Date created: | Last Updated:

Category: Analysis

Description: Review question To examine the efficacy, acceptability and tolerability of both first- and second-generation antipsychotic drugs for treatment-resistant schizophrenia patients of all ages by applying a network meta-analysis approach. Searches 1. Electronic databases: We will search the Cochrane Schizophrenia Group’s Study-Based Register of Trials with no date/time, language, document type, and publication status limitations. This register is compiled of regular searches in multiple electronic databases, ClinicalTrials.gov, WHO register of clinical trials and more. Details on the register can be found in (Shokraneh and Adams 2017, 2019, 2020, 2021). 2. Hand searching: The Cochrane Schizophrenia Group’s Study-Based Register of Trials also includes hand searches. We will additionally inspect references in previous reviews on treatment-resistant schizophrenia, as well as previous reviews from our team[1-3]. 3. Personal contact: We will contact via email the first and/or corresponding authors of each included study published in the last 20 years for missing information about their studies. Types of study to be included We will include all randomized trials (RCTs) comparing one antipsychotic drug with another antipsychotic agent or placebo in treatment-resistant schizophrenia, but open and single-blind RCTs will be excluded in a sensitivity analysis (see below the Analysis of subgroups or subsets). Trials in which antipsychotic drugs were used as an augmentation- or combination strategy will be excluded as well as studies with a high risk of bias in the randomization process, and studies from mainland China[4]. In the case of cross-over studies we will only use only the first cross-over phase to avoid the problem of carry-over effects which are very likely in schizophrenia. We will include cluster-RCTs applying the appropriate correction of the unit of analysis, or if such a correction can be applied post-hoc. But we anticipate it could be a very rare case. The minimum duration of trials will be 3 weeks. Condition or domain being studied Schizophrenia Participants/population We will include participants with a treatment-resistant form of schizophrenia, schizophreniform disorder, or schizoaffective disorder (no age limit, any definition of treatment resistance, no restriction in diagnostic criteria, setting, gender, and ethnicity). Intervention(s), exposure(s) We will include all antipsychotic drugs that are available in at least one country, including first-generation (“typical”, “conventional”) antipsychotics as well as second-generation (“atypical”) antipsychotics and placebo. We will include all these compounds at any dose and in any oral forms of administration (for example tablets, liquid) or as intramuscular depot formulations. If an antipsychotic is available in both oral and depot forms, both formulations will be used as separate interventions in the network. Comparator(s)/control In a network meta-analysis, any treatment can be a comparator as all antipsychotics and placebo will be used to compare with each other. Placebo will be used as reference for presentation. Context There are no restrictions in terms of setting, for example, we will include in- and outpatients. Main outcome(s) The primary outcome will be the overall symptoms of schizophrenia as measured by rating scales such as the Positive and Negative Syndrome Scale (PANSS)[5], the Brief Psychiatric Rating Scale (BPRS)[6] or of any other validated scale (e.g. the Manchester Scale[7]) for the assessment of overall schizophrenic symptomatology. As not all studies will have used the same scale, we will apply the following hierarchy: mean change of the PANSS total score from baseline to endpoint, if not available mean change of the BPRS, or if again not available the mean values at endpoint of the PANSS/ BPRS. The results of other rating scales will only be used if the instrument has been published in a peer-reviewed journal, because it has been shown that unvalidated schizophrenia scales exaggerate differences[8]. Measures of effect The effect size for the primary outcome will be the standardized mean difference (SMD), presented with their 95% CIs, because we expect different rating scales of schizophrenia symptomatology can be used in studies. We will give preference to the estimates based on imputation methods to handle missing data (used by the original authors) over completers’ data. Additional outcome(s) 1. Response to treatment. The following hierarchy of the response definitions will be applied: at least 20% reduction of the baseline score of the PANSS, 20% reduction of the BPRS or 20% reduction of any other global schizophrenia rating scale, at least “minimally improved” (score of 3) on the Clinical Global-Impressions-Improvement Scale (CGI). We choose this cutoff because even minimal improvement can be clinically important for treatment-resistant patients[8]. If none of these definitions is available, we will use the original authors’ primary definition. 2. Positive symptoms, measured by published rating scales. 3. Negative symptoms, measured by published rating scales. 