Flaws in animal studies exploring statins and impact on meta‐analysis. (6th May 2014)

• Record Type: Journal Article
• Authors: Moja, Lorenzo; Pecoraro, Valentina; Ciccolallo, Laura; Dall'Olmo, Luigi; Virgili, Gianni; Garattini, Silvio
• Is Part Of: European journal of clinical investigation. Volume 44:Number 6(2014:Jun.)
• Page Start: 597
• Page End: 612
• Publication Date: 2014-05-06
• Subjects: Pathology -- Periodicals; Medical research -- Periodicals; 616.075
• DOI: 10.1111/eci.12264
• Languages: English
• ISSNs: 0014-2972

Abstract

<abstract abstract-type="main" id="eci12264-abs-0001"> <title>Abstract</title> <sec id="eci12264-sec-0001" sec-type="section"> <title>Background</title> <p>Animal experiments should be appropriately designed, correctly analysed and transparently reported to increase their scientific validity and maximise the knowledge gained from each experiment. This systematic review of animal experiments investigating statins evaluates their quality of reporting and methodological aspects as well as their implications for the conduction of meta‐analyses.</p> </sec> <sec id="eci12264-sec-0002" sec-type="section"> <title>Methods</title> <p>We searched <sc>medline</sc> and <sc>embase</sc> for studies reporting research on statins in mice, rats and rabbits. We collected detailed information about the characteristics of studies, animals and experimental methods.</p> </sec> <sec id="eci12264-sec-0003" sec-type="section"> <title>Results</title> <p>We retrieved 161 studies. A little over half did not report randomisation (55%) and most did not describe blinding (88%). All studies reported details on the experimental procedure, although many omitted information about animal gender, age or weight. Four percent did not report the number of animals used. None reported the sample size. Fixed‐ and random‐effects models gave different results (ratio of effect size increased by five folds). Heterogeneity was consistently substantial within animal models, for which accounting for covariates had minimal impact. Publication bias is highly suspected across studies.</p> </sec> <sec id="eci12264-sec-0004" sec-type="section"> <title>Conclusions</title> <p>Although statins showed efficacy in animal models, preclinical studies highlighted fundamental problems in the way in which such research is conducted and reported. Results were often difficult to interpret and reproduce. Different meta‐analytic approaches were highly inconsistent: a reliable approach to estimate the true parameter was imperceptible. Policies that address these issues are required from investigators, editors and institutions that care about the quality standards and ethics of animal research.</p> </sec> </abstract>