Olanzapine for schizophrenia

Abstract

Background

Olanzapine is an atypical antipsychotic reported to be effective without producing disabling extrapyramidal adverse effects associated with older, typical antipsychotic drugs.

Objectives

To determine the clinical effects and safety of olanzapine compared with placebo, typical and other atypical antipsychotic drugs for schizophrenia and schizophreniform psychoses.

Search methods

We updated the first search [Biological Abstracts (1980‐1999), The Cochrane Library (Issue 2, 1999), EMBASE (1980‐1999), MEDLINE (1966‐1999), PsycLIT (1974‐1999) and The Cochrane Schizophrenia Group's Register (October 2000)] in October 2004 using the Cochrane Schizophrenia's Group's register of trials. We also searched references of all included studies for further trials, and contacted relevant pharmaceutical companies and authors.

Selection criteria

We included all randomised clinical trials comparing olanzapine with placebo or any antipsychotic treatment for people with schizophrenia or schizophreniform psychoses.

Data collection and analysis

We independently extracted data and, for homogeneous dichotomous data, calculated the random effects relative risk (RR), the 95% confidence intervals (CI) and the number needed to treat (NNT) on an intention‐to‐treat basis. For continuous data we calculated weighted mean differences.

Main results

Fifty five trials are included (total n>10000 people with schizophrenia). Attrition from olanzapine versus placebo studies was >50% by six weeks, leaving interpretation of results problematic. Olanzapine appeared superior to placebo at six weeks for the outcome of 'no important clinical response' (any dose, 2 RCTs n=418, RR 0.88 CI 0.8 to 0.1, NNT 8 CI 5 to 27). Although dizziness and dry mouth were reported more frequently in the olanzapine‐treated group, this did not reach statistical significance. The olanzapine group gained more weight.

When compared with typical antipsychotic drugs, data from several small trials are incomplete. With high attrition in both groups (14 RCTs, n=3344, 38% attrition by six weeks, RR 0.81 CI 0.65 to 1.02) the assumptions included in all data are considerable. For the short term outcome of 'no important clinical response', olanzapine seems as effective as typical antipsychotics (4 RCTs, n=2778, RR 0.90 CI 0.76 to 1.06). People allocated olanzapine experienced fewer extrapyramidal adverse effects than those given typical antipsychotics. Weight change data for the short term are not statistically significant but results between three to 12 months suggest a clinically important average gain of four kilograms for people given olanzapine (4 RCTs, n=186, WMD 4.62, CI 0.6 to 8.64).

Twenty three percent of people in trials of olanzapine and other atypical drugs left by eight weeks; 48% by three to12 months (11 RCTs, n=1847, RR 0.91 CI 0.82 to 1.00). There is little to choose between the atypicals, although olanzapine may cause fewer extrapyramidal adverse effects than other drugs in this category. Olanzapine produces more weight gain than other atypicals with some differences reaching conventional levels of statistical significance (1 RCT, n=980, RR gain at 2 years 1.73 CI 1.49 to 2.00, NNH 5 CI 4 to 7). There are very few data for people with first episode illness (1 RCT, duration 6 weeks, n=42). For people with treatment‐resistant illness there were no clear differences between olanzapine and clozapine (4 RCTs, n=457).

Authors' conclusions

The large proportion of participants leaving studies early in these trials makes it difficult to draw firm conclusions on olanzapine's clinical effects. For people with schizophrenia it may offer antipsychotic efficacy with fewer extrapyramidal adverse effects than typical drugs, but more weight gain. There is a need for further large, long‐term randomised trials with more comprehensive data.

Plain language summary

Olanzapine for schizophrenia

Since the early 1950s the mainstay of treatment for schizophrenia has been the typical antipsychotics such as chlorpromazine and haloperidol. Although they are effective in controlling voices and delusions for many people with schizophrenia, they have a smaller effect on symptoms such as apathy and social withdrawal. They also have disabling adverse effects such as tremor, stiffness and slowing of movement. The newer drugs, such as olanzapine, are reputed to have fewer adverse motor effects and are as effective, if not more so, than the older drugs. This review examines the randomised controlled trials of olanzapine compared with placebo, typical and atypical antipsychotic drugs.

Most of the studies in this review were either sponsored by or run by the pharmaceutical companies responsible for the marketing of these drugs. Olanzapine seems to be an effective antipsychotic that produces less adverse effects for movement. It does however tend to cause more weight gain than the older drugs. Its effectiveness appears similar to other atypical drugs and it is not markedly beneficial for apathy, demotivation and social withdrawal.

Background

Antipsychotic drugs have been the mainstay of treatment for schizophrenia since the 1950s. These conventional antipsychotics are largely effective against the 'positive' symptoms such as fixed false beliefs (delusions) and perceptions without a cause (hallucinations). They seem relatively ineffective against the 'negative' symptoms of schizophrenia (apathy, slowing and paucity of movement and thought). It is believed that the conventional or 'typical' antipsychotics exert their antipsychotic effect by blocking specific sites in the brain (dopamine D2 receptors ‐ Seeman 1987). However, this D2 blockade is also responsible for the occurrence of extrapyramidal symptoms, such as stiffness, tremor, repetitive movements (parkinsonism), that may contribute to poor compliance with medication.

The search for antipsychotic agents to manage both 'positive' and 'negative' symptoms has produced a class of drugs called 'atypical' antipsychotics. These compounds, including olanzapine, are purported to have addressed the limitations of traditional antipsychotics by promising better efficacy and/or adverse effect profiles (McEwan 1993, Anonymous 1997).

Technical background 
 Recent research suggests that the pathophysiology of schizophrenia is likely to be heterogeneous and may involve neurotransmitter systems other than D2 receptors. Tamminga 1998, Lieberman 1998 suggested that the addition of 5‐HT2 antagonists, for example ritanserin, to traditional anti dopamine‐D2 drug therapy improved negative symptoms and attenuate pre‐existing drug‐induced extrapyramidal effects (Tamminga 1998, Lieberman 1998).

Olanzapine, a thienobenzodiazepine derivative, is an 'atypical' antipsychotic showing affinity at D1‐D5, serotinergic (5HT2,3,6), muscarinic (subtypes 1‐5), adrenergic (alpha 1‐2) and histaminergic (H1) binding sites (Conley 1998, Tollefson 1997(HGAJ), Reus 1997, Anonymous 1997). It is structurally similar to clozapine but has a slightly different binding site affinity. It is weaker than clozapine as an alpha‐1 and alpha‐2 adrenergic agonist relative to D2, D4 or 5HT2a antagonism.

Objectives

The primary objective of this review is to determine the clinical effects of olanzapine compared with placebo, typical antipsychotic agents, and other atypical antipsychotics, for treating schizophrenia and related psychoses.

As secondary objectives, we also investigated: 
 1. whether people with schizophrenia described as 'treatment resistant' differed in their treatment response from those whose illness was not classified as such 
 2. whether people having predominantly negative symptoms of schizophrenia were more responsive to olanzapine than those without this classification 
 3. whether people experiencing their first episode of a psychotic illness differed in response from those with a previous history.

Methods

Criteria for considering studies for this review

Types of studies

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

Types of participants

We included people with schizophrenia and other types of schizophrenia‐like psychoses (schizophreniform and schizoaffective disorders). There is no clear evidence that the schizophrenia‐like psychoses are caused by fundamentally different disease processes or require different treatment approaches (Carpenter 1994).

Types of interventions

1. Olanzapine: any dose 
 2. Any other antipsychotic agent, divided into typical and atypical 
 3. Placebo.

Types of outcome measures

Outcomes were divided into short‐term (less than three months) medium term (three to 12 months) and long term (over one year).

Primary outcomes

1. Clinical response 
 1.1 No clinically significant response in global state ‐ as defined by each of the studies

Secondary outcomes

1. Death, by suicide or natural causes

2. Leaving the study early 
 
 3. Clinical response 
 3.1 Average score/change in global state 
 3.2 No clinically significant response on psychotic symptoms ‐ as defined by each of the studies 
 3.3 Average score/change on psychotic symptoms 
 3.4 No clinically significant response on positive symptoms ‐ as defined by each of the studies 
 3.5 Average score/change in positive symptoms 
 3.6 No clinically significant response on negative symptoms ‐ as defined by each of the studies 
 3.7 Average score/change in negative symptoms 
 3.8 Use of additional medication (other than anticholinergics) for psychiatric symptoms 
 
 4. Extrapyramidal adverse effects 
 4.1 Incidence of use of antiparkinson drugs 
 4.2 No clinically significant extrapyramidal adverse effects ‐ as defined by each of the studies 
 4.3 Average score/change in extrapyramidal adverse effects 
 
 5. Other adverse effects, general and specific 
 
 6. Service utilization outcomes 
 6.1 Hospital admission 
 6.2 Days in hospital 
 
 7. Economic outcomes 
 
 8. Quality of life/satisfaction with care for either recipients of care or carers 
 8.1. Significant change in quality of life/satisfaction ‐ as defined by each of the studies 
 8.2 Average score/change in quality of life/satisfaction

Search methods for identification of studies

Electronic searches

1. Electronic searching for update (October 2004)

1.1. We searched the Cochrane Schizophrenia Group's Register (October 2004) with the phrase:

[olanzapine* or Zyprexa* in title or *olanzapine* or *zyprexa* in abstract, index terms of REFERENCE] or [olanzapine* 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.

2. Details of previous searches:

2.1. We searched the Cochrane Schizophrenia Group's Register (October 2000) using the phrase:

[zyprex* or olanzapine* or #42 = 427]

(#42 is the field within this Register in which each intervention is coded and 427 is olanzapine.)

2.2. We searched the Cochrane Library (June 1999) using the phrase:

[zyprex* or olanzapine*]

2.3. We searched biological Abstracts (Jan 1980 ‐ June 1999) using the Cochrane Schizophrenia Group's phrase for randomised controlled trials and schizophrenia (see Group search strategy) combined with the phrase:

[and (zyprex* or olanzapine*)]

2.4. We searched EMBASE (Jan 1980 ‐ July 1999) using the Cochrane Schizophrenia Group's phrase for randomised controlled trials and schizophrenia (see Group search strategy) combined with the phrase:

[and (zyprex* or olanzapine* or explode "OLANZAPINE"/ all subheadings)]

2.5. We searched MEDLINE (Jan 1966 ‐ September 1999) using the Cochrane Schizophrenia Group's phrase for randomised controlled trials and schizophrenia (see Group search strategy) combined with the phrase:

[and (zyprex* or olanzapine*)]

2.6. We searched PsycLIT (Jan 1974 ‐ June 1999) using the Cochrane Schizophrenia Group's phrase for randomised controlled trials and schizophrenia (see Group search strategy) combined with the phrase:

[and (zyprex* or olanzapine*)]

3. Conference abstracts: 
 3.1 the American Psychiatric Association (1998) conference, XXIst Collegium Internationale Neuro‐psychopharmacologicum (Glasgow, Scotland ‐ July 1998)

3.2 the 11th European College of Neuropsychopharmacology Congress (Paris, France, October 31 ‐ November 4, 1998)

3.3 We searched the World Psychiatric Association (1999) Abstracts of the XI World Congress of Psychiatry (Hamburg, Germany ‐ August 6‐11, 1999) using the phrase:

[olanzapine and randomi*]

Searching other resources

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

2. Authors of studies 
 We contacted authors of studies when necessary to clarify data, and requested additional studies.

3. Pharmaceutical company 
 We contacted Eli Lilly to obtain unpublished data and unpublished trials.

Data collection and analysis

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

1. Selection of studies 
 We (LD, MF) independently inspected all reports of identified studies. It was usually possible to resolve any disagreement by consensus; however, where doubt remained we acquired the full article. We independently decided whether these met the review criteria. No blinding to the names of authors, institutions and journal of publication took place. Again, we resolved any disagreements by consensus. When this proved impossible, we sought further information and, in the interim, added these trials to the 'Awaiting assessment' list. LD, MF and JR independently inspected citations from the subsequent updated search (2004) and identified relevant abstracts. We obtained and inspected full reports of the abstracts meeting the review criteria.

2. Assessment of methodological quality 
 We allocated trials to three quality categories, as described in the Cochrane Collaboration guidelines (Alderson 2004). We only included trials in Category A or B in the review.

3. Data management 
 3.1 Data extraction 
 We independently extracted data and resolved disagreement by discussion. When this was not possible we sought further information from trial authors.

3.2 Intention to treat analysis 
 Where possible we analysed data on an intention‐to‐treat basis and assumed that those who had not been accounted for had the less positive outcome. This rule did not include the outcome of 'death'. We tested this assumption with a sensitivity analysis. For continuous data it is impossible to manage the data in this way therefore we presented 'completer' data. Where feasible, we converted continuous scores to dichotomous data.

If, for a given outcome, more than 50% of the total numbers randomised were not accounted for, we did not present results as such data are impossible to interpret with authority. If, however, more than 50% of those in one arm of a study were lost, but the total loss was less than 50%, we marked data with '*' to indicate the result may well be prone to bias.

4. Data analysis 
 4.1 Binary data 
 When summation was appropriate, with binary outcomes such as improved/not improved, we calculated the relative risk (RR) statistic with a 95% confidence interval (CI) and used a random effects model. In addition, as a measure of efficiency, we estimated the number needed to treat (NNT) or the number needed to harm (NNH) from the pooled totals.

4.2 Continuous data 
 4.2.1 Normally distributed data 
 Continuous data on clinical and social outcomes are often not normally distributed. To avoid the pitfall of applying parametric tests to non‐parametric data, we applied the following standards to all data before inclusion: (a) standard deviations and means were reported in the paper or were obtainable from the authors; (b) when a scale started from the finite number zero, the standard deviation, when multiplied by two, was less than the mean (as otherwise the mean is unlikely to be an appropriate measure of the centre of the distribution, (Altman 1996); (c) if a scale started from a positive value (such as PANSS which can have values from 30 to 210) the calculation described above was modified to take the scale starting point into account. In these cases skew is present if 2SD>(S‐Smin), where S is the mean score and Smin is the minimum score. Endpoint scores on scales often have a finite start and end point and these rules 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 difficult to tell whether data are non‐normally distributed (skewed) or not. Skewed data are presented in the 'Other data' tables rather than included in the analysis.

For change data (endpoint minus baseline), the situation is even more problematic. In the absence of individual patient data it is impossible to know if data are skewed, though this is likely. After consulting the ALLSTAT electronic statistics mailing list, we presented change data in order to summarise available information. In doing this, it was assumed either that data were not skewed or that the analyses could cope with the unknown degree of skew. Again, without individual patient data it is impossible to test this assumption. Where both change and endpoint data were available for the same outcome category, we have presented only endpoint data. We acknowledge that by doing this, much of the published change data could have been excluded, but argue that endpoint data is more clinically relevant and that if change data were to be presented along with endpoint data, it would be given undeserved equal prominence. We contacted authors of studies that only reported change for endpoint figures.

4.2.2 Summary statistic 
 For continuous outcomes we estimated a weighted mean difference (WMD) between groups. Again this was based on the random effects model, as this takes into account any differences between studies even if there is no statistically significant heterogeneity. We did not consider continuous data presented without use of summary statistics (i.e. mean, SD, SE, median, interquartile range), although we noted the existence of these data in the text.

4.2.3 Valid Scales 
 Many rating scales are available to measure outcomes in mental health trials (Marshall 2000). These scales vary in quality and many are poorly validated. It is generally accepted that measuring instruments should have the properties of reliability (the extent to which a test effectively measures anything at all) and validity (the extent to which a test measures that which it is supposed to measure). Before publication of an instrument, most scientific journals insist that its reliability and validity be demonstrated to the satisfaction of referees. As a minimum standard, we excluded data from unpublished rating scales. In addition, the rating scale should be either: (i) a self report; or (ii) completed by an independent rater or relative. More stringent standards for instruments may be set in future editions of this review.

Continuous data may be presented from different scales, rating the same outcome. In this event, we presented all data without summation and inspected the general direction of effect.

4.2.4 Conversion to a common metric 
 To facilitate comparison between trials, we converted variables (such as days in hospital) that could be reported in different metrics (mean days per year, per week or per month) to a common metric (e.g. mean days per month).

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

Where clustering was not accounted for in primary studies, we presented the data in a table, with a (*) symbol to indicate the presence of a probable unit of analysis error. In subsequent versions of this review we will seek to contact first authors of studies to obtain intra class correlation co‐efficients of their clustered data and to adjust for this by 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 intra‐class correlation co‐efficient (ICC) Design effect = 1+(m‐1)*ICC (Donner 2002). If the ICC was not reported it was assumed to be 0.1 (Ukoumunne 1999).

5. Investigation for heterogeneity 
 Firstly, we considered all the included studies within any comparison to judge clinical heterogeneity. We then visually inspected the graphs to investigate the possibility of statistical heterogeneity and supplemented this using, primarily, the I‐squared statistic. This provides an estimate of the percentage of variability due to heterogeneity rather than to chance alone. Where the I‐squared estimate was greater than or equal to 75%, we interpreted this as indicating the presence of high levels of heterogeneity (Higgins 2003). If inconsistency had been high, data would not have been summated, but we would have presented them separately and investigated reasons for heterogeneity.

6. Addressing publication bias 
 We entered all data from all identified and selected trials into a funnel graph (trial effect against trial size) in an attempt to investigate the likelihood of overt publication bias (Egger 1997).

7. General 
 Where possible, we entered data in such a way that the area to the left of the line of no effect indicated a favourable outcome for olanzapine.

Results

Description of studies

For substantive descriptions of studies please see Included and Excluded studies tables.

1. Excluded studies 
 When this review was first undertaken we excluded a total of over 160 studies. Of these, more than 60 were excluded because they were not randomised trials, including nine open‐label, single‐arm studies supplied by Eli Lilly. Nine were review articles and one reported validity data for a questionnaire. Twelve studies did not include people with schizophrenia. More than 70 studies presented outcomes that were not usable. One study was originally excluded (Conley 2001) following the withdrawal of the study by Janssen‐Cilag. Amended data have now been presented and are included (Conley 2001). For the 2004 update, even fewer studies were added to the excluded studies due to the increasing specificity of the Cochrane Schizophrenia register of controlled trials.

2. Awaiting assessment 
 No studies are waiting assessment.

3. Ongoing studies 
 Eight RCTs comparing olanzapine with other antipsychotics are ongoing (Centorrino 2001, Hirsch 2000, Krakowski 2001, Ramamurthy 2000a, Reveley 2000, Reynolds 2003, Singh 2000b, Turner 2002).

4. Included studies 
 Fifty five studies are included in the October 2004 update and this represents an additional 26 studies since the 2003 update. These 55 studies randomised over 10000 people with schizophrenia, schizophreniform and schizoaffective disorders. All included studies are reported as randomised and all but 16 are described as double blind. In almost half of the studies, Eli Liiy (the pharmaceutical company that produces olanzapine) carried out randomisation at the individual investigation sites; Janssen‐Cilag undertook randomisation in the Conley 2001 study. For the remaining included studies, no additional information is available.

4.1 Length of trials 
 Schizophrenia is a lifelong illness that affects young people. Thirty one studies presented data on 'short‐term' follow‐up (less than three months). Three of these acute phase studies (Beasley 1996a (HGAD), Beasley 1997 (E003), Tomlinson 1997(HGAJ)) lasted six weeks with 'responders' entering a 46‐week extension. Tomlinsons (HGDY) study was the shortest trial (three to five days) following abrupt discontinuation of cocaine. Twenty three studies fell into the 'medium term' (three to 12 months) category and only two trials (Jones 1998 (P022), Meltzer (Intercept)) fulfilled our criterion for 'long‐term' (more than one year). The included trials of treatment resistant patients provided only short‐term and medium‐term data.

4.2 Participants 
 All but eleven studies included participants with operationalised diagnoses (Chang 2003, Chen 2003, Jakovljevic1999 HGCH, Jones 1998 (P022), Kern 2001, Lecrubier 1999, Littrell 1999, McQuade 2003, Namjoshi 2002, Ritchie 2003, Wang 2002) by Diagnostic and Statistical Manual (APA 1994) or The International Statistical Classification of Diseases and Related Health Problems (ICD‐10). The participants in all studies bar one were diagnosed as suffering from schizophrenia, although twelve studies also included people with schizoaffective disorder and schizophreniform psychosis. Wang 2003 used CLMD‐3 (Chinese Classification of Mental Disorders Version 3). HGCQ (Turkey) 2000 did not mention the diagnosis but Eli Lilly listed the study's participants as suffering from 'schizophrenia and related psychoses' and hence it is included. As Beasley 1996a (HGAD) only allowed women of 'childbearing potential' to enter after approximately two‐thirds of the enrolment had been completed the participants were predominantly male. Svestka 2003 recruited only women.

The participants in Tollefson 1997(HGAJ) were less severely ill than was generally seen in the other studies (mean BPRS score at baseline in the olanzapine group: Tollefson 1997(HGAJ) ‐ 33.1, Beasley 1996a (HGAD) ‐ 42.6, Beasley 1997 (E003) ‐ 42.3, Jakovljevic1999 HGCH ‐ 43.7). This large study also contained a large number of people who were intolerant of their current medication.

Four studies selected people whose illness was described as resistant to treatment (Beuzen 1998 (HGCF), Bitter 2004 (HGCK), Conley 1998, Kelly 2003). All used different definitions of treatment resistance and treatment intolerance as set out in the included studies table.

4.3 Setting 
 Twenty two trials took place in both inpatient and outpatient settings. Eighteen used an inpatient design and only five studies an outpatient setting. Eleven studies did not report the setting.

4.4 Study size 
 Tollefson 1997(HGAJ) was by far the largest study (1996 participants) whilst Kelly 2003 was the smallest study, randomising only 13 people. Seventeen studies had fewer than fifty participants, eleven had between 50 and 100 participants, 24 studies had between 100 and 400 participants and four randomised more than four hundred people. There were no cluster trials.

4.5 Intervention 
 The trialists gave olanzapine in a wide range of doses from 1mg/day to 50 mg/day. Beasley 1996b (HGAP) used 1 mg/day stating that this was equivalent to a placebo. Beasley 1997 (E003), however, started with the same assertion but altered this during the study. Because of this we stopped using 1mg/day as additional placebo data.

The comparators were placebo, chlorpromazine, haloperidol, fluphenazine, flupentixol, perphenazine, clozapine, amisulpride, risperidone, zisprasidone, aripiprazole and sonepiprazole. (Pharmacia Corporation recently announced that it was discontinuing the development of the atypical antipsychotic sonepiprazole (PNU 101387)).

4.6 Outcomes 
 4.6.1 Improvement ‐ Definition of improvement often consisted of a 20‐40% reduction in BPRS or PANSS scores. For treatment resistant participants, the lower level of change in mental state scores was considered as an improvement. Individual papers reported different cut offs for improvement. Pragmatic outcomes were reported in Meltzer (InterSept) (suicide attempt, death by suicide, change in hospital status etc.).

Studies often dichotomised their overall measure of efficacy from continuous scales (BPRS, PANSS). This dichotomising is welcome as data begins to become clinically understandable. It is, however, wasteful of time and effort, if a simple yes/no question on clinical efficacy would suffice rather than hundreds of BPRS and PANSS questions. The validity of dichotomising is unclear.

Seven studies supplied by Eli Lilly, present data for loss to follow up and adverse effects. As far as we understand, effects on global functioning and mental state are not currently available, even to the funding company. We have contacted Eli Lilly to ascertain exactly what outcomes were recorded in addition to those for which data are available, but have not received a response to date. However, additional data is included in this update from three studies HGCQ (Turkey) 2000, HGDV (Morocco) 1999 and Loza 1999 (HGDT).

4.6.2 Adverse effects ‐ Many studies sponsored by pharmaceutical companies reported data only when the differing incidence was statistically significant. The authors regarded these data as potentially misleading, as chance events were likely to be given prominence and equally frequent adverse effects to be ignored. Eli Lilly provided published and unpublished adverse event data from the COSTART terms. These have been included in this version of the review. 
 
 4.6.3 Outcome scales ‐ Details of scales that provided usable data are shown below. Reasons for exclusion of data from other instruments are given under 'Outcomes' in the 'Included studies' section.

4.6.3.1 Global state scales 
 4.6.3.1.1 Clinical Global Impression Scale ‐ CGI Scale (Guy 1976) 
 This is used to assess both severity of illness and clinical improvement, by comparing the conditions of the person standardised against other people with the same diagnosis. A seven‐point scoring system is usually used with low scores showing decreased severity and/or overall improvement.

4.6.3.2 Mental state scales 
 4.6.3.2.1 Brief Psychiatric Rating Scale ‐ BPRS (Overall 1962) 
 This is used to assess the severity of abnormal mental state. The original scale has 16 items, but a revised 18‐item scale is commonly used. Each item is defined on a seven‐point scale varying from 'not present' to 'extremely severe', scoring from 0‐6 or 1‐7. Scores can range from 0‐126, with high scores indicating more severe symptoms. Four studies extracted the BPRS scores from the PANSS Beasley 1996b (HGAP), Beasley 1997 (E003), Tollefson 1997, COSTART 1990, Tran 1997 (HGBG). We have not found any validation of this method of extracting scores.

4.6.3.2.2 Hamilton Rating Scale for Depression ‐ HDRS (Hamilton 1960) 
 This instrument is designed to be used only on people already diagnosed as suffering from an affective disorder of depressive type. It is used for quantifying the results of an interview, and its value depends entirely on the skill of the interviewer in eliciting the necessary information. The scale contains 17 variables measured on either a five or a three‐point rating scale, the latter being used where quantification of the variable is either difficult or impossible. Among the variables are: depressed mood, suicide, work and loss of interest, retardation, agitation, gastro‐intestinal symptoms, general somatic symptoms, hypochondriasis, loss of insight, and loss of weight. It is useful to have two raters independently scoring the person at the same interview. The scores of the person are obtained by summing the scores of the two physicians. High scores indicate greater severity of depressive symptoms.

4.6.3.2.3 Hamilton Anxiety Rating Scale ‐ HAMA (Hamilton 1969) 
 The Hamilton Anxiety Scaleis a rating scale developed to quantify the severity of anxiety symptomatology, often used in psychotropic drug evaluation. It consists of 14 items, each defined by a series of symptoms. Each item is rated on a 5‐point scale, ranging from 0 (not present) to 4 (severe). The 14 items consist of: Anxious Mood, Tension, Fears, Insomnia, Intellectual, Depressed Mood, Somatic Complaints (Muscular), Somatic Complaints (Sensory), Cardiovascular Symptoms, Respiratory Symptoms, Gastrointestinal symptoms, Genitourinary symptoms, Autonomic Symptom and Behaviour at Interview. 
 
 4.6.3.2.4 Mini Mental State Examination ‐ MMSE (Folstein 1975) 
 This clinician‐administered clinical evaluation assesses cognition in five areas: orientation, immediate recall, attention and calculation, delayed recall, and language. The test takes 15 minutes to administer and the score ranges from 0 (severe impairment) to 30 (normal).

4.6.3.2.5 Positive and Negative Syndrome Scale ‐ PANSS (Kay 1986) 
 This schizophrenia scale has 30 items, each of which can be defined on a seven‐point scoring system varying from 1 ‐ absent to 7 ‐ extreme. It can be divided into three sub‐scales for measuring the severity of general psychopathology, positive symptoms (PANSS‐P), and negative symptoms (PANSS‐N). A low score indicates lesser severity.

4.6.3.2.6 Scale for the Assessment of Negative Symptoms ‐ SANS (Andreasen 1983) 
 This six‐point scale gives a global rating of the following negative symptoms; alogia, affective blunting, avolition‐apathy, anhedonia‐asociality and attention impairment. Higher scores indicate more symptoms.

4.6.3.2.7 Montgomery Asberg Depression Rating Scale ‐ MADRS (Montgomery 1979) 
 A 65‐item comprehensive psychopathology scale was used to identify the 17 most commonly occurring symptoms in primary depressive illness. Ratings are based on 10 items, with higher scores indicating more symptoms.

4.6.3.2.7 Subjective Well‐being on Neuroleptics Scale ‐ SWN (de Haan 2002a) 
 The SWN is an instrument to measure the subtle subjective changes, such as restrictions in emotionality, the clarity of thinking and spontaneity that are often referred to as 'pharmacogenic depression' or the 'neuroleptic induced deficit syndrome'.

4.6.3.3 Adverse effects scales 
 4.6.3.3.1 Abnormal Involuntary Movement Scale ‐ AIMS (Guy 1976) 
 This has been used to assess tardive dyskinesia, a long‐term, drug‐induced movement disorder and short‐term movement disorders such as tremor.

4.6.3.3.2 Association for Methodology and Documentation in Psychiatry (AMDP‐5 1981) 
 This is a 40‐item adverse event questionnaire.

4.6.3.3.3 Barnes Akathisia Scale ‐ BAS (Barnes 1989) 
 The scale comprises items rating the observable, restless movements that characterise akathisia, a subjective awareness of restlessness, and any distress associated with the condition. These items are rated from 0 ‐ normal to 3 ‐ severe. In addition, there is an item for rating global severity (from 0 ‐ absent to 5 ‐ severe). A low score indicates low levels of akathisia.

4.6.3.3.4 COSTART Terms (COSTART 1990) 
 This is a list drawn up by the US Food and Drug Administration in order to help consistent description of adverse reactions. It is not a scoring system. Eighteen studies reported usable data.

4.6.3.3.5 Drug‐Induced Extrapyramidal Symptom Scale ‐ DIEPSS (Inada 1996) 
 This is a nine item scale used to assess the extrapyramidal syndromes parkinsonism, akathisia, dystonia and dyskinsia.

4.6.3.3.6 Extrapyramidal Symptom Rating Scale ‐ ESRS (Chouinard 1980) 
 This consists of a questionnaire relating to parkinsonian symptoms (nine items), a physician's examination for parkinsonism and dyskinetic movements (eight items), and a clinical global impression of tardive dyskinesia. High scores indicate severe levels of movement disorder.

4.6.3.3.7 Hillside Akathisia Scale ‐ HAS (Fleischhacker 1989) 
 The Hillside Akathisia Scale (HAS) has two subjective and three objective items for which anchored rating points are provided.

4.6.3.3.8 Simpson Angus Scale ‐ SAS (Simpson 1970) 
 This ten‐item scale, with a scoring system of 0‐4 for each item, measures drug‐induced parkinsonism, a short‐term drug‐induced movement disorder. A low score indicates low levels of parkinsonism.

4.6.3.3.9 UKU Side Effects Rating Scale ‐ UKU‐SERS (Lingjaerde 1987) 
 The UKU rates four major topics: psychological side effects (10 items), neurological side effects (eight items), autonomic side effects (11 items) and other side effects (19 items). Each item is defined by means of a four‐point scale where zero means not/doubtfully present. Scoring range is 0‐144.

4.6.3.3.10 Leeds Sleep Evaluation Questionaire ‐ LSEQ (Parrott 1980) 
 The SEQ is a simple 10 item bipolar VAS type questionnaire that can be used to evaluate sedative effects in psychotropic drugs. Four areas are tapped 1) sleep onset, 2) quality of sleep, 3) ease of awakening and 4) integrity of morning behaviour. The SEQ has been factor analysed and four factors were identified, 1) getting asleep, 2) perceived sleep quality, 3) awakening and 4) early morning behaviour. A high score indicates improvement.

4.6.3.4 Quality of life scales 
 4.6.3.4.1 Quality of Life Scale (Lehman 1983) 
 This standardised assessment includes areas such as living situation, leisure activities, relationships and finances. This is rated on a seven‐point scale, with higher scores indicating a better quality of life.

4.6.3.4.2 Quality of Life Scale ‐ QLS (Carpenter 1984) 
 This semi‐structured interview is administered and rated by trained clinicians. It contains 21 items rated on a seven‐point scale based on the interviewer's judgement of patient functioning. A total QLS and four sub‐scale scores are calculated, with higher scores indicating less impairment.

4.6.3.5 Satisfaction 
 4.6.3.5.1 Van Putten Scale 
 This is designed to measure patient satisfaction with treatment. This was used in the HGCQ (Turkey) 2000 study. A fuller description of this scale is being sought.

4.6.3.5.2 Drug Attitude Inventory ‐ DAI‐30 (Hogan 1983) 
 This is a self‐report 30‐item scale for assessing patient satisfaction with antipsychotic treatment in‐patients with schizophrenia. This inventory is designed to measure the person's subjective responses to medications as well as values and attitudes toward illness and health. Higher scores indicate a more favourable attitude towards antipsychotic drug treatment. A positive total final score means a positive subjective response (compliant). A negative total score means a negative subjective response (non‐compliant).

5. Missing outcomes 
 Only Beasley 1996a (HGAD), Meltzer (InterSept) and Namjoshi 2002 reported hospital admission data. No useful data about daily functioning, such as 'employed' or 'trouble with the police', or satisfaction with care was reported. Eli Lilly provided economic data and this will be incorporated into further updates of the review. Six studies reported mortality data.

Risk of bias in included studies

1. Randomisation 
 Reporting of randomisation seemed poor in the published papers. As studies have shown that poor reporting of randomisation increases the odds of presenting 'significant' outcomes (Chalmers 1983, Schulz 1995) this was brought to the attention of the companies responsible for the trials (Eli Lilly and Janssen‐Cilag). They have since provided full reports of randomisation and the studies now fall into quality category A (adequate concealment of allocation). Eli Lilly randomised using computer‐generated blocks, for investigative sites or investigator and concealed randomisation from the investigators. Janssen‐Cilag used a randomisation ratio of 1:1 using the minimisation technique of Pocock and Simon (Pocock 1975) and a dynamic randomisation procedure stratified by body mass index (BMI), PANSS total and ESRS scores.

