The new antipsychotics, and their potential for early intervention in schizophrenia

Haloperidol versus placebo for schizophrenia

BACKGROUND

Haloperidol was developed in the late 1950s for use in the field of anaesthesia. Research subsequently demonstrated effects on hallucinations, delusions, aggressiveness, impulsiveness and states of excitement and led to the introduction of haloperidol as an antipsychotic.

OBJECTIVES

To evaluate the clinical effects of haloperidol for the management of schizophrenia and other similar serious mental illnesses compared with placebo.

SEARCH METHODS

Initially, we electronically searched the databases of Biological Abstracts (1985-1998), CINAHL (1982-1998), The Cochrane Library (1998, Issue 4), The Cochrane Schizophrenia Group's Register (December 1998), EMBASE (1980-1998), MEDLINE (1966-1998), PsycLIT (1974-1998), and SCISEARCH. We also checked references of all identified studies for further trial citations and contacted the authors of trials and pharmaceutical companies for further information and archive material.For the 2012 update, on 15 May 2012, we searched the Cochrane Schizophrenia Group's Trials Register.

SELECTION CRITERIA

We included all relevant randomised controlled trials comparing the use of haloperidol (any oral dose) with placebo for those with schizophrenia or other similar serious, non-affective psychotic illnesses (however diagnosed). Our main outcomes of interest were death, loss to follow-up, clinical and social response, relapse and severity of adverse effects.

DATA COLLECTION AND ANALYSIS

We evaluated data independently and extracted, re-inspected and quality assessed the data. We analysed dichotomous data using risk ratio (RR) and calculated their 95% confidence intervals (CI). For continuous data, we calculated mean differences (MD). We excluded continuous data if loss to follow-up was greater than 50% and inspected data for heterogeneity. We used a fixed-effect model for all analyses. For the 2012 update, we assessed risk of bias of included studies and used the GRADE approach to create a 'Summary of findings' table.

MAIN RESULTS

Twenty-five trials randomising 4651 people are now included in this review. We chose seven main outcomes of interest for the 'Summary of findings' table. More people allocated haloperidol improved in the first six weeks of treatment than those given placebo (4 RCTs n = 472, RR 0.67 CI 0.56 to 0.80, moderate quality evidence). A further eight trials also found a difference favouring haloperidol across the six weeks to six months period (8 RCTs n = 307 RR 0.67 CI 0.58 to 0.78, moderate quality evidence). Relapse data from two trials favoured haloperidol at < 52 weeks but the evidence was very low quality (2 RCTs n = 70, RR 0.69 CI 0.55 to 0.86). Moderate quality evidence showed about half of those entering studies failed to complete the short trials (six weeks to six months), although, at up to six weeks, 16 studies found a difference that marginally favoured haloperidol (n = 1812, RR 0.87 CI 0.80 to 0.95). Adverse effect data does, nevertheless, support clinical impression that haloperidol is a potent cause of movement disorders, at least in the short term. Moderate quality evidence indicates that haloperidol caused parkinsonism (5 RCTs n = 485, RR 5.48 CI 2.68 to 11.22), akathisia (6 RCTs n = 695, RR 3.66 CI 2.24 to 5.97, and acute dystonia (5 RCTs n = 471, RR 11.49 CI 3.23 to 10.85). Discharge from hospital was equivocal between groups (1 RCT n = 33, RR 0.85 CI 0.47 to 1.52, very low quality evidence). Data were not reported for death and patient satisfaction.

AUTHORS' CONCLUSIONS

Haloperidol is a potent antipsychotic drug but has a high propensity to cause adverse effects. Where there is no treatment option, use of haloperidol to counter the damaging and potentially dangerous consequences of untreated schizophrenia is justified. However, where a choice of drug is available, people with schizophrenia and clinicians may wish to prescribe an alternative antipsychotic with less likelihood of adverse effects such as parkinsonism, akathisia and acute dystonias. Haloperidol should be less favoured as a control drug for randomised trials of new antipsychotics.

