Haloperidol versus chlorpromazine for schizophrenia

Abstract

Background

Chlorpromazine and haloperidol are benchmark antipsychotic drugs. Both are said to be equally effective when used at equivalent doses, but have different side‐effect profiles.

Objectives

To compare the effects of haloperidol and chlorpromazine for people with schizophrenia and schizophrenia‐like psychoses.

Search methods

We searched the Cochrane Schizophrenia Group's register (August 2006). We searched references of all included studies for further trials. We contacted pharmaceutical companies and authors of relevant trials.

Selection criteria

We included all randomised controlled trials that compared haloperidol with chlorpromazine for people with schizophrenia and/or schizophrenia‐like psychoses.

Data collection and analysis

Citations and, where possible, abstracts were independently inspected by at least two reviewers, papers ordered, re‐inspected and quality assessed. We independently extracted data. For dichotomous data we calculated the relative risk (RR), 95% confidence interval (CI) and, where appropriate, the number needed to treat (NNT) on an intention‐to‐treat basis using a random‐effects model. For continuous data, we calculated weighted mean differences (WMD).

Main results

We found 14 relevant studies, mostly of short duration, poorly reported and conducted in the 1970s (total n=794 participants). Nine of these compared oral formulations of both compounds, and five compared intramuscular formulations.

Haloperidol was associated with significantly fewer people leaving the studies early (13 RCTs, n=476, RR 0.26 CI 0.08 to 0.82). The efficacy outcome 'no significant improvement' tended to favour haloperidol, but this difference was not statistically significant (9 RCTs, n=400, RR 0.81 CI 0.64 to 1.04). Movement disorders were more frequent in the haloperidol groups ('at least one extrapyramidal side effect': 6 RCTs, n=37, RR 2.2 CI 1.1 to 4.4, NNH 5 CI 3 to 33), while chlorpromazine was associated with more frequent hypotension (5 RCTs, n=175, RR 0.31 CI 0.11 to 0.88, NNH 7 CI 4 to 25). Similar trends were found when studies comparing intramuscular formulations and studies comparing oral formulations were analysed separately.

Authors' conclusions

Given that haloperidol and chlorpromazine are global standard antipsychotic treatments for schizophrenia, it is surprising that less than 800 people have been randomised to a comparison and that incomplete reporting still makes it difficult for anyone to draw clear conclusions on the comparative effects of these drugs. However, it seems that haloperidol causes more movement disorders than chlorpromazine, while chlorpromazine is significantly more likely to lead to hypotonia. We are surprised to have to say that we feel further, large, well designed, conducted and reported studies are required.

Plain language summary

Haloperidol versus chlorpromazine for schizophrenia

Schizophrenia can be a long‐term, chronic illness with a worldwide lifetime prevalence of about one per cent. The most common treatment of this condition is using antipsychotics. In the developed world there is a large choice of antipsychotics including some that are quite expensive, whereas in the developing world the older and cheaper drugs such as haloperidol and chlorpromazine are still used for the majority of the people. In addition, most new medications are tested for their effectiveness against haloperidol or chlorpromazine. This review looks at clinical trials comparing people with schizophrenia who have been treated with either chlorpromazine or haloperidol, in tablet form or as injection into muscle but not by long acting injection. There were 14 trials identified containing a total of 794 people. The trials varied in length from ‘several hours’ to 36 weeks but only two were six months or longer. The most recent trial was published in 1994, and the earliest 1962. As diagnosis of schizophrenia has changed over the years, some people in the early trials may have diagnoses other than schizophrenia by today’s criteria.

Overall the mental health and general functioning of people taking chlorpromazine or haloperidol improved. Compared to nowadays fewer people in both arms of the trial left the study early but those taking chlorpromazine were statistically more likely to do so. This was also the case when the oral medications were analysed on their own, but not in the case of the injected form. There was a suggestion that people may leave the study because of either adverse events or because the treatment did not work well. Haloperidol had statistically more movement side effects while chlorpromazine was statistically more likely to cause low blood pressure (hypotension). Although these trials show both haloperidol and chlorpromazine to be effective drugs for schizophrenia, a lot of data were not able to be used because some measures were not reported. Therefore the area would benefit from a new trial with a large number of people and lasting at least a year.

(Plain language summary prepared for this review by Janey Antoniou of RETHINK, UK www.rethink.org).

Background

Chlorpromazine and haloperidol are both benchmark antipsychotics (Adams 2007; Joy 2006). When Delay and Deniker discovered chlorpromazine's antipsychotic properties in 1952 (Delay 1952) this revolutionised the treatment of schizophrenia. Despite the development of numerous other antipsychotic agents chlorpromazine is still listed among the "essential drugs" by the World Health Organization (WHO 2002). It is estimated by the WHO (WHO 2001) that there are 24 million people suffering from schizophrenia worldwide, 50% of whom are untreated, 90% living in the developing world. On the WHO Essential Medicines List, (WHO 2002), chlorpromazine, haloperidol and fluphenazine are advocated for the treatment of psychotic disorders. Haloperidol was developed later in the 1950s (Ayd 1972), but seemingly it has replaced chlorpromazine as a benchmark treatment for the evaluation of new agents. According to Cochrane Reviews on new, so called 'atypical', antipsychotics (amisulpride, olanzapine, quetiapine, olanzapine, risperidone, ziprasidone (see issue 1, 2003, The Cochrane Library) 67.7% (n=44) of the trials published since 1990 used haloperidol, 10.8% (n=7) used chlorpromazine, 6.1% (n=4) used "standard treatment" and 15.4% (n=10) used other conventional antipsychotics as benchmarks.

Both drugs are said to be equally effective when used at equivalent doses (Davis 1989). According to textbooks (Benkert 2000) chlorpromazine is associated with more so‐called "autonomic" side effects such as sedation or hypotension, and anticholinergic side‐effects such as dry mouth, whereas haloperidol may cause more movement disorders, muscular stiffness or overwhelming feelings of restlessness (akathisia). However, to our best knowledge, both compounds have never been compared in a systematic review. The aim of this publication is to fill this gap. The results are of potential interest not only for users of these drugs, but they might also help researchers to find the best benchmark for the evaluation of new antipsychotics.

Technical background 
 The differences in terms of side‐effects between both compounds can be explained by their receptor binding profiles. Haloperidol has a very high affinity to central D2 receptors (Seeman 1976). The theory is that binding to these receptors in the limbic system is associated with the alleviation of schizophrenic symptoms. Binding to dopamine receptors in the basal ganglia, however, is the reason for the movement disorders associated with haloperidol. Chlorpromazine must be given in higher doses to obtain the same dopamine receptor occupancy as haloperidol (Seeman 1976). In contrast to haloperidol it has a relatively high affinity to a number of receptors which are associated with other side‐effects: histamine receptors (sedation and weight gain), alpha‐receptors (orthostatic hypotension and vertigo) and cholinergic receptors (obstipation, urinary retention, dry mouth and blurred vision; Benkert 2000).

Objectives

To compare the effects of haloperidol and chlorpromazine for people with schizophrenia and schizophrenia‐like psychoses.

Methods

Criteria for considering studies for this review

Types of studies

Relevant randomised controlled trials.

Types of participants

People with schizophrenia, schizophreniform psychoses, delusional disorder, schizoaffective psychoses and non‐affective serious/chronic mental illness irrespective of mode of diagnosis, age, sex, chronicity of illness.

Types of interventions

1. Haloperidol: any dose and route of administration with the exception of depot formulations (haloperidol decanoate). 
 2. Chlorpromazine: any dose and route of administration.

Types of outcome measures

We reported all outcomes for the short term (up to 12 weeks), medium term (13 to 26 weeks) and long term (over 26 weeks).

Primary outcomes

1. Leaving the study early 
 1.1 Due to any reason

2. Clinical response 
 2.1 No clinically significant response ‐ as defined by each of the studies

3. Adverse events 
 3.1 Anticholinergic adverse events (blurred vision, dry mouth, obstipation, urinary retention) 
 3.2 Central nervous system (ataxia, nystagmus, fits, diplopia, sedation) 
 3.3 Extrapyramidal adverse events (general, akathisia, dystonia, dyskinesia, malignant neuroleptic syndrome, muscular stiffness, parkinsonism, tardive dyskinesia, tremor)

Secondary outcomes

1. Leaving the study early 
 1.1 Due to adverse events 
 1.2 Due to inefficacy of treatment

2. Service utilisation 
 2.1 Hospital admission 
 2.2 Days in hospital 
 2.3 Change in hospital status

3. Clinical response: 
 3.1 No clinically significant improvement of global state 
 3.2 Average score/change in global state 
 3.3 Relapse ‐ as defined by each of the studies 
 3.4 No clinically significant improvement in mental state ‐ as defined by each of the studies 
 3.5 Average score/change in mental state 
 3.6 No clinically significant response in positive symptoms ‐ as defined by each of the studies 
 3.7 Average score/change in positive symptoms 
 3.8 No clinically significant response in negative symptoms ‐ as defined by each of the studies 
 3.9 Average score/change in negative symptoms 
 3.10 No clinically significant response in depressive symptoms ‐ as defined by each of the studies 
 3.11 Average score/change in depressive symptoms 
 3.12 No clinically significant response in manic symptoms ‐ as defined by each of the studies 
 3.13 Average score/change in manic symptoms

4. Behaviour 
 4.1 General behaviour 
 4.2 Specific behaviour 
 4.2.1 Social functioning 
 4.2.2 Employment status during trial (employed / unemployed) 
 4.2.3 Occurrence of violent incidents (to self, others, or property)

5. Adverse events 
 5.1 General adverse events 
 5.2 Specific adverse events 
 5.2.1 Allergic reactions 
 5.2.2 Blood dyscrasia such as agranulocytosis 
 5.2.3 Death 
 5.2.4 Gastrointestinal (nausea, vomiting, diarrhoea)

6. Additional drug use 
 6.1 Antipsychotics 
 6.2 Benzodiazepines

7. Economic (cost of care)

Search methods for identification of studies

Electronic searches

We searched the Cochrane Schizophrenia Group's Register (last search August 2006) using the phrase:

[(*anadep* or *chloractil* or *chlorazin* or *chlorpromados* or*chlorpromazine* or *chlorprom‐ez‐ets* or *(chlor p‐z)* or *chromedazine* or *cpz* or *elmarine* or *esmind* or *fenactil* or *hibanil* or *hibernal* or *klorazin* or *klorproman* or *klorpromez* or *largactil* or *megaphen* or *neurazine* or *plegomazine* or *procalm* or *promachel* or *promacid* or *promapar* or *promexin* or *promosol* or *prozil* or *psychozine* or *psylactil* or *serazone* or *sonazine* or *thoradex* or *thorazine* or *tranzine*)] AND [((*haloperi* or *R‐1625* or *haldol* or *alased* or *aloperidi* or *bioperido* or *buterid* or *ceree* or *dozic* or *duraperido* or *fortuna* or *serena* or *serenel* or *seviu* or *sigaperid* or *sylad* or *zafri* in title, abstract or index terms of REFERENCE)) or [haloperidol and chlorpromazine in interventions of STUDY)].

