Group Module). Incoming trials are assigned to relevant existing or new review titles.

Searching other resources

1. Reference searching 
 We inspected the references of all identified studies for more trials. SCISEARCH (Science Citation Index (1974 to September 2000)) was used to trace papers that had cited included trials. We inspected these reports in order to identify further studies.

2. Personal contact 
 We contacted the first author of each included study for information regarding unpublished trials.

3. Pharmaceutical companies 
 We contacted the manufacturer of penfluridol (Jansen‐Cilag Ltd) for additional data.

Data collection and analysis

1. Selection of trials 
 We inspected all reports (BGOS and CJSR) and MSL re‐inspected a 10% sub‐sample in order to ensure reliable selection. Where agreement could not be reached, we acquired the full report for more detailed scrutiny. Once the full articles were obtained we (BGOS and CJSR) decided whether they met the review criteria. MSL was sent 10% of all acquired articles to help promote reliability at this stage. Again, if disagreement could not be resolved from published information, we added the trial to those awaiting assessment and contacted the authors for further clarification.

2. Assessment of methodological quality 
 We assessed the methodological quality of trials included in this review using the criteria described in the Cochrane Handbook (Higgins 2005) which is based on evidence of a strong relationship between allocation concealment and the potential for bias in the results. The categories are defined below:

A. Low risk of bias (adequate allocation concealment) 
 B. Moderate risk of bias (some doubt about the results) 
 C. High risk of bias (inadequate allocation concealment)

For the purpose of the analysis in this review, we included trials if they met the criteria A or B.

3. Data collection 
 We (CJSR with BGOS) independently extracted data from selected trials and MSL re‐extracted information from a sample of 20%. When disputes arose, resolution was attempted by discussion. If this was not possible and further information was necessary to resolve the dilemma, we did not enter the data.

4. Data synthesis 
 4.1 Incomplete data 
 With the exception of the outcome of leaving the study early, we did not include trial outcomes if more than 50% of people were not reported in the final analysis.

4.2 Dichotomous data 
 We used an intention to treat analysis, on the condition that more than 50% of people completed the study and everyone allocated to the intervention was counted regardless of whether they completed the follow up. We assumed that those leaving the study early had the negative outcome, with the exception of death. We calculated the Relative risk (RR) and their 95% confidence interval (CI) based on the random effects model. This takes into account any differences between studies even if there is no statistically significant heterogeneity. We inspected the data to see if analysis using a fixed effects model made any substantive difference. Where possible we estimated the number needed to treat (NNT).

4.3 Continuous data 
 4.3.1 Normal data: data on continuous outcomes are frequently skewed, the mean not being the centre of the distribution. The statistics for meta‐analysis are thought to be able to cope with some skew, but were formulated for parametric data. To avoid this potential pitfall, we applied the following standards to all data before inclusion: (a) standard deviations and means were reported or obtained from authors, (b) for data with finite limits, such as endpoint scale data, the standard deviation (SD), when multiplied by two, was less than the mean. Otherwise the mean is unlikely to be an appropriate measure of the centre of the distribution (Altman 1996). We reported data that did not meet the first or second standard in 'other data' tables.

For change data (endpoint minus baseline), the situation is even more problematic. In the absence of individual patient data it is impossible to know if data are skewed, though this is likely to be the case. As carried out in other CSG reviews, we presented change data in MetaView in order to summarise available information. In doing this, it was assumed either that data were not skewed or that the analyses could cope with the unknown degree of skewness. Without individual patient data it is impossible to test this assumption. Where both change and endpoint data were available for the same outcome category, we only presented endpoint data. We acknowledge that by doing this much of the published change data were excluded, but argue that endpoint data is more clinically relevant and that if change data were to be presented along with endpoint data it would be given undeserved equal prominence. Authors of studies reporting only change data are being contacted for endpoint figures. We reported non‐normally distributed data in the 'Other data types' tables.

4.3.2 Rating scales: a wide range of instruments are available to measure mental health outcomes. These instruments vary in quality and many are not valid, or even ad hoc. For outcome instruments some minimum standards have to be set. These were that: (a) the psychometric properties of the instrument should have been described in a peer‐reviewed journal, (b) the instrument should either be: (i) a self‐report, or (ii) completed by an independent rater or relative (not the therapist) and (c) the instrument should be a global assessment of an area of functioning (Marshall 2000).

4.3.3 Cluster trials 
 Studies increasingly employ 'cluster randomisation' (such as randomisation by clinician or practice) but analysis and pooling of clustered data poses problems. Firstly, authors often fail to account for intra class correlation in clustered studies, leading to a 'unit of analysis' error (Divine 1992) whereby p values are spuriously low, confidence intervals unduly narrow and statistical significance overestimated. This causes type I errors (Bland 1997, Gulliford 1999).

Where clustering was not accounted for in primary studies, we presented the data in a table, with a (*) symbol to indicate the presence of a probable unit of analysis error. In subsequent versions of this review we will seek to contact first authors of studies to obtain intra‐class correlation co‐efficients of their clustered data and to adjust for this using accepted methods (Gulliford 1999). Where clustering has been incorporated into the analysis of primary studies, we will also present these data as if from a non‐cluster randomised study, but adjusted for the clustering effect.

We have sought statistical advice and have been advised that the binary data as presented in a report should be divided by a 'design effect'. This is calculated using the mean number of participants per cluster (m) and the intraclass correlation co‐efficient (ICC) [Design effect = 1+(m‐1)*ICC] (Donner 2002). If the ICC was not reported it was assumed to be 0.1 (Ukoumunne 1999).

4.4 Sensitivity analyses 
 4.4.1 Dose: results for high doses (however 'high' was defined in the study), or >60 mg/day if not defined, were compared to those for lower doses. 
 4.4.2 Treatment resistance: results from those whose illnesses were described as 'treatment resistant' were compared to those whose illnesses were not designated as such. 
 4.4.3 First episode: results from those experiencing their first episode of schizophrenia were compared to those with illness of longer duration. 
 4.4.4 The assumption of negative outcome: the primary outcome analysis employing the assumption of negative outcome for those who left early was compared to an analysis where complete data only were analysed.

5. Heterogeneity 
 Firstly, we undertook consideration of all the included studies within any comparison to estimate clinical heterogeneity. Then we used visual inspection of graphs to investigate the possibility of statistical heterogeneity. This was supplemented employing, primarily, the I‐squared statistic. This provides an estimate of the percentage of inconsistency thought to be due to chance. Where the I‐squared estimate was greater than or equal to 75% this was interpreted as evidence of high levels of heterogeneity (Higgins 2003). We did not summate data with 75% or greater I‐squared statistic, but we presented these separately and investigated reasons for heterogeneity.

6. Addressing publication bias 
 Data from all identified and selected trials were entered into a funnel graph (trial effect versus trial size) in an attempt to investigate the likelihood of overt publication bias.

Results

Description of studies

1. Excluded studies 
 We excluded 25 studies. Eleven were case series including only penfluridol (Bobon 1970, Floru 1973, Glatzel 1973, Hasse 1971, Kingstone 1977, Morales 1981, Richards 1976, Sharma 1988, Shopsin 1977, Vaidyalingam 1990, Woggon 1978); twelve were not described as randomised, did not have information on allocation procedure, or did not present usable clinical data (Abuzzahab 1977, Angst 1975, Bao 1988, Baro 1972, Chouinard 1977, Cooper 1975, Eklund 1976, Floru 1975, Gallant 1975, Nistico 1974, Tanghe 1972, Van Lommel 1974).Two randomised controlled trials (Li 1987, Linden 1973) did not present usable clinical outcomes.