Cannabinoids for the treatment of schizophrenia? A balanced neurochemical framework for both adverse and therapeutic effects of cannabis use

Recent studies have found that cannabinoids may improve neuropsychological performance, ameliorate negative symptoms, and have antipsychotic properties for a subgroup of the schizophrenia population. These findings are in contrast to the longstanding history of adverse consequences of cannabis use, predominantly on the positive symptoms, and a balanced neurochemical basis for these opposing views is lacking. This paper details a review of the neurobiological substrates of schizophrenia and the neurochemical effects of cannabis use in the normal population, in both cortical (in particular prefrontal) and subcortical brain regions. The aim of this paper is to provide a holistic neurochemical framework in which to understand how cannabinoids may impair, or indeed, serve to ameliorate the positive and negative symptoms as well as cognitive impairment. Directions in which future research can proceed to resolve the discrepancies are briefly discussed.

Systematic overview of Cochrane reviews for anticholinergic effects of antipsychotic drugs

BACKGROUND

Despite much being written on the topic, there are few surveys investigating the prevalence of anticholinergic adverse effects of antipsychotic drugs. One study, however, used trial-derived data to calculate estimates.

OBJECTIVES

To investigate the prevalence/incidence rates of anticholinergic effects as viewed from within relevant randomized trials.

METHODS

Data were extracted from each relevant study included in Cochrane reviews. Data were checked, extracted, and simple frequencies, and 95% confidence intervals (CIs) were calculated.

RESULTS

Many trials in relevant reviews reported no data on anticholinergic effects (estimate 40,000 participants). However, data were extracted from 177 studies within 54 reviews (N = 27,328 participants). Most data are short-term (<12 weeks). For blurred vision, the newer generations of drugs have rates of between 10% and 20% (eg, risperidone, n = 1460, 6 randomized controlled trials [RCTs], 11.9% prevalence; CI, 10-14; olanzapine, n = 1584; 4 RCTs, 12.2% prevalence; CI, 11-14). These estimates are similar to those of sulpiride (n = 186; 2 RCTs, 12.4%; CI, 8-18) and chlorpromazine (n = 294; 10 RCTs, 11.2%; CI, 8-15), less than trifluoperazine (n = 167; 8 RCTs, 31.1%; CI, 25-39), but considerably more than perphenazine (n = 410; 8 RCTs, 3.7%; CI, 2-6). Data are presented on a range of anticholinergic effects across different periods.

CONCLUSIONS

Anticholinergic symptoms are common adverse effects associated with the use of all antipsychotic drugs, and newer-generation drugs are not clearly distinguishable from many older compounds. Adverse effect data should be more accessible.

Schizophrenic patients treated with clozapine or olanzapine perform better on theory of mind tasks than those treated with risperidone or typical antipsychotic medications

Theory of mind (ToM), the ability to attribute mental states to others, is associated with medial prefrontal cortical (mPFC) activity and is impaired in schizophrenia. Olanzapine or clozapine but not typical antipsychotics or risperidone preferentially affect c-fos expression in mPFC in animals. We tested the hypothesis that schizophrenic patients treated with different antipsychotics would perform differently on ToM tasks. Groups receiving Typicals (n=23), Clozapine (n=18), Olanzapine (n=20) or Risperidone (n=23) and a Control group of healthy volunteers (n=24) were matched for age, gender, handedness and education. ToM functioning was assessed with picture sequence, second-order belief and faux-pas tests. Schizophrenic groups performed similarly to controls on non-ToM conditions. The Olanzapine and Clozapine groups performed similarly to Controls on ToM tasks. The Typicals and Risperidone groups performed worse than the other groups on ToM tasks. We concluded that ToM performance of schizophrenic patients is influenced by the antipsychotic they are taking. Our results suggest that olanzapine or clozapine but not typicals or risperidone may improve or protect ToM ability.