The Schizophrenia Groups trials register is based on regular searches of BIOSIS Inside; CENTRAL; CINAHL; EMBASE; MEDLINE and PsycINFO; the hand searching of relevant journals and conference proceedings, and searches of several key grey literature sources. A full description is given in the Group's module.

Searching other resources

1. Reference lists 
 We searched all references of articles selected for inclusion for further relevant trials.

2. Pharmaceutical companies 
 We contacted companies producing haloperidol or chlorpromazine directly to obtain data on unpublished trials.

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

Data collection and analysis

[For definitions of terms used in this, and other sections, please refer to the Glossary]

1. Study selection 
 All study citations identified by the searches were independently inspected by at least two reviewers, and we obtained full reports of the studies of agreed relevance. Where agreement could not be reached, we acquired the full report for more detailed scrutiny. These articles were then inspected independently by at least two reviewers to assess their relevance to the review. Again, if the disagreement could not be resolved from published information, we added the article to those awaiting assessment and contacted the authors of the study for clarification.

2. Quality assessment 
 We assessed he methodological quality of the trials included in this review using the criteria described in The Cochrane Handbook (Higgins 2005). These criteria are based on the evidence of a strong relationship between the potential for bias in the results and the allocation concealment (Schulz 1995) and defined as 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 criteria A or B of the Handbook.

3. Data extraction 
 Two reviewers independently extracted the data from included studies. Again, we discussed any disagreement documented the decisions and, if necessary, contacted the authors of the studies for clarification. We documented justification for excluding references from the review.

4. Data management 
 4.1 Intention to treat 
 For studies that did not specify the reasons for people leaving the study early (drop‐outs), we assumed that these people had no change in the clinical outcome variables. Wahlbeck and colleagues highlighted the problem of high drop‐out rates in randomised trials of drug treatments for schizophrenia (Wahlbeck 2001). Since there is no evidence on the degree of attrition which makes a reasonable analysis of the data possible, we included all trials in the main analysis. However, we carried out a sensitivity analysis to test whether the exclusion of trials with drop‐out rates of higher than 50% significantly changed the results of the primary outcome parameters.

4.2 Crossover design 
 We expected that some trials would use a crossover design. In order to exclude the potential additive effect in the second or later stages of these trials, we only analysed data from the first stage.

4.3 Data types 
 Outcomes were assessed using continuous (for example changes on a behaviour scale), categorical (for example, one of three categories on a behaviour scale, such as 'little change', 'moderate change' or 'much change') or dichotomous measures (for example, either 'no important changes' or 'important changes' in a person's behaviour). Currently RevMan does not support categorical data so they could not be analysed as such.

4.3.1 Dichotomous data: Where possible we attempted to convert outcome measures to dichotomous data. This can be done by identifying cut off points on rating scales and dividing subjects accordingly into 'clinically improved' or 'not clinically improved'. If the authors of a study had used a predefined cut off point for determining clinical effectiveness we used this where appropriate. Otherwise we generally assumed that if there had been a 50% reduction in a scale‐derived score, this could be considered as a clinically significant response. Similarly, a rating of 'at least much improved' according to the Clinical Global Impression Scale (Guy 1976) could be considered as a clinically significant response.

For dichotomous outcomes, we estimated a relative risk (RR) with the 95% confidence interval (CI) based on a random effects model. We inspected data, however, to see if an analysis using a fixed effects model for homogeneous results made a substantive difference. When overall results were significant, we calculated the Number Needed to Treat (NNT) and/or the Number Needed to Harm (NNH) as the inverse of the risk reduction. It has been shown that RR is more intuitive (Boissel 1999) than odds ratios and that odds ratios tend to be interpreted as RR by clinicians (Deeks 2000). This misinterpretation then leads to an overestimate of the impression of the effect.

4.3.2 Continuous data 
 4.3.2.1 Normal distribution: Continuous data on outcomes in trials relevant to mental health issues are often not normally distributed. To avoid the pitfall of applying parametric tests to non‐parametric data we applied the following standards to data derived from continuous measures of endpoint ('state' data) before inclusion: 
 When a scale starts from zero, the standard deviation, when multiplied by two, is less than the mean (as otherwise the mean is unlikely to be an appropriate measure of the centre of the distribution ‐ Altman 1996). Endpoint scores on scales often have a finite start and end point and this rule can be applied to them.

When continuous data are presented on a scale which includes a possibility of negative values (such as change on a scale) it is impossible to tell whether data are non‐normally distributed (skewed) or not. It is thus preferable to use scale end point data, which typically cannot have negative values. If end point data were not available, reviewers chose to use change data, because the statistics used in Metaview are rather robust towards skewness. If a scale starts from a positive value (such as PANSS, which can have values from 30‐210) the calculation described above should be modified to take the scale starting point into account. In these cases skewness is present if 2SD>(S‐Smin), where S is the mean score and Smin is the minimum score.

4.3.2.2 Intention‐to‐treat versus completer analyses: We assumed that for incomplete continuous data intention‐to‐treat analyses would be impossible. In these cases we analysed data as they were presented in the original publications.

4.3.2.3 Summary statistic: For continuous outcomes, we estimated a weighted mean difference (WMD) between groups. Again, a random effects model was used in the primary analysis, but it was inspected whether the use of a fixed‐effects model made a substantive difference for statistically homogeneous outcomes. Whenever possible we took the opportunity to make direct comparisons between trials that used the same measurement instrument to quantify specific outcomes. Where continuous data were presented from different scales rating the same effect, we presented both sets of data and inspected the general direction of effect.

4.3.2.4 Rating scales: A wide range of instruments is 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. It has been shown that the use of rating scales which had not been described in a peer‐reviewed journal (Marshall 2000) is associated with bias; therefore we excluded the results of such scales. Furthermore, the instrument should either be a self report or be completed by an independent rater or relative (not the therapist), and the instrument could be considered a global assessment of an area of functioning. However, as it was expected that therapists would frequently also be the rater, we included on such data but commented on the data as 'prone to bias'.

4.4 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 1 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]. If the ICC was not reported it was assumed to be 0.1 (Ukoumunne 1999). 
 
 If cluster studies had been appropriately analysed taking into account intra‐class correlation coefficients and relevant data documented in the report, synthesis with other studies would have been possible using the generic inverse variance technique.

4.5 Data display 
 We entered data into RevMan in such a way that the area to the left of the line of no effect indicated a favourable outcome for haloperidol. Haloperidol was considered to be the intervention, because it was developed later than chlorpromazine.

5. Heterogeneity 
 Firstly, we undertook consideration of all the included studies within any comparison to judge clinical heterogeneity. Then we visually inspected graphs to investigate the possibility of statistical heterogeneity. This was supplemented using, primarily, the I‐squared statistic. This provides an estimate of the percentage of variability due to heterogeneity rather than chance alone. Where the I‐squared estimate was greater than or equal to 50% we interpreted this as indicating the presence of considerable levels of heterogeneity (Higgins 2005). If inconsistency was high and clear reasons explaining the heterogeneity were found we presented the data separately. If not, we commented on the heterogeneity of the data.

6. Publication bias 
 We entered data from all included trials into a funnel graph (trial effect versus trial size or 'precision') in an attempt to investigate the likelihood of overt publication bias. Where appropriate we undertook a formal test of funnel plot asymmetry (suggesting potential publication bias) (Egger 1997). Significance levels of p < 0.1 were set a priori to accept the presence of asymmetry.

7. Sensitivity analysis 
 We examined whether the exclusion of trials with drop‐out rates of higher than 50% significantly changed the results of the primary outcome parameters.

Results

Description of studies

For more detailed descriptions please see 'Included' and 'Excluded' studies tables.

1. Excluded studies 
 We had to exclude nineteen studies because they were not (adequately) randomized or because the interventions relevant to this review were not randomized. In two studies the participants were not appropriate for the review, and in one it was unclear how many had schizophrenia. We had to exclude ten studies because they did not compare appropriate interventions (bromperidol decanoate, risperidone versus typical neuroleptics, chlorpromazine versus tryptophan‐benserazide, chlorpromazine versus placebo, chlorpromazine versus clozapine, clozapine versus chlorpromazine or haloperidol, chlorprothixene versus haloperidol, haloperidol combined with chlorpromazine versus haloperidol alone, haloperidol versus levomepromazine versus perphenazine, buronil vs promazine). Five studies had no usable data.

2. Studies awaiting assessment 
 We have not identified any studies awaiting assessment.

3. Ongoing studies 
 We have not identified any ongoing studies.

4. Included studies 
 Fourteen studies fulfilled the inclusion criteria and presented data that could be used for at least one comparison.

4.1 Length of trials 
 The duration of the included trials covered a range from 48 hours to three years, with the most common duration between four to 12 weeks.

4.2 Design 
 All trials used a parallel‐group design. Clark 1969; Gallant 1967; Rompel 1978; Serafetinides 1972 and Teja 1975 were single‐centred. Nishizono 1994 was multi‐centred.

4.3 Participants 
 Most studies used a 'clinical diagnosis' for the assessment of the participants. Bailine 1981 and Bi 1994 applied Research Diagnostic Criteria for schizophrenia and schizoaffective disorder. Bi 1994 applied the Chinese Classification of Mental Disorders (CCMD‐2) in addition. Teja 1975 used DSM‐II. Lempérière 1962 applied the French classification for paranoid schizophrenia. Nishizono 1994 (schizophrenia) and Ritter 1972 (acute or chronic schizophrenia, most paranoid schizophrenia, 1 involutional melancholia, 1 manic‐depressive) did not describe their classification system. Shopsin 1969 also did not mention a clear diagnosis, but the participants suffered from 'acute psychotic behaviour with excessive psychomotor hyperactivity'.