2. Awaiting assessment 
 Twelve studies are awaiting assessment (Baro 1972a, Duval 2006, Itoh 1976, Jiang 1995, Jus 1976, Kondo 1975, Shu 1983, Strous 2007, Xun 2006, ปรีชา 2533, 孙轻骑 2007)

3. Ongoing 
 We are not aware of any ongoing studies.

4. Included Studies 
 We included 27 studies in this review. It is peculiar that 82% (22 studies) were published in the 70s, when the drug was launched, and the remaining five studies were published in the 1980s. Most trial centres were based in North America (seven in USA, and five in Canada), nine studies in Europe (four in Netherlands, two in Sweden, one in Romania, one in Belgium, and one in Denmark), three in India, one in Israel, one in New Zealand and one in Australia.

4.1 Lengths of Trials 
 Duration of studies varied widely, from 3 to 52 weeks. Five trials (Chouinard 1976, D'Elia 1977, Jacobsson 1976, Munitz 1986, Van Praag 1971) were classified as short term studies (less than three months). The remaining twenty two were medium term trials (three months to one year), out of which nine lasted exactly three months. There were no long term trials (more than one year).

4.2 Participants 
 A total of 1024 people participated in the 27 trials with sample sizes ranging from 15 to 60 (mean 37.9). Fifty‐two percent of studies had 40 or fewer participants. Standardised diagnostic criteria were described in only nine trials (DSM‐II, DSM‐III, RDC) though it is possible that people in the other trials had a clinical diagnosis of schizophrenia. Many studies (n=19) involved only people with chronic illness. The remaining studies included acutely ill and first episode participants; one in particular specified people with acute psychosis (Van Praag 1971). Only one study (Zwanikken 1973) included participants who had schizophrenia (n=17) and mental retardation (n=20). From the 21 studies which provided information on gender distribution there were 495 males and 269 females. Ages ranged from 17 to 70 years (mean from 18 studies was about 39.2 years). None of the studies focused on a specific age group.

4.3 Setting 
 Seventeen studies had inpatient setting (Baro 1970, Charalampous 1977, Chouinard 1976, D'Elia 1977, Deutsch 1976, Gallant 1974, Jacobsson 1976, James 1977, Jus 1974, Lapierre 1978, Munitz 1986, Ota 1974, Quitkin 1977, Roelofs 1974, Vandecasteele 1974, Van Praag 1971 and Zwanikken 1973). Four included only those who were not in hospital (Claghorn 1979, Donlon 1978, Iqbal 1978 and Kurland 1975). Two studies included inpatients and outpatients (Gerlach 1975, Ionescu 1983). The setting was unclear in four of the included trials (Channabasavanna 1987, Doongaji 1988, Khorana 1988, Wang 1982).

4.4 Interventions 
 The mean dose of penfluridol, based on 13 studies which reported it was about 39 mg/week and the range, taken from 18 studies, was 5 to 160 mg/week. In Munitz 1986 participants allocated to placebo received one dose of penfluridol 20 mg on the first day post‐randomisation. This study was included in order to increase the applicability of the data (such inclusion did not change the overall findings). Ten studies compared penfluridol to placebo (Baro 1970, Channabasavanna 1987, Deutsch 1976, Gallant 1974, Kurland 1975, Munitz 1986, Roelofs 1974, Vandecasteele 1974, Van Praag 1971, Zwanikken 1973). Ten studies (Charalampous 1977, Chouinard 1976, Claghorn 1979, D'Elia 1977, Donlon 1978, Jacobsson 1976, Jus 1974, Khorana 1988, Lapierre 1978, Wang 1982) compared penfluridol with typical oral antipsychotics such as fluphenazine or chlorpromazine. Six studies (Doongaji 1988, Gerlach 1975, Ionescu 1983, Iqbal 1978, James 1977, Quitkin 1977) compared penfluridol with depot antipsychotics such as fluphenazine decanoate or flupenthixol decanoate, and one study (Ota 1974) compared penfluridol to the previously used antipsychotics (without naming them). None of the trials compared penfluridol with an atypical antipsychotic.

4.5 Outcomes 
 Most outcomes were reported as dichotomous (yes‐no/ binary outcomes) and were presented as such. Most scale derived data were unusable as they were usually presented as a graph or as mean without standard deviation or standard error. The main efficacy outcome, described as 'no marked improvement', was derived from Clinical Global Impression (Guy 1976), dichotomised in 'no improvement or worse'. Two studies (Ionescu 1983 and Lapierre 1978) presented data from mental state as BPRS, with mean and standard deviation. Included studies did not present any data on economic outcomes.

4.5.1 Scales presented in analysis 
 4.5.1.1. Global state 
 4.5.1.1.1 Clinical Global Impression ‐ CGI (Guy 1976) 
 A rating instrument commonly used in studies on schizophrenia that enables clinicians to quantify severity of illness and overall clinical improvement during therapy. A seven point scoring system is usually used with low scores indicating decreased severity and/or greater recovery. Usable data from this scale was reported by Charalampous 1977, Deutsch 1976, Donlon 1978, Gallant 1974, Gerlach 1975, Ionescu 1983 and Kurland 1975.

4.5.2. Mental state 
 4.5.2.1. Brief Psychiatric Rating Scale ‐ BPRS (Overall 1961) 
 A brief rating scale used to assess the severity of a range of psychiatric symptoms, including psychotic symptoms. The original scale has 16 items, but a revised 18‐item scale is commonly used. Each item is defined on a seven‐point scale varying from 'not present' to 'extremely severe', scoring from 0‐6 or 1‐7. Scores can range from 0‐126, with high scores indicating more severe symptoms. Usable data from this scale was reported by Ionescu 1983 and Lapierre 1978.

Risk of bias in included studies

1. Randomisation 
 Two studies (Munitz 1986, Lapierre 1978) described adequate methods of allocation concealment. In Munitz 1986, participants were recruited in the emergency unit by a physician who did not take part in the study. Nursing staff recorded the random numbers and administered the medication. The treatment physician was blind to the assignment. Lapierre 1978 gave a detailed description of the allocation procedure that could be summarised as: medication packaged and numbered with random numbers, sealed bottles and sealed envelopes. In reply to our request, Dr Simpson (Iqbal 1978), confirmed that there was allocation concealment, but did not give further information. These three studies were given the quality score 'A'. The remaining 24 studies were classified as 'B', regarding the allocation concealment. Nineteen studies were described as randomised; two studies (Roelofs 1974 and Vandecasteele 1974) had the allocation done by an 'unbiased statistician' without further details. Three studies (D'Elia 1977, Ionescu 1983, Jacobsson 1976) with unclear allocation procedure were included after a sensitivity analysis (there were no substantive differences within primary outcomes). 
 
 2. Blinding 
 All included trials used a double blind design, except for D'Elia 1977, Ionescu 1983, Ota 1974, which were open studies. Seven studies (Channabasavanna 1987, Claghorn 1979, Donlon 1978, Doongaji 1988, Jus 1974, Khorana 1988, Wang 1982) also described forms of blinding, such as use of identical capsules.

3. Leaving the study early 
 Eighteen studies reported participants leaving the study early. Reasons for leaving early were only described by Iqbal 1978.

Effects of interventions

1. The search 
 The electronic databases search identified 54 references, reference cross‐checking provided ten citations, contact with a pharmaceutical company (Jansen) provided a further six, and contacting authors of included studies has not yielded further studies. Twenty seven studies are included in this review, 25 are described as excluded, and four are awaiting assessment.

2. COMPARISON 1. PENFLURIDOL versus PLACEBO 
 Ten studies allocated 305 people to this comparison.

2.1 Global state: 1. no marked improvement (CGI) 
 Four medium‐term studies presented results for change in global state (Deutsch 1976, Gallant 1974, Kurland 1975, Zwanikken 1973), and pooled data favoured penfluridol (n=159, RR 0.69 CI 0.6 to 0.8, NNT 4 CI 3 to 8).