Effects of acute antipsychotic treatment on brain activation in first episode psychosis: an fMRI study

This study aimed to assess the neurophysiological effects of acute atypical antipsychotic treatment on cognitive functioning in subjects presenting with a first episode of psychosis. We used functional MRI to examine the modulatory effects of acute psychopharmacological intervention on brain activation during four different cognitive tasks: overt verbal fluency, random movement generation, n-back and a spatial object memory task. Treatment with atypical antipsychotics was associated with alterations in regional activation during each task and also when task demands were manipulated within paradigms. The initial treatment of psychosis with atypical antipsychotics thus appears to be associated with modifications of the neurofunctional correlates of executive and mnemonic functions. These effects need to be considered when interpreting group differences in activation between medicated patients and controls.

Haloperidol versus placebo for schizophrenia

BACKGROUND

Haloperidol was developed in the late 1950s for use in the field of anaesthesia. Research subsequently demonstrated effects on hallucinations, delusions, aggressiveness, impulsiveness and states of excitement and led to the introduction of haloperidol as an antipsychotic.

OBJECTIVES

To evaluate the clinical effects of haloperidol for the management of schizophrenia and other similar serious mental illnesses compared to placebo.

SEARCH STRATEGY

We initially electronically searched the databases of Biological Abstracts (1985-1998), CINAHL (1982-1998), The Cochrane Library (1998, Issue 4), The Cochrane Schizophrenia Group's Register (December 1998), EMBASE (1980-1998), MEDLINE (1966-1998), PsycLIT (1974-1998), and SCISEARCH. We also checked references of all identified studies for further trial citations and contacted the authors of trials and pharmaceutical companies for further information and archive material. For the 2005 update we searched The Cochrane Library (2005, Issue 6).

SELECTION CRITERIA

We included all relevant randomised controlled trials comparing the use of haloperidol (any oral dose) with placebo for those with schizophrenia or other similar serious, non-affective psychotic illnesses (however diagnosed). Our main outcomes of interest were death, loss to follow up, clinical and social response, relapse and severity of adverse effects.

DATA COLLECTION AND ANALYSIS

We evaluated data independently and analysed on an intention-to-treat basis, assuming that people who left the study early, or were lost to follow-up, had no improvement. Where possible and appropriate, we analysed dichotomous data using Relative Risk (RR) and calculated their 95% confidence intervals (CI). If appropriate, the number needed to treat (NNT) or number needed to harm (NNH) was estimated. For continuous data, we calculated weighted mean differences. We excluded continuous data if loss to follow up was greater than 50% and inspected data for heterogeneity.

MAIN RESULTS

Twenty-one trials randomising 1519 people are now included in this review. One new trial, Kane 2002 (n=414) has been added but it did not affect the overall results. More people allocated haloperidol improved in the first six weeks of treatment than those given placebo (3RCTs n=159, RR failing to produce a marked improvement 0.44 CI 0.3 to 0.6, NNT 3 CI 2 to 5). A further eight trials also found a difference favouring haloperidol across the 6-24 week period (8 RCTs n=308 RR no marked global improvement 0.68 CI 0.6 to 0.8 NNT 3 CI 2.5 to 5) but this may be an over estimate of effect as small negative studies were not identified. About half of those entering studies failed to complete the short trials, although, at 0-6 weeks, 11 studies found a difference that marginally favoured haloperidol (11 RCTs n=898, RR 0.8 CI 0.7 to 0.9, NNT 59 CI 38 to 200). Adverse effect data does, nevertheless, support clinical impression, that haloperidol is a potent cause of movement disorders, at least in the short term. Haloperidol promotes acute dystonia (3 RCTs n=93, RR 4.7 CI 1.7 to 44, NNH 5 CI 3 to 9), akathisia (4 RCTs n=333, RR 2.6 CI 1.4 to 4.8, NNH 7 CI 3 to 25) and parkinsonism (4 RCTs n=163, RR 11.7 CI 2.9 to 47, NNH 3 CI 2 to 5).