In those studies which reported the gender of the participants there were a total of 255 males and 311 females. Their mean ages ranged from 25.2 to 52 years.

4.4 Setting 
 Ten studies were conducted in inpatient settings; four studies did not provide information on the setting.

4.5 Study size 
 The number of people in the included studies ranged from 20 to 169 (mean 56.7, SD 36.4). 
 A total of 794 people participated in the 14 trials. Most of them had a diagnosis of schizophrenia, but four studies included other diagnoses (see 'Included studies table').

4.6 Interventions 
 Five studies compared intramuscular haloperidol (n=109) versus intramuscular chlorpromazine (n=84). Nine studies compared oral haloperidol (n=183) versus oral chlorpromazine (n=121).

4.6.1 Haloperidol 
 A total of 292 people received haloperidol in the trials. For oral haloperidol two studies allowed doses up to 100 mg/day but in most of the studies doses ranged from 2 to 45 mg/day. In most of the studies investigating intramuscular haloperidol, doses were between 1 to 35 mg/day, but one study had a maximum dose of 100mg/day. One study used a fixed dose (Ritter 1972 ‐ intramuscular haloperidol); all other studies used flexible doses. 
 
 4.6.2 Control intervention ‐ chlorpromazine 
 A total of 205 participants received chlorpromazine. Oral chlorpromazine doses usually ranged from 50 to 800 mg/day; but two studies administered a maximum dose of 1800 mg/day or 1600 mg/day respectively. For intramuscular chlorpromazine, doses varied between 25 and 350 mg/day; only one study had a maximum dose of 500 mg/day.

4.7 Outcomes 
 4.7.1 Acceptability and efficacy 
 The number of patients 'leaving the studies early', which can be used as a measure of acceptability of treatment, could be extracted from all trials. Many studies reported a global clinical judgement as to whether the participants were improved (Bailine 1981; Clark 1969; Gallant 1967; Man 1973; McCreadie 1977; Reschke 1974; Rompel 1978; Bi 1994 and Serafetinides 1972). Only Bi 1994 presented efficacy data based on the Brief Psychiatric Rating Scale (BPRS).

4.7.2 Adverse events 
 The most frequently recorded adverse effects were extrapyramidal side‐effects. Other relatively frequently reported adverse events were hypotension, sedation and GPT increase.

4.7.3 Outcome scales 
 Details of scales that provided usable data are shown below.

4.7.3.1 Global state 
 Clinical Global Impression ‐ CGI (Guy 1976) 
 A rating instrument 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 low severity or greater recovery.

4.7.3.2 Mental state 
 Brief Psychiatric Rating Scale ‐ BPRS (Overall 1962) 
 This scale is used to assess the severity of abnormal mental states. 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', ranging from 0 to 6 or 1 to 7. Scores can range from 0 to 108 or 18 to 126 respectively. High scores indicate more severe symptoms. The BPRS‐positive cluster comprises four items, which are conceptual disorganisation, suspiciousness, hallucinatory behaviour and unusual thought content. The BPRS‐negative cluster comprises only three items, which are emotional withdrawal, motor retardation, and blunted affect.

4.7.4 Missing outcomes 
 Many important outcomes such as costs, issues of hospital admission and satisfaction with care were not addressed in the included studies.

Risk of bias in included studies

1. Randomisation 
 Although all trials were described as 'randomised', no adequate information about allocation concealment was given. Therefore, all of the 14 included studies fell into category 'B' (unclear allocation concealment), because it was unknown how precisely the randomisation had been undertaken.

2. Blindness 
 Nine studies were conducted in a double‐blind fashion. Five studies used capsules, bottles or ampoules of identical appearance (Gallant 1967; Ritter 1972; Rompel 1978; Serafetinides 1972 and Shopsin 1969). In four studies there were no further details about the blinding methods (Clark 1969; Man 1973; McCreadie 1977 and Teja 1975). In three studies, blinding was not mentioned; it is probable that they were open studies (Bailine 1981; Bi 1994; Nishizono 1994). Reschke 1974 was initially double‐blind but later switched to open administration. Lempérière 1962 was an open‐label study.

3. Loss to follow up 
 All 14 studies described those participants who were unable to complete the trial protocol. Thirteen studies mentioned whether they discontinued due to adverse events or inefficacy. Only Nishizono 1994 described leaving early due to any reason. Clark 1969 only reported how many participants discontinued prematurely due to adverse events or inefficacy.

Effects of interventions

1. The search 
 The original search yielded 126 records, and we analysed the abstract of each identified reference. Fifty two possibly relevant results were ordered and inspected. We considered all these papers to sufficiently match the review's criteria to be mentioned in either the 'Included studies' (14 references) or 'Excluded studies' (38 references) section.

2. Comparison 1: HALOPERIDOL versus CHLORPROMAZINE 
 Fourteen studies compared haloperidol (n=292) with chlorpromazine (n=205).

2.1 Leaving the study early 
 The combination of 13 studies revealed that those randomised to haloperidol were significantly less likely to leave the study early for any reason (13 RCTs, n=476, RR 0.26 CI 0.08 to 0.82, NNT could not be calculated because the risk difference was not statistically significant). On considering specific reasons for leaving the study early, we found that fewer haloperidol treated participants left early due to both adverse events (13 RCTs, n=446, RR 0.3 CI 0.08 to 1.16) and inefficacy of treatment (13 RCTs, n=446, RR 0.36 CI 0.06 to 2.15); but these results were not statistically significant.

2.2 Global state 
 2.2.1 No clinically significant improvement ‐ as defined by the individual studies 
 Overall results tended to favour haloperidol (9 RCTs, n=400, RR 0.81 CI 0.64 to 1.04), but this difference was not statistically significant. As regards specific definitions of clinical response, only one study reported 'less than adequately improved' and found a superiority of haloperidol (n=29, RR 0.57 CI 0.34 to 0.97, NNT 3 CI 1 to 10). None of the studies considering 'less than moderate improvement' (4 RCTs, n=159, RR 0.65 CI 0.36 to 1.19), 'less than marked improvement' (2 RCTs, n=147, RR 0.95 CI 0.8 to 1.14), 'less than much improvement' (1 RCT, n=28, RR 1.08 CI 0.93 to 1.25), nor 'less than 40% BPRS reduction' (1 RCT, n=37, RR 0.84 CI 0.33 to 2.17) found significant differences between groups.

2.2.2 Average scores 
 One study considering the mean CGI severity score at endpoint reported no significant difference between oral haloperidol and oral chlorpromazine (n=37, WMD ‐0.2 CI to 0.98‐0.58).

2.3 Mental state: Average scores 
 Bi 1994 found no significant differences between haloperidol and chlorpromazine concerning the mean BPRS total score at endpoint (n=37, WMD ‐2.70 CI 7.28 to 1.88).

2.4 Adverse events 
 2.4.1 At least one adverse event 
 Clark 1969 found significant difference concerning the category 'at least one adverse event' (1 RCT, n=28, RR 0.92 CI 0.65 to 1.30).

2.4.2 Anticholinergic 
 No significant differences between haloperidol and chlorpromazine concerning anticholinergic adverse events were observed. Obstipation (1 RCT, n=28, RR 0.2 CI 0.01 to 3.82), dry mouth (2 RCTs, n=77, RR 0.68 CI 0.27 to 1.75) and blurred vision (1 RCT, n=38, RR 1 CI 0.16 to 6.38) did not occur significantly more often in either group.

2.4.3 Cardiovascular 
 Hypertension did not occur significantly more often among the haloperidol or chlorpromazine treated participants (2 RCTs, n=69, RR 1.5 CI 0.09 to 25.78). Hypotension was significantly more frequent in the chlorpromazine group (5 RCTs, n=175, RR 0.31 CI 0.11 to 0.88, NNH 7 CI 4 to 25).

2.4.4 Central nervous system 
 Only Serafetinides 1972 reported the adverse event 'excitement' but not a single participant suffered from this adverse event (RR not estimable). There was no significant difference in terms of sedation (5 RCTs, n=165, RR 0.49 CI 0.14 to 1.75). In Man 1973 one seizure occurred in the chlorpromazine group (1 RCT, n=30, RR 0.33 CI 0.01 to 7.58).

2.4.5 Endocrine ‐ galactorrhea 
 In Bi 1994 no significant difference in terms of galactorrhea was found (1 RCT, n=37, RR 1.88 CI 0.76 to 4.68).

2.4.6 Metabolic ‐ weight change 
 Four participants in the chlorpromazine group as oppose to none in the haloperidol group gained weight (2 RCTs, n=48, RR=0.11 CI 0.01 to 1.89). Weight loss occurred more often in the haloperidol group (six events compared to two, 1 RCT, n=28, RR=3 CI 0.73 to 12.39). None of these results was significant.

2.4.7 Movement disorders 
 Haloperidol was associated with significantly more participants suffering from 'at least one extrapyramidal side effect' than chlorpromazine (6 RCTs, n=37, RR 2.2 CI 1.1 to 4.4, NNH 5 CI 3 to 33). The use of antiparkinson medication showed no significant difference between haloperidol and chlorpromazine (1 RCT, n=20, RR 3 CI 0.37 to 24.17). Significant results could be observed neither for akathisia (3 RCTs, n=86, RR 1.9 CI 0.29 to 12.55), dystonia (2 RCT, n=58, RR 3.6 CI 0.63 to 20.56), nor for parkinsonism (1 RCT, n=38, RR 4 CI 0.49 to 32.57).

Bi 1994 used a rating scale to measure one extrapyramidal side effects and found a significant difference favouring chlorpromazine (1 RCT, n=37, WMD 2.10 CI 1.38 to 2.82).

2.4.8 Skin 
 There were no significant differences in terms of photosensitivity (2 RCTs, n=48, RR 0.17 CI 0.02 to 1.3), oedema (1 RCT, n=28, no events, thus RR not estimable), and rash (1 RCT, n=38, RR 3 CI 0.13 to 69.31).

2.4.9 Laboratory data 
 No significant differences concerning the following laboratory values were found: the liver enzyme GPT (4 RCTs, n=156, RR 0.20 CI 0.04 to 1.12), abnormal blood count (no events), leukocytosis (no events), leukopenia (1 RCT, n=50, RR 3.0 CI 0.13 to 70.3), white blood cells in urine (1 RCT, n=28, RR 3.0 CI 0.13 to 67.91), glucosyria (no events), urea increase (1 RCT, n=28,nhnb RR 3.0 CI 0.13 to 67.91), high cholesterol (no events) and hypertriglyceridemia (1 RCT, n=28, RR 1.0 CI 0.07 to 14.45).

