Impact of prenatal stress on offspring glucocorticoid levels: A phylogenetic meta-analysis across 14 vertebrate species

Abstract

Prenatal exposure to maternal stress is commonly associated with variation in Hypothalamic Pituitary Adrenal (HPA)-axis functioning in offspring. However, the strength or consistency of this response has never been empirically evaluated across vertebrate species. Here we meta-analyzed 114 results from 39 studies across 14 vertebrate species using Bayesian phylogenetic mixed-effects models. We found a positive overall effect of prenatal stress on offspring glucocorticoids (d' = 0.43) though the 95% Highest Posterior Density Interval overlapped with 0 (-0.16-0.95). Meta-regressions of potential moderators highlighted that phylogeny and life history variables predicted relatively little variation in effect size. Experimental studies (d' = 0.64) produced stronger effects than observational ones (d' = -0.01), while prenatal stress affected glucocorticoid recovery following offspring stress exposure more strongly (d' = 0.75) than baseline levels (d' = 0.48) or glucocorticoid peak response (d' = 0.36). These findings are consistent with the argument that HPA-axis sensitivity to prenatal stress is evolutionarily ancient and occurs regardless of a species' overall life history strategy. These effects may therefore be especially important for mediating intra-specific life-history variation. In addition, these findings suggest that animal models of prenatal HPA-axis programming may be appropriate for studying similar effects in humans.

Figure 1

Phylogeny of study species with mean effect sizes for each species indicated by color and ranges given in parantheses. Internal branches are color-coded to reflect ancestral states inferred by maximum likelihood. 100my = 100 million years.

Figure 2

Overall weighted effect size and 95% Highest Posterior Density Intervals and the independent influence of each moderator variable in explaining variation in effect size. Experimental studies and those measuring glucocorticoid recovery stand out as producing higher effect sizes. Width of lines and size of points are proportional to the number of effect sizes in each category.

Figure 3

Funnel plots on (a) the raw effect sizes and (b) the residuals from the full model. Solid vertical line indicates zero and dashed vertical line indicates overall weighted effect size from the intercept-only model. Asymmetry in the funnel plot, specifically missing null or negative effects at low precision (lower left corner of the funnel), indicate publication bias. Egger’s regression on the residuals (b) revealed a trend towards publication bias, and the trim and fill method imputed 14 missing studies (brown points) in the lower left corner.