2.2 Global state: 2. needing additional antipsychotic 
 Subjects taking penfluridol needed less additional antipsychotics than those taking placebo, either in one short‐term study (Van Praag 1971, n=34, RR 0.45 CI 0.2 to 0.9, NNT 2 CI 1 to 7), or with the pooled data from five medium‐term studies (Channabasavanna 1987, Kurland 1975, Roelofs 1974, Vandecasteele 1974, Zwanikken 1973, n=138, RR 0.43 CI 0.2 to 0.8, NNT 3 CI 2 to 20).

2.3 Leaving the study early 
 One short‐term (Van Praag 1971) presented result for this data, with no differences between interventions. Three medium‐term studies evaluated this outcome. Baro 1970 and Kurland 1975 favoured penfluridol, and Channabasavanna 1987 did not show differences between penfluridol and placebo. Pooled data of these three studies favoured penfluridol (n=91, RR 0.13, CI 0.02 to 0.9).

2.4. Adverse events: 1. Extrapyramidal 
 Extrapyramidal adverse effects were reported in only two studies, and the meta‐analysis of these trials was equivocal. Medium term data from Gallant 1974 were significant (n=50, RR 5.33 CI 1.8 to 16.1), while Munitz 1986 short term data were found to be equivocal (n=20, RR of 0.50 CI 0.1 to 4.7).

2.5. Adverse events: 2. Other side effects 
 Two medium‐term studies (Channabasavanna 1987, Vandecasteele 1974) evaluated the need for antiparkinson medication and reported no difference between penfluridol and placebo (n=51, RR 1.04 CI 0.6 to 1.9). Gallant 1974 referred a higher incidence of ECG alterations in those patients treated with penfluridol, but this difference was not statistically significant (n=50, RR 3.0 CI 0.7 to 13.5). Vandecasteele 1974 reported equivocal results (n=21, RR of 2.73 CI 0.3 to 22.2) for insomnia when comparing penfluridol with placebo.

3. COMPARISON 2. PENFLURIDOL versus TYPICAL (ORAL) ANTIPSYCHOTIC 
 A total of 449 patients from 11 studies were randomised to this comparison.

3.1 Global state: 1. no marked improvement (CGI) 
 Two medium‐term studies reported changes in global state (Charalampous 1977, Donlon 1978). They found no statistically significant difference between interventions (n=80, RR 0.92 CI 0.7 to 1.2).

3.2 Global state: 2. needing additional antipsychotic 
 Two short‐term studies (Chouinard 1976, Jacobsson 1976) reported data on this outcome, with no statistically significant difference between penfluridol and other typical antipsychotics (n=90, RR 1.35 CI 0.9 to 2.0). Regarding the need for additional antipsychotic medium‐term studies, again, no statistical difference was observed (n=105, RR 1.08 CI 0.7 to 1.8). The studies were Donlon 1978, Jus 1974 and Wang 1982.

3.3 Mental state: BPRS endpoint score 
 Only one study supplied continuous data which could be re‐analysed for this comparison. We contacted study authors by letter, but only Lapierre 1978 has supplied data from the BPRS scale with means and standard deviations from this one‐year study. No statistically significant differences between penfluridol and fluphenazine were observed in endpoint scores (n=41, WMD 1.24 CI ‐4.4 to 6.9).

3.4 Leaving the study early 
 Three short‐term studies (Chouinard 1976, Jacobsson 1976, D'Elia 1977) reported data on this outcome, with no statistically significant difference between penfluridol and other typical antipsychotics (n=118, RR 0.46 CI 0.2 to 1.4). Six medium‐term studies presented data regarding this comparison without specifying the reason for leaving the study early (Charalampous 1977, Claghorn 1979, Donlon 1978, Lapierre 1978, Ota 1974, Wang 1982). Pooled data from these studies showed that people were no more likely to leave the study early with penfluridol or with oral typical antipsychotic (n=235, RR 1.16 CI 0.8 to 1.7). 
 
 3.5 Adverse events: 1. needing antiparkinsonian medication 
 There were no differences between penfluridol and oral typical antipsychotics in the additional use of antiparkinsonian drugs, neither in two short‐term studies (Chouinard 1976, Jacobsson 1976, n=90, RR 1.09 CI 0.6 to 2.0), nor in three medium‐term studies (Charalampous 1977, Khorana 1988, Lapierre 1978, Wang 1982, n=169, RR 1.20 CI 0.9 to 1.6).

3.6 Adverse events: 2. Other side effects 
 There were no differences favouring penfluridol either in short or medium term treatment for the side effects: insomnia, akathisia, acute dystonia, lack of facial expression, drowsiness, nausea, fatigue, tremors, postural hypotension, rigidity and muscle span.

4. COMPARISON 3. PENFLURIDOL versus TYPICAL (DEPOT) ANTIPSYCHOTIC 
 A total of 274 patients from six studies were randomised to this comparison.

4.1 Global state: 1. no marked improvement (CGI) 
 Only one medium‐term study (Ionescu 1983) reported data on this outcome, with no statistically significant difference between penfluridol and fluphenazine decanoate (n=60, RR 0.60 CI 0.2 to 2.3).

4.2 Global state: 2. needing additional antipsychotic 
 Two medium‐term studies (Ionescu 1983 and Iqbal 1978) showed equivocal findings (n=98, RR 0.67 CI 0.2 to 1.9).

4.3 Mental state ‐ BPRS 
 Only one medium‐term study (Ionescu 1983) has supplied data from the BPRS scale with means and standard deviations. There was no statistically significant difference between penfluridol and fluphenazine decanoate (n=60, WMD ‐2.60 CI ‐5.5 to 0.3).

4.4 Leaving the study early 
 Data from five medium‐term studies (Doongaji 1988, Gerlach 1975, Iqbal 1978, James 1977, Quitkin 1977) showed that more patients dropped out of the depot typical antipsychotics group than the penfluridol one (n=218, RR 0.55 CI 0.3 to 0.97, NNT 6 CI 3 to 50). In Iqbal 1978, one person was categorised as 'leaving the study early' due to attempted suicide in the penfluridol group. In this same trial no‐one in the penfluridol group left the study early due to side‐effects.

4.5 Adverse events: 1. needing antiparkinsonian medication 
 Data pooled from three medium‐term studies (Ionescu 1983, Iqbal 1978 and James 1977) did not show a difference between groups for this outcome (n=114, RR 0.87 CI 0.7 to 1.0).

4.6 Adverse events: other side effects 
 Several adverse effects were described in two trials, with no statistically significant differences between interventions for insomnia, akathisia, rigidity, lack of facial expression, difficulty with visual accommodation, hypersalivation, fatigue, tremors and lethargy.

Discussion

1. General 
 This systematic review is retrospective research and, after an exhaustive search for studies, it was possible to include in the analysis almost all publications of controlled clinical trials in which penfluridol was compared to placebo or to other antipsychotics. We found 27 relevant studies, but data are limited due to small sample sizes and poor reporting of outcomes. This may be partially explained because most of the trials were published in the 1970s, when there were no clear concerns and guidance for conducting trials were not yet published (Moher 2001). Moreover, many outcomes could not be re‐analysed with the conservative inclusion and methodological quality criteria that give strength to the Cochrane reviews. Information supplied by the study authors helped to improve the quality and reliability of this review.

The population of these primary studies may be peculiar in some aspects because most trials included patients that were hospitalised, with a long course of the disease, but excluded clinical co‐morbidities that are common in people with schizophrenia.

2. PENFLURIDOL versus PLACEBO 
 Penfluridol was superior compared to placebo in the efficacy outcomes (global state, need for additional antipsychotics). Nevertheless, small sample sizes demands caution in the interpretation of these findings. If analysing the data from placebo controlled studies, penfluridol seems to be a drug that presents the same profile as other typical antipsychotics.

3. PENFLURIDOL versus TYPICAL ANTIPSYCHOTICS 
 When compared to other antipsychotics (oral and depot), penfluridol showed similar efficacy and tolerance. Penfluridol did not lead to excessive sedation and in general terms it was well tolerated. When penfluridol was compared to depot medications, it showed better results in terms of keeping the patients in treatment, providing lower dropout rates in medium‐term (16.4% vs 31.5%, respectively, p=0.04, NNT 6).

