What are the impacts of in utero ACS administration on brain function, the underlying mechanisms, and the HPA axis in nonhuman mammalian species?

An initial search of PubMed and PROSPERO (International Prospective Register of Systematic Reviews) was undertaken to identify articles on the topic and to assess knowledge gaps. A full search strategy was developed using indexed terms and text words within the titles and abstract (Table 1). This search string was then modified to include syntax specific to each individual database. This search string was validated by checking for the inclusion of a set of 11 key references reporting outcomes spanning brain function, determinants, and HPA axis measures in the output of each search [25-35]. We searched the following electronic databases for published evidence: PubMed, Embase, MEDLINE (Ovid), Web of Science, Scopus, and ProQuest. Searches will be updated in June-July 2026 to identify and include all sources published or accepted for publication to the end of 2025. Publication is planned for 2027.

Database search results will be imported to Covidence (Covidence systematic review software, Veritas Health Innovation), where reviewers can collaboratively screen sources for inclusion. Two assessors (any two of the review authors) will independently review sources using eligibility criteria (Table 2) at title, abstract, and full-text levels. Assessors will be blinded during screening. Uncertainties or conflicting decisions at each stage will be resolved through discussion and consensus, with at least one other member of the review team.

Data extraction will be performed by two independent reviewers using author-designed fields in Covidence that will reflect the standardized JBI data extraction [36]. Uncertainties or conflicting decisions at each stage will be resolved through discussion and consensus with at least one other member of the review team. Study authors will be contacted to request any missing or additional data. The extracted data will include citation information, species, strain, and normal term gestation of experimental animals used and information regarding study design including sample size (including number of mothers and number of offspring), method of allocation to treatments, timing of delivery, and assessment blinding. Extracted data relating to the ACS exposure will include the type of corticosteroid used, the dose including the formulation and route of administration, and the timing and duration of administration expressed in days of gestation and as a fraction of term. Extracted data related to outcomes will include sample size, sex, and postnatal age at assessment, details of the functional outcomes or brain regions analyzed, and information regarding the methodology used to measure outcomes. Where possible, the means, standard deviations, and percentage differences will be extracted for each treatment group and outcome. Where outcomes are only presented in figures, the numeric data (means and variance) will be extracted using WebPlotDigitizer v5 (Automeris LLC). Where further data or clarification are required for a study, the authors will be contacted via email. A minimum of three attempts at monthly intervals will be made to contact authors. Following extraction, we will cross-reference author names, experimental design, and study dates to identify multiple publications originating from the same experiment to prevent redundant data extractions arising from multiple publications. Where studies report overlapping data, such as multiple reports of the same outcome from an experimental cohort, only data from the largest study or more relevant exposures will be included. Where the same outcomes are reported in identical cohorts, the earliest publication will be retained. In studies where multiple treatment groups are compared against a single shared control group, the control group sample size will be divided equally across the comparison arms.

Methodological quality will be assessed through the use of the Systematic Review Center for Laboratory Animal Experimentation Risk of Bias (SYRCLE RoB) tool [37]. Critical appraisal will be conducted by two independent reviewers, and judgment disagreements will be resolved through discussion and consensus or a third review if necessary. Appraisal results will be reported in table format.

A narrative synthesis will be presented following standard guidelines. A dose-response meta-analysis will be performed where at least 3 studies report the same outcome following in utero exposure to the same corticosteroid with comparable timing and data are made available by authors. The litter will be treated as the primary unit of analysis. For studies that report data from multiple offspring per litter without adjustment, the litter mean will be used as a single data point for each measured outcome. To ensure that these dose-response analyses are biologically meaningful across different mammalian species, doses will be standardized using allometric scaling (based on body surface area) to calculate human-equivalent dosing where possible. Where sufficient data are available, subgroup analyses will be performed within different species and biological sexes. Studies with missing data where authors do not respond to data requests will be included in the narrative synthesis but will be excluded from meta-analyses. Studies that report mean values for groups without measures of variance will be excluded from the meta-analysis but will be included in the narrative synthesis. For instance, where studies in the functional review report overlapping cohorts or outcomes with the underlying mechanisms and stress response reviews, we will cross-reference these findings to highlight consistent patterns of association across the different domains.

For any outcomes where meta-analysis is able to be performed, the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach will be used to grade the certainty of evidence [38]. Following this, a Summary of Findings will be created using GRADEpro GDT (McMaster University). Independent, blinded assessment of the certainty of evidence will be carried out by two reviewers, with any potential disagreements resolved through discussion or the involvement of a third reviewer. The Summary of Findings will present, where appropriate, mean differences, a ranking of the quality of the evidence based on the risk of bias, heterogeneity, precision, and risk of publication bias of the review results.

The results of this systematic review will be disseminated through peer-reviewed publications.

We hypothesize that this systematic review will reveal a dose-dependent relationship between ACS exposure and alterations in neurodevelopmental outcomes. Specifically, ACS exposure may be associated with increased anxiety-like behaviors and deficits in spatial learning and memory across various mammalian species. We expect to find evidence of reduced neuronal density and altered myelination patterns in brain regions relevant to behavior and altered HPA axis sensitivity. We hypothesize that the impacts of ACS will differ by biological sex, potentially suggestive of greater vulnerability in one sex regarding cognitive tasks or stress-reactivity. Finally, we anticipate that, while basic mechanisms of glucocorticoid action are conserved across species, the timing of exposure relative to brain maturation may contribute to differences in impacts in different species.

Previous scoping and systematic reviews of the effects of ACS on neurodevelopment and stress responses have primarily focused on clinical outcomes in human cohorts [12,15,18]. The systematic or scoping reviews describing these outcomes in nonhuman species are of narrower scope, do not fully reflect the clinical context, and/or require updating [20,23,24]. Unlike clinical studies that may limited by confounding environmental variables, synthesis of preclinical evidence allows for a more controlled evaluation of dose-response relationships and tissue-specific molecular changes.

This systematic review will have several limitations. Given that we are excluding publications not in English at screening, there is a possibility that we will miss relevant studies published in other languages, although these will be listed in the exclusion reasons for transparency [39]. It is not clear whether meta-analyses of outcomes will be possible given the heterogeneity of timings of steroid administration and outcome evaluation and the variety of species, antenatal steroid regimens, and outcome assessments evident in previous reviews [20]. Effect sizes may also be overestimated due to the potential nonpublication of null results in preclinical research [40].