Protocol for a Systematic Review and Meta-Analysis of Data from Preclinical Studies Employing Forced Swimming Test: an Update
Total Page:16
File Type:pdf, Size:1020Kb
BMJ Open Science: first published as 10.1136/bmjos-2017-000043 on 31 May 2019. Downloaded from Open access Protocol Protocol for a systematic review and meta-analysis of data from preclinical studies employing forced swimming test: an update A B Ramos-Hryb,1 Z Bahor,2 S McCann,2 E Sena,2 M R MacLeod,2 C Lino de Oliveira1 This article has received a OSF ABSTRACT Strengths and limitations of this study badge for Open data. Objective Forced swimming test (FST) in rodents is a widely used behavioural test for screening antidepressants To cite: Ramos-Hryb AB, ► This protocol for systematic review will collect, with in preclinical research. Translational value of preclinical Bahor Z, McCann S, et al. broad inclusion criteria, preclinical studies employ- studies may be improved by appraisal of the quality of Protocol for a systematic review ing forced swimming test (FST). and meta-analysis of data from experimental design and risk of biases, which remains to ► The present protocol has been preregistered with preclinical studies employing be addressed for FST. The present protocol of a systematic Open Science Framework. forced swimming test: an review with meta-analysis aims to investigate the quality ► A preliminary version of the present protocol has update. BMJ Open Science of preclinical studies employing FST to identify risks of 2019;3:e000035. doi:10.1136/ been preregistered with Systematic Review Facility bias in future publications. In addition, this protocol will bmjos-2017-000043 (Collaborative Approach to Meta-Analysis and help to determine the effect sizes (ES) for primary and Review of Animal Data from Experimental Studies). ► Provenance and Open peer secondary outcomes according to several aspects of the ► Results obtained with this systematic review and review Prepublication and FST study design. Review History is available meta-analysis may help to create specific and ratio- Search strategy, Screening annotation, Data nal methodological guidelines for application of FST at http:// dx. doi. org/ 10. 1136/ management Publications reporting studies testing bmjos- 2017- 000043. in rodents. different classes of antidepressants in FST will be collected ► High levels of heterogeneity between studies may Received 30 December 2017 from Medline, SCOPUS and Web of Science databases. limit the external validity of our results. Revised 24 November 2018 A broad list of inclusion criteria will be applied excluding ► The summary effect size may be overestimated by http://openscience.bmj.com/ Accepted 1 February 2019 those studies whereby FST is used as a stressor or studies publication bias. reporting data from co-treatments. For assessing the quality of the included publications, the quality checklist adapted by Collaborative Approach to Meta-Analysis and INTRODUCTION Review of Animal Data from Experimental Studies will Major depression disorder (MDD) in humans be used. If the meta-analysis seems feasible, the ES and is characterised by depressed mood and the 95% CI will be analysed. The heterogeneity between behavioural inhibition and often comes with studies will be assessed by using the χ2statistic with n−1 degrees of freedom. Subgroup meta-analysis (meta- social avoidance, generalised anxiety, eating regression, and if necessary, stratified regression) will be disorders, sleeping and problems to cope with 1 on September 30, 2021 by guest. Protected copyright. performed when possible according to characteristics of stress. Despite the difficulty in finding suit- © Author(s) (or their study design and study quality to assess their impact on able models to mimic subjective, behavioural employer(s)) 2019. Re-use efficacy of the treatments. In addition, funnel plotting, Egger and neurobiological aspects of MDD, there permitted under CC BY. Published by BMJ. regression, and ‘trim and fill’ will be used to assess the are several animal models predictive of MDD risk of publication bias. Results of this protocol will help to 2 3 1Physiological Sciences treatment. Most of these animal models are Deptartment, Biological create rational methodological guidelines for application based on behavioural responses of an animal Sciences Center, Federal of FST in rodents and improve the quality and translational to inescapable stress, providing a framework University of Santa Catarina, value of preclinical research on antidepressant discovery. for several laboratory tests.4–7 Usually, ines- Florianópolis, Brazil Reporting A preliminary version of the present protocol capable stress induces behavioural inhibition 2Centre for Clinical Brain has been preregistered with Systematic Review Facility (or immobility) that can be counteracted Sciences, University