4. Quality of reporting 
 Although only randomised and controlled clinical trials were included, 85% of them did not report the randomisation procedure for the participants. This implies that the quality of the randomised allocation process might have been affected. The trials were also limited due to small sample sizes and none of the included studies reported how the sample size was calculated.

All studies were carried out in the 1970s and 1980s, when methodological quality and the amount of data reported in the publications were generally poor. In the 1990s a number of papers highlighted the importance of detailing data, mainly data relating to methodology, quantity and quality of information published (Begg 1996, Schulz 1995). Scientific publications currently present stricter criteria for publishing clinical trials.

Authors' conclusions

Implications for practice.

1. For people with schizophrenia 
 Penfluridol has shown to be an adequate treatment option for people with schizophrenia, especially for those who do not respond to oral medication on a daily basis and who do not adapt to depot drugs. It is also effective in treating those chronic patients with residual psychotic symptoms who nevertheless need continuous use of antipsychotic medication. Penfluridol is also attractive in terms of cost, since it is an affordable drug with a monthly cost (40 mg/week) of US $4 on the Brazilian market.

2. For managers or policy makers 
 Data relating to service utilisation, satisfaction with care and economic outcomes were not reported.

3. For clinicians 
 Penfluridol is a suitable therapeutic alternative. Its efficacy is similar to other typical antipsychotics, both oral and depot, and its adverse effects seem to be similar to those from butyrophenones. Based on the doses of penfluridol used in the studies included in this review (mean dose ˜40 mg/ week from 13 studies), we suggest that a dose ranging from 40 to 80 mg/week may be enough to result in clinical improvement.

Implications for research.

1. General 
 All trials reviewed predated the CONSORT statement (Begg 1996, Moher 2001). Had this been anticipated, possibly far more data would have been available to inform practice. Allocation concealment gives the assurance that selection bias is kept to a minimum and should be properly described. Only three studies in this review reported independent allocation or allocation concealment. For the other studies, we were given little assurance that bias was minimised during the allocation procedure. Well reported and tested blinding could have encouraged confidence in the control of performance and detection bias. Included trials used a double blind design, except for D'Elia 1977, Ionescu 1983 and Ota 1974, which were open studies. Seven studies (Channabasavanna 1987, Claghorn 1979, Donlon 1978, Doongaji 1988, Jus 1974, Khorana 1988, Wang 1982) even described forms of blinding, such as use of identical capsules. It is also important to note how many, and from which groups, people were withdrawn in order to evaluate exclusion bias. Raters should be independent of treatment. Continuous data was poorly reported in the reviewed studies. It would have been helpful if study authors had presented data in a way which reflected associations between intervention and outcome, for example, relative risk, odds‐ratio, risk or mean differences, as well as raw numbers. Binary outcomes should be calculated in preference to continuous results, as they are easier to interpret. If p‐values are used, the exact value should be reported. Service utilisation data as well as satisfaction with care and economic outcomes should also be reported in trials.

2. Specific ‐ is more research needed? 
 All studies in this review were carried out in the 1970s and 1980s, and the last clinical trial included in this review took place in 1988. New studies comparing penfluridol to both typical and atypical antipsychotics could be undertaken. More precise studies, with larger sample sizes, could better define efficacy and comparative tolerance of this drug. 
 
 3. Note: the eight citations in the awaiting classification section of the review may alter the conclusions of the review once assessed.

What's new

Date	Event	Description
10 September 2012	Amended	Update search of Cochrane Schizophrenia Group's Trial Register (see Search methods for identification of studies), 8 studies added to Studies awaiting classification.
31 October 2008	Amended	Converted to new review format.

Data and analyses

Comparison 1. PENFLURIDOL versus PLACEBO.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Global state: 1. No marked improvement	4	159	Risk Ratio (M‐H, Random, 95% CI)	0.69 [0.56, 0.85]
1.1 medium‐term (CGI)	3	122	Risk Ratio (M‐H, Random, 95% CI)	0.62 [0.49, 0.80]
1.2 medium‐term (Zwanikken scale)	1	37	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.59, 1.22]
2 Global state: 2. Needing additional antipsychotic	6		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
2.1 short‐term	1	34	Risk Ratio (M‐H, Random, 95% CI)	0.45 [0.23, 0.86]
2.2 medium‐term	5	138	Risk Ratio (M‐H, Random, 95% CI)	0.43 [0.24, 0.77]
3 Leaving the study early	4		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
3.1 short‐term	1	34	Risk Ratio (M‐H, Random, 95% CI)	2.1 [0.24, 18.17]
3.2 medium‐term	3	91	Risk Ratio (M‐H, Random, 95% CI)	0.13 [0.02, 0.85]
4 Adverse events: 1. Extrapyramidal	2	70	Risk Ratio (M‐H, Random, 95% CI)	2.01 [0.20, 19.73]
4.1 short term	1	20	Risk Ratio (M‐H, Random, 95% CI)	0.5 [0.05, 4.67]
4.2 medium term	1	50	Risk Ratio (M‐H, Random, 95% CI)	5.33 [1.77, 16.05]
5 Adverse events: 2. Other side effects (medium term)	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
5.1 needing antiparkinson medication	2	51	Risk Ratio (M‐H, Random, 95% CI)	1.04 [0.57, 1.91]
5.2 insomnia	1	21	Risk Ratio (M‐H, Random, 95% CI)	2.73 [0.34, 22.16]
5.3 alterations in ECG	1	50	Risk Ratio (M‐H, Random, 95% CI)	3.0 [0.67, 13.46]

1.1. Analysis.

Comparison 1 PENFLURIDOL versus PLACEBO, Outcome 1 Global state: 1. No marked improvement.

1.2. Analysis.

Comparison 1 PENFLURIDOL versus PLACEBO, Outcome 2 Global state: 2. Needing additional antipsychotic.

1.3. Analysis.

Comparison 1 PENFLURIDOL versus PLACEBO, Outcome 3 Leaving the study early.

1.4. Analysis.

Comparison 1 PENFLURIDOL versus PLACEBO, Outcome 4 Adverse events: 1. Extrapyramidal.

1.5. Analysis.

Comparison 1 PENFLURIDOL versus PLACEBO, Outcome 5 Adverse events: 2. Other side effects (medium term).

Comparison 2. PENFLURIDOL versus TYPICAL ANTIPSYCHOTICS (ORAL).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Global state: 1. No marked improvement (CGI)	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.1 medium‐term	2	80	Risk Ratio (M‐H, Random, 95% CI)	0.92 [0.68, 1.24]
2 Global state: 2. Needing additional antypsychotic	5		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
2.1 short‐term	2	90	Risk Ratio (M‐H, Random, 95% CI)	1.35 [0.90, 2.01]
2.2 medium‐term	3	105	Risk Ratio (M‐H, Random, 95% CI)	1.08 [0.66, 1.75]
3 Mental state: 1. BPRS mean endpoint score	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
3.1 medium‐term	1	41	Mean Difference (IV, Fixed, 95% CI)	1.24 [‐4.40, 6.88]
4 Leaving the study early	9		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
4.1 short‐term	3	118	Risk Ratio (M‐H, Random, 95% CI)	0.46 [0.16, 1.35]
4.2 medium‐term	6	235	Risk Ratio (M‐H, Random, 95% CI)	1.16 [0.82, 1.65]
5 Adverse events (short term)	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
5.1 needing antiparkinsonism medication	2	90	Risk Ratio (M‐H, Random, 95% CI)	1.09 [0.61, 1.97]
5.2 insomnia	1	57	Risk Ratio (M‐H, Random, 95% CI)	1.07 [0.51, 2.24]
5.3 drowsiness	1	33	Risk Ratio (M‐H, Random, 95% CI)	0.19 [0.03, 1.40]
6 Adverse events (medium term)	7		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
6.1 needing antiparkinsonism medication	4	169	Risk Ratio (M‐H, Random, 95% CI)	1.20 [0.91, 1.58]
6.2 insomnia	2	88	Risk Ratio (M‐H, Random, 95% CI)	2.14 [0.43, 10.57]
6.3 akathisia	4	154	Risk Ratio (M‐H, Random, 95% CI)	1.28 [0.48, 3.41]
6.4 acute dystonia	2	57	Risk Ratio (M‐H, Random, 95% CI)	1.77 [0.23, 13.52]
6.5 lack of facial expression	1	32	Risk Ratio (M‐H, Random, 95% CI)	1.32 [0.57, 3.07]
6.6 drowsiness	2	81	Risk Ratio (M‐H, Random, 95% CI)	1.02 [0.48, 2.15]
6.7 nausea	1	25	Risk Ratio (M‐H, Random, 95% CI)	1.08 [0.27, 4.37]
6.8 extrapyramidal	3	135	Risk Ratio (M‐H, Random, 95% CI)	0.86 [0.44, 1.66]
6.9 fatigue	1	32	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.36, 2.47]
6.10 tremors	3	128	Risk Ratio (M‐H, Random, 95% CI)	1.41 [0.95, 2.08]
6.11 postural hypotension	1	25	Risk Ratio (M‐H, Random, 95% CI)	0.54 [0.06, 5.24]
6.12 rigidity	1	32	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.30, 2.77]
6.13 muscle spasm	1	56	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.01, 7.85]