In terms of prolactin increase, in Bi 1994, 17 out of 17 haloperidol treated participants and 17 out of 20 chlorpromazine treated participants had increased prolactine levels (1 RCT, n=37, RR1.18 CI 0.98 to 1.41). Again in Bi 1994 the mean prolactin levels at endpoint no did not differ significantly (1 RCT, n=37, WMD 21.1 CI 7.47 to 34.73).

2.5 Missing outcomes 
 Data on important outcomes such as hospital admission, relapse, satisfaction with care, social functioning, family burden and employment were not available.

2.6 Publication bias 
 Due to the small number of RCTs, funnel plots made sense only for the outcome 'no clinically important improvement'. The funnel plot looked asymmetrical, indicating that some studies in favour of chlorpromazine may not have been published; but due to the small number of studies this finding is not conclusive.

3. Subgroups: INTRAMUSCULAR HALOPERIDOL versus INTRAMUSCULAR CHLORPROMAZINE and ORAL HALOPERIDOL versus ORAL CHLORPROMAZINE

Five studies compared intramuscular haloperidol (n=109) with intramuscular chlorpromazine (n=84). Nine studies compared oral haloperidol (n=183) with oral chlorpromazine (n=121).

In both subgroups (intramuscular and oral) the direction of effects were the same as in the overall analysis. However, statistically significant differences were even rarer than in the overall results. The only significant results in the oral subgroup were 'leaving the study early for any reason', showing that more participants discontinued prematurely in the chlorpromazine groups (8 RCTs, n=283, RR 0.15 CI 0.04 to 0.64, NNT could not be calculated because the risk difference was not statistically significant), and 'less than adequate improvement' which favoured haloperidol in one study (1 RCT, n=29, RR 0.57 CI 0.34 to 0.97, NNT 2 CI 1 to 10). In the subgroup of studies comparing intramuscular formulations, none of the results were statistically significant.

Discussion

1. General remarks 
 1.1 Generalisability 
 Fourteen studies with a total of 794 participants were included in this review. This number is not very high, given that both haloperidol and chlorpromazine are the prototype antipsychotic drugs for the treatment of schizophrenia.

Many studies did not use operational criteria to diagnose the participants, but 'clinical diagnoses'. As the definition of schizophrenia has changed several times since the earliest studies were conducted in the 1960s, the use of such clinical classifications may have led to the inclusion of some participants who would nowadays not be considered to have schizophrenia. Otherwise the participants seemed to be comparable to those in clinical practice as regards some of their characteristics and duration of illness. Therefore the studies might well reflect clinical practice.

Regarding interventions and doses, the studies in this review presented strategies that are similar to those used today but with a tendency towards higher dosages than are currently administered in most settings. In most of the studies investigating intramuscular haloperidol, doses were between 1 and 35 mg/day; only one study had a maximum dose of 100 mg/day. For oral haloperidol, two studies had doses up to 100 mg/day but in most of the studies the doses ranged between 2 and 45 mg/day. Oral chlorpromazine doses usually ranged from 50 to 800 mg/day; only two studies administered a maximum dose of 1800 mg/day and 1600mg/day respectively. For intramuscular chlorpromazine, doses varied between 25 and 350 mg/day; one study had a maximum dose of 500 mg/day.

Another general problem is that the included studies mainly fell into the 'short term' category. As schizophrenia is often a chronic illness, the focus on short‐term studies is especially disadvantageous. Only one trial had a duration of more than six months (Teja 1975). More long‐term studies are needed in order to be able to analyze the long‐term effects of haloperidol and chlorpromazine. Finally, outcomes such as hospital admission, relapse, satisfaction with care, social functioning, family burden and employment, which are of high clinical relevance, are currently completely lacking.

1.2 Quality of reporting 
 The included studies were mostly reported deficiently, e.g. standard deviations were often missing from the data description. All included studies had to be classified as category B (unclear allocation concealment). This may be a source of bias.

2. Comparison 1: HALOPERIDOL versus CHLORPROMAZINE

When drawing conclusions on the effects of haloperidol compared to chlorpromazine, the low number of studies and participants included must be kept in mind.

2.1 Acceptability of treatment (leaving the studies early) 
 Very few participants left the studies early, which is remarkable, because such low rates are quite uncommon in studies conducted nowadays. Studies on new, so‐called 'atypical' antipsychotics often have drop‐out rates around 40‐50% (Wahlbeck 2001).

The participants randomised to haloperidol were significant less likely to leave the studies early for any reason than the participants in the chlorpromazine group. This difference may have been caused both by more discontinuations due to adverse events and more discontinuations due to inefficacy of treatment in the chlorpromazine groups. Haloperidol appears to be more satisfactory for people with schizophrenia than chlorpromazine, when considering the acceptability of treatment assessed indirectly by the number of people leaving early for any reason.

2.3 Clinical improvement 
 Haloperidol appears to be slightly more effective than chlorpromazine as a treatment for schizophrenia, but the results only showed a trend favouring haloperidol which did not turn out to be significant. Only one study used the CGI or the BPRS total score at endpoint. These results were based on too few patients to be meaningful.

2.4 Tolerability 
 According to the analysis, the main disadvantage of haloperidol was a higher frequency of movement disorders, at least at the doses used. Concerning the criterion 'at least one extrapyramidal side effect' there were significantly more adverse events in the haloperidol group, although the single adverse events 'use of antiparkinson medication' (one study), 'akathisia' (three studies), 'dystonia' (two studies) and 'parkinsonism' (one study) did not reach significance level. Limited power may have contributed to the impossibility of finding significant differences.

In terms of cardiovascular adverse events, the results confirmed a clinical impression that low potency antipsychotics such as chlorpromazine cause significantly more hypotension than high‐potency drugs such as haloperidol.

Although a number of other outcomes related to tolerability were analysed (at least one adverse event, anticholinergic events, excitement, galactorrhea, laboratory data, prolactin increase, prolactin level at endpoint, sedation, seizures, skin problems and weight change), the results did not show significant differences between groups. This does not necessarily mean that there are no differences between the two compounds, because again the numbers of studies reporting on such adverse events were usually low.

3. Subgroups: INTRAMUSCULAR HALOPERIDOL versus INTRAMUSCULAR CHLORPROMAZINE and ORAL HALOPERIDOL versus ORAL CHLORPROMAZINE

Both subgroups showed the same direction of effects as the overall analysis, although in the subgroup analysing intramuscular formulations, none of the results turned out to be significant. In the subgroup of studies examining oral formulations the only significant differences were 'leaving the study early for any reason', where again fewer events occurred in the haloperidol group, and 'less than adequately improved', where a single study found a superiority of haloperidol. Whether this lack of statistically significant differences was simply due to reduced power needs to be established.

Authors' conclusions

Implications for practice.

1. For people with schizophrenia 
 Currently available data suggest that haloperidol and chlorpromazine are similarly effective for treatment of schizophrenia. But at least at higher doses haloperidol seems to be associated with more extrapyramidal side effects while hypotension appears to occur more frequently when chlorpromazine is used. Overall, haloperidol may also be more acceptable for those afflicted by the illness. 
 
 2. For clinicians 
 Although haloperidol and chlorpromazine have been the standard drugs in the treatment of schizophrenia for many years, there are relatively few studies investigating their efficacy and tolerability head to head. There were no statistically significant efficacy differences between the two compounds, but the side‐effect profiles differ. Extrapyramidal side effects seem to be more frequent when haloperidol is used, in comparison to which more hypotension is associated with chlorpromazine treatment.

3. For managers or policy makers 
 Haloperidol and chlorpromazine are relatively cheap drugs, at least compared to new, so called 'atypical' antipsychotics. From the results of this review no definite conclusions on the relative efficacy of the two compounds can be drawn, but their side‐effect profiles differ.

Implications for research.

1. General 
 The vast majority of the included studies were carried out at least 20 years ago, which explains the lack of modern standards regarding methods and reporting. This general limitation must be kept in mind when analyzing the data and needs to be improved by further studies. The fact that haloperidol and chlorpromazine are still standard drugs administered in the treatment of schizophrenia, and that they are thus also potential gold standards for the evaluation of new antipsychotic drugs, justifies further research to compare the properties of these compounds.

2. Specific 
 Future studies, perhaps set in situations where there are only these two drugs available, should be larger and of longer term and use clinically meaningful outcomes such as such hospital admission, relapse, satisfaction with care, social functioning, family burden and employment (Table 1).

1. Suggested design of study.

Methods	Participants	Interventions	Outcomes	Notes
Allocation: centralised sequence generation with table of random numbers or computer generated code, stratified by severity of illness, sequence concealed till interventions assigned. In the context of limited provision, randomisation may be the only equitable way of distributing care. Blindness: double ‐ tested. Duration: 1 year. Design: parallel groups.	Diagnosis: it may be preferred not to use diagnostic categories such as DSM IV and just to include those whose mental health problem is designated as schizophrenia or psychosis and depression. Perhaps people for whom there is no other choice than haloperidol or chlorpromazine. N=300. Age˜any. Sex: either. Setting: anywhere. History: any.	1. Haloperidol: dose and choice at clinician's and patient's discretion. N=150. 2. Chlorpomazine: dose ‐ clinician's and patient's discretion. N=150.	Qualtiy of life: healthy days. Service outcomes: days in hospital, time attending psychiatric outpatient clinic. Satisfaction with care: patients/carers. Global state: CGI.*** Mental state: CGI, relapse.** Functioning: engagement with services, leaving the study early, living independently. Adverse effects: including mortality. Economic outcomes: cost‐effectiveness and cost‐benefit.	* size of study to detect a 10% difference in improvement with 80% certainity. *** Primary outcome. If scales are used to measure outcome then there should be binary cut off points, defined before study start, of clinically important improvement.

What's new

Date	Event	Description
31 January 2013	Amended	Contact details updated.

History

Protocol first published: Issue 3, 2005
 Review first published: Issue 1, 2008

Date	Event	Description
17 October 2012	Amended	Contact details updated.
11 May 2011	Amended	Contact details updated.
15 February 2010	Amended	Contact details updated.
15 April 2009	Amended	Plain Language Summary added
30 October 2008	Amended	Converted to new review format.