AUTHORS' CONCLUSIONS

Haloperidol is a potent antipsychotic drug but has a high propensity to cause adverse effects. Where there is no treatment option, use of haloperidol to counter the damaging and potentially dangerous consequences of untreated schizophrenia is justified. However, where a choice of drug is available, people with schizophrenia and clinicians may wish to prescribe an alternative antipsychotic with less likelihood of adverse effects such as parkinsonism, akathisia and acute dystonias. Haloperidol should not be a control drug of choice for randomised trials of new antipsychotics.

Combination of "atypical" antipsychotic medication in the management of treatment-resistant schizophrenia and schizoaffective disorder

BACKGROUND

This article reviews the published clinical data on treatment-resistant schizophrenic and schizoaffective patients managed with combinations of "atypical" antipsychotic medication.

METHOD

A computerized MEDLINE literature search covering an 18-year period (1985-2003) was conducted. All pertinent papers on the subject of the use of combination "atypical" antipsychotic medication in the management of treatment-resistant schizophrenia and schizoaffective disorder were obtained with subsequent analysis and discussion of the retrieved data.

RESULTS

The search identified 29 case reports and case series reports (172 patients) and one double-blind placebo-controlled trial (28 patients) describing the use of combination "atypical" antipsychotic medication (clozapine-risperidone; clozapine-sulpiride; clozapine-olanzapine; clozapine-quetiapine; olanzapine-sulpiride; olanzapine-quetiapine; risperidone-olanzapine; risperidone-quetiapine) in the treatment of resistant schizophrenic and schizoaffective patients. An overview of results suggests that the combinations were beneficial in the described patients with reduction of positive symptoms and occasionally negative symptoms. Significant adverse effects, while rare, were reported in a few cases and did not appear to different in nature from those managed on monotherapeutic regimens.

CONCLUSION

Combinations of "atypical" antipsychotic medications are well tolerated and may be effective in the management of treatment refractory schizophrenia and schizoaffective disorder. However, further double-blind placebo-controlled trials are required in order to test and confirm these observations.

Is earlier better? At the beginning of schizophrenia: timing and opportunities for early intervention

The fundamental tenet is treating psychotic patients as quickly and as effectively as possible. Few would oppose this idea. Increasing community awareness of the services, enhancing accessibility, optimizing the treatment approaches, improving compliance, and addressing substance misuse should hopefully translate into improved outcomes for the patients and their families and are extremely encouraging and welcome developments. However, the field urgently needs properly designed randomized controlled trials to definitively determine their efficacy. If they are shown to be efficacious the emphasis should then shift to randomized controlled trials of prodromal intervention. If prodromal intervention is proven to be successful then earlier might indeed be better and primary prevention within reach.

Prospective analysis of premature mortality in schizophrenia in relation to health service engagement: a 7.5-year study within an epidemiologically complete, homogeneous population in rural Ireland

While premature death in schizophrenia is well recognised, mortality risk has received little longitudinal study in relation to population representativeness and patient engagement with health services. Within a rural Irish catchment area of socioeconomic, ethnic and geographical homogeneity and low residential mobility, an epidemiologically complete population of 72 patients with schizophrenia was followed up over 7.5 years in order to quantify mortality prospectively. Information was obtained in relation to 99% of the cohort, with 94% of those surviving retained in engagement with psychiatric care. There were 25 deaths (35% of cohort). A relative risk of 2.06 (95% CI, 1.40-2.80; P < 0.001) among this epidemiologically complete population may constitute an estimate of risk for mortality inherent to schizophrenia when disengagement from health services, residential mobility and socioeconomic, ethnic and geographical diversity are minimised. On long-term prospective evaluation, risk for death in schizophrenia was doubled on a background of enduring engagement in psychiatric care with increasing provision of community-based services and introduction of second-generation antipsychotics.