2. Blinding to interventions and outcomes 
 Again authors poorly reported their blinding procedures with only (Conley 1998) clearly describing adequate precautions. Eli Lilly and Janssen‐Cilag supplied us with details of blinding that we have included in the 'Included studies' table. Several studies gave people a medication kit containing their medication of allocation in a form that was not clearly different from the comparison drug or used identical capsules (Conley 2001). No study tested blinding. He 2003 and Meltzer (InterSept) were not blind.

3. Follow‐up 
 The numbers leaving the study early were high, albeit comparable to trials of other atypical antipsychotics (Fabre 1995, Hunter 2003, Srisurapanont 2004, Thornley 2003). The majority of trials that had been published in peer reviewed journals described the participant disposition well, clearly informing the reader as to individual reasons for withdrawal from the study. Studies that had been presented as posters or presentations often did not include data on the follow‐up of participants, leaving the reader uninformed as to the whereabouts of the participants.

Nearly all the studies stated that they analysed their data on an intention‐to‐treat basis using the last observation carried forward. This practice may well overestimate any treatment effect. Once participants leave a study, unless the trialists continue to follow and collect data, assumptions have to be made about outcome. As many people left these studies early, the results do incorporate considerable assumptions (see 'Results', 'Discussion' and 'Implications for research'). Eli Lilly kindly provided drop out data from their trials.

In this review those leaving the studies early were classified as a 'bad outcome' and analysed accordingly (except for the outcome of death). Whatever the management of lost data, interpretation of results with large degrees of attrition must be undertaken with caution. Where greater than 50% of those randomised were lost to follow up, outcomes are marked with '*' in order to draw the attention of the reader to the 'prone to bias' footnote in the graphical display of results.

Effects of interventions

1. The search 
 We found more than 800 citations using the search strategies and 55 studies for inclusion. Twenty six of these are additional studies in the 2004 update. Eli Lilly identified 11 randomised controlled trials, four of which were unpublished data on file. We found many trials reported as presentations at conferences or meetings with the full report being published in an academic journal (Table 1). There is a close correlation between study size and number of presentations. If regression analysis of the data in Table 1 is undertaken on a log‐graph the least‐squares line makes a close fit (R2=0.92). For every 10 people that are randomised in an olanzapine versus typical drugs comparison there is a report. A study of 10 could be predicted to merit a report, of 100‐110 presentations and of 1000‐1100 reports. Considering the likely under‐identification of presentations of the larger studies this pattern of comprehensive dissemination does not seem too unlikely. Janssen‐Cilag withdrew data from Conley 2001 in April 1999, due to site irregularities. It has now been published and is included in this review.

1. Olanzapine versus typical antipsychotics: Study size and number of reports.

Study tag	Participants	Published reports	Unpublished reports
Tollefson (HGAJ)1997	1996	143
Beasley 1997	431	16
Beasley 1996a	267	35
HGFH (Korea) 1998	104	1
Conley 1998a	84	5
Bernardo 2001	27	1
HGCU (Taiwan) 1998	54	3
HGBJ (Finland)	46	2	1
Jones 1998	65	10	2
Loza 1999	41	2	1
HGDV (Morocco) 1999	40	0	2
HGCJ (Hong Kong) 1998	31	5
HGBL 1997	30	0	1
HGCQ (Turkey) 2000	30	3	1
Altamura 1999	24	1
Avasti 2001	27	1
Barak 2002	20	1
Chang 2003	62	1
Chaudry 2003	276	1
Conley 1998	84	7
de Hann 2003	24	2
Esel 2001	29	1
He 2003	80	1
Ishigooka 2001	182	4
Jakovljevic1999 HGCH	60	11
Lieberman 2003 (HGDH)	263	28
Lima 2003 (HGHS)	197	3	1
Malyarov 1999	43	1
Namjoshi 2002	364	4
Rosenheck 2003(HGFI)	309	3

2. COMPARISON 1. OLANZAPINE versus PLACEBO 
 Five trials used placebo groups in their comparisons (Beasley 1996a (HGAD), Beasley 1996b (HGAP), Corrigan 2004, Lecrubier 1999, Tollefson 1999(HGDY)). Tollefson 1999(HGDY) is so short term (duration five days) that we have reported data in a separate comparison. The loss to follow up in the other four studies is such that most data within this comparison is marked with '*' and 'prone to bias' in the graphical presentations. The degree of assumption within these data is very great and must be interpreted with a high degree of caution. Tollefson 1999(HGDY) was the only trial to have small numbers of people leaving, but this trial lasted for just five days.

2.1 Global effect 
 2.1.1 No important clinical response 
 The trialists defined 'no important clinical response' as the number of people per treatment group who did not present at least a 40% reduction in psychotic symptoms measured by any scale. Olanzapine was superior to placebo at six weeks (any dose, 2 RCTs, n=418, RR 0.88 CI 0.8 to 0.1, NNT 8 CI 5 to 27). When analysed by dose, greatest effect occurred at 15 mg/day olanzapine (1 RCT n=137, RR 0.74 CI 0.6 to 0.1, NNT 5 CI 3 to 42), was smaller for 10 mg/day (2 RCTs, n=232, RR 0.8 CI 0.7 to 0.1, NNT 7 CI 4 to 40) and lost in the single study that reported data on 5 mg/day (1 RCT, n=133, RR 0.94 CI 0.7‐1.2) and 1 mg/day (1 RCT, n=102, RR 0.98 CI 0.9 to 1.1).

2.1.2 Change in overall state as measured at endpoint by CGI (Clinical Global Impression) 
 CGI continuous data shows a significant difference in favour of olanzapine (2 RCTs, n=230, WMD (weighted mean difference) ‐0.53, CI ‐0.89 to ‐0.17).

2.1.3 Needing additional benzodiazepines 
 At six weeks, similar proportions of each group had received benzodiazepines as an additional medication (2 RCTs, n=418, RR 0.96 CI 0.9 to 1.1). 
 
 2.2 Mental state 
 2.2.1 Total scores 
 At six weeks global measures of mental state favoured olanzapine over placebo. Both non‐skewed BPRS total and PANSS total data favoured olanzapine (by about seven points on BPRS (1 RCT, n=98, MD ‐6.86 CI ‐12.3 to ‐1.38) and 12 points (1 RCT, n=98, MD ‐12.43 CI ‐22.54to ‐2.32) on PANSS. Beasley 1996a (HGAD) reported skewed data for the BPRS scale. These are not graphically displayed.

2.2.2 Negative/positive sub‐scores 
 The SANS, a scale specifically designed and validated for reporting negative symptoms showed no difference in olanzapine and placebo groups (1 RCT, n=94, MD ‐0.50, CI ‐2.87 to 1.87). The PANSS negative sub‐scale score was also equivocal. Skewed data from the negative sub‐scores of the BPRS and SANS tended to favour olanzapine but differences were not significant.

Positive symptoms were measured using sub‐scores of the BPRS and PANSS. Skewed data from the former are difficult to interpret but one study reported PANSS data that suggest a statistically significant advantage for olanzapine of four points (n=98, MD ‐4.0 CI ‐7.10 to ‐0.90). 
 
 2.3 Leaving the study early 
 2.3.1 Any reason 
 In the short term (six weeks), the levels of attrition reported by Beasley 1996a (HGAD), Beasley 1996b (HGAP) were similarly high, olanzapine (52%) placebo (53%). Pooling the data favoured olanzapine (3 RCTs, n=598, RR 0.86 CI 0.75 to 0.98, NNT 14 CI 8 to 95). Nevertheless, these attrition rates are considerable. Beasley 1996a (HGAD) reported a maintenance phase. This study counted people as 'dropouts' if they were ineligible, or eligible and unwilling, to continue the study medication beyond the acute phase (six weeks). At the start of the maintenance phase, 153/198 of those taking olanzapine and 55/68 had left the study (RR 0.96 CI 0.8 to 1.1). At 52 weeks only 17 people on olanzapine and two on placebo were continuing within their allocated groups (RR 0.94 CI 0.9 to 1.0). The data at six months are equivocal from Lecrubier 1999.

2.3.2 Due to lack of efficacy 
 Participants taking olanzapine were significantly less likely to leave studies early due to lack of efficacy at six weeks (RR 0.61 CI 0.43 to 0.89, NNT 7 CI 5 to124). Lecrubier 1999 data at six months are equivocal. At one year those who entered the maintenance phase of the original six‐week acute phase study (Beasley 1996a (HGAD)), however, were significantly less likely to leave the study early due to lack of efficacy (n=58, RR 0.41 CI 0.2 to 0.9, NNT 3 CI 2 to 52).

2.4 Adverse effects 
 We did not agree with some of the trialists' criteria for adverse events. For example 'exacerbation of schizophrenia' is classed as adverse event in Beasley 1996b (HGAP). This is reported as 'lack of efficacy' in this review and strengthens the assertion that adding those who left the study early to the adverse outcome may be justified. In these trials, adverse events needed to reach a frequency of at least 10% or be statistically significant before being reported. As a result of this, less common, but possibly important adverse effects were not reported.

2.4.1 Anticholinergic effects 
 Beasley 1996a (HGAD) reported that dry mouth and dizziness, especially, were more frequent in the olanzapine treated group, although neither difference reached conventional levels of statistical significance.

2.4.2 Extrapyramidal effects 
 The need for anticholinergic medication, used as a proxy‐measure for extrapyramidal adverse effects, was the same in those allocated to olanzapine and those taking placebo (n=418, RR 0.90 CI 0.3 to 2.8). Scales that provide normally distributed data measured akathisia, dyskinesia and parkinsonism. Againthere are no clear differences between groups.

2.4.3 Other effects 
 In the doses used within these studies, olanzapine did not clearly increase nausea or vomiting, problems with sedation or agitation, hostility and withdrawal. Again, it is debatable that these should be reported as treatment emergent adverse effects. Weight was increased by six weeks in the olanzapine group but not to a statistically significant extent (2 RCTs, n=227, WMD 3.6kg CI ‐1.2 to 8.4) and at six months there was no clear difference in weight gain.

2.5 Quality of life 
 Hamilton 1998 recorded, but did not report, usable quality of life data from Beasley 1996a (HGAD). Lecrubier 1999, however, did present some useful information suggesting that the olanzapine group did report a higher quality of life than those allocated placebo. This did not reach conventional levels of statistical significance (n=80, MD 8.2 CI ‐2.5 to 18.9).

2.6 Missing outcomes 
 Data were collected regarding inpatient or outpatient status, but again they reported no data. No study reported mortality, cognitive functioning, satisfaction with treatment, cost effectiveness, social functioning or self‐harm.

3. COMPARISON 2. OLANZAPINE versus PLACEBO (three to five days)

Tollefson 1999(HGDY) reported very short‐term data and it did not seem reasonable to synthesize these with the data presented above in comparison number one as these record outcomes at about six weeks. It appears to be a switching study from a group taking clozapine to either olanzapine or placebo.

3.1 Global effect 
 Tollefson 1999(HGDY) did not achieve conventional levels of statistical significance for the outcome of change in the CGI severity scale (n=106, MD ‐0.23 CI ‐0.52 to 0.06). However, for the outcome of relapse at five days, they did find a difference in favour of olanzapine, although we are unclear what a relapse at five days means (n=106, RR 0.13 CI 0.02 to 0.96).

3.2 Mental state 
 For the outcome of change in PANSS total scores at five days, results favour olanzapine (n=106, MD ‐5.23 CI ‐9.98 to ‐0.48), however, on positive and negative symptoms, results were equivocal. Tollefson 1999(HGDY) reported improvement on change scores using the MADRS for those taking olanzapine at five days (n=106, MD ‐4.52 CI ‐6.99 to ‐2.05), but found no difference on the MMSE.

3.3 Leaving the study early. Any reason 
 Equivocal results for leaving the study early are reported in this very short term study (n=105, RR 0.38 CI 0.11 to 1.34).

3.4 Adverse events 
 Equivocal five day results for headache and gastrointestinal adverse events are reported.

4. COMPARISON 3. OLANZAPINE versus TYPICAL ANTIPSYCHOTICS

Twenty nine trials compared olanzapine with typical antipsychotics. Altamura 1999 (HGBQ), Avasthi 2001, Barak 2002, Beasley 1996a (HGAD), Beasley 1997 (E003), Bernardo 2001 (HGDD), Chaudhry 2003, de Haan 2003, Esel 2001, Ishigooka 2001, Jones 1998 (P022), Lieberman 2003 (HGDH, Malyarov 1999, Namjoshi 2002; Rosenheck 2003(HGFI), Tollefson 1997(HGAJ), HGCJ (Hong Kong), HGCU (Taiwan) 1998, HGFH (Korea) 1998, used haloperidol as the comparator, while Chang 2003, Conley 1998, He 2003, Loza 1999 (HGDT), HGDV (Morocco) 1999, HGCQ (Turkey) 2000 used chlorpromazine. Jakovljevic1999 HGCH compared olanzapine with fluphenazine, HGBJ (Finland) with perphenazine and HGBL 1997, flupentixol. Lima 2003 (HGHS) compared olanzapine with typical antipsychotics decided upon by the clinicians involved (mainly haloperidol or chlorpromazine).

4.1 Death 
 For the outcome of death, data were reported in only one study (Ishigooka 2001) and supplied by Eli Lilly for (Beasley 1997 (E003) and Tollefson 1997(HGAJ)). Death from 'all causes' was not more likely in either group (2 RCTs, n=2427, RR 1.84 CI 0.48 to 7.11). Death from suicide and accidental injury (3 RCTs, n=2609, RR 1.28 CI 0.17 to 9.70), 'other causes' (2 RCTs, n=2427, RR 1.48 CI 0.15 to 14.22) and cause 'unknown' (1 RCT, n=1996, RR 0.16 CI 0.01 to 4.04) were all equally likely whether allocated to olanzapine or the typical drug.

4.2 Global effect 
 4.2.1 No important clinical response 
 Data relevant to this outcome are not available for many of the studies listed above (see 'Description of studies'). At six to eight weeks, Beasley 1996a (HGAD), Beasley 1997 (E003) and Tollefson 1997(HGAJ) defined 'important clinical response' as those who presented at least a 40% reduction in psychotic symptoms as measured by any scale. Conley 1998, for the treatment resistant participants, defined 'important clinical response' as at least a 20% reduction in mental state ratings. The large Tollefson 1997(HGAJ) trial dominates heterogeneous results for 'any dose olanzapine', but if random effects relative risk is calculated, there is no clear difference between the typical drugs and olanzapine for this outcome (4 RCTs, n=2778, RR 0.93 CI 0.85 to 1.01, Chi square 21.8, df=3). In an earlier version of this review, we calculated the effect size using Peto Odds and a fixed effect model produced an unequivocal result favouring olanzapine (Peto OR 0.6 CI 0.5 to 0.7, NNT 7 CI 5.3 to 8.8, Chi‐square 13.5, df 3). The random effects relative risk data, especially in the presence of heterogeneity and common events, are likely to be the more reliable (see 'Discussion'). Data relating to specific doses do not show differences between olanzapine and typical drugs. At 12 weeks Lieberman 2003 (HGDH) reports equivocal data (n=263, RR 0.82 CI .64 to 1.06). At nine months Lima 2003 (HGHS) also defined 'important clinical response' as those who presented at least a 40% reduction in psychotic symptoms on the PANSS (n=197, RR 0.85 CI 0.69 to 1.05), or a endpoint CGI score of </=2, (n=197, RR 0.82, CI 0.67 to 1.00).

4.2.2 Relapse/rehospitalisation 
 Only Beasley 1996a (HGAD) and Namjoshi 2002 reported hopitalisation rates which were higher in the olanzapine arm, but the drop outs in the olanzapine arm were >50% and prone to bias and data are heterogeneous (I‐squared 96%).

4.2.3 Clinical Global Impression endpoint score 
 In the short term, endpoint data do favour olanzapine over typical drugs (9 RCTs, n=2503, WMD ‐0.31 CI ‐0.57o ‐0.06) as do three to twelve month data (5 RCTs, n=411, WMD ‐0.46 CI ‐0.83 to ‐0.09).

4.3 Mental state 
 4.3.1 Needing additional medication 
 Less people taking olanzapine required additional benzodiazepines than those taking haloperidol (3 RCTs, n=2526, RR 0.91 CI 0.85 to 0.96, NNT 17 CI 10 to 37).

4.3.2 Total scores 
 Bernardo 2001 (HGDD), Chang 2003, Conley 1998, Esel 2001, Ishigooka 2001 and Loza 1999 (HGDT) presented endpoint data for BPRS total scores that were equivocal between olanzapine and typical antipsychotics (6 RCTs, n=397, WMD 0.00 CI ‐3.38 to 3.38). Much of the skewed BPRS data, presented in 'Other data' tables, are also equivocal. PANSS total scores however, clearly favour olanzapine. Outcomes reported for the short and medium term significantly favour olanzapine (8 RCTs, n=2427, WMD ‐6.04 CI ‐11.11 to ‐.96; 6 RCTs n=563, WMD ‐8.48 CI ‐13.84 to ‐3.12 respectively). Longer‐term data are again equivocal but are based on small numbers (n=44, Jones 1998 (P022)).

4.3.3 Negative symptoms 
 PANSS negative sub‐score data are consistently in favour of olanzapine (short term 4 RCTs, n=2297, WMD ‐1.20 CI ‐2.37 to ‐0.02; medium term 5 RCTs, n=408, WMD ‐2.11 CI ‐3.34 to ‐0.88; long term 1 RCT, n=44 MD ‐4.75 CI ‐8.10 to ‐1.40). When measured using BPRS negative sub‐scores and the SANS, all data still favoured olanzapine although the clinical significance of these data is not clear (see Discussion). 
 
 4.3.4 Positive symptoms 
 When olanzapine was compared with typicals and rated using the BPRS positive symptom sub‐score trialists found no clear difference (short term, 2 RCTs, n=110, WMD ‐0.97 CI ‐2.21 to 0.28)). PANSS positive sub‐score rating in the short term was in favour of olanzapine (4 RCTs, n=2297, WMD ‐1.03 CI ‐1.63 to ‐0.43). In the medium term however, this was not continued (3 RCTs , n=334, WMD ‐2.20 CI ‐5.42 to 1.02 respectively). Long‐term study data (Jones 1998 (P022) were not significant. Skewed data tended to favour olanzapine.

4.3.5 Depression 
 Avasthi 2001, Tollefson 1997(HGAJ), HGCU (Taiwan) 1998 and HGCJ (Hong Kong) rated depression using the Montgomery and Asberg Depression Scale. All data are skewed but tend to be moderately in favour of olanzapine. They are reported in 'Other data' tables.

4.3.6 Subjective well being 
 No difference was found between the two groups in this small study (de Haan 2003).

4.4 Leaving the study early 
 4.4.1 Any reason 
 The results for this outcome (at four to eight weeks) vary in terms of whether they are statistically significant depending on the statistical test applied to them. The conservative random effects relative risk suggests that the difference is not statistically significant (14 RCTs, n=3344, RR 0.81 CI 0.65 to 1.02). If, however, a Peto odds ratio is employed, the results are statistically significant (14 RCTs, n=3344, OR 0.54 CI 0.46 to 0.63, NNT 8 CI 6 to 10). We suggest that the conservative approach, in this high event rate group, is more likely to be accurate (see 'Discussion'). In any event, by four/eight weeks, 34% in the olanzapine group had dropped out compared to 45% in the comparator groups. By three to 12 months, significantly fewer people had left the olanzapine group (11 RCTs, n=1084, RR 0.79 CI 0.65 to 0.95). Beyond 12 months the findings were again equivocal (n=2738, RR 0.93 CI 0.86 to1.02). At this stage 83% of those originally allocated to olanzapine had left versus 90% randomised to typical drugs.

4.4.2 Due to lack of efficacy 
 When lack of efficacy is cited as reason for attrition there is a difference at three to 12 months in favour of those taking olanzapine (6 RCTs, n=482 RR 0.57 CI 0.34 to 0.96, NNT 16 CI 11 to 167). However, there is no difference between olanzapine and other drugs up to 8 weeks and over 12 months.

4.5 Adverse effects 
 4.5.1 Anticholinergic effects ‐ specific symptoms 
 Fewer people taking olanzapine experienced orthostatic changes (2 RCTs, n=82, RR 0.13 CI 0.1 to 0.4) and hypersalivation (4 RCTs, n=2636, RR 0.37 CI 0.20 to 0.71), than those on typical drugs. Other effects such as blurred vision, constipation, dry mouth, dizziness and urination difficulties were equivocal.

4.5.2 Extrapyramidal effects 
 Beasley 1997 (E003), Conley 1998 and Tollefson 1997(HGAJ) report on the incidence of any extrapyramidal event. Results favour olanzapine (3 RCTs, n=2511, RR 0.41 CI 0.28 to 0.60). Heterogeneous ESRS data (Chi squared=57.25 df=2) was equivocal (4 RCTs, n=139, WMD ‐3.62 CI ‐8.11 to 0.87). Those taking olanzapine required less anticholinergic medication compared with those taking haloperidol at six to 12 weeks (4 RCTs, n=2957, RR 0.28 CI 0.21 to 0.38, NNT 3 CI 3 to 4), and by three to 12 months (n=276, RR 0.55 CI 0.35 to 0.86, NNT 8 CI 6 to 24). Data on a whole series of specific extrapyramidal symptoms favours olanzapine. Participants taking olanzapine experienced less acute dystonia, akathisia, bradykinesia, hypertonia, hypokinesia, new onset parkinsonism, rigidity, and tremor than people allocated to haloperidol. For the outcomes of dyskinetic movements and choreoathetosis data were equivocal. Interestingly, in the single study (Conley 1998) with chlorpromazine as the comparator and reporting data for this outcome, the superiority of olanzapine is less convincing. Lieberman 2003 (HGDH) reports that fewer participants on olanzapine required propranolol. Continuous albeit skewed rating‐scale data for akathisia, dyskinesia and parkinsonism all favour olanzapine but the statistical or clinical significance is unclear (see 'Discussion'). Bernardo 2001 (HGDD) provided skewed data for the Simpson and Angus Scale and the Barnes Akathisia Scale which are not reported, due to the large SD's. Ishigooka 2001 reported extrapyramidal data from the DIEPSS scale; all were skewed and are shown in 'Other Data' tables.

4.5.3 Other adverse effects 
 Ishigooka 2001 reported equivocal data for adverse drug reactions which required dose reduction or additional medication. Data from Ishigooka 2001 for the outcome of anorexia favoured olanzapine (n=182, RR 0.26 CI 0.09 to 0.74). Tollefson 1997(HGAJ) reported that people taking olanzapine had an increase in appetite (n=1996, RR 1.65 CI 1.35 to 2.01, NNH 10 CI 7 to15). Reports of nausea were equivocal while data on vomiting favoured olanzapine over haloperidol (n=1996, RR 0.59 CI 0.45 to 0.78).

4.5.3.1 Drowsiness ‐ There is no suggestion of a problem of insomnia with olanzapine. However, those taking typicals are more likely to be drowsy than those on olanzapine, (6 RCTs, n=2636, RR 0.82 CI 0.73 to 0.93, NNH 18 CI 12 to 57). Continuous data on the Leeds Sleep Evaluation Questionnaire were equivocal between those on olanzapine and chlorpromazine.

4.5.3.2 Weight changes ‐ This update of the review contains some new data on weight changes. Twenty of 29 studies report weight data in various ways, not all of them amenable to comparison. Data for the short term are inconclusive (8 RCTs, n=2455, WMD 0.81kg CI ‐0.56 to 2.18) but the three to 12 month results suggest an average gain of four kilograms (4 RCTs, n=186, WMD 4.62, CI 0.6 to 8.64). Two studies, Barak 2002 and Lieberman 2003 (HGDH), report change data suggesting an increase in weight of 3.44 Kgs at three months (2 RCTS, n=283, WMD 3.44 CI 2.57 to 4.31). Five studies (Avasthi 2001, Chaudhry 2003, Ishigooka 2001, Lieberman 2003 (HGDH), He 2003) report heterogeneous results for weight gain (medium term) which was significantly higher for those receiving olanzapine (5 RCTs, n=428, RR 1.78 CI 1.44 to 2.19, NNH 6 CI 4 to 11), but most authors did not specify the criteria used for weight increase. One study reported this to be any weight increase >= 7%. Ishigooka 2001 also reported on weight loss with equivocal results (p=0.06). Weight changes reported beyond 12 months were equivocal (1 RCT, Jones 1998 (P022)).

4.5.3.3 Other effects ‐ One study reported treatment emergent agitation, hostility and withdrawn behaviour (Beasley 1996a (HGAD)) and found no differences between olanzapine and haloperidol. Also, Beasley 1996a (HGAD) reported equivocal data for gastrointestinal disturbances for the olanzapine and haloperidol groups. One study, HGCQ (Turkey) 2000, reported on rash with equivocal finding for both olanzapine and chlorpromazine. Avasthi 2001 reported equivocal data for asthenia. Ishigooka 2001 reported abnormal gait which appeared less common in the olanzapine group (n=182, RR 0.11 CI 0.3 to 0.5 NNH 6 CI 6 to 10). Esel 2001 reports laboratory results for prolactin levels at 6 weeks in favour of olanzapine (n=29, MD ‐15.01 CI ‐17‐94 to ‐12.08).

4.6 Quality of life 
 Beasley 1996a (HGAD) and Tollefson 1997(HGAJ) both reported data on quality of life in a form that we could not use in this review.

4.7 Economic outcomes 
 Three studies reported cost data, however, none were in a usable form (Jones 1998 (P022), Namjoshi 2002 and Rosenheck 2003(HGFI).

4.8 Missing outcomes 
 No data are reported for social functioning, employment status, death and family satisfaction with care.

5. COMPARISON 4. OLANZAPINE versus TYPICAL ANTIPSYCHOTIC FOR THOSE WITH TREATMENT RESISTANT ILLNESS

Conley 1998, compared olanzapine with chlorpromazine in hospitalised treatment resistant patients in an eight‐week trial and Altamura 1999 (HGBQ) compared olanzapine with haloperidol in a 14 week trial, defining non‐responders according to Breier 1994; essentially those with treatment resistant illnesses. Many outcomes were not reported by Altamura 1999 (HGBQ). All participants were inpatients in Conley 1998, however, Altamura 1999 (HGBQ) did not describe the setting of the study.

5.1 Global effect 
 5.1.1 No important clinical response 
 Conley 1998 defined this as the number of people per group who did not present at least a 20% reduction in psychotic symptoms measured by any scale. At eight weeks thee was no difference reported between the two groups (n=84, RR 0.93 CI 0.85 to 1.01).

5.1.2 CGI (Clinical Global Impression) 
 No difference between groups was apparent in Conley 1998 (n=81, MD ‐0.10 CI ‐0.49 to 0.29).

5.2 Mental state 
 5.2.1 Total score 
 Conley 1998 reported no discernible difference between the olanzapine and chlorpromazine groups on the total score of the BPRS (n=84, MD ‐2.80 CI ‐8.43 to 2.83).

5.2.2 Negative symptoms 
 Skewed SANS data favoured olanzapine but there were only 0.9 points difference between groups and the statistical and clinical significance were unclear.

5.2.3 Positive symptoms 
 Parametric BPRS positive sub‐scale ratings favoured olanzapine but not to a degree that was statistically significant (n=84, MD ‐1.30 CI ‐2.90 to 0.30).

5.3 Leaving the study early 
 5.3.1 Any reason 
 No differences between those allocated olanzapine and people taking either chlorpromazine or haloperidol were apparent (Altamura 1999 (HGBQ) and Conley 1998).

5.3.2 Due to lack of efficacy 
 No difference was reported between the groups in the two trials.

5.4 Adverse events 
 Essentially the data for adverse effects reflect those already presented for the larger comparison between olanzapine and typical antipsychotic drugs for those whose illness is not treatment resistant.

5.4.1 Anticholinergic effects 
 Conley 1998 reported fewer people on olanzapine experienced dry mouth (RR 0.52 CI 0.3 to 0.8, NNH 3 CI ‐6.3 to ‐1.8). Blurred vision, dizziness and palpitations were equivocal. Orthostatic changes (1 RCT, n=84, RR 0.13 CI 0.13 to 0.35, NNT 3CI 2 to 3) and dry mouth (1 RCT, n=84, RR 0.52 CI 0.3 to 0.8, NNT 3 CI 3 to 7) both favoured olanzapine to chlorpromazine.

5.4.2 Extrapyramidal events 
 In Conley 1998 fewer people taking olanzapine experienced 'any extrapyramidal effects' (n=84, RR 0.57 CI 0.3 to1.0) or dyskinetic movements (n=84, RR 0.10 CI 0.01 to 0.5, NNH 3 CI ‐5.5 to ‐2.1). Skewed data from both the Barnes Akathisia Scale and Simpson and Angus (parkinsonism) Scale were uninformative but did not highlight clear differences between groups.

5.4.3 Other adverse events 
 Those allocated to olanzapine and chlorpromazine reported similar rates of nausea/vomiting and difficulty in getting to sleep or drowsiness. Altamura 1999 (HGBQ) (n=47) reported body weight at the end of the 14 week study. No suggestion of a difference between those randomised to olanzapine and those allocated haloperidol were seen at this time (MD 0.35kg CI ‐9.7 to 10.4). 
 
 5.4.4 Missing outcomes 
 No data were reported on quality of life, social functioning, employment, death and harm, cognitive functioning, satisfaction with treatment and cost effectiveness. Trials supplied no clear data on the use of additional medication but Conley 1998 stated "no difference between the groups in the amount and duration of lorazepam use during the study".

6. COMPARISON 5. OLANZAPINE versus ATYPICAL ANTIPSYCHOTICS

Twenty seven trials compared olanzapine with atypical antipsychotics (Beuzen 1998 (HGCF), Bitter 2004 (HGCK), Casey 2003, Chan 2003, Chen 2003, Conley 2001, Corrigan 2004, de Haan 2002, Jeste 2003, Jones 1998 (P022), Kelly 2003, Kern 2001, Lecrubier 1999, Littrell 1999, Malyarov 1999, Martin 2002, McQuade 2003, Meltzer (InterSept), Naber 2001 (HGBF), Namjoshi 2002, Ritchie 2003, Simpson 2004, Svestka 2003, Thomas 1998 (HGBU), Tran 1997 (HGBG), Wang 2002, Wang 2003). Thirteen trials compared olanzapine with risperidone, eight with clozapine, two with amisulpiride, two with aripiprazole and one each with sonepiprazole and zisprasidone. Some of the trials relate to those with treatment resistant illness (Beuzen 1998 (HGCF), Bitter 2004 (HGCK), Kelly 2003 and Naber 2001 (HGBF)), people in their first episode (Svestka 2003) and the elderly (Jeste 2003, Ritchie 2003).

6.1 Death 
 Unpublished data from Beuzen 1998 (HGCF) and Conley 2001 did not reveal deaths in either group. Three other studies reported mortality data; one whose primary aim was to measure if olanzapine vs clozapine prevented death (Meltzer (InterSept)), and two reporting on interventions in older people (Jeste 2003, Ritchie 2003). No statistical difference was detected between the two groups (any cause, 5 RCTs, n=1778, RR 0.70 CI 0.31 to 1.60; by suicide, 1 RCT, n=980, RR 1.67 CI 0.40 to 6.94).

6.2 Global 
 6.2.1 Global effect 
 No difference was found for the outcome of no important clinical response (as measured by CGI, or by Kane 1988 criteria), when olanzapine was compared to other atypicals at eight weeks or three months, (7 RCTs).

6.2.2 Relapse/rehospitalisation 
 Namjoshi 2002 and Meltzer (InterSept) were the only studies to provide data on rehospitalisation, however the Meltzer (InterSept) data relates to this being used as a rescue intervention to prevent suicide and is reported in this comparison as a service outcome (no 40 see later). By 12 months the number of people rehospitalised was significantly higher in the risperidone group (n=278, RR 0.49 CI 0.31 to 0.79, NNT 7 CI 5 to 16; Namjoshi 2002).

6.2.3 CGI endpoint score 
 Pooled results from three trials are equivocal (3 RCTs, n=570, WMD 0.00 CI ‐0.20 to 0.19).

6.3 Mental state 
 6.3.1 Needing additional benzodiazepines 
 Pooled data at six to eight weeks, and by eighteen weeks, shows similar numbers of participants receiving olanzapine or other atypicals (clozapine, risperidone or ziprasidone) required additional benzodiazepines. At two years, Meltzer (InterSept) reports a barely significant result in favour of clozapine (n=980, RR 1.11 CI 1.01 to 1.22, NNH 16 8 to 167).

6.3.2 Needing an antidepressant by two years 
 Meltzer (InterSept) reports on data from two years that more participants on olanzapine needed an antidepressant than people allocated clozapine, however, this did not reach statistical significance (n=980, RR 1.12 CI 0.99 to 1.27). The same study reported antidepressants being used as a rescue intervention to prevent suicide (n=980, RR 2.20 CI 1.21 to 4.00, NNH 26 CI 11 to 145).