Data and analyses

Comparison 1. HALOPERIDOL versus CHLORPROMAZINE.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Leaving the study early	14		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.1 any reason	13	476	Risk Ratio (M‐H, Random, 95% CI)	0.26 [0.08, 0.82]
1.2 due to adverse events	13	446	Risk Ratio (M‐H, Random, 95% CI)	0.30 [0.08, 1.16]
1.3 due to inefficacy	13	446	Risk Ratio (M‐H, Random, 95% CI)	0.36 [0.06, 2.15]
2 Global state: 1. No clinically significant improvement ‐ as defined by the individual studies	9	400	Risk Ratio (M‐H, Random, 95% CI)	0.81 [0.64, 1.04]
2.1 less than moderate improvement	4	159	Risk Ratio (M‐H, Random, 95% CI)	0.65 [0.36, 1.19]
2.2 less than 40% BPRS reduction	1	37	Risk Ratio (M‐H, Random, 95% CI)	0.84 [0.33, 2.17]
2.3 less than adequately improved	1	29	Risk Ratio (M‐H, Random, 95% CI)	0.57 [0.34, 0.97]
2.4 less than markedly improved	2	147	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.80, 1.14]
2.5 less than much improved	1	28	Risk Ratio (M‐H, Random, 95% CI)	1.07 [0.89, 1.30]
3 Global state: 2. Average score at endpoint (CGI, high = poor)	1	37	Mean Difference (IV, Random, 95% CI)	‐0.20 [‐0.98, 0.58]
4 Mental state: Average score at endpoint (BPRS, high = poor)	1	37	Mean Difference (IV, Random, 95% CI)	‐2.70 [‐7.28, 1.88]
5 Adverse events: 1. General ‐ at least one adverse event	1	28	Risk Ratio (M‐H, Random, 95% CI)	0.92 [0.65, 1.30]
6 Adverse events: 2. Specific ‐ anticholinergic	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
6.1 blurred vision	1	38	Risk Ratio (M‐H, Random, 95% CI)	1.0 [0.16, 6.38]
6.2 constipation	1	28	Risk Ratio (M‐H, Random, 95% CI)	0.2 [0.01, 3.82]
6.3 dry mouth	2	77	Risk Ratio (M‐H, Random, 95% CI)	0.68 [0.27, 1.75]
7 Adverse events: 3. Specific ‐ cardiovascular	6		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
7.1 hypertension	2	69	Risk Ratio (M‐H, Random, 95% CI)	1.50 [0.09, 25.78]
7.2 hypotension	5	175	Risk Ratio (M‐H, Random, 95% CI)	0.31 [0.11, 0.88]
8 Adverse events: 4. Specific ‐ central nervous system	6		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
8.1 excitement	1	28	Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
8.2 sedation	5	165	Risk Ratio (M‐H, Random, 95% CI)	0.49 [0.14, 1.75]
8.3 seizure	1	30	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.01, 7.58]
9 Adverse events: 5. Specific ‐ endocrine ‐ galactorrhoea	1	37	Risk Ratio (M‐H, Random, 95% CI)	1.88 [0.76, 4.68]
10 Adverse events: 6. Specific ‐ metabolic ‐ weight change	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
10.1 gain	2	48	Risk Ratio (M‐H, Random, 95% CI)	0.11 [0.01, 1.89]
10.2 loss	1	28	Risk Ratio (M‐H, Random, 95% CI)	3.00 [0.73, 12.39]
11 Adverse events: 7a. Specific ‐ movement disorders	7		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
11.1 general ‐ use of antiparkinson medication	1	20	Risk Ratio (M‐H, Random, 95% CI)	3.00 [0.37, 24.17]
11.2 specific ‐ akathisia	3	86	Risk Ratio (M‐H, Random, 95% CI)	1.91 [0.29, 12.55]
11.3 specific ‐ dystonia	2	58	Risk Ratio (M‐H, Random, 95% CI)	3.60 [0.63, 20.56]
11.4 specific ‐ >1 extrapyramidal side effect	6	212	Risk Ratio (M‐H, Random, 95% CI)	2.20 [1.11, 4.40]
11.5 specific ‐ parkinsonism	1	38	Risk Ratio (M‐H, Random, 95% CI)	4.0 [0.49, 32.57]
12 Adverse events: 7b. Specific ‐ movement disorders ‐ EPS score (high=poor)	1	37	Mean Difference (IV, Fixed, 95% CI)	2.10 [1.38, 2.82]
13 Adverse events: 8. Specific ‐ skin problems	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
13.1 oedema	1	28	Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
13.2 photosensitivity	2	48	Risk Ratio (M‐H, Random, 95% CI)	0.17 [0.02, 1.30]
13.3 rash	1	38	Risk Ratio (M‐H, Random, 95% CI)	3.00 [0.13, 69.31]
14 Adverse events: 9a. Laboratory data (binary)	6		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
14.1 haematology ‐ abnormal blood count	1	20	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
14.2 haematology ‐ leukocytosis	1	28	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
14.3 haematology ‐ leukopenia	1	50	Risk Ratio (M‐H, Fixed, 95% CI)	3.0 [0.13, 70.30]
14.4 liver function ‐ GPT increase	4	156	Risk Ratio (M‐H, Fixed, 95% CI)	0.2 [0.04, 1.12]
14.5 metabolic ‐ cholesterol elevated	1	28	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
14.6 metabolic ‐ glucosyria	1	28	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
14.7 metabolic ‐ prolactin increased	1	37	Risk Ratio (M‐H, Fixed, 95% CI)	1.17 [0.95, 1.43]
14.8 metabolic ‐ triglycerides elevated	1	28	Risk Ratio (M‐H, Fixed, 95% CI)	1.0 [0.07, 14.45]
14.9 renal function ‐ urea increase	1	28	Risk Ratio (M‐H, Fixed, 95% CI)	3.0 [0.13, 67.91]
14.10 renal function ‐ white blood cells in urine	1	28	Risk Ratio (M‐H, Fixed, 95% CI)	3.0 [0.13, 67.91]
15 Adverse events: 9b. Laboratory data (continuous)	1	37	Mean Difference (IV, Random, 95% CI)	21.10 [7.47, 34.73]
15.1 metabolic ‐ prolactin levels (high = poor)	1	37	Mean Difference (IV, Random, 95% CI)	21.10 [7.47, 34.73]

1.1. Analysis.

Comparison 1 HALOPERIDOL versus CHLORPROMAZINE, Outcome 1 Leaving the study early.

1.2. Analysis.

Comparison 1 HALOPERIDOL versus CHLORPROMAZINE, Outcome 2 Global state: 1. No clinically significant improvement ‐ as defined by the individual studies.

1.3. Analysis.

Comparison 1 HALOPERIDOL versus CHLORPROMAZINE, Outcome 3 Global state: 2. Average score at endpoint (CGI, high = poor).

1.4. Analysis.

Comparison 1 HALOPERIDOL versus CHLORPROMAZINE, Outcome 4 Mental state: Average score at endpoint (BPRS, high = poor).

1.5. Analysis.

Comparison 1 HALOPERIDOL versus CHLORPROMAZINE, Outcome 5 Adverse events: 1. General ‐ at least one adverse event.

1.6. Analysis.

Comparison 1 HALOPERIDOL versus CHLORPROMAZINE, Outcome 6 Adverse events: 2. Specific ‐ anticholinergic.

1.7. Analysis.

Comparison 1 HALOPERIDOL versus CHLORPROMAZINE, Outcome 7 Adverse events: 3. Specific ‐ cardiovascular.

1.8. Analysis.

Comparison 1 HALOPERIDOL versus CHLORPROMAZINE, Outcome 8 Adverse events: 4. Specific ‐ central nervous system.

1.9. Analysis.

Comparison 1 HALOPERIDOL versus CHLORPROMAZINE, Outcome 9 Adverse events: 5. Specific ‐ endocrine ‐ galactorrhoea.

1.10. Analysis.

Comparison 1 HALOPERIDOL versus CHLORPROMAZINE, Outcome 10 Adverse events: 6. Specific ‐ metabolic ‐ weight change.

1.11. Analysis.

Comparison 1 HALOPERIDOL versus CHLORPROMAZINE, Outcome 11 Adverse events: 7a. Specific ‐ movement disorders.

1.12. Analysis.

Comparison 1 HALOPERIDOL versus CHLORPROMAZINE, Outcome 12 Adverse events: 7b. Specific ‐ movement disorders ‐ EPS score (high=poor).

1.13. Analysis.

Comparison 1 HALOPERIDOL versus CHLORPROMAZINE, Outcome 13 Adverse events: 8. Specific ‐ skin problems.

1.14. Analysis.

Comparison 1 HALOPERIDOL versus CHLORPROMAZINE, Outcome 14 Adverse events: 9a. Laboratory data (binary).

1.15. Analysis.

Comparison 1 HALOPERIDOL versus CHLORPROMAZINE, Outcome 15 Adverse events: 9b. Laboratory data (continuous).

Comparison 2. SUBGROUP: HALOPERIODL INTRAMUSCULAR versus CHLORPROMAZINE INTRAMUSCULAR.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Leaving the study early	5		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.1 any reason	5	193	Risk Ratio (M‐H, Random, 95% CI)	0.66 [0.10, 4.18]
1.2 adverse events	5	193	Risk Ratio (M‐H, Random, 95% CI)	0.45 [0.05, 3.83]
1.3 inefficacy	5	193	Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2 Global state: No clinically significant improvement ‐ as defined by the individual studies	4		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
2.1 less than moderate improvement	4	159	Risk Ratio (M‐H, Random, 95% CI)	0.65 [0.36, 1.19]
3 Adverse events: 1. Specific ‐ anticholinergic	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
3.1 dry mouth	1	39	Risk Ratio (M‐H, Random, 95% CI)	1.38 [0.17, 10.93]
4 Adverse events: 2. Specific ‐ cardiovascular	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
4.1 hypertension	1	40	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.01, 7.72]
4.2 hypotension	3	109	Risk Ratio (M‐H, Random, 95% CI)	0.37 [0.10, 1.41]
4.3 tachycardia	1	40	Risk Ratio (M‐H, Random, 95% CI)	0.2 [0.01, 3.92]
5 Adverse events: 3. Specific ‐ central nervous system	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
5.1 sedation	2	79	Risk Ratio (M‐H, Random, 95% CI)	0.70 [0.37, 1.33]
5.2 seizure	1	30	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.01, 7.58]
6 Adverse events: 4. Specific ‐ movement disorders ‐ at least one EPS	3	109	Risk Ratio (M‐H, Random, 95% CI)	3.49 [0.84, 14.44]
7 Adverse events: 5. Laboratory data	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
7.1 haematology ‐ leukopenia	1	50	Risk Ratio (M‐H, Random, 95% CI)	3.00 [0.13, 70.30]
7.2 liver function ‐ GPT increase	2	90	Risk Ratio (M‐H, Random, 95% CI)	0.25 [0.03, 2.22]

2.1. Analysis.

Comparison 2 SUBGROUP: HALOPERIODL INTRAMUSCULAR versus CHLORPROMAZINE INTRAMUSCULAR, Outcome 1 Leaving the study early.