Newer atypical antipsychotic medication in comparison to clozapine: a systematic review of randomized trials

The aim of this study was to evaluate the effectiveness of newer atypical antipsychotic drugs in comparison to clozapine for schizophrenia. Publications in all languages were searched from all relevant databases and all randomized controlled trials comparing clozapine with newer atypical drugs were included. The review and meta-analysis includes eight studies, most of them short in duration. Newer atypical drugs were broadly similar to clozapine when improvement was measured using a psychosis symptom rating scale or a global index. There was a trend for clozapine to be more effective than the others for positive symptoms, and less effective for the negative symptoms. The adverse effect profile of clozapine and newer atypical drugs was dissimilar: while clozapine produced more fatigue, hypersalivation, and orthostatic dizziness, new atypical drugs, with the exception of olanzapine, produced more extrapyramidal symptoms. As these results were obtained from few studies and a relatively small amount of patients, the equal effectiveness and tolerability of new atypical drugs in comparison with clozapine is not yet demonstrated. More trials of sufficient power, with longer duration, and measuring clinically important outcomes are urgently needed.

Childhood-onset schizophrenia: premorbid and prodromal diagnostic and treatment histories

OBJECTIVE

There is increasing interest in the possible relationship between the early diagnosis and treatment of schizophrenia during adolescence and improved long-term outcome. This study reviews the premorbid and prodromal diagnostic and treatment histories for childhood-onset schizophrenia, to assess whether early identification and treatment is possible in this school-age group.

METHOD

Parents of 17 children with childhood-onset schizophrenia or schizoaffective disorder were questioned retrospectively regarding symptoms, exposure to mental health professionals, diagnoses, and treatments.

RESULTS

Initial presenting symptoms clustered around violent aggression and school problems. Age of first recognized psychotic symptoms ranged from 2 to 11 years, followed 2.0+/-2.0 years later by a diagnosis of schizophrenia. Prior to a schizophrenia diagnosis, these children were exposed to stimulants, antidepressants, lower-dose typical neuroleptics, mood stabilizers, alternative treatments, and individual and family therapy.

CONCLUSION

Early diagnosis of childhood-onset schizophrenia is met with caution in the psychological and medical community. These children received many diagnoses before schizophrenia or schizoaffective disorder was diagnosed. A diagnosis of schizophrenia or schizoaffective disorder and utilization of effective atypical neuroleptic treatment was delayed until evaluation by a child and adolescent psychiatrist. Obstacles to early identification and treatment are discussed.

Amisulpride: a review of its use in the management of schizophrenia

Amisulpride, a substituted benzamide derivative, is a second-generation (atypical) antipsychotic. At low doses, it enhances dopaminergic neurotransmission by preferentially blocking presynaptic dopamine D2/D3 autoreceptors. At higher doses, amisupride antagonises postsynaptic dopamine D2 and D3 receptors, preferentially in the limbic system rather than the striatum, thereby reducing dopaminergic transmission. In patients with acute exacerbations of schizophrenia, the recommended dosage of amisulpride is 400 to 800 mg/day, although dosages < or =1200 mg/day may be administered. In comparative trials, amisulpride administered within this range (400 to 1200 mg/day) was as effective as haloperidol 5 to 40 mg/day, flupenthixol 25 mg/day and risperidone 8 mg/day in patients with acute exacerbations of schizophrenia with predominantly positive symptoms. Amisulpride was more effective than haloperidol but equally effective as risperidone in controlling negative symptoms. Amisulpride 400 to 800 mg/day was more effective than haloperidol, risperidone and flupenthixol in controlling affective symptoms in these patients. In randomised, double-blind trials involving patients with predominantly negative symptoms of schizophrenia, amisulpride 50 to 300 mg/day was more effective than placebo. Amisulpride is effective as maintenance therapy in patients with chronic schizophrenia. Long-term treatment with amisulpride was associated with improvements in quality of life and social functioning. Amisulpride is generally well tolerated. In well-controlled trials, the neurological tolerability profile (including ratings on extrapyramidal symptom scales) of amisulpride 400 to 1200 mg/day was superior to that of the conventional antipsychotics (haloperidol or flupenthixol), but was similar to that of the atypical antipsychotic risperidone. At low dosages of amisulpride (< or =300 mg/day), the incidence of adverse events (including extrapyramidal symptoms) reported with amisulpride was similar to that with placebo.

