Rodent chronic variable stress procedures: A disjunction between stress entity and impact on behaviour.

Chronic variable stress (CVS) procedures are widely used to model depression in laboratory mice and rats. In order to explore how study design might impact experimental outcomes, we systematically documented characteristics of study design in a series of published rodent CVS studies and, in a subset of studies, measured effect sizes in the behavioural tests used to evaluate the effects of CVS. We hypothesised that CVS procedures that were longer or involved more stressors would be associated with larger effect sizes in five commonly used behavioural tests: the sucrose preference test (SPT), the tail suspension test (TST), the forced swim test (FST), the open field test (OFT) and the elevated plus maze (EPM). We also hypothesised that effect sizes would positively correlate between the behavioural tests that are believed to measure the same consequences of CVS. We searched PubMed for articles using CVS procedures with mice or rats and systematically documented the duration (the length of the CVS procedure), burden (the total number of stressors experienced by the animal) and diversity (the total number of different types of stressors used) of the CVS procedures used. We also systematically documented the design of the behavioural tests used to evaluate the effects of CVS in each study and calculated the effect sizes obtained in these tests. To ask whether effect sizes in these tests correlated with characteristics of the CVS procedure used, we used a linear model of the effect of duration, burden, and diversity on the effect size, then calculated the Euclidean distance between studies' characteristics and correlated those with the differences in effect size between studies. To explore whether effect sizes correlated between different behavioural tests, we calculated a pairwise Pearson correlation. We observed that most studies used a unique CVS procedure. In contrast to our hypothesis, the most evident impact of CVS procedure design was on FST effect sizes, where longer-duration CVS procedures with more diverse types of stressors were associated with a smaller effect size in behavioural tests. When exploring correlations between behavioural test effect sizes, we found a positive correlation between effect sizes in the TST and FST, and in the OFT and EPM, but the strongest positive correlations were between the EPM and TST, and between the EPM and FST. These data uncover complex relationships that are not necessarily in concordance with current understanding of what these tests measure. Accordingly, our data raise scientific questions around the design of CVS procedures used and the behavioural tests used to evaluate them.