2.2. Analysis.

Comparison 2 SUBGROUP: HALOPERIODL INTRAMUSCULAR versus CHLORPROMAZINE INTRAMUSCULAR, Outcome 2 Global state: No clinically significant improvement ‐ as defined by the individual studies.

2.3. Analysis.

Comparison 2 SUBGROUP: HALOPERIODL INTRAMUSCULAR versus CHLORPROMAZINE INTRAMUSCULAR, Outcome 3 Adverse events: 1. Specific ‐ anticholinergic.

2.4. Analysis.

Comparison 2 SUBGROUP: HALOPERIODL INTRAMUSCULAR versus CHLORPROMAZINE INTRAMUSCULAR, Outcome 4 Adverse events: 2. Specific ‐ cardiovascular.

2.5. Analysis.

Comparison 2 SUBGROUP: HALOPERIODL INTRAMUSCULAR versus CHLORPROMAZINE INTRAMUSCULAR, Outcome 5 Adverse events: 3. Specific ‐ central nervous system.

2.6. Analysis.

Comparison 2 SUBGROUP: HALOPERIODL INTRAMUSCULAR versus CHLORPROMAZINE INTRAMUSCULAR, Outcome 6 Adverse events: 4. Specific ‐ movement disorders ‐ at least one EPS.

2.7. Analysis.

Comparison 2 SUBGROUP: HALOPERIODL INTRAMUSCULAR versus CHLORPROMAZINE INTRAMUSCULAR, Outcome 7 Adverse events: 5. Laboratory data.

Comparison 3. SUBGROUP: HALOPERIDOL ORAL versus CHLORPROMAZINE ORAL.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Leaving the study early	9		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.1 any reason	8	283	Risk Ratio (M‐H, Random, 95% CI)	0.15 [0.04, 0.64]
1.2 due to adverse events	8	253	Risk Ratio (M‐H, Random, 95% CI)	0.23 [0.04, 1.32]
1.3 due to inefficacy	8	253	Risk Ratio (M‐H, Random, 95% CI)	0.36 [0.06, 2.15]
2 Global state: 1. No clinically significant improvement ‐ as defined by the individual studies	5	241	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.73, 1.13]
2.1 less than 40% BPRS reduction	1	37	Risk Ratio (M‐H, Random, 95% CI)	0.84 [0.33, 2.17]
2.2 less than adequately improved	1	29	Risk Ratio (M‐H, Random, 95% CI)	0.57 [0.34, 0.97]
2.3 less than markedly improved	2	147	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.80, 1.14]
2.4 less than much improved	1	28	Risk Ratio (M‐H, Random, 95% CI)	1.07 [0.89, 1.30]
3 Global state: 2. Average score at endpoint (CGI, high = poor)	1	37	Mean Difference (IV, Random, 95% CI)	‐0.20 [‐0.98, 0.58]
4 Mental state: Average score at endpoint (BPRS, high = poor)	1	37	Mean Difference (IV, Random, 95% CI)	‐2.70 [‐7.28, 1.88]
5 Adverse events: 1. General ‐ at least one adverse event	1	28	Risk Ratio (M‐H, Random, 95% CI)	0.92 [0.65, 1.30]
6 Adverse events: 2. Specific ‐ anticholinergic	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
6.1 blurred vision	1	38	Risk Ratio (M‐H, Random, 95% CI)	1.0 [0.16, 6.38]
6.2 constipation	1	28	Risk Ratio (M‐H, Random, 95% CI)	0.2 [0.01, 3.82]
6.3 dry mouth	1	38	Risk Ratio (M‐H, Random, 95% CI)	0.57 [0.20, 1.63]
7 Adverse events: 3. Specific ‐ cardiovascular	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
7.1 hypertension	1	29	Risk Ratio (M‐H, Random, 95% CI)	6.07 [0.32, 116.33]
7.2 hypotension	2	66	Risk Ratio (M‐H, Random, 95% CI)	0.22 [0.04, 1.24]
8 Adverse events: 4. Specific ‐ central nervous system	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
8.1 excitement	1	28	Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
8.2 sedation	3	86	Risk Ratio (M‐H, Random, 95% CI)	0.35 [0.02, 5.73]
9 Adverse events: 5. Specific ‐ endocrine ‐ galactorrhoea	1	37	Risk Ratio (M‐H, Random, 95% CI)	1.88 [0.76, 4.68]
10 Adverse events: 6. Specific ‐ metabolic ‐ weight change	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
10.1 gain	2	48	Risk Ratio (M‐H, Random, 95% CI)	0.11 [0.01, 1.89]
10.2 loss	1	28	Risk Ratio (M‐H, Random, 95% CI)	3.00 [0.73, 12.39]
11 Adverse events: 7a. Specific ‐ movement disorders	4		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
11.1 general ‐ use of antiparkinson medication	1	20	Risk Ratio (M‐H, Random, 95% CI)	3.00 [0.37, 24.17]
11.2 specific ‐ akathisia	3	86	Risk Ratio (M‐H, Random, 95% CI)	1.91 [0.29, 12.55]
11.3 specific ‐ dystonia	2	58	Risk Ratio (M‐H, Random, 95% CI)	3.60 [0.63, 20.56]
11.4 specific ‐ >1 extrapyramidal side effect	3	103	Risk Ratio (M‐H, Random, 95% CI)	1.96 [0.93, 4.10]
11.5 specific ‐ parkinsonism	1	38	Risk Ratio (M‐H, Random, 95% CI)	4.0 [0.49, 32.57]
12 Adverse events: 7b. Specific ‐ movement disorders ‐ EPS score (high=poor)	1	37	Mean Difference (IV, Fixed, 95% CI)	2.10 [1.38, 2.82]
13 Adverse events: 8. Specific ‐ skin problems	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
13.1 oedema	1	28	Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
13.2 photosensitivity	2	48	Risk Ratio (M‐H, Random, 95% CI)	0.17 [0.02, 1.30]
13.3 rash	1	38	Risk Ratio (M‐H, Random, 95% CI)	3.00 [0.13, 69.31]
14 Adverse events: 9a. Laboratory data (binary)	4		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
14.1 haematology ‐ abnormal blood count	1	20	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
14.2 haematology ‐ leukocytosis	1	28	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
14.3 liver function ‐ GPT increase	2	66	Risk Ratio (M‐H, Fixed, 95% CI)	0.14 [0.01, 2.59]
14.4 metabolic ‐ cholesterol elevated	1	28	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
14.5 metabolic ‐ glucosyria	1	28	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
14.6 metabolic ‐ prolactin increased	1	37	Risk Ratio (M‐H, Fixed, 95% CI)	1.17 [0.95, 1.43]
14.7 metabolic ‐ triglycerides elevated	1	28	Risk Ratio (M‐H, Fixed, 95% CI)	1.0 [0.07, 14.45]
14.8 renal function ‐ urea increase	1	28	Risk Ratio (M‐H, Fixed, 95% CI)	3.0 [0.13, 67.91]
14.9 renal function ‐ white blood cells in urine	1	28	Risk Ratio (M‐H, Fixed, 95% CI)	3.0 [0.13, 67.91]
15 Adverse events: 9b. Laboratory data (continuous)	1	37	Mean Difference (IV, Random, 95% CI)	21.10 [7.47, 34.73]
15.1 metabolic ‐ prolactin levels (high = poor)	1	37	Mean Difference (IV, Random, 95% CI)	21.10 [7.47, 34.73]

3.1. Analysis.

Comparison 3 SUBGROUP: HALOPERIDOL ORAL versus CHLORPROMAZINE ORAL, Outcome 1 Leaving the study early.

3.2. Analysis.

Comparison 3 SUBGROUP: HALOPERIDOL ORAL versus CHLORPROMAZINE ORAL, Outcome 2 Global state: 1. No clinically significant improvement ‐ as defined by the individual studies.

3.3. Analysis.

Comparison 3 SUBGROUP: HALOPERIDOL ORAL versus CHLORPROMAZINE ORAL, Outcome 3 Global state: 2. Average score at endpoint (CGI, high = poor).

3.4. Analysis.

Comparison 3 SUBGROUP: HALOPERIDOL ORAL versus CHLORPROMAZINE ORAL, Outcome 4 Mental state: Average score at endpoint (BPRS, high = poor).

3.5. Analysis.

Comparison 3 SUBGROUP: HALOPERIDOL ORAL versus CHLORPROMAZINE ORAL, Outcome 5 Adverse events: 1. General ‐ at least one adverse event.

3.6. Analysis.

Comparison 3 SUBGROUP: HALOPERIDOL ORAL versus CHLORPROMAZINE ORAL, Outcome 6 Adverse events: 2. Specific ‐ anticholinergic.

3.7. Analysis.

Comparison 3 SUBGROUP: HALOPERIDOL ORAL versus CHLORPROMAZINE ORAL, Outcome 7 Adverse events: 3. Specific ‐ cardiovascular.

3.8. Analysis.

Comparison 3 SUBGROUP: HALOPERIDOL ORAL versus CHLORPROMAZINE ORAL, Outcome 8 Adverse events: 4. Specific ‐ central nervous system.

3.9. Analysis.

Comparison 3 SUBGROUP: HALOPERIDOL ORAL versus CHLORPROMAZINE ORAL, Outcome 9 Adverse events: 5. Specific ‐ endocrine ‐ galactorrhoea.

3.10. Analysis.

Comparison 3 SUBGROUP: HALOPERIDOL ORAL versus CHLORPROMAZINE ORAL, Outcome 10 Adverse events: 6. Specific ‐ metabolic ‐ weight change.