CONCLUSION

In comparative trials, amisulpride 400 to 1200 mg/day showed efficacy in reducing overall symptomatology and positive symptoms similar to that of conventional antipsychotics and newer atypical antipsychotics in patients with acute exacerbations of schizophrenia. Moreover, its effective alleviation of negative and affective symptoms, its lower association with extrapyramidal symptoms and loss of cognitive function than conventional antipsychotics and its long-term efficacy justifies consideration of the use of higher dosages of amisulpride in this group of patients. Consequently, the dosage of amisulpride that is recommended in patients with acute exacerbations of schizophrenia is 400 to 800 mg/day, although dosages < or =1200 mg/day may be administered. Lower dosages of amisulpride (50 to 300 mg/day) should be considered for the management of patients with negative symptoms of schizophrenia. Amisulpride is a first-line treatment option in the management of schizophrenia in the acute phase and for the maintenance of treatment response.

Atypical antipsychotics: new directions and new challenges in the treatment of schizophrenia

"Atypical" antipsychotics represent a new generation of antipsychotics with a significantly lower incidence of extrapyramidal side effects (EPS), as well as little or no effect on prolactin elevation. These advantages constitute a major improvement in the treatment of patients with schizophrenia. The exact mechanisms that make these drugs atypical is not clear. However, a preferential action on serotonin 5-HT2 or D4 receptors, or a more rapid dissociation from the dopamine D2 receptor, may account for atypicality. Although the atypical antipsychotics have overcome EPS, other side effects such as weight gain and impaired glucose tolerance/lipid abnormalities have come to the fore. Thus, the challenges are far from over. The current atypicals are much more effective against the psychosis of schizophrenia than against the other, more enduring aspects of this disorder, e.g. negative symptoms and cognitive dysfunction. At present, the atypicals use a "pharmacological shotgun" strategy to treat aspects of the disease in all patients. A more sophisticated and perhaps effective approach to schizophrenia may lie in independently targeting the pathophysiological mechanisms of each clinical dimension (i.e. positive, negative, cognitive, and affective) with more selective drugs that can be combined and individually titrated to the needs of each patient.

Two double-blind placebo-controlled pilot studies of eicosapentaenoic acid in the treatment of schizophrenia

Evidence that the metabolism of phospholipids and polyunsaturated fatty acids (PUFA) is abnormal in schizophrenia provided the rationale for intervention studies using PUFA supplementation. An initial open label study indicating efficacy for n-3 PUFA in schizophrenia led to two small double-blind pilot studies. The first study was designed to distinguish between the possible effects of two different n-3 PUFA: eicosapentaenoic acid (EPA) and docohexaenoic acid (DHA). Forty-five schizophrenic patients on stable antipsychotic medication who were still symptomatic were treated with either EPA, DHA or placebo for 3 months. Improvement on EPA measured by the Positive and Negative Syndrome Scale (PANSS) was statistically superior to both DHA and placebo using changes in percentage scores on the total PANSS. EPA was significantly superior to DHA for positive symptoms using ANOVA for repeated measures. In the second placebo-controlled study, EPA was used as a sole treatment, though the use of antipsychotic drugs was still permitted if this was clinically imperative. By the end of the study, all 12 patients on placebo, but only eight out of 14 patients on EPA, were taking antipsychotic drugs. Despite this, patients taking EPA had significantly lower scores on the PANSS rating scale by the end of the study. It is concluded that EPA may represent a new treatment approach to schizophrenia, and this requires investigation by large-scale placebo-controlled trials.

Health care utilization in patients with schizophrenia maintained on atypical versus conventional antipsychotics

1. Patients with schizophrenia who had been stabilized on their antipsychotic medication and subsequently maintained on it for a period of at least 18 months were identified: clozapine (N=15); risperidone (N=15); depot conventional (N=18); oral conventional (N=18). 2. Groups were compared on a clinical measure as well as the use of various health care services: hospitalizations; days in hospital, emergency room visits; physician and non-physician visits. 3. No differences between groups were found for hospitalizations, days in hospital, or emergency room visits, while physician and non-physician visits were highest in the clozapine group, in keeping with the need for routine hematologic monitoring in this population. The clozapine group had the highest baseline clinical scores and greatest number of previous hospitalizations. These treatment groups may reflect different clinical populations. However, the findings suggest that in drawing conclusions regarding long-term benefits of different agents, clinical or economic, it would prove useful to include in the evaluation a comparison of patients who have been stabilized on each of the treatments.