2.1. Analysis.

Comparison 2 PENFLURIDOL versus TYPICAL ANTIPSYCHOTICS (ORAL), Outcome 1 Global state: 1. No marked improvement (CGI).

2.2. Analysis.

Comparison 2 PENFLURIDOL versus TYPICAL ANTIPSYCHOTICS (ORAL), Outcome 2 Global state: 2. Needing additional antypsychotic.

2.3. Analysis.

Comparison 2 PENFLURIDOL versus TYPICAL ANTIPSYCHOTICS (ORAL), Outcome 3 Mental state: 1. BPRS mean endpoint score.

2.4. Analysis.

Comparison 2 PENFLURIDOL versus TYPICAL ANTIPSYCHOTICS (ORAL), Outcome 4 Leaving the study early.

2.5. Analysis.

Comparison 2 PENFLURIDOL versus TYPICAL ANTIPSYCHOTICS (ORAL), Outcome 5 Adverse events (short term).

2.6. Analysis.

Comparison 2 PENFLURIDOL versus TYPICAL ANTIPSYCHOTICS (ORAL), Outcome 6 Adverse events (medium term).

Comparison 3. PENFLURIDOL versus TYPICAL ANTIPSYCHOTICS (DEPOT).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Global state: 1. No marked improvment (CGI)	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.1 medium‐term	1	60	Risk Ratio (M‐H, Random, 95% CI)	0.6 [0.16, 2.29]
2 Global state: 2. Needing additional antipsychotic	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
2.1 medium‐term	2	98	Risk Ratio (M‐H, Random, 95% CI)	0.67 [0.23, 1.94]
3 Mental state: 1. BPRS mean endpoint score	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
3.1 medium term	1	60	Mean Difference (IV, Fixed, 95% CI)	‐2.60 [‐5.45, 0.25]
4 Leaving the study early	5		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
4.1 medium‐term	5	218	Risk Ratio (M‐H, Random, 95% CI)	0.55 [0.31, 0.97]
5 Adverse events (medium term)	4		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
5.1 needing antiparkinsonism medication	3	114	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.74, 1.02]
5.2 insomnia	1	40	Risk Ratio (M‐H, Random, 95% CI)	0.45 [0.04, 4.60]
5.3 akathisia	1	40	Risk Ratio (M‐H, Random, 95% CI)	6.36 [0.35, 115.73]
5.4 lethargy	2	76	Risk Ratio (M‐H, Random, 95% CI)	0.57 [0.07, 4.93]
5.5 lack of facial expression	1	40	Risk Ratio (M‐H, Random, 95% CI)	0.30 [0.07, 1.32]
5.6 hypersalivation	1	60	Risk Ratio (M‐H, Random, 95% CI)	19.00 [1.16, 312.42]
5.7 fatigue	2	100	Risk Ratio (M‐H, Random, 95% CI)	0.98 [0.36, 2.64]
5.8 tremors	2	100	Risk Ratio (M‐H, Random, 95% CI)	0.38 [0.08, 1.78]
5.9 difficulty with visual accomodation	2	76	Risk Ratio (M‐H, Random, 95% CI)	2.36 [0.34, 16.62]
5.10 rigidity	2	100	Risk Ratio (M‐H, Random, 95% CI)	0.18 [0.03, 1.02]

3.1. Analysis.

Comparison 3 PENFLURIDOL versus TYPICAL ANTIPSYCHOTICS (DEPOT), Outcome 1 Global state: 1. No marked improvment (CGI).

3.2. Analysis.

Comparison 3 PENFLURIDOL versus TYPICAL ANTIPSYCHOTICS (DEPOT), Outcome 2 Global state: 2. Needing additional antipsychotic.

3.3. Analysis.

Comparison 3 PENFLURIDOL versus TYPICAL ANTIPSYCHOTICS (DEPOT), Outcome 3 Mental state: 1. BPRS mean endpoint score.

3.4. Analysis.

Comparison 3 PENFLURIDOL versus TYPICAL ANTIPSYCHOTICS (DEPOT), Outcome 4 Leaving the study early.

3.5. Analysis.

Comparison 3 PENFLURIDOL versus TYPICAL ANTIPSYCHOTICS (DEPOT), Outcome 5 Adverse events (medium term).

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Baro 1970.

Methods	Allocation: randomised, no further details. Blindness: double, no further details. Duration: 10 months. Design: 2 parallel groups, two phases ‐ 1st uncontrolled study for 8 months; 2nd RCT for more than 8 weeks. Analysis: non ITT.
Participants	Diagnosis: schizophrenia, no further details. N=26. Age: 24‐70 years. Sex: male. Setting: inpatient. History: unclear.
Interventions	1. Penfluridol: dose 10‐40 mg/week. N=13. 2. Placebo. N=13. Dexbenzetimide for EPSE as required.
Outcomes	Leaving the study early. Unable to use ‐ Global state: FCRS (no data). Mental state: PEQ (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Channabasavanna 1987.

Methods	Allocation: randomised. Blindness: double, identical capsules. Duration: 12 weeks. Analysis: non ITT.
Participants	Diagnosis: schizophrenia (DSM‐III). N=30. Age: mean ˜ 36 years. Sex: 16M, 12F*. History: mean length of illness ˜ 11 years, well maintained on antipsychotics, physically healthy. Setting: unclear.
Interventions	1. Penfluridol: dose 20 mg on day 1, 40 mg on day 7, 60 mg on day 14 and weekly thereafter. N=15. 2. Placebo. N=15. Trihexyphenidyl as required for EPSE.
Outcomes	Global state: needing additional antipsychotic drugs. Leaving the study early. Adverse events: medication for parkinsonian symptoms. Unable to use ‐ Global state: time to use additional antipsychotic drugs (idiosyncratic reporting of non‐pre‐specified outcome). Mental state: SANS, SAPS (no data). Adverse events: Simpson and Angus Scale (no data).
Notes	* sex of 2 participants not given.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Charalampous 1977.

Methods	Allocation: randomised, no further details. Blindness: double, no further details. Duration: 26 weeks. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia, no further details. N=55. Age: 21‐55 years, mean ˜ 35 years. Sex: male. History: chronic schizophrenia, mean length of illness 7.9 years. Setting: inpatient.
Interventions	1. Penfluridol: dose 20‐140 mg/week. N=28. 2. Fluphenazine: dose oral 21‐147 mg/day N=27. Benztropine mesylate for EPSE, and chloral hydrate for insomnia.
Outcomes	Global state: CGI improved/not improved. Leaving the study early. Adverse events: medication for parkinsonian symptoms. Unable to use ‐ Mental state: BPRS (no data). Behaviour: NOSIE (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Chouinard 1976.