of (http:// syrf. org. uk/). A preprint version of the current 4–7 Edinburgh, Edinburgh, UK protocol has been registered with Open Science by antidepressant treatment. Therefore, Framework (https:// osf. io/ 9kxm4/). Results will be behavioural tests in animals are employed as Correspondence to communicated in scientific meetings and peer-reviewed screening steps during the preclinical phase C Lino de Oliveira, Dept de journals. We plan to conduct an anonymous and online of antidepressant drug discovery. Forced Ciências Fsisiológicas, Centro 4 de Ciências Biológicas, survey within the scientific community to ask researchers swimming test (FST) in rats and mice is Universidade Federal de Santa about their perception of risk of bias and their experience used in preclinical trials of antidepressants. Catarina; cilene. lino@ ufsc. br with the publication of negative results. FST is easy to run, inexpensive, sensitive and Ramos-Hryb AB, et al. BMJ Open Science 2019;3:e000035. doi:10.1136/bmjos-2017-000043 1 BMJ Open Science: first published as 10.1136/bmjos-2017-000043 on 31 May 2019. Downloaded from Open access relatively selective to known antidepressants (for review experimental groups generating significant and non-sig- see Cryan et al2). One criticism that may apply to FST is nificant results in a single experiment. The incidences of the abundance of ‘positive results'8 that contrasts with significant results for primary and secondary outcomes the failure of antidepressant treatments in some clinical within the experiments were 88.2% and 84.6%, respec- trials or therapeutics.9–11 There is an estimation that up tively. Interestingly, the experiments also showing non-sig- to 50% of patients are resistant to the treatment with the nificant results for primary outcomes were only 29.4% antidepressants currently available.9 10 Many different whereas most of the experiments (92.3%) also reported reasons may account for the contrasting findings between non-significant results for secondary outcomes. The high preclinical and clinical data11 including individual vari- number of significant results as compared with the nega- ability, poor quality of the studies as well as publication tive ones was also found in another study.8 In summary, bias. Publication bias in a preclinical field may inflate the these preliminary analyses indicated that quality scores estimated effect size (ES)12 13 leading to inflated expec- will be independent of the publication date, the experi- tations of efficacy in clinical trials, which may explain mental species as well as the type of antidepressant tested. partially the perceived contrast between fields.11 There- The differences in sex, strains and ages may be a source fore, the aim of the present study is to evaluate the quality of heterogeneity. In addition, it is expected that ‘random of published literature applying FST to detect effects of allocation to a treatment’, ‘concealment of treatment the treatment with antidepressants and the risk of bias in allocation’ and ‘sample size calculation’ will be neglected this research field. in this field of research. Moreover, these interim data Initially, a pilot study was performed to create a suggest the existence of publication bias. Considering database and to standardise the methods for a system- the sample used in the pilot study as the representative atic review.14 15 This pilot study started with a review in sample, the screening process may generate enough data Medline and Embase retrieving more than 7000 publica- (50% of publications in the database will fit inclusion tions by using expressions commonly found in the liter- criteria, ie, 2200 publications) for reliable estimation of ature such as ‘forced swimming test’ OR ‘forced swim the quality of the studies, ES, heterogeneity and publica- test’ OR ‘Porsolt test' OR ‘fst’. The combination of these tion bias in the field. with medical subject heading (MeSH) terms related to A preliminary version of this protocol was deposited ‘rodents'16 and ‘antidepressants'17 retrieved the publica- in the Systematic Review Facility in February 2016.14 15 tions more relevant for the present study. For screening The current version of the protocol was updated based on purposes, a database containing bibliographical informa- procedures available in CAMARADES following instruc- tion from retrieved publications was built. We applied tions by de Vries et al.20 A preprint version of the current inclusion and exclusion criteria in the screening steps. text was preregistered with Open Science Framework ( Forty references, randomly selected from the database, osf. io/ 9kxm4). Meta-analysis will be performed after http://openscience.bmj.com/ generated 20 references to the pilot study, that is, one publication of the protocol in a journal with a peer-review reference in every two fitted the inclusion