Haloperidol versus placebo for schizophrenia

BACKGROUND

Haloperidol was developed in the late 1950s for use in the field of analgesia. Research subsequently demonstrated effects on hallucinations, delusions, aggressiveness, impulsiveness and states of excitement and led to the introduction of haloperidol as an antipsychotic.

OBJECTIVES

To evaluate the clinical effects of haloperidol for the management of schizophrenia and other similar serious mental illnesses compared to placebo.

SEARCH STRATEGY

Electronic searches of Biological Abstracts (1985-1998), CINAHL (1982-1998), The Cochrane Library (1998, Issue 4), The Cochrane Schizophrenia Group's Register (December 2000), EMBASE (1980-1998), MEDLINE (1966-1998), PsycLIT (1974-1998), and SCISEARCH (January 1974-December 1998) were undertaken. References of all identified studies were searched for further trial citations. Authors of trials and pharmaceutical companies were contacted for further information and archive material.

SELECTION CRITERIA

All relevant randomised controlled trials comparing use of haloperidol (any dose) with placebo for those with schizophrenia or other similar serious, non-affective psychotic illnesses (however diagnosed). The main outcomes of interest were death, loss to follow up, clinical and social response, relapse and severity of adverse effects.

DATA COLLECTION AND ANALYSIS

Reviewers evaluated data independently and analysed on an intention-to-treat basis, assuming that people who left the study early, or were lost to follow up, had no improvement. Where possible and appropriate, dichotomous data were analysed using relative risk (RR) and their 95% confidence intervals (CI) calculated. If appropriate, the number needed to treat (NNT) or number needed to harm (NNH) was estimated. For continuous data, weighted mean differences were calculated. Continuous data were excluded if loss to follow up was greater than 50%. All data were inspected for heterogeneity.

MAIN RESULTS

Seventy-four trials were identified but only 20 included. More people allocated to haloperidol improved in the first six weeks of treatment than those given placebo (three trials, n=159, RR failing to produce a marked improvement 0.44 CI 0.3 to 0.6, NNT 3 CI 2 to 5). A further eight trials (n=313) also found a difference favouring haloperidol across the 6-24 week period (RR no marked global improvement 0.68 CI 0.6 to 0.8 NNT 3 CI 2.5 to 5) but this may be an overestimate of effect as small negative studies were not identified. About half of those entering studies failed to complete the short trials, although, at 0-6 weeks, 10 studies found a difference that favoured haloperidol (n=686, RR 0.82 CI 0.7 to 0.95, NNT 8 CI 5 to 17). Limited adverse effect data do, nevertheless, support the clinical impression that haloperidol is a potent cause of movement disorders, at least in the short term. Haloperidol promotes acute dystonia (three trials, n=135, RR 4.7 CI 1.7 to 44, NNH 5 CI 3 to 9 - not assuming those who left early from placebo suffered dystonis), akathisia (three trials, n=129, RR 6.5 CI 1.5 to 28, NNH 6 CI 4 to 14) and parkinsonism (four trials, n=165, RR 8.9 CI 2.6 to 31, NNH 3 CI 2 to 5).

REVIEWER'S CONCLUSIONS

Haloperidol is a potent antipsychotic drug but with a high propensity to cause adverse effects. Given no choice of drug, use of haloperidol to counter the damaging and potentially dangerous consequences of untreated schizophrenia is justified. If a choice of drug is available, however, people with schizophrenia and clinicians may wish to start another antipsychotic with less likelihood of causing parkinsonism, akathisia and acute dystonias. For countries where haloperidol is not widely used, it should not be a control drug of choice for randomised trials of new antipsychotics.