3.11. Analysis.

Comparison 3 SUBGROUP: HALOPERIDOL ORAL versus CHLORPROMAZINE ORAL, Outcome 11 Adverse events: 7a. Specific ‐ movement disorders.

3.12. Analysis.

Comparison 3 SUBGROUP: HALOPERIDOL ORAL versus CHLORPROMAZINE ORAL, Outcome 12 Adverse events: 7b. Specific ‐ movement disorders ‐ EPS score (high=poor).

3.13. Analysis.

Comparison 3 SUBGROUP: HALOPERIDOL ORAL versus CHLORPROMAZINE ORAL, Outcome 13 Adverse events: 8. Specific ‐ skin problems.

3.14. Analysis.

Comparison 3 SUBGROUP: HALOPERIDOL ORAL versus CHLORPROMAZINE ORAL, Outcome 14 Adverse events: 9a. Laboratory data (binary).

3.15. Analysis.

Comparison 3 SUBGROUP: HALOPERIDOL ORAL versus CHLORPROMAZINE ORAL, Outcome 15 Adverse events: 9b. Laboratory data (continuous).

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Bailine 1981.

Methods	Allocation: randomised, no further details. Blindness: not mentioned, probably open. Duration: 72 hours. Design: parallel.
Participants	Diagnosis: schizophrenia or schizoaffective disorder, Research Diagnostic Criteria, clinical diagnosis. History: Participants were either violent or assaultive or too psychotically disorganized to be maintained on a unit without significant extra supervision. N= 34. Age: mean 25.2 years. Sex: 26 males, 8 females. Setting: inpatients. Exclusion criteria: not mentioned.
Interventions	1. I.m. haloperidol: dose range 5‐100 mg/day. N= 20. 2. I.m. chlorpromazine: dose range 50‐500 mg/day. N= 14.
Outcomes	Leaving the study early. Unable to use: GAS, Nursing Observation Scale for Inpatient Evaluation, BPRS, Simpson‐Angus‐Scale (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Bi 1994.

Methods	Allocation: randomised, no further details. Blindness: not mentioned, probably open. Duration: 4 weeks. Design: parallel.
Participants	Diagnosis: schizophrenia, RDC and CCMD‐2 criteria (Chinese classification of mental disorders). History: a part of the patients was antipsychotic naive, no further details. N= 50. Age: 18‐45 years. Sex: 24 males, 26 females. Setting: not mentioned. Exclusion criteria: other mental disorders, severe physical diseases.
Interventions	1. Oral haloperidol: dose range 5‐100 mg/day. N= 20. 2. Oral chlorpromazine: dose range 50‐500 mg/day. N= 14.
Outcomes	Leaving the study early. Less than 40% BPRS reduction. Global state: mean CGI at endpoint. Mental state: mean BPRS at endpoint. Adverse events: EPS (mean value of used EPS‐ score, number of patients with at least one EPS), serum‐ Prolactine (mean serum‐PRL at endpoint, number of patients over the normal cut‐off of prolactin), galactorrhea (number of patients with galactorrhea at endpoint).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Clark 1969.

Methods	Allocation: randomised, no further details. Blindness: double, no further details. Duration: 12 weeks. Design: parallel, single‐centre.
Participants	Diagnosis: chronic schizo‐ phrenia, clinical diagnosis. History: Haloperidol group: participants with previous treatment 93%, subjects with previous hospitalization 79%, subjects currently hospitalized more than one year 93%, subjects with history of family mental illness 93%. Chlorpromazine group: subjects with previous treatment 64%, subjects with previous hospitalization 71%, subjects currently hospitalized more than one year 93%, subjects with history of family mental illness 21%. Placebo group: subjects with previous treatment 92%, subjects with previous hospitalization 85%, subjects currently hospitalized more than one year 100%, subjects with history of family mental illness 31%. N= 43. Age: 22‐61 years. Sex: 18 males, 26 females. Setting: inpatients. Exclusion criteria: not mentioned.
Interventions	1. Haloperidol: dose range 3‐15 mg/day. N= 14. 2. Chlorpromazine: dose range 200‐1600 mg/day. N= 14. 3. Placebo: N= 16.
Outcomes	Leaving the study early due to adverse events or inefficacy. Adverse events: At least one side effect. Unable to use: Global state: CGI at endpoint, but no results given, no usable data. Mental state: BPRS at endpoint, but no results given, no usable data. Nurses observation scale: no usable data. Clinical improvement: no usable data. Laboratory findings (hereby most cited: hematology), adverse behavioral effects, adverse effects concerning central nervous system (rigidity as most cited symptom), weight loss and other, not further specified miscellaneous side‐effects (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Gallant 1967.

Methods	Allocation: randomised, prearranged randomisation procedure. Blindness: double, identical capsules. Duration: 4 weeks. Design: parallel, single‐centre.
Participants	Diagnosis: acute schizophrenia, clinical diagnosis. History: first admissions or admission no more than twice previously with a history of excellent remission between acute schizophrenic episodes. N= 58. Age: mean 33.4 years. Sex: 30 males, 28 females. Setting: inpatients. Exclusion criteria: not mentioned.
Interventions	1. Trifluperidol: increasing dose, maximum 4 mg/day. N= 20. 2. Haloperidol, increasing dose, maximum 16 mg/day. N= 19. 3. Chlorpromazine: increasing dose, maximum 800 mg/day. N= 19.
Outcomes	Leaving the study early. Less than markedly improved. Adverse events: At least one EPS, akathisia, dystonia, parkinsonism, dry mouth, blurred vision, hypotension, GPT increase, sedation, rash. Unable to use: Beckomberga scale, MACC Behavioral Adjustment scale, Tulane test battery (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Lempérière 1962.

Methods	Allocation: randomised, no further details. Blindness: open. Duration: 8 weeks. Design: parallel.
Participants	Diagnosis: paranoid schizophrenia, French classification. History: not mentioned. N= 82. Age: not mentioned. Sex: 82 females. Setting: not mentioned. Exclusion criteria: some specific schizophrenia subtypes.
Interventions	1. Chlorpromazine: increasing dose, maximum dose 400 mg/day. N= 22. 2. Prochlorpemazine: increasing dose, maximum dose 120 mg/day. N= 20. 3. Thioproperazine: increasing dose, maximum dose 60mg/day. N= 17. 4. Haloperidol: increasing dose, maximum dose 10 mg/day. N= 24.
Outcomes	Leaving the study early. Unable to use: Wittenborn scale (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Man 1973.

Methods	Allocation: not mentioned. Blindness: double, no further details. Duration: 3 days. Design: parallel.
Participants	Diagnosis: manic‐ depression psychosis, manic type and acute uncontrollable schizophrenia of various types, clinical diagnosis. History: 90 % of the participants had a record of previous admission to a psychiatric hospital. Participants presented a clear danger to themselves and others. N= 30. Age: 18‐56 years, mean 33 years. Sex: 15 males, 15 females. Setting: inpatients. Exclusion criteria: acute or chronic brain syndromes, epilepsy, neuroses, drug addiction and personality disorders.
Interventions	1. I.m. haloperidol: dose range 5 mg/30 min, maximum 7 injections. N= 15. 2. I.m. chlorpromazine: dose range 50 mg/30 min, maximum 7 injections. N= 15.
Outcomes	Leaving the study early. Less than moderate improvement. Adverse events: At least one EPS, hypotension, seizures. Unable to use: Target Symptom Rating Scale, BPRS (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

McCreadie 1977.

Methods	Allocation: randomised, no further details. Blindness: double, no further details. Duration: 12 weeks. Design: parallel, single‐centre.
Participants	Diagnosis: schizophrenia, clinical diagnosis. History: all participants had had many previous courses of psychotropic medication, mean length of stay in hospital 20 years, range 1‐34. Participants were physically fit and under 65 years of age, a diagnosis of schizophrenia had consistently been made during stay in hospital, presence of one or more "Schneiderian" first‐rank symptoms. N= 20. Age: 38 ‐ 61 years, mean 52 years. Sex: 20 males. Setting: inpatients. Exclusion criteria: participants in a token economy research project, participants who in the opinion of either the medical or the nursing staff were controlled on and benefiting from medication.
Interventions	1. Liquid haloperidol + placebo chlorpromazine: dose range 15‐100 mg/day. N= 10. 2. Chlorpromazine + placebo haloperidol: dose range 100‐600 mg/day. N= 10.
Outcomes	Leaving the study early. Adverse events: Use of antiparkinson medication, akathisia, dystonia, abnormal blood count, sedation, photosensitivity, weight gain. Unable to use: Hamilton scale, Nurse's Observation Scale for In‐Patient Evaluation, assessment of extrapyramidal side effects on a 4‐point scale (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Nishizono 1994.

Methods	Allocation: randomised, no further details. Blindness: not mentioned, probably open. Duration: 4 weeks. Design: multi‐centre.
Participants	Diagnosis: schizophrenia, classification not mentioned. History: not mentioned. N= 169. Age: not mentioned. Sex: not mentioned. Setting: not mentioned. Exclusion criteria: not mentioned.
Interventions	1. Zotepine: maximum dose 450 mg/day. N= not mentioned. 2. Chlorpromazine: maximum dose 450 mg/day. N= not mentioned. 3. Haloperidol: maximum dose 21 mg/day. N= not mentioned.
Outcomes	Leaving the study early. Less than markedly improved. Unable to use: BPRS, GAS (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Reschke 1974.

Methods	Allocation: randomised, no further details. Blindness: initially double, later switch to open administration. Duration: several hours. Design: parallel.
Participants	Diagnosis: schizophrenia, clinical diagnosis. History: participants showed symptoms as agitation, excitement and assaultiveness of such severity that rapid control was necessary. N= 50. Age: 19‐57 years. Sex: 2 males, 48 females. Setting: inpatients. Exclusion criteria: pregnant women, acute or chronic brain syndrome, acute alcoholic intoxication, epilepsy, psychoneurosis, drug addiction, personality disorder.
Interventions	1. I.m. haloperidol: dose range 1‐5 mg/ml, maximum 4 injections. N= 29. 2. I.m. chlorpromazine: dose range 25 mg/ml, maximum 4 injections. N= 10. 3. Placebo. N= 11.
Outcomes	Leaving the study early. Less than moderate improvement. Adverse events: At least one EPS, dry mouth, hypotension, sedation. Unable to use: BPRS (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Ritter 1972.