Methods	Allocation: randomised, no further details. Blindness: double, no further details. Duration: 3 weeks. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia (DSM‐II). N=33. Age: unclear. Sex: male and female, numbers unclear. History: unclear. Setting: inpatients (brief therapy unit).
Interventions	1. Penfluridol: dose 40 mg on day 1, 80 mg on day 7, 120 mg on day 14. N=14. 2. Chlorpromazine: dose 300 mg/day for week 1, 600 mg/day for week 2, 900 mg/day for week 3. N=19. Procyclidine HCL for EPSE.
Outcomes	Global state: needing additional antipsychotic drugs. Leaving the study early. Adverse events: medication for parkinsonian symptoms. Unable to use ‐ Global state: CGI (no data). Mental state: BPRS, IMPS (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Claghorn 1979.

Methods	Allocation: randomised, no further details. Blindness: double, identical capsules. Duration: 52 weeks. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia, no further details. N=56. Age: mean ˜ 37 years. Sex: 32M, 24F. History: mean length of illness 10 years. Setting: outpatients.
Interventions	1. Penfluridol: dose maximum 160 mg/week. N=28. 2. Chlorpromazine: dose maximum 5320 mg/week. N=28.
Outcomes	Leaving the study early. Adverse events. Unable to use ‐ Global state: CGI (no data). Mental state: BPRS (no data). Social function: ESF (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

D'Elia 1977.

Methods	Allocation: rated by a physician, no further details. Blindness: open. Duration: 4 weeks. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia, no further details. N=28. Age: 19‐63 years. Sex: 22M, 6F. History: length of illness, 1‐38 years, duration of hospitalisation 5 days‐32 years. Setting: inpatient.
Interventions	1. Penfluridol: dose no further details. N=14. 2. Thiothixene: dose no further details. N=14.
Outcomes	Leaving the study early. Unable to use ‐ Changes in psychopathology: (no usable data). EEG variables: (no usable data). Visual averaged evoked responses: (no usable data).
Notes	The main aim of the study was to find out if outcome could be predicted by the EEG or VAER (visual averaged evoked response results obtained before treatment.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Deutsch 1976.

Methods	Allocation: randomised, no further details. Blindness: double, no further details. Duration: 12 weeks. Design: 2 parallel groups, two phases ‐ 1st uncontrolled study for 12 weeks; 2nd RCT for 12 weeks. Analysis: non ITT.
Participants	Diagnosis: chronic schizophrenia, no further details N=37. Age: unclear. Sex: unclear. History: unclear. Setting: inpatient.
Interventions	1. Penfluridol: dose 20‐140 mg/week. N=18. 2. Placebo: N=19.
Outcomes	Global state: CGI improved/not improved. Adverse events: number of participants presenting at least one adverse effect. Unable to use ‐ Mental state: BPRS (no data). Behaviour: NOSIE (no data). Adverse events: TESS (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Donlon 1978.

Methods	Allocation: randomised, no further details. Blindness: double, identical capsules. Duration: 52 weeks. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia, no further details. N=25. Age: mean ˜ 40 years. Sex: 12M, 13F. History: chronic, healthy, none had been hospitalised six months previously. Setting: outpatient.
Interventions	1. Penfluridol: dose 20‐160 mg/week. N=12. 2. Trifluoperazine: dose 14‐40 mg/day N=13. Thioridazine for extra‐medication. Benztropine mesylate or trihexyphenidyl HCL for EPSE.
Outcomes	Global state: CGI improved/not improved, needing additional antipsychotic drugs. Leaving the study early. Adverse events. Unable to use ‐ Mental state: BPRS (no data). Adverse events: TESS (no data). Social functioning: ESFR (no data). Depression: HPRSD (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Doongaji 1988.

Methods	Allocation: randomised, no further details. Blindness: double, all patients received capsules and injections. Duration: 12 weeks. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia (RDC). N=40. Age: mean ˜ 28 years. Sex: 27M, 13F. History: chronic and sub‐chronic schizophrenia, mean length of illness ˜2.3 years, physically healthy. Setting: unclear.
Interventions	1. Penfluridol: dose 20 mg on day 1, 40 mg on day 7, 60 mg on day 14 and thereafter. N=21. 2. Fluphenazine decanoate: dose 12.5 mg/IM on days 1& 7, 14.25 mg/IM weekly thereafter. N=19. Promethazine hydrochloride, benzhexol hydrochloride as needed for EPSE, and chloral hydrate or nitrazepam for insomnia.
Outcomes	Leaving the study early. Adverse events. Unable to use ‐ Global state: CGI (no SD). Mental state: MRS, BPRS (no SD).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Gallant 1974.

Methods	Allocation: randomised, no further details. Blindness: double, no further details. Duration: 12 weeks. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia, no further details. N=50. Age: 24‐59 years, mean ˜ 41 years. Sex: 25M, 25F. History: severely chronically ill patients, median years of hospitalisation ˜ 15 years, 12 weeks previous to the trial patients were stabilised on haloperidol (n=31) or penfluridol (n=19). Setting: inpatient.
Interventions	1. Penfluridol: dose maximum 160 mg/week N=25. 2. Placebo N=25.
Outcomes	Global state: CGI improved/not improved. Adverse events. Unable to use ‐ Mental state: BPRS (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Gerlach 1975.

Methods	Allocation: randomised, no further details. Blindness: double, no further details. Duration: 12 weeks. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia, no further details. N=60. Age: 18‐70 years. Sex: males. History: illness <2 years. Setting: in/outpatients.
Interventions	1. Penfluridol: dose 20 mg/week N=30. 2. Flupenthixol decanoate: dose 20‐100 mg/biweekly. N=30. Additional: chlorprothixene, procyclidine, amitriptyline, nitrazepam, choralodol.
Outcomes	Global state: CGI improved/not improved. Leaving the study early. Adverse events: various side effects, use of additional antipsychotic drugs. Unable to use ‐ Mental state: BPRS (no data). Behaviour: NOSIE (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Ionescu 1983.

Methods	Allocation: no information on allocation procedure. Blindness: open. Duration: one year. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: Kreplian sense schizophrenia. N=60. Age: 18‐49 years, mean ˜27 years. Sex: 36M, 24F. History: mean length of illness ˜3.8 years. Setting: in/outpatients.
Interventions	1. Penfluridol: dose 20‐60 mg/week, mean dose ˜30 mg/week. N=30. 2. Fluphenazine decanoate: dose mean 7.01 mg/week. N=30.
Outcomes	Global state: CGI improved/not improved, needing additional antipsychotic drugs. Mental state: BPRS. Adverse events: several, medication for parkinsonian symptoms.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Iqbal 1978.

Methods	Allocation: randomised, no further details. Blindness: double, no further details. Duration: 52 weeks. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia, no further details. N=38. Age: 18‐70 years, mean ˜37 years. Sex: 19M, 19F. History: length of illness 1.5 ‐ 28 years. Setting: outpatient.
Interventions	1. Penfluridol: dose mean 67 mg/week. N=17. 2. Fluphenazine decanoate: dose mean 25.5 mg/week. N=21. Additional: chloral hydrate for insomnia, benztropine mesylate for EPSE.
Outcomes	Global state: use of additional antipsychotics Leaving the study early. Adverse events: number of participants presenting at least one side effect, medication for parkinsonian symptoms. Unable to use ‐ Global state: CGI (no data). Mental state: BPRS (no data) . Social function: ESFR (no data).
Notes	Author informed that there was allocation concealment, but did not provide details.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Jacobsson 1976.