6.3.3 No important clinical response (20 and 40% decrease in PANSS total score; 20 and 40% decrease in BPRS total score) 
 Short term data from four studies (Chen 2003, Conley 2001, Jeste 2003, Wang 2003) were equivocal for the olanzapine and atypical groups (risperidone or clozapine). The criterion for clinical improvement was 20% to 25% or more reduction in total PANSS score. Svestka 2003 used a 50% reduction in total PANSS score over six weeks in people who were experiencing their first episode of psychosis. Results are again equivocal. Medium term outcomes by (Beuzen 1998 (HGCF), Bitter 2004 (HGCK) and (Tran 1997 (HGBG)), favoured olanzapine to the other atypical antipsychotic groups (clozapine and risperidone, 3 RCTs, n=669, RR 0.89 CI 0.81 to 0.99, NNT 13 CI 8 to 139) using a 40% or more cut off point. BPRS data at six weeks from Simpson 2004 and Wang 2002 did not distinguish between the two groups.

6.3.3 Total endpoint scores 
 BPRS endpoint data from the clozapine (Beuzen 1998 (HGCF) and risperidone (Tran 1997 (HGBG), Thomas 1998 (HGBU) comparisons were skewed and difficult to interpret. All findings favour olanzapine but by very minimal differences. Wang 2002 reports a statistical significance on BPRS endpoint score in favour of olanzapine (n=61, MD ‐1.60 CI ‐2.96 to ‐0.24). Martin 2002 reported equivocal BPRS end‐point data for the medium term (n=377) between olanzapine and amisulpride. PANSS total scores, at three to 12 months are heterogeneous (Chi squared 37.44 df=4), and equivocal (5 RCTs, n=715, WMD 3.01 CI ‐5.74 to 11.76). Twelve months data are also equivocal (n=42 MD 1.86 CI ‐9.08 to 12.8).

6.3.4 Negative symptoms 
 No difference was seen between olanzapine and risperidone at eight weeks (n=278, MD 0.20 CI ‐1.1 to 1.5). Endpoint PANSS negative sub‐scale scores are similar between atypicals at 3‐12 months (4 RCTs, n=593, WMD ‐1.52 CI ‐2.08 to 5.13) and over 12 months (n=42, MD ‐0.20 CI ‐3.76 to 3.36). Lecrubier 1999 (amisulpiride) reported endpoint SANS scores as identical. Other skewed data support the parametric findings.

6.3.5 Positive symptoms 
 Both skewed BPRS and SANS data and parametric PANSS positive symptom scores do not suggest any differences between olanzapine and other atypical antipsychotics. 
 
 6.3.6 Depression 
 We entered skewed data from the HAM‐D (Conley 2001) and MADRS (Martin 2002) in 'Other data' tables, and they do not suggest any difference between those taking olanzapine, risperidone or amisulpiride.

6.4 Leaving the study early 
 6.4.1 Any reason 
 There was no difference in the numbers of people leaving the study early for either short, medium or long‐term studies. By eight weeks, 21% of those taking olanzapine had left the study, compared to 25% taking the other atypicals. By three to 12 months, this had increased to 46% and 52% respectively.

6.4.2 Due to lack of efficacy 
 There was no notable difference between atypical drugs.

6.5 Adverse effects 
 6.5.1 Anticholinergic effects ‐ specific symptoms 
 Statistically significant differences were found for the outcomes of constipation by six to eight weeks (2 RCTs, n=247, RR 0.47 CI 0.25 to 0.89, NNH 9 CI 7 to 44), 18 weeks (n=180, RR 0.35 CI 0.15 to 0.85, NNT 9 CI 7 to 36) and by two years (n=980, RR 0.45 CI 0.34 to 0.60, NNT 7 CI 6 to 10). No difference was detected for the outcome of dizziness at six to eight, or by 18 weeks, however, at two years, less people on olanzapine experienced dizziness (n=480, RR 0.51 CI 0.40 to 0.66, NNH 8 CI 6 to 11), did not report a problem with libido (n=66, RR 0.78 CI 0.61 to 0.99, NNT 5 CI 3 to 110). At all time points, those taking olanzapine experienced less salivation (6 weeks, n=121, RR 0.02 CI 0.00 to 0.36, NNT 3 CI 3 to 5; by 18 weeks, 2 RCTs, n=330, RR 0.32 CI 0.15 to 0.67, NNT 4 CI 3 to 8; by 2 years, n=980, RR 0.17 CI 0.13 to 0.23, NNT 3 CI 3 to 3) and urinary incontinence (n=980, RR 0.45 CI 0.27 to 0.77, NNT 21 CI 16 to 49) and salivation (3 RCTs, n=1192, RR 0.20 CI 0.06 to 0.63, NNT 3 CI 3 to 7). Participants taking olanzapine were more likely to experience dry mouth by two years (n=490, RR 1.51 CI 1.02 to 2.55, NNH 12 CI 7 to 715). Data for vision abnormalities at eight weeks and rhinitis were equivocal.

6.5.2 Extrapyramidal effects 
 6.5.2.1 Needing additional anticholinergic medication 
 At both eight to 12 weeks and three to 12 months, significantly fewer people taking olanzapine had received anticholinergic medication compared to those taking comparators, (4 RCTs, n=923, RR 0.64 CI 0.46 to 0.91, NNT 7 CI 5 to 27; 2 RCTs, n=519, RR 0.56 CI 0.4 to 0.79, NNH 9 CI 7 to 18) respectively.

6.5.2.2 Any extrapyramidal symptoms 
 Dichotomised data show no difference at eight weeks (4 RCT's, n=657, RR 0.76 CI 0.54 to 1.08), however at three to 12 months, fewer participants taking olanzapine reported any extrapyramidial symptoms (n=339, RR 0.6 CI 0.41 to 0.88, NNT 9 CI 6 to 27).

6.5.2.3 Parkinsonism 
 Skewed data from two studies using the ESRS (Conley 2001, Jones 1998 (P022)), when pooled, gave equivocal results, as does Martin 2002 which uses the SAS.

6.5.2.4 Specific extrapyramidal symptoms 
 Beuzen 1996 (HGCE) and Tran 1997 (HGBG) reported only those adverse events that were statistically significantly different between groups, whereas Conley 2001 only reported those occurring in 10% of either group. Beuzen 1998 (HGCF) reported results at a 5% incidence. Rates were similar for akathisia (by three to 12 months, 4 RCTs, n=924, RR 0.98 CI 0.64 to 1.49), dyskinetic movements (2 RCTs, n=519, RR 0.99 CI 0.40 to 2.46), restlessness (n=122, RR 1.00 CI 0.15 to 6.87) and tremor (2 RCTs, n=297, RR 0.54 CI 0.15 to 1.91). Akathisia at two years however appears more prevalent in those who have taken olanzapine (n=980, RR 1.63 CI1.07 to 2.48, NNH 27 CI 12 to 239). The incidence of parkinsonism appeared less in the olanzapine group (1 RCT, n=339, RR 0.58 CI 0.36 to 0.94, NNH 11 CI 6 to 75).

6.5.2.3 Other adverse effects 
 Treatment emergent adverse events were less common in the olanzapine group (3 RCTs, n=569, RR 0.9 CI 0.82 to 0.98, NNT 13 CI 8 to 64).

There was no difference between those on olanzapine and those on comparators in levels of asthenia at siz to eight weeks and 18 to 52 weeks. However at two years, those taking olanzapine were less likely to have asthenia than those on clozapine (n=980, RR 0.51 CI 0.32 to 0.82, NNT 21 CI 15 to 56), or low white cell counts by two years (n=980, RR 0.44 CI 0.25 to 0.76, NNT 23 CI 17 to 53), but not at four or 18 to 26 weeks. People in the olanzapine group were also less likely to have dyspepsia at two years (n=980, RR 0.69 CI 0.50 to 0.95, NNT 22 CI 14 to 134), headache at four to eight weeks (4 RCTs, n=677, RR 0.72 CI 0.54 to 0.96, NNT 15 CI 9 to 100), hyperlactinaemia by six weeks (n= 42, RR 0.24 CI 0.10 to 0.58, NNT 2 CI 2 to 3), nausea at 18 to 26 weeks and two years (n=435, RR 0.27 CI 0.09 to 0.75, NNT 10 CI 8 to 2; n=980, RR 0.67 CI 0.49 to 0.90, NNT 17 CI 11 to 56), tachycardia at six weeks (n= 122, RR 0.06 CI 0.01 to 0.40, NNT 4 CI 4 to 6) and vomiting at two years (n= 980, RR 0.61 CI 0.45 to 0.83, NNT 15 CI 11 to 33). The longest study, Meltzer (InterSept), found no difference between groups for the outcome of diabetes mellitus at two years. Equivocal results were reported for agitation, anxiety, fever, flu syndrome, increase in SGPT, increase in transaminase, sweating, tooth disorder and postural hypotension.

6.5.3 Sleep problems 
 No statistical differences were reported in the trials between olanzapine and risperidone at six to eight weeks for insomnia or somnolence. At 18 to 26 weeks, people on olanzapine had less insommia (n=255, RR 0.34 CI 0.20 to 0.60, NNT 5 CI 4 to 8), whilst those on olanzapine at two years were more likely to suffer greater insomnia (n=980, RR 1.57 CI 1.28 to 1.93, NNH 9 CI 6 to 18). At 18 to 26 weeks, there was no difference for the outcome of somnolence, whilst at two years, people on olanzapine experienced less somnolence (n=980, RR 0.57 CI 0.48 to 0.67, NNT 5 CI 5 to 7).

6.5.4 Weight 
 Differences in weight in the short and medium term were equivocal. However, at two years, significantly more participants had gained weight on olanzapine than clozapine (1 RCT, n=980, RR 1.73 CI 1.49 to 2.00, NNH 5 CI 4 to 7). End point data reported in three trials at three to 12 months, and one study at over 12 months were equivocal. However, change data in the short term revealed an increase in weight for those taking olanzapine compared to comparator (3 RCTs, n= 463, WMD 1.42 CI 0.62 to 2.23).

6.5.5 Other 
 A single study (de Haan 2002) reported equivocal skewed data at six weeks for the outcome of obsessive compulsive disorder.

6.6 Quality of life 
 Lecrubier 1999, Thomas 1998 (HGBU) and Tran 1997 (HGBG) employed two continuous data scales relating to quality of life and results were equivocal at three to 12 months

6.7 Service outcomes 
 Martin 2002 reported equivocal data for discharged from hospital (short term: two months). Meltzer (InterSept) reported equivocal results for an increase in levels of surveillance for those who were suicidal. Those taking olanzapine were more likely to be admitted to hospital as a rescue intervention than those on clozapine at two years (n=980, RR 1.28 CI 1.02 to 1.61, NNH 18 CI 9 to 251).

6.7 Missing data 
 No data were reported on additional medication (other than anticholinergics or benzodiazepines), social functioning, employment status, cognitive functioning and cost effectiveness. Data on satisfaction with treatment were provided by the Drug Awareness Inventory, but has yet to be included as we are still trying to identify the validity of the measure and its meaning.

7. COMPARISON 6. OLANZAPINE versus ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS

Four trials compare olanzapine to clozapine for people whose illness were resistant to treatment (Beuzen 1998 (HGCF), Bitter 2004 (HGCK), Kelly 2003, Naber 2001 (HGBF).

7.1 Death 
 Unpublished data supplied by the trialists of Beuzen 1998 (HGCF) reveal no deaths in either group (n=180).

7.2 Global effect ‐ No important response 
 Beuzen 1998 (HGCF) and Bitter 2004 (HGCK) used two criteria for improvement when comparing olanzapine with clozapine (CGI and Kane 1988 criteria). Results from both favour neither drug at 18 weeks.

7.3 Mental state 
 7.3.1 No important clinical response (20% reduction in PANSS total score) 
 There is no clear difference between olanzapine and clozapine (2 RCTs, n=330, RR 1.0 CI 0.68 to 1.49).

7.3.2 Total score 
 Beuzen 1998 (HGCF) reports BPRS and PANSS total endpoint scores. BPRS data are skewed and therefore problematic to present by graph, but do not suggest a clear difference between groups. PANSS total scores are equivocal (n=176, MD 0.03 CI ‐8.07 to 8.13).

7.3.3 Negative symptoms 
 When skewed negative sub‐scores are extracted from the BPRS, no differences between olanzapine and clozapine are revealed. PANSS negative sub‐scores are parametric and also show no clear differences (1 RCT, n=176, MD ‐0.75 CI ‐2.9 to 1.4).

7.3.4 Positive symptoms 
 Positive symptoms sub‐scores of the BPRS are also skewed and uninformative. Data are equivocal.

7.4 Leaving the study early 
 7.4.1 Any reason 
 Data are equivocal at eight weeks and by three to 12 months almost half had left both groups.

7.4.2 Due to lack of efficacy 
 Data from Beuzen 1998 (HGCF) and Bitter 2004 (HGCK) are equivocal.

7.5 Adverse effects 
 7.5.1 Anticholinergic effects 
 Results for salivation are equivocal. Constipation is less of a problem for people taking olanzapine compared with those allocated clozapine (1 RCT, n=180, RR 0.35 CI 0.15 to 0.85, NNT 9 CI 7 to 38).

7.5.2 Extrapyramidal effects 
 Mean data from the HAS and the AIMS and the SAS do not demonstrate a difference between groups, neither do the data on akathisia.

7.5.3 Other 
 Those on olanzapine suffer less nausea (n=180, RR 0.10 CI 0.01 to 0.77, NNT 11 CI 10 to 40) and somnolence (2 RCTs, n=310, RR 0.54 CI 0.32 to 0.88, NNT 10 CI 7 to 38) compared with clozapine. Results are equivocal for the outcomes of agitation, anxiety, asthenia, back pain, decrease in white blood cells, fever, flu syndrome, postural hypotension, sweating, tooth disorder, vomiting and weight gain.

8. COMPARISON 7. OLANZAPINE versus ATYPICAL ANTIPSYCHOTICS FOR FIRST EPISODE

One six week trial (Svestka 2003) with 42 females is included in this comparison.

8.1 Global effect ‐ No important response 
 Equivocal data from this one small study are reported.

8.2 Adverse events 
 8.2.1 Anticholinergic effects 
 Less people taking olanzapine complained of rhinitis (n=42, RR 0.13 CI 0.02 to 0.91, NNT 4 CI 3 to 30).

8.2.2 Extrapyramidal ‐ needing anticholinergic medication 
 Fewer participants taking olanzapine required anticholinergic medication (n=42, RR .38 CI 0.17 to 0.89, NNT 3 CI 2 to 15).

8.2.3 Extrapyramidal ‐ any report 
 People taking risperidone reported more extrapyramidal adverse events (n=42, RR 0.4 CI 0.19 to 0.83, NNT 3 CI 2 to 9). 
 
 8.2.4 Hyperprolactineamia 
 Hyperprolactineamia was more common in people taking risperidone (n=42, RR 0.24 CI 0.10 to 0.58, NNT 2 CI 2 to 3).

Discussion

1. Applicability of results 
 The largest trials were multi‐centre and conducted primarily within the western world, although smaller studies have been reported from countries such as Morocco and Taiwan. Participants had operationally diagnosed disorders in the main uncomplicated by co‐morbidity. The large Tollefson 1997(HGAJ) study, dominant within the important olanzapine versus typical antipsychotics comparison, however, appeared more representative of day‐to‐day practice because those with co‐morbid disorders were not excluded. More of the recently included studies come from China, Korea, and east Europe, with less originating from the USA and the UK. Meltzer (InterSept) is the longest study (two years), and uses outcomes based on more real world relevance to people with schizophrenia and clinicians alike. The inclusion criteria were more closely matched to those seen in clinical practice than in previous research trials. We welcome the inclusion of many more trials from China and are seeking further information on a number of these trials so we can include additional outcomes. Data which have been included from these trials show a marked difference in the ability of their participants to remain in the studies compared with the high attrition rates seen in the west. To summarise we get some impression of change over time with the setting, participants and outcomes of trials becoming slightly more like everyday practice. 
 
 2. Reporting of adverse event data 
 Most trials only reported adverse events if their incidence was either statistically significant or had a frequency of 10% or greater which results in rare adverse effects being ignored. If this policy had been implemented rigorously in other drug trials, the agranulocytosis of clozapine and ocular problems associated with thioridazine would have gone unreported. Eli Lilly has kindly provided us with unpublished COSTART adverse event data for many of the studies and so we are able to report more adverse effect data than is seen in the final reports of the studies. It is, however still very problematic that the >10% frequency rule for reporting of adverse effects is still used as a device in these influential trials.

3. Limited data 
 A great deal of global effect/mental state data are missing from the studies supplied by Eli Lilly and we have been informed that rights to these data are not available even to the company. Efforts have been made to contact authors to acquire the necessary information for inclusion in this review. It is frustrating that despite considerable investment in clinical trials, data are so limited on tolerability, social functioning, employment, death, legal problems, cognitive functioning and satisfaction. In addition, data on hospital admission, outpatient/day hospital visits seems to have been collected by several studies but only reported in a useable form in one study, Beasley 1996a (HGAD). We find it surprising that a compound worth in excess of $3.69bn per year (Sellers 2003) is so successful even though there is a lack of important data for people with schizophrenia, their families and professional carers.

4. The search 
 Researchers have studied a remarkably successful dissemination of results through academic presentations both at conferences and in journals. It is likely that even the hundreds of citations to included studies within this review represent only a proportion of presentations of these trials. Understandably the larger studies have attracted more publications or conference presentations. We have found no evidence of an attempt to conceal the fact that the hundreds of presentations relate to a limited number of studies but inclusion of the unique company trial codes in each presentation would have helped piece together the few sausages from the very thinly sliced salami.

Sponsorship or involvement with a pharmaceutical company was apparent in 38 of the 55 included studies (69%). Although high, this figure may be an underestimate, as in many reports, potential conflicts of interest or acknowledgements were not recorded. Sponsorship of trials by the pharmaceutical industry continues to cause concern (Offerhaus 2004) and the continued need for independent research is advocated (Hoadley 2004).

5. COMPARISON 1: OLANZAPINE versus PLACEBO 
 At the end of these trials the extent of those who had 'dropped out' (olanzapine 52%, placebo 53%) was of such a magnitude that all results, except for those of 'leaving the study early', are overly full of assumptions making it impossible to reach a confident interpretation.

5.1 Global effect 
 This outcome is based on only two studies. Beasley 1996a (HGAD), a study that did not extrapolate BPRS from the PANSS, gives 75% of the weighting to the final result. The fact that approximately eight people have to be treated with olanzapine, as compared with placebo, in order for one person to improve at six weeks is encouraging of olanzapines' antipsychotic properties. However, it would be more reassuring if this finding was based on more direct measures of all participants and was sustained in the long term. The CGI continuous data also favour olanzapine. Clinical interpretation of a weighted mean difference of ‐0.523 is, however, problematic. When analysed by dose, the effect was only seen with the 15 mg/day and 10 mg/day olanzapine groups.

5.2 Mental state 
 5.2.1 Total scores 
 Change in total BPRS and PANSS scores favour olanzapine over placebo. However, interpretation of an average change of seven BPRS points or 12 in the PANSS scoring is problematic, especially considering that over 50% of data are 'last‐observation‐carried‐forward',where the trialists assume that the person remains stable after their last rating. These scales also do not produce 'interval' data (so a 20% decline from 100 to 80 is not the same clinically as a 20% decline from a score of 50 to 40). These scales dont report clinically meaningful outcomes.

5.2.2 Negative/positive sub‐scores 
 The same applies to the negative and positive sub‐scores as measured with the BPRS and PANSS. The SANS data, for negative symptoms, were equivocal and other information sheds doubt on olanzapine's advantage over placebo for these difficult symptoms.

Incomplete data from ratings of positive symptoms suggest some effect for olanzapine. It is, however, unclear how to use these data to inform those with schizophrenia or their carers.

5.3 Leaving the study early 
 There was great attrition from these studies. Over a six‐week period, fewer of those given olanzapine left the study early (52%) than those given placebo (53%), NNT 14 CI 8 to 95, but the numbers are still high. Even if this result is secondary to clinical improvement, such a degree of loss must suggest that there was some factor within the design of the study that was unacceptable to participants. By one year almost everyone had left the key study in this area (Beasley 1996a (HGAD)) making any other data impossible to interpret. When lack of efficacy is cited as the reason for attrition, data are supportive of olanzapine's short term antipsychotic effects.

5.4 Adverse effects 
 Olanzapine has anticholinergic adverse effects, causing dry mouth and dizziness, but this is not significantly more apparent than placebo from analysis of the current data in this review. Specific extrapyramidal symptoms (akathisia and tremor) appeared higher in the olanzapine group, although this did not reach statistical or, probably, clinical significance. This is supported by the similar use of anticholinergic medication in the two groups. In addition all scale data relating to movement disorders showed no important differences between olanzapine and placebo. Data on nausea/vomiting and sleepiness are equivocal, but all are based on information that makes a great deal of assumptions.

Data supplied by Eli Lilly show that weight increases with olanzapine, even in short studies, although this did not reach conventional levels of statistical significance (at six weeks WMD 3.58kg CI ‐1.18 to 8.34). Nearly four kilograms in six weeks would probably be of clinical significance.

5.5 Quality of life 
 It is unfortunate that Beasley 1996a (HGAD) did not fully report quality of life data (Hamilton 1998). What limited data there are come from Lecrubier 1999. The data are interesting and should encourage Eli Lilly to find more powerful data for dissemination.

5.6 Missing outcomes 
 Data were collected regarding inpatient or outpatient status, but, again, the published reports do not tell us the results. This must undermine confident use of olanzapine.

6. COMPARISON 2. OLANZAPINE verus PLACEBO (three to five days)

6.1 Global effect 
 It is difficult to know what to make of many of the outcomes in this very short term study on a population taking clozapine. One could surmise that this is a treatment resistant or treatment intolerant group as they were switched from clozapine to olanzapine. (Clozapine is only licensed in this group.) It would be unreasonable to be looking for major change in mental state/behaviour in those diagnosed as acutely ill.

6.2 Mental state 
 PANSS total scores at five days favours olanzapine, however, the clinical significance of a mean difference of ‐5.23 on this scale is unclear, and results of the positive and negative subscales were equivocal. We find the utility of the MMSE for a group of people acutely ill with schizophrenia puzzling. Depression scores favoured olanzapine.

6.3 Leaving the study early 
 Unsurprisingly, few dropped out of this very short study.

6.4 Adverse events 
 Olanzapine‐treated participants reported similar levels of adverse events as those taking placebo.

6.5 Missing outcomes 
 The single trial did not report on hospitalisation status or quality of life.

7. COMPARISON 3: OLANZAPINE versus TYPICAL ANTIPSYCHOTICS

7.1 Death 
 Only one study reported this important outcome. Eli Lilly supplied us with data from a further two trials. Results indicate there is no substantive difference in the numbers of people who have died due to suicide, accidental injury or other causes between olanzapine and other typical antipsychotic.

7.2 Global effect 
 From data presented in this review there are no clear differences between olanzapine and typical drugs for the outcome of 'no important clinical response at six to eight weeks'. Using odds ratios does make a statistically significant difference, but in these heterogeneous data, random effects relative risk analyses are advisable. The latter makes the result not statistically significant and this is still true at twelve weeks and nine months.

CGI endpoint data favours olanzapine over the typical antipsychotics both at six weeks and at three to twelve months.

7.3 Mental state 
 People taking olanzapine required less additional benzodiazepines than those taking typical drugs implying that they are less disturbed as a result of taking the olanzapine.

Scale derived data differs in the presented outcomes. BPRS did not favour the olanzapine group when compared with typical antipsychotics (n=397), but more powerful PANSS data do favour olanzapine. Gains are an average of a decline of six PANSS points in the short term and medium term. The clinical relevance of these scores remains unclear. Whether the PANSS is an interval scale, with a six‐point decline having the same clinical meaning for someone who started with a score of 70, as for a person whose initial rating was 30, is not explained by the trials.

Both positive and negative symptom PANSS‐derived data favours olanzapine, which underlines the impression that olanzapine is at least as effective an antipsychotic as typical drugs. Montgomery and Asberg Depression Scale data were skewed and with the data provided in the studies, clinical interpretation of these scale data is not possible.

7.4 Leaving the study early 
 People allocated to olanzapine were less likely to leave the study than those taking typical antipsychotics; the results being statistically significant or not dependent on the analysis undertaken. We suggest the most conservative random effects relative risk is more applicable (and only statistically significant at three to twelve months). At four to eight weeks 34% of those taking olanzapine and 45% of those taking a typical antipsychotic had left, and beyond 52 weeks, this had risen to 83% and 90% respectively. Such attrition makes interpretation of all the other data virtually impossible. These data suggest that clinicians should be confident that all but a few of those given olanzapine will still be taking their medication at the end of a year, provided the same study conditions apply.

Supporting the unclear general effects, results for the comparison of olanzapine versus typical drugs, when lack of efficacy is cited as reason for attrition there are no differences between groups except at three to 12 months where olanzapine is favoured.

7.5 Adverse effects 
 Current data pertaining to anticholinergic adverse effects, suggests that olanzapine causes less hypersalivation and orthostatic changes than typical drugs. Other effects such as dry mouth, dizziness, palpitations, constipation, dry mouth and urination difficulties were as common as they were with the older drugs.

Data on a series of specific extrapyramidal symptoms favours olanzapine ‐ especially when compared to haloperidol. Within the chlorpromazine comparison the differences were less marked. Heterogeneous rating scale data do not suggest that extrapyramidal effects are less marked with olanzapine, but clinical interpretation of these heterogeneous data is not possible. We view the use of additional anticholinergic medication as a more useful marker of extrapyramidal effects (both for those taking the antipsychotics and their carers) and this is lower in those taking olanzapine than the typical drugs.

More people treated with olanzapine reported an increase in appetite compared to those taking conventional antipsychotics. This supports clinical experience. The data on weight gain is equivocal in the short‐term studies and shows a statistically significant difference in the medium term, with those on olanzapine gaining more weight. Jones 1998 (P022), a small study, is equivocal regarding weight gain in the long‐term and may be an example of a Type II Error.

7.6 Quality of life 
 It is disappointing that large and important studies, such as Beasley 1996a (HGAD) and Tollefson 1997(HGAJ), report data on quality of life in a form that is not possible to use in this review. We hope that further collaboration with Eli Lilly will shed light on these data.

7.7 Economic outcomes 
 Statistical significance favours olanzapine for cost effectiveness, but again no usable data are provided to support this claim.

7.8 Cognitive function 
 Jones 1998 (P022) reported data on cognitive functioning that are not possible to use in this review. As this topic gains interest, it is likely that more data will be forthcoming in future studies.

7.9 Missing outcomes 
 Conclusions cannot be drawn for outcomes of social functioning or satisfaction for this popular commonly used drug.

8. COMPARISON 4. OLANZAPINE versus TYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS

Both Conley 1998 and Altamura 1999 (HGBQ) compared olanzapine with typical antipsychotics (chlorpromazine and haloperidol respectively) in people whose illnesses were rated as treatment resistant. Olanzapine has been marketed as being structurally similar to clozapine; a drug with efficacy in treatment resistant patients (Wahlbeck 1999). The effects of olanzapine in this sub‐group of people with schizophrenia are important. Although both studies focused on treatment‐resistant participants, who would, by definition, have long‐standing illness, the studies were classified as short‐term (Conley 1998) and medium‐term Altamura 1999 (HGBQ). Altamura 1999 (HGBQ) did not describe the setting of the study.

8.1 Global effect 
 Conley 1998 defined 'no important clinical response' as the number of people per group who did not present at least a 20% reduction in psychotic symptoms measured by any scale. At eight weeks there was no difference between the two groups. The continuous measure (Clinical Global Impression) corroborated this finding. It appears that olanzapine has similar global effects to typical antipsychotics on those with treatment resistant illness but this may reflect a Type II Error and these findings require replication.

8.2 Mental state 
 In keeping with the equivocal result for 'global effect', mental state measures (total BPRS, SANS, positive symptom sub‐score BPRS) do not appear significantly different.

8.3 Leaving the study early 
 Similar numbers of people left the studies early for 'any reason' and due to 'lack of efficacy', although rates were less than for those with non‐treatment‐resistant illness. This may reflect the severity of their illness, and the likelihood that they are inpatients.

8.4 Adverse events 
 Data were only available from Conley 1998 on the outcome of anticholinergic effects, where more people on chlorpromazine experienced dry mouth. It is clear, however, that olanzapine has important anticholinergic effects, as the incidence of blurred vision and dizziness was similar in both groups. More people taking chlorpromazine experienced orthostatic changes and palpitations than those on olanzapine.

For extrapyramidal events again data were only available from Conley 1998 where fewer people taking olanzapine experienced these symptoms. Until the findings of this trial are replicated, it is inadvisable to draw definitive conclusions regarding the incidence of movement disorders in relation to olanzapine. Much of the scale data were uninformative.

In this small study (Conley 1998), olanzapine was no less sedating than chlorpromazine although the results suggested that, with larger numbers, and more power and precision, the experimental drug might have shown significantly less drowsiness than chlorpromazine. Data on weight gain were only available from the Altamura 1999 (HGBQ) medium‐term study. Although when comparing baseline weights with endpoint weight, those taking olanzapine gained weight whilst those on haloperidol did not, this did not reach statistical significance. Data relating to the question of weight gain remain limited but none suggest that olanzapine causes weight to decrease.

8.5 Missing outcomes 
 No quality of life data or cost effectiveness data are reported. Conclusions cannot be drawn for the outcomes of social functioning, employment status, death and cognitive functioning or family satisfaction.

9. COMPARISON 5. OLANZAPINE versus ATYPICAL ANTIPSYCHOTICS

Twenty seven studies compared olanzapine with atypical antipsychotics. These trials contained a mix of patient populations including first episode (Siesta 2003) and elderly people (Jeste 2003, Ritchie 2003), whereas Tran 1997 (HGBG) excluded people whose illness was 'treatment resistant' and Jones 1998 (P022) studied 'stable, early phase outpatients'. Four studies included only treatment‐resistant participants (Beuzen 1998 (HGCF), Bitter 2004 (HGCK), Kelly 2003 and Naber 2001 (HGBF)) and used clozapine, a drug effective in this group (Wahlbeck 1999). Thomas 1998 (HGBU) did not expand on description further than DSM‐IV criteria, whereas Lecrubier 1999 described the participants as suffering from 'primarily negative symptoms'. Martin 2002 utilised amisulpride as the comparator antisychotic, and randomised both inpatients and outpatients.

9.1 Death 
 Again this important outcome is poorly reported. There were no significant differences in deaths reported in any study. Within well‐conducted trials there are likely to be fewer deaths than amongst people with the same condition who are not part of a controlled study.

9.2 Global effect 
 At this point there are no clear differences between olanzapine and other atypicals for CGI data or clinical response. Data from a single study does suggest that rehospitalisation is lower for those given olanzapine when compared with risperidone.

9.3 Mental state 
 In the short and medium term similar proportions on olanzapine and other atypicals required additional benzodiazepines, however, by two years more people taking olanzapine required additional benzodiazepines than those on clozapine. At two years antidepressants had been used more frequently in the olanzapine treated group.

When mental state percentage improvements were used to classify response there was no difference between olanzapine and other atypicals. In the short term however, medium term data favoured olanzapine.

Endpoint data from ratings scales reveal inconsistencies with the majority of comparisons being equivocal. Olanzapine does not appear to be any different to the other atypical antipsychotics on the outcome measured by mental state rating scales. Negative symptoms were not clearly more amenable to olanzapine than other drugs of the atypical class.

9.4 Leaving the study early 
 This was a prevalent behaviour in both groups for all time periods with 49% lost in the olanzapine group and 56% in the comparator group by 12 months. Nevertheless attrition of such a degree must draw into question the external validity of all other data. Similar proportions left due to lack of efficacy of trials.

9.5 Adverse effects 
 Some anticholinergic effects appear less prevalent in the olanzapine treated group (constipation, hypersalivation) whereas dry mouth is more of a problem for those taking olanzapine.

Fewer people taking olanzapine received anticholinergic medication compared to those taking other atypicals. At eight weeks similar proportions in the two groups had complained of any extrapyramidal event, however, by three to 12 months this was more apparent in those taking other atypicals. New occurrences of parkinsonism were fewer in the olanzapine group. At three to 12 months akathisia was similar in the two groups, by two years it was higher in the olanzapine treated group.

Treatment emergent effects were less common in the olanzapine group.

Data reported at two years by Meltzer (InterSept) revealed lower rates of dyspepsia, asthenia, nausea, vomiting and reduction in white blood cell count in people taking olanzapine compared with those allocated to clozapine.

Sleep problems reveal inconsistencies with both insomnia and somnolence being side effects.

Weight gain appears to be more of a problem with olanzapine than other atypicals, although dichotomous data reached conventional levels of statistical significance only at two years. By eight weeks those taking olanzapine had gained significantly more weight than those taking other atypicals.

9.6 Quality of life 
 No difference is apparent between atypicals and olanzapine on two continuous measures of quality of life. It is good to see an attempt to measure this elusive but important issue. However, simple outcomes such as 'gaining employment' or 'avoiding trouble with the police' or 'feeling that life has improved in quality' may have resulted in more comprehensible information.

9.7 Service outcomes 
 Those people taking olanzapine in the Meltzer (InterSept) study required significantly more admissions to hospital and increased levels of surveillance to maintain their safety compared to the clozapine group.

9.8 Missing data 
 Social function outcomes were absent as were data relating to harm, cognitive functioning and cost effectiveness. Data on satisfaction with treatment were provided by the Drug Awareness Inventory, but has yet to be included while we try to identify validity and meaning.