4. Dropout due to any reason. Premature discontinuation due to any reason combines efficacy, tolerability, and other factors and can therefore be considered as a measure of “overall acceptability of treatment”. 5. Dropout due to specific reasons. Dropout due to inefficacy of treatment will be considered as an additional outcome of the efficacy of treatment. Dropout due to the occurrence of adverse events will be used as a measure of overall tolerability. 6. Specific adverse events. We will focus on five most common adverse events, use of antiparkinson medication (dichotomous outcome), weight gain (kg, continuous outcome), sedation (dichotomous outcome), prolactin levels (ng/mL, continuous outcome) and QTc prolongation (ms, continuous outcome). 7. Quality of life. We will accept any published rating scale such as Heinrichs quality of life scale, Quality of Life Scale (QOLS), or any other published rating scale. 8. Functioning. Functioning will be measured by rating scales such as the Global Assessment of Functioning, the Psychosocial Performance Scale, or any other published rating scale. Measures of effect The effect size for continuous outcomes will be the standardized mean difference (SMD), presented with their 95% CIs, since we expect different rating scales of schizophrenia symptomatology, quality of life and functioning can be used in studies. Nevertheless, we will use mean differences (MD) for weight gain (kg), prolactin levels (ng/ml) and QTc prolongation (ms), since we can convert values of these outcomes into the same metric. We will give preference to the estimates based on imputation methods to handle missing data (used by the original authors) over completers’ data. The effect size for dichotomous outcomes will be the odds ratio (OR) and its 95% confidence intervals (CIs), because odds ratio has better mathematical properties. But we will convert back to relative risks (RRs) and percentages in treatment and control groups for presentation of the results. Data extraction (selection and coding) 1. Selection of trials: Two reviewers will independently inspect all abstracts identified in the searches. Disagreement will be resolved by discussion, and where doubt still remains, we will acquire the full article for further inspection. Once the full articles are obtained, at least two reviewers will independently decide whether the studies meet the review criteria. If disagreement cannot be resolved by discussion, we will resolve it with a third reviewer or seek further information from the study authors. 2. Data extraction: At least two reviewers will independently extract data from all selected trials on specifically customized digital forms in the Microsoft Access database. Disagreement will be resolved by discussion with a third reviewer or by contacting the study authors. Risk of bias (quality) assessment Two independent reviewers will assess risk of bias of individual studies using the Cochrane Risk of Bias tool, RoB 2.0[9]. Strategy for data synthesis 1. Two-step procedure. In a first step we will perform series of conventional pair-wise meta-analyses by combining studies that compared the same interventions. In a second step we will then perform network meta-analysis within a frequentist framework[10]. The analysis and presentation of results will be performed using R (meta and netmeta packages). 2. The heterogeneity (variability in relative treatment effects within the same treatment comparison) will be measured with the tau-squared (the variance of the random effects distribution). The heterogeneity variance will be assumed common across the various treatment comparisons and we will compare the empirical distribution with predictive distributions[11, 12]. Potential reasons for heterogeneity will be explored by subgroup analysis (see below the Analysis of subgroups or subsets). 3. Assessment of the transitivity. Intransitive networks can lead to misleading estimates. Therefore, we will assess the transitivity assumption by investigating the distribution of clinical and methodological variables that can act as effect modifiers across treatment comparisons. Potential effect modifiers are listed under “Analysis of subgroups or subsets “below. We will investigate if these variables are similarly distributed across studies grouped by comparison. 4. Network meta-analysis. We will conduct a random effects network meta-analysis to synthesize all evidence for each outcome, and obtain a comprehensive ranking of all treatments. Treatments will be ranked for each outcome using P-scores[13]. A key assumption is coherence, meaning the agreement between direct and indirect evidence. This will be assessed locally for each closed loop using the separating indirect from direct evidence approach, and globally for the whole network using a design-by-treatment interaction model. In case incoherence, we will investigate possible sources of it (mistakes in data entry, clear differences in study characteristics) and utilize analytical approaches [14]. If the requirements of network meta-analysis are not met (low likelihood of transitivity and/or large unexplained inconsistency) we will use pairwise meta-analysis for data synthesis. 