Methods	Allocation: no information on allocation procedure. Blindness: double. Duration: 10 weeks. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia or paranoid syndromes, no further details. N=57. Age: 19‐63 years, mean ˜41 years. Sex: 42M, 15F. History: length of illness 1‐38 years, mean ˜17 years. Setting: inpatient.
Interventions	1. Penfluridol: dose 20‐80 mg/week, mean dose 46.7 mg/week. N=29. 2. Thiothixene: dose 30‐100 mg/day, mean dose 48.3 mg/day. N=28. Additional: hypnotics for insomnia, and biperiden for EPSE.
Outcomes	Global state: use of additional antipsychotic drugs. Leaving the study early. Adverse events: several, medication for parkinsonian symptoms. Unable to use ‐ Nurses scale (no data). Doctors rating (S‐scale) (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

James 1977.

Methods	Allocation: randomised, no further details. Blindness: double, no further details. Duration: 24 weeks (week 1‐8 dosage adjustment period). Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia, no further details. N=16. Age: 20‐55 years. Sex: unclear. History: chronic illness (ill at least 1 year), physically healthy. Setting: inpatient.
Interventions	1. Penfluridol: dose 20‐60 mg/week, mean 35 mg/week. N=8. 2. Fluphenazine decanoate: dose 12.5‐35.5 mg/month, mean 35.5 mg/month. N=8.
Outcomes	Leaving study early. Adverse events: several, medication for parkinsonian symptoms. Unable to use ‐ Mental state: BPRS, FCRS (no data). Behaviour: NOSIE (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Jus 1974.

Methods	Allocation: randomised, no further details. Blindness: double, identical capsules. Duration: 12 weeks, preceded by 20 weeks open phase with penfluridol. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia, no further details. N=42*. Age: unclear. Sex: unclear. History: chronic illness. Setting: inpatient.
Interventions	1. Penfluridol: dose mean 68 mg/week. N=20. 2. Thioridazine: dose mean 312 mg/day, max 700 mg/day N=19. Antiparkinsonians for EPSE (no more data).
Outcomes	Global state: use of additional antipsychotic drug. Adverse events. Unable to use ‐ Global state: CGI (no data for separated intervention).
Notes	*Three participants dropped out in RCT with no information from which group.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Khorana 1988.

Methods	Allocation: randomised, no further details. Blindness: double, identical capsules and administration. Duration: 12 weeks. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia, no further details. N=50. Age: mean ˜ 29 years*. Sex: 21M, no further details*. History: chronic illness. Setting: unclear.
Interventions	1. Penfluridol: dose 20 mg/week. N=unclear. 2. Trifluoperazine: dose 10 mg/day. N=unclear.
Outcomes	Adverse events: several, medication for parkinsonian symptoms. Unable to use ‐ Mental state: BPRS, Hamilton, Wing, Venables, and O'Connor (no data).
Notes	* Authors described data just for completers (Penfluridol N=15 and Trifluoperazine N=17).
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Kurland 1975.

Methods	Allocation: randomised, no further details. Blindness: double, no further details. Duration: 12 weeks. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia, no further details. N=35. Age: 21‐60 years, mean ˜ 43 years. Sex: 19M, 16F. History: chronically ill, stabilised for 12 weeks using penfluridol. Setting: outpatient.
Interventions	1. Penfluridol: dose 20‐120 mg/week. N=18. 2. Placebo: N=17.
Outcomes	Global state: CGI improved/not improved, use of additional antipsychotic drugs. Leaving the study early. Unable to use ‐ Mental state: BPRS (no data). Behaviour: NOSIE (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Lapierre 1978.

Methods	Allocation: randomised, sealed bottles and envelopes. Blindness: double, no further details. Duration: 52 weeks. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia, no further details. N=41. Age: 18 ‐52 years, mean ˜ 37 years. Sex: 24M, 17F. History: mean duration of illness ˜8.9 years, physically healthy. Setting: inpatient.
Interventions	1. Penfluridol: dose 20‐120 mg/week, mean dose ˜60.9 mg/week N=20. 2. Fluphenazine: dose oral 3‐18 mg/day, mean dose 9.2 mg/day. N=21.
Outcomes	Mental state: BPRS. Leaving the study early. Adverse events: several, medication for parkinsonian symptoms. Unable to use ‐ Social function: Katz social adjustment scale (no data).
Notes	Author was contacted and informed randomisation procedure.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Munitz 1986.

Methods	Allocation: randomised*. Blindness: double, no further details. Duration: 6 weeks. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia (DSM‐III). N=20. Age: mean ˜ 33 years. Sex: 14M, 6F. History: newly admitted, drug free for at least a month. Setting: inpatient.
Interventions	1. Penfluridol: dose 20 mg for weeks 1‐6. N=10. 2. Penfluridol: dose 20 mg for weeks 2‐6 (preceded by one week of placebo). N=10. Benzhexol for EPSE.
Outcomes	Adverse events. Unable to use ‐ Global state: CGI (no data). Mental state: BPRS (no data). Behaviour: NOSIE (no data).
Notes	* Recruitment of patients was carried out in the emergency unit by a physician who did not participate in the study. The random numbers were kept by the nursing staff. The treatment physician was blind to the assignment. The medication was administered by the nursing staff. Data obtained by the author of the study.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Ota 1974.

Methods	Allocation: randomised, divided in 2 groups with 2 randomisation procedures performed. Blindness: open. Duration: 17 weeks. Design: 4 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia, no further details. N=18. Age: 21‐60 years, mean ˜ 43 years. Sex: unclear. History: length of illness at least 1.5 years, good physical conditions. Setting: inpatients.
Interventions	1. Penfluridol: dose 5‐200 mg/week. N=12. 2. Atypical antipsychotics. N=6. Additional: haloperidol as extra antipsychotic; benztropine mesylate for EPSE.
Outcomes	Leaving the study early. Unable to use ‐ Global state: CGI (no data), needing additional antipsychotic drugs (no data). Mental state: BPRS (no data). Behaviour: NOSIE (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Quitkin 1977.

Methods	Allocation: randomised, no further details. Blindness: double, no further details. Duration: 52 weeks. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: RDC schizophrenia. N=60. Age: 17‐48 years. Sex: 41M, 19F. History: unclear. Setting: inpatients.
Interventions	1. Penfluridol: dose 20‐160 mg/week. N=30. 2. Fluphenazine decanoate: dose 0.5‐4 ml/fortnightly. N=30. Flurazepam HCL for insomnia, and procyclidine HCL for EPSE. Psychotherapy, and recreation groups.
Outcomes	Leaving the study early. Adverse events. Unable to use ‐ Global state: CGI (no data). Mental state: BPRS (no data). Social function: SAS (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Roelofs 1974.

Methods	Allocation: divided into two groups by an unbiased statistician, no further details. Blindness: double, no further details. Duration: 6 months. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: chronic psychotic patients, no further details. N=15. Age: 25‐64 years, mean ˜ 54 years. Sex: 6M, 9F. History: hospitalisation 1‐31 years, mean ˜5 years. Setting: inpatients.
Interventions	1. Penfluridol: dose mean 40 mg/week. N=7. 2. Placebo. N=8. Dexitimine 1 mg/day (mean dose).
Outcomes	Global state: needing additional antipsychotic drugs. Unable to use ‐ Rating scale for psychiatric evaluation (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Van Praag 1971.

Methods	Allocation: randomised, no further details. Blindness: double, no further details. Duration: 4 weeks. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: acute psychosis, no further details. N=34. Age: 19‐70 years, mean ˜ 41 years. Sex: unclear. History: admitted to a university psychiatric clinic with acute psychosis. Setting: unclear.
Interventions	1. Penfluridol: dose 30 mg/week + dexitimine 0.5 mg/day. N=20. 2. Placebo. N=14.
Outcomes	Global state: needing additional antipsychotic drugs. Leaving the study early.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Vandecasteele 1974.