10. COMPARISON 6. OLANZAPINE versus ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS

As olanzapine's structural similarity to clozapine is cited in marketing, the authors decided to incorporate a comparison with clozapine for people whose illness is treatment resistant.

10.1 Death 
 Again this important outcome was not reported. That no deaths occurred in this trial (Beuzen 1998 (HGCF)) of 180 participants is reassuring but longer‐term real‐world studies are needed before any firm conclusions can be drawn.

10.2 Global effect 
 No differences on global effect as measured by the CGI or Kane 1988 criteria were apparent, although this may be a Type II Error because of the small sample size. It is nevertheless interesting that there are no clear differences between olanzapine and clozapine in a treatment‐resistant group of people for this outcome but more data are needed.

10.3 Leaving the study early 
 At 18 weeks (medium term) similar numbers had left the study early (46%). Olanzapine, in the context of a randomised trial, is just as acceptable or unacceptable as clozapine.

10.4 Adverse events 
 Constipation, nausea and somnolence are less of a problem for those on olanzapine compared with clozapine.

10.5 Missing data 
 Again, no data were reported relating to social functioning, quality of life or cost effectiveness.

11. COMPARISON 7. OLANZAPINE versus ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH FIRST EPISODE

One small study of women with first episode of illness was included (Svestka 2003) and compared olanzapine with risperidone.

11.1 Global effect 
 Equivocal data from this one small study are reported.

11.2 Adverse events 
 Fewer participants taking olanzapine required anticholinergic medication or reported any extrapyramidal adverse events.

Hyperprolactineamia was more common in people taking risperidone though the clinical significance of this is unclear.

Authors' conclusions

Implications for practice.

1. For people with schizophrenia 
 For people with less severe schizophrenia, olanzapine offers antipsychotic efficacy with fewer movement disorders. If the illness is more severe, then there is little to choose between any of the typical or atypical drugs except on the issue of adverse effects such as parkinsonism. Olanzapine is effective in reducing negative symptom scores on a mental state rating scale, but the clinical relevance of this result is unclear. These findings alone may be enough to support people with schizophrenia trying olanzapine before the typical antipsychotic drugs. Olanzapine's propensity to cause weight gain may, however, be unacceptable to many. 
 
 2. For clinicians 
 The very great losses to follow up make recommendations difficult. In these trials most people given olanzapine stopped taking the drug within six months to one year. Global impression suggests that 10‐15 mg/day of olanzapine is antipsychotic, being better than placebo, but, for people with severe illness, when compared with typicals and atypicals there is little difference for the same outcome. On one sub‐scale score, of one mental state rating scale, olanzapine shows superiority over typical antipsychotics for negative symptoms. This result is difficult to interpret clinically. Such findings need replication in large simple studies and should not form the evidence base of treatment recommendations. Olanzapine may have fewer extrapyramidal effects than chlorpromazine and haloperidol and, perhaps, than risperidone. Despite poor data regarding the effects of olanzapine on weight, it does seem to encourage rapid weight gain. However, it is unclear as to whether this weight gain continues while the person is taking olanzapine and if the weight is lost if the drug is stopped. Currently not enough data relating to those with treatment‐resistant schizophrenia are available to draw definitive conclusions.

3. For managers/policy makers 
 Olanzapine is expensive when compared with the typical antipsychotics. For people making cost‐effectiveness and cost‐utility decisions, there is little useful trial‐derived data to help decision making.

Implications for research.

1. General 
 Clear and strict adherence to the CONSORT statement (Moher 2001) for all outcomes would have resulted in this review being more informative. Denominator data were not always clearly presented and some results were described as 'significant' or not, but raw data were not reported (Beasley 1996a (HGAD), Tollefson 1997(HGAJ)).

2. Specific 
 2.1 Trial identifiers 
 The authors have had to inspect a very great number of presentations of the few trials included in this review and there are likely to be more that we have not yet discovered. Using specific trial identifiers, within the multiple publications, would greatly decrease confusion over identification of the source trial.

2.2 More studies 
 Additional well planned, longer, conducted and reported randomised controlled trials are needed that address important, relevant outcomes. Conflict free funding is justified. High‐quality trial‐derived data relating to hospital admission, satisfaction with care, family burden, employment, and important adverse effects such as weight gain are lacking. Dichotomous, clinically useful data relating to mental state and behaviour need replication and expansion. Useful cost‐effectiveness data needs to be a research priority.

What's new

Date	Event	Description
12 December 2012	Amended	Comment added, see Published notes.

History

Protocol first published: Issue 1, 1998
 Review first published: Issue 3, 1998

Date	Event	Description
13 April 2011	Amended	Contact details updated.
5 August 2009	Amended	Contact details updated.
31 October 2008	Amended	Converted to new review format.

Notes

Cochrane Schizophrenia Group internal peer review complete (see Module). 
 External peer review scheduled.

It has been drawn to our attention that there are parts of the data of this review that may need correction (1). All comments relate to the placebo comparison and, so far in our investigations, we have found no substantive issue that changes conclusions. Reviewers interested in a new review covering this specific comparison are welcome to contact the editorial base. Amendments to this review will be made shortly.

1. Hutton P, Morrison AP, Yung AR, Taylor PJ, French P, Dunn G. Effects of drop‐out on efficacy estimates in five Cochrane reviews of popular antipsychotics for schizophrenia. Acta Psychiatr Scand. 2012 Jul;126(1):1–11.

Data and analyses

Comparison 1. OLANZAPINE vs PLACEBO.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Global effect: 1. No important clinical response ‐ by 6 weeks*	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.1 any dose of olanzapine	2	418	Risk Ratio (M‐H, Random, 95% CI)	0.88 [0.80, 0.98]
1.2 15mg olanzapine	1	137	Risk Ratio (M‐H, Random, 95% CI)	0.74 [0.56, 0.97]
1.3 10mg olanzapine	2	232	Risk Ratio (M‐H, Random, 95% CI)	0.82 [0.71, 0.95]
1.4 5mg olanzapine	1	133	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.74, 1.18]
1.5 1mg olanzapine	1	102	Risk Ratio (M‐H, Random, 95% CI)	0.98 [0.87, 1.11]
2 Global effect: 2. CGI endpoint score ‐ by 6 weeks (high score=poor)*	2	230	Mean Difference (IV, Random, 95% CI)	‐0.53 [‐0.89, ‐0.17]
3 Global effect: 3. Needing additional benzodiazepines ‐ by six weeks*	2	418	Risk Ratio (M‐H, Random, 95% CI)	0.96 [0.86, 1.08]
4 Mental state: 1. BPRS total ‐ by 6 weeks (endpoint, high score=poor)*	1	98	Mean Difference (IV, Random, 95% CI)	‐6.86 [‐12.34, ‐1.38]
5 Mental state: 2. BPRS total ‐ by 6 weeks (endpoint, high score=poor, skewed data)*			Other data	No numeric data
6 Mental state: 3. PANSS total ‐ by 6 weeks (endpoint, high score=poor)*	1	98	Mean Difference (IV, Random, 95% CI)	‐12.43 [‐22.54, ‐2.32]
7 Mental state: 4. Negative symptoms ‐ by 6 months (SANS, endpoint, high score=poor)*	1	94	Mean Difference (IV, Random, 95% CI)	‐0.5 [‐2.87, 1.87]
8 Mental state: 5. Negative symptoms ‐ by 6 weeks (PANSS, endpoint, high score=poor)*	1	98	Mean Difference (IV, Random, 95% CI)	‐1.39 [‐4.42, 1.64]
9 Mental state: 6. Negative symptoms ‐ by 6 weeks (skewed data)*			Other data	No numeric data
9.1 as measured by BPRS (endpoint, high=poor)			Other data	No numeric data
9.2 as measured by SANS (endpoint, high=poor)			Other data	No numeric data
10 Mental state: 7. Positive symptoms ‐ by 6 weeks (endpoint, high score=poor, skewed data)*			Other data	No numeric data
10.1 BPRS			Other data	No numeric data
10.2 PANSS			Other data	No numeric data
11 Mental state: 8. Positive symptoms ‐ by 6 weeks (PANSS endpoint, high score=poor)*	1	98	Mean Difference (IV, Random, 95% CI)	‐4.0 [‐7.10, ‐0.90]
12 Leaving the study early: 1. Any reason	4		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
12.1 by 6 weeks	3	598	Risk Ratio (M‐H, Random, 95% CI)	0.86 [0.75, 0.98]
12.2 by 6 months	1	175	Risk Ratio (M‐H, Random, 95% CI)	1.34 [0.87, 2.07]
12.3 not eligible, or eligibile and unwilling to continue on study medication beyond 6 weeks	1	266	Risk Ratio (M‐H, Random, 95% CI)	0.96 [0.83, 1.10]
12.4 by 52 weeks	1	266	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.89, 1.00]
13 Leaving the study early: 2. Due to lack of efficacy	4		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
13.1 by 6 weeks	3	598	Risk Ratio (M‐H, Random, 95% CI)	0.61 [0.43, 0.89]
13.2 by 6 months	1	175	Risk Ratio (M‐H, Random, 95% CI)	0.65 [0.25, 1.70]
13.3 by 52 weeks	1	58	Risk Ratio (M‐H, Random, 95% CI)	0.41 [0.20, 0.87]
14 Adverse events: 1. Anticholinergic effects ‐ specific symptoms by 6 weeks *	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
14.1 dry mouth	1	266	Risk Ratio (M‐H, Random, 95% CI)	1.60 [0.47, 5.41]
14.2 dizziness	1	266	Risk Ratio (M‐H, Random, 95% CI)	3.95 [0.96, 16.31]
15 Adverse events: 2a. Extrapyramidal ‐ needing anticholinergic medication by 6 weeks*	2	418	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.29, 2.79]
16 Adverse events: 2b. Extrapyramidal ‐ specific symptoms by 6 weeks*	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
16.1 akathisia	1	266	Risk Ratio (M‐H, Random, 95% CI)	4.12 [0.55, 31.11]
16.2 tremor	1	266	Risk Ratio (M‐H, Random, 95% CI)	2.40 [0.30, 19.19]
17 Adverse events: 2c. Extrapyramidal ‐ specific symptoms ‐ by 6 weeks (various scales, skewed data)*			Other data	No numeric data
17.1 akathisia ‐ (Barnes Akathisia Scale, low score=good)			Other data	No numeric data
17.2 dyskinesia (AIMS, endpoint, high=poor)			Other data	No numeric data
17.3 parkinsonism (Simpson‐Angus Scale, endpoint, low score=good)			Other data	No numeric data
18 Adverse events: 3. Gastrointestinal ‐ sickness ‐ nausea/vomiting ‐ by 6 weeks*	1	266	Risk Ratio (M‐H, Random, 95% CI)	0.40 [0.14, 1.15]
19 Adverse events: 4. Other problems ‐ by 6 weeks*	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
19.1 agitation	2	418	Risk Ratio (M‐H, Random, 95% CI)	1.20 [0.70, 2.07]
19.2 hostility	2	418	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.57, 1.63]
19.3 withdrawl	1	266	Risk Ratio (M‐H, Random, 95% CI)	0.86 [0.44, 1.69]
20 Adverse events: 5. Sleep problems ‐ by 6 weeks*	2	418	Risk Ratio (M‐H, Random, 95% CI)	0.93 [0.59, 1.45]
21 Adverse events: 6. Weight*	3		Mean Difference (IV, Random, 95% CI)	Subtotals only
21.1 by 6‐8 weeks	2	227	Mean Difference (IV, Random, 95% CI)	3.58 [‐1.18, 8.34]
21.2 by 3 ‐12 months	1	104	Mean Difference (IV, Random, 95% CI)	‐0.52 [‐6.14, 5.10]
22 Quality of Life: 1. QLS total ‐ by six months (endpoint, high score=good)*	1	80	Mean Difference (IV, Random, 95% CI)	8.20 [‐2.47, 18.87]

1.1. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 1 Global effect: 1. No important clinical response ‐ by 6 weeks*.

1.2. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 2 Global effect: 2. CGI endpoint score ‐ by 6 weeks (high score=poor)*.

1.3. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 3 Global effect: 3. Needing additional benzodiazepines ‐ by six weeks*.

1.4. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 4 Mental state: 1. BPRS total ‐ by 6 weeks (endpoint, high score=poor)*.

1.5. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 5 Mental state: 2. BPRS total ‐ by 6 weeks (endpoint, high score=poor, skewed data)*.

Mental state: 2. BPRS total ‐ by 6 weeks (endpoint, high score=poor, skewed data)*
Study	Intervention	Mean	SD	N
Beasley 1996a (HGAD)	Olanzapine	27.43	17.28	65
Beasley 1996a (HGAD)	Placebo	36.00	20.50	62

1.6. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 6 Mental state: 3. PANSS total ‐ by 6 weeks (endpoint, high score=poor)*.

1.7. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 7 Mental state: 4. Negative symptoms ‐ by 6 months (SANS, endpoint, high score=poor)*.

1.8. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 8 Mental state: 5. Negative symptoms ‐ by 6 weeks (PANSS, endpoint, high score=poor)*.

1.9. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 9 Mental state: 6. Negative symptoms ‐ by 6 weeks (skewed data)*.

Mental state: 6. Negative symptoms ‐ by 6 weeks (skewed data)*
Study	Intervention	Mean	SD	N
as measured by BPRS (endpoint, high=poor)
Beasley 1996a (HGAD)	Olanzapine	4.42	3.56	65
Beasley 1996a (HGAD)	Placebo	6.55	4.98	65
Beasley 1996b (HGAP)	Olanzapine	5.71	3.06	49
Beasley 1996b (HGAP)	Placebo	6.37	3.88	49
as measured by SANS (endpoint, high=poor)
Beasley 1996a (HGAD)	Olanzapine	9.32	5.39	65
Beasley 1996a (HGAD)	Placebo	12.40	6.95	65

1.10. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 10 Mental state: 7. Positive symptoms ‐ by 6 weeks (endpoint, high score=poor, skewed data)*.

Mental state: 7. Positive symptoms ‐ by 6 weeks (endpoint, high score=poor, skewed data)*
Study	Intervention	Mean	SD	N
BPRS
Beasley 1996a (HGAD)	Olanzapine	9.12	6.21	65
Beasley 1996a (HGAD)	Placebo	11.29	6.90	63
Beasley 1996b (HGAP)	Olanzapine	9.96	5.17	49
Beasley 1996b (HGAP)	Placebo	12.29	4.96	49
PANSS
Tollefson 1999(HGDY)	Olanzapine	‐0.06	3.38	53
Tollefson 1999(HGDY)	Placebo	0.88	5.52	53

1.11. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 11 Mental state: 8. Positive symptoms ‐ by 6 weeks (PANSS endpoint, high score=poor)*.

1.12. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 12 Leaving the study early: 1. Any reason.

1.13. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 13 Leaving the study early: 2. Due to lack of efficacy.

1.14. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 14 Adverse events: 1. Anticholinergic effects ‐ specific symptoms by 6 weeks *.

1.15. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 15 Adverse events: 2a. Extrapyramidal ‐ needing anticholinergic medication by 6 weeks*.

1.16. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 16 Adverse events: 2b. Extrapyramidal ‐ specific symptoms by 6 weeks*.

1.17. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 17 Adverse events: 2c. Extrapyramidal ‐ specific symptoms ‐ by 6 weeks (various scales, skewed data)*.

Adverse events: 2c. Extrapyramidal ‐ specific symptoms ‐ by 6 weeks (various scales, skewed data)*
Study	Intervention	Mean	SD	N
akathisia ‐ (Barnes Akathisia Scale, low score=good)
Beasley 1996a (HGAD)	Olanzapine	0.42	0.66	65
Beasley 1996a (HGAD)	Placebo	0.47	0.87	64
dyskinesia (AIMS, endpoint, high=poor)
Beasley 1996a (HGAD)	Olanzapine	2.09	3.53	65
Beasley 1996a (HGAD)	Placebo	2.56	4.25	64
parkinsonism (Simpson‐Angus Scale, endpoint, low score=good)
Beasley 1996a (HGAD)	Olanzapine	1.35	2.44	63
Beasley 1996a (HGAD)	Placebo	1.16	2.74	64

1.18. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 18 Adverse events: 3. Gastrointestinal ‐ sickness ‐ nausea/vomiting ‐ by 6 weeks*.

1.19. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 19 Adverse events: 4. Other problems ‐ by 6 weeks*.

1.20. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 20 Adverse events: 5. Sleep problems ‐ by 6 weeks*.

1.21. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 21 Adverse events: 6. Weight*.

1.22. Analysis.

Comparison 1 OLANZAPINE vs PLACEBO, Outcome 22 Quality of Life: 1. QLS total ‐ by six months (endpoint, high score=good)*.

Comparison 2. OLANZAPINE vs PLACEBO (3 to 5 days).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Global effect: 1. CGI severity of illness (very short term 3‐5 days, change, high=poor)	1	106	Mean Difference (IV, Random, 95% CI)	‐0.23 [‐0.52, 0.06]
2 Global effect: 2. Relapse (very short term 3‐5 days)	1	106	Risk Ratio (M‐H, Fixed, 95% CI)	0.13 [0.02, 0.96]
3 Mental state: 1. PANSS total (very short term 3 to 5 days, change, high score=poor)	1	106	Mean Difference (IV, Random, 95% CI)	‐5.23 [‐9.98, ‐0.48]
4 Mental state: 2. Negative symptoms (very short term 3 to 5 days, PANSS, endpoint, high score=poor)	1	106	Mean Difference (IV, Random, 95% CI)	0.01 [‐1.55, 1.57]
5 Mental state: 3. Positive symptoms (very short term 3 to 5 days, PANSS 3‐5 days)	1	106	Mean Difference (IV, Random, 95% CI)	‐0.94 [‐2.68, 0.80]
6 Mental state: 4. Depression (very short term 3 to 5 days, MADRS, change, high score=poor)	1	106	Mean Difference (IV, Random, 95% CI)	‐4.52 [‐6.99, ‐2.05]
7 Mental state: 5. Mental state (very short term 3 to 5 days, MMSE, high score=poor)	1	106	Mean Difference (IV, Random, 95% CI)	0.30 [‐1.05, 1.65]
8 Leaving the study early. Any reason ‐ by 5 days	1	106	Risk Ratio (M‐H, Random, 95% CI)	0.38 [0.11, 1.34]
9 Adverse events: 1. Anticholinergic effects ‐ headache (very short term 3 to 5 days,)*	1	106	Risk Ratio (M‐H, Random, 95% CI)	3.0 [0.32, 27.93]
10 Adverse events: 2. Gastrointestinal (very short term 3 to 5 days,)*	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
10.1 diarrhoea	1	106	Risk Ratio (M‐H, Random, 95% CI)	0.75 [0.18, 3.19]
10.2 nausea	1	106	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.01, 8.00]
10.3 vomiting	1	106	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.04, 3.10]
10.4 nausea and vomiting	1	106	Risk Ratio (M‐H, Random, 95% CI)	3.0 [0.12, 72.02]

2.1. Analysis.

Comparison 2 OLANZAPINE vs PLACEBO (3 to 5 days), Outcome 1 Global effect: 1. CGI severity of illness (very short term 3‐5 days, change, high=poor).

2.2. Analysis.

Comparison 2 OLANZAPINE vs PLACEBO (3 to 5 days), Outcome 2 Global effect: 2. Relapse (very short term 3‐5 days).

2.3. Analysis.

Comparison 2 OLANZAPINE vs PLACEBO (3 to 5 days), Outcome 3 Mental state: 1. PANSS total (very short term 3 to 5 days, change, high score=poor).

2.4. Analysis.

Comparison 2 OLANZAPINE vs PLACEBO (3 to 5 days), Outcome 4 Mental state: 2. Negative symptoms (very short term 3 to 5 days, PANSS, endpoint, high score=poor).

2.5. Analysis.

Comparison 2 OLANZAPINE vs PLACEBO (3 to 5 days), Outcome 5 Mental state: 3. Positive symptoms (very short term 3 to 5 days, PANSS 3‐5 days).

2.6. Analysis.

Comparison 2 OLANZAPINE vs PLACEBO (3 to 5 days), Outcome 6 Mental state: 4. Depression (very short term 3 to 5 days, MADRS, change, high score=poor).

2.7. Analysis.

Comparison 2 OLANZAPINE vs PLACEBO (3 to 5 days), Outcome 7 Mental state: 5. Mental state (very short term 3 to 5 days, MMSE, high score=poor).

2.8. Analysis.

Comparison 2 OLANZAPINE vs PLACEBO (3 to 5 days), Outcome 8 Leaving the study early. Any reason ‐ by 5 days.

2.9. Analysis.

Comparison 2 OLANZAPINE vs PLACEBO (3 to 5 days), Outcome 9 Adverse events: 1. Anticholinergic effects ‐ headache (very short term 3 to 5 days,)*.

2.10. Analysis.

Comparison 2 OLANZAPINE vs PLACEBO (3 to 5 days), Outcome 10 Adverse events: 2. Gastrointestinal (very short term 3 to 5 days,)*.

Comparison 3. OLANZAPINE vs TYPICAL ANTIPSYCHOTICS.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Death: deaths during study or within 30 days of study discontinuation	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.1 all causes	2	2427	Risk Ratio (M‐H, Random, 95% CI)	1.84 [0.48, 7.11]
1.2 other causes	2	2427	Risk Ratio (M‐H, Random, 95% CI)	1.48 [0.15, 14.22]
1.3 suicide and accidental injury	3	2609	Risk Ratio (M‐H, Random, 95% CI)	1.28 [0.17, 9.70]
1.4 unknown	1	1996	Risk Ratio (M‐H, Random, 95% CI)	0.16 [0.01, 4.04]
2 Global effect: 1. No important clinical response ‐ by 6‐8 weeks	4		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
2.1 any dose of olanzapine	4	2778	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.76, 1.06]
2.2 15mg olanzapine	2	308	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.70, 1.08]
2.3 10 mg olanzapine	2	300	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.79, 1.19]
2.4 5mg olanzapine	2	302	Risk Ratio (M‐H, Random, 95% CI)	1.08 [0.89, 1.31]
2.5 1mg olanzapine	1	169	Risk Ratio (M‐H, Random, 95% CI)	1.13 [0.87, 1.48]
3 Global effect: 2. No important clinical response ‐ by 12 weeks	1	263	Risk Ratio (M‐H, Random, 95% CI)	0.82 [0.64, 1.06]
4 Global effect: 3. No important clinical response by 9 months	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
4.1 >/= 40% reduction in PANSS endpoint score	1	197	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.69, 1.05]
4.2 >/=2 CGI	1	197	Risk Ratio (M‐H, Random, 95% CI)	0.82 [0.67, 1.00]
5 Global effect: 4. Relapse/hospitalisation by 52 weeks*	2	495	Risk Ratio (M‐H, Random, 95% CI)	0.68 [0.19, 2.38]
6 Global effect: 5. CGI total (endpoint, high=poor)	14		Mean Difference (IV, Random, 95% CI)	Subtotals only
6.1 by 6 weeks	9	2503	Mean Difference (IV, Random, 95% CI)	‐0.31 [‐0.57, ‐0.06]
6.2 by 3‐12 months	5	411	Mean Difference (IV, Random, 95% CI)	‐0.46 [‐0.83, ‐0.09]
7 Global effect: 6. CGI severity of illness (endpoint, high=poor, skewed data)			Other data	No numeric data
8 Mental state: 1. Needing additional benzodiazepines ‐ by 12 weeks	3	2526	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.85, 0.96]
9 Mental state: 2. BPRS total by 3 months (endpoint, high score=poor)	6	397	Mean Difference (IV, Random, 95% CI)	0.00 [‐3.38, 3.38]
10 Mental state: 3. BPRS total (endpoint, high score=poor, skewed data)			Other data	No numeric data
11 Mental state: 4. PANSS total (endpoint, high score=poor)	15		Mean Difference (IV, Random, 95% CI)	Subtotals only
11.1 by 8 weeks	8	2427	Mean Difference (IV, Random, 95% CI)	‐6.04 [‐11.11, ‐0.96]
11.2 by 3‐12 months	6	563	Mean Difference (IV, Random, 95% CI)	‐8.48 [‐13.84, ‐3.12]
11.3 over 12 months	1	44	Mean Difference (IV, Random, 95% CI)	‐7.32 [‐18.63, 3.99]
12 Mental state: 5. Negative symptoms (PANSS, endpoint, high score=poor)	10		Mean Difference (IV, Random, 95% CI)	Subtotals only
12.1 by 3 months	4	2297	Mean Difference (IV, Random, 95% CI)	‐1.20 [‐2.37, ‐0.02]
12.2 by 3‐12 months	5	408	Mean Difference (IV, Random, 95% CI)	‐2.11 [‐3.34, ‐0.88]
12.3 over 12 months	1	44	Mean Difference (IV, Random, 95% CI)	‐4.75 [‐8.10, ‐1.40]
13 Mental state: 6. Negative symptoms (various scales, endpoint, high score=poor, skewed data)			Other data	No numeric data
13.1 as measured by BPRS			Other data	No numeric data
13.2 as measured by PANSS‐N			Other data	No numeric data
13.3 as measured by SANS			Other data	No numeric data
14 Mental state: 7. Positive symptoms ‐ 8 weeks (BPRS, endpoint, high score=poor)	2	110	Mean Difference (IV, Random, 95% CI)	‐0.97 [‐2.21, 0.28]
15 Mental state: 8. Positive symptoms (PANSS, endpoint, high score=poor)	8		Mean Difference (IV, Random, 95% CI)	Subtotals only
15.1 by 3 months	4	2297	Mean Difference (IV, Random, 95% CI)	‐1.03 [‐1.63, ‐0.43]
15.2 by 3 ‐ 12 months	3	334	Mean Difference (IV, Random, 95% CI)	‐2.20 [‐5.42, 1.02]
15.3 over 12 months	1	44	Mean Difference (IV, Random, 95% CI)	‐0.26 [‐3.42, 2.90]
16 Mental state: 9. Positive symptoms (various scales, endpoint, high score=poor, skewed data)			Other data	No numeric data
16.1 as measured by BPRS ‐ by 6 weeks			Other data	No numeric data
16.2 as measured by PANSS‐P ‐ by 6 to 12 weeks			Other data	No numeric data
17 Mental state: 10. Depression (MADRS, endpoint, high score=poor, skewed data)			Other data	No numeric data
17.1 by 3 months			Other data	No numeric data
17.2 by 3 to 12 months			Other data	No numeric data
18 Mental state: 11. Anxiety ‐ by 14 weeks (HAMA, endpoint, high score=poor, skewed data)			Other data	No numeric data
19 Mental state: 12. Subjective Well‐Being Under Neuroleptic Scale ‐ by 6 weeks (high score=	1	20	Mean Difference (IV, Random, 95% CI)	‐10.58 [‐38.34, 17.18]
20 Leaving the study early: 1a. Any reason	26		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
20.1 by 4‐8 weeks	14	3344	Risk Ratio (M‐H, Random, 95% CI)	0.81 [0.65, 1.02]
20.2 not eligible, or eligibile and unwilling to continue on study medication beyond 6 weeks	3	2694	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.75, 1.08]
20.3 by 3 ‐ 12 months	12	1112	Risk Ratio (M‐H, Random, 95% CI)	0.76 [0.61, 0.94]
20.4 over 12 months	4	2738	Risk Ratio (M‐H, Random, 95% CI)	0.93 [0.86, 1.02]
21 Leaving the study early: 1b. Any reason (Peto Odds ratio)	14	3344	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.54 [0.46, 0.63]
21.1 by 4‐8 weeks	14	3344	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.54 [0.46, 0.63]
22 Leaving the study early: 2. Due to lack of efficacy	15		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
22.1 by 4‐8 weeks	8	2981	Risk Ratio (M‐H, Random, 95% CI)	0.86 [0.59, 1.25]
22.2 by 3 ‐ 12 months	6	482	Risk Ratio (M‐H, Random, 95% CI)	0.57 [0.34, 0.96]
22.3 over 12 months	4	1012	Risk Ratio (M‐H, Random, 95% CI)	0.79 [0.57, 1.10]
23 Adverse events: 1. Anticholinergic effects ‐ specific symptoms ‐ by 6 to 12 weeks*	6		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
23.1 blurred vision	3	2107	Risk Ratio (M‐H, Random, 95% CI)	0.60 [0.31, 1.18]
23.2 cardiovascular ‐ dizziness	2	351	Risk Ratio (M‐H, Random, 95% CI)	1.20 [0.61, 2.37]
23.3 cardiovascular ‐ palpitations	2	2080	Risk Ratio (M‐H, Random, 95% CI)	0.43 [0.11, 1.66]
23.4 cardiovascular ‐ orthostatic changes	1	84	Risk Ratio (M‐H, Random, 95% CI)	0.13 [0.05, 0.35]
23.5 constipation	1	27	Risk Ratio (M‐H, Random, 95% CI)	1.18 [0.37, 3.70]
23.6 salivation ‐ dry mouth	4	2374	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.51, 1.74]
23.7 salivation ‐ hypersalivation	4	2636	Risk Ratio (M‐H, Random, 95% CI)	0.37 [0.20, 0.71]
23.8 urination difficulties	2	2023	Risk Ratio (M‐H, Random, 95% CI)	0.50 [0.23, 1.05]
24 Adverse events: 2a. Extrapyramidal ‐ ESRS ‐ by 6 weeks (endpoint, high score=poor)	4	139	Mean Difference (IV, Random, 95% CI)	‐3.62 [‐8.11, 0.87]
25 Adverse events: 2b. Extrapyramidal ‐ needing additional anticholinergic medication*	5		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
25.1 by 6 to 12 weeks	4	2957	Risk Ratio (M‐H, Random, 95% CI)	0.28 [0.21, 0.38]
25.2 by 3 to 12 months	1	276	Risk Ratio (M‐H, Random, 95% CI)	0.55 [0.35, 0.86]
26 Adverse events: 2c. Extrapyramidal ‐ specific symptoms*	9		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
26.1 abnormal gait ‐ by 12 weeks	1	182	Risk Ratio (M‐H, Random, 95% CI)	0.11 [0.03, 0.47]
26.2 akathisia ‐ by 6 to 12 weeks	7	3193	Risk Ratio (M‐H, Random, 95% CI)	0.28 [0.21, 0.38]
26.3 any extrapyramidal event ‐ by 6 to 8 weeks	3	2511	Risk Ratio (M‐H, Random, 95% CI)	0.41 [0.28, 0.60]
26.4 bradykinesia ‐ by 8 weeks	1	182	Risk Ratio (M‐H, Random, 95% CI)	0.11 [0.03, 0.47]
26.5 choreoathetosis ‐ by 9 months	1	197	Risk Ratio (M‐H, Random, 95% CI)	0.18 [0.01, 3.68]
26.6 dyskinetic movements ‐ by 6 to 8 weeks	3	2511	Risk Ratio (M‐H, Random, 95% CI)	0.20 [0.03, 1.36]
26.7 dystonia, acute ‐ by 6 to 8 weeks	4	2721	Risk Ratio (M‐H, Random, 95% CI)	0.11 [0.03, 0.41]
26.8 dystonia, acute ‐ by 9 months	1	197	Risk Ratio (M‐H, Random, 95% CI)	0.60 [0.17, 2.05]
26.9 hypertonia ‐ by 6 weeks	3	2694	Risk Ratio (M‐H, Random, 95% CI)	0.25 [0.11, 0.55]
26.10 hypokinesia ‐ by 6 to 8 weeks	2	2023	Risk Ratio (M‐H, Random, 95% CI)	0.43 [0.33, 0.55]
26.11 parkinsonism ‐ by 4 to 12 weeks	3	2286	Risk Ratio (M‐H, Random, 95% CI)	0.39 [0.31, 0.50]
26.12 rigidity ‐ by 12 weeks	1	27	Risk Ratio (M‐H, Random, 95% CI)	0.15 [0.04, 0.56]
26.13 tardive dyskinesia ‐ by 9 months	1	197	Risk Ratio (M‐H, Random, 95% CI)	0.15 [0.03, 0.65]
26.14 tremor ‐ by 6 to 12 weeks	4	907	Risk Ratio (M‐H, Random, 95% CI)	0.23 [0.10, 0.50]
27 Adverse events: 2d. Extrapyramidal ‐ specific symptoms (various scales, endpoint, skewed data)*			Other data	No numeric data
27.1 akathisia ‐ by 4 to 12 weeks (Barnes Akathisia Scale)			Other data	No numeric data
27.2 dyskinesia ‐ by 6 weeks (AIMS)			Other data	No numeric data
27.3 parkinsonism (ESRS) ‐ over 12 months			Other data	No numeric data
27.4 parkinsonism ‐ by 4 to 12 weeks (Simpson‐Angus Scale)			Other data	No numeric data
28 Adverse events: 2e. Extrapyramidal ‐ Van Putten ‐ by 6 weeks (endpoint, skewed data)			Other data	No numeric data
29 Adverse events: 2f. Extrapyramidal ‐ DIEPSS scale ‐ by 8 weeks (endpoint, skewed data)			Other data	No numeric data
29.1 DIEPSS total score			Other data	No numeric data
29.2 akathisia			Other data	No numeric data
29.3 dyskinesia			Other data	No numeric data
29.4 dystonia			Other data	No numeric data
29.5 overall severity			Other data	No numeric data
29.6 parkinsonism			Other data	No numeric data
30 Adverse events: 3. Requiring propranolol ‐ by 12 weeks	1	263	Risk Ratio (M‐H, Random, 95% CI)	0.12 [0.04, 0.40]
31 Adverse events: 4. Adverse drug reaction, requiring dose reduction or additional medication ‐ by 8 weeks	1	182	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.74, 1.12]
32 Adverse events: 5. Gastrointestinal ‐ specific symptoms*	4		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
32.1 anorexia ‐ by 12 weeks	1	182	Risk Ratio (M‐H, Random, 95% CI)	0.26 [0.09, 0.74]
32.2 appetite increase ‐ by 6 weeks	1	1996	Risk Ratio (M‐H, Random, 95% CI)	1.65 [1.35, 2.01]
32.3 sickness ‐ nausea ‐ by 6 to 12 weeks	4	2529	Risk Ratio (M‐H, Random, 95% CI)	0.71 [0.38, 1.32]
32.4 sickness ‐ vomiting ‐ by 6 weeks	1	1996	Risk Ratio (M‐H, Random, 95% CI)	0.59 [0.45, 0.78]
33 Adverse events: 6. Other problems*	5		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
33.1 agitation ‐ by 6 weeks	1	267	Risk Ratio (M‐H, Random, 95% CI)	0.80 [0.52, 1.23]
33.2 anxiety ‐ by 12 weeks	1	182	Risk Ratio (M‐H, Random, 95% CI)	0.74 [0.42, 1.30]
33.3 asthenia ‐ by 12 weeks	1	27	Risk Ratio (M‐H, Random, 95% CI)	0.78 [0.38, 1.60]
33.4 excitement ‐ by 12 weeks	1	182	Risk Ratio (M‐H, Random, 95% CI)	0.96 [0.51, 1.80]
33.5 hostility ‐ by 6 weeks	1	267	Risk Ratio (M‐H, Random, 95% CI)	1.63 [0.70, 3.76]
33.6 incapacitation ‐ by 9 months	1	197	Risk Ratio (M‐H, Random, 95% CI)	0.26 [0.10, 0.68]
33.7 malaise ‐ by 12 weeks	1	182	Risk Ratio (M‐H, Random, 95% CI)	0.96 [0.42, 2.19]
33.8 rash ‐ by 6 weeks	1	30	Risk Ratio (M‐H, Random, 95% CI)	0.5 [0.03, 7.19]
33.9 withdrawal ‐ by 12 weeks	1	267	Risk Ratio (M‐H, Random, 95% CI)	0.58 [0.33, 1.04]
34 Adverse events: 7. Sleep problems*	7		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
34.1 difficulty in getting to sleep by 6 to 12 weeks	5	2960	Risk Ratio (M‐H, Random, 95% CI)	0.88 [0.55, 1.40]
34.2 drowsiness ‐ by 6 to 12 weeks	6	2636	Risk Ratio (M‐H, Random, 95% CI)	0.82 [0.73, 0.93]
35 Adverse events: 8. Sleep problems by 6 weeks (LSEQ)	1	90	Mean Difference (IV, Random, 95% CI)	7.12 [‐3.21, 17.45]
35.1 Awaking from sleep	1	30	Mean Difference (IV, Random, 95% CI)	5.30 [‐11.31, 21.91]
35.2 Getting to sleep score	1	30	Mean Difference (IV, Random, 95% CI)	0.60 [‐18.68, 19.88]
35.3 Quality of sleep	1	30	Mean Difference (IV, Random, 95% CI)	15.0 [‐3.07, 33.07]
36 Adverse events: 9. Sleep problems, Behaviour following waking ‐ by 6 weeks (skewed data)			Other data	No numeric data
37 Adverse events: 10. Weight average increase*	13		Mean Difference (IV, Random, 95% CI)	Subtotals only
37.1 weight increase by 3 months	8	2455	Mean Difference (IV, Random, 95% CI)	0.81 [‐0.56, 2.18]
37.2 weight increase by 3‐12 months	4	186	Mean Difference (IV, Random, 95% CI)	4.62 [0.60, 8.64]
37.3 weight increase over 12 months	1	47	Mean Difference (IV, Random, 95% CI)	0.96 [‐10.53, 12.45]
38 Adverse events: 11. Weight: average increase ‐ by 12 weeks (change data)*	2	283	Mean Difference (IV, Random, 95% CI)	3.53 [1.28, 5.79]
39 Adverse events: 12. Weight: gain	5		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
39.1 by 6 to 12 weeks	4	552	Risk Ratio (M‐H, Random, 95% CI)	1.64 [0.71, 3.77]
39.2 by 24 weeks	1	276	Risk Ratio (M‐H, Random, 95% CI)	1.38 [1.01, 1.88]
40 Adverse events: 13. Weight decrease ‐ by 12 weeks	1	182	Risk Ratio (M‐H, Random, 95% CI)	0.14 [0.02, 1.09]
41 Adverse events: 14. Plasma prolactin levels, endpoint 6 weeks	1	29	Mean Difference (IV, Random, 95% CI)	‐15.01 [‐17.94, ‐12.08]
42 Quality of life: endpoint ‐ by 6 to 12 weeks (high score=good, skewed data)			Other data	No numeric data
43 Drug compliance: DAI‐30, mean change ‐ by 22 weeks (skewed data)			Other data	No numeric data