5. Assessment of the confidence in the evidence. The confidence in the relative treatment effects for the primary outcome will be evaluated using the Confidence in Network Meta-Analysis framework[15, 16], implemented in the web application (http://cinema.ispm.ch/model/CINeMA_paper.pdf). This tool evaluates the credibility of the findings across the domains of within-study bias, across-study bias, indirectness, imprecision, heterogeneity and incoherence. Analysis of subgroups or subsets The following potential effect moderators of the primary outcome will be explored by subgroup analyses: 1. The criteria of treatment-resistant definitions 2. Mean participant age 3. Dose of the antipsychotics in chlorpromazine-equivalents according to Gardner et al. [17] 4. Publication date (to address the effect of possibly generally decreasing effect sizes over time) 5. Severity of illness at baseline (PANSS or BPRS score at baseline) 6. Study duration Sensitivity analyses will be performed as follow: 1. Exclusion of non-double-blind studies (open and single-blind studies) 2. Exclusion of studies that presented only completer analyses 3. Exclusion of studies that did not use operationalized criteria to diagnose schizophrenia 4. Exclusion of studies with an overall assessment of high risk of bias 5. Exclusion of studies only included children and/or adolescents Contact details for further information Stefan Leucht stefan.leucht@tum.de Organisational affiliation of the review Department of Psychiatry and Psychotherapy, School of Medicine, Technical University of Munich, Munich, Germany https://www.tum.de/ Review team members and their organisational affiliations Ms Shimeng Dong. Department of Psychiatry and Psychotherapy, School of Medicine, Technical University of Munich, Munich, Germany Dr Johannes Schneider-Thoma. Department of Psychiatry and Psychotherapy, School of Medicine, Technical University of Munich, Munich, Germany Mr Spyridon Siafis. Department of Psychiatry and Psychotherapy, School of Medicine, Technical University of Munich, Munich, Germany Mr Dongfang Wang. Department of Psychiatry and Psychotherapy, School of Medicine, Technical University of Munich, Munich, Germany Professor Stefan Leucht. Department of Psychiatry and Psychotherapy, School of Medicine, Technical University of Munich, Munich, Germany Type and method of review Network meta-analysis Anticipated or actual start date 01 June 2021 Anticipated completion date 31 June 2022 Funding sources/sponsors China Scholarship Council (CSC) Language English Country Germany References 1. Huhn, M., et al., Comparative efficacy and tolerability of 32 oral antipsychotics for the acute treatment of adults with multi-episode schizophrenia: a systematic review and network meta-analysis. Lancet, 2019. 394(10202): p. 939-951. 2. Samara, M.T., et al., Efficacy, Acceptability, and Tolerability of Antipsychotics in Treatment-Resistant Schizophrenia: A Network Meta-analysis. JAMA Psychiatry, 2016. 73(3): p. 199-210. 3. Krause, M., et al., Efficacy, acceptability, and tolerability of antipsychotics in children and adolescents with schizophrenia: A network meta-analysis. Eur Neuropsychopharmacol, 2018. 28(6): p. 659-674. 4. Wu, T., et al., Randomized trials published in some Chinese journals: how many are randomized? Trials, 2009. 10: p. 46. 5. Kay, S.R., A. Fiszbein, and L.A. Opler, The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr Bull, 1987. 13(2): p. 261-76. 6. Overall, J.E. and D.R. Gorham, The Brief Psychiatric Rating Scale. Psychological Reports, 1962. 10(3): p. 799-812. 7. Krawiecka, M., D. Goldberg, and M. Vaughan, A standardized psychiatric assessment scale for rating chronic psychotic patients. Acta Psychiatr Scand, 1977. 55(4): p. 299-308. 8. Marshall, M., et al., Unpublished rating scales: a major source of bias in randomised controlled trials of treatments for schizophrenia. Br J Psychiatry, 2000. 176: p. 249-52. 9. Sterne, J.A.C., et al., RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ, 2019. 366: p. l4898. 10. Rücker, G., Network meta-analysis, electrical networks and graph theory. Res Synth Methods, 2012. 3(4): p. 312-24. 11. Turner, R.M., et al., Predicting the extent of heterogeneity in meta-analysis, using empirical data from the Cochrane Database of Systematic Reviews. Int J Epidemiol, 2012. 41(3): p. 818-27. 12. Rhodes, K.M., R.M. Turner, and J.P. Higgins, Predictive distributions were developed for the extent of heterogeneity in meta-analyses of continuous outcome data. J Clin Epidemiol, 2015. 68(1): p. 52-60. 13. Rücker, G. and G. Schwarzer, Ranking treatments in frequentist network meta-analysis works without resampling methods. BMC Med Res Methodol, 2015. 15: p. 58. 14. Salanti, G., Indirect and mixed-treatment comparison, network, or multiple-treatments meta-analysis: many names, many benefits, many concerns for the next generation evidence synthesis tool. Res Synth Methods, 2012. 3(2): p. 80-97. 15. Nikolakopoulou, A., et al., CINeMA: An approach for assessing confidence in the results of a network meta-analysis. PLoS Med, 2020. 17(4): p. e1003082. 16. Salanti, G., et al., Evaluating the quality of evidence from a network meta-analysis. PLoS One, 2014. 9(7): p. e99682. 17. Gardner, D.M., et al., International consensus study of antipsychotic dosing. Am J Psychiatry, 2010. 167(6): p. 686-93.