Methods	Allocation: divided into two groups by an unbiased statistician, no further details. Blindness: double, no further details. Duration: 6 months. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: chronic schizophrenia (DSM‐II). N=21. Age: 25‐64 years, mean ˜59 years. Sex: females. History: hospitalisation 2‐41 years, mean 27 years. Setting: inpatients.
Interventions	1. Penfluridol: dose 20‐120 mg/week, mean 40 mg/week. N=11. 2. Placebo. N=10.
Outcomes	Global state: needing additional antipsychotic drugs. Adverse events: several, medication for parkinsonian symptoms. Unable to use ‐ General symptoms rating scale (no data, Zwanikken 1971).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Wang 1982.

Methods	Allocation: randomised, no further details. Blindness: double, capsules of identical appearance. Duration: 12‐20 weeks, preceded by 4‐12 weeks on penfluridol. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia, no further details. N=41. Age: 20‐60 years. Sex: males. History: unclear. Setting: unclear.
Interventions	1. Penfluridol: dose 20‐140 mg/week. N=21. 2. Chlorpromazine: dose 150‐1050 mg/day. N=20. Medication on twice‐a‐day schedule, benztropine mesylate and chloral hydrate as needed.
Outcomes	Global state: use of additional antipsychotic drugs. Leaving the study early. Adverse events: medication for parkinsonian symptoms. Unable to use ‐ Global state: CGI (no SD). Mental state: BPRS (no SD). Behaviour: NOSIE (no SD). Social functioning: ESF (no SD).
Notes	It is described in the paper that there were no statistical differences in scale‐derived data, and that people on penfluridol had akathisia more frequently than those on chlorpromazine, but no data presented.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Zwanikken 1973.

Methods	Allocation: randomised, no further details. Blindness: double, no further details. Duration: 6 months preceded by 3 months of an open phase. Design: 2 parallel groups. Analysis: non ITT.
Participants	Diagnosis: schizophrenia (DSM‐II), (n=17) and mental retardation (n=20). N=37. Age: 22‐74 years, mean ˜ 46 years. Sex: females. History: length of illness 1‐53 years, mean 16.5 years. Setting: inpatients.
Interventions	1. Penfluridol: dose mean 57.3 mg/week. N=20. 2. Placebo. N=17. Haloperidol as a possible extra medication; dexitimine 0.5 mg/week for extrapyramidal symptoms.
Outcomes	Global state: Zwanikken scale, use of additional antipsychotic drugs. Unable to use ‐ Adverse events: medication for parkinsonian symptoms (no data).
Notes	Author contacted.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Diagnostic tools 
 DSM‐II ‐ Diagnostic Statistical Manual, version 2 
 DSM‐III ‐ Diagnostic Statistical Manual, version 3 
 RDC ‐ Research Diagnostic Criteria

Rating scales:

Global state ‐ 
 CGI ‐ Clinical global impression 
 FCRS ‐ Factor construct rating scale 
 Zwanikken scale

Mental state ‐ 
 BPRS ‐ Brief psychiatric rating scale 
 IMPS ‐ Impatient multidimensional psichiatric scale 
 PEQ ‐ Psychiatric Evaluation Questionnaire 
 SANS ‐ Scale for assessment of negative synptoms 
 SAPS ‐ Scale for assessment of positive symptoms 
 Scale A ‐ Winng (1961) modified 
 Scale B ‐ of authors own design

Behaviour ‐ 
 NOSIE ‐ Nurses observation for patient evaluation

Adverse events ‐ 
 EPSE ‐ Extra‐pyramidal side effects MRS ‐ Modified Rating Scale 
 TESS ‐ Treatment Emergent Symptom Scale

Depression ‐ 
 HPRSD ‐ Hamilton Psychiatric Rating Scale for Depression

Social function ‐ 
 SAS ‐ Social adjustment scale 
 ESF ‐ Evaluation of Social Functioning 
 ESFR ‐ Evaluation of Social Functioning Factors

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Abuzzahab 1977	Allocation: not described. Participants: people with schizophrenia. Interventions: penfluridol versus trifluoperazine versus stelazine. Outcomes: no usable data.
Angst 1975	Allocation: not described. Participants: people with schizophrenia. Interventions: fluphenazine versus fluspirilene versus penfluridol versus perphenazine versus pipotiazine. Outcomes: no usable data.
Bao 1988	Allocation: not randomised, sequential.
Baro 1972	Allocation: not randomised.
Bobon 1970	Allocation: not randomised, case series.
Chouinard 1977	Allocation: randomised. Participants: people with schizophrenia. Interventions: penfluridol versus chlorpromazine and the use of procyclidine. Outcomes: only data for penfluridol and procyclidine were reported.
Cooper 1975	Allocation: not randomised.
Eklund 1976	Allocation: not described. Participants: people with schizophrenia. Interventions: penfluridol versus perphenazine. Outcomes: data not usable "As evident from the table, the penfluridol patients are older and have a longer period of illness behind them" (author of the study).
Floru 1973	Allocation: not randomised, case series.
Floru 1975	Allocation: not randomised.
Gallant 1975	Allocation: not randomised.
Glatzel 1973	Allocation: not randomised, case series.
Hasse 1971	Allocation: not randomised, case series.
Kingstone 1977	Allocation: not randomised, case series.
Li 1987	Allocation: randomised. Participants: people with schizophrenia. Interventions: clozapine versus chlorpromazine versus penfluridol. Outcomes: no usable data.
Linden 1973	Allocation: randomised. Participants: people with schizophrenia. Intervention: penfluridol versus fluspirilene. Outcomes: no usable data.
Morales 1981	Allocation: not randomised, case series.
Nistico 1974	Allocation: not described. Participants: people with schizophrenia. Interventions: flupenthixol versus penfluridol. Outcomes: no usable data.
Richards 1976	Allocation: not randomised, case series.
Sharma 1988	Allocation: not randomised, case series.
Shopsin 1977	Allocation: not randomised, case series.
Tanghe 1972	Allocation: not randomised.
Vaidyalingam 1990	Allocation: not randomised, case series.
Van Lommel 1974	Allocation: not described. Participants: people with schizophrenia. Interventions: penfluridol versus haloperidol versus placebo. Outcomes: no usable data.
Woggon 1978	Allocation: not randomised, case series.

Characteristics of studies awaiting assessment [ordered by study ID]

Baro 1972a.

Methods
Participants
Interventions
Outcomes
Notes	To be assessed.

Duval 2006.

Methods
Participants
Interventions
Outcomes
Notes	To be assessed.

Intoh 1990.

Methods
Participants
Interventions
Outcomes
Notes	To be assessed.

Itoh 1976.

Methods
Participants
Interventions
Outcomes
Notes	To be assessed.

Jiang 1995.

Methods
Participants
Interventions
Outcomes
Notes	To be assessed.

Jus 1976.

Methods
Participants
Interventions
Outcomes
Notes	To be assessed.

Kondo 1975.

Methods
Participants
Interventions
Outcomes
Notes	To be assessed.

Shu 1983.

Methods
Participants
Interventions
Outcomes
Notes	To be assessed.

Strous 2007.

Methods
Participants
Interventions
Outcomes
Notes	To be assessed.

Xun 2006.

Methods
Participants
Interventions
Outcomes
Notes	To be assessed.

ปรีชา 2533.

Methods
Participants
Interventions
Outcomes
Notes	To be assessed.

孙轻骑 2007.

Methods
Participants
Interventions
Outcomes
Notes	To be assessed.

Contributions of authors

Bernardo Soares ‐ review writing, data collection, analysis, final report.

Maurício Lima ‐ review writing, data collection, analysis, final report.

Sources of support

Internal sources

• Universidade da Região de Joinville ‐ UNIVILLE, Brazil.

• Brazilian Cochrane Centre, Brazil.

External sources

• Centro de Medicina Baseada em Evidências ‐ Universidade Federal de Pelotas, Brazil.

Declarations of interest

Bernardo Soares and Mauricio Lima are employees of Eli Lilly Brazil.

Edited (no change to conclusions)