Methods	Allocation: randomised, no further details. Blindness: double, identical appearing ampules. Duration: 48 hours. Design: parallel.
Participants	Diagnosis: acute or chronic schizophrenia. 48 participants were primarily paranoid schizophrenics. 1 participant had involutional melancholia and 1 was manic‐depressive. History: all exhibited symptoms such as aggressiveness, assaultiveness or unccoperativeness that disrupted normal hospital operation and interfered with attempts at therapy. N= 50. Age: 21‐65 years, mean 36 years. Sex: 50 males. Setting: not mentioned. Exclusion criteria: not mentioned.
Interventions	1. I.m. haloperidol: fixed dose 5 mg/6‐8h. N= 25. 2. I.m. chlorpromazine: fixed dose 50 mg/6‐8h. N= 25.
Outcomes	Leaving the study early. Less than moderate improvement. Adverse events: Leukopenia, GPT increase. Unable to use: hypotension (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Rompel 1978.

Methods	Allocation: randomised, no clear details. Blindness: double, as tablets were readily recognizable, they were issued by the nursing staff, who was strictly forbidden to tell the investigators. Drugs were in their ordinary commercial form and were bottled in numbered bottles with no other information. Duration: 8 weeks. Design: parallel, single‐centre.
Participants	Diagnosis: chronic schizophrenia, clinical diagnosis. History: Participants had been in hospital for at least 5 years and have been firmly diagnosed as schizophrenics. 19 participants were diagnosed with schizophrenia, 1 hebephrenic, 2 katatonic, 3 paranoid schizophrenia. N= 25. Age: mean 45.3 years. Sex: 15 males, 10 females. Setting: inpatients. Exclusion criteria: not mentioned.
Interventions	1. Serenace: probably dose range 5 mg/day, maximum not mentioned. N= 12. 2. Chlorpromazine: dose range 50‐300 mg/day. N= 13.
Outcomes	Leaving the study early. Less than adequately improved. Adverse events: Hypertension.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Serafetinides 1972.

Methods	Allocation: randomised, no further details. Blindness: each participant received medication from his own individual stock bottle, labeled only with his name; capsules appeared identical. Duration: 12 weeks. Design: parallel, single‐centre.
Participants	Diagnosis: chronic schizophrenia, clinical diagnosis. History: participants were without evidence of organic disease, ill for 2 years or longer. N= 57. Age: 21‐61 years. Sex: 25 males, 32 females. Setting: inpatients. Exclusion criteria: not mentioned.
Interventions	1. Haloperidol: dose range 3‐15 mg/day. N= 14. 2. Chlorpromazine: dose range 200‐1000 mg/day. N= 14. 3. Clopenthixol: dose range 50‐250 mg/day. N= 15. 4. Placebo. N= 14.
Outcomes	Leaving the study early. Less than much improved. Adverse events: At least one EPS, akathisia, obstipation, hypotension, excitement, leukocytosis,GPT increase, white blood cells in urine, glycoseria, urea increase, high cholesterol, hypertriglyceridemia, sedation, photosensitivity, edema, weight gain, weight loss. Unable to use: CGI, BPRS, Nurse's Observation Scale for Inpatient Evaluation, Oklahoma Behavior Rating Scale (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Shopsin 1969.

Methods	Allocation: randomised, no further details. Blindness: double; medications were distributed in packets labelled A, B or C. Haloperidol and chlorpromazine ampules were identical. Duration: 6 months. Design: parallel.
Participants	Diagnosis: not mentioned. History: participants were newly admitted to the hospital and showed acute psychotic behaviour with exessive psychomotor hyperactivity, agitation and hallucinations. N= 60. Age: 21‐60 years. Sex: males and females, no further details. Setting: inpatients. Exclusion criteria: acute or chronic brain syndromes, alcoholics, drug addicts, epileptics, psychoneurotics, females thought to be pregnant.
Interventions	1. I.m. haloperidol: dose range 5‐30 mg/day. N= 20. 2. I.m. chlorpromazine: dose range 25‐300 mg /day. N= 20. 3. I.m. sodium amobarbital: dose range 125‐750 mg/day. N= 20.
Outcomes	Leaving the study early. Less than moderate improvement. Adverse events: At least one EPS, hypertension, hypotension, GPT increase, sedation. Unable to use: BPRS, Global Rating scales (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Teja 1975.

Methods	Allocation: randomised, no further details. Blindness: double, no further details. Duration: 36 weeks. Design: parallel, single‐centre.
Participants	Diagnosis: chronic schizophrenia, DSM‐II. History: participants had been hospitalized for 10.01 years. (mean). 16 participants had received ECT. Participants had failed to respond to various treatments administered over the years. N= 66. Age: 22‐57 years, mean 38.01 years. Sex: 30 males, 36 females. Setting: inpatients. Exclusion criteria: epileptic seizures, leucotomy, mental subnormality, schizo‐affective illness, schizophrenia with minimal, well controlled symptoms.
Interventions	All participants entered a 4‐week placebo wash‐out phase before randomisation. 1. Placebo, dose range 25‐450 mg/day. N= 10. 2. Chlorpromazine:dose range 100‐1800 mg/day. N= 14. 3. Trifluoperazine: dose range 5‐90 mg/day. N= 15. 4. Thiothixene: dose range 5‐90 mg/day. N= 14. 5. Haloperidol: dose range 2‐45 mg/day. N= 13. At the end of 20 weeks half of the participants were given amitriptylin, dose range 25‐150 mg/day, or amitriptylin placebo in addition to their antipsychotic. These results were not included in our analysis.
Outcomes	Leaving the study early. Unable to use: BPRS, Burdock's Ward Behavior Rating Scale (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Diagnostic tools: 
 GAS ‐ Global Adjustment Scale. 
 DSM‐II and DSM‐IV ‐Diagnostic Statistical Manual version 2 and 4. 
 CCMD ‐ Chinese Classification of Mental Diseases. 
 RDC ‐ Research Diagnostic Criteria.

WHO ‐ World Health Organisation. 
 I.m. ‐ intramuscular.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Bernardi 1992	Allocation: unclear. Participants: people with schizophrenia and schizoaffective disorder. Intervention: bromperidol decanoate.
Bouchard 2002	Allocation: randomised. Participants: people with schizophrenia. Intervention: risperidone versus classical neuroleptics.
Bruno 1965	Allocation: not randomised.
Chen 2001	Allocation: unclear.
Chouinard 1978	Allocation: randomised. Participants: people with chronic schizophrenia. Intervention: chlorpromazine versus tryptophan ‐ benserazide.
Cosar 1999	Allocation: randomised. Participants: people with schizophrenia. Intervention: haloperidol versus chlorpromazine versus sulpiride. Outcomes: no usable data.
Eitan 1992	Allocation: randomised. Participants: people with schizophrenia. Intervention: placebo versus chlorpromazine versus thioridazine versus haloperidol versus trifluoperazine. Outcomes: no usable data.
Ekdawi 1966	Allocation: unclear randomisation for neuroleptics.
Enoch 1960	Allocation: not randomised.
Fransella 1960	Allocation: randomised. Participants: people with chronic schizophrenia. Intervention: chlorpromazine versus placebo.
Gerstenzang 1977	Allocation: randomised. Participants: unclear how many of the participants had schizophrenia.
Herrera 1988	Allocation: randomised. Participants: people with schizophrenia. Intervention: haloperidol phase followed by chlorpromazine versus clozapine phase.
Hogan 1992	Allocation: randomised. Participants: people with schizophrenia. Intervention: haloperidol versus chlorpromazine. Outcomes: no usable data.
Holt 1984	Allocation: not randomised.
Honigfeld 1984	Allocation: randomised. Participants: people with schizophrenia. Intervention: clozapine versus chlorpromazine or haloperidol.
Horodnicki 1985	Allocation: unclear. Participants: people with schizophrenia. Intervention: chlorprothixene versus haloperidol.
Jones 1968	Allocation: randomised. Participants: people with chronic schizophrenia. Intervention: haloperidol combined with chlorpromazine versus haloperidol alone.
Kurland 1981	Allocation: the interventions relevant for this review were not randomised.
Lejoyeux 1993	Allocation: not randomised.
Lempérière 1985	Allocation: randomised. Participants: people with schizophrenia. Intervention: haloperidol versus chlorpromazine. Outcomes: no usable data.
Lewis 1973	Allocation: randomised. Participants: people with neurosis (10), psychosis (4), situational disorder (3) and behavioral disorders (5).
Marjerrison 1969	Allocation: the interventions relevant for this review were not randomised.
Martin 1975	Allocation: not randomised.
Nishikawa 1982	Allocation: randomised. Participants: people with schizophrenia. Intervention: haloperidol versus chlorpromazine versus placebo versus diazepam versus imipramine. Outcomes: no usable data.
Palma 1997	Allocation: the interventions relevant for this review were not randomised.
Patterson 1981	Allocation: not randomised.
Schipper 1971	Allocation: not randomised.
Shalev 1993	Allocation: randomised. Participants: people with schizophrenia. Intervention: haloperidol versus levomepromazine versus perphenazine.
Singh 1975	Allocation: not randomised.
Singh 1976	Allocation: not randomised.
Singh MM//Kay S2	Allocation: not randomised.
Singh MM//Kay S3	Allocation: not randomised.
Small 1987	Allocation: not randomised.
Terminska 1989	Allocation: not randomised.
van Praag 1975	Allocation: not randomised.
Vangtorp 1968	Allocation: randomised. Participants: people with schizophrenia. Intervention: buronil versus promazine.
Wang 2000	Allocation: not randomised.
Yu 2002	Allocation: quasi randomised (alternate allocation).

Contributions of authors

Claudia Leucht: protocol development, data extraction and writing of the report. 
 Maria Kitzmantel: data entry, writing of the report. 
 Lian Chua: revision of the protocol. 
 John Kane: revision of the protocol. 
 Stefan Leucht: protocol development, data extraction, writing of the report.

Sources of support

Internal sources

• The Zucker Hillside Hospital Intervention Research Center for Schizophrenia (MH 60575), USA.

• Freistaat Bayern, Germany.

External sources

• No sources of support supplied

Declarations of interest

Claudia Leucht: none known. 
 Maria Kitzmantel: none known. 
 Lian Chua: none known. 
 John Kane: has received honoraria from Janssen‐Cilag (manufacturer of haloperidol). 
 Stefan Leucht: has received honoraria from Janssen‐Cilag (manufacturer of haloperidol) and SanofiAventis (manufacturer of chlorpromazine).

Edited (no change to conclusions)