3.1. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 1 Death: deaths during study or within 30 days of study discontinuation.

3.2. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 2 Global effect: 1. No important clinical response ‐ by 6‐8 weeks.

3.3. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 3 Global effect: 2. No important clinical response ‐ by 12 weeks.

3.4. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 4 Global effect: 3. No important clinical response by 9 months.

3.5. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 5 Global effect: 4. Relapse/hospitalisation by 52 weeks*.

3.6. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 6 Global effect: 5. CGI total (endpoint, high=poor).

3.7. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 7 Global effect: 6. CGI severity of illness (endpoint, high=poor, skewed data).

Global effect: 6. CGI severity of illness (endpoint, high=poor, skewed data)
Study	Intervention	Mean	SD	N
Avasthi 2001	Olanzapine	3.19	0.98	16
Avasthi 2001	Haloperidol	2.86	1.57	7

3.8. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 8 Mental state: 1. Needing additional benzodiazepines ‐ by 12 weeks.

3.9. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 9 Mental state: 2. BPRS total by 3 months (endpoint, high score=poor).

3.10. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 10 Mental state: 3. BPRS total (endpoint, high score=poor, skewed data).

Mental state: 3. BPRS total (endpoint, high score=poor, skewed data)
Study	Intervention	Mean	SD	N
Avasthi 2001	Olanzapine	9.50	7.06	16
Avasthi 2001	Haloperidol	12.57	13.39	7
Beasley 1996a (HGAD)	Olanzapine	27.43	17.28	65
Beasley 1996a (HGAD)	Haloperidol	28.99	15.86	68
Beasley 1997 (E003)	Olanzapine	25.95	16.20	85
Beasley 1997 (E003)	Haloperidol	28.78	15.41	79
HGCJ (Hong Kong)	Olanzapine	12.75	8.81	16
HGCJ (Hong Kong)	Haloperidol	26.86	13.71	14
HGCQ (Turkey) 2000	Olanzapine	20.2	17.9	20
HGCQ (Turkey) 2000	Chlorpromazine	24.7	18.4	10
HGCU (Taiwan) 1998	Olanzapine	15.83	11.33	24
HGCU (Taiwan) 1998	Haloperidol	20.32	12.21	28
HGDV (Morocco) 1999	Olanzapine	6.7	6.4.	27
HGDV (Morocco) 1999	Chlorpromazine	20.7	12.9	12
Jakovljevic1999 HGCH	Olanzapine	17.9	14.1	27
Jakovljevic1999 HGCH	Fluphenazine group	25.6	19.2	27
Tollefson 1997(HGAJ)	Olanzapine	22.16	13.36	1312
Tollefson 1997(HGAJ)	Haloperidol	26.16	13.34	636

3.11. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 11 Mental state: 4. PANSS total (endpoint, high score=poor).

3.12. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 12 Mental state: 5. Negative symptoms (PANSS, endpoint, high score=poor).

3.13. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 13 Mental state: 6. Negative symptoms (various scales, endpoint, high score=poor, skewed data).

Mental state: 6. Negative symptoms (various scales, endpoint, high score=poor, skewed data)
Study	Intervention	Mean	SD	N
as measured by BPRS
Beasley 1996a (HGAD)	Olanzapine	4.42	3.56	65
Beasley 1996a (HGAD)	Haloperidol	4.87	3.44	68
Beasley 1997 (E003)	Olanzapine	5.26	3.27	85
Beasley 1997 (E003)	Haloperidol	6.23	2.97	79
Esel 2001	Olanzapine	6.46	1.59	15
Esel 2001	Haloperidol	8.2	5.6	14
Jakovljevic1999 HGCH	Olanzapine	3.8	2.7	27
Jakovljevic1999 HGCH	Fluphenazine	5.1	4.2	27
Tollefson 1997(HGAJ)	Olanzapine	4.62	3.0	1312
Tollefson 1997(HGAJ)	Haloperidol	5.55	3.22	636
as measured by PANSS‐N
Avasthi 2001	Olanzapine	15.62	7.93	16
Avasthi 2001	Haloperidol	16.86	8.71	7
as measured by SANS
Avasthi 2001	Olanzapine	21.87	19.47	16
Avasthi 2001	Haloperidol	27.43	19.48	7
Beasley 1996a (HGAD)	Olanzapine	9.32	5.39	65
Beasley 1996a (HGAD)	Haloperidol	11.22	4.80	68
Conley 1998	Olanzapine	13.1	8.2	42
Conley 1998	Chlorpromazine	14.0	9.5	39

3.14. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 14 Mental state: 7. Positive symptoms ‐ 8 weeks (BPRS, endpoint, high score=poor).

3.15. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 15 Mental state: 8. Positive symptoms (PANSS, endpoint, high score=poor).

3.16. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 16 Mental state: 9. Positive symptoms (various scales, endpoint, high score=poor, skewed data).

Mental state: 9. Positive symptoms (various scales, endpoint, high score=poor, skewed data)
Study	Interventions	Mean	SD	N
as measured by BPRS ‐ by 6 weeks
Beasley 1996a (HGAD)	Olanzapine	9.12	6.21	65
Beasley 1996a (HGAD)	Haloperidol	8.59	5.32	68
Beasley 1997 (E003)	Olanzapine	7.87	4.89	85
Beasley 1997 (E003)	Haloperidol	7.91	4.74	79
Jakovljevic1999 HGCH	Olanzapine	5.6	5.8	27
Jakovljevic1999 HGCH	Fluphenazine	8.8	7.1	28
Tollefson 1997(HGAJ)	Olanzapine	6.83	4.66	1312
Tollefson 1997(HGAJ)	Haloperidol	7.61	4.54	636
as measured by PANSS‐P ‐ by 6 to 12 weeks
Avasthi 2001	Olanzapine	11.44	4.11	16
Avasthi 2001	Haloperidol	10.86	8.49	7
Barak 2002	Olanzapine	15.4	7.8	10
Barak 2002	Haloperidol	17.3	6.10	10
He 2003	Olanzapine	10.2	4.9	40
He 2003	Chlorpromazine	11.7	6.2	40
Jakovljevic1999 HGCH	Olanzapine	15.0	9.4	27
Jakovljevic1999 HGCH	Fluphenazine	19.6	10.9	28

3.17. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 17 Mental state: 10. Depression (MADRS, endpoint, high score=poor, skewed data).

Mental state: 10. Depression (MADRS, endpoint, high score=poor, skewed data)
Study	Intervention	Mean	SD	N
by 3 months
Avasthi 2001	Olanzapine	3.00	2.42	16
Avasthi 2001	Haloperidol	5.00	4.58	7
de Hann 2003	Olanzapine	10.56	9.81	11
de Hann 2003	Haloperidol	7.91	5.39	9
HGDV (Morocco) 1999	Olanzapine	3.3	2.0	27
HGDV (Morocco) 1999	Chlorpromazine	6.9	4.8	12
Lieberman 2003 (HGDH	Olanzapine	6.95	7.01	125
Lieberman 2003 (HGDH	Haloperidol	8.38	8.21	126
Tollefson 1997(HGAJ)	Olanzapine	10.65	8.3	1053
Tollefson 1997(HGAJ)	Haloperidol	13.64	9.82	428
by 3 to 12 months
HGCJ (Hong Kong)	Olanzapine	4.06	4.39	16
HGCJ (Hong Kong)	Haloperidol	12.29	10.56	14
HGCU (Taiwan) 1998	Olanzapine	3.67	5.43	24
HGCU (Taiwan) 1998	Haloperidol	4.79	5.14	28

3.18. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 18 Mental state: 11. Anxiety ‐ by 14 weeks (HAMA, endpoint, high score=poor, skewed data).

Mental state: 11. Anxiety ‐ by 14 weeks (HAMA, endpoint, high score=poor, skewed data)
Study	Intervention	Mean	SD	N
HGDV (Morocco) 1999	Olanzapine	3.3	1.8	27
HGDV (Morocco) 1999	Chlorpromazine	7.6	5.8	12

3.19. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 19 Mental state: 12. Subjective Well‐Being Under Neuroleptic Scale ‐ by 6 weeks (high score=.

3.20. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 20 Leaving the study early: 1a. Any reason.

3.21. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 21 Leaving the study early: 1b. Any reason (Peto Odds ratio).

3.22. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 22 Leaving the study early: 2. Due to lack of efficacy.

3.23. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 23 Adverse events: 1. Anticholinergic effects ‐ specific symptoms ‐ by 6 to 12 weeks*.

3.24. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 24 Adverse events: 2a. Extrapyramidal ‐ ESRS ‐ by 6 weeks (endpoint, high score=poor).

3.25. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 25 Adverse events: 2b. Extrapyramidal ‐ needing additional anticholinergic medication*.

3.26. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 26 Adverse events: 2c. Extrapyramidal ‐ specific symptoms*.

3.27. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 27 Adverse events: 2d. Extrapyramidal ‐ specific symptoms (various scales, endpoint, skewed data)*.

Adverse events: 2d. Extrapyramidal ‐ specific symptoms (various scales, endpoint, skewed data)*
Study	Intervention	Mean	SD	N
akathisia ‐ by 4 to 12 weeks (Barnes Akathisia Scale)
Avasthi 2001	Olanzapine	0.0	0.0	16
Avasthi 2001	Haloperidol	0.29	0.49	7
Beasley 1996a (HGAD)	Olanzapine	0.42	0.66	65
Beasley 1996a (HGAD)	Haloperidol	0.87	1.12	68
Beasley 1997 (E003)	Olanzapine	0.38	0.79	85
Beasley 1997 (E003)	Haloperidol	0.91	1.24	79
Bernardo 2001 (HGDD)	Olanzapine	0.4	1.1	14
Bernardo 2001 (HGDD)	Haloperidol	0.8	1.0	13
Conley 1998	Olanzapine	0.60	1.40	42
Conley 1998	Chlorpromazine	0.60	1.70	39
dyskinesia ‐ by 6 weeks (AIMS)
Beasley 1996a (HGAD)	Olanzapine	2.09	3.53	65
Beasley 1996a (HGAD)	Haloperidol	2.49	3.93	68
Beasley 1997 (E003)	Olanzapine	1.38	2.80	85
Beasley 1997 (E003)	Haloperidol	1.90	3.64	78
parkinsonism (ESRS) ‐ over 12 months
Jones 1998 (P022)	Olanzapine	1.19	1.86	21
Jones 1998 (P022)	Haloperidol	2.57	2.61	23
parkinsonism ‐ by 4 to 12 weeks (Simpson‐Angus Scale)
Avasthi 2001	Olanzapine	0.75	1.39	16
Avasthi 2001	Haloperidol	0.86	1.86	7
Beasley 1996a (HGAD)	Haloperidol	3.34	4.39	67
Beasley 1996a (HGAD)	Olanzapine	1.35	2.44	63
Beasley 1997 (E003)	Olanzapine	1.8	3.10	85
Beasley 1997 (E003)	Haloperidol	5.59	7.93	79
Bernardo 2001 (HGDD)	Olanzapine	0.2	1.7	14
Bernardo 2001 (HGDD)	Haloperidol	4.5	3.4	13
Conley 1998	Olanzapine	1.50	2.40	41
Conley 1998	Chlorpromazine	2.10	2.30	39

3.28. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 28 Adverse events: 2e. Extrapyramidal ‐ Van Putten ‐ by 6 weeks (endpoint, skewed data).

Adverse events: 2e. Extrapyramidal ‐ Van Putten ‐ by 6 weeks (endpoint, skewed data)
Study	Intervention	mean	SD	N
HGCQ (Turkey) 2000	Olanzapine	6.7	7.2	20
HGCQ (Turkey) 2000	Chlorpromazine	9.6	7.2	10

3.29. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 29 Adverse events: 2f. Extrapyramidal ‐ DIEPSS scale ‐ by 8 weeks (endpoint, skewed data).

Adverse events: 2f. Extrapyramidal ‐ DIEPSS scale ‐ by 8 weeks (endpoint, skewed data)
Study	Intervention	Mean	SD	N
DIEPSS total score
Ishigooka 2001	Olanzapine	2.0	3.0	81
Ishigooka 2001	Haloperidol	3.30	3.90	80
akathisia
Ishigooka 2001	Olanzapine	0.20	0.50	81
Ishigooka 2001	Haloperidol	0.50	0.90	80
dyskinesia
Ishigooka 2001	Olanzapine	0.10	0.40	81
Ishigooka 2001	Haloperidol	0.20	0.50	80
dystonia
Ishigooka 2001	Olanzapine	0.10	0.30	81
Ishigooka 2001	Haloperidol	0.10	0.40	80
overall severity
Ishigooka 2001	Olanzapine	0.70	0.70	81
Ishigooka 2001	Haloperidol	1.00	0.90	80
parkinsonism
Ishigooka 2001	Olanzapine	1.70	2.50	81
Ishigooka 2001	Haloperidol	2.60	3.30	80

3.30. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 30 Adverse events: 3. Requiring propranolol ‐ by 12 weeks.

3.31. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 31 Adverse events: 4. Adverse drug reaction, requiring dose reduction or additional medication ‐ by 8 weeks.

3.32. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 32 Adverse events: 5. Gastrointestinal ‐ specific symptoms*.

3.33. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 33 Adverse events: 6. Other problems*.

3.34. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 34 Adverse events: 7. Sleep problems*.

3.35. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 35 Adverse events: 8. Sleep problems by 6 weeks (LSEQ).

3.36. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 36 Adverse events: 9. Sleep problems, Behaviour following waking ‐ by 6 weeks (skewed data).

Adverse events: 9. Sleep problems, Behaviour following waking ‐ by 6 weeks (skewed data)
Study	Intervention	mean	SD	N
HGCQ (Turkey) 2000	Olanzapine	55.9	18.8	20
HGCQ (Turkey) 2000	Chlorpromazine	55.1	28.3	10

3.37. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 37 Adverse events: 10. Weight average increase*.

3.38. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 38 Adverse events: 11. Weight: average increase ‐ by 12 weeks (change data)*.

3.39. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 39 Adverse events: 12. Weight: gain.

3.40. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 40 Adverse events: 13. Weight decrease ‐ by 12 weeks.

3.41. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 41 Adverse events: 14. Plasma prolactin levels, endpoint 6 weeks.

3.42. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 42 Quality of life: endpoint ‐ by 6 to 12 weeks (high score=good, skewed data).

Quality of life: endpoint ‐ by 6 to 12 weeks (high score=good, skewed data)
Study	Intervention	mean	SD	N
Avasthi 2001	Olanzapine	51.19	23.38	16
Avasthi 2001	Haloperidol	49.14	33.88	7
Loza 1999 (HGDT)	Olanzapine	56.6	31.3	27
Loza 1999 (HGDT)	Chlorpromazine	47.7	26.6	14

3.43. Analysis.

Comparison 3 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS, Outcome 43 Drug compliance: DAI‐30, mean change ‐ by 22 weeks (skewed data).

Drug compliance: DAI‐30, mean change ‐ by 22 weeks (skewed data)
Study	Intervention	mean change	SD	N
Jakovljevic1999 HGCH	Olanzapine	1.0	1.9	26
Jakovljevic1999 HGCH	Fluphenazine	‐0.2	2.6	26

Comparison 4. OLANZAPINE vs TYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Global effect: 1. No important clinical response ‐ by 8 weeks	1	84	Risk Ratio (M‐H, Random, 95% CI)	0.93 [0.85, 1.02]
2 Global effect: 2. CGI total ‐ by 8 weeks (endpoint, high score=poor)	1	81	Mean Difference (IV, Random, 95% CI)	‐0.10 [‐0.49, 0.29]
3 Mental state: 1. BPRS total ‐ by 8 weeks (endpoint score, high score=poor)	1	81	Mean Difference (IV, Random, 95% CI)	‐2.80 [‐8.43, 2.83]
4 Mental state: 2. Negative symptoms ‐ 8 weeks (SANS, endpoint, high score=poor, skewed data)			Other data	No numeric data
5 Mental state: 3. Positive symptoms by 8 weeks (BPRS, endpoint, high score=poor)	1	81	Mean Difference (IV, Random, 95% CI)	‐1.30 [‐2.90, 0.30]
6 Leaving the study early: 1. Any reason	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
6.1 by 8 weeks	1	84	Risk Ratio (M‐H, Random, 95% CI)	0.92 [0.48, 1.78]
6.2 by 3‐12 months	1	28	Risk Ratio (M‐H, Random, 95% CI)	0.08 [0.00, 1.22]
7 Leaving the study early: 2. Due to lack of efficacy	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
7.1 by 8 weeks	1	84	Risk Ratio (M‐H, Random, 95% CI)	2.5 [0.51, 12.17]
7.2 by 3‐12 months	1	28	Risk Ratio (M‐H, Random, 95% CI)	0.16 [0.01, 2.89]
8 Adverse events: 1. Anticholinergic effects ‐ specific symptoms ‐ by 8 weeks	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
8.1 blurred vision	1	84	Risk Ratio (M‐H, Random, 95% CI)	0.8 [0.23, 2.77]
8.2 cardiovascular ‐ dizziness	1	84	Risk Ratio (M‐H, Random, 95% CI)	0.86 [0.31, 2.34]
8.3 cardiovascular ‐ orthostatic changes	1	84	Risk Ratio (M‐H, Random, 95% CI)	0.13 [0.05, 0.35]
8.4 cardiovascular ‐ palpitations	1	84	Risk Ratio (M‐H, Random, 95% CI)	0.14 [0.02, 1.11]
8.5 salivation ‐ dry mouth	1	84	Risk Ratio (M‐H, Random, 95% CI)	0.52 [0.34, 0.79]
9 Adverse effects: 2a. Extrapyramidal ‐ specific symptoms ‐ by 8 weeks	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
9.1 any extrapyramidal event	1	84	Risk Ratio (M‐H, Random, 95% CI)	0.57 [0.32, 1.01]
9.2 dyskinetic movements	1	84	Risk Ratio (M‐H, Random, 95% CI)	0.07 [0.01, 0.48]
10 Adverse effects: 2b. Extrapyramidal ‐ specific symptoms ‐ by 8 weeks (various scales, endpoint, high=poor, ske			Other data	No numeric data
10.1 akathisia (Barnes Scale)			Other data	No numeric data
10.2 parkinsonism (Simpson‐Angus Scale)			Other data	No numeric data
11 Adverse effects: 3. Gastrointestinal ‐ nausea / vomiting ‐ by 8 weeks	1	84	Risk Ratio (M‐H, Random, 95% CI)	0.63 [0.22, 1.75]
12 Adverse events: 4. Sleep problems ‐ by 8 weeks	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
12.1 difficulty getting to sleep	1	84	Risk Ratio (M‐H, Random, 95% CI)	3.0 [0.64, 14.02]
12.2 drowsiness	1	84	Risk Ratio (M‐H, Random, 95% CI)	0.68 [0.41, 1.12]

4.1. Analysis.

Comparison 4 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 1 Global effect: 1. No important clinical response ‐ by 8 weeks.

4.2. Analysis.

Comparison 4 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 2 Global effect: 2. CGI total ‐ by 8 weeks (endpoint, high score=poor).

4.3. Analysis.

Comparison 4 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 3 Mental state: 1. BPRS total ‐ by 8 weeks (endpoint score, high score=poor).

4.4. Analysis.

Comparison 4 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 4 Mental state: 2. Negative symptoms ‐ 8 weeks (SANS, endpoint, high score=poor, skewed data).

Mental state: 2. Negative symptoms ‐ 8 weeks (SANS, endpoint, high score=poor, skewed data)
Study	Intervention	Mean	SD	N
Conley 1998	Olanzapine	13.1	8.2	42
Conley 1998	Chlorpromazine	14.0	9.5	39

4.5. Analysis.

Comparison 4 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 5 Mental state: 3. Positive symptoms by 8 weeks (BPRS, endpoint, high score=poor).

4.6. Analysis.

Comparison 4 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 6 Leaving the study early: 1. Any reason.

4.7. Analysis.

Comparison 4 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 7 Leaving the study early: 2. Due to lack of efficacy.

4.8. Analysis.

Comparison 4 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 8 Adverse events: 1. Anticholinergic effects ‐ specific symptoms ‐ by 8 weeks.

4.9. Analysis.

Comparison 4 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 9 Adverse effects: 2a. Extrapyramidal ‐ specific symptoms ‐ by 8 weeks.

4.10. Analysis.

Comparison 4 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 10 Adverse effects: 2b. Extrapyramidal ‐ specific symptoms ‐ by 8 weeks (various scales, endpoint, high=poor, ske.

Adverse effects: 2b. Extrapyramidal ‐ specific symptoms ‐ by 8 weeks (various scales, endpoint, high=poor, ske
Study	Intervention	Mean	SD	N
akathisia (Barnes Scale)
Conley 1998	Olanzapine	0.60	1.40	42
Conley 1998	Chlorpromazine	0.60	1.70	39
parkinsonism (Simpson‐Angus Scale)
Conley 1998	Olanzapine	1.50	2.40	41
Conley 1998	Chlorpromazine	2.10	2.30	39

4.11. Analysis.

Comparison 4 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 11 Adverse effects: 3. Gastrointestinal ‐ nausea / vomiting ‐ by 8 weeks.

4.12. Analysis.

Comparison 4 OLANZAPINE vs TYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 12 Adverse events: 4. Sleep problems ‐ by 8 weeks.

Comparison 5. OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Death	5		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.1 any cause	5	1778	Risk Ratio (M‐H, Random, 95% CI)	0.70 [0.31, 1.60]
1.2 suicide	1	980	Risk Ratio (M‐H, Random, 95% CI)	1.67 [0.40, 6.94]
2 Global effect: 1. No clinically important response	7		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
2.1 as measured by CGI by 8 weeks	5	1258	Risk Ratio (M‐H, Random, 95% CI)	0.93 [0.85, 1.03]
2.2 as measured by CGI by 3 to 12 months	2	330	Risk Ratio (M‐H, Random, 95% CI)	0.89 [0.73, 1.08]
2.3 as measured by Kane criteria by 3 to 12 months	2	330	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.80, 1.17]
3 Global effect: 2. Relapse/hospitalisation by 52 weeks	1	278	Risk Ratio (M‐H, Random, 95% CI)	0.49 [0.31, 0.79]
4 Global effect: 3. CGI total by 3 ‐12 months (endpoint, high score=poor)	3	570	Mean Difference (IV, Random, 95% CI)	‐0.00 [‐0.20, 0.19]
5 Mental state: 1a. Needing addtional benzodiazepines	5		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
5.1 by 6 to 12 weeks	3	706	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.68, 1.19]
5.2 by 18 weeks	1	180	Risk Ratio (M‐H, Random, 95% CI)	1.09 [0.88, 1.36]
5.3 by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	1.11 [1.01, 1.22]
6 Mental state: 1b. Needing antidepressants by 2 years	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
6.1 Any antidepressant	1	980	Risk Ratio (M‐H, Random, 95% CI)	1.12 [0.99, 1.27]
6.2 as a rescue intervention to prevent suicide	1	980	Risk Ratio (M‐H, Random, 95% CI)	2.20 [1.21, 4.00]
7 Mental state: 2. No important clinical response (decrease in PANSS total score)	8		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
7.1 short term ‐ by 6 to 8 weeks (20 to 25% decrease in PANSS total score)	4	738	Risk Ratio (M‐H, Random, 95% CI)	1.00 [0.87, 1.15]
7.2 short term ‐ by 6 weeks (50% decrease in PANSS total score)	1	42	Risk Ratio (M‐H, Random, 95% CI)	0.83 [0.30, 2.31]
7.3 medium term ‐ by 18‐28 weeks (40% decrease in PANSS total score)	3	669	Risk Ratio (M‐H, Random, 95% CI)	0.88 [0.80, 0.97]
8 Mental state: 3. No important clinical response (decrease in BPRS total score)	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
8.1 >20% improvement ‐ 6 to 8 weeks	2	332	Risk Ratio (M‐H, Random, 95% CI)	1.08 [0.86, 1.35]
8.2 >40% improvement ‐ by 8 weeks	1	269	Risk Ratio (M‐H, Random, 95% CI)	1.26 [0.80, 1.99]
9 Mental state: 4. BPRS total at 8 weeks (endpoint, high score=worse)	1	61	Mean Difference (IV, Random, 95% CI)	‐1.60 [‐2.96, ‐0.24]
10 Mental state: 5. BPRS total at 3‐12 months (endpoint, high score=poor, skewed data)			Other data	No numeric data
11 Mental state: 6. PANSS total (endpoint, high score=poor)	6		Mean Difference (IV, Random, 95% CI)	Subtotals only
11.1 by 3‐12 months	5	715	Mean Difference (IV, Random, 95% CI)	3.01 [‐5.74, 11.76]
11.2 over 12 months	1	42	Mean Difference (IV, Random, 95% CI)	1.86 [‐9.08, 12.80]
12 Mental state: 7. Negative symptoms (PANSS, endpoint, high score=poor)	6		Mean Difference (IV, Random, 95% CI)	Subtotals only
12.1 by 8 weeks	1	278	Mean Difference (IV, Random, 95% CI)	0.20 [‐1.11, 1.51]
12.2 by 3‐12 months	4	593	Mean Difference (IV, Random, 95% CI)	1.52 [‐2.08, 5.13]
12.3 over 12 months	1	42	Mean Difference (IV, Random, 95% CI)	‐0.20 [‐3.76, 3.36]
13 Mental state: 8. Negative symptoms (various scales, endpoint, high score=poor, skewed data)			Other data	No numeric data
13.1 as measured by BPRS ‐ by 18 weeks			Other data	No numeric data
13.2 as measured by SANS ‐ by 28 weeks			Other data	No numeric data
13.3 PANSS ‐ by 6 to 8 weeks			Other data	No numeric data
14 Mental state: 9. Negative symptoms change data at 6 months (SANS, high score=poor)	1	131	Mean Difference (IV, Random, 95% CI)	0.0 [‐1.78, 1.78]
15 Mental state: 10. Positive symptoms (PANSS, endpoint, high score=poor)	4		Mean Difference (IV, Random, 95% CI)	Subtotals only
15.1 by 8 weeks	1	278	Mean Difference (IV, Random, 95% CI)	0.30 [‐1.09, 1.69]
15.2 3‐12 months	2	86	Mean Difference (IV, Random, 95% CI)	3.47 [‐3.87, 10.81]
15.3 over 12 months	1	42	Mean Difference (IV, Random, 95% CI)	0.62 [‐2.43, 3.67]
16 Mental state: 11. Positive symptoms (various scales, endpoint, high score=poor, skewed data)			Other data	No numeric data
16.1 as measured by BPRS ‐ by 18 weeks			Other data	No numeric data
16.2 as measured by PANSS ‐ by 6 to 8 weeks			Other data	No numeric data
16.3 as measured by PANSS at 28 weeks			Other data	No numeric data
17 Mental state: 12. Depression ‐ by 8 weeks (Hamilton or MADRS rating scale, high score=poor, skewed data)			Other data	No numeric data
18 Mental state: 13. Depression ‐ by 18 weeks (MADRS, high score=poor, skewed data)			Other data	No numeric data
19 Leaving the study early: 1. Any reason	23		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
19.1 by 8 weeks	12	2304	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.74, 1.03]
19.2 by 3‐12 months	11	1847	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.82, 1.00]
19.3 over 12 months	2	1004	Risk Ratio (M‐H, Random, 95% CI)	0.98 [0.84, 1.14]
20 Leaving the study early: 2. Due to lack of efficacy	11		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
20.1 by 8 weeks	6	1643	Risk Ratio (M‐H, Random, 95% CI)	0.93 [0.51, 1.69]
20.2 by 3‐12 months	4	556	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.52, 1.37]
20.3 over 12 months	1	42	Risk Ratio (M‐H, Random, 95% CI)	0.25 [0.03, 2.05]
21 Adverse events: 1. Anticholinergic effects ‐ specific symptoms	9		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
21.1 constipation by 6 to 8 weeks	2	247	Risk Ratio (M‐H, Random, 95% CI)	0.47 [0.25, 0.89]
21.2 constipation ‐ by 18 weeks	1	180	Risk Ratio (M‐H, Random, 95% CI)	0.35 [0.15, 0.85]
21.3 constipation ‐ by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	0.45 [0.34, 0.60]
21.4 dizziness by 6 to 8 weeks	3	624	Risk Ratio (M‐H, Random, 95% CI)	0.52 [0.20, 1.35]
21.5 dizziness ‐ by 18 weeks	2	330	Risk Ratio (M‐H, Random, 95% CI)	0.39 [0.14, 1.11]
21.6 diizziness ‐ by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	0.51 [0.40, 0.66]
21.7 dry mouth by 6 to 8 weeks	3	562	Risk Ratio (M‐H, Random, 95% CI)	1.35 [0.76, 2.41]
21.8 dry mouth by 18 weeks	1	180	Risk Ratio (M‐H, Random, 95% CI)	1.79 [0.99, 3.21]
21.9 dry mouth ‐ by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	1.51 [1.02, 2.25]
21.10 did not report improvement in libido ‐ by 4 weeks	1	66	Risk Ratio (M‐H, Random, 95% CI)	0.78 [0.61, 0.99]
21.11 rhinitis ‐ by 6 to 8 weeks	3	544	Risk Ratio (M‐H, Random, 95% CI)	0.39 [0.09, 1.61]
21.12 rhinitis ‐ by 18 weeks	1	180	Risk Ratio (M‐H, Random, 95% CI)	1.67 [0.41, 6.77]
21.13 salivation ‐ by 6 weeks	1	122	Risk Ratio (M‐H, Random, 95% CI)	0.02 [0.00, 0.36]
21.14 salivation by 18 weeks	2	330	Risk Ratio (M‐H, Random, 95% CI)	0.32 [0.15, 0.67]
21.15 salivation by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	0.17 [0.13, 0.23]
21.16 urinary incontinence ‐ by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	0.45 [0.27, 0.77]
21.17 vision abnormalities ‐ by 8 weeks	1	377	Risk Ratio (M‐H, Random, 95% CI)	1.57 [0.79, 3.15]
21.18 vision abnormalities by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	3.0 [0.12, 73.46]
22 Adverse events: 2a. Extrapyramidal ‐ needing anticholinergic medication	7		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
22.1 by 8 weeks	5	923	Risk Ratio (M‐H, Random, 95% CI)	0.64 [0.46, 0.91]
22.2 by 3 to 12 months	2	519	Risk Ratio (M‐H, Random, 95% CI)	0.56 [0.40, 0.79]
23 Adverse events: 2b. Extrapyramidal ‐ any	5		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
23.1 any extrapyramidal event by 8 weeks	4	657	Risk Ratio (M‐H, Random, 95% CI)	0.76 [0.54, 1.08]
23.2 any extrapyramidal event by 3‐12 months	1	339	Risk Ratio (M‐H, Random, 95% CI)	0.60 [0.41, 0.88]
24 Adverse events: 2c. Extrapyramidal ‐ parkinsonism ESRS (endpoint, high score=poor, skewed data)			Other data	No numeric data
24.1 by 8 weeks			Other data	No numeric data
24.2 by 1 year			Other data	No numeric data
25 Adverse events: 2d. Extrapyramidal ‐ specific symptoms	7		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
25.1 akathisia ‐ by 3 to 12 months	4	924	Risk Ratio (M‐H, Random, 95% CI)	0.98 [0.64, 1.49]
25.2 akathisia ‐ by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	1.63 [1.07, 2.48]
25.3 dyskinetic movements ‐ by 8 weeks	1	175	Risk Ratio (M‐H, Random, 95% CI)	0.70 [0.42, 1.15]
25.4 dyskinetic movements ‐ by 3 to 12 months	2	519	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.40, 2.46]
25.5 new parkinsonism ‐ by 3 to 12 months	2	519	Risk Ratio (M‐H, Random, 95% CI)	0.61 [0.40, 0.94]
25.6 restlessness ‐ by 6 weeks	1	122	Risk Ratio (M‐H, Random, 95% CI)	1.0 [0.15, 6.87]
25.7 tardive dyskinesia	1	180	Risk Ratio (M‐H, Random, 95% CI)	1.5 [0.44, 5.14]
25.8 tremor ‐ by 6 to 8 weeks	2	297	Risk Ratio (M‐H, Random, 95% CI)	0.54 [0.15, 1.91]
26 Adverse events: 3. Treatment emergent adverse events ‐by 4 to 8 weeks	3	569	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.82, 0.98]
27 Adverse events: 4. Other:	9		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
27.1 agitation ‐ by 8 weeks	2	552	Risk Ratio (M‐H, Random, 95% CI)	0.96 [0.48, 1.93]
27.2 agitation ‐ by 18 to 26 weeks	2	435	Risk Ratio (M‐H, Random, 95% CI)	1.10 [0.25, 4.97]
27.3 anxiety ‐ by 6 to 8 weeks	2	502	Risk Ratio (M‐H, Random, 95% CI)	0.80 [0.30, 2.14]
27.4 anxiety ‐ by 18 to 26 weeks	2	435	Risk Ratio (M‐H, Random, 95% CI)	0.46 [0.11, 1.94]
27.5 asthenia ‐ by 4 weeks	1	125	Risk Ratio (M‐H, Random, 95% CI)	0.69 [0.16, 2.97]
27.6 asthenia ‐ by 18 weeks	2	330	Risk Ratio (M‐H, Random, 95% CI)	0.66 [0.32, 1.37]
27.7 asthenia ‐ by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	0.51 [0.32, 0.82]
27.8 decrease in white blood cells ‐ 4 weeks	1	122	Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
27.9 decrease in white blood cells ‐ by 18 to 26 weeks	2	330	Risk Ratio (M‐H, Random, 95% CI)	0.20 [0.03, 1.13]
27.10 decrease in white blood cells ‐ by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	0.44 [0.25, 0.76]
27.11 diabetes mellitus ‐ by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	1.26 [0.77, 2.05]
27.12 dyspepsia ‐ by 4 weeks	1	125	Risk Ratio (M‐H, Random, 95% CI)	0.92 [0.43, 1.97]
27.13 dyspepsia ‐ by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	0.69 [0.50, 0.95]
27.14 fever ‐ by 18 weeks	1	180	Risk Ratio (M‐H, Random, 95% CI)	0.2 [0.02, 1.68]
27.15 flu syndrome ‐ by 18 weeks	2	330	Risk Ratio (M‐H, Random, 95% CI)	0.84 [0.39, 1.82]
27.16 headache ‐ 4 to 8 weeks	3	677	Risk Ratio (M‐H, Random, 95% CI)	0.72 [0.54, 0.96]
27.17 headache ‐ 18 to 26 weeks	2	405	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.57, 1.44]
27.18 hyperprolactinaemia ‐ by 6 weeks	1	42	Risk Ratio (M‐H, Random, 95% CI)	0.24 [0.10, 0.58]
27.19 increased transaminase ‐ by 6 weeks	1	122	Risk Ratio (M‐H, Random, 95% CI)	0.2 [0.02, 1.66]
27.20 nausea ‐ by 4 weeks	1	125	Risk Ratio (M‐H, Random, 95% CI)	0.51 [0.18, 1.44]
27.21 nausea ‐ by 18 to 26 weeks	2	435	Risk Ratio (M‐H, Random, 95% CI)	0.27 [0.09, 0.75]
27.22 nausea ‐ by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	0.67 [0.49, 0.90]
27.23 pain ‐ by 4 weeks	1	125	Risk Ratio (M‐H, Random, 95% CI)	0.59 [0.24, 1.42]
27.24 postural hypotension ‐ by 18 weeks	1	150	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.33, 2.23]
27.25 SGPT increased ‐ by 4 weeks	1	125	Risk Ratio (M‐H, Random, 95% CI)	10.17 [0.57, 180.04]
27.26 sweating ‐ by 18 weeks	1	180	Risk Ratio (M‐H, Random, 95% CI)	0.40 [0.08, 2.01]
27.27 tachycardia ‐ by 6 weeks	1	122	Risk Ratio (M‐H, Random, 95% CI)	0.06 [0.01, 0.40]
27.28 tooth disorder	1	180	Risk Ratio (M‐H, Random, 95% CI)	0.11 [0.01, 2.03]
27.29 vomiting ‐ by 4 weeks	1	125	Risk Ratio (M‐H, Random, 95% CI)	0.35 [0.10, 1.24]
27.30 vomiting ‐ by 18 weeks	1	180	Risk Ratio (M‐H, Random, 95% CI)	0.8 [0.22, 2.88]
27.31 vomiting ‐ by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	0.61 [0.45, 0.83]
28 Adverse events: 8. Simpson and Angus Scale ‐ by 2 months (endpoint, high score=poor, skewed data)			Other data	No numeric data
29 Adverse events: 9. Yale Brown Obsessive Compulsive Scale ‐ by 6 weeks (endpoint, high score=poor, skewed da			Other data	No numeric data
30 Adverse events: 10. Sleep problems	8		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
30.1 insomnia by 6 to 8 weeks	2	499	Risk Ratio (M‐H, Random, 95% CI)	1.43 [0.20, 10.09]
30.2 insomnia by 18 to 26 weeks	1	255	Risk Ratio (M‐H, Random, 95% CI)	0.34 [0.20, 0.60]
30.3 insomnia ‐ by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	1.57 [1.28, 1.93]
30.4 somnolence by 6 to 8 weeks	4	687	Risk Ratio (M‐H, Random, 95% CI)	0.74 [0.45, 1.20]
30.5 somnolence by 18 to 26 weeks	3	585	Risk Ratio (M‐H, Random, 95% CI)	0.86 [0.33, 2.28]
30.6 somnolence ‐ by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	0.57 [0.48, 0.67]
31 Adverse events: 6. Weight gain	7		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
31.1 4 to 8 weeks	3	250	Risk Ratio (M‐H, Random, 95% CI)	0.92 [0.42, 2.02]
31.2 3 to 12 months	3	585	Risk Ratio (M‐H, Random, 95% CI)	1.76 [0.68, 4.59]
31.3 over 12 months	1	980	Risk Ratio (M‐H, Random, 95% CI)	1.73 [1.49, 2.00]
32 Adverse events: 7. Weight gain	4		Mean Difference (IV, Random, 95% CI)	Subtotals only
32.1 by 3‐12 months	3	535	Mean Difference (IV, Random, 95% CI)	2.18 [‐0.63, 4.98]
32.2 over 12 months	1	44	Mean Difference (IV, Random, 95% CI)	10.09 [‐0.68, 20.86]
33 Adverse events: 8. Weight ‐ by 8 weeks (change data)	3	463	Mean Difference (IV, Random, 95% CI)	1.42 [0.62, 2.23]
34 Quality of life: 1. QOL total ‐ by 3 to 12 months(endpoint, high score=good)	1	240	Mean Difference (IV, Random, 95% CI)	2.42 [‐3.93, 8.77]
35 Quality of life: 2. QLS total by 3 to 12 months (endpoint, high score=good)	2	172	Mean Difference (IV, Random, 95% CI)	‐1.39 [‐9.80, 7.01]
36 Service outcome:1. Discharged from hospital ‐ by 8 weeks	1	214	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.88, 1.07]
37 Service outcome: 2. Hospitalisation for imminent risk of suicide ‐ by 2 years	1	980	Risk Ratio (M‐H, Random, 95% CI)	1.28 [1.02, 1.61]
38 Service outcome: 3. Increase in level of surveillance ‐ by two years	1	980	Risk Ratio (M‐H, Random, 95% CI)	1.39 [0.95, 2.04]

5.1. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 1 Death.

5.2. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 2 Global effect: 1. No clinically important response.

5.3. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 3 Global effect: 2. Relapse/hospitalisation by 52 weeks.

5.4. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 4 Global effect: 3. CGI total by 3 ‐12 months (endpoint, high score=poor).

5.5. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 5 Mental state: 1a. Needing addtional benzodiazepines.

5.6. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 6 Mental state: 1b. Needing antidepressants by 2 years.

5.7. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 7 Mental state: 2. No important clinical response (decrease in PANSS total score).

5.8. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 8 Mental state: 3. No important clinical response (decrease in BPRS total score).

5.9. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 9 Mental state: 4. BPRS total at 8 weeks (endpoint, high score=worse).

5.10. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 10 Mental state: 5. BPRS total at 3‐12 months (endpoint, high score=poor, skewed data).

Mental state: 5. BPRS total at 3‐12 months (endpoint, high score=poor, skewed data)
Study	Intervention	Mean	SD	N
Beuzen 1998 (HGCF)	Olanzapine	27.16	17.10	89
Beuzen 1998 (HGCF)	Clozapine	27.05	14.93	87
Thomas 1998 (HGBU)	Olanzapine	15.5	14.2	32
Thomas 1998 (HGBU)	Risperidone	19.4	8.2	30
Tran 1997 (HGBG)	Olanzapine	19.66	14.86	166
Tran 1997 (HGBG)	Risperidone	21.05	12.73	165

5.11. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 11 Mental state: 6. PANSS total (endpoint, high score=poor).

5.12. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 12 Mental state: 7. Negative symptoms (PANSS, endpoint, high score=poor).

5.13. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 13 Mental state: 8. Negative symptoms (various scales, endpoint, high score=poor, skewed data).

Mental state: 8. Negative symptoms (various scales, endpoint, high score=poor, skewed data)
Study	Intervention	Mean	SD	N
as measured by BPRS ‐ by 18 weeks
Beuzen 1998 (HGCF)	Olanzapine	5.15	3.22	89
Beuzen 1998 (HGCF)	Clozapine	5.28	2.57	87
as measured by SANS ‐ by 28 weeks
Tran 1997 (HGBG)	Olanzapine	7.94	5.19	157
Tran 1997 (HGBG)	Risperidone	8.70	4.69	151
PANSS ‐ by 6 to 8 weeks
Martin 2002	Olanzapine	‐2.2	5.8	186
Martin 2002	Amisulpride	‐2.0	5.7	186
Wang 2003	Olanzapine	9.37	5.70	31
Wang 2003	Clozapine	10.01	7.73	31

5.14. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 14 Mental state: 9. Negative symptoms change data at 6 months (SANS, high score=poor).

5.15. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 15 Mental state: 10. Positive symptoms (PANSS, endpoint, high score=poor).

5.16. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 16 Mental state: 11. Positive symptoms (various scales, endpoint, high score=poor, skewed data).

Mental state: 11. Positive symptoms (various scales, endpoint, high score=poor, skewed data)
Study	Intervention	Mean	SD	N
as measured by BPRS ‐ by 18 weeks
Beuzen 1998 (HGCF)	Olanzapine	8.16	5.29	89
Beuzen 1998 (HGCF)	Clozapine	8.40	5.07	87
as measured by PANSS ‐ by 6 to 8 weeks
Martin 2002	Olanzapine	‐9.9	7.3	186
Martin 2002	Amisulpride	‐10.8	7.3	186
Wang 2003	Olanzapine	6.07	5.03	31
Wang 2003	Risperidone	7.35	5.98	31
as measured by PANSS at 28 weeks
Tran 1997 (HGBG)	Olanzapine	15.35	7.74	166
Tran 1997 (HGBG)	Risperidone	15.55	6.60	165

5.17. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 17 Mental state: 12. Depression ‐ by 8 weeks (Hamilton or MADRS rating scale, high score=poor, skewed data).

Mental state: 12. Depression ‐ by 8 weeks (Hamilton or MADRS rating scale, high score=poor, skewed data)
Study	Intervention	Mean	SD	N
Conley 2001	Olanzapine	8.3	6.0	189
Conley 2001	Risperidone	7.4	5.4	188
Martin 2002	Olanzapine	‐6.3	7.7	186
Martin 2002	Amisulpride	‐6.0	8.8	186

5.18. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 18 Mental state: 13. Depression ‐ by 18 weeks (MADRS, high score=poor, skewed data).

Mental state: 13. Depression ‐ by 18 weeks (MADRS, high score=poor, skewed data)
Study	Intervention	mean	SD	N
Martin 2002	Olanzapine	‐6.3	7.7	186
Martin 2002	Amisulpride	‐6.0	8.8	186

5.19. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 19 Leaving the study early: 1. Any reason.

5.20. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 20 Leaving the study early: 2. Due to lack of efficacy.

5.21. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 21 Adverse events: 1. Anticholinergic effects ‐ specific symptoms.

5.22. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 22 Adverse events: 2a. Extrapyramidal ‐ needing anticholinergic medication.

5.23. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 23 Adverse events: 2b. Extrapyramidal ‐ any.

5.24. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 24 Adverse events: 2c. Extrapyramidal ‐ parkinsonism ESRS (endpoint, high score=poor, skewed data).

Adverse events: 2c. Extrapyramidal ‐ parkinsonism ESRS (endpoint, high score=poor, skewed data)
Study	Intervention	Mean	SD	N
by 8 weeks
Conley 2001	Olanzapine	3.8	4.6	145
Conley 2001	Risperidone	4.2	4.9	133
by 1 year
Jones 1998 (P022)	Olanzapine	1.19	1.86	21
Jones 1998 (P022)	Risperidone	3.24	2.84	21

5.25. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 25 Adverse events: 2d. Extrapyramidal ‐ specific symptoms.

5.26. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 26 Adverse events: 3. Treatment emergent adverse events ‐by 4 to 8 weeks.

5.27. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 27 Adverse events: 4. Other:.

5.28. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 28 Adverse events: 8. Simpson and Angus Scale ‐ by 2 months (endpoint, high score=poor, skewed data).

Adverse events: 8. Simpson and Angus Scale ‐ by 2 months (endpoint, high score=poor, skewed data)
Study	Intervention	mean	SD	N
Martin 2002	Olanzapine	0.21	0.36	185
Martin 2002	Amisulpride	0.24	0.40	186

5.29. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 29 Adverse events: 9. Yale Brown Obsessive Compulsive Scale ‐ by 6 weeks (endpoint, high score=poor, skewed da.

Adverse events: 9. Yale Brown Obsessive Compulsive Scale ‐ by 6 weeks (endpoint, high score=poor, skewed da
Study
de Haan 2002	Olanzapine
de Haan 2002	Risperidone

5.30. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 30 Adverse events: 10. Sleep problems.

5.31. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 31 Adverse events: 6. Weight gain.

5.32. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 32 Adverse events: 7. Weight gain.

5.33. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 33 Adverse events: 8. Weight ‐ by 8 weeks (change data).

5.34. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 34 Quality of life: 1. QOL total ‐ by 3 to 12 months(endpoint, high score=good).

5.35. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 35 Quality of life: 2. QLS total by 3 to 12 months (endpoint, high score=good).

5.36. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 36 Service outcome:1. Discharged from hospital ‐ by 8 weeks.

5.37. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 37 Service outcome: 2. Hospitalisation for imminent risk of suicide ‐ by 2 years.

5.38. Analysis.

Comparison 5 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS, Outcome 38 Service outcome: 3. Increase in level of surveillance ‐ by two years.

Comparison 6. OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Death ‐ by 18 weeks	1	180	Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2 Global effect: No clinically important response ‐ by 18 weeks	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
2.1 as measured by CGI	2	330	Risk Ratio (M‐H, Random, 95% CI)	0.89 [0.73, 1.08]
2.2 as measured by Kane criteria	2	330	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.80, 1.17]
3 Mental state: 1. No important clinical response ‐ by 18 weeks (20% decrease in PANSS)	2	330	Risk Ratio (M‐H, Random, 95% CI)	1.00 [0.68, 1.49]
4 Mental state: 2. BPRS total ‐ by 18 weeks (endpoint, high score=poor, skewed data)			Other data	No numeric data
5 Mental state: 3. PANSS total ‐ by 18 weeks (endpoint, high score=poor)	1	176	Mean Difference (IV, Random, 95% CI)	0.03 [‐8.07, 8.13]
6 Mental state: 4. Negative symptoms ‐ by 18 weeks (BPRS, endpoint, high score=poor, skewed data)			Other data	No numeric data
7 Mental state: 5. Negative symptoms ‐ by 18 weeks (PANSS, endpoint, high score=poor)	1	176	Mean Difference (IV, Random, 95% CI)	‐0.75 [‐2.94, 1.44]
8 Mental state: 6. Positive symptoms ‐ by 18 weeks (BPRS, endpoint, high score=poor, skewed data)			Other data	No numeric data
9 Leaving the study early. Any reason	4	457	Risk Ratio (M‐H, Random, 95% CI)	0.98 [0.81, 1.19]
9.1 by 8 weeks	1	13	Risk Ratio (M‐H, Random, 95% CI)	4.67 [0.29, 75.02]
9.2 by 3‐12 months	3	444	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.80, 1.18]
10 Leaving the study early. due to lack of efficacy ‐ by 18 weeks	2	330	Risk Ratio (M‐H, Random, 95% CI)	1.32 [0.65, 2.70]
11 Adverse events: 1. HAS ‐ by 18 weeks (change, high score=poor)	1	137	Mean Difference (IV, Random, 95% CI)	0.40 [‐2.30, 3.10]
12 Adverse events: 2. AIMS ‐ by 18 weeks (change, high score=poor)	1	139	Mean Difference (IV, Random, 95% CI)	0.30 [‐0.58, 1.18]
13 Adverse events: 3. Simpson and Angus Scale ‐ by 18 weeks (change, high score=poor)	1	139	Mean Difference (IV, Random, 95% CI)	‐0.10 [‐1.56, 1.36]
14 Adverse events: 4. Anticholinergic effects ‐ by 18 weeks	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
14.1 constipation	1	180	Risk Ratio (M‐H, Random, 95% CI)	0.35 [0.15, 0.85]
14.2 dry mouth	1	180	Risk Ratio (M‐H, Random, 95% CI)	1.67 [0.94, 2.95]
14.3 dizziness	2	330	Risk Ratio (M‐H, Random, 95% CI)	0.39 [0.14, 1.11]
14.4 headache	2	330	Risk Ratio (M‐H, Random, 95% CI)	1.67 [0.82, 3.38]
14.5 rhinitis	1	180	Risk Ratio (M‐H, Random, 95% CI)	1.67 [0.41, 6.77]
14.6 salivation	2	330	Risk Ratio (M‐H, Random, 95% CI)	0.40 [0.13, 1.21]
15 Adverse events: 5. Extrapyramidal ‐ by 18 weeks	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
15.1 akathisia	2	330	Risk Ratio (M‐H, Random, 95% CI)	1.38 [0.53, 3.64]
15.2 dyskinesia	1	180	Risk Ratio (M‐H, Random, 95% CI)	1.25 [0.35, 4.50]
15.3 new parkinsonism	1	180	Risk Ratio (M‐H, Random, 95% CI)	0.78 [0.30, 2.00]
15.4 tardive dyskinesia	1	180	Risk Ratio (M‐H, Random, 95% CI)	1.5 [0.44, 5.14]
16 Adverse events: Other ‐ by 18 weeks	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
16.1 agitation ‐ by 18 weeks	1	180	Risk Ratio (M‐H, Random, 95% CI)	2.5 [0.81, 7.68]
16.2 anxiety	1	180	Risk Ratio (M‐H, Random, 95% CI)	1.0 [0.30, 3.34]
16.3 asthenia	2	330	Risk Ratio (M‐H, Random, 95% CI)	0.66 [0.32, 1.37]
16.4 backpain	1	150	Risk Ratio (M‐H, Random, 95% CI)	2.43 [0.80, 7.42]
16.5 decrease in white blood cells	1	180	Risk Ratio (M‐H, Random, 95% CI)	0.2 [0.02, 1.68]
16.6 fever	1	180	Risk Ratio (M‐H, Random, 95% CI)	0.2 [0.02, 1.68]
16.7 flu syndrome	2	330	Risk Ratio (M‐H, Random, 95% CI)	0.84 [0.39, 1.82]
16.8 nausea	1	180	Risk Ratio (M‐H, Random, 95% CI)	0.10 [0.01, 0.77]
16.9 postural hypotension	1	150	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.33, 2.23]
16.10 somnolence	2	330	Risk Ratio (M‐H, Random, 95% CI)	0.54 [0.32, 0.88]
16.11 sweating	1	180	Risk Ratio (M‐H, Random, 95% CI)	0.40 [0.08, 2.01]
16.12 tooth disorder	1	180	Risk Ratio (M‐H, Random, 95% CI)	0.11 [0.01, 2.03]
16.13 vomiting	1	180	Risk Ratio (M‐H, Random, 95% CI)	0.8 [0.22, 2.88]
16.14 weight gain	2	330	Risk Ratio (M‐H, Random, 95% CI)	1.10 [0.60, 2.03]

6.1. Analysis.

Comparison 6 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 1 Death ‐ by 18 weeks.

6.2. Analysis.

Comparison 6 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 2 Global effect: No clinically important response ‐ by 18 weeks.

6.3. Analysis.

Comparison 6 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 3 Mental state: 1. No important clinical response ‐ by 18 weeks (20% decrease in PANSS).

6.4. Analysis.

Comparison 6 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 4 Mental state: 2. BPRS total ‐ by 18 weeks (endpoint, high score=poor, skewed data).

Mental state: 2. BPRS total ‐ by 18 weeks (endpoint, high score=poor, skewed data)
Study	Intervention	Mean	SD	N
Beuzen 1998 (HGCF)	Olanzapine	27.16	17.10	89
Beuzen 1998 (HGCF)	Clozapine	27.05	14.93	87

6.5. Analysis.

Comparison 6 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 5 Mental state: 3. PANSS total ‐ by 18 weeks (endpoint, high score=poor).

6.6. Analysis.

Comparison 6 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 6 Mental state: 4. Negative symptoms ‐ by 18 weeks (BPRS, endpoint, high score=poor, skewed data).

Mental state: 4. Negative symptoms ‐ by 18 weeks (BPRS, endpoint, high score=poor, skewed data)
Study	Intervention	Mean	SD	N
Beuzen 1998 (HGCF)	Olanzapine	5.15	3.22	89
Beuzen 1998 (HGCF)	Clozapine group	5.28	2.57	87

6.7. Analysis.

Comparison 6 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 7 Mental state: 5. Negative symptoms ‐ by 18 weeks (PANSS, endpoint, high score=poor).

6.8. Analysis.

Comparison 6 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 8 Mental state: 6. Positive symptoms ‐ by 18 weeks (BPRS, endpoint, high score=poor, skewed data).

Mental state: 6. Positive symptoms ‐ by 18 weeks (BPRS, endpoint, high score=poor, skewed data)
Study	Intervention	Mean	SD	N
Beuzen 1998 (HGCF)	Olanzapine	8.16	5.29	89
Beuzen 1998 (HGCF)	Clozapine	8.40	5.07	87

6.9. Analysis.

Comparison 6 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 9 Leaving the study early. Any reason.

6.10. Analysis.

Comparison 6 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 10 Leaving the study early. due to lack of efficacy ‐ by 18 weeks.

6.11. Analysis.

Comparison 6 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 11 Adverse events: 1. HAS ‐ by 18 weeks (change, high score=poor).

6.12. Analysis.

Comparison 6 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 12 Adverse events: 2. AIMS ‐ by 18 weeks (change, high score=poor).

6.13. Analysis.

Comparison 6 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 13 Adverse events: 3. Simpson and Angus Scale ‐ by 18 weeks (change, high score=poor).

6.14. Analysis.

Comparison 6 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 14 Adverse events: 4. Anticholinergic effects ‐ by 18 weeks.

6.15. Analysis.

Comparison 6 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 15 Adverse events: 5. Extrapyramidal ‐ by 18 weeks.

6.16. Analysis.

Comparison 6 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR PEOPLE WITH TREATMENT RESISTANT ILLNESS, Outcome 16 Adverse events: Other ‐ by 18 weeks.

Comparison 7. OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR FIRST EPISODE.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Mental state: 1. No important clinical response ‐ by 6 weeks (50% decrease in PANSS total score)	1	42	Risk Ratio (M‐H, Random, 95% CI)	0.83 [0.30, 2.31]
2 Adverse events: 1. Extrapyramidal ‐ needing anticholinergic medication ‐ by 6 weeks	1	42	Risk Ratio (M‐H, Random, 95% CI)	0.38 [0.17, 0.89]
3 Adverse events: 2. Exprapyramidal ‐ any ‐ by 6 weeks	1	42	Risk Ratio (M‐H, Random, 95% CI)	0.40 [0.19, 0.83]
4 Adverse events: 3. Anticholinergic effects ‐ rhinitis ‐ by 6 weeks	1	42	Risk Ratio (M‐H, Random, 95% CI)	0.13 [0.02, 0.91]
5 Adverse events: 4. hyperprolactinaemia ‐ by 6 weeks	1	42	Risk Ratio (M‐H, Random, 95% CI)	0.24 [0.10, 0.58]

7.1. Analysis.

Comparison 7 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR FIRST EPISODE, Outcome 1 Mental state: 1. No important clinical response ‐ by 6 weeks (50% decrease in PANSS total score).

7.2. Analysis.

Comparison 7 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR FIRST EPISODE, Outcome 2 Adverse events: 1. Extrapyramidal ‐ needing anticholinergic medication ‐ by 6 weeks.

7.3. Analysis.

Comparison 7 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR FIRST EPISODE, Outcome 3 Adverse events: 2. Exprapyramidal ‐ any ‐ by 6 weeks.

7.4. Analysis.

Comparison 7 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR FIRST EPISODE, Outcome 4 Adverse events: 3. Anticholinergic effects ‐ rhinitis ‐ by 6 weeks.

7.5. Analysis.

Comparison 7 OLANZAPINE vs ATYPICAL ANTIPSYCHOTICS FOR FIRST EPISODE, Outcome 5 Adverse events: 4. hyperprolactinaemia ‐ by 6 weeks.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Altamura 1999 (HGBQ).

Methods	Allocation: randomised, computer‐generated, blocks for each investigator 1:1, concealed from investigators. Blindness: double, medication kits issued. Duration: 14 weeks (preceded by screening phase, unspecified). Design: multicentre, 1 week wash out period.
Participants	Diagnosis: schizophrenia, paranoid (DSM‐IV). History: informed consent obtained. N=28.* Age: not reported. Sex: not clearly reported. Setting: not reported. Inclusion criteria: partial or non‐responders to treatment according to preset criteria.
Interventions	1. Olanzapine: dose range 5‐20 mg, mean 12.4 mg/day. N=13. 2. Haloperidol: dose range 5‐20 mg, mean 12.3 mg/day. N=15.
Outcomes	Leaving the study early. Unable to use ‐ Global state: CGI (no data). Mental state: BPRS, SANS (no usable data). Adverse events: AIMS, BAS, SAS (no usable data), weight gain (unpublished data with unclear denominator).
Notes	*Data supplied by company suggests 48 people were randomised and presents detailed results. Professor Altamura, however, in a personal communication suggests that these are projected figures and not true ratings.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Avasthi 2001.

Methods	Allocation: randomised. Blindness: open label. Duration: 12 weeks. Design: parallel groups, with washout period.
Participants	Diagnosis: schizophrenia (DSM‐IV). History: informed consent obtained. N=27. Age: between 18‐65 years. Sex: male and female. Setting: hospital & community. Inclusion criteria: =/> 3 on the CGI severity scale. Excluded: hepatitis, jaundice, current agranulocytosis.
Interventions	1. Olanzapine: dose range 5‐20 mg/day. N=17. 2. Haloperidol: dose range 5‐20 mg/day. N=10.
Outcomes	Leaving the study early. Global state: CGI. Mental state: BPRS, PANSS‐P, PANSS‐N, MADRS, SANS. Adverse events: BAS, SAS, treatment emergent side‐effects. Quality of Life: QoL. Weight gain. Unable to use ‐ Mental state: HAM‐S (no usable data). Adverse events: UKU (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Barak 2002.

Methods	Allocation: randomised ‐ no further details. Blindness: open label. Duration: 3 months. Design: parallel group.
Participants	Diagnosis: schizophrenia (DSM IV). Elderly. History: mean length of illness 35.3 years, informed consent obtained. N=20. Age: range 68‐83, mean 72.7 years. Sex: 10 M, 10 F. Setting: hospitalised. Inclusion criteria: PANSS total score of >50.
Interventions	1. Olanzapine: dose 13.1 mg/day. N=10. 2. Haloperidol: dose 7.2 mg/day. N=10.
Outcomes	Leaving study early. Global state: CGI. Mental state: PANSS. Weight change. Additional medication.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Beasley 1996a (HGAD).

Methods	Allocation: randomised, computer generated, block by investigator, concealed from investigator. Blindness: double, medication kits issued. Duration: 6 weeks (preceded by placebo lead‐in of 4‐7 days; 46 week extension for responders). Investigators: trained on BPRS & SANS. Design: dosage study.
Participants	Diagnosis: schizophrenia (DSM‐III‐R). History: acute exacerbation, informed consent obtained. N=335. Age: mode ˜ 30's, range 18‐65. Sex: ˜ 90% M. Setting: hospital & community. Inclusion criteria: BPRS >23, CGI >3, off neuroleptics prior to study, lead‐in period responders (BPRS total decreased by >24% / <24) excluded. Excluded: substance misuse within 3 months, serious suicidal risk, unstable medical illness, parkinsons disease, myasthenia gravis, hepatitis, jaundice, history of seizures, leukopenia.
Interventions	1. Olanzapine low: dose 2.5‐7.5 mg/day. N=65. 2. Olanzapine medium: dose 7.5‐12.5 mg/day. N=64. 3. Olanzapine high: dose 12.5‐17.5mg/day*. N=69. 4. Haloperidol: dose 10‐20 mg/day. N=69. 5. Placebo: N=68. Up to 10mg/day lorazepam (day 1‐21) & benztropine 6mg/day as required.
Outcomes	Leaving study early. Global effect: CGI‐S. Mental state: BPRS‐anchored version**, needing additional benzodiazepines, SANS. Adverse events: AIMS, BAS, SAS, requiring benztropine, COSTART list, weight change. Unable to use ‐ Hospital status: (no data). Laboratory tests and physiological measures: (no data).
Notes	*Chosen as comparator with other trials as mean dose = 13.2mg/day. **A priori efficacy >39 decrease from baseline or to <19 BPRS total.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Beasley 1996b (HGAP).

Methods	Allocation: randomised, computer‐generated, block by investigative site, concealed from investigator. Blindness: double, medication kits issued. Duration: 6 weeks (preceded by placebo lead‐in of 4‐7 days). Investigators: trained in PANSS. Design: multicentre, 12 sites in USA.
Participants	Diagnosis: schizophrenia (DSM‐III‐R). History: informed consent obtained. N=152. Age: mean 38 years. Sex: ˜ 70% M. Setting: initially all in hospital.* Inclusion criteria: BPRS >23, CGI >3, not involved in previous olanzapine trial, off neuroleptics prior to entering study, lead‐in period responders (BPRS total decreased by >24% / <24) excluded. Excluded: serious unstable medical conditions, parkinsons disease, myasthenia gravis, hepatitis, jaundice, history of seizures, leukopenia.
Interventions	1. Olanzapine: dose 1 mg/day. N=52. 2. Olanzapine: dose 10mg/day. N=50. 3. Placebo: N=50. Up to 10mg/day lorazepam (day 1‐21) & benztropine 6mg/day allowed as required.
Outcomes	Leaving study early. Global state: CGI‐S. Mental State: BPRS**, needing additional benzodiazepines, PANSS. Adverse events: AIMS, BAS, requiring benztropine, SAS, COSTART list, weight change. Unable to use ‐ Hospital status: (no data). Global state: PGI (no data). Laboratory tests & physiological measures: (no data).
Notes	* Eligible for discharge if BPRS total decreased by >24% from baseline or was <24. ** BPRS (scored 0‐6) extracted from PANSS ‐ no reference given for validity of procedure. ** A priori efficacy >39 decrease from baseline or to <19 BPRS total.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Beasley 1997 (E003).

Methods	Allocation: randomised, computer‐generated ‐ blocks of 5, concealed from investigators. Blindness: double, medication kits issued. Duration: 6 weeks (preceded by placebo lead‐in of 4‐7 days: 46 week extension for responders). Design: multicentre, 50 sites, Europe, South Africa, Israel and Australia.
Participants	Diagnosis: schizophrenia (DSM‐III‐R). History: informed consent obtained. N=431. Age:18‐65. Sex: 275 M, 156 F. Setting: initially all in hospital.* Inclusion criteria: BPRS >23, CGI >3, off neuroleptics prior to entering study, lead‐in period responders (BPRS total decreased by >24% / <24) excluded. Excluded: substance misuse within last 3 months, suicide risk, parkinsons disease, myasthenia gravis, hepatitis, jaundice, history of seizures, leukopenia.
Interventions	1. Olanzapine (OLZ‐1): dose 1mg/day. N=88. 2. Olanzapine (OLZ‐L): dose 2.5‐7.5mg/day. N=87. 3. Olanzapine (OLZ‐M): dose 7.5‐12.5mg/day. N=86. 4. Olanzapine (OLZ‐H): dose 12.5‐17.5mg/day**. N=89. 5. Haloperidol (HAL): dose 10‐20mg/day. N=81. Up to 10mg/day benzodiazipine (day 1‐21) & biperiden up to 6mg/day as required.
Outcomes	Leaving study early. Global state: CGI‐S. Mental State: BPRS***, needing additional benzodiazepines, PANSS. Death.**** Adverse events: AIMS, BAS, requiring benztropine, SAS, COSTART list, weight change. Unable to use ‐ Hospital status: (no data). Global state: PGI (no data). Laboratory tests & physiological measures: (no data).
Notes	* Eligible for discharge if BPRS total decreased by >24% from baseline or was <24. ** Chosen as the comparitor with other trials as mean dose = 13.2mg/day. *** BPRS (scored 0‐6) extracted from PANSS ‐ no reference given for validity of procedure. *** A priori efficacy >39 decrease from baseline or to <19 total. ****Deaths reported during the study or within 30 days of discontinuation.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Bernardo 2001 (HGDD).

Methods	Allocation: randomised ‐ no further details. Blindness: double. Duration: 4 weeks. Design: parallel group.
Participants	Diagnosis: schizophrenia (DSM IV). History: all experiencing acute psychosis, (n=11, first episode, neuroleptic naive), (n=16 neuroleptic free after minimum washout of 7 days), informed consent obtained. N=27. Age: range 19‐44 years, mean 28. Sex: M 17, F 10. Setting: inpatients. Inclusion criteria: not reported.
Interventions	1. Olanzapine: dose 10 mg/day. N=14. 2. Haloperidol: dose 10 mg/day. N=13.
Outcomes	Leaving the study early. Global improvement: CGI. Mental state: BPRS, PANSS, PANSS‐P, PANSS‐N. Adverse events: SAS, BAS.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Beuzen 1998 (HGCF).

Methods	Allocation: randomised, block design, computer‐generated, concealed from investigators. Blindness: double. Duration: 18 weeks (preceeded by washout period). Design: multicentre, 14 sites, Europe and Africa.
Participants	Diagnosis: schizophrenia (DSM IV). History: informed consent obtained. N=180. Age: range 18‐70 years. Sex: male and female. Setting: inpatients and outpatients. Inclusion criteria: treatment resistant, >3 on at least 2 items of PANSS positive sub‐scale. Excluded: previously treated with olanzapine or were olanzapine or clozapine non‐responders.
Interventions	1. Olanzapine: dose range 15‐25 mg/day. N=90. 2. Clozapine: dose range 200‐600 mg/day. N=90. Benzodiazipines, biperiden, benztropine & chloral hydrate as required.
Outcomes	Leaving study early. Global improvement: CGI. Mental state: BPRS, Kane criteria, PANSS, COSTART list, weight change. Unable to use ‐ Global improvement: use of additional medication (no data). Laboratory tests & physiological measures (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Bitter 2004 (HGCK).

Methods	Allocation: randomised ‐ no further details. Blindness: double. Duration: 18 weeks. Design: multicentre, Hungary and South Africa.
Participants	Diagnosis: treatment resistant or treatment intolerant schizophrenia (DSM IV). History: informed consent obtained. N=150. Age: 18‐65 years. Sex: 88 M, 59 F. Setting: hospitalised. Inclusion criteria: minimum score of at at least 42 of the BPRS, extracted from the PANSS scale; failure to respond to standard antipsychotic treatment (at least one treatment trial of 4‐6 weeks duration at 400‐600mg‐equivalents of chlorpromazine, because of either insufficient effectiveness or intolerable side effects). Excluded: failure on clozapine due to adverse event or lack of efficacy. Substance related disorder, serious suicide risk, pregnancy, breast feeding, seizures, leukopenia, leucocytes <3.5 GI/L, neutrophils < 2.0 GI/L, jaundice, or active hepatitis B.
Interventions	1. Olanzapine: dose range 5‐25 mg/day. N=76. 2. Clozapine: dose range 100‐500 mg/day. N=74.
Outcomes	No important clinical response. Leaving the study early. Adverse events. Unable to use ‐ Mental state: PANSS (no usable data). Global state: CGI (no usable data). Addtional medication (data unclear).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Casey 2003.

Methods	Allocation: randomised ‐ no further details. Blindness: double; raters proficiency had to meet predefined criteria. Duration: 28 days. Informed consent obtained. Design: multicentre, parallel group with 4 week washout phase.
Participants	Diagnosis: schizophrenia (DSM IV). N=125*. Age: 18‐65. Sex: 91 M, 29 F. Setting: hospitalised with leave allowed for max. 7 days. Inclusion criteria:PANSS score =/>60 and have had a positive response to treatment with antipsychotics within the previous 2 years. Excluded: current diagnosis of schizoaffective disorder, drug induced psychosis, manic episode, depressive episode, violent, homicidal ideation; pregnancy, lactating females, clinically significant abnormal laboratory data, unstable medical condition or an underlying condition that would confound the interpretation of study results.
Interventions	1. Olanzapine: dose 15 mg/day. N=65. 2. Risperidone: dose 6 mg/day. N=60.
Outcomes	Adverse events. Unable to use ‐ Adverse events: SAS, BAS, AIMS (no usable data). Global state: CGI (no usable data). Mental state: PANSS (no usable data).
Notes	*total randomised was 249 with 4 arms, only 2 arms usable.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Chan 2003.

Methods	Allocation: randomised ‐ no further details. Blindness: single 'rater blind' Duration: 8 weeks. Design: parallel group.
Participants	Diagnosis: schizophrenia (DSM IV). History: informed consent obtained. N=60. Age: 18‐65. Sex: 30 M, 30 F. Setting: hopitalised. Inclusion criteria: not reported.
Interventions	1. Olanzapine: dose range 2.5‐20 mg/day. N=30. 2. Risperidone: dose range 0.5‐6 mg/day. N=30.
Outcomes	Leaving the study early. Needing addtional benzodiazapines. Weight change. Unable to use ‐ Global state: CGI (no usable data). Mental state: BPRS (no usable data). Adverse effects: EPRS (no usable data). Needing addtional antipsychotics (unclear data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Chang 2003.

Methods	Allocation: randomised ‐ no further details. Blindness: not reported. Duration: 8 weeks. Design: parallel group.
Participants	Diagnosis: schizophrenia. History: length of illness 4‐5 years. N=62. Age: 18‐53. Sex: male and female. Setting: hospital and community. Inclusion criteria: not reported.
Interventions	1. Olanzapine: 5‐20 mg/day. N=32. 2. Chlorpromazine: 25‐600 mg/day. N=30.
Outcomes	Leaving the study early. Mental state: BPRS.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Chaudhry 2003.

Methods	Allocation: randomised ‐ no further details. Blindness: double. Duration: 24 weeks. Design: multicentre Phillipines, Malaysia, Pakistan, Thailand, Singapore.
Participants	Diagnosis: schizophrenia (DSM IV). History: informed consent obtained. N=276. Age: mean 32.3 years. Sex: male and female. Setting: outpatients. Inclusion criteria: BPRS score > 18 at baseline.
Interventions	1. Olanzapine: dose range 5‐20 mg/day. N=144. 2. Haloperidol: dose range 5‐20 mg/day. N=132.
Outcomes	Adverse events. Additional medication: anticholinergics. Weight gain.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Chen 2003.

Methods	Allocation: randomised ‐ no further details. Blindness: double. Duration: 8 weeks. Design: parallel group.
Participants	Diagnosis: schizophrenia. History: not reported. N=64. Age: mean 63.3 years. Sex: male and female. Setting: hospitalised. Inclusion criteria: not reported.
Interventions	1. Olanzapine: dose range 5‐25 mg/day. N=30. 2. Clozapine: dose range 25‐500 mg/day. N=34.
Outcomes	Leaving the study early. Mental state: PANSS.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Conley 1998.

Methods	Allocation: randomised ‐ no further details. Blindness: double ‐ no further details. Duration: 8 weeks (preceded by 6 weeks of haloperidol and 1‐2 weeks washout). Investigators: trained on BPRS & SANS. Design: multicentre, 3 sites USA.
Participants	Diagnosis: schizophrenia (DSM‐III‐R). History: treatment resistant, informed consent obtained. N=84. Age: mean 42.7 years. Sex: 62 M, 22 F. Setting: hospital. Inclusion criteria: minimum BPRS of 45, CGI‐S score >3, treatment resistant, non‐responders during haloperidol phase, not responded to 2 previous antipsychotics.
Interventions	1. Olanzapine: dose 25 mg/day. N=42. 2. Chlorpromazine: dose 1200 mg/day & benztropine 4mg/day. N=42. Benzodiazipine during washout & first 3 weeks of trial as required.
Outcomes	Leaving study early. Global state: CGI. Mental state: BPRS*, SANS. Adverse events: BAS, SAS. Unable to use ‐ Global state: use of benzodiazipines (no data). Adverse events: COSTART list, weight change (no data). Hospital status: (no data). Laboratory tests & physiological measures: (no data).
Notes	* A priori efficacy >19 decrease from baseline or to <34 total score. Treatment resistance 1. >2 periods of treatment in last 5 years with antipsychotics (from >2 different classes, excluding haloperidol, dose >/= 1g/day chlorpromazine for 6 weeks) without significant relief; 2. No period of good functioning in last 5 years; and 3. BPRS total >/=45, CGI >/=4, >2 BPRS psychosis items >/=4.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Conley 2001.

Methods	Allocation: randomised, stratified by site , BMI, Investigator, baseline PANSS total and ESRS scores. Blindness: double ‐ capsules identical in appearance, taste and smell. Investigators were provided with a sealed envelope for each subject. Placebo capsules were used to supplement certain doses of risperidone in order to match the number of olanzapine capsules. Duration: 8 weeks (preceeded by 1 week washout). Design: multicentre, 41 sites, 2 of which were excluded (n=30) because of data quality concerns.
Participants	Diagnosis: schizophrenia or schizoaffective disorder (DSM‐IV). History: 48% in RISP group and 51% in OLZ group had previously taken atypical antipsychotics. 16% and 14% respectively had taken depot antipsychotic medication. N=377. Age: 18‐64. Sex: 274 M, 103 F. Setting: hospital (within 4 weeks of admission) or outpatient. Inclusion criteria: PANSS between 60 and 120, not treatment resistant, or refractory to RIS or OLZ, no previous CLOZ.
Interventions	1. Olanzapine: mean modal dose 12.43 mg/day, range 5‐20 mg. N=189. 2. Risperidone: mean modal dose 4.78 mg/day, range 2‐6 mg. N=188. Additional use of anticholinergics and lorazepam allowed.
Outcomes	Leaving study early. Mental state: PANSS, use of addtional lorazepam. Adverse events: ESRS, use of anticholinergic drugs, weight change. Unable to use ‐ Laboratory tests & physiological measures: (no usable data).
Notes	Data from two sites not included because of concerns over data quality, (n=30). All individual participants files audited post study.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Corrigan 2004.

Methods	Allocation: randomised ‐ no further details. Blindness: double, identical capsules. Duration: 6 weeks. Design: multicentre. Informed consent obtained.
Participants	Diagnosis: schizophrenia (DSM IV). N=467. Age: 18‐65. Sex: male and female. Setting: hospitalised. Inclusion criteria: PANSS score >/=60. Excluded: clinically significant disease, clinically significant abnormalities on physical examination, electrocardiogram, lab tests, history of failure to respond to antipsychotic drugs, drug and alcohol abuse.
Interventions	1.Olanzapine: dose 15 mg/day. N=93. 2. Sonepiprazole: dose 1.5 mg/day. N=96. 3. Sonepiprazole: dose 10 mg/day. N=99. 4. Sonepiprazole: dose 60 mg/day. N=91. 5. Placebo: OD. N=87. Additional medication allowed included. chloral hydrate, anticholinergics, benzotropine mesylate, benzhexol hydrochloride, oral diphenhydramine, lorazepam, zolpidemor flunitrazepam.
Outcomes	Leaving the study early. Unable to use ‐ Mental state: PANSS, BPRS, Calgary Depression Scale (no SD). Global state: CGI (no usable data). Adverse events: SAS, BAS, (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

de Haan 2002.

Methods	Allocation: randomised ‐ no further details. Blindness: not reported. Duration: 6 weeks. Design: parallel groups.
Participants	Diagnosis: schizophrenia (DSM IV) or related disorders. History: recent onset, informed consent obtained. N=36*. Age: mean 22.4 years. Sex: 92 M, 21 F. Setting: hospitalised. Inclusion criteria: not reported. Excluded: neurological or endocrine diseases and mental retardation.
Interventions	1. Olanzapine: mean dose 14.2 mg/frequency not reported. N=19. 2. Risperidone: mean dose 4.2 mg/frequency not reported. N=17.
Outcomes	Behaviour: Yale Brown Obsessive Compulsive Scale.
Notes	*113 randomised, but only 36 were not receiving olanzapine or risperidone before the study, or were switched from olanzapine to risperidone or vise versa.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

de Hann 2003.

Methods	Allocation: randomised, block design. Blindness: double. Duration: 6 weeks. Design: parallel group.
Participants	Diagnosis: schizophrenia (DSM IV). History: recent onset schizophrenia, informed consent obtained. N=24. Age: 17‐28 years. Sex: male and female. Setting: hopitalised and community. Inclusion criteria: not reported. Excluded: neurological or endocrine disease, mental retardation, use of adjunctive medication such as mood stabilizers or antidepressants, history of clozapine treatment, or history of unresponsiveness to haloperidol or olanzapine and intramuscular antipsychotic treatment within a year.
Interventions	1. Olanzapine: dose 7.5 mg/day. N=12. 2. Haloperidol: dose 2.5 mg/day. N=12. Additional medication, benzodiazepine oxazepine oxazepam.
Outcomes	Leaving the study early. Mental state: PANSS, MADRS. Subjective Wellbeing Under Neuroleptic Scale. Unable to use ‐ Glaobal state: CGI (no usable data). Adverse events: SAS, BAS (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Esel 2001.

Methods	Allocation: randomised ‐ no further details. Blindness: not reported. Duration: 6 weeks. Design: parallel group.
Participants	Diagnosis: schizophrenia (DSM IV). History: not reported. N=29. Age: range 17‐28, mean 29.3 years. Sex: male. Setting: hospital. Inclusion criteria: not reported.
Interventions	1. Olanzapine: dose 10 mg/day. N=15. 2. Haloperidol: dose 10 mg/day. N=14.
Outcomes	Leaving the study early. Mental state: BPRS Global state: CGI. Adverse events: ESRS.
Notes	Author contacted who stated the study was randomised.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

He 2003.

Methods	Allocation: randomised ‐ no further details. Blindness: open label. Duration: 8 weeks. Design: parallel groups.
Participants	Diagnosis: schizophrenia (CCMD‐2‐R). History: mean length of illness, olanzapine 1.5 years; chlorpromazine 2.2 years. N=80. Age: 18‐60. Sex: male and female. Setting: hospitalised. Inclusion criteria: not reported.
Interventions	1. Olanzapine: dose 5‐30 mg/day. N=40. 2. Chlorpromazine: dose 100‐600 mg/day. N=40.
Outcomes	Leaving the study early. Unable to use ‐ Mental state: PANSS (data unclear). Adverse events: TESS (data unclear).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

HGBJ (Finland).

Methods	Allocation: randomised, computer‐generated, blocks for each investigator 1:1, concealed from investigators. Blindness: double, medication kits issued. Duration: 26 weeks (preceded by screening phase, unspecified). Design: multicentre, Finland.
Participants	Diagnosis: schizophrenia (DSM‐IV). History: acute symptoms. N=46. Age: range 18‐70 years, mean 37.6. Sex: M 28, F 18. Setting: not reported Inclusion criteria: severity of illness at least moderately ill (CGI 4).
Interventions	1. Olanzapine: dose range 5‐20 mg/day. N=23. 2. Perphenazine: dose range 8‐32 mg/day. N=23.
Outcomes	Leaving study early. Adverse events: COSTART list, weight change.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

HGBL 1997.

Methods	Allocation: randomised, computer‐generated, blocks, 1:1 for each investigator, concealed from investigators. Blindness: double, medication kits issued. Duration: 4 weeks (preceded by placebo lead‐in of 4‐7 days. Design: parallel groups.
Participants	Diagnosis: schizophrenia (DSM‐IV). History: not reported. N=30. Age: not reported. Sex: not reported. Setting: inpatients. Inclusion criteria: not reported.
Interventions	1. Olanzapine: dose range 5‐20 mg/day. N=15. 2. Flupentixol: dose range 5‐20 mg/day. N=15.
Outcomes	Leaving study early. Adverse events: COSTART list, weight change. Unable to use ‐ Adverse events: extrapyramidal (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

HGCJ (Hong Kong).

Methods	Allocation: randomised, computer‐generated, blocks, 1:1 for each investigator, concealed from investigators. Blindness: double, medication kits issued. Duration: 14 weeks. Design: parallel groups.
Participants	Diagnosis: schizophrenia, schiziphreniform disorder, schizoaffective disorder (DSM‐IV). History: not reported. N=31. Age: not reported. Sex: not reported. Setting: inpatients and outpatients. Inclusion criteria: screening used, no further details.
Interventions	1. Olanzapine: dose range 5‐20 mg/day. N=17. 2. Haloperidol: dose range 5‐20 mg/day. N=14.
Outcomes	Leaving study early. Global state: CGI‐S. Mental state: BPRS, MADRS, PANSS. Adverse events: COSTART list, weight change. Unable to use ‐ Adverse events: extrapyramidal (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

HGCQ (Turkey) 2000.

Methods	Allocation: randomised, blocks, computer‐generated, 2:1 for each investigator, concealed from investigators. Blindness: double, medication kits issued. Duration: 6 weeks. Design: 2 centres, Turkey.
Participants	Diagnosis: schizophrenia (DSM‐IV). History: informed consent obtained. N=30. Age: range 18‐47 years, mean 32.8. Sex: M 17, F 13. Setting: inpatients and outpatients. Inclusion criteria: CGI severity at least 4. Excluded: pregnancy, seriously unstable illness, including hepatic, renal, gastroenterologic, respiratory, cardiovascular, endocrinologic, neurologic or hematologic disease; glaucoma, uncontrolled thyroid disease, myasthenia gravis, urinary retention, seizures, leucopenia.
Interventions	1. Olanzapine: dose range 5‐20 mg/day. N=20. 2. Chlorpromazine: dose range 200‐800 mg/day. N=10.
Outcomes	Leaving the study early. Mental state: BPRS, PANSS. Global state: CGI. Adverse events: ESRS, COSTART list, weight change. Quality of Life: VPS, LSEQ. Unable to use ‐ Adverse events: UKU, weight change (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

HGCU (Taiwan) 1998.

Methods	Allocation: randomised, blocks, computer‐generated, 1:1 for each investigator, concealed from investigators. Blindness: double, medication kits issued. Duration: 14 weeks. Design: multicentre, 4 sites.
Participants	Diagnosis: schizophrenia, schiziphreniform disorder, schizoaffective disorder (DSM‐IV). History: not reported. N=54. Age: not reported. Sex: not reported. Setting: inpatients and outpatients. Inclusion criteria: not reported.
Interventions	1. Olanzapine: dose range 5‐20 mg/day. N=26. 2. Haloperidol: dose range 5‐20 mg/day. N=28.
Outcomes	Leaving study early. Global state: CGI‐S. Mental state: BPRS, MADRS, PANSS. Adverse events: COSTART list, weight change. Unable to use ‐ Adverse events: extrapyramidal (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

HGDV (Morocco) 1999.

Methods	Allocation: randomised, computer‐generated, blocks for each investigator, 2:1, olanzapine to chlorpromazine. Blindness: open‐label, medication kits issued. Duration: 6 weeks (preceded by washout phase; extension for responders). Design: single centre Morocco.
Participants	Diagnosis: schizophrenia (DSM‐IV). History: previously hospitalised (mean ˜ 1.5 times), informed consent obtained. N=40. Age: 18‐47. Sex: 6 male, 33 female. Setting: inpatients and outpatients. Inclusion criteria: initial CGI severity score of 4. Excluded: pregnancy, seriously unstable illness, including hepatic, renal, gastroenterologic, respiratory, cardiovascular, endocrinologic, neurologic or hematologic disease; glaucoma, uncontrolled thyroid disease, myasthenia gravis, urinary retention, seizures, leucopenia.
Interventions	1. Olanzapine: dose range 5‐20 mg/day. N=27. 2. Chlorpromazine: dose range 200‐800 mg/day. N=13.
Outcomes	Leaving study early. Mental state: BPRS, PANSS, MADRS, HAM‐S. Global state: CGI. Adverse events: EPRS, COSTART list, weight gain.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

HGFH (Korea) 1998.

Methods	Allocation: randomised, computer‐generated, blocks for each investigator, 1:1. Blindness: open label. Duration: 6 weeks (preceded by washout phase, unspecified; extension for responders). Design: single centre, Korea.
Participants	Diagnosis: schizophrenia, schizophreniform and schizoaffective disorder (DSM‐IV). History: not reported. N=104. Age: not reported. Sex: not reported. Setting: inpatients & outpatients. Inclusion criteria: not reported.
Interventions	1. Olanzapine: dose range 5‐20 mg/day. N=53. 2. Haloperidol: dose range 1.5‐20 mg/day. N=51.
Outcomes	Leaving study early. Adverse events: COSTART list, weight gain.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Ishigooka 2001.

Methods	Allocation: randomised ‐ no further details. Blindness: double. Duration: 8 weeks. Design: multicentre, 67 investigative institutions in Japan.
Participants	Diagnosis: schizophrenia (ICD 10). History: informed consent obtained. N=182. Age: 18‐65 years. Sex: male and female. Setting: inpatients and outpatients. Inclusion criteria: not reported. Excluded: previous olanzapine treatment or haloperidol within previous 2 weeks, serious illness, seizures, pregnancy, leukopenia, granulocytopenia, Parkinson's disease.
Interventions	1. Olanzapine: dose range 5‐15 mg/day. N=93. 2. Haloperidol: dose range 4‐12 mg/day. N=89.
Outcomes	Leaving the study early. Mental state: BPRS, PANSS, PANSS‐N, PANSS‐P. Adverse events: ADR, DIEPSS, treatment emergent side‐effects. Death. Unable to use ‐ Global state: FGIR (scale not validated). Physiological measurements: (no SD).
Notes	Contacted author for further information on 20/04/04
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Jakovljevic1999 HGCH.

Methods	Allocation: randomised, blocks 1:1, computer‐generated, concealed from investigators. Blindness: double, medication kit issued. Duration: 6 weeks followed by extension for 22 weeks (preceded by washout period of 2‐9 days). Design: multicentre 3 sites, Croatia.
Participants	Diagnosis: schizophrenia, no further details. History: informed consent obtained. N=60. Age: between 18 and 65 years. Sex: male and female. Setting: hospitalised. Inclusion criteria: minimum PANSS score of at least 42 and at least 4 on the CGI severity scale. Excluded: serious unstable illness, history of intolerance to olanzapine, substance dependence within last 30 days, risk of suicide, pregnancy, breast feeding.
Interventions	1. Olanzapine: dose range 5‐20 mg/day. N=30. 2. Fluphenazine: dose range 6‐21 mg/day. N=30.
Outcomes	Leaving study early. Global state: CGI‐S. Mental state: BPRS, PANSS. Adverse events: treatment emergent side‐effects: COSTART list, weight change. Quality of life: DAI, LSEQ. Unable to use ‐ Adverse events: extrapyramidal (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Jeste 2003.

Methods	Allocation: randomised, using a minimization technique to ensure no imbalances occurred in either group based on scores from the ESRS, PANSS and MMSE scales. Blindness: double. Duration: 8 weeks, 1 week washout. Design: mulitcentre, 32 sites, USA, Poland, Israel, Norway, Netherlands and Austria. Informed consent obtained,
Participants	Diagnosis: schizophrenia or schizoaffective disorder (DSM‐IV). History: mean duration of illness 36 years. N=175. Age: over 60 years (mean over 71 years). Sex: M 62, F 113. Setting: inpatients and outpatients. Inclusion criteria: PANSS baseline score between 50‐120. Excluded: DSM‐IV diagnosis of substance abuse in previous 3 months, major depressive episode as defined by DSM‐IV criteria within 6 months before screening, MMSE score of less than 18, CNS disorders, QTc >500msec, unresponsive to risperidone or olanzapine in the past, any acute unstable, significant, or untreated medical illnesses and renal, hepatic, or gastrointestinal disease that could potentially interfere with the absorption, excretion or metabolism of trial medications.
Interventions	1. Olanzapine: dose lowest dose not stated, range ˜5‐20 mg/day. N= 88. 2. Risperidone: dose range 1‐3 mg/day.N= 87. Flexible doses used to acheive maximum efficacy with minimum adverse events. Additional medication, Lorazepam max 4 mg/day.
Outcomes	Death. Leaving the study early. Global state: CGI. Mental state: PANSS. Adverse effects: ESRS. Other adverse events. Use of antiparkinsonian medication. Unable to use ‐ Mental state: HAM‐D (no usable data). Adverse events: weight change: (no usable data).
Notes	Concomitant medications disallowed during the study period included antidepressants, anticonvulsants, mood stabilizers other antipsychotics, sedatives, anxiolytics, and other psychoactive drugs (except lorazepam during washout period, for agitation).
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Jones 1998 (P022).

Methods	Allocation: random, computer‐generated, blocks by investigator, concealed from investigator. Blindness: double. Duration: 54 weeks (preceded by 1 month stabilization phase followed by 1 week washout, screening period). Design: multicentre, Canada.
Participants	Diagnosis: schizophrenia. History: 'early phase', first 5 years of illness. N=65. Age: mean 28.9 years. Sex: male and female. Setting: outpatients. Inclusion criteria: not reported. Excluded: if PANSS>90.
Interventions	1. Olanzapine: dose range 5‐20 mg/day. N=21. 2. Risperidone: dose range 4‐10 mg/day. N=21. 3. Haloperidol: dose range 5‐20 mg/day. N=23.
Outcomes	Leaving the study early. Mental state: PANSS. Adverse events: treatment emergent side‐effects, ESRS, COSTART list, weight change. Unable to use ‐ Cost: (no data). Cognitive function: GCIS, neuropsychological test battery (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Kelly 2003.

Methods	Allocation: randomised‐ no further details. Blindness: double. Duration: 8 weeks. Design: crossover.
Participants	Diagnosis: treatment resistant schizophrenia (DSM IV). History: infomed consent obtained. N=13. Age: mean 37 years. Sex: not reported. Setting: hospital. Inclusion criteria: treament resistant schizophrenia as evidenced by persistent positive psychotic symptoms: item score >/=4 (moderate) on at least two of four positive symptom items on BPRS, the current presence of at least moderately severe illness as rated by the total BPRS score (score>/=45 on the 18‐item scale) and a score >/=4 (moderate) on the CGI, two failed historical trials of antipsychotics of at least 6‐weeks duration at doses of at least 600mg/day chlorpromazine equivalents and no stable period of good social and/or occupational functioning within the last 5 years. Excluded: previously failed clozapine treatment.
Interventions	1. Olanzapine: dose range 50 mg/day. N=8. 2. Clozapine: dose range 450 mg/day. N=5. Additional medication: lorazepam and benztropine mesylate.
Outcomes	Leaving the study early. Unable to use ‐ Global state: CGI (no usable data prior to crossover). Mental state: BPRS (no usable data prior to crossover).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Kern 2001.

Methods	Allocation: randomised ‐ no further details. Blindness: open‐label. Duration: 26 weeks. Design: multicentre, parallel group.
Participants	Diagnosis: schizophrenia or schizoaffective disorder. History: chronic stable, not hospitalised for > 2 months prior to randomisation, previously on stable dose of antipsychotic for > 2 months. N=255. Age: mean ˜40, range 18‐65 years. Sex: M 164, F 92. Setting: outpatient. Inclusion criteria: not reported.
Interventions	1. Olanzapine: dose range 15 mg/day, after 10 mg/day 1‐7. N=127. 2. Aripiprazole: dose range 30 mg/day. N=128.
Outcomes	Adverse effects: spontaneously reported adverse effects occurring in >10%, clinically significant weight gain. Unable to use ‐ Adverse effects: weight change (no SD). Cognitive functioning: California Verbal Learning test, Benton visual retention test, Winsconsin card sorting test, Trail making A and B, Continuous perfomance test, Verbal Fluency, Letter‐number sequencing from the WAIS III, Grooved pegboard test (no usable data/no SDs).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Lecrubier 1999.

Methods	Allocation: randomised, computer‐generated, blocks for each investigator, 2:2:2:1, concealed from investigator. Blindness: double, medication kits issued. Duration: 6 months. Design: 4 treatment groups.
Participants	Diagnosis: "schizophrenic patients with primarily negative symptoms" ‐ no further details. History: primarily negative symptoms. N=244. Age: not reported. Sex: not reported. Setting: not reported. Inclusion criteria: minimum SANS summary score of 10 (excluding attention subscore) & no score > 4 on hallucination & delusion items of PANSS (normalized, scored 0‐6).
Interventions	1. Olanzapine: dose 5 mg/day. N=70. 2. Olanzapine: dose 20 mg/day. N=70. 3. Amisulpride: dose 150 mg/day. N=70. 4. Placebo: N=34.
Outcomes	Leaving study early. Mental state: SANS. Adverse events: COSTART list, weight change. Quality of life: QLS. Unable to use ‐ Mental state: PANSS (no usable data). Adverse events: extrapyramidal (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Lieberman 2003 (HGDH.

Methods	Allocation: randomised ‐ no further details. Blindness: double. Duration. 12 weeks (acute phase). Design: multicentre, 14 sites, North America and Europe.
Participants	Diagnosis: schizophrenia (DSM IV). History: first episode schizophrenia, informed consent obtained. N=263. Age: 16‐40. Sex: M 215, F 48. Setting: not clear. Inclusion criteria: scored =/> 4 on at least two positive and negative syndrome scale and CGI =/> 4. Excluded: 1 month of substance abuse; previoulsy revceived antipsychotics for more than 16 cumulative weeks, had been treated with clozapine at anytime, or had an depot neuroleptic within three dosing intervals before study entry; pregnancy/nursing; serious unstable illness.
Interventions	1. Olanzapine: dose range 5‐20 mg/day. N=132. 2. Haloperidol: dose range 2‐20 mg/day. N=131. Additional medication, chloral hydrate, lorazepam, diazepam; benztropine, biperiden, propranolol, procyclidine.
Outcomes	Leaving the study early. Global state: CGI. Mental state: PANSS, MADRS. Adverse events: requiring anticholinergics, benzodiazepines, propranolol, treatment emergent parkinsonism, akathisia, weight gain.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Lima 2003 (HGHS).

Methods	Allocation: randomised ‐ no further details. Blindness: open label. Duration: 9 months. Design: multicentre, Brazil.
Participants	Diagnosis: schizophrenia (DSM IV), or related disorders. History: participants experiencing an acute psychiatric admission or recently discharded, informed consent obtained. N=197* Age: mean 33.8. Sex: M 154, F 43. Setting: not reported. Inclusion criteria: not reported.
Interventions	1. Olanzapine: dose as clinically indicated, no further details. N=104. 2. Conventional antipsychotics (mainly haloperidol or chlorpromazine, no further details. N=93.
Outcomes	Leaving the study early. Global state: CGI. Mental state: PANSS. Adverse events. Unable to use ‐ Additional medication (data unclear).
Notes	*Abstract reports 190 participants.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Littrell 1999.

Methods	Allocation: randomised ‐ no further details. Blindness: open label. Duration: 1 year. Design: parallel groups.
Participants	Diagnosis: schizophrenia. History: chronic illness. N=24. Age: not reported. Sex: not reported. Setting: outpatients. Inclusion criteria: not reported.
Interventions	1. Olanzapine: dose 19.2 mg/day. N=12. 2. Risperidone: dose 5.2 mg/day. N=12.
Outcomes	Leaving the study early. Mental state: PANSS, PANSS‐N, PANSS‐P. Unable to use ‐ Adverse events: SAS (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Loza 1999 (HGDT).

Methods	Allocation: randomised, computer‐generated, blocks for each investigator, 2:1, olanzapine to chlorpromazine, concealed from investigators. Blindness: double, medication kits issued. Duration: 6 weeks (preceded by washout phase of 2‐9 days; extension for responders). Design: multicentre: 2 sites, Egypt.
Participants	Diagnosis: schizophrenia (DSM‐IV). History: informed consent obtained. N=41. Age: range 17‐47 years, mean 32.3. Sex: M 33, F 8. Setting: inpatients & outpatients. Inclusion criteria: initial score of at least 4 on the CGI severity scale. Excluded: pregnancy, seriously unstable illness, including hepatic, renal, gastroenterologic, respiratory, cardiovascular, endocrinologic, neurologic or hematologic disease; glaucoma, uncontrolled thyroid disease, myasthenia gravis, urinary retention, seizures, leucopenia.
Interventions	1. Olanzapine: dose range 5‐20 mg/day. N=27. 2. Chlorpromazine: dose range 200‐800 mg/day. N=14.
Outcomes	Leaving study early. Mental state: BPRS, PANSS. Global state: CGI‐S. Quality of Life Scale: QoL. Adverse events: ESRS, COSTART list, weight change. Unable to use ‐ Adverse events: UKU (no data). Hospital status: (no data). Laboratory tests & physiological measures: (no data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Malyarov 1999.

Methods	Allocation: randomised ‐ no further details. Blindness: not reported. Duration: 6 months. Design: three treatment groups.
Participants	Diagnosis: schizophrenia (ICD 10). History: acute psychotic state. N=43. Age: mean 24.5 years. Sex: M 28, F 15. Setting: inpatients. Inclusion criteria: not reported.
Interventions	1. Olanzapine: dose range 5‐15 mg/day. N=15. 2. Haloperidol: dose range 5‐20 mg/day. N=18. 3. Risperidone: dose range 3‐6 mg/day. N=10.
Outcomes	Leaving the study early. Unable to use ‐ Mental state: PANSS (no SD). Global state: GAF (no SD).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Martin 2002.

Methods	Allocation: randomised ‐ no further details. Blindness: double. Duration: 2 months. Design: multicentre, 76 sites, Belgium, Switzerland, Denmark, France, Great Britain, Czech Republic, Tunisia, Hungary, Morocco and Portugal.
Participants	Diagnosis: schizophrenia (DSM‐IV). History: informed consent obtained. N=377. Age: mean 37.8 years. Sex: M/F (˜2:1) ratio. Setting: inpatients and outpatients. Inclusion criteria: 18‐65 years with BPRS score minimum of 36 (extracted from PANSS) and a positive PANSS rating higher than the negative PANSS rating. Excluded: patients whose BPRS score improved by more than 40% between the screening and baseline visit; pregnant and lactating women and inadequate use of contraception in fertile women.
Interventions	1. Olanzapine: mean dose 12 mg/day. N=188. 2. Amisulpride: mean dose 489 mg/day. N=189.
Outcomes	Leaving the study early. Global state: CGI. Mental state: BPRS , PANSS, MADRS. Adverse events: SAS. Unable to use ‐ Weight change: (no SD). Mental state: PANSS change score ‐ endpoint data takes precendence over change.
Notes	BPRS Data calculated from graphic figure 1.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

McQuade 2003.

Methods	Allocation: randomised ‐ no further details. Blindness: double. Duration: 26 weeks. Design: parallel, multi‐centre.
Participants	Diagnosis: schizophrenia. History: acute relapse, mean age at first episode ˜25 years, mean time since current acute episode began ˜21 days. N=317. Age: mean ˜38 years. Sex: M 229, F 88. Setting: not reported. Inclusion criteria: not reported.
Interventions	1. Olanzapine: dose range 10‐20 mg/day. N=161. 2. Aripiprazole: dose range 15‐30 mg/day. N=156.
Outcomes	Leaving the study early. Unable to use ‐ Global state: CGI (no usable data). Mental state: PANSS (no SDs). Adverse effects: mean change in weight (no SDs), clinically significant weight gain, serum prolactin, EPS, plasma lipids outside normal limits (no usable data).
Notes	Greater than 60% discontinuation rate.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Meltzer (InterSept).

Methods	Allocation: randomised, block allocated 1:1 ratio. Blindness: open label with raters blinded. Duration: 2 years. Informed consent obtained. Design: suicide prevention, parallel groups, multicentre, 67 sites, 11 countries USA, Canada, France, Italy, the United Kingdom, the Czech Republic, Hungary, Croatia, South Africa, Argentina and Chile.
Participants	Diagnosis: schizophrenia and schizoaffective disorder (DSM IV). N=980*. Age: range 18‐66 years, mean 37. Sex: M 602, F 378. Setting: hospital and community. Inclusion criteria: attempted suicide or were hospitalised to prevent sucide within the three years before their baseline evaluation or demonstrated moderate to severe suicidal ideation with a depressive or delusinal component within 1 week before their baseline evaluation were considered at high risk of suicidality.
Interventions	1. Olanzapine: dose 16.4 mg/day. N=477. 2. Clozapine: dose 274 mg/day. N=479.
Outcomes	Death. Leaving the study early. Adverse events. Rescue interventions including hospitalisation, increased level of surveillance and addition of an antidepressant. Unable to use ‐ Global state: CGI‐SS (no usable data). Mental state: PANSS, Calgary Depression Scale (no usable data).
Notes	*24 never treated for administrative reasons, unknown from which arm.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Naber 2001 (HGBF).

Methods	Allocation: randomised 1:1 block allocation. Blindness: double. Duration: 26 weeks. Design: multicentre, parallel group.
Participants	Diagnosis: schizophrenia (DSM IV). History: all participants had either failed to respond to at least one antipsychotic other than clozapine and olanzapine or had experienced intolerable side effects during these prior antipsychotic treatments. N=114. Age: range 18‐59, mean 34. Sex: M 69, F 45. Setting: not reported. Inclusion criteria: minimum BPRS score of 24. Excluded: prior olanzapine treatment at any time or prior clozapine treatment within the last 3 months.
Interventions	1. Olanzapine: dose range 5‐25 mg/day. N=57. 2. Clozapine: dose range 100‐400 mg/day. N=57.
Outcomes	Leaving the study early. Unable to use ‐ (>50% drop out). Global state: CGI. Mental state: PANSS, BPRS. Quality of life: MLDL, SWN. Adverse events: SAS, side effects.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Namjoshi 2002.

Methods	Allocation: randomised ‐ no further details. Blindness: not described. Duration: 1 year. Design: three treatment groups; six participants were randomised, though not reported on in the study reports; we have made an assumption that two were missing from each group and added them back in.
Participants	Diagnosis: schizophrenia. History: not reported. N=364. Age: mean 39.4 years. Sex: male and female. Setting: outpatients. Inclusion criteria: not reported.
Interventions	1. Olanzapine: dose range 5‐20 mg/day. N=142. 2. Risperidone: dose range 2‐10 mg/day. N=136. 3. Haloperidol: dose range 2‐9 mg/day. N=86.
Outcomes	Hospital admission. Unable to use ‐ Mental state: PANSS (no usable data). Cost effectiveness: (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Ritchie 2003.

Methods	Allocation: randomised, blocks of six by study site. Blindness: allocation concealed from local investigators, open label; investigators trained in use of scales using videos. Duration: first arm approx. four weeks (unclear). Design: multicentre, 8 sites Australia, crossover. Informed consent obtained.
Participants	Diagnosis: schizophrenia. N=66. Age: mean 69.6 years. Sex: M 19, 47 F. Setting: not reported. Inclusion criteria: people >60 years who were taking conventional antipsychotics. Excluded: idiopathic parkinson's disease, epilepsy or previous treatment failure with olanzapine or risperidone.
Interventions	1. Olanzapine: dose ˜9.9 mg/day. N=34. 2. Risperidone: dose ˜1.7 mg/day. N=32.
Outcomes	Leaving the study early. Global state: CGI. Mental state: PANSS. Adverse events. Additional medication(data unclear).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Rosenheck 2003(HGFI).

Methods	Allocation: randomised, random sequence numbers. Blindness: double, medication kits issued. Duration: 12 months. Design: parallel group.
Participants	Diagnosis: schizophrenia (DSM IV). History: informed consent obtained. N=309. Age: mean age 46.5 years. Sex: M 153, F 144. Setting: hospital and community. Inclusion criteria: BPRS score >36. Excluded: previous participation in olanzapine research.
Interventions	1. Olanzapine: dose range 5‐20 mg/day. N=159. 2. Haloperidol: dose range 5‐20 mg/day. N=150.
Outcomes	Leaving the study early. Unable to use ‐ (>50% lost to follow in both arms). Global state: CGI. Mental state: PANSS. Quality of Life. Adverse events: SAS, AIMS, BAS, check lists. Assessment of Health Care Costs.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Simpson 2004.

Methods	Allocation: randomised ‐ no further details. Blindness: double. Duration: 6 weeks. Design: multicentre. Informed consent obtained.
Participants	Diagnosis: schizophrenia (DSM IV). N=269. Age: 18‐55. Sex: M 176, F 93. Setting: inpatients. Inclusion criteria: persistant psychotic symptoms for a week before hospital admission; CGI score =/>4 and score =/>4 on PANSS positive scale; normal laboratory test results and normal ECG and negative urine drug test at entry.
Interventions	1. Olanzapine: dose range 5‐15 mg/day. N=133. 2. Ziprasidone: dose range 40‐80 mg/bid. N=136. Additional medication, lorazepam permitted for control of agiitation or insomnia and benzotropine for extrapyramidal symptoms.
Outcomes	Leaving the study early. Global state: CGI. Mental state: BPRS. Additional medication (benzodiazepines, anticholinergics). Adverse events. Unable to use ‐ Mental state: PANSS, Calgary Depression Scale (no usable data). Adverse events: ESRS, BAS, AIMS (no usable data). Weight changes (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Svestka 2003.

Methods	Allocation: randomised ‐ no further details. Blindness: double. Duration: 6 weeks. Design: parallel group.
Participants	Diagnosis: schizophrenia (ICD 10). History: first episode. N=42. Age: not reported. Sex: female. Setting: not reported. Inclusion criteria: not reported.
Interventions	1. Olanzapine: dose range 10‐20 mg/day. N=21. 2. Risperidone: dose range 4‐6 mg/day. N=21.
Outcomes	Mental state: PANSS. Adverse events: EPS, rhinitis, use of antiparkinsonics, hyperprolactinaemia. Unable to use ‐ Global state: CGI (no usable data). Adverse events: EPS (no usable data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Thomas 1998 (HGBU).

Methods	Allocation: randomised, computer‐generated, blocks for each investigator, 1:1, concealed from investigator. Blindness: double, medication kits issued. Duration: 30 weeks. Design: multicentre, Australia (6 sites), New Zealand (2 sites).
Participants	Diagnosis: schizophrenia, schizophreniform and schizoaffective disorder (DSM‐IV). History: informed consent obtained. N=65. Age: 18 years or over. Sex: M 38, F 27. Setting: patients could begin the study as either in or out patients. Inclusion criteria: scores greater than 36 (on items scored 1 to 7) on the BPRS; excluded pregnant females and patients who did not respond to >400 mg/day of clozapine taken for a minimum of 6 weeks. Excluded: pregnacy, failure to show a response to clozapine at a dose of >400mg/day for at least 6 weeks.
Interventions	1. Olanzapine: dose range 10‐20 mg/day. N=32. 2. Risperidone: dose 4‐8 mg/day. N=33.
Outcomes	Leaving study early. Global state: CGI‐S. Mental state: BPRS, PANSS. Adverse events: COSTART list, weight change. Quality of life: QLS. Unable to use ‐ Quality of life: SF‐36 (no total score).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Tollefson 1997 Conti.

Methods
Participants
Interventions
Outcomes
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Tollefson 1997(HGAJ).

Methods	Allocation: randomised, computer‐generated, blocks,ratio 2:1, concealed from investigators. Blindness: double. Duration: 6 weeks (preceded by a screening phase of 2‐9 days, maintenance phase of 46/52 for responders), and a 3 year, double blind extension phase. Design: multicentre, 174 sites in 17 countries.
Participants	Diagnosis: schizophrenia, schizophreniform or schizoaffective disorder (DSM‐III‐R). History: informed consent obtained. N=1996. Age: mean ˜ 39 years. Sex: male and female. Setting: not reported. Inclusion criteria: >18 BPRS and/or intolerant of current antipsychotics.
Interventions	1. Olanzapine: dose range 5‐20 mg/day. N=1336. 2. Haloperidol: dose range 5‐20 mg/day. N=660. Benztropine & benzodiazepine as required.
Outcomes	Leaving study early. Global state: (CGI). Mental state: (BPRS*, needing additional benzodiazepines, MADRS, PANSS). Death.** Adverse effects: (AMDP, Barnes Akathisia Scale, requiring benztropine, SAS). Other adverse events: (COSTART list, weight change). Unable to use ‐ Hospital status: (no data). Laboratory tests & physiological measures: (no data). Weight loss: (undefined subgroup reported from 3 year extension phase).
Notes	* BPRS (scored 0‐6) extracted from PANSS ‐ no reference given for validity of procedure. * A priori efficacy response was 40% improvement in BPRS score and three weeks in study. **Deaths reported during the study or within 30 days of discontinuation.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Tollefson 1999(HGDY).

Methods	Allocation: randomly assigned ‐ no further details. Blindness: double. Duration: 3‐5 days (preceeded by 2‐12 days clozapine discontinuation). Design: multicentre, 11 sites in USA.
Participants	Diagnosis: schizophrenia (DSM‐IV). History: informed consent obtained. N=106. Age: mean 38.8 years. Sex: M 75, F 31. Setting: inpatients and outpatients. Inclusion criteria: received clozapine maximum dose 900mg/day for minimum of 4 weeks prior to study entrance. Excluded: substance use or substance dependence, risk of suicide, unstable medical condition or previously treated with olanzapine.
Interventions	1. Olanzapine: dose range 10 mg/day. N=53. 2. Placebo. N=53.
Outcomes	Leaving the study early. Relapse. Global state: (CGI). Mental state: (PANSS, MADRS, MMSE).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Tran 1997 (HGBG).

Methods	Allocation: randomised, computer‐generated, blocks, concealed from investigators. Blindness: double, medication kits issued. Duration: 28 weeks (preceeded by 2‐9 day washout). Investigators: trained on PANSS. Design: multicentre, 9 countries.
Participants	Diagnosis: schizophrenia, schizophreniform disorder or schizoaffective disorder (DSM‐IV). History: informed consent obtained. N=339. Age: range 16‐65 years. Sex: M 220, F 119. Setting: inpatients or outpatients. Inclusion criteria: not reported. Excluded: "treatment resistant".
Interventions	1. Olanzapine: dose range 10‐20 mg/day, mean modal dose 17.2, +/‐ 3.6 mg/day. N=172. 2. Risperidone: dose 4‐12 mg/day, mean modal dose 7.2 +/‐2.7 mg/day. N=167. Benzodiazepines, chloral hydrate, benztropine mesylate, biperiden as required.
Outcomes	Leaving study early. Global state: (CGI). Mental state: (BPRS*, PANSS, SANS). Adverse effects: (AIMS, Barnes Akathisia Scale, requiring benztropine or biperiden, SAS). Other adverse effects: (AMDP, COSTART list, neutrophil counts, prolactin, weight change). Hospital status. Quality of life: (QOL). Unable to use ‐ Laboratory tests & physiological measures: (no data). Economic burden: (no SD).
Notes	* BPRS (scored 0‐6) extracted from PANSS ‐ no reference given for validity of procedure. PANSS response rates reported >/= 20%,>/= 30%,>/= 40%>/= 50%.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Wang 2002.

Methods	Allocation: randomised ‐ no further details. Blindness: double. Duration: 8 weeks. Design: parallel groups.
Participants	Diagnosis: schizophrenia. History: not reported. N=63. Age: 18‐55. Sex: male and female. Setting:hospital and community. Inclusion criteria: not reported.
Interventions	1. Olanzapine: dose 5‐20 mg/day. N=30. 2. Clozapine: dose: 25‐400 mg/day. N=33.
Outcomes	Leaving the study early. Mental state: BPRS. Adverse events.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Wang 2003.

Methods	Allocation: randomised ‐ no further details. Blindness: double. Duration: 6 weeks. Design: parallel groups.
Participants	Diagnosis: schizophrenia (CLMD‐3). History: not reported. N=122. Age: 18‐60. Sex: male and female. Setting: hospitalised. Inclusion criteria: not reported.
Interventions	1.Olanzapine: dose 5‐20 mg/day. N=61. 2. Clozapine: dose 50‐450 mg/day. N=61.
Outcomes	Leaving the study early. Mental state: PANSS. Adverse events.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Diagnostic tool 
 DSM III‐R and DSM‐IV ‐ Diagnostic Statistical Manual version 3 Revised and version 4. 
 ICD 10 ‐ The International Statistical Classification of Diseases and Related Health Problems. 
 BMI ‐ Body Mass Index.

Rating Scales:

Cognitive function ‐ 
 GCIS ‐ General Cognitive Index Score.

Behaviour ‐ 
 YBOCS ‐ Yale Brown Obsessive Compulsive Scale

Global rating scales: 
 CGI ‐ Clinical Global Impressions. 
 CGI‐S ‐ Clinical Global Impression‐Severity. 
 CGI‐I ‐ Clinical Global Impression‐Improvement. 
 PGI ‐ Patient Global Impression.

Mental state: 
 BPRS ‐ Brief Psychiatric Rating Scale. 
 HAM‐S ‐Hamilton Anxiety Scale. 
 MADRS ‐ Montgomery‐Asberg Depression Rating Scale. 
 MMSE ‐ Wiing Mini Mental State Examination. 
 PANSS ‐ Positive and Negative Syndrome Scale. 
 SANS ‐ Scale for the Assessment of Negative Symptoms.

Side effects: 
 ADR ‐ Adverse Drug Reaction. 
 AIMS ‐ Abnormal Involuntary Movement Scale. 
 AMDP‐5 ‐ Association for Methodology and Documentation in Psychiatry adverse event questionnaire. 
 BAS ‐ Barnes Akathisia Scale. 
 BMI ‐ Body mass index. 
 COSTART ‐ Coding Symbols and Thesaurs for Adverse Reaction Terms. 
 DIEPSS ‐ Drug‐Induced Extrapyramidal Symptoms Scale. 
 ESRS ‐ Extrapyramidal Syndrome Rating Scale. 
 HAS ‐ Hillside Akathisia Scale. 
 SAS ‐ Simpson‐Angus Index ‐ for neurological side effects. 
 UKU ‐ Udvalg for kliniske ndersogelser Side Effect Rating Scale ‐side effect rating scale.

Quality of Life: 
 QoL ‐ Quality of Life Scale. 
 Van Putten Scale. 
 Drug Attitude Inventory. 
 Leeds Sleep Evaluation Questionaire. 
 MLDL‐ Munich Life Dimension List 
 SWN ‐Subjective Well‐being List

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Addington 1998	Allocation: not randomised.
Allison 1998	Allocation: not randomised, meta‐analysis.
Apicella 2001	Allocation: not described.
Apiquian 2003	Allocation: not randomised.
Aquila 2000 (HGEC)	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus 'comparator antipsychotic' versus risperidone. Outcomes: no usable data.
Arango 2001	Allocation: randomised. Participants: people with schizophrenia. Intervention: olanzapine versus haloperidol. Outcomes: no usable data.
Awad 2002b	Allocation: randomised. Participants: people with schizophrenia. Intervention: participants switiched to ziprasidone.
Baker 2003	Allocation: randomised. Participants: people with schizophrenia. Intervention: olanzapine dosage study.
Beasley 1996c	Allocation: not randomised, meta‐analysis.
Beasley 2003 (HGGI)	Allocation: randomised. Participants: people with schizophrenia. Intervention: olanzapine versus clozapine. Outcomes: no usable data.
Beuzen 1996 (HGCE)	Allocation: randomised. Participants: healthy elderly volunteers.
Birkett 1994	Allocation: not randomised.
Breier 2003	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus risperidone versus haloperidol. Outcomes: no usable data.
Britto 2002	Allocation: randomised. Participants: people with schizophrenia. Intervention: olanzapine versus ziprasadone. Outcomes: no usable data.
Buchanan 2003	Allocation: unclear. Participants: people with schizophrenia. Intervention: olanzapine versus haloperidol. Outcomes: no usable data.
Burgoyne 1996	Allocation: not described.
Castilla 2002	Allocation: randomised. Participants: children with psychosis.
Clouth 1999	Allocation: not randomised, economic drug analysis.
Corya 2002 (HGGA)	Allocation: randomised. Participants: major depression with psychosis.
Cutler 2002	Allocation: randomised. Participants: people with schizophrenia. Intervention: olanzapine versus ziprasidone. Outcomes: no usable data.
Dalheim 1997	Allocation: randomised crossover study.
Davis 1998	Allocation: not randomised, meta‐analysis.
Devanand 1998	Allocation: randomised. Participants: people with dementia.
Docherty 2002	Allocation: randomised. Participants: people with schizophrenia. Intervention: olanzapine versus risperidone. Outcomes: no usable data.
Dolnak 2000	Allocation: randomised. Participants: people with schizophrenia. Intervention: olanzapine versus haloperidol. Outcomes: no usable data.
Dossenbach 1997	Allocation: not randomised.
Finzen 2002	Allocation: not randomised.
Fleming 1999	Allocation: unclear. Participants: people with schizophrenia. Interventions: olanzapine versus placebo. Outcomes: no usable data.
Gallinat 2001	Allocation: double blind. Participants: people with schizophrenia and healthy volunteers. Interventions: olanzapine or clozapine (schizophrenia group) versus healthy volunteers. Outcomes: data for treatment group (olanzapine or clozapine) not differentiated.
Gothelf 2003	Allocation: not randomised.
Hrdlicka 2001	Allocation: not randomised, retrospective analysis of patients records.
Irwin 2003	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus risperidone. Outcomes: no usable data.
Jarboe 2001	Allocation: not randomised.
Javitt 2001	Allocation: randomised. Participants: people with schizophrenia. Intervention: current antipsychotic medication and glycine versus current antipsychotic medication and placebo.
Jerrell 2002	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus risperisone versus haloperidol. Outcomes: no usable data.
Kalali 2000	Allocation: not randomised.
Kinon (HGFW)	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine switching study.
Kinon 2003	Allocation: randomised. Participants: people with schizophrenia. Interventions: current therapy or risperidone vesus olanzapine.
Knegtering 2000	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus risperidone. Outcomes: no usable data.
Kolff 2000	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus risperidone. Outcomes: no usable data.
Kucerova 2002	Allocation: not randomised, case reports.
Lahti 1999a	Allocation: not randomised.
Lahti 1999b	Allocation: not randomised.
Lilly (HGBM) 1999	Allocation: not randomised, open‐label, single‐arm study of olanzapine.
Lilly (HGCL) 1999	Allocation: not randomised, open‐label, single‐arm study of olanzapine.
Lilly (HGCV) 1999	Allocation: not randomised, open‐label, single‐arm study of olanzapine.
Lilly (HGDU) 1999	Allocation: not randomised, open‐label, single‐arm study of olanzapine.
Lilly (HGDZ) 1999	Allocation: not randomised, open‐label, single‐arm study of olanzapine.
Lilly (HGEP) 1999	Allocation: not randomised, open‐label, single‐arm study of olanzapine.
Lilly (HGFM) 1999	Allocation: not randomised, open‐label, single‐arm study of olanzapine.
Lindenmayer 1998	Allocation: not randomised, review.
Loza 2001	Allocation: randomised. Participants: people with schizophrenia. Interventions: several typical antipsychotics versus several atypical antipsychotics. Outcomes: no usable data, full publication pending in Psychiatricia, ISSN 1643/0956.
Mahmoud 1999	Allocation: randomised. Participants: not described.
Marder 1998	Allocation: not randomised, case series.
Mauri 2002	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus quetiapine. Outcomes: no usable data.
McElroy 1998	Allocation: not randomised, review.
McGurk 1998	Allocation: not described.
Meltzer 2003	Allocation: not randomised.
Mintzer 1998	Allocation: not randomised, review.
Mitchell 2003	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine dosage study, no comparator drug.
Mosolov 1998	Allocation: not randomised, review.
Mulqueen 2000	Allocation: not described.
Naber 1997	Allocation: not randomised.
Narendran 2003	Allocation: not randomised.
Oliemeulen 2000	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus clozapine. Outcomes: no usable data.
Ortega‐Solo (HGBS)	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus risperidone. Outcomes: no usable data.
Parellada 1999	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus haloperidol. Outcomes: no usable data.
Perro 1999	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus risperidone versus zotepin versus sertindole. Outcomes: no usable data.
Pivac 2002	Allocation: not randomised.
Ratakonda 1998	Allocation: not randomised.
Reus 1997	Allocation: not randomised, review.
Sacchetti 2003	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus risperidone versus quetiapine. Outcomes: no data available.
Sacristan (HGGS)	Allocation: not randomised.
Sangar 1998 (HGEH)	Allocation: unclear. Participants: those with rapid‐cycling bipolar disorder.
Sharma 2003	Allocation: unclear. Participants: people with schizophrenia. Interventions: olanzapine versus haloperidol. Outcomes: no data available.
Sheitman 1997	Allocation: not randomised, case series.
Sheitman 2000	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus risperidone. Outcomes: no usable data.
Sikich 2004	Allocation: randomised. Participants: people with schizophrenia (8‐19 years).
Smelson 2003	Allocation: randomised. Participants: people with schizophrenia and cocaine addiction. Interventions: olanzapine versus haloperidol. Outcomes: no usable data.
Smith 1998	Allocation: not randomised.
Smith 2001	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus haloperidol. Outcomes: no usable data.
Soutullo 1998	Allocation: not randomised, case series.
Stephenson 2000	Allocation: not randomised, review article.
Szafranski 1999	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus perphenazine. Outcomes: no usable data.
Tamura 1996 (F1D‐JE)	Allocation: not randomised, testing a questionnaire.
Tang 2003	Allocation: randomised. Participants: people with schizophrenia. Intervention: olanzapine versus risperidone. Outcomes: no usable data.
Tohen 2000b (HGGW)	Allocation: randomised. Participants: not described.
Tran 1998 (HGEJ)	Allocation: not randomised, case series.
Trandafir 1998	Allocation: randomised. Participants: people with drug‐induced psychosis.
Uzun 2002	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus ziprasidone. Outcomes: no usable data.
Volavka 2002	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus clozapine versus haloperidol versus risperidone. Outcomes: olanzapine arm added during study without equal randomisation.
Voruganti 2002	Allocation: naturalistic cohort study.
Woods 2002 (HGGF)	Allocation: randomised. Participants: people with suspected prodromal schizophrenia.
Woodward 2001	Allocation: not randomised.
Wright 2003 (HGAB)	Allocation: randomised. Participants: people with schizophrenia. Interventions: IM olanzapine, followed by oral olanzapine versus IM haloperidol followed by oral haloperidol.
Wudarsky 1999	Allocation: double blind and open. Participants: childhood onset schizophrenia.
Yagdiran 2000	Allocation: randomised. Participants: people with schizophrenia. Interventions: olanzapine versus risperidone versus amisulpride, clozapine. Outcomes: no usable data.
Zhang 1999	Allocation: not randomised, open‐label, single‐arm study of olanzapine.
Ziherl (HGCZ) 1999	Allocation: not randomised, open‐label, single‐arm study of olanzapine.

Characteristics of ongoing studies [ordered by study ID]

Centorrino 2001.

Trial name or title	A double‐blind, multi‐center controlled study comparing the safety and efficacy of ziprasidone and olanzapine in patients with schizophrenia or schizoaffective disorder needing inpatient care
Methods
Participants	People with schizophrenia.
Interventions	Olanzapine versus ziprasidone.
Outcomes	Not known.
Starting date	Not known.
Contact information
Notes

Hirsch 2000.

Trial name or title	Ziprasidone in the treatment of chronic schizophrenia (maintenance) versus olanzapine ‐ efficacy and side effects
Methods
Participants	People with schizophrenia.
Interventions	Olanzapine versus ziprasidone.
Outcomes	Not known.
Starting date	Not known.
Contact information
Notes

Krakowski 2001.

Trial name or title	Clozapine and olanzapine in violent schizophrenics.
Methods
Participants	People with schizophrenia.
Interventions	Olanzapine vs. Haloperidol vs. Clozapine.
Outcomes	Behaviour Mental State Cognitive function Clinical Global Impression
Starting date	Not known.
Contact information
Notes

Ramamurthy 2000a.

Trial name or title	A multi‐centre, double blind, randomised comparative study of aripiprazole and olanzapine in the treatment of patients with acute schizophrenia.
Methods
Participants	People with schizophrenia.
Interventions	Aripiprazole vs. olanzapine.
Outcomes	Not known.
Starting date	Not known.
Contact information
Notes

Reveley 2000.

Trial name or title	Effectiveness trials of antipsychotic drugs.
Methods
Participants	People with schizophrenia.
Interventions	Olanzapine versus risperidone.
Outcomes	Not known.
Starting date	Not known.
Contact information
Notes

Reynolds 2003.

Trial name or title	Ziprasidone observational study of cardiac outcomes (zodiac) large simple trial: rationale and design.
Methods
Participants	People with schizophrenia.
Interventions	Olanzapine versus ziprasidone.
Outcomes	Not known.
Starting date	Not known.
Contact information
Notes

Singh 2000b.

Trial name or title	A six month international controlled trial of the therapeutic activity of amisulpride 200 to 800mg/day verses olanzapine 5 to 20mg/day in patiients with schizophrenic disorders.
Methods
Participants	People with schizophrenia.
Interventions	Olanzapine versus amisulpride.
Outcomes	Mental state.
Starting date	Not known.
Contact information
Notes

Turner 2002.

Trial name or title	Quality of life ‐ Pfizer trial
Methods
Participants	People with schizophrenia.
Interventions	Olanzapine versus ziprasidone.
Outcomes	Not known.
Starting date	Not known.
Contact information
Notes

Contributions of authors

Lorna Duggan ‐ searching, study selection, data extraction and assimilation, report writing and updating.

Mark Fenton ‐ searching, study selection, data extraction and assimilation, report writing and updating.

Roland Dardennes ‐ protocol development, searching, study selection.

Ahmed El‐Dosoky ‐ protocol development, searching.

Saroja Indran ‐ protocol development, searching.

John Rathbone ‐ study selection, data extraction and assimilation, report writing (update 2004).

Sources of support

Internal sources

• Centre Hospitalier Sainte‐Anne, France.

• Behman Hospital, Egypt.

• University Kebangsaan, Malaysia.

• St Andrew's Hospital, UK.

External sources

• NHS Centre for Reviews and Dissemination, York, UK.

Declarations of interest

Lorna Duggan ‐ has attended functions sponsored by Lundbeck, Janssen, Pfizer, Bristol Myers Squibb and Zeneca and has accepted sponsorship from Eli Lilly for internal flights in the United States.

Mark Fenton ‐ has led Janssen, Lilly and Zeneca sponsored workshops for clinicians.

Ahmed El‐Dosoky ‐ has participated in Eli Lilly sponsored research (Loza 1999 (HGDT)).

John Rathbone ‐ no known conflicts of interest. 
 
 The Cochrane Schizophrenia Group editorial base in Leeds has received general support funding from Eli Lilly during the years 1996‐1999 (see Group Module). This, along with some funds from other pharmaceutical companies, is used to support any ongoing work of the editorial base and is not linked to any particular review (annual report available on request).

Edited